Professional Documents
Culture Documents
Many will agree that sampling and sample preparation are key parts of the ana-
lytical process. The reliability of analysis is based on the sampling process, stor-
age and preservation of samples, isolation of the analytes, clean-up and final
determination. In all these operations, sampling and sample preparation still
determine the overall analysis time and are the real bottleneck. This book on
Sampling and Sample PreparationTechniques for Fieldand Laboratory, edited
by J. Pawliszyn, addresses the above. The book contains 33 chapters covering
sampling, fundamentals and applications of extraction techniques, and new
materials. Particular attention has been given to the unification of the theory of
extraction, including solid phase micro-extraction, which became very popular
in the late 1990s because of its simplicity. Instrumentation aspects and integra-
tion with other techniques such as chromatography, capillary electrophoresis
and mass spectrometry are covered, followed by a variety of applications in envi-
ronmental, clinical, food and pharmaceutical analysis.
I am glad to introduce this new book in the Comprehensive Analytical Chem-
istry series, since it is the first in the series that comprehensively covers the topic
of sampling and sample preparation. This book is a good example of the "teach-
ing the teachers" approach by introducing sampling and sample preparation
methodologies in a step-by-step way, and at the same time being useful for
undergraduates and graduates as well as practising analytical chemists.
Finally, I am extremely grateful to everyone who has contributed to this
book. In particular, I would especially like to thank Janus Pawliszyn for doing a
great job as Editor. I am sure that this book will help to establish the science of
sampling and sample preparation with all its associated technologies in the field
of analytical chemistry.
D. Barcel
xxxi
Preface
Sampling and sample preparation are integral parts of the analytical process.
They should be part of the analytical chemistry teaching curriculum, but often
are not even mentioned during graduate or undergraduate analytical courses.
The primary reason for this situation is that sample preparation is not consid-
ered a separate part of analytical science with its unique challenges, but rather
the "thing" you do when you develop and perform analytical methods.
The main difficulty in recognizing the scientific principles of sample prepara-
tion is that the fundamentals of extraction, involving natural and frequently
complex samples are much less developed and understood, compared to physico-
chemically simpler systems used in the separation and quantification steps of
the analytical process, such as chromatography and mass spectrometry. This sit-
uation creates an impression that rational design and optimization of extraction
for complex systems is not possible. Therefore, the development of sample prep-
aration procedures is frequently considered to be an "art" but not a "science".
Relative to its significance in the overall success of analysis, there has been
little scientific interest in sample preparation. Little scientific interest in sample
preparation has also been associated with the fact that the technology has been
based, until quite recently, on very simple, "low tech", approaches such as
sample-solvent or sample-headspace partitioning, while underlying more scien-
tifically challenging problems associated with the sample matrix have been
ignored. This situation is currently changing with the introduction of non-
traditional technologies, which address the need for solvent-free alternatives,
automation and miniaturization. These approaches are frequently simpler to
operate, but more difficult to optimize, requiring more fundamental knowledge
by analytical chemists not only about equilibrium conditions, but, more impor-
tantly, about the kinetics of mass transfer in the systems. For some years, I have
been actively involved in teaching the fundamental aspects of modern sample
preparation technology to practitioners of analytical chemistry, mainly indus-
trial chemists. I recognized a need to provide the fundamental background, not
only to assist users, but also to help educators in developing their undergraduate
and graduate programs. Designing teaching programs to address the new devel-
opments in extraction technologies is quite difficult considering that the scien-
tific literature's emphasis is placed on the differences between techniques rather
than on their common features, which would facilitate general understanding.
The present trend in analytical instrumentation is towards miniaturization,
which results in portability. This development will eventually enable the attain-
ment of a major goal of the analytical chemist: to perform analysis at the place a
xxxiii
sample is located, rather than moving the sample to a laboratory, as is currently
the common practice in many cases. The new approach reduces errors and the
time associated with sample transport and storage, resulting in more accurate,
precise and faster analytical data. Simplification of sample preparation technolo-
gies and their integration with sampling and the convenient introduction of
extracted components to analytical instrumentation are significant challenges
to, and opportunities for, the contemporary analytical chemist. The design of
easy-to-use, but powerful, sample preparation tools will facilitate the convenient
implementation of integrated analytical technologies. The results of current
research will have a profound effect on the future of analytical methodology.
The purpose of this book is to address the needs and challenges outlined
above. To facilitate recognition of sample preparation as a separate science with
its unique challenges and research opportunities, the primary focus of the book
is on the fundamental aspects of extraction technology. Leading scientists in this
area have contributed chapters on modern aspects of liquid, solid phase and
membrane extractions with and without derivatization step as well as the chal-
lenges associated with different types of matrices. In the chapter on Unified The-
ory of Extraction, an attempt has been made to outline common features among
extraction technologies. The application chapters are dedicated to different
types of matrices, and focus on the impact of new technologies on the practice of
the science of sample preparation. There are some contributions that are con-
cerned with the characteristics of sample matrices, and their effects on method
development. Also, there are separate chapters focusing on sampling strategies
and equipment; many authors discuss this topic within their own contributions,
emphasizing the fact that extraction technologies should not be considered in
isolation but should be well integrated with the sampling and the introduction to
analytical instrument steps. This is particularly important when implementing
analytical technology directly on-site. Finally, there are a some chapters that
focus on new extraction materials, which could facilitate further improvement in
the technology.
The book is not intended to provide a comprehensive review of the topic of
sample preparation, but rather to be a first step towards a unified treatment of
analytical sample preparation technologies. It is hoped that it will be helpful for
learning more about sample preparation, as well as for encouraging an interest
in it by outlining the present practice of the technology and by indicating
research opportunities.
The book was developed during my scientific visit to Germany sponsored by
the Alexander von Humboldt Foundation award. I would like to thank my Ger-
man hosts, Karsten Levsen from Fraunhofer ITA, Hanover, and Peter Popp,
UFZ, Leipzig, for their hospitality, patience and encouragement during this
project.
Janusz Pawliszyn
Waterloo, Ontario
xxxiv
Chapter 1
The recorded history of public air quality awareness goes back to the smoke laws
in London in the Middle Ages. The modern era of air quality research started in
the early 1950s with the tragic London Fog episode which prompted epidemio-
logical studies and risk assessments of polluted outdoor air. Some early measure-
ments of selected indoor air pollutants started in the early 1960s. One of the first
gases measured was nitrogen dioxide (NO2 ) originating in combustion devices.
This was followed by measurements of some components of environmental
tobacco smoke (ETS) and radon in the 1970s. By the early 1980s it was generally
understood that indoor air pollution sources of volatile organic compounds
(VOCs), particulate matter (PM), and NO2 were typically far more significant in
overall personal exposure. Concerns about indoor formaldehyde, bioaerosols,
and allergens peaked in the 1980s and 1990s resulting in the publication of many
monographs, reports, and journal publications. In recent years, the wide use of
heating/air conditioning and energy-efficient building ventilation, the introduc-
tion of engineered building materials, and attached garages have often nega-
tively affected indoor air quality.
Four decades of research carried out by engineers, chemists, toxicologists,
epidemiologists, industrial hygienists, occupational health physicians, and psy-
chologists has resulted in much progress. Still, many unknowns are present and
many challenges remain. These include determination of sources, better under-
standing of release and migration mechanisms, improvements in real-time and
long-term monitoring, improved sampling and analysis methods, lowering of
detection limits, linking exposures with human health risk, improvements in
indoor air modeling and control, improvements in building materials, appli-
ances, consumer products, building codes, and public education.
2
Sampling and sample preparation strategy are the major part of the indoor
air quality assessment. The strategy considers (1) type of effect, e.g., acute vs.
chronic, (2) patterns of occupant exposure, (3) pollutant type and potential
health risks, (4) pollutant sources, (5) applicable regulations (for occupational
air), (6) applicable sampling and analysis methods, (7) site selection, (8) number
of samples and their statistical significance, and (8) budgetary, site access, and
logistical limitations. The choice of applicable sampling and analysis methods
continues to be one of the most challenging tasks. As sampling devices and ana-
lytical instruments become more sensitive and economical, air quality profes-
sionals are provided with a greater variety of choices. Ideal methods for indoor
air sampling and analysis could provide a direct (instantaneous/real-time) and
integrated (long-term, time-weighted average) automated reading in real time
for multiple pollutants of concern at multiple locations with no disturbance to
the sampled microenvironments. However, such a sampling is neither possible
nor practical, and the actual solution is the result of a compromise. On-site moni-
tors are often calibrated only to one or a few target pollutants. On-site monitors
for occupational monitoring cannot achieve the low detection limits required for
residential air and can be affected by interferences. Samples based on extraction
to a sorbent media via active or passive sampling or capture to a container
require sample preservation for later analysis in a laboratory. Attempts are
made to speed up and automate the sampling process by on-site analysis with a
portable gas chromatograph and/or mass spectrometer, liquid chromatograph,
and others. In such cases, various sample introduction methods can be used, e.g.,
solid phase extraction, thermal desorption, solid phase microextraction (SPME),
denuders, membrane extraction with solid interface (MESI), membrane inlet,
and other technologies as discussed in this book.
There are several types and categories of indoor air pollutant measuring sys-
tems depending on the sampler mobility, operating characteristics, and output
characteristics [1]. Personal, portable, and stationary samplers can be defined
using the mobility criteria. Personal devices can be worn or carried. These
devices are used to assess workplace exposure to target pollutants and are capa-
ble of a time-weighted average (TWA) measurement. Portable units can be
hand-carried during a survey. Stationary units must operate from a fixed loca-
tion. Depending on the operation mode, air sampling can be active or passive.
Active sampling requires a power source, e.g., pump or vacuum, to draw sample
to a sensor/analyzer or collector. Passive sampling relies on diffusion to acquire
samples. Finally, the measurement can be obtained from analyzers or collectors.
Analyzers provide an instantaneous/real-time (or nearly real-time) signal that
corresponds to the pollutant concentration. Analyzers are very useful in provid-
ing concentration including peak values. The output can be integrated to obtain
TWA concentrations. In contrast, collectors are often limited to TWA measure-
ments. No peak concentrations can be measured unless the averaging period is
very short and the detection methods are sensitive to analyze the lower mass col-
lected. TWA and peak concentrations are very important in indoor air sampling,
3
TABLE 1.1
Summary exposure guidelines for occupational air in the U.S.
Regulating/issuing agency Criteria for peak and averaged concentrations
ACGIH (American TLV: threshold limit value time weighted average (TWA). The
Conference of Governmental TLV is the concentration below which a healthy worker could
Industrial Hygienists) be exposed for 8 h/day, 5 days/week, for 20 years without
developing any disease. TWA means that the TLV can be
exceeded a time if it is counter-weighted by lower exposure.
STEL: shorfor t term exposure limit is a 15 min TWA exposure
that should not be exceeded at any time during a work-day.
OSHA (Occupational Safety PEL TWA: permissible exposure limit TWA. PEL TWAs are
and Health Administration) very similar to ACGIH TLVs but are issued as a regulation by
OSHA to protect workers.
NIOSH (National Institute REL TWA: recommended exposure limits TWA. REL TWAs are
for Occupational Safety very similar to ACGIH TLV TWAs. NIOSH conducts research
and Health) and issues recommendations but not regulations.
IDLH: immediately dangerous to life and health. IDLH values
reflect a concentration that is likely to cause death or
immediate or delayed permanent health effects.
4
main factors affecting the preconcentration process on passive samplers are [2]:
(1) temperature, pressure, humidity, and air (face) velocities; (2) concentration
of analytes, response time and degree of sorbent or reagent saturation; and (3)
back-diffusion, loss, and decomposition of analytes.
This chapter attempts only to summarize sampling and sample preparation
methods for the most important indoor air pollutants with the focus on residen-
tial air. However, the reader is encouraged to seek more detailed information in
several excellent books written specifically on the background, occurrences,
measurements, health effects and controls of indoor air pollutants [1,3-11].
These and other resources contain up-to-date extensive summaries of the com-
mercial air sampling instrumentation that is widely available. Brown and
Monteith compiled gas and vapor sample collectors [12]. Saltzman and Caplan
summarized detector tubes, direct reading passive badges, and dosimeter tubes
[13]. Namiesnik and Gorecki summarized references for applications of passive
samplers to determination of 16 inorganic and 30 organic compounds or group of
compounds in air [2]. Woebkenberg and McCammon discussed direct-reading
gas and vapor instruments [14]. These summaries contain information about
specific compounds, analysis methods, detection limits, instrument weight and
dimensions, and the name of manufacturer and/or supplier. Internet resources
have become very useful in searching for standard methods. These include the
U.S. Environmental Protection Agency (EPA), National Institute for Occupa-
tional Safety and Health (NIOSH), materials and chemicals, instrumentation
and buyers guides (Air &Waste Management Association), and up-to-date litera-
ture searches.
Nearly 1000 organic pollutants were known to be present in indoor air by the
end of the 1980s [15,16]. The list is continuously growing as new and more sensi-
tive detection methods are being developed. The concentrations of many of these
organics are often greater indoors than outdoors. Boiling point range is one of
the criteria for classification, e.g., very volatile organic compounds (WOC), vola-
tile organic compounds (VOCs), semivolatile organic compounds (SVOC), and
organic compounds associated with particulate matter (POM). Additional crite-
ria include reactivity, environmental importance, sampling considerations, and
chemical structure. Protocols for collection and analysis of organic pollutants
are being standardized by NIOSH, U.S. EPA, and other agencies [5,17]. This
subsection summarizes sampling and sampling methodologies for major groups
of organic compounds in indoor air.
5
Tenax -1.5 grams (60 mm bed depth)
Glass wool plugs 5 mm long
6
sampling sampling I nermal
durtc Trap GC
pump pump undesorption
unit
Fig. 1.2. Schematic of sampling with adsorbent tube and analysis with a thermal
desorption-GC-MS system.
the background contamination of Tenax with toluene and benzene from manu-
facturing, production of false nitrosoamines on Tenax that is in contact with
NOx, low breakthrough volumes for very volatile compounds (e.g., vinyl chlo-
ride), dilution of analytes when solvent desorption is used, accuracy of air flow
rate estimates and pump stability, and often incomplete recovery. The repro-
ducibility is generally between 10 and 30%. Method detection limits depend on
detector responses for each compound, and can be as low as 0.1 ng for MS.
The U.S. EPA Method IP-1A uses pre-evacuated SUMMA canisters for sam-
ple collection (Fig. 1.3) followed by trapping on a cryogenic trap and analysis on a
GC-MS or other GC detector [5]. The advantages of the canister collection
method include the ease of collecting 24 h samples, and ease of deployment,
transportation, and shipping. The disadvantages include incomplete recovery
and incomplete canister clean-up, up-front capital cost and the need to use a spe-
cialized GC interface for sample introduction.
In recent years, solid-phase microextraction has been used as a precon-
centrating device and sample introduction device [25] or as a field sampler capa-
ble of collecting a sample in the field, retaining it, and introducing it into a
conventional GC [26,27]. Sampling times [28] as short as 30 s for spot sampling
to as long as 8 h for time-weighted average can be used for selected VOCs and
SVOCs and quantitative analysis [29-31]. The limits of SPME are associated
with sample losses that can be faster than for many conventional methods and
selected analytes, and the need for determination of several physico-chemical
parameters in addition to typically measured sampling temperature, pressure,
and relative humidity [32].
Passive samplers are capable of collecting samples at a specific rate con-
trolled by the diffusion through a static layer of air or membrane [33,34]. Pas-
sive/diffusive samplers do not require pumps or any other means to move air,
thus they are robust and easily deployable. Samplers use sorption or reactive
sampling to trap analytes. The uptake rate depends on the analyte concentration
in air, the sampler area and length of the static air gap, and the diffusion coeffi-
7
Fig. 1.3. Sampler configuration for SUMMA canister sampling.
cient of the analyte in air. Air velocity, excessive humidity, sorbent starva-
tion/overload, sample loss, and reverse diffusion may interfere with the accurate
measurement. Typically, passive samplers such as a badge or tube for workplace
sampling can be adapted to indoor air sampling by extending exposure time and
selecting a more sensitive analytical instrumentation. Badges have a higher dif-
fusive uptake than tubes and use stronger absorbents, thus the reverse diffusion
is less likely. However, badges use solvent desorption that is difficult to auto-
mate, and solvent dilution decreases sensitivity [35]. Tube samplers have lower
diffusive uptake rates, use thermal desorption which can be automated, and the
tubes can be reused if the VOC recoveries are satisfactory. Comprehensive guid-
ance on designing a sampling strategy using passive samplers and applications of
diffusive samplers are presented elsewhere [35,36].
Typical badges consist of a charcoal wafer with 1 cm diffusion length, e.g.,
OVM 3500 (3M). Sampling times range from 24 h to 3 weeks. The surrogate
uptake rate per 8 h period as reported by the manufacturer is approximately 30
ml/min. This uptake depends on the analyte. Several research groups used pas-
sive samplers to study VOCs in indoor air [37-40]. Reported mean concentra-
tions of VOCs ranged from less than 1 to several hundred ALg/m 3 for individual
VOCs. Detection limits depend on the sampling time and analyte and typically
start at approximately 1 g/m 3 .
One of the most popular tube diffusive samplers (from Perkin-Elmer) con-
sists of a 90 mm long steel tube (5 mm ID) with a Tenax TA (or other sorbent)
retained by a stainless-steel mesh 15.3 mm from the opening. The opening is also
protected by another stainless-steel mesh to minimize convection inside the
tube. The opposite end of the tube is sealed with a Swagelok fitting. Sampling
time can range from 24 h to 4 weeks. Analysis is typically completed using a ther-
8
mal desorption (TD) CG and FID or MS detector. Several comprehensive indoor
air studies of VOCs involving homes and offices are described in the literature
[33,41,42].
Real-time monitoring devices are very advantageous for screening indoor air
quality because they can detect the presence of intermittent sources, determine
the most current concentration(s), monitor concentration trends, and monitor
effects of variables influencing air quality. Some may be hand-held and can be
used to find a source by following a concentration gradient. However, their
response is often non-specific because the output signal is the sum of instrument
responses to many compounds. The most popular types of detector are based on
photoacoustic spectroscopy (PAS), FID, photoionization detector (PID), non-
dispersive infrared (NDIR), and Fourier transform infrared (FTIR). The PAS
uses the fact that many VOCs absorb infrared light at 3.4 Aum. Absorption of the
pulsed light by the hydrocarbons that pass through a PAS cell generate changes
in pressure that is detectable by microphones and can be related to a surrogate
total VOC concentration [43]. Interferences from water vapor are compensated
by subtracting its concentration measured by another infrared (IR) wavelength.
The PAS detects a wide range of organic compounds from VVOCs to VOCs. The
PAS is very sensitive to n-alkanes, e.g., methane, while responding very poorly to
chlorinated organics [23]. It is possible to adjust the IR wavelength to provide an
enhanced sensitivity for a specific compound, e.g., formaldehyde. However,
interferences from other analytes should be considered [42].
The PID detector ionizes VOCs and VVOCs with UV radiation and detects an
electric current caused by ions and related to the concentrations of analytes
present. Typically, gases with a double bond, e.g., aromatic hydrocarbons,
respond very well. Responses of chlorinated organics are typically very low. Dif-
ferences in response between individual VOCs can be greater than one order of
magnitude. In a flame ionization detector, organic compounds are burned in a
hydrogen flame and produce ions. These ions are collected on an electrode where
they can be amplified. The produced current can be related to the concentration.
The signal is also dependent on the number of carbons and the type of analyte.
For example, response to n-alkanes and aromatic hydrocarbons is very similar,
while it is much lower for oxidized and chlorinated compounds. The NRIR and
FTIR can be calibrated for a specific compound [45,46].
1.2.2.2 Aldehydes
The aldehydes include a large family of compounds that have H-C =O functional
groups. Of these, formaldehyde, acetaldehyde, glutaraldehyde, and acrolein are
the most important aldehydes with respect to indoor air quality because of their
ubiquitous nature and toxicity. Formaldehyde is released from adhesives used in
particle-board and hardwood manufacturing, from textiles that were subjected
to permanent press treatment, from acid-cured coatings, and from selected
paper, insulation, and hard plastic products. It is also produced during combus-
tion or thermal oxidation in combustion appliances, wood burning, and tobacco
9
smoke. Formaldehyde is classified as a probable carcinogen. Several studies have
shown a significant dose-response relationship between formaldehyde concen-
trations in mobile homes and houses with particle-board underlay and health
effects which ranged from eye irritation and sore throat to chest pain, fatigue,
sleeping problems and depression [47,48]. Indoor air acetaldehyde, acrolein, and
glutaraldehyde are produced when organic matter is combusted. Similar to
formaldehyde, these aldehydes are also off-gassed from selected hard plastic
products, and many consumer products such as cleaning agents, cosmetics,
drugs, and textiles.
The most popular sampling methods for airborne aldehydes involve derivati-
zation because of the relatively low concentrations and high reactivity [49]. The
U.S. EPA proposed three methods for sampling and analysis of aldehydes in air:
sorbent tube sampling followed by analysis with HPLC, continuous colorimetric
gas analyzer, and passive sampling followed by analysis with HPLC [5]. The U.S.
EPA Method IP-6A utilizes tubes with 2,4-dinitrophenylhydrazine (DNPH)-
coated sorbent. This method is specific for formaldehyde, but it can be modified
to detect fourteen other aldehydes by optimizing chromatographic conditions,
e.g., using two columns in a series or by changing mobile phase composition
through a gradient program. Derivatizing agent reacts with a carbonyl group to
form more stable derivatives. The sampling time can vary from 5 min to 24 h.
Recommended sampling flow rate is between 500 and 1200 ml/min. It is also rec-
ommended that the derivatizing agent is added in situ on the silica gel adsorbent
because of abnormally high formaldehyde background in some of the commer-
cial pre-coated tubes [5]. Used tubes are desorbed with acetonitrile and stored
until analysis. The DNPH-aldehyde derivative is determined using isocratic
reverse phase HPLC and UV absorption operated at 360 nm. In some cases, an
ozone denuder may be used if high ozone concentrations may interfere with
DNPH and the derivatization product. The method detection limits vary from
approximately 1 ppbv for 10 1 sample volume to 0.01 ppbv for 1000 1 sample
volume.
The U.S. EPA Method IP-6B uses a continuous colorimetric analyzer for
determination of formaldehyde and other aldehydes [5]. Any aldehyde present in
air is scrubbed with sodium pararosaniline (PRA) in the presence of sodium sul-
fite (Fig. 1.4). The formed dye is directed through a flow cell where transmission
of light is measured by photodetectors at 550 nm. The intensity of color is pro-
portional to formaldehyde concentration. The method detection limits are 3
ppbv with the standard range up to 5 ppm. Method IP-6C uses passive diffusive
badges for determination of formaldehyde and other aldehydes (Fig. 1.5) [5].
This method is best suited for long sampling times up to weeks to typical formal-
dehyde concentrations of less than 100 ppbv because of low sampling rates. The
method utilizes a glass filter coated with DNPH that is placed behind a diffusion
gap. After analysis, the filter is desorbed with acetonitrile and analyzed with
HPLC and UV detector at 360 nm. The linear range for this method is approxi-
mately 0.05 to 20 Ag/l.
10
Reagents Flow meter
Urall
=>_____________~_
_~Retaining clip
Other derivatizing agents and methods can be used to extract aldehydes, e.g.,
chromotropic acid method, MBTH method, AHMT method, and acetylacetone
method [50]. The MBTH method (3-methyl-2-benzothiazolinone-hydrazone) is a
non-selective colorimetric method for the determination of aldehydes of low
molecular weight. This method quantifies total aldehydes as equivalents of
formaldehyde. MBTH reacts with aldehydes forming an azide. In parallel, a reac-
tive cation is formed by oxidation of MBTH with Fe(III). The resulting blue ionic
dye can be analyzed for absorbance at 628 nm [50]. This method is less sensitive
11
than the pararosaniline method and strong reducing agents can interfere with
aldehydes. AHMT (4-amino-3-hydrozino-5-mercapto-4H- 1,2,4-triazole) is used
in a strong alkaline media to facilitate reaction with aldehydes to a colorless
product which is then oxidized to form a purple-colored dye [50]. The dye can be
analyzed by colorimetry at 550 nm. This method is most sensitive for formalde-
hyde. The acetylacetone (2,4-pentanedione) method works very well for formal-
dehyde. Air is pumped through a solution, where the Hantzsch synthesis that
involves cyclization of acetylacetone in the presence of formaldehyde and ammo-
nium nitrate to form dihydropiridine-3,5-diacetyl-1,4-dihydrolutidine (DDL)
occurs. The quantification is performed using colorimetry at 412 nm or fluores-
cence at 510 nm [51].
A novel method for sampling and determination of formaldehyde based on
SPME is also available. In this method o-(2,3,4,5,6-) pentafluorobenzyl hydroxy-
amine (PFBHA) is first coated on a PDMS/DVB 65 gm fiber using headspace
extraction for approximately 2 min immediately before sampling. On-fiber
derivatization of formaldehyde occurs when the fiber is exposed to air. Sampling
times can be as low as few minutes for spot sampling with an exposed fiber to as
long as 8 h for time-weighted average sampling with a retracted fiber [30,52].
SPME-based samples are analyzed on a GC-FID. Method detection limits are
approximately 5 ppbv. Thermal desorption tubes have also been used for sam-
pling and determination of aldehydes, albeit with a lesser success than afore-
mentioned methods. Studies with gas standards have shown that the Tenax TA
sorbent is the most effective sorbent for aldehydes from butanal to decanal [53].
However, Tenax TA is not appropriate for acetic aldehyde, propenal, and
propanal.
1.2.2.3 Pesticides
Pesticides are used to control, repel, or kill pests, e.g. insects, weeds, fungus, or
rodents [54]. There are at least 600 pesticides and approximately 50,000 pestici-
dal formulations. Pesticides have bioaccumulating properties and can cause
chronic health effects [5]. On average, the major exposure to pesticides occurs
inside the home [55]. Most pesticides are semi-VOCs and are present in air pri-
marily as gases. These gases can also be volatilized from particulate matter after
sample collection [55]. Pesticides can also be simple inorganic substances,
organometals, VOCs, and nonvolatile organic compounds. Indoor air sources
include treatments for cockroaches, termites, ants, and other insects, disinfec-
tants for kitchen and bathroom, room deodorizers, laundry aids, outdoor lawn
treatments, and others [55].
Because indoor concentrations are typically low, long sampling times, e.g. 24
h, are usually selected to collect enough detectable mass. The U.S. EPA Method
IP-8 for organochlorine pesticides uses active sampling and adsorption on poly-
urethane foam (PUF) [5] (Fig. 1.6). This method is a modified Method EPA 608
3
and is recommended for concentrations greater than 0.01 Atg/m for 4-24 h sam-
pling time and sampling rate of 1-5 1/min. Pesticides are extracted from the PUF
12
Glass tube Polyurethane foam PTFE filter gasket Screw cap
Fig. 1.6. Schematic of simple air sampling cartridge with PUF sorbent and particulate filter for
pesticide sampling.
13
summary of common indoor PAHs and nitro-PAHs and their concentrations can
be found elsewhere [58]. Lower molecular weight PAHs, e.g., 2 to 4-ring, are
likely to be found in gaseous phase, while heavier PAHs occur in the particulate
phase. Sampling of PAHs typically requires the use of quartz fiber filters and
sorbents because of the partitioning of PAHs in gaseous and particulate phase.
U.S. EPA Method IP-7 uses a combination of quartz fiber filters and XAD-2
or PUF sorbents [5]. Approximately 30 m3 is drawn through a filter situated
upstream from an adsorbent cartridge at 20 1/min. The filter and cartridge are
extracted by Soxhlet extraction with either methylene chloride (for XAD-2) or
hexane (for PUF), concentrated by a Kuderna-Danish evaporator, and followed
by analysis on GC-MS or HPLC and UV or fluorescence detection. In some cases,
e.g., flame ionization detection, an additional step of silica gel cleanup may be
needed. Method detection limits range from approximately 5-50 g/l for HPLC-
fluorescence, 50-250 Alg/l for HPLC-UV, to as low as 100 ng/l for GC-MS [5].
Phthalates, a.k.a. phthalate esters, are the most commonly used plasticizers
in PVC resin [58]. Phthalates are used to manufacture a wide variety of con-
sumer products and building materials, e.g., food containers and wraps, toys,
tiles, shower curtains, vinyl upholstery, lubricants, sealers, and adhesives. Many
phthalates are classified as endocrine disrupters. Reported indoor concentra-
tions are as high as 1.5 ng/il compared to 0.05 ng/l for outdoor air [58]. Typical
sampling includes particulate collection on a glass fiber filter and gas phase col-
lection on PUF or XAD-2 [59].
Phenols in indoor air typically originate from phenol-formaldehyde bonded
wood products [50]. Alkylphenol polyethoxylates are common surfactants used
in detergents, cleaners, and personal hygiene products [58]. Some of these com-
pounds are likely involved in endocrine disruption. There are several methods of
sampling and analysis of indoor air phenolic compounds. Quartz filters can be
used in combination with PUF or XAD-2 sorbents. Filters and sorbents are then
extracted with methylene chloride, followed by derivatization and analysis on a
GC-MS using single ion monitoring [59]. Another method is based on pumping
air at 2 l/min through 0.1 M NaOH using Muencke-type absorbers [50]. After
neutralization with hydrochloric acid, a p-nitroaniline reagent is added to form
4-nitro-4'-hydroxyazobenzene (dye). The absorbance of total phenolic com-
pounds is measured at 490 nm. Interferences include aromatic amines, hydrogen
sulfide, and sulfur dioxide. The MDLs are approximately 0.8 Ag/m 3 for a total
sampling volume of 100 1 [50]. A variation to this method uses direct analysis
(without derivatization) of phenols after a neutralization step using HPLC and
UV (at 254 nm and 280 nm) and/or fluorescence detection. Fluorescence is pre-
ferred for alkylphenols [60].
14
from 0.01gm to more than 100 Am. Larger particles are typically solids that were
abraded or ground, whereas smaller size particle often originate in combustion,
condensation, or reaction as a fine nuclei, e.g., 0.05 jAm, which tend to grow to
more stable size below 1 m [5]. Coarse particles are defined as those particles
for which the aerodynamic diameter is less than 10 Am (PM-10). Coarse particles
can be inhaled; however, they are not capable of reaching deep into the lungs
where only fine particulate with an aerodynamic diameter less than 2.5 Am
(PM-2.5) can enter. PM concentrations indoors are affected by combustion
sources (environmental tobacco smoke, wood stoves, gas, and appliances), occu-
pant activities (cooking, vacuuming, and cleaning), and infiltration of outdoor
air. Indoor-to-outdoor ratios often exceed 1, depending on the type of activity,
time of day, ventilation type, and existence of sources [61]. Characterization of
air particles and aerosols is described in a separate chapter in this book.
The U.S. EPA proposed two methods for determination of respirable
particulates (defined as PM-10) [5]. U.S. EPA Method IP-1OA uses a micro-
environmental exposure monitor (MEM) as a fixed site monitor (Fig. 1.7) and a
personal exposure monitor (PEM) for estimation of personal exposure to parti-
cles. Both devices use a specially made nozzle and an impactor plate designed to
provide a predetermined size cut-off at a constant air flow rate of 4 1/min. A
tarred particulate filter is mounted downstream from the separator. For the
MEM device, the total volume sampled is approximately 5.5 m3 per day. The
average concentration is estimated by dividing the weight gain of the filter by
the total volume sampled. A 12 h level of detection (LOD) is typically 8 jg/m3 for
the PEM and 4 jg/m3 for the MEM. These LODs can be affected by the precision
of the collection and size cut-off efficiency, particle resuspension, microbalance
measurements, flow rate estimates, interactions between collected particles, and
interactions with the filter media. In cases where chemical analysis is also con-
duced in addition to mass gravimetric analysis, special filter materials that do
not interfere with sampled gases or analytical procedures should be used.
over
-10 cut-off
action plate
-2.5 cut-off
)action plate
rticle
lection filter
Pump
Fig. 1.7. Schematic of microenvironmental exposure monitor (MEM) for particulate matter.
15
Diffusion denuders are used for selective collection of individual gases or
classes of gases upstream from the particulate filter [62,63]. Denuders allow for
separation of gases from particulates and minimizing interferences by preclud-
ing reactions between collected particulate and sampled gases. Sodium carbon-
ate, sodium hydroxide, and potassium hydroxide can be coated on the surface of
a denuder where they can effectively remove acidic gases (formic and acidic
acids), HNO3 , HC1, and SO2. There are several methods for analysis of collected
particulate: these include X-ray fluorescence, proton-induced X-ray emission,
neutron activation analysis, atomic absorption, and ion chromatography [11].
U.S. EPA Method IP-1OB uses a tapered element oscillating microbalance
(TEOM) continuous particulate monitor [5]. This instrument is equipped with
an interchangeable filter cartridge where particulate matter collects. A hollow
glass oscillating tube is attached to the cartridge. As particulates collect on the
filter, the tube oscillation frequency decreases and can be related to the mass
deposited. Drawn air is heated to a designated temperature and the air flow is
measured and adjusted. The sampling rate is between 0.5 to 5 1/min. A pre-
separator for PM-2.5 or PM-10 can be mounted upstream from the heated inlet.
Dissolved semivolatiles, mechanical noise, and rapid changes in air temperature
may interfere with the readings. The LODs are approximately 5 Ag/m 3 .
Stationary or personal nephelometers can be also used [64]. Nephelometers
measure light scattering and relate it to particulate matter concentration. An
averaging time as low as 1 min is possible. Nephelometers illuminate a fixed
sample volume from one side and observe the changes in the amount of light that
is scattered by particles and gas molecules in the direction of a photomultiplier
tube. A particle filter can be inserted periodically into the sample stream to mea-
sure the light scattered by gas molecules and can be subtracted from the total
scattered signal to determine the contribution from the particles alone. The
instruments are calibrated by filling the sample volume with CO2 gas, which has
a known scattering coefficient. Beta attenuation monitors detect changes in beta
radiation scatter inside a fixed volume cell with moving sampled air. Electrical
counters are based on charging passing aerosols and measuring their capability
to cross a controlled electrical field inside a sample cell.
1.2.4 Radon
Radon is a radioactive and inert gas originating from the decay of uranium 238 and
thorium 232 . Radon originates from soil bedrock and often penetrates to the house
envelope through the basement floor, cracks, and deep well water. Radon exists in
several isotopic forms which, shortly after their formation, combine with small
aerosol particles [65]. These particles can be inhaled in the lungs where daughter
radon molecules decay and release alpha, beta, and gamma rays that increase the
probability of cancer, cardiac/gastrointestinal, and nervous system distress.
Indoor air concentrations are often several orders of magnitude higher than those
of ambient air and typically range from 10 to 140 bequerels per cubic meter
16
(bq/m3) [66]. Concentrations as high as 100,000 bq/m3 have been detected. Pre-
ferred measurements last from several months to a year and use passive dosime-
ters because emission rates for radon can vary significantly, hourly, diurnally, and
seasonally. Grab sampling is only used for initial screening and spot check of con-
trol methods.
Most measuring methods are based on detection of the alpha, beta, and
gamma particles produced by radon and radon decay products and include scin-
tillation cells, alpha track detectors, charcoal detectors, electret detectors, and
electronic monitors [65]. Scintillation cells consist of a small metal cylinder
coated on the inside with ZnS(Ag). After the air sample fills the cylinder either
via diffusion or with a pump, alpha particles strike the ZnS(Ag), the signal is
multiplied, and the decay rate is related to the radon concentration [67].
Alpha track detectors (ATDs) are inexpensive and ideal for long-term sam-
pling. ADTs use special plastic or polymeric materials, e.g. LR-115 cellulose
nitrate film, CR-39 thermoset polymer plastic, or Makrofol polycarbonate plas-
tic, that are damaged specifically by the passage of alpha particles. The damage
trail can be made visible by chemical or electrochemical etching and counted
under microscope or by software in a laboratory [68]. Most ADTs consist of few
cm2 of coating inside an almost airtight container to which air enters via diffu-
sion. Charcoal detectors are suitable for short-term (several days to a week) sur-
veys. Radon is adsorbed to charcoal inside a passive, pocket-size sampler. The
gamma radiation emitted by radon and its decay products are later analyzed in
the laboratory with a gamma ray detector such as sodium iodide (NAI(T1)) or
with a thermoluminescent detector [69].
Electret detectors use a charged material such as aluminized Teflon which
will retain a charge for as long as a year [70,71]. Typically, an electret is mounted
inside a small plastic container to which air diffuses. Charged radon decay prod-
ucts decrease the electret charge over time. The charge reduction can be related
to radon concentration on-site. Electrets can be recharged and therefore reused.
The presence of water vapor and other charged particles may interfere with mea-
surements. Electronic monitors are typically active samplers that detect alpha
radiation using solid-state detectors. This type of detector is capable of grab sam-
pling and on-site analysis [72].
1.2.5 Asbestos
Asbestos is the generic name for a wide variety of hydrated silicate minerals such
as chrysolite or amosite with a simplified chemical formula of MenSi 2mOp(OH)2 r,
where Me might be Mg, Na, Fe, or Ca, n = 2,5, or 7, m = 1 or 4,p = 5 or 22, r = 1
or 2. All types of asbestos are considered to be carcinogens due to their capability
of tissue scarring [73,74]. Asbestos has extremely good insulation, fire retarda-
tion, and chemical resistance. Indoor air quality can be adversely affected when
asbestos is released from building materials.
Measurements of asbestos concentrations are very challenging because of
17
possible interferences from other particulate matter. The range of typical asbes-
tos concentrations can vary from as low as 0.1 ng/m3 to 1 m/m 3 , which are often
several orders of magnitude lower that PM concentrations [3]. The "critical"
asbestos fibers are equal or longer than 5 mm, and have a diameter up to 3 gm.
Typical asbestos concentrations expressed in critical fibers per m3 (F/m 3 ) vary
from approximately 100 F/m3 in buildings without specific asbestos sources up to
10,000 F/m 3 in buildings with friable asbestos [3]. Occupational air concentra-
tions can be as high as 100,000,000 F/m 3 .
The most common approach is to sample with a PM sampler using a
polycarbonate cellulose ester filter with 0.4 /m or 0.8 Aum pore size [75]. A seg-
ment of used filter is then analyzed with an electron microscope where asbestos
is identified, counted, and sized. Often an X-ray diffraction analyzer is attached
to the electron microscope to positively identify asbestos fibers. The results are
extrapolated to a number of critical fibers per air volume based on sample vol-
ume and filter area analyzed. Conversion factors for estimation of asbestos mass
per volume of air can also be used [76]. One of the new developments is the use of
PDMS-7 am SPME fibers to collect airborne asbestos followed by Raman micro-
spectroscopy analysis [77]. SPME coating is used as a "glue" to capture airborne
asbestos, which can be positively identified by Raman microspectroscopy.
1.2.6 Ozone
18
The workable range varies from approximately 30 ppb to 1 ppm. Presence of
humidity and changes in pressure can interfere with the method sensitivity. Dry
colorimetry uses paper tape impregnated with 0O-specific dry reagent which
changes color upon contact with ozone. The color change can be measured with a
photometer and related to 03 concentration. Working range for this method var-
ies from approximately 30 ppb to 300 ppb. The same approach can be used for
other gaseous pollutants by use of other gas-specific dry reagents. The chemo-
luminescence method is based on reaction of O0 with ethylene to produce specific
wavelength in the 300 to 600 nm range. The lower detection limit (LDL) for this
method is approximately 1 ppb. Solid-state ozone sensors are also applicable for
indoor air applications [80].
19
I
A\ ~
o/ Rml~~hre
Air sample /
Up
inlet
/Acrylic tube
ting
on
3 stainless steel
screens coated with TEA
W162
-Fixed cap
20
Continuous monitoring for NO2 concentrations as low as 1 ppb to 10 ppb can
be achieved with chemoluminescence [86]. This method is based on a two-stage
process initiated by reaction of NO with internally generated ozone to form NO2 .
At this stage, the NO concentration can be estimated based on the characteristic
luminescence. Next, nitrogen dioxide is converted at 325°C to NOx and the signal
represents the NOx (as NO + NO2 ) concentration. The NO2 concentration is cal-
culated as a difference between the NOx and NO concentration. Some monitors
are capable of measuring ammonia concentrations when the second converter is
used, and the ammonia concentrations are calculated as a difference between
the total nitrogen and the NOx.
Sulfur dioxide (SO 2) is a colorless gas with a strong pungent odor detectable at
approximately 0.5 ppm. It is classified as a primary air pollutant that is generated
during combustion of sulfur-containing fuels, e.g., natural gas, coal, wood, and
kerosene. SO2 can form acid aerosols when SO2 reacts with water vapor, and sec-
ondary fine particulate matter when SO2 and SO3 react with ammonia. Sulfur
dioxide is highly soluble in water, thus it irritates mucous membranes and causes
respiratory problems. Indoor SO2 concentrations are typically lower than out-
doors and range from approximately 2 Atm/m3 for houses without kerosene heat-
ers to 150 Am/m 3 for households with kerosene heaters and/or gas stoves [88].
Concentrations as high as 2660 tkm/m 3 (approximately 1 ppm) were measured in
houses with unvented kerosene heaters [89].
There are several approaches for SO2 sampling in ambient air that can be
modified for indoor air sampling including U.S. EPA Method 6 for "stationary
sources" by titration [90], U.S. EPA Method 15 for sulfur compounds by GC/ FPD
[91], direct reading methods using pararosaniline reaction and electrochemical
reaction [92]. Method 6 uses a heated probe to collect the sample and filter out
particulates and sulfuric acid mist. Ammonia and water-soluble cations may inter-
fere. Air samples are collected in impingers filled with isopropanol and hydrogen
peroxide. The concentration of SO2 is determined by titration. There are other
variations of this method involving different absorbing solutions, e.g., tetrachloro-
mercurate and pararosaniline, formaldehyde and pararosaniline, and different
analyzers, e.g., spectrophotometer. Typical LDLs are approximately 10 ppb.
Continuous monitoring of SO2 can also be achieved using amperometric
methods [90]. In this method SO2 entering the detector cell (via diffusion or
pump) reacts with free iodine or bromine which serve as a titrant. Any change in
titrant concentration produces a potential on anode/cathode electrodes that is
restored by a current generated by a pair of sensor/reference electrodes. This
current is proportional to the SO concentration. The LDL of this method is
approximately 10 ppb. A flame photometric detector (FPD) can be also used as a
stand-alone detector for continuously flowing air or as a GC detector [85,93]. In
this case, the LDLs are close to 1 ppb.
21
1.2.9 Carbon monoxide and carbon dioxide
Carbon monoxide (CO) is a colorless, odorless, and tasteless gas generated pri-
marily by incomplete combustion of carbon-containing materials. It is one of the
most common air pollutants. Carbon monoxide is typically generated in mal-
functioning or inadequately vented cooking/ heating appliances, fire and tobacco
smoke. In same cases, penetration of vehicular emissions from underground or
attached garages can also be a source of indoor air CO. When inhaled, CO can
quickly combine with blood hemoglobin and drastically reduce the oxygen-
carrying capacity. Health effects vary from minor aggravation to death, depend-
ing on the concentration, exposure time, and the level of physical activity. Typi-
cal concentrations recorded in residential air varied from less than 1 ppm to
nearly 30 ppm, with the average of 2.8 ppm [3]. These concentrations are
approximately one order of magnitude lower than ambient air CO concentra-
tions found in extremely polluted large cities, where automotive emissions are
the major CO source [94]. The NAAQS for ambient air are set to 9 ppm for an 8 h
period, and 35 ppm for 1 h period.
Carbon dioxide (CO 2 ) is a colorless and odorless gas. It is a principal combus-
tion product. The main sources indoors are natural gas, coal, wood-burning
appliances, and exhaled air. Concentrations of CO2 are often used to evaluate
ventilation and air turnover rates indoors. The ASHRAE Standard 62-1981 sug-
gests that enough outdoor air is supplied per occupant so that resulting CO2 lev-
els are not greater than 2,500 ppm.
There are several methods that are applicable to CO and CO 2 sampling and
analysis. Most are based on CO/CO 2 detection via a non-dispersive infrared ana-
lyzer (NDIR) and electrochemical or catalytic oxidation. The NDIR uses infrared
radiation that passes through two parallel cells, the first with an air sample
(enclosed or continuously flowing through), the second filled with CO/CO 2 -free
reference air. Concentrations of CO/CO2 are related to the difference in
absorbance of one or more characteristic CO/CO 2 wavelength IR spectrums
using Beer-Lambert Law. The LDLs of this method are approximately 1 ppm.
Another variation of the IR-based approach is the use of gas filter correlation
(GFC) where an optical filter limits the photometers' sensitivity to the IR wave-
lengths of interest. In the GFC, the IR beam passes through a gas filter alternat-
ing between CO/CO 2 and N2 . The CO/CO 2 gas filter produces a beam that cannot
be attenuated by CO/CO2 in the sample. The N2 is transparent to the IR radia-
tion and can be absorbed by CO/CO 2 . The chopped detector signal is modulated
by the alteration between CO and N2 and is proportional to the CO/CO2 concen-
tration [5]. The MDLs of this method are approximately 0.020 ppm.
A more sensitive and specific detector for CO/CO2 is a photoacoustic-IR [3].
The air sample is irradiated with light of a wavelength coinciding with an
absorption maxima of CO/CO 2 and the gas temperature increases due to absorp-
tion. The gas pressure will increase if the volume is constant and vary if the light
is modulated by a chopper. Pressure variations produce a sound wave which can
22
be detected by a condenser microphone and related to the CO/CO2 concentra-
tions. Electrochemical or catalytic oxidation analyzers are often small in size and
used as personal exposure monitors that can be attached to a person [85]. The
monitor samples either via diffusion or convection aided by a small diaphragm
pump. The CO is converted to CO 2 in a liquid cell producing a small electrical
current that can be related to CO concentration. The typical MDL is approxi-
mately 1 ppm. Carbon dioxide and CO can also be sampled with colorimetric
direct reading tubes where typical MDLs are approximately 5 ppm and 100 ppm
for CO and CO2, respectively. Carbon dioxide can also be sampled with a canister
and analyzed on a GC equipped with a helium ionization detector and/or thermal
conductivity detector, resulting in MDLs of approximately 1 ppm and 100 ppm,
respectively.
23
C r inlet
;k spring
Glass wool
,ent section
orbent section
, Personal pump
Fig. 1.10. Schematic of sorbent tube for sampling of nicotine and other VOCs.
Bioaerosols include a wide range of particles and aerosols that affect indoor air.
These include endotoxins, allergens derived from insects, domestic animals, and
natural rubber, pollen, fungi, bacteria and viruses [104,105]. Inhalation of air-
borne endotoxins have been linked to occupational diseases suspected in the
development of sick building syndrome and triggering asthma responses [105].
Endotoxins are composed of lipopolysaccharides (LPS) that are shed (and subse-
quently can become airborne) from the outer cell membrane of gram-negative
bacteria. The potency of endotoxins is measured by comparing the biological activ-
ity of a sample with that of a reference such as Limulus amebocyte lysate (LAL)
assay [105]. The U.S. reference standard is purified from E. coli 0113:H10:K. By
convention, 1 endotoxin unit (EU) is equivalent to 0.1 ng of the reference standard
endotoxin and it does not represent a measure of concentration of a single chemi-
cal [105]. Averaged endotoxin concentrations measured varied from a few tens to
a few hundreds EU/m3 in agricultural and related workplaces. Concentrations
24
exceeding 1000 EU/m 3 were found in machine shops where metal lubricating fluid
was infected and aerosolized [106]. Typical sampling methods use filter media or
glass impingers to collect airborne endotoxins using the LAL assay. Filter selec-
tion depends on the type of aerosol and source sampled. Teflon, cellulose mixed
ester, and PVC can alter, inhibit, or deactivate endotoxin activity [107]. To date,
only polycarbonate capillary pore membrane filters are known to be acceptable for
a wide range of endotoxins [108]. Glass impingers are not sensitive samplers for
submicron particulates, which consist the majority of airborne endotoxins.
Extraction of filters involves sonication and rocking in pyrogen-free water. Sam-
ple analysis usually involves the use of an assay containing an enzyme that is spe-
cifically activated by endotoxin, which in turn, activates other enzymes and
enhances response [105]. The variability of assays can be reduced by the use of
GC-MS, which is based on the analysis of specific markers that are characteristic
of LPS, e.g., 3-hydroxy fatty acids which are derivatized from samples [109]. The
main disadvantage is the lower detection limits compared to the assay method.
Airborne allergens contain antigen proteins that induce immune responses
in humans. The major sources of residential allergens include cats and dogs,
mites, and German cockroaches [110]. Cat and dog allergens are usually
attached to small particles, which cannot be settled easily and can attach them-
selves to walls and clothing, while mites proteins tend to attach themselves to
larger particles of 10 to 20 Am which are less mobile. Electron microscopic analy-
sis identified these particles as mite fecal pellets [111]. The protein identified as
a mite allergen (Der p 1) has MW equal to 24,000 daltons and could be radio-
labeled with 125I and measured using the inhibition radioimmunoassay (RIA)
[112]. This technique allows the measurement of mite allergens in house dust
from approximately 0.4 to 500 Ag of Der p 1 per gram of dust. Special nasal
masks can be used to trap, monitor, and identify particles entering the nose
[113]. Typical sampling methods for airborne allergens use cascade samplers or
liquid impingers, followed by the radioallergosorbent (RAST) or radioimmuno-
assay (RIA) analysis. It is currently possible to measure a single representative
allergen for cats (Mancini technique), and mites, cats, dogs, and cockroaches
using two site monoclonal antibody-based assays [110].
There is growing evidence that immunologically foreign proteins can also be
generated by insects and pests such as cockroaches [114]. These include the Bla g
1 and Bla g 2 associated with German cockroaches which are likely excreted in
their feces. Particles carrying these proteins are likely greater than 5 Am and can
become airborne only when disturbed, e.g., during vacuuming, and likely cannot
get attached to clothing. In contrast, cat allergens are typically attached to
smaller particles [115]. Typical cat allergen concentrations for houses with a cat
range from 2 to as high as 80 Am Fel d 1 per gram of dust. Both cat and dog aller-
gens are sticky, easily transferable to houses without them, and have a potential
to accumulate [116].
Pollen is a male reproductive part of flowering plants that can cause hay
fever, allergenic asthma, and other allergenic symptoms. Pollen can enter indoor
25
spaces and air via vents, cracks, open windows and doors, clothing, and pets. Pol-
len can cause discomfort in persons whose immune system overreacts to pollen
grains that come in contact with mucous membranes in areas such as eyes or the
respiratory system [117]. The typical ratio of outdoor-to-indoor pollen concen-
trations depends on the time of year, time of day, and ventilation scheme, and
ranges from approximately 1 to more than 20 [118,119]. Average measured con-
centrations in homes were 1.2 grains per m3 (gr/m3 ) for airborne pollen and
5.8x 105 gr/m3 for dustborne pollen [120]. The simplest way to sample airborne
pollen consists of a microscope slide coated with a sticky film exposed to settling
pollen for sampling times ranging from hours to days, e.g. Durham sampler. Set-
tling samplers tend to be biased toward particles with larger settling velocity and
the difficulty to relate the number of grains per surface area to the actual air-
borne concentrations. Another variation of the settling sampler is a Taber trap
consisting of a 10 x 10 cm cylinder covered with a specially shaped cover to allow
pollen to settle inside to a glycerol solution. This solution can be periodically
removed and analyzed for pollen settling rate.
Rotating arm impactors use narrow bars wrapped with a clear plastic or
adhesive-coated tape that are rotated around the axis collecting particles [117].
These particles can be stained, e.g. Calberla stain, and counted under a light
microscope. Another popular sampler is the Rotorod sampler capable of collect-
ing particles down to approximately 15 ,Am. This sampler is typically deployed
for periods from 15 to 60 s at 10 min intervals for a 24 h period. Hirst-type sam-
plers use a narrow slit through which air is pulled and impacts a moving
impaction surface coated with an adhesive. Hirst-type samples allow for time
resolution down to 2 h and particle size cut-off at 5 Aum which allows for collection
of the majority of airborne pollen. Filtration samplers use high volume pumps
and a variety of filters to effectively collect airborne pollen. Passive filtration
devices, typically used in outdoor air, e.g. Cour traps, use two 5-layered gauze fil-
ters with an exposed area of approximately 400 cm 2 that can be exposed for days
up to several weeks. Pollen grains can be counted and identified using light
microscopy at magnifications from 400 x to 1000 x. Sample preparation usually
involves dissolving organic matter in a procedure called acetolysis, which does
not affect the resistant pollen outer layer [121]. In the last decade, the use of
labeled antibodies and detection of the actual allergens with fluorescence was
developed. This method is faster than conventional methods and amenable to
automation of counting [122].
Fungi (mold, mildew) are ubiquitous in indoor air environments. Fungi
release water, C0 2 , ethanol, organic acids, enzymes, VOCs, and non-volatile tox-
ins (mycotoxins) [123J. Some of the fungi produce proteases which are likely to
be allergens. VOCs such as aldehydes, ketones, and mycotoxins including afla-
toxins, satratoxins, verrucarins, roridins, griseolvium, and fumonisis are of
interest to indoor air quality. Indoor air sampling often involves the use of an
impactor sampler or liquid impinger. In the impactor sampler, moving air passes
through jets (small nozzles) that impact air against a Petri dish with a nutrient-
26
Air inlet-*
Separating
nozzles
Fig. 1.11. Schematic of two-stage biological aerosols sampler for collection of non-respirable and
respirable fraction.
coated surface (Fig. 1.11). Staged samplers, where each stage is designed to cap-
ture a predetermined size or range of sizes, can be used to determine the size dis-
tribution of airborne aerosols [124]. Cultures appearing after 3-7 day incubation
can be visually counted. Results are reported as colony forming units per m3 of
air (CFU/m 3 ). Liquid impingers, such as the all-glass impinger (AGI), capture
and retain airborne microorganisms, which can be cultured and counted on a
specific (or non-specific) agar. Mycotoxins are often analyzed with a combination
of analytical methodologies including thin layer chromatography (TLC), HPLC,
SPE, GC-MS, LC-MS, and NMR which are summarized in the handbook of the
Association of Official Analytical Chemists [125]. Airborne bacteria can be
labeled and analyzed using direct count methods based on staining and fluores-
cence [126]. Another method, the polymerase chain reaction (PRC), is based on
detection of selected regions of DNA specific to a particular kind of bacteria of
interest, amplified by 10 4 to 106 times for identification [127]. Mobile and near
real time systems for field sampling of aerosol protein and detection via
absorbance promise automation and on-site analysis [128].
REFERENCES
1 N.L. Nagda, H.E. Rector and M.D. Koontz, Guidelines for Monitoring Indoor Air
Quality. Hemisphere Publishing, New York, NY, 1987.
2 J. Namiesnik and T. Gorecki, LC-GC Europe, 9 (2001) 679.
3 M. Maroni, B. Seifert and T. Lindvall (Eds.), Indoor Air Quality: A Comprehensive
Reference Book. Elsevier, Amsterdam, The Netherlands, 1995.
4 J.S. Spengler, J.M. Samet and J.F. McCarthy (Eds.), Indoor Air Quality Handbook.
McGraw-Hill, New York, NY, 2000.
5 W.T. Winberry, L. Forehand, N.T. Murphy, A. Ceroli, B. Phinney and A. Evans
(Eds.), Methods for Determination of Indoor Air Pollutants: EPA Methods. Noyes
Data Corporation, Park Ridge, NJ, 1993.
6 T. Salthammer (Ed.), Organic Indoor Air Pollutants: Occurrence, Measurement,
Evaluation.Wiley-VCH, Weinheim, Germany, 1999.
27
7 M.J. Boss and D.W. Day, Air Sampling and IndustrialHygiene Engineering.Lewis
Publishers, Boca Raton, FL, 2001.
8 E.W. Miller and R.M. Miller, Indoor Pollution:A Reference Handbook. ABC-CLIO,
Santa Barbara, CA, 1998.
9 L.H. Keith and M.M. Walker, Handbook of Air Toxics: Sampling, Analysis, and
Properties.CRC Press, Boca Raton, FL, 1995.
10 B.S. Cohen and S.V. Hering (Eds.), Air Sampling Instruments for Evaluation of
Atmospheric Contaminants. ACGIH, Cincinnati, OH, 1995.
11 R. Wilson and J. Spengler (Eds.), Particlesin Our Air: Concentrations and Health
Effects. Harvard University Press, Cambridge, MA, 1996.
12 R.H. Brown and L.E. Monteith, in: B.S. Cohen and S.V. Hering (Eds.), Air Sampling
Instruments for Evaluation of Atmospheric Contaminants. ACGIH, Cincinnati, OH,
1995, Ch. 17, p. 369.
13 B.E. Saltzman and P.E. Caplan, in: B.S. Cohen and S.V. Hering (Eds.), Air Sampling
Instrumentsfor Evaluation of Atmospheric Contaminants.ACGIH, Cincinnati, OH,
1995, Ch. 18, p. 401.
14 M.L. Woebkenberg and C.S. McCammon, in: B.S. Cohen and S.V. Hering (Eds.), Air
Sampling Instruments for Evaluation of Atmospheric Contaminants. ACGIH,
Cincinnati, OH, 1995, Ch. 19, p. 439.
15 U.S. Environmental Protection Agency, Report to Congress on Indoor Air Quality,
Vol. II, Assessment and Control of Indoor Air Pollution, Report no EPA
400/1-89-001C, Washington, DC, 1989.
16 L. Molhave, IndoorAir, 4 (1996) 357.
17 National Institute for Occupational Safety and Health, Manual of Analytical
Methods, 4th edn. Centers for Disease Control and Prevention, Cincinnati, OH, 1994.
18 W.G. Tucker, in: J.S. Spengler, J.M. Samet and J.F. McCarthy (Eds.), Indoor Air
QualityHandbook. McGraw-Hill, New York, NY, 2000, Ch. 31, p. 31.1.
19 P. Wolfoff, Volatile organic compounds-sources, measurements, emissions and the
impact on indoor air quality. IndoorAir suppl., 3/95, (1995) 1.
20 American Society for Testing and Materials (ASTM), D5466-95: Standard Text
Method for Determination of Volatile Organic Compounds in Atmospheres (Canister
Sampling Methodology), West Conshohocken, PA, American Society for Testing and
Materials, 1995.
21 American Society for Testing and Materials (ASTM), D6196-97: Standard Practice
for the Sampling and Analysis of VOCs in Air by Pumped Sorbent Tube and Thermal
Desorption, West Conshohocken, PA, American Society for Testing and Materials,
1997.
22 M. Harper, J. Chromatogr.A, 885 (2000) 129.
23 A.T. Hodgson, IndoorAir, 5 (1995) 247.
24 E. Ilgen, N. Karfich, K. Levsen, J. Angerer, P. Schneider, J. Heinrich, H-E.
Wichmann, L. Dunemann and J. Begerow, Atmos. Environ., 35 (2001) 1235.
25 K. Elke, E. Jermann, J. Begerow and L. Dunemann, J. Chromatogr. A, 826 (1998)
191.
26 D. Gorlo, B. Zygmunt, M. Dudek, A. Jaszek, M. Pilarczyk and J. Namiesnik,
FreseniusJAnal. Chem., 363 (1999) 696.
27 J.A. Koziel and J. Pawliszyn, J. Air Waste Manage. Assoc., 51 (2001) 173.
28 F. Augusto, J.A. Koziel and J. Pawliszyn, Anal. Chem., 73 (2001) 481.
29 P.A. Martos and J. Pawliszyn, Anal. Chem., 71 (1999) 1513.
30 J. Koziel, J. Noah and J. Pawliszyn, Environ. Sci. Technol., 35 (2001) 1481.
31 M. Jia, J.A. Koziel and J. Pawliszyn, Field Anal. Chem. Technol., 4 (2000) 73.
32 J. Pawliszyn, Solid Phase Microextraction: Theory and Practice. Wiley-VCH, New
York, NY, 1997.
28
33 V.M. Brown, D.R. Crump and D. Gardiner, Environ. Technol., 14 (1992) 367.
34 R.H. Brown and H.J. Saunders (Eds.), Diffusive Sampling. Royal Society for
Chemistry, London, 1987.
35 D. Crump, in T. Salthammer (Ed.), Organic Indoor Air Pollutants: Occurrence,
Measurement, Evaluation.Wiley-VCH, Weinheim, Germany, 1999 Ch. 1.5, p. 57.
36 ECA, Sampling Strategies for volatile organic compounds (VOCs). European
collaborative action-Indoor Air Quality and Its Impact on Man, Report No 14,
European Commission, EUR 16051 EN, Luxemburg, 1994.
37 M.A. Cohen, P.B. Ryan, Y. Yanagisawa and S. Hammond, J. Air Waste Manage.
Assoc., 40 (1990) 993.
38 R. Otson, P. Fellin and Q. Tran, Atmos. Environ., 28 (1994) 3563.
39 K. Levsen, E. Schimming, J. Angerer, H.E. Wickman and J. Heinrich, Proceedingsof
IndoorAir 96, July 1996, Nagoya, Japan, 1996, Vol. 1, p. 1061.
40 C. Krause, R. Nael, C. Schultz, B. Seifert and D. Ulrich, Proceedingsof IndoorAir 87,
August 1987, Berlin, Germany, 1987, Vol. 1, p. 102.
41 D. Cavallo, D. Alcini, P. Carrer, A. Basso, D. Bollini, L. Lovato, F. Vercelli, F. Visigalli
and M. Marconi, Proceedings of the Healthy Buildings/IAQ 97, September/October
1997, Washington, DC, 1997, Vol. 3, p. 141.
42 V.M. Brown, A.H. Cockram, D.R. Crump and H.S. Mann, Proceedings of IndoorAir
96, July 1996, Nagoya, Japan, 1996, vol. 2, p. 115.
43 A. Rosencwaig, in: P.J. Elving, J.D. Winefordner, I.M. Kolthoff (Eds.), Chemical
Analysis. Vol. 57, Chichester, UK, 1980.
44 L.E. Ekberg, in T. Salthammer (Ed.) Organic Indoor Air Pollutants: Occurrence,
Measurement, Evaluation. Wiley-VCH, Weinheim, Germany, 1999, Ch. 1.6, p. 73.
45 J.B. Hicks, E.D. Winegar and C.S. Sims, Proceedingsof IAQ '92: Environments and
People, San Francisco, 1992. American Society of Heating Refrigerating and
Air-Conditioning Engineers, Inc. Atlanta, GA, 1992, p. 251.
46 K.A. Bunding Lee, A.L. Clobes, A.L. Hood, J.A., Schroeder and L. Hawkins, Proc-
eedings of Indoor Air '93, 1993, Helsinki, Finland, 1993, Vol. 2, p. 245.
47 T. Godish, T.W. Zollinger and V. Konopinski, J. Environ. Health., 53 (1990) 34.
48 K-S. Liu, F-Y. Huang and S.B. Hayes, Proceedings of the 4th Int. Conf. Indoor Air
Quality and Climate, 1987, Berlin, Germany, 1987, Vol. 2, p. 610.
49 A. Vairamurthy, J.M. Roberts and L. Newman, Atmos. Environ., 26A (1992) 1965.
50 M. Schultz and T. Salthammer, in: T. Salthammer (Ed.), OrganicIndoorAir Pollut-
ants: Occurrence, Measurement,Evaluation.Wiley-VCH, Weinheim, Germany, 1999,
Ch. 1.2, p. 15.
51 T. Salthammer, JPhotochem. Photobiol. A., 74 (1993) 195.
52 P. Martos and J. Pawliszyn, Anal. Chem., 70 (1998) 2311.
53 E. Uhde and A. Borgschulte, Proceedings of the 8th International Conference on
Indoor Air and Climate. Edinburgh, Scotland, 1999.
54 R.G. Lewis, in J.S. Spengler, J.M. Samet, J.F. McCarthy (Eds.), Indoor Air Quality
Handbook. McGraw-Hill, New York, NY, 2000, Ch. 35, p. 35.1.
55 R.G. Lewis and S.M. Gordon, in: L.H. Keith (Ed.), Principles of Environmental
Sampling. ACS Professional Reference Book. American Chemical Society,
Washington, DC, 2nd ed. 1996, p. 401.
56 American Society for Testing and Materials (ASTM), D4861: Standard practice for
sampling and selection of analytical techniques for pesticides and polychlorinated
biphenyls in air. In: Annual Book of ASTM Standards. West Conshohocken, PA,
American Society for Testing and Materials, Vol. 11.03, 2000.
57 American Society for Testing and Materials (ASTM), D4947: Standard test method
for chlordane and pentachlor residues in indoor air. In: Annual Book of ASTM
Standards. West Conshohocken, PA, American Society for Testing and Materials,
29
Vol.11.03, 2000.
58 R. Rudel, in: J.S. Spengler, J.M. Samet and J.F. McCarthy (Eds.), IndoorAir Quality
Handbook. McGraw-Hill, New York, NY, 2000, Ch. 34, p. 34.1.
59 R.A. Rudel, P.W. Geno, G. Sun, A. Yau, J.D. Spengler, J. Vallarino and J.G. Brody, J.
Air Waste Manage. Assoc., 51 (2001) 499.
60 K. Ogan and E. Katz, Anal. Chem., 53 (1981) 160.
61 L. Wallace, J. Air Waste Manage. Assoc., 46 (1996) 98.
62 P. Koutrakis, J.M. Wolfson, J.L. Slater, M. Brauer and J.D. Spengler, Environ. Sci.
Technol., 22 (1988) 1463.
63 S.K. Poruthoor, P.K. Dasgupta and Z. Genfa, Environ. Sci. Technol., 29 (1995) 1534.
64 C. Howard-Reed, A.W. Rea, M.J. Zufall, J.M. Burke, R.W. Williams, J.C. Suggs, L.S.
Sheldon, D. Walsh and R. Kwok, J. Air Waste Manage. Assoc., 50 (2000) 1125.
65 W.W. Nazaroff and A.V. Nero (Eds.), Measurement Techniques, Radon and Its
Productsin IndoorAir. Wiley Interscience, New York, NY, 1988.
66 A.V. Nero, M.B. Schwehr, W.W. Nazaroff and K.L. Rezan, Science, 234 (1986) 992.
67 H.F. Lukas, Rev. Sci. Instrum., 28 (1957) 680.
68 B.G. Cartwright, E.K. Shirk and P.B. Price, Nuclear Inst. Meth., 153 (1978) 45.
69 E. Stranden, A.K. Kolstad and B. Lind, Radiat. Prot.Dosim., 5 (1983) 241.
70 P. Chittaporn, M. Eisenbud and N.H. Harley, Health Physics, 41 (1981) 405.
71 P. Kotrappa, J. Bigu, Proceedings of 38th Annual Meeting of the Health Physics
Society, Atlanta, GA, July 1993.
72 W.E. Simon and M.B. Schell, Proceedings of the 1990 International Symposium on
Radon and Radon Reduction Technology, U.S. EPA, EPA/600/ 9-90, 1990.
73 P.J. Landrigan, New Engl. J. Med., 338 (1998) 1618.
74 W.N. Rom, in: Rom (Ed.), Environmental and Occupational Medicine. Lippincott-
Raven Publishers, Philadelphia, PA, 1998, p. 349.
75 A.V. Samudra, C.F. Harwood and J.D. Stockham, Electron microscope measurement
of airborne asbestos concentrations. Research Triangle Park, NC. US EPA, EPA
Report No. 60012-77-178-Rev., PB 285945, 1978.
76 B.T. Commings, The Significance of Asbestos and Other Mineral Fibers in Environ-
mental Ambient Air. Commings Assoc. Maidenhead, UK, 1985.
77 M. Odziemkowski, J.A. Koziel, D. Irish and J. Pawliszyn, Anal. Chem., 73 (2001)
3131.
78 U.S. Environmental Protection Agency (EPA), Air quality criteria for ozone and
other photochemical oxidants, Report No EPA-600/8-84-020F, Washington, DC,
1986.
79 American Society for Testing and Materials (ASTM), Atmospheric Analysis; Occupa-
tional Health and Safety; Protective Clothing. Annual Book of ASTM Standards.
West Conshohocken, PA, 1999, Vol. 11.03.
80 P. McGheehin, P.T. Mosely and D.E. Williams, Sensor Review, 1 (1994) 13.
81 World Health Organization (WHO), Air Quality Guidelines for Europe. WHO
Regional Publications, European Series No. 23, WHO Regional Office for Europe,
Copenhagen Denmark, 1987.
82 W.A. Wade, W.A. Cote and J.E. Yocom, J. Air Pollut. Control Assoc., 25 (1975) 933.
83 E.D. Palmes, A.F. Gunnison, J. Di Mattio and C. Tomczyk, Am. Ind. Hyg. Assoc. J.,
37 (1976) 570.
84 Y. Yagisawa and H. Nashimura, Environ. Int., 8 (1982) 235.
85 J.E, Yocom and S.M. McCarthy, Measuring Indoor Air Quality: A Practical Guide.
Wiley-VCH, Chichester, UK, 1991.
86 J.P. Lodge, Methods of Air Sampling and Analysis. Lewis Publishers, Boca Raton,
FL, 1988.
87 E.E Rickman, A.H. Groen, R.S. Wright and J.E. Sickles, Laboratory and Field
30
Evaluation of Extrasensitive SO2 and NO2 Analyzers for Acid Deposition Monitoring,
U.S. EPA, Quality Assurance Division, Research Triangle Park, NC, RTI/3999/
18-04F, Draft, May 1989.
88 P.B. Leaderer, J.A.J. Stolwijk, R.T. Zagraniski and M. Quin-Shan, Proceedingsof the
77th Ann. Meeting Air Poll. Control Assoc., June 1984, San Francisco, CA, 1984,
Paper No 84-33.3.
89 K.P. Cooper and R.R. Alberti, Am. Rev. Resp. Dis., 129 (1984) 629.
90 U.S. Environmental Protection Agency, Code of Federal Regulations, Title 40, Part
60, Appendix A, Washington, DC, Office of Federal Register, July 1, 1987.
91 U.S. Environmental Protection Agency, Standards of Performance of New Station-
ary Sources. Compilation, EPA-340/1-77-015, 1987.
92 S. Buckestein, K.A. Tucker and P.R. Gifford, Anal. Chem., 52 (1980) 2396.
93 R.K. Stevens, J.D. Mulik, A.E. O'Keefe and K.J. Frost, Anal. Chem., 43 (1971) 827.
94 J.H. Seinfeld, Atmospheric Chemistry and Physics of Air Pollution. John Wiley &
Sons, New York, NY, 1986.
95 U.S. Department of Health and Human Services (USDHHS), A Report of the
Surgeon General: The Health Consequences of Smoking - Chromic and Obstructive
Lung Disease, Washington, DC, U.S. Government Printing Office, 1989.
96 M.R. Guerin, R.A. Jenkins and B.A. Tomkins, in: Center for Indoor Air Research
(Ed.), The Chemistry of Environmental Tobacco Smoke: Composition and Measure-
ment. Lewis Publishers, Chelsea, MI, 1992.
97 C.L. Benner, J.M. Bayona, F.M. Caka, H. Tang, L.D. Lewis and D.J. Eatough,
Environ. Sci. Technol., 23 (1989) 688.
98 California Environmental Protection Agency (Cal EPA), Health Effects of Exposure
to Environmental Tobacco Smoke, Office of Environmental Health Hazard Assess-
ment, 1997.
99 U.S. Environmental Protection Agency (USEPA), Respiratory Health Effects of
Passive Smoking: Lung Cancer and Other Disorders, USEPA-600-006F,
Washington, DC, U.S. Government Printing Office, 1992.
100 National Research Council (NRC), Environmental Tobacco Smoke: Measuring Expo-
sures and Assessing Health Effects, Committee on Passive Smoking, National
Academy Press, Washington, DC, 1986.
101 J.D. Spengler, D.W. Dockery, W.A. Turner, J.M. Wolfson and B.G. Ferris, Atmos.
Environ., 15 (1981) 23.
102 F.W. Henderson, H.F. Reid, R. Morris, O.L. Wang, P.C., Hu, R.W. Helms, L.
Forehand, J. Mumford, J. Lewtas, N.J. Haley and K.S. Hammond, Am. Rev. Resp.
Dis., 140 (1989) 183.
103 K.S. Hammond, Environmental Health Persp., 107 (1999) 329.
104 C.S. Cox and C.M. Wathes (Eds.), Bioaerosols Handbook. Lewis Publishers, Boca
Raton, F1, 1995.
105 T.A. Myatt and D.K. Milton, in: J.S. Spengler, J.M. Samet and J.F. McCarthy (Eds.),
IndoorAir Quality Handbook. McGraw-Hill, New York, NY, 2000, Ch. 42, p. 42.1.
106 N.L. Sprince, P.S. Thorne, W. Poppendorf, C. Zwerling, E.R. Miller and J.A.
DeKostner, Am. J. Indust. Med., 31 (1997) 403.
107 D.K. Milton, R.J. Gere, H.A. Feldman and I.A. Greaves, Am. Indust. Hyg. Assoc. J.,
51 (1990) 331.
108 M. Walters, D.K. Milton, L. Larsson and T. Ford,Appl. Environ. Microbiol., 60 (1994)
996.
109 Z. Mielniczuk, E. Mielniczuk and L. Larson, J. Microbiol. Meth., 17 (1993) 91.
110 T.A.E. Platts-Mills, in: J.S. Spengler, J.M. Samet and J.F. McCarthy (Eds.), Indoor
Air Quality Handbook. McGraw-Hill, New York, NY, 2000, Ch. 43, p. 43.1.
111 F. De Blay, P.W. Heymann, M.D. Chapman and T.A.E. Platts-Mills, J. Allergy Clin.
31
Immunol., 88 (1992) 919.
112 M.D. Chapman and T.A.E. Platts-Mills, J. Immunol., 125 (1980) 587.
113 S. DeLucca, R. Sporik, T. O'Meara and E.R. Tovey, J. Allergy Clin. Immunol., 102
(1999) 174.
114 D.L. Rosenstreich, P. Eggleston, M. Kattan, D. Baker, R.G. Slavin, P. Gergen, H.
Mitchel, K. McNiff-Mortimer, H. Lynn, D. Ownby and F. Malveaux, New Engl. J.
Med., 336 (1997) 1356.
115 C.M. Luczynska, Y. Li, M.D. Chapman and T.A.E. Platts-Mills, Am. Rev. Resp. Dis.,
141 (1990) 361.
116 A. Custovic, R. Green, A. Fletcher, A. Smith, C.A. Pickering, M.D. Chapman and A.A.
Woodstock, Am. J. Resp. Crit. CareMed., 155 (1997) 94.
117 M.S. Muilenberg, in: J.S. Spengler, J.M. Samet and J.F. McCarthy (Eds.), Indoor Air
Quality Handbook. McGraw-Hill, New York, NY, 2000, Ch. 44, p. 44.1.
118 D.A. Sterling and R.D. Lewis, Ann. Allergy Asthma Immunol., 80 (1998) 279.
119 R. Yankova, Grana, 30 (1991) 171.
120 M.D. Lebowitz, M.K. O'Rourke, R. Dodge, G.C. Holberg, R.W. Hoshaw, J.L. Pinnas,
R.A. Barbee and M.R. Sneller, Environ. Int., 8 (1982) 375.
121 P.D. Moore, J.A. Webb and M.E. Collinson, Pollen Analysis. Blackwell Scientific
Publications, Oxford, Boston, 1991, Ch. 4.
122 Y. Takahashi, T. Nagoya, M. Watanabe, S. Sakaguchi and S.A. Katagiri, Allergy, 48
(1993) 94.
123 H.A. Burge, in: J.S. Spengler, J.M. Samet and J.F. McCarthy (Eds.), Indoor Air
Quality Handbook. McGraw-Hill, New York, NY, 2000, Ch. 45, p. 45.1.
124 M.P. Buttner and L.D. Stetzenbach, Appl. Environ. Microbiol., 59 (1993) 219.
125 P.M. Scott, in AOAC Official Methods of Analysis, 1990, Ch. 49, p. 1185.
126 J.L. Lange, P.S. Thorne and N.L. Lynch, Appl. Environ. Microbiol., 63 (1997) 1557.
127 American Society for Testing and Materials (ASTM), D5952-96: Standard Guide for
Inspecting Water Systems for Legionellae and Investigating Possible Outbreaks of
Legionellosis (Legionnaires' disease or Pontiac fever). In: Annual Book of ASTM
Standards,ASTM Committee on Standards.ASTM, West Conshohocken, PA, 1997,
Vol. 11.03.
128 S.K. Poruthoor and P.K. Dasgupta, Environ. Sci. Technol., 32 (1998) 1147.
32
Chapter2
2.1 INTRODUCTION
The sampling strategy should specify the types of samples to be collected. The
type of sample essentially depends on the character of water or wastewater, flow
variability of water and wastewater streams and required accuracy of the mea-
surements. The most common liquid samples are [3,4]: discrete samples, simple
composite samples, flow proportional composite samples, and sequential com-
posite samples.
A discrete sample is collected separately in its own individual container. It
represents the source at the time the sample was taken and is appropriate if
conditions at the source are essentially constant. The discrete samples are
collected for determination, e.g., pH, total organic carbon or dissolved oxygen.
A simple composite sample, also known as a time proportional composite
sample, comprises a series of identical volume samples (Vc) taken at constant
time intervals (tc) and deposited into the same container. It is defined by the
product of tcVc. In sampling a lake or tank waters a space proportional sample
can be also distinguished.
A flow proportional composite sample is collected depending on the water
flow during the sampling in such a way that it represents the average conditions
of sampling. It can be realized in two ways: (a) the sampler is triggered by a
signal from stream recorder, intervals between sampling of equal volume sub-
samples are changed depending on stream flow; (b) by changing the volume of
each sample depending on stream flow when the intervals between the sequence
operations are constant.
A sequential composite sample is formed from a series of individual samples
(typically from 2 to 8) placed in one container; each container represents a
determined interval (usually 1 h). This way of sampling is useful when the
stream of water or wastewater is not constant (when discharge of wastewater is
expected or self-purification conditions take place). In such conditions, for
example, high or small pH values would be difficult to determine in simple
composite sample.
Representative and appropriate samples of water or wastewater may be
collected in different ways. The general principle of an elementary selection of
sampling technique is given in Table 2.1 [4] and possible types of water sampling
are presented in Fig. 2.1 [5].
The discrete samples are usually taken from pipelines, reservoirs, rivers,
streams and sewers. The discrete sample can be collected either manually or by
using automatic samplers. Each sample should be a source of information on
water quality only at the time when it was sampled.
The composite sample may be obtained by manual mixing of individual
samples or by using an automatic system for sampling. Composite samples
represent the average concentration and constituent loads of the source during
the longer time. However, in most cases, abrupt changes of analyte concentra-
tions may not be noticed in the analysis of such samples. The collection of time
34
TABLE 2.1
Principles of preselection of sampling technique [3]
Rate of fluctuation of Small changes of stream flow Large changes of stream flow
analyte concentration
Small Discrete sample Discrete sample
Large Time proportional composite Volume proportional composite
sample samples
a) v/ , -1 -C)
f t
b) v d) v
111111]1111111111111111. t
iiilll iiiilllili II -
t
Fig. 2.1. Possible types of water sampling techniques: (a) discharge, (b) time dependent
sampling, (c) flow dependent sampling, (d) volume dependent sampling (based on Ref. [5]).
Due to the great diversity of possible pollutants that can be present in water, a
full analysis of water is not usually performed. Usually water is analysed in order
to evaluate its usefulness for a particular purpose and only specific and harmful
components are determined. The results obtained allow selection of the proper
purification methods or they are used for checking if a process was conducted
properly or if the water fulfils the requirements. Full descriptions of water sam-
pling and analytical methods may be found elsewhere [8-10].
The sample volume depends mainly on the substances to be analyzed and
their expected concentrations but 1 to 5 1 samples are commonly collected [11].
35
When developing a sampling strategy it is important to consider not only the
nature of water sample. The composition of the water in terms of the dissolved
compounds and trace fractions associated with colloids and suspended solids,
and the temporal and spacial variability at the sampling site should be also
considered. Furthermore, the chemical properties of the determinants and the
expected concentrations in the samples also influence the choice of sampling
methodology.
Water samples should be taken to completely sealed bottles, carefully rinsed
with sampling water. Any bubbles of air should be removed under the closing
system of the sampling vessels, otherwise, during transport or sample storage,
volatile components may pass into the headspace leading to their losses. In turn,
some components of gas phase may be dissolved in liquid phase. Both processes
may proceed up to reaching the equilibrium, determined by value of partition
coefficients of each analyte.
Care must be taken in all methods of collecting samples to ensure that
neither the sample storage containers nor any collecting vessels contaminate or
alter the sample. This can occur by:
- leaching of contaminants from the container surfaces
- leaching of chemicals (organic and inorganic) from container materials
- adsorption of trace metals or chemical compounds onto container surfaces
- reaction of the sample with container material
- change in equilibrium between pollutants present in particulate and
solution phases.
36
a) b) 6 ,N1l
Left: Fig. 2.2. Bottle for sampling from different depths. 1: Tag line, 2: spring, 3: plastic tube, 4:
rubber stopper, 5: rubber stopper, 6: 1-1 plastic bottle with broad neck, 7: plastic tube, 8: brass
belt bracket (based on Refs. [11-13]).
Right: Fig. 2.3. (a) Collector for sampling from optional depths. 1: Tag line for lowering container
for determined depth, 2: rope for container opening, 3: metal reservoir for sample, 4: lead
basement. (b) Ruttner's sampler. 1: Plexiglass cylinder, 2: metallic lid connected to a rod, 3:
second lid, 4: spring catch on a barrel, 5: thermometer, 6: sinker, 7: tap water (based on Refs.
[11-13]).
bottles and transported to the laboratory. Figures 2.2 and 2.3 give examples of
sampler configuration for sampling from different depths [11-13].
When gases such as 02, CO2 or H2S dissolved in water are determined, a
problem can arise if water flowing into the bottle is mixed with the air inside the
empty bottle. The impossibility of rinsing the bottle with sampling water before
the final filling may generate additional errors. The sampler shown in Fig. 2.4
eliminates such inconveniences [14]. It can be deployed to a depth of up to 40 m.
The system consists of three connected bottles inserted in a lead holder and
closed with polished cork. The first bottle is connected to the atmosphere by a
polyethylene tube. A 40 m nylon tube linked to the third bottle is used for:
- lowering the sampler to the required depth,
- rinsing the system with gas during lowering, e.g., with nitrogen from a
pressurized tank, to avoid filling the bottle at the wrong depth,
Fig. 2.4. Sampler for collecting water from depths of up to 40 m. 1: Hold-fast system to regulate
force of bottle tightening, 2: spring, 3: lead sinker (based on Ref. [141).
37
- connecting the sampler to the atmosphere during filling,
- preventing the losses of sample by tight closing after the sampling but before
its withdrawing to the surface.
Purging with gas stops when the sampler reaches the required level and the end
of the nylon tube above the water surface is open. The first bottle is filled, fol-
lowed by the two others. In this way, the first bottle is rinsed with a water vol-
ume equal to the volume of the other two bottles. The content of the first bottle is
used for determination of gaseous components. Samples from the remaining bot-
tles can be used for different chemical and biological analyses.
The modification of the original Bodman's sampler, called Bodega-
Bodman's sampler [15] (Fig. 2.5), made of high purity aluminium, stainless
steel, Teflon and Viton allows the collection of samples of up 90 litres through
the whole column of water, reducing sample contamination to minimum. An
easily dismantled swagelock type valve at the lid of the sampler for connecting
the purging nitrogen line from the cylinder and stainless steel ball valve placed
at the bottom enables the transfer of the sampler content to an aseptic room for
further treatment without contact with the atmosphere. The outlet valve at the
centre of the bottom lid reduces the possibility of losses of greater, easily sedi-
mentated particulate fraction. A 50 kg sinker, in the form of lead disks, is
attached to the bottom of the sampler. A manual sampler made of transparent
acryl may be also used for the same purpose [16].
Figure 2.6 shows a multichannel (1-4 channels) sampler for collecting sam-
ples for the determination of volatile organic compounds [17]. Each syringe has
.I ra
-U.. 1 I I - I I 1
i2
I1 -
II <2
r___ ------
- 2_ __
c---
- - ---
__
_:_ -
<-3
--- ___ __
I
-~A dK' In -4
-
_ f
/4 p#iajr---1PR
-j.1 _L~-
Left: Fig. 2.5. Bodega-Bodman's sampler: (a) open, (b) closed. 1: Upper construction plate, 2:
upper lid, 3: pressure dropping valve, 4: holding rods, 5: cylinder collar, 6: stainless steel leading
rods, 7: sampler cylinder, 8: 12 kHz pulsator, 9: bottom collar, 10: teflon o-ring, 11: bottom
construction plate, 12: ball valve, 13: ratchet mechanism, 14: magnet of pulsator switcher (based
on Ref. [15]).
Right: Fig. 2.6. Schematic representation of multichannel syringe sampler of water. 1: Syringes,
2: screw for regulation of water volume, 3: circular cam, 4: DC motor (based on Ref. [171).
38
to shin
·-;
-2
Left: Fig. 2.7. Modified Slocum's sampler for collecting and in situ preconcentration of organic
analytes from water. 1: Teflon stopper (removed under the water surface), 2: steel ropes, 3:
holder of stainless steel filter, 4: reduction connector, 5: supporting frame (steel rods), 6: inlet
tube with blades for directing inlet according to water stream, 7: swagelock-type connection, 8:
connector to sorption tube, 9: teflon column filled with high density polyurethane foam, 10:
nylon collar, 11: column connector to sorbent, 12: one-way valve, 13: connection to tube thread,
14: steel catch, 15: steel catch belts, 16: vacuum rubber tube leading to surface (based on Ref.
[151).
Right: Fig. 2.8. Equipment for water sampling by the "drop-bottle" method. 1: Displacement
buoy, 2: sea level, 3: frame holder of glass bottle, 4: lead sinker (based on Refs. [18,19]).
its own sampling line (Teflon tubes) equipped with a three-way electromagnetic
valve that enables the sampled water to be delivered to a container with a
floating lid. This construction eliminates the headspace above the liquid surface
and avoids the losses connected with partition of organic analytes between
aqueous and gas phases. The use of Teflon and glass as the construction
materials minimizes losses of volatile components by reducing the surface
adsorption. The action of the sampler is controlled by computer using radio-
telemetry.
Frequently, the amount of particulate fraction in water samples is equally
important as compounds dissolved in water. Modified Slocum's sampler [15],
shown in Fig. 2.7, enables independent collection of suspended particulates on
stainless steel filter inserted in the sampler holder and preconcentration of
dissolved organic compounds on high density polyurethane foam, placed in
Teflon adsorption tube. The total weight of the sampler is about 23 kg.
39
Left: Fig. 2.9. DHI sampler (Deutsche Hydrographisches Institut) for sampling deep-water. 1:
Glass tubes, 2: tight closure, 3: movable metallic arm for breaking glass tubes, 4: glass containers
(based on Refs. [18,20]).
Right: Fig. 2.10. Schematic representation of sampler for collecting water from depths of up 3 m.
1: Operational handle, 2: casing of operational rod, 3: holding belts, 4: rod, 5: spring catch, 6:
polyethylene bag, 7: hold-fast element, 8: sampling bottle, 9: grip, 10: deep indicator, 11:
operational rod (based on Ref. [21]).
Water samples from open water bodies from depths of 0.5-1.5 m may be
collected using the "drop-bottle" system [18,19]. The equipment for such samp-
ling is presented in Fig. 2.8. For sampling from deeper layers, bathometers
elaborated by Stadler, commonly known as DHI samplers (Fig. 2.9) are
frequently used [18,20]. The samples are collected to exchangeable glass flasks,
which can be washed and protected against contamination by hermetic closing.
The glass tubes that protect the holes are broken at the required depths by
moving a lever. Extraction and separation of mixtures are performed in the same
flask, which eliminates the need for liquid pouring and a separatory funnel, thus
reducing the possibility of sample contamination. Thick-walled flasks are used
for collecting the samples from depths of up several thousands meters.
Another type of sampler for determination of volatile organic compounds is
shown in Fig. 2.10 [21]. It is essentially made of stainless steel and aluminium
and partly of brass and polypropylene. During the sampling an empty bottle (100
ml capacity) closed by a polypropylene stopper, and provided with an injection
40
a) b)
r---q
Fig. 2.11. (a) Czapla-1 sampler; 1: Grip, 2: telescoping shaft, 3: striction system, 4: o-ring, 5: body,
6: screw, 7: dipper. (b) Nurnik-1 sampler; 1: regulation of lock force, 2: lock, 3: lock, spring catch,
4: striction system, 5: bearing ring, 6: sinker, 7: dipper (based on Ref. [22]).
41
2.4.2 Sampling of groundwater from small boreholes
Groundwater is sampled to determine its quality and the quality changes with
time [23-25]. Equipment for this purpose was developed for the U.S. Geological
Survey. Samplers for collecting groundwater must fulfil several specific de-
mands [26]:
- the sampler must enable the collection of samples from high depths,
- it must reach the bottom of the drilling hole and collect the sample,
- the material of which the sampler is made must not contaminate the water
sample,
- the sampler must be rinsed with water before collecting the sample,
- the sampler must be capable of being closed immediately after finishing
sampling in order to avoid losses of volatile compounds and protect the
sample from contact with air,
- the sampler must be able to avoid the degassing of water samples by sudden
changes of pressure during the sampling,
- the sampler must be easy to transport to the laboratory.
Therefore, during the designing of such samplers the following requirements
should be considered [26]:
- a small diameter drilling hole (ca. 5 cm),
- the inertness of material for the sampler (stainless steel or Teflon),
- the ability to collect samples from a depth of about 60 meters,
- the container for rinsing water should be three times larger than the
container for the sample,
- the system should be light, portable and easy to operate,
- minimizing the usage of pressure-operated elements (pressurized tank
containers and pumps for gases should be eliminated).
Figures 2.12 and 2.13 show schemes of eight samplers for collecting the ground-
water from drilling holes [26]. In first four samplers, the piston is used to pump
water through the cylinder for samples.
The Geological Survey of Canada developed a sampler for collecting samples
from very small diameter holes drilled by diamond drills [27]. The sampler
shown in Fig. 2.14, made of stainless steel and equipped with a portable travel-
ling block, enables the collection of samples of about 350 ml from small holes at
any depth. A steel rope passes through the central pipe of the sampler and is
fastened to its bottom. Water flows through the sampler during its lowering into
a drill hole thanks to spring-operated jaws installed in holes at the top of the
sampler. After achieving the required depth, a so-called messenger is dropped at
the bottom to a hole along the steel rope. At that moment the jaws are opened
and a sampling tube is liberated which can drop onto a rubber stopper at the
bottom of sampler. Although the top of the sampler is still open, the mixing of
water is minimized during the withdrawal of the sampler. A properly designed
sampler enables the collection of samples from drilling holes that are at angles of
up to 45° .
42
a) 1 b) c) d)
1 1 1
g2 .2 2 2
3 3
3
3 4
4
5
5
4
~4~ ~6 6
5 1 5 I 7 7
6 8 8
7 9 59
7
Fig. 2.12. Schematic representation of the construction of four samplers for collecting water
from drilling holes. (a) Sampler with coil driven by piston; 1: coiled cable, 2: returning spring, 3:
piston, 4: valve, 5: sample container, 6: valve, 7: controlling valve. (b) Sampler driven by manual
pump; 1: hydraulic pressure, returning spring, 3: piston, 4: valve, 5: sample reservoir, 6: valve, 7:
controlling valve. (c) Sampler with two tubes; 1: sampling tube, 2: tube of pump, 3: double piston,
4: returning spring, 5: controlling valve, 6: valve, 7: sample reservoir, 8: valve, 9: controlling
valve. (d) One-way sampler driven by manual pump; 1: tube for sampling and pump, 2:
controlling valve, 3: double piston, 4: controlling valve, 5: returning spring, 6: valve, 7: sample
container, 8: controlling valve (based on Ref. [26]).
a) b) c) d) 1
2
22
3
~2 ~ ~
3 4
4 5
6
4
Fig. 2.13. Schematic representation of the construction of four samplers for collecting water
from drilling holes. (a) Sampler driven by electric motor; 1: electric motor in sampler head, 2:
thread, 3: piston, 4: sample reservoir, 5: valve, 6: controlling valve. (b) Sampler with electric gear
wheel pump; 1: gear wheel pump, 2: pressure controlling valve, 3: valve, 4: sample container, 5:
valve, 6: controlling valve. (c) Sampler driven by motor; 1: electric motor in sampler head, 2:
thread, 3: upper piston, 4: membrane, 5: cylinder, 6: bottom piston. (d) Manual sampler; 1: cable,
2: opening system, 3: piston sampler, 4: piston, 5: valve, 6: piston, 7: sample reservoir, 8: valve, 9:
controlling valve (based on Ref. [26]).
43
a) b)
2,5 cm
membrane
=Z.
E
M
nn
C
/1
W/2
r•
3
3
:4
U
I P 5~
5t_ 5
Left: Fig. 2.14. Schematic representation of the construction of a sampler for collecting a sample
from small diameter holes. 1: Casing, 2: sampler tube, 3: rubber closing cap, 4: spring-operated
jaws (based on Ref. [271).
Right: Fig. 2.15. (a) Sampler for in situ collecting of water samples from drilling holes; A: tip
filter, B: disposable double hypodermic needle, C: sterilized, pre-evacuated glass ampoule. (b)
Schematic representation of a sampler for sampling and simultaneous enrichment of analytes.
The flow of water through sorption tubes is controlled by a reversing valve, needle of syringe,
placed at the outflow of each sorption tube. 1: Steel tube (to surface) of 0.45 cm diameter, 2:
swagelock-type connectors, 3: brass sinker, 4: reversing valve with needle, 5: sorption tubes, 6
stream of water from drilling hole (based on Ref. [28]).
44
a)
The quality of groundwater can easily be spoiled by polluted surface water leach-
ing into deeper aquifers often used as sources of high-quality drinking water.
The leaching water may be contaminated by, e.g., pesticides used in agriculture,
migration of pollutants from landfills or from drainage. Such seepage waters
form an unsaturated water zone. A knowledge of the chemistry of waters con-
45
D vacuum vacuum
Fig. 2.17. Schematic representation of a system for collecting water using a lysimeter for
determination of unsaturated water zones. Numbers in circles from 1 to 10 indicate a particular
route of the sample in a system; 1: 5-cm3 syringe, 2: silicon tube, 3: three-way valve, 4: 5-cm3
ampoule, 5: scrubber with porous partition, 6: sorption tube, 7: burette (based on Ref. [31]).
The surface microlayer is the part of each water body in which a large number of
contaminants, including insoluble organics and metals are normally concen-
trated. Representative sampling of this layer is rather difficult because the film
is only some 100 m thick. Therefore, a great deal of research is focused on devel-
oping efficient methods of sampling the surface microfilms, frequently thin oily
films, concentrated in organic compounds, both polar and non-polar [32-35].
Organic films on the surface of seawater, biogenic and anthropogenic in
origin, play an important role in seashore waters. Such films affect the physical
and chemical processes taking place in surface seawaters. Special sampling
46
l
C .' 1
Fig. 2.18. (a) A dip-type sampler with double walls for sampling the surface layers of water. (b) A
section of sampler against A-B line, 1: rectangular plates, 2: cutoff frame, 3: Langmuir's channel
section, 4: glass slide, 5: surface with microfilm, 6: Wilhelm's plate for surface tension measure-
ment, 7: thread, 8: valves, F: connection with torsion balance. The water level in the sampler, C:
during detaching, D: in stand-by position, E: before measurement by using Langmuir's channel
(based on Ref. [36]).
devices are required for collecting of such films. Figure 2.18 presents a scheme of
a submersible rectangular sampler with double walls which cuts off the surface
microlayer from the adjacent subsurface water layer [36]. The sampler consists
of a rectangular plate (40-50 cm) of 6 cm depth made of plastic material with a
shallow Langmuir channel section (30x40x0.7 cm) at the lower part. However,
the so-called Garret's sampler, a metal screen (50x65 cm), is more frequently
used for collecting microfilms of water. It was found that the sampler collected a
surface layer of 0.3 mm thickness [17,37]. The equipment for collecting samples
of the surface microlayer by Garret's method is shown in Fig. 2.19.
Another type of sampler, designed at the Chemical Faculty of the Technical
University of Gdansk, for collecting the surface microlayer is presented in Fig.
2.20. The sampler allows the determination of the oil film thickness on the
surface of water. The water sample is taken from cylinder (1) to glass tube (5)
with a diameter several or more times smaller than the cylinder. The thickness
of the collected film can be measured very simply, depending only on the
diameter of the exchangeable cylinder. Film thickness can be calculated on the
basis of the distance between the pointers (9). The ball of the closing valve (4) is
lifted to about half the height of the glass tube using rope (12). Next, the sampler
is placed in the sampling site perpendicularly to the water surface and immersed
deeply enough that the buoys (7) cause the sampler to float. When the rope slack,
a special spring mechanism prepares to action the rope with the ball attached to
it. After the second tightening of the rope (12) for withdrawing the sampler, the
spring mechanism is liberated, the rope is freed, the ball is lowered and the
cutting-off socket is closed.
47
1
Left: Fig. 2.19. Sampler for collecting microfilms of water by Garret's method. 1: Garret's
sampler metal gauze, 2: gauze holder, 3: bottle for the sample (based on Refs. [17,37]).
Right: Fig. 2.20. Sampler for collecting a surface microlayer. 1: Exchangeable cylinder, 2: cut-off
sockets, 3: metal o-ring (hung on cut-off rope), 4: ball, 5: glass tube, 6: guarding and mounting
tube, 7: buoys, 8: scale, 9: scale of oil level, 10: lid with special handle, 11: cutoff ball and device
suspension, 12: line.
The sampling of rainwater needs special care and equipment. The simplest sam-
plers for this purpose [38-40] are open containers with known surface, that in
stand-by state are closed with lids. Samples collected in such samplers are used
for typical physicochemical analysis. Samplers of a more complex constructions
are used for the determination of dissolved organic compounds in water; such a
sampler is shown in Fig. 2.21 [29], consisting of the following elements:
- a collecting surface (0.81 m2 gives an 8.1-1 sample for 1 cm rainwater); it is a
specially profiled Teflon plate with a hole in its centre,
- a glass container to which two stainless steel electrodes are installed at
different heights on the side wall, acting as sensors of the water level inside
the container,
- a unit for isolation/preconcentration of organic compounds (adsorption
tubes) from collected water. The flow of rainwater through the adsorption
system is forced by peristaltic pump. The pump is switched on and off by a
signal from the lower electrode sensor. When the rainwater level exceeds the
upper sensor the flow is increased twofold.
48
20
Fig. 2.21. Sampler for collecting rainwater for determination of volatile organic compounds. 1:
Glass reservoir, 2: glass fibre filter, 3: filter casing, 4: union, 5: glass T-tubes, 6: Teflon unions, 7:
ball joining, 8: top electrode, 9: set of electrodes (front view), 10: stainless steel rod, 11: bottom
electrode, 12: teflon guarding, 13: electric cable, 14: sorbent tube (Tenax-GC) for liquid
extraction, 15: sorbent tube (Tenax-GC) for thermal desorption, 16: connecting glass tube, 17:
stream splitter (capillary), 18: Teflon tube (to pump), 19: Teflon connector, 20: collecting surface
(based on Ref. [29]).
A typical configuration for sampling water and vapour from closed circuits con-
sists of a probe, a tube connecting a probe to a condenser, a condenser and a
choke appliance. When the installation is designed for continuous measure-
ments it should be equipped with an additional screen filter with 0.25 mm mesh
and a surface of 25-30 cm 2 . A thermic fuse is also needed to protect the control
device from damage by high temperatures. A system for sampling water and
vapour for periodical analyses is presented schematically in Fig. 2.22 [41], and
for continuous measurements in Fig. 2.23 [41].
Figure 2.24 shows a probe for collecting samples from pipelines [42]. The
internal diameter of the probe depends on the diameter of the pipeline in which
it will be installed. The probe cannot be used for pipelines with diameters
smaller than 50 mm. In such cases, a tube leading to a condenser should be
connected directly to a connector welded to the pipeline wall. The diameter of
the hole in the pipeline wall and the internal diameter of the connector should be
equal to the internal diameter of the tube leading to the condenser. A probe
should be installed on the straight segment of the pipeline in a such way that the
distance between the site of its installation and a valve, a slide or pipeline curve
is at least six times higher than the diameter of the pipeline. A probe can be
installed, depending on the local situation, to perpendicular or horizontal pipe-
lines, with a downward or upward flow of water. If possible, the probe should be
placed either horizontally to a perpendicular pipeline with a downward flow of
water, or perpendicularly or horizontally to a horizontal pipeline.
49
6
Fig. 2.23. Schematic representation of an installation for sampling water or vapour for conti-
nuous measurements. 1: Probe, 2: pipeline, 3: blocking valves, 4: tube connecting probe with
cooler, 5: by-pass for rinsing of probe and tube 6: filter, 7: controlling valve or throttle, 8: fuse, 9:
sensor of measuring device, 10: cooler (based on Ref. [41]).
A schematic design of a probe and method of sampler installation for the direct
collection of water from containers without a temporary flow of water is shown in
Fig. 2.25. A tube with an internal diameter of 10-12 mm is used as a probe. Its
opening should be placed at the height of the pipeline outlet. For containers with
water flow or to which different waters flow in, a probe located on the outlet pipe-
line, possibly in the vicinity of a container (Fig. 2.24) can be used.
50
direction of
water flow
Fig. 2.24. Installation of a probe for sampling water from pipelines (based on Ref. [42]).
inflow pipe
010
I
to condenser
For sampling water from horizontal barrel boilers a probe such as that
shown in Fig. 2.26 is recommended [43]. A probe cannot be installed in the
vicinity of a tube delivering correction chemicals to a boiler. It should also be
noted that the probe must be placed about 50 mm beneath the allowable level of
the water in the boiler. The part of the probe installed on the supports inside the
boiler should be resistant to vibration and forces appearing during the water
flow. The external part of the probe should be strengthened and equipped with
exchangeable connectors of different diameters.
_ S·-~-99~
2
I%
<4
I'~r~~
Fig. 2.26. Schematic representation of a probe for
sampling water from barrel boilers. 1: Orifices, 2:
maximum level of water, 3: common level of water, 4:
minimum level of water (based on Ref. [43]). to condenser
51
2.4.7 Sampling of piped water for determination of volatile
compounds and dissolved gases
Because the concentration of contaminants in water from pipes may change with
distance from the distribution tank and the residence time in the pipe, the condi-
tions for sampling of such water should be defined with respect to sampling site
(along the distribution grid), time of sampling (first open or after stagnation),
number of samples, sampling intervals, pipe material and residence time in the
pipe. The sampling of piped water should be conducted in such a way that the
sample is protected from contact with the atmosphere. After connecting a clean
rubber tube to a tap, a valve is opened and water flows for 10 min. Next, the sec-
ond end of the tube is placed on the bottom of the bottle and water flows into the
bottle until it is full. Before the sampling is finished, the water in the bottle
should be exchanged several times. After that the tube is withdrawn and the bot-
tle is immediately closed without leaving any headspace.
52
lifting turning measuring bottling
Fig. 2.27. Schematic representation of automatic collection of water, based on free water outflow
with delivering the constant volume samples. (a) Sample drawing, (b) turning away the excess of
water, (c) measuring sample volume, (d) transferring the sample into the reservoir (based on Ref.
[51).
automatic valve and a burette. A sampler is connected to the water source (in the
simplest case, a tap) by Teflon tubing. The stream of water flows continuously
into the sink through a valve and a burette with a total capacity of 32 ml. After a
predetermined time established with the control unit, the motor changes the po-
sition of the valve and the water is directed from the burette to a reservoir. An
electronic counter records the collection of the water sample. The combination of
many discrete samples in defined time intervals allows us to obtain time sequen-
tial composite samples.
Recently, a new device for automatic sampling of runoff during rainfall
events has been described [49]. It is easily made, simple to use, not vulnerable to
operational disturbances and inexpensive. The basic principle is to use the
weight of the rainwater to trigger the sample collection.
53
rounding the dosimeter into the inside part of a trap placed in the sampler is
completely free (according to Fick's first law of diffusion). The main driving
force and separation mechanism are based on the differences in concentration.
The equipment used for sampling is therefore not complicated, which is of great
importance since sampling sites are very often situated far from the laboratory.
Another advantage of the passive approach over grab sampling is that only
one device is necessary at a given sampling location for the duration of sampling.
A passive sampler can cover a long sampling period, integrating the pollutant
concentration over time. Moreover, decomposition of the sample by transport
and storage and/or changes during the sample enrichment step (e.g. contamina-
tion, degradation, adsorption) are minimized. Besides, in contrast to dynamic
techniques, passive sampling is simple to perform and after isolation and/or
enrichment, no further sample preparation treatment is usually required. Fur-
thermore, passive sampling is less sensitive to accidental extreme variations of
the organic pollutant concentration in natural waters.
Passive methods may generally be classified as adsorptive or absorptive.
Adsorptive methods take advantage of the physical or chemical retention by sur-
faces and key parameters involve surface binding and/or surface area. Absorp-
tive methods involve not only surface phenomena but also analyte permeation in
the interceding material. This latter approach provides the possibility of com-
pound discrimination due to the membrane's physicochemical characteristics.
The membrane, constituting of a diffusion barrier for transported molecules
of the analyte, is the most important element of permeation passive dosimeters;
hence, the material from which membranes are made should meet specific
conditions such as a large permeability coefficient for the analytes, which is
directly related to the kind of membrane material and to membrane thickness
and homogeneity, so that there are no differences between dosimeters used
during simultaneous exposure. Because permeation is also determined by
dissolution in the membrane, the transport of pollutants through the membrane
is influenced to a large extent by the nature of the membrane material.
54
TABLE 2.2
General specification of SPMD devices [51]
Parts of SPMD devices Material
Membranes: thin-walled Polyethylene; Silicone or silastic (option of
(25-250 pum) non-porous polymer tubes plasma-treated surface); Polypropylene;
Polyvinylchloride
Sequestration phase: high-molecular- weight Neutral lipids; Silicone fluids; Other lipid-like
(>600 Da) non-polar liquids or fluids organic fluids
Samples of collected water are usually pretreated in different ways before final
analysis. Some steps of the treatment can be carried out either during or imme-
diately after the sampling process. The main aims of these procedures, which are
frequently labour- and time-consuming, are [53-56]: ensuring the appropriate
physical properties of the samples and removal of any interfering components;
55
TABLE 2.3
Sample pretreatment methods for water [57]
Initial operation of sample Procedure
pretreatment
Removal of suspended Filtration, centrifugation
matter
Preservation pH adjustment to make more generally applicable (e.g. acidif-
ication); biocide addition; derivatization of analytes; application
of UW radiation; storing of sample at a temperature of 4°C.
Isolation and/or Solvent extraction; supercritical fluid extraction; solid phase
preconcentration of extraction; extraction into gaseous phase; membrane extraction;
analytes osmosis processes and ultrafiltration; freezing and lyophilisation.
Enrichment of extracts Evaporation of solvent excess, e.g. in Kuderna-Denish apparatus.
Purification of extracts Liquid chromatography; gel chromatography.
Extracts drying Salt adding, e.g. NaSO,4.
Usually, it is difficult to begin the analysis of water immediately after sample col-
lection. However, it should be noted that during storage many changes in the
sample may occur, and ensuring the integrity of the sample is as important as
the analysis itself. The character of the sample, the diversity and differing rates
of processes induced by many factors mean that there is no one universal method
of ensuring the stability of sample composition. Changes in sample composition,
and their reasons, between the sampling and final analysis are different and can
be caused by different reactions; the most important are chemical reactions,
physical processes, biochemical and photochemical reactions.
For chemical reactions, oxidation and reduction processes or hydrolysis of
chemical substances are quite possible. The reaction of chlorine with humic
materials and formation of halomethanes may occur. Depolymerization of con-
densed inorganic phosphates and polymeric silicic acids are also possible.
Changes of pH, as a result of absorption of CO 2 from air, lead to decreasing hard-
ness by precipitation of calcium carbonate. Some components of the sample may
form undissolved salts. Instability of the sample caused by chemical reactions
may be prevented or minimized mainly by the addition of an appropriate chemi-
cal reagent or by pH control.
Among physical changes, the most important are volatilization, adsorption
and diffusion. Volatilization can be minimized by filling the sample collection
vessels, leaving no headspace. Any contact with air is prevented and distribution
of volatile compound between the surface of sample and the headspace is
56
prohibited. Losses by adsorption can be minimized by rapid preservation of the
sample or by eliminating its exposure to the atmosphere.
Diffusion of some organic compounds (e.g. phthlate esters or plasticizers)
may be controlled by collecting samples in glass containers and by using bottles
caps or liners that minimize this process.
Samples can contain organisms (e.g., bacteria or algae) that may consume a
number of substances required for their growth, e.g., nitrogen, phosphorus,
silicon and organic compounds leading to their losses and introducing to the
sample metabolic products of such processes. These changes are generally
minimized by pH and temperature control or by chemical addition. Extreme pH
conditions (low or high) and low temperature are effective for minimizing
degradation. The addition of some chemicals to the sample (e.g., mercuric
chloride and pentaphenol) kills the microorganisms and effectively preserves the
sample, but their toxicity creates environmental hazards.
Generally, in order to prevent significant changes in water and wastewater
samples between sampling and analysis, one of the following preservation
methods is recommended: cooling to a temperature of 2-5°C; acidification to pH
< 2; addition of sodium hydroxide to pH > 11; addition of chloroform or
formaldehyde. The applicability of each of these methods of sample preservation
for determination of inorganic constituents and physicochemical parameters is
given in Table 2.4 [58].
Sampling and storage conditions and preservation methods of water samples
for determination of selected analytes and parameters are tabulated in Table 2.5
[59,60]. It must be noted, however, that complete stability of every sample's
constituents or parameters can never be guaranteed and at best, chemical,
physical and biological processes affecting the sample can only be slowed down.
TABLE 2.4
Applicability of preservation methods for aqueous samples [58]
Preservation method Determined substances (parameters)
Cooling to 2-5°C Organic, total, Kjeldahl and ammonium nitrogen, free and ionized
ammonia, nitrates, nitrites, colour, bromides and bromine
compounds, BOD, phosphates, iodides, acidity and alkalinity,
conductivity, sulphates, cationic surfactants, smell
Acidification to H2SO4 Ammonium and Kjeldahl nitrogen, free and
pH < 2 ionized ammonia, nitrates, COD, permanganate
index, dissolved silicates, silica
HNO3 Total hardness, metals
Addition of NaOH to Cyanides, iodides,
pH >11 sulphides
Addition of chemicals Chloroform Nitrates, nitrites, suspended matter
Formaldehyde Non-ionic surfactants
Other Phenol index (CuSO4 addition), sulphides (addition of zinc or
cadmium acetate)
57
TABLE 2.5
Sample preservation methods for some determinants of water [59,60]
Substance or parameter Sample Preservation Min. Preser- Max.
con- sample vol. vation holding
tainer (ml) temp. (C) time
_
Alkalinity-acidity P,GB 4 24 h
Ammonia P,GB H2SO4 (pH<2) 200 48 h
Arsenic P,G HCI (pH<2) 500 2 months
Nitrogen (Kjeldahl's PG H2SO4 (pH<2) 48 h
method)
Organic carbon G H2SO 4 (pH<2) 24 h
Chlorides P,G 100 15 days
Conductivity P,G In situ measure- 200 48 h
ment recommended
Cyanides GB NaOH (pH<12) 100 48 h
Biochemical oxygen P,G 1000 24 h
demand
Chemical oxygen P,G H2SO 4(pH<2) 100 24 h
demand
Hardness P,G HNO, (pH<2) 100 1 month
Fluorides P 300 7 days
Hydrocarbons G CCI4 (10 ml) 300 7 days
Polynuclear aromatic G C6 H1 4 (10 ml) 100( 46 days
hydrocarbons
Oils and fats G HCl (pH<2) 100( 4 15 days
Lithium G 15 days
Suspended matter Al, P,G HNO, (pH< 1.5) 100( 4 6h
Ag, Cd, Cr, Mn, Pb, Zn 2 month
Mercury GB HNO, (pH<5) 1 month
Nitrates P,G 100 4 48 h
Nitrites P,G 100 4 24 h
Colour, smell, taste G 500 4 24 h
Oxygen, dissolved GB In situ measure- 300 4 24h
ment recommended
Pesticides G 200( 4 7 days
pH GB 4 24 h
Phenols GB CuSO4 (1 g/ml) + H3PO4 500 4 7 days
(pH 4)
Phosphates P,G 100 4 48 h
Radioactivity P 100( species-
dependen
t
Dry residue P,G 500 4 7 days
Ether extract G CHC1I (1 ml/l) or HgC12 4 48 h
(40 mg/l)
Silica P 4 7 days
Sulphates P 4 6 days
58
TABLE 2.5 (continuation)
REFERENCES
59
25 J. Unwim and V. Maltby, ASTM STP, 963 (1988) 240.
26 J.H. Ficken, ASTM STP, 963 (1988) 253.
27 W. Dyck, J.C. Pelchat and G.A. Meilleur, Equipment and Procedures for the
Collection and Determination of Dissolved Gases in Natural Waters, Ecological
Survey of Canada, Paper 75-34, 1976.
28 B.A. Torstensson and A.M. Petsonk, ASTM STP, 963 (1988) 274.
29 M.E. Rosen, J.F. Pankow, J. Gibs and T.E. Imbrigiotta, Ground Water Monit. Rev.
(GWMR), 126, Winter 1992.
30 J.F. Pankow, L.M. Isabelle, J.P. Hewetson and J.A. Cherry, Ground Water, 22 (1984)
330.
31 A. De Schrijver, G.Van Hoydonck, L. Nachtergale, L.De Keesmaeker, S. Mussche and
N. Lust, Water Air Soil Pollut., 122 (2000) 77.
32 J.A. Smith, H.J. Cho, P.R. Jaffe, C.L. MacLeod and S.A. Koehnlein, J. Environ.
Qual., 21 (1992) 264.
33 S.M. Chernyak, C.P. Rice and L.L. McConnell, Marine Pollut. Bull., 32 (1996) 410.
34 K.W. Lin and R.M. Dickhut, Environ. Sci. Technol., 31 (1997) 2777.
35 J.M. Southwood, D.C.G. Muir and D. Mackay, Chemosphere, 38 (1999) 121.
36 D.T.Waite, A.J. Cessna, R. Grover and E.J. Woodsworth, J. Environ. Qual., 29 (2000)
901.
37 S.J. Pogorzelski, Rev. Sci. Instrum., 63 (1992) 3773.
38 IOC/GEMSI Operational Manual for Sampling the Sea Surface Microlayers in the
Marine Pollution Monitoring Progamme for Petroleum (MARPOLMON-P), 1983.
39 W.M.J. Strachan and H. Huneault, Environ. Sci. Technol., 18 (1984) 127.
40 W.M.J. Strachan and H. Huneault, J. Great Lakes Res., 5 (1979) 61.
41 J.F. Pankow, L.M. Isabelle and W.E. Asher, Environ. Sci. Technol., 18 (1984) 310.
42 Polska Norma, PN-74/C-04621. Woda i scieki. Pobieranie pr6bek pary wodnej do
analizy fizycznej i chemicznej (in Polish).
43 E. Sierakowski and J. Mrozek, Kontrola chemiczna obieg6w wodnych i wodno-
parowych w elektrowniach. WNT, Warsaw, 1974 (in Polish).
44 J. Dojlido, Instrumentalne metody badan wody i sciek6w. Arkady, Warsaw, 1980 (in
Polish).
45 D. Hirst, J. Hydrol., 134 (1992) 95.
46 L.V. Parker, Ground WaterMonit. Remed., Spring (1994) 130.
47 T.E. Imbrigiotta, J. Gibs, T.V. Fussillo, G.R. Kish and J.J. Hochreiter, ASTM STP,
963 (1988) 258.
48 J. Turcotte, J.E. Cot, E. Fraser, H. Oullette and S. Langlois, Poll. Atmos., 33 (1991)
192.
49 M. Nieminen, Boreal Env. Res., 5 (2000) 133.
50 J.D. Petty, C.E. Orazio, J.N. Huckins, R.W. Gale, J.A. Lebo, J.C. Meadows, K.R.
Echols and W.L. Cranor, J. Chromatogr.A, 879 (2000) 83.
51 J.N. Huckins, M.W. Tuburgen and G.H. Manuweera, Chemosphere, 33 (1990) 533.
52 A. Kot, B. Zabiegala, J. Namiesnik, Trends Anal. Chem., 19 (2000) 446.
53 D. Barcelo, W.A. House, E.A. Maeir and B. Griepink B., Int. J. Environ. Anal. Chem.,.
57 (1994) 237.
54 K.R. Bull, K.H. Lakhani and A.P. Rowland, Chem. Ecol., 9 (1994) 47.
55 J.T Creed, T.D. Martin and M. Sivaganesan, JAWWA, 87 (1994) 104.
56 M.J. Mogollon, A.J. Raminez and C. Bifano, Chem. Geol., 121 (1995) 263.
57 J. Namiesnik, Z. Jamr6giewicz, M. Pilarczyk, L. Thorez, Przygotowanie pr6bek
srodowiskowych. WNT, Warsaw, 2000 (in Polish).
58 J. Namiesnik and Z. Jamrugiewicz (Eds.), Fizykochemiczne metody kontroli
zanieczyszczen. WNT, Warsaw, 1998 (in Polish).
59 R. Jeannot, Int. J. Environ. Anal. Chem., 57 (1994) 231.
60 A. Kot-Wasik, M. Morawska and J. Namiesnik, Chem. Inz. Ekol, 8 (2001) 177 (in
Polish).
60
Chapter3
3.1 INTRODUCTION
Solid samples are characterised by even larger variability than liquid samples.
Typical solid materials sampled for analysis include: raw materials for indus-
try/manufacturing, soils, sediments, snow and ice, wastes (industrial, municipal,
hazardous, etc.), biological plant material, electrofilter dusts, road dusts, tissues
and organs of animals, etc.
Analysis can give reliable results only if a sample has the same quantitative
and qualitative composition as the material studied [1-4]. Even the most
accurate analysis of the test sample (Section 3.8 Glossary) is useless if primary
samples are not correctly collected and a gross sample is not representative of
the material studied. The same importance should be given to sampling as is
usually given to subsequent steps of an analytical process. Also, a quality
assurance program is required for this step to ensure that information derived
from analytical data is correct. These sampling problems are discussed in a
number of books and papers [1,3-6].
Generally, analytical protocols contain detailed description of this step of
analysis. Sample collection procedures and systems are also included in techni-
cal standards. If this is not the case general procedures for collecting a primary
sample and producing a representative laboratory sample should be applied.
General discussion of sampling and nomenclature can be found elsewhere
[7-11]. The brief list of definitions of basic terms as used in this chapter is given
in a short glossary (Section 3.8).
Some sampling problems are universal, and general principles for sampling
bulk materials can often be directly applied to numerous matrices. A compre-
hensive description of sample collection at industrial wastes dumps, for example,
has been given by Rasemann and Markert [12].
The aim of sampling is to obtain a small and representative portion of the
population studied. If the samples do not reflect the properties of interest of the
population under study then they are of little or no use. Therefore planning for
informative sampling should be an integral part of any study.
Sampling is a part of the overall analytical process, and before formulating a
sampling procedure it is necessary to define: objectives of sampling, analytes of
ComprehensiveAnalytical Chemistry XXXVIr
J. Pawliszyn (Ed.)
© 2002 Elsevier Science B.V. All rights reserved 61
interest, analytical methods to be used, way of reporting the results, and way of
using data. When preparing the sampling plan the following questions should be
answered:
- What kinds of samples are needed?
- When, where and how should samples be collected?
- What containers should be used to collect samples?
- How should the samples be preserved?
- What preparations are necessary for sample analysis?
- What precision of sampling is required?
Each sampling protocol should specify the number and size of samples. They
depend on many factors, including lot size, material granulation, analytical
method to be used, variability within and between particles, concentration of the
component of interest, etc. The related problems should be discussed individu-
ally for each sampling task.
If the product to be sampled is in packages, a specified number should be
randomly selected from the lot, i.e. from the total amount of material which is to
be characterised. One of the factors determining the number of sampled pack-
ages is the number in the lot; generally, if it is less than five, samples should be
taken from each package in the lot.
Statistics were used to answer questions concerning the number and size of
the samples to be collected. Although most work was aimed at solving specific
sampling problems some treatises deal with more unified approaches [13,14]. In
the widely known Ingamells approach describing non-segregated objects the
mass of a sample is calculated from relation (3.1) [15]:
WR 2 = K (3.1)
where W is the mass which should be taken, R is a sampling relative standard
deviation, expressed in percent, found for the material experimentally, and K. is
the sampling constant, corresponding to the mass of sample required to limit the
sampling uncertainty of 1% with 68% confidence.
The number of samples necessary to achieve a given level of confidence can
be estimated from relation (3.2) given by Kratochvil and Taylor [7]:
2 2 2 2
n = t S /iR X (3.2)
2
where t is the Student t-table value for the level of confidence required, s (the
standard deviation of the individual samples) and x (the average of analytical
measurements) are estimated from preliminary measurements or from previous
knowledge of the bulk material, and R is the percent standard deviation in the
average.
For segregated populations Eq. (3.3) was proposed by Wisman [16]:
s 2 = A/wn + B/n (3.3)
where s2 is the variance of the average of n samples using a total mass w of
sample, and A and B are constants for a given bulk material.
62
As pointed out by Benedetti-Pichler [17] random sampling errors can also
occur in well-mixed particulate mixtures if particles differ considerably in
composition and a sample is relatively small. The relationship between the
allowable error and the number of particles to be counted is as follows [17,18]:
n = (1-p)/Gr
- (3.4)
where p is the fraction of one kind of particles and (1 - p) is the fraction of the
other kind of particles.
The above approach is also extended to describe the bulk material composed
of two kinds of particles differing in content of a component of interest-a higher
percentage is P1 and a lesser content is P2 giving an average percentage of P; the
corresponding densities are d, and d2 giving an average density of d. In this case
Eq. (3.5) can be applied [17-19]:
63
This property can be the content of a selected component, moistness as well as
particle size.
Let us consider lumpy material, e.g. raw material composed of particles of
different size. Lumps can differ in chemical composition and density as well as
size. When such material is unloaded into a pile, segregation occurs; a majority of
the larger lumps gather at the outer bottom part of a cone formed during the
process. Similar segregation takes place during unloading bulk material onto
rail cars; additional segregation occurs during transport. Fine particles tend to
move to the bottom while larger particles remain in the upper layer. If material
piled on the ground or in a rail car is soaked with precipitation, the moisture
content will not be the same throughout but generally will decrease with depth.
The opposite situation can also happen; moist material subjected to sunlight
loses water from outer layers faster, resulting in moisture content increases with
distance from the surface. Such variability can be observed on a conveyor belt
supplying material from a pile or rail car. Transported bulk material undergoes
partial re-mixing during unloading and reloading; however, the effect is
generally not sufficient to obtain the previous state of homogeneity. Often
additional segregation occurs during transport on belt conveyors due to belt
vibrations.
From the above considerations it is clear that variability in a given property
can differ greatly from case to case. This must be strongly stressed since the
nature of the variability can have a significant influence on the sampling
procedure. The difficulty arises mainly from the fact that little is generally
known about the nature and extent of variability before sampling. Usually,
because of time and cost restraints, it is also difficult to conduct corresponding
studies to acquire this knowledge. In such situations experience and knowledge
of the behaviour of materials having similar characteristics can be very helpful.
Furthermore, a thorough examination of a product lot with respect to physical
characteristics, transport history (including extent of loading and reloading),
etc. should be conducted, especially for materials from new suppliers.
The degree of variability or difference in properties among segments of the
same lot shows whether the material homogeneity is fit for the purpose or not. If
differences in a given property among different lot parts are not significant for
the purpose, material can be regarded as homogeneous for that purpose. The
definition does not strictly define the size of the parts. It also allows for
variability within as well as between parts. In other words, if a given part were
divided into a number of smaller portions, the difference in a property among
them can be significant and particular parts could be heterogeneous. If, how-
ever, average values of a property for the segments differed to a statistically
insignificant degree, then the material lot with respect to these segments could
be regarded as homogeneous. The practical importance of this problem for
sampling will be exemplified below.
Let us consider a certain lot of material, composed of a large number of pack-
ages from which several tens were selected for sampling. Measurements on pri-
64
mary samples collected from different points within individual packages showed
significant variations in values of a selected property. The material in these
packages was not homogeneous in the component measured. However, the aver-
age values between packages were not significantly different, so the lot could be
regarded as homogeneous in this component with respect to the packages. A
practical conclusion from this observation is that the number of packages
selected for analysis can be decreased provided that a sufficient number of pri-
mary samples are taken from each package selected for analysis.
A similar situation can happen for a bulk material lot transported in rail
cars. If primary samples from particular cars differ considerably in composition
then the product in them is heterogeneous, and the number of primary samples
collected from individual cars should be increased. If, however, average values of
a given property for different cars differ negligibly, then analytical costs can be
reduced because the number of cars subjected to analysis can be decreased. This
assumes the precision and accuracy of analytical data meet required levels.
A different situation can occur where primary samples from within particu-
lar packages do not differ significantly, but average values between packages are
significantly different. Then the number of samples from a package can be
reduced but the number of packages subjected to analysis should be increased.
If a material is homogeneous throughout the whole lot then collecting a
portion of material from any place in the lot will give a representative sample.
However, assuming that material is fully homogeneous without properly
planned investigations can be risky and expensive. Therefore, even in the case of
liquid materials, well averaged by thorough mixing, more than one primary
sample should be taken.
The effect of variability of the properties studied on sampling is not limited
to the problems described above. Material heterogeneity can manifest itself in
many different ways. Several heterogeneity varieties are recognized [21]: ran-
dom, directional (e.g. axial) and periodic. Heterogeneity is random if a property
value in one portion of the population studied is independent of the values in
neighbouring portions and independent of the position in the whole bulk of the
material. For example, a given lot of material in drums from the same plant but
different production lines and/or different periods may give a final product
which differs in content of a given component. If the products were not properly
mixed before packaging and, due to multiple reloading, packages were mixed up,
it is not possible to tell the characteristics of the product in a given drum even if
the content of a given component is known in all neighbouring drums.
A similar situation can arise in the case of lumpy materials, though the
additional heterogeneity comes from segregation with respect to particle size.
For instance, raw material can come from the same mine but from different
beds, or it can pass through different enrichment equipment; this can lead to
heterogeneity. Although it can be mixed during reloading some heterogeneity
can remain; it will only be random heterogeneity if the material does not
undergo segregation due to difference in particle size.
65
In the case of random heterogeneity, the location of a sampling point is not
very important; however, the number of primary samples taken is crucial, and
should increase as the difference between the particular samples increases.
Problems with heterogeneity become more complex when property variation
with direction and time needs to be taken into account. If systematic heterogene-
ity occurs, a property value for a given material segment can be obtained from
values in the neighbouring segments. Let us consider again a drum with granu-
lar material of different particle sizes. Once the drum content is thoroughly
mixed, averaging of properties can be sufficient for heterogeneity to be random.
In practice, however, particle-size segregation can occur due to reloading and
transport. Then samples collected from different layers will show axial variabil-
ity, i.e. a property value (e.g. content of small particles) can linearly increase in
bottom direction. Such distribution of a property requires that a given sample is
collected along the whole drum height or samples are taken from several layers
along the height.
On a conveyor belt, vertical and horizontal segregation can occur; this makes
it necessary to take samples from the whole cross-section of the material.
Directional variability of material in a storage place or on a rail car is not as
simple as for the cases described above. Property variation can be multi-
directional, often dependant on how the material was reloaded. Such variability
requires careful selection of sampling points and a larger number of primary
samples. The idea is to collect samples which reflect the distribution of a
property in the whole bulk of the material.
Figure 3.1 shows the distribution of a fine particle fraction (< 10 mesh) in the
cone of coal containing 20% of the above fraction when a chute pipe in a vertical
position is raised slowly to minimise segregation (a) and when a chute pipe is in
an oblique position (b) [21].
Figure 3.2 shows the principle of sampling points location when particle-size
segregation occurs [21]. This is a scheme proposed by the Association of Official
Analytical Chemists for material on, e.g., a tipper trailer or rail car. According to
this scheme, a material lot should be divided into ten parts and samples collected
from each. The total of these samples gives a gross sample which is to be
representative of the material studied. Directional heterogeneity can also be
found in liquids containing suspensions.
Periodic heterogeneity should also be discussed: this occurs in flow manufac-
turing processes and can result from periodicity in feeding raw materials or
a) b)
Fig. \ Distribution
23.1.of fine13,
Fig. 3.1. Distribution of fine 29,6 '27, ' 28,0'
particle fraction (<10 mesh) in / , ,
a cone of coal containing 20% of 8 20 9 35,8/ \14.4\
this fraction (based on Ref.
[21]). vertical shute side shute
66
7 8
3 1 2 1 r,
6 1_
Fig. 3.2. Principles of sampling points location when 10 9
particle-size segregation occurs (based on Ref. [21]).
67
a) A-A b)
S2
08
c)
I
1
Fig. 3.3. Samplers for greasy and doughy, and fusible materials (see text for description) (based
on Ref. [22]).
The other ends of the tube halves are tapered. The open sampler is driven down
to a required depth and closed by turning the handles 180° . After withdrawal the
sample is transferred to a container with a rod or spatula.
For more doughy materials the sampler shown in Fig. 3.3b may be used. Its
principle of operation is similar to a gimlet bit: after removal of the upper layer
the sampler is driven into the material by rotating it. Again, the sample collected
is removed from the sampler with a small spatula.
68
material remaining in the winding should be removed, e.g., with a spatula or
stirring rod and included in the sample.
For sampling brittle and easily breakable materials, tools such as axes or
chisels can also be used; these tools are not convenient for sampling materials
that adhere to them. Before chipping off pieces of material for sampling pur-
poses, the surface layer-which can have changed in composition-is removed.
The pieces of material from different spaces of the package or lot are used as
primary samples or to prepare a gross sample.
The main factors determining selection of sampling equipment for loose and
lumpy solids are granulation, material quantity and whether it is moving or
stationary during sampling. As in the case of other sampled objects, samplers
must be made of a material which does not interact with sample components.
The simplest tools for sampling loose and lumpy material are shovels whose
dimensions depend on particle size and primary sample size. Under ordinary
conditions samples from the outer layer or from small depth can be collected
with these samplers. Therefore, they can be used when material is homogeneous
or of random variability.
Shovels are also used for collecting samples from belt conveyors of small
capacity when the material particle size is not larger than 50 mm. Care should be
taken when collecting samples from the whole cross-section of material to move
the shovel at a constant speed. The shovel should be 2.5-3 times wider than the
largest lumps. It should be noted, however, that shovel sampling directly from a
moving conveyor belt is prohibited.
The most common samplers for fine-grained (<20 mm), granulated and
powdered materials are pipe samplers of different design; they can be either
single-tube or double-tube samplers.
The sampler presented in Fig. 3.4a [22] is made of a metal tube bevelled at
45° at one end; the tube edges are slightly rounded off. At the other end is a
handle closing the tube. Since the sampler has an open end, it should be pulled
out of the material in a horizontal position or at an angle smaller than the
natural slip angle of the material. The sampler length should allow samples to be
taken along the whole diagonal of a container. The tube diameter should be at
least 2.5 times larger than the maximum particle size but never smaller than 20
mm. When the sampler is open at both ends one can collect samples in a
continuous way by allowing the material to flow from the package to a sample
container; this, however, gives unrepresentative samples if the material is
heterogeneous. Moreover, sampling must be done in a horizontal position or at a
rather small angle; an additional disadvantage is that some of the sample is lost
when pulling the sampler from the material. A certain degree of segregation of
the material of different granulation can also occur when sliding the sampler
into the material.
69
a) b) a)
a)~ b,)~ la).,
I I
A A - ! '
X~~~~~~~~~~~~~~~~~~~~~~~j
__1
o 0
I X I I
il
I' ,,, I
,1 8 I
III I o
1,5 At
iI
I
Left: Fig. 3.4. Samplers for loose materials (see text for description) (based on Ref. [221).
Right: Fig. 3.5. Samplers for loose materials (see text for description) (based on Ref. [22]).
The sampler shown in Fig. 3.4b [22] differs from that shown in Fig. 3.4a in
that one side of the tube is cut out leaving a U shape. The width of the cut-out
opening should not be less than 2.5 times the maximum particle size. The
sampler should be pushed horizontally into the material with the opening
downward, three half-turns should be made, and it should be tapped slightly,
then pulled out with the opening upwards. Particularly in the case of powdered
and very loose materials, the package should be positioned in such a way that the
sampler can be withdrawn horizontally, otherwise some of the sample is lost.
Another type of sampler is presented in Fig. 3.5a [22]. It is T-shaped and
made of tubing terminated at one end in a rounded cone; the other end is open
and has a handle fastened. A rod or tube handle is welded on or fastened in
another way to the outer side of the tube. Along the whole tube length, an
opening is cut out, whose width should not be smaller than half the outer
diameter of the tube. Similarly to the previous one, this sampler is applied in a
horizontal position or at a small angle despite having a closed tip. If used in a
vertical position it may collect more material from the upper layers and part of
the sample may be lost when pulling the sampler out of the material, mainly due
to friction with the material in the package. The disadvantage of this sampler is
that the lowest layer of the material (between the cone tip and the opening) is
not sampled. To minimize this effect, samplers with a cone of very small height
are constructed. Sampling is done in the same way as in the case of the sampler
presented in Fig. 3.4a. Care should be taken during emptying and cleaning the
sampler-in the process the sampler is gently tapped; the sample part which
70
a) b)
closed open
Fig. 3.6. Samplers for loose materials (see text for description) (based on Ref. [22]).
remains is taken out with a spoon. The sampler can also be applied in such a way
that a fitted metal or wooden rod is slid into the tube before pushing it into the
material. Filling proceeds when the rod is slowly withdrawn while the sampler is
gently rotated.
Other samplers for loose materials are presented in Figs. 3.5b and 3.6 [22].
They consist of two tubes, one inside the other, which are well fitted but can
move with respect to each other. The lower end of the outer tube is terminated
with a slightly rounded cone to protect the package from puncturing. The end of
the inner tube, which extends up to the cone base is closed. Rod or tube handles
are fastened perpendicularly to the upper ends of the tubes. The inner tube
handle does not go across the inner tube bore; this makes emptying and cleaning
easier. Along the length of the outer tube, from below the handle to the cone
base, a longitudinal opening is cut. The width of the opening is generally equal to
half the diameter of the tube; other sizes are allowed, but they should be no less
than 2.5 times maximum particle size. Often, several short longitudinal open-
ings along the same line are made at small distances from each other (Fig. 3.5b).
The inner tube also has one or more openings identical to the outer ones with
respect to dimensions and arrangement. The adjacent edges of openings should
be cut at right angles to protect particles from getting between tubes. The
diameter of the sampler should be at least five times larger than the largest
particle and never smaller than 20 mm. On both tubes marks are made close to
the handles to show sampler position (closed or open).
Before use one should check that both tubes move easily with respect to each
other. The closed sampler is pushed into the material horizontally or at a small
angle with the outer tube openings directed upwards. The sampler is opened by
turning the inner tube to a position such that the openings of both tubes
correspond. To facilitate filling, the sampler is gently tapped or rotated in both
directions. Then the sampler is closed by turning the inner tube 180° , withdraw-
ing it and emptying it through the open upper tube end. Lastly, the outer tube is
pulled out and cleaned up if necessary. For materials with a tendency to enter
between the tubes and make tube rotating difficult, the inner tube can be
71
replaced by a metal or wooden rod of slightly smaller diameter and the sampler
used as a one-tube sampler (Fig. 3.5a). The advantage of this design over
non-closed ones is that the sample is not collected until the sampler is completely
immersed in the material. Also, no sample is lost during removal from the
material. These samplers have some disadvantages, but generally give more
representative samples than the previous ones. The main disadvantages are: if
used in a vertical position the proportion of material from the upper layers is
increased; particles can crumble during sampler closing; the conical end of the
sampler does not allow the lowest layer (from tip to the first opening) to be
sampled.
An improved design of pipe sampler is shown in Fig. 3.6b. The difference
from the previous designs is the division of the inner tube into segments with
lateral barriers. Each segment has a separate opening in both tubes. The
sampler is used as previously described but can be pushed into the material
vertically since segmentation results in more representative sampling along the
sampling axis. With care during emptying, separate samples from individual
layers can be obtained.
The sampler recommended for sampling fertilisers is shown in Fig. 3.6a. It is
constructed of two halves of fitted tubes, the outer one ending with a flattened
cone. The halves are joined by axles around which they can be turned. After
immersion in the material, it is closed by turning the outer half. During sam-
pling the material does not move at all or only negligibly; this is very important
since material movement results in segregation which leads to additional errors.
Selecting the correct sampling place in the soil profile requires the following
initial operations: (1) digging out soil profile, i.e., uncovering the upper part of
72
the soil, at least one layer (usually three), if possible without disturbance, to a
depth of 1.5 m; (2) determination of the genetic classification unit of soil (type,
soil textural group); (3) detailed description of the profile (thickness of strata,
colour, structure, granulation, etc.).
Samples are collected from particular genetic levels or soil horizons,
sometimes also from geological strata. To determine granulation, apparent
density (ratio of mass of sample of undisturbed structure to sample volume),
hygroscopicity, acidity, carbonate content, chemical and mineralogical compo-
sition, samples can be collected without structure preservation. To determine air
and water conditions, i.e. capacity with respect to water (capillary, pore and total
water) and to air (porosity, pore size) as well as soil structure and density,
samples should be taken without natural structure disturbance.
When profile sampling of mineral soils, sampling places are selected after
careful cleaning of the profile wall; sampling starts from the lowest part of the
profile.
Samples without structurepreservation are cut out with a shovel, spatula or
knife from the middle part which is the most typical for a given layer. A sample
from the lowest part of the profile should represent its lower level down to a
depth of 2 m. The soil beyond the profile reach (below the level of 1.5 m) is
sampled with a special drill. The sample from the top organic layer is collected
from an undisturbed area nearby.
The sample mass of mineral soil should be ca 1.5 kg (ca 1 kg air dry soil of
granulation smaller than 1 mm) and of organic soil ca 0.3-0.5 kg. Samples of
gravelly and stony soils should be generally larger than 1.5 kg.
Samples from each genetic level are packed in linen or plastic bags. A plastic
or paper tag is inserted into each bag giving: area description (place, etc.), profile
number, characteristics of genetic level (symbol, thickness, colour, etc.) and
sampling date. The closed bags from a given profile should be tied together.
Samples of undisturbedstructure are obtained with metal or plastic cutting
cylinders of 100 ml capacity, closed at both ends with rubber caps. Before
sampling, numbered cylinders should be cleaned, dried and weighed.
After cleaning a profile wall and identifying particular layers, sampling
points are selected. Starting from the profile bottom, cylinders are driven into
the soil at each level at the point chosen. The cylinders filled with soil are taken
out with a shovel or knife, the soil surface is levelled out at both ends and the
cylinders capped. The cylinders are stood next to each other in a special
container to protect them from falling over during transport. Each sample is
described (as above) in a logbook.
Samples from organic soil layers (humus, peat, etc.) are also collected using
cutting cylinders, mainly of a volume of 250-500 ml.
The collection of undisturbed structure samples of gravelly and stony soils is
particularly difficult. In this case, large cylinders of a volume of up to 1000 ml are
used. If a soil contains large rocky pieces, then their content should be evaluated
and samples collected from spaces of fine earth fraction.
73
3.5.2 Sampling arable layers
1
2
74
boxes and air-dried for approximately two weeks in a dry and well ventilated
space, with the boxes open (slight soil dusting can occur). For determination of
volatile organic compounds in soil, separate standard procedures for collecting
and storing samples are described in Ref. [29]. The samples are generally stored
in glass jars closed with PTFE caps at a temperature of 4°C.
For preparation of a test sample, depending on needs, part of the laboratory
sample is taken and ground in a soil mill with a 2 mm mesh sieve. Particle size
reduction can also be done in an agate or porcelain mortar followed by sieving
through a plastic or iron screen of 2 mm mesh. After sieving and discarding plant
parts, stones and gravel, the soil must be thoroughly mixed. For determination
of general forms of mineral components, part of the laboratory sample is
powdered to such a degree that 80% passes through a sieve of 0.25 mm mesh.
The very fine fraction (<0.25 mm) and that which remains on the screen are
accurately mixed and poured out into a plastic vessel.
75
from individual samples whose number depends on area size, e.g. for areas lower
than 1 ha 5-10 samples are sufficient, for 10-20 ha 20-30 primary samples are
necessary. Sampling sites should reflect the variability of the area studied as
much as possible. Most often samples are collected along a zigzag line (Section
3.5.6).
The most typical technique of sampling soil from the root layer is the
following. In selected sites a soil portion is dug out to a depth of 25-30 cm in a
single operation with a shovel. After removal, the soil is left on the shovel and
samples of 100-150 g are collected with a small shovel or knife from each clearly
distinguished genetic level starting from the lowest. Each successive level is
sampled after removing the previous one. Samples from each level are collected
in separate plastic buckets. Samples from the organic layer should be taken in
such a way that duffs (non-decayed organic matter) are not contained in them.
After collecting the required number of samples from each level, the material in
the buckets should be thoroughly mixed and a portion of ca 1 kg taken from each
bucket. For the area studied three composite samples are generally taken, each
characterising a selected genetic level. If a toxic substance is to be determined in
the root layer, composite samples from all levels are mixed together to give an
composite sample for the whole root layer. In some root layers only two genetic
levels are discerned; this is mainly the case for orchards and parks, where former
arable layers and subsoil are clearly marked. Composite samples of these two
layers are put into linen or plastic bags. Identification tags with additional
description (subsection, section, district, locality) of the area studied are inserted
into each bag. The bags are tied up and kept in dark and cool space.
76
handled, labelled and prepared in the ways discussed above.
The average sample should represent the area of agricultural land of close
natural conditions (soil type, soil textural group, topographic features, etc.). For
similar areas with respect to soil, topographical features, etc., the sampling area
to prepare one composite sample should be ca 0.2-0.4 ha. Sampling sites should
be selected along a zigzag line or along a diagonal (Fig. 3.8) [29].
The approaches presented schematically in Fig. 3.9 are also proposed [30].
They include normal sampling, test sampling, diagonal sampling, and cross
sampling.
77
a) along zigzag line b) along diagonal
iF\\ / /
\\\ / / / /
Fig. 3.8. Zigzag and diagonal arrangement of sampling points for soils (based on Ref. [29]).
a)
Fig. 3.9. Arrangement of soil sampling sites proposed by Rump and Kirst [30]. (a) Normal
sampling, (b) test sampling, (c) diagonal sampling, (d) cross sampling.
Methods and dates of sampling biological plant materials depend on the aim of
the studies. Most often this material is sampled to determine: (a) crop quality
(pollution level included); nourishment needs; cycling of nutrients in the envi-
ronment; degree of use of soil nutrients on the basis of their accumulation in a
given plant species in a given period; (b) level of supply of nutrients to plants,
distribution of nutrients in plants; effect of different parameters (e.g. fertilisa-
tion, SO2 emission) on the basis of accumulation of these substances in different
parts of a plant at various phases of growth; and (c) direct plant pollution from
given substances accumulation in the whole plant or in some parts.
According to the intended aim of the studies, the sample collected can char-
acterise: a given plant (e.g. tree, bush)-the sample is collected from that
particular plant; a given area occupied by one or more plant species-individual
samples are collected from points representative of the area: the samples can be
analysed separately, however, they are most often combined to give a composite
sample.
3.6.1 Sampling
When sampling from shaded areas, e.g. in parks, orchards, etc., it is important to
take samples that represent the specific area under study (e.g. ecological niches).
When insolation is highly differentiated, separate samples of plants from shaded
and insolated areas should be collected. The sampling principles for plants in
growth phase and for mature plants are the same. The whole plant or only leaves
are gathered; full-grown leaves must be taken from the middle part of a stem;
growth phase should be noted. From leaves collected from single plants repre-
78
sentative of a given area, a composite sample is prepared for analysis. Leaves
should be gathered proportionally to the plant occurrence-each plant should be
properly represented in the composite sample. A similar procedure is applied for
sampling other plant parts (e.g. fruit, flowers). The samples are placed in linen
bags or in paper envelopes (never in plastic bags) and delivered to the laboratory
as soon as possible. The samples should be properly labelled; an identification
tag should be inserted in each bag; sampling site, plant species, plant part
collected and sampling date should be given.
To determine total content of particular components accumulated in a
period of growth by cereals and oil plants, samples of full-grown plants are
taken. If the assimilation rate of a given component, e.g. from a fertiliser or
emission, is to be determined, then samples are collected at different growth
phases. Independently of the aim of the studies, the area sampled should always
be characterised. Over an area of ca 1 ha 30-50 sites (more or less equally
distributed) are selected; at each site at least five plants are cut close to the
ground and placed in a linen bag. The total material collected in the field is
thoroughly mixed and a sample (0.3-0.5 kg) is taken for analysis. For full-grown
plants, seeds and straw are separated and stored for analysis as separate
samples.
Obtaining a representative composite sample of root crops is especially
difficult: plants of different characteristics should be gathered so that the total is
representative of the area studied. As a rule, several tens of individual samples
are collected by digging out individual plants. After separation of bulbs and
above-ground parts, they are packed into separate linen bags. After samples
have been collected from the whole area they are poured out (on paper or sand),
thoroughly mixed, bulbs and above-ground parts separately, and composite
samples are collected.
Primary samples of fodder plants are collected in a similar way to sampling
cereals. Plants from 30-50 sampling sites regularly distributed over an area of 1
ha are cut off at the ground over squares of 10 x 10 cm. After mixing the material
collected a composite sample of ca 0.3-0.5 kg is taken for analysis.
Depending on the aim of the studies, primary samples of wild plants are
taken either as collective samples containing all plant species cut off over a
so-called micro-area or as samples of particular species from given micro-areas.
The sampling procedure is the same as that described above. The number of
samples depends on the area studied, density of coverage, occurrence of
particular species, etc.
In the case of forests, mainly leaves or needles are subjected to analysis,
although sometimes shoots, bark, wood and ground cover are collected. The
principles of sampling ground cover are the same as for wild plants. Plants
higher in evolution are gathered during blooming, and bryophytes (mosses,
lichens) in autumn. Leaves are collected during growth periods, needles after a
period of growth, from the outer part of the middle of the tree head. Leaves from
taller trees are collected with special tools or by shooting twigs.
79
The time of sampling leaves depends on weather conditions, phases of
growth, growth conditions, and other factors, e.g. the necessity for immediate
determination of pollution in a given region. Therefore, depending on the aim of
the studies and the climatic conditions, the date of sampling leaves should be
decided individually for each case.
The age of needles on a coniferous tree can range from a year to tens of years
depending on the tree's age; which needles are selected for analysis depends on
the aim of the study. Generally, older needles are analysed to determine the
long-term effect of anthropogenic impact on coniferous plants. If the aim of the
study is to determine the quantity of particular components assimilated during a
growing period, needles from the last growing period are collected. Needles are
always gathered during non-growing periods, i.e., November to February in the
northern temperate zone. Which part of the tree head is sampled also depends on
the aim of the analysis: to determine the pollution of needles, samples are
collected from the windward side, i.e. from the prevailing direction of pollutant
influx.
Due to variability in populations of particular tree species, the leafage of each
tree is treated as an individual sampling area; several twigs from each tree are
used to prepare the composite sample. To characterise a selected forest area,
samples are gathered from several trees representative of the area studied. Each
tree is analysed separately. If bark is subject to analysis it must come from the
same tree that the needles or leaves were gathered from. Bark is collected at a
height of ca 1 m above ground level; care should be taken not to damage the
phloem, which is living conducting tissue.
Samples taken from trees are placed in linen or paper bags. Each sample
must be described (sampling date, characteristics and tree number) and the
identification tag placed in the bag.
When sampling leaves and bark of trees and bushes in orchards or parks, the
procedures are as for forests; samples of herbaceous plants are gathered in the
same manner as plants growing wild. The sampling procedures for fruit depend
on the aim of the studies and should be developed separately for each case.
In the laboratory material should be cleaned and, if needed, dried and reduced in
size. Roots and bulbs should be washed with running water to remove soil
completely and then rinsed with distilled water.
The method of preparation of above-ground parts is strictly related to the
aim of the studies. If a polluting substance taken in by plants is to be determined,
then above-ground parts must be washed before drying. The same amounts of
distilled water should be used for samples being compared. When the effect of
pollutants on plants is to be determined, then samples are not washed because
substances on the plant surface may affect conditions of plant growth, e.g. by
reducing the rate of photosynthesis.
80
Regardless of whether they are washed or not, plants can be dried, first in air
in well ventilated rooms to constant mass and then in driers or directly in driers
with air circulation and temperature control. In the beginning the temperature
should be kept at 30°C; at the end it can be raised maximally to 60°C. Roots,
bulbs and thick stems should be cut to shorten drying time. Dried material is
milled to dust of uniform particle size. The samples obtained should be stored in
tightly closed glass jars for further study.
Generally, sampling sediments requires much greater care than sampling other
objects. As always, the sample must resemble the original population as closely
as possible. The sampling device should be lowered slowly to avoid setting up
currents or causing shock waves that might disturb fresh sediments. Samplers
specially designed for the purpose are available and new designs are proposed in
the specialist literature [32,33]. Those commercially available are dredges,
grabs, and corers [34]. Dredges are used only for qualitative studies due to the
impossibility of controlling depth or location of the sampler. Grab samplers are
used to collect surficial samples 1-3 cm thick, which can be used for spatial
variability of a given property. Corers can be applied to obtain both surficial as
well as sediment column samples. With corers mud can be collected with the
least amount of disturbance.
Determination of some properties, e.g. total parameters such as chemical
oxygen demand or total organic content in sediments or solid wastes, requires
not only a proper sampler to collect representative samples but also the use of
correct sample preparation technique before the analysis proper [35].
The problem of sediment sampling has increased in significance because of
growing interest in the relationship between the quality of aqueous environ-
ments and chemical composition of sediment surface layers. The rule is to select
a sampling tool and technique which can give sediment samples with undis-
turbed surface layer structure. Such requirements are satisfied, e.g. by the
Niemisto corer [36] which is commonly used in Baltic studies to collect soft sedi-
ment samples. It consists of a metal tube, inside which a smaller diameter tube
made of plexiglass is fitted. It has stabilising fins near the top which assist verti-
cal penetration into the mud. The auto-valve is open during free fall and sedi-
ment penetration. The plunger is pushed into its machined seat; a vacuum
develops and prevents loss of the sediment on retrieval. This sampler is pre-
sented in Fig. 3.10.
Besides corers, different scoops such as the van Veen grab [37] (Fig. 3.11), or
the McIntyre-Day scoop [37] are used. They give samples whose structure is
partially disturbed during closing and raising the grab. After opening the grab
the sample is collected with a spoon or spatula, scraping the surface layer of
required thickness. Samples from a depth of 40 m or less can be taken by a
81
a) b) c)
Left: Fig. 3.10. Niemisto corer. : Stabilising fins, 2: catch, 3: metal tube with inner tube made of
plexiglass (based on Ref. [36]).
Right: Fig. 3.11. Van Veen grab used for sampling sediments (based on Ref. [37]). (a) Before
lowering, (b) at the bottom, (c) during withdrawing. 1: Stainless steel jaws, 2: covers, 3: lines
opening covers, 4: central weight, 5: shutting mechanism, 6: weight on longer lever arm.
skin-diver who carefully drives a sampler into the bottom or scrapes the surface
layer of sediment.
For coarse-grained, gravelly and stony sediments, sufficiently heavy grabs
should be used, e.g. the McIntyre-Day grab. Obtaining repeatable results is
difficult in the case of such sediments. Coarse-grained fractions are discarded
before analysis by dry or wet sieving. The most representative results are
obtained when the fraction below 63 Cm is subjected to analysis. This fraction
accumulates organic and inorganic compounds and is very useful in ecological
studies.
Figure 3.12 is a schematic of a cryogenic sampler for sediments [38]. The
sampler is composed of a cold finger (5 cm in diameter) and protecting tube (i.d.
10 cm). The protection tube is used to thermally insulate a frozen sample. The
cold finger must be filled with freezing mixture (dry ice + methanol) before it is
pushed into the sediment layer. Immediately afterwards the protecting tube
must be slid over the finger to thermally insulate the sample. After a freezing
time of 10-40 min, the sampler is pulled out and the sample collected from the
finger.
Sediments are stored wet in glass or metal vessels at a temperature of-20OC.
Drying, if needed, is done by lyophilisation.
Figure 3.13 presents a sampler with a shutter closure [38]. It consists of a
rectangular tube made of plastic, whose bottom end can be closed with a shutter.
The lower part of the tube is divided into sections 5 cm high to collect sample
layers.
82
Fig. 3.12. Cryogenic sampler for sediments. 1: Cold finger, 2: connector pipe, 3: extension rod, 4:
connector pipe supplying cooling medium, 5: rubber flange, 6: ring for fastening transport line, 7:
stiffener, 8: protecting pipe, 9: ring seal, 10: rubber ring (based on Ref. [38]).
a) b)
Fig. 3.13. Shutter sampler for sediments. (a) Rear view, (b) end view. 1: Rod for manipulating
with closure, 2: tube part for fractionation, 3: shutter (based on Ref. [38]).
83
et al. 40]. An appraisal of a simple sampling device for collected time-integrated
fluvial suspended sediment samples was given by Russell et al. [41].
As in the case of soils, special care is needed when sampling sediments for
analysis of volatile compounds. An important measurement for aquatic systems
is determination of bubble gas, whose composition can change during migration
from the sediment through the water column. To obtain gas directly from
profundal sediments of aquatic ecosystems a special sampler was proposed by
Huttunen et al. [42].
To get a sample representative of the whole population studied, and proper
sampler must be selected and sampling must be properly designed. For instance,
the effects of sampling design on estimating the magnitude and distribution of
contaminated sediments in a large reservoir were described by Pearson and Rose
[43].
3.8 GLOSSARY
84
Segregation: a process or state in which a substance or particles undergo
distribution throughout a bulk material in a non-random way.
Test sample: a sample prepared from a laboratory sample, e.g., by means of
grinding, reduction, etc. Portions of test sample, termed test portions, are
taken for single measurements
Unit sample: a portion of a material composed of all primary samples collected
from one package.
REFERENCES
85
24 T.E. Lewis, A.B. Crockett and R.L. Siegrist, Environ. Monitor. Assess., 30 (1994) 213.
25 A.D. Hewitt, T.F. Jenkins and C.L. Grant, Am. Environ. Lab., February 1995.
26 S.P. Theocharopoulos, G. Wagner, J. Sprengart, M-E. Mohr, A. Desaules, H. Muntau,
M. Christou and P. Quevauviller, Sci. Total Environ., 264 (2001) 63.
27 Polish Technical Standards: BN-74/9180-02, Analiza chemiczno-rolnicza gleby,
Pobieranie pr6bek (Chemical-agricultural analysis of soil. Sampling) (in Polish).
28 R.L. Siegrist and P.D. Jenssen, Environ. Sci. Techn., 24 (1990) 1387.
29 J. Siepak (Ed.), Fizykochemiczna analiza w6d i grunt6w, University of Adam
Mickiewicz, Poznan, 1992 (in Polish).
30 H.H. Rump and H.H. Kirst, LaboratoryManualfor the Examination of Water, Waste
and Soil. VCH, Weinheim, 1988.
31 J. Namiesnik, J. Lukasiak and Z. Jamr6giewicz, Pobieraniepr6beksrodowiskowych
do analizy. Wydawnictwa Naukowe PWN, Warszawa, 1995 (in Polish).
32 B.E. Markert, M.G. Tesmer and P.E. Parker, Water Res., 17 (1983) 603.
33 A. Murdoch, S.D. MacKnight, CRC Handbook of Techniques for Sediment Sampling.
CRC Press, Boca Raton, Florida, 1991.
34 U.M. Cowgill, Sampling and analysis of lake sediments. In: B. Markert, (Ed.),
Environmental Sampling for Trace Analysis. VCH, Weinheim, 1994.
35 S.D. Wachasunder and S. Satyanarayan, Ind. J. Environ. Protect., 9 (1989) 19.
36 L. Niemistb, A gravity corer for studies of softy sediments, Merentutkimuslait, Julk./
Harseforskiningsinst.Skr. 238, 1974.
37 N.A. Holme and A.D. McIntyre (Eds.), Methods of the Study of Marine Benthos.
Blackwell Scientific, Oxford and Edinburgh, 1971.
38 E. Ristenpart, R. Gitzel and M. Uhl, Water Sci. Technol., 25 (1992) 63.
39 A.J. Horowitz, A Primer on Sediment-Trace Element Chemistry. Lewis Publishers,
Chelsea, 1991.
40 A.J. Horowitz, F. Rinella, P. Lamothe, T. Miller, T. Edwards, R. Roch and D. Rickert,
Environ. Sci. Technol., 24 (1990) 13113.
41 M.A. Russell, D.E. Walling and R.A. Hodgkinson, The Role of Erosion and Sediment
Transport in Nutrient and Contaminant Transfer. Proceedingsof a Symposium held
at Waterloo, Canada, July, 2000. IAHS Publ. No. 263, 2000.
42 J.T. Huttunen, K.M. Lappalainen, E. Saarijarvi, T. Vaisanen and J. Martikainen,
Sci. Total Environ., 266 (2001) 153.
43 S.M. Pearson and K.A. Rose, Environnmetrics, 12 (2001) 81.
86
Chapter4
4.1 INTRODUCTION
Often too little attention is given to the handling of environmental samples after
their collection and before actual instrumental analysis. This could lead to
errors, despite increasingly sensitive analytical systems. How and for how long
different samples can be stored to preserve their integrity for chemical, physical,
and biological testing is the subject matter of this chapter. This issue is particu-
larly important for the measurements of various organic compounds in ambient
air or in emission sources. Organic compounds are often highly reactive, some-
times unstable, and exhibit a wide range of volatility and are hence distributed
in the atmosphere between the gas and particle phases. Compounds having satu-
rated vapor pressure at 25°C greater than 10' mm Hg are generally classified as
volatile organic compounds (VOC) and are present entirely in the gas phase.
Compounds with saturated vapor pressure at 25°C between 10-1 and 10 - 7 mm Hg
are generally called semi-volatile organic compounds (SVOC) and are distrib-
uted between the gas and particle phases. Compounds having saturated vapor
pressure at 25°C less than 10 - 7 mm Hg are nonvolatile and are predominantly
adsorbed on particles. Different sampling techniques are required for the quan-
titative collection of VOC, SVOC, and nonvolatile organics (see, for example,
Zielinska and Fujita [1] for a review of sampling techniques). The storage and
transport requirements for samples collected by each of these techniques are
briefly reviewed below.
Electropolished canisters have been used for more than two decades to collect air
samples; they can hold sufficient sample for replicate analyses and are reusable.
Made of stainless steel, canisters passivated by the patented Summa® process
provide a stable environment for a large number of volatile organic compounds
(VOCs). They have been widely used for sampling source and ambient air, and
their use is recommended in the USEPA Method TO-14A [2]. It has been found
repeatedly that humid samples provide improved reproducibility and storage
ComprehensiveAnalytical ChemistryXXXVII
J. Pawliszyn (Ed.)
© 2002 Elsevier Science B.V. All rights reserved 87
over dry. Water is thought to occupy active sites on the canister walls, increasing
the passivity of the surface.
88
synthetic mixture of ten polar organic compounds (methanol, acetone, isoprene,
acrylonitrile, vinyl acetate, methyl ethyl ketone, t-butyl methyl ether, ethyl
acetate, n-butanol, and ethyl acrylate) and two nonpolar compounds (methyl
chloroform and toluene, as controls) in unpolished and polished canisters under
humid and dry conditions over a 31-day period. They found good recovery (100 ±
10%) with exception of methanol (120% after 7 days and 145% after 31 days) in
humid canisters (passivated and unpassivated). The stability of the polar
compounds in the dry canisters was found to be very poor. Kelly et al. [11]
examined the stability of a humidified synthetic mixture of 16 compounds
(1,3-butadiene, methanol, ethanol, acetonitrile, acetone, 2-propanol, acrylonitrile,
methyl t-butyl ether, vinyl acetate, methyl ethyl ketone, ethyl acetate, ethyl
t-butyl ether, benzene, 1-butanol, ethyl acrylate, and toluene) over a 12-day
period, and observed good precision for most of the compounds (with a coefficient
of variation below 25%), with the exception of vinyl acetate and the alcohols.
Batterman et al. [12] investigated a synthetic mixture of seven aldehydes and four
terpenes in electropolished humidified and dry canisters for 16 days. They found
poor stability in all canisters studied, although humidified samples were reproduc-
ible for a longer period than those in dry environments. In the study of Goliff and
Zielinska [13], the stability of oxygenated compounds in whole air samples in
electropolished canisters was investigated. Three canisters-one each spherical
and cylindrical 6-liter, and one cylindrical 3-liter-were used to collect whole air
samples. Their contents were tested over a four-month period for stability of
selected oxygenated compounds, including acetaldehyde, methanol, acetone,
2-propanol, MTBE, hexanal, octanal, nonanal and benzaldehyde. They found
hexanal and MTBE to be the only oxygenated compounds to show good stability
(relative standard deviation less than 0.10) in all three canisters, while methanol
and acetone displayed the poorest precision.
Gholson et al. [14] studied sample storage of VOCs in aluminum canisters,
passivated and non-passivated, and with and without water. They' found that the
stability of VOCs in aluminum canisters was dependent on the humidity of the
sample and the reactivity of the compounds analyzed. The authors recom-
mended against the usage of aluminum canisters for polar compounds due to
their decreasing concentrations over time. In the USEPA Method TO-14A,
stainless steel canisters are recommended for sampling.
The other method of choice for sampling of VOCs in air is collection on solid
adsorbents, where they are removed from possible reaction with other collected
species. A method of preconcentration followed by separation by gas chromatog-
raphy has been standardized by the U.S. EPA (TO-1) [15] for sample collection
and analysis using Tenax, a polymer of 2,6-diphenyl-p-phenylene oxide for
hydrocarbons and halocarbons from C6 to C20 , and by USEPA Method TO-17 [16]
using Carbotrap, Carbosieve, and other adsorbents for more volatile compounds.
89
While there have been many publications regarding artifacts and degradation of
species with a variety of adsorbents, such as Tenax GC, Tenax TA, Carbotrap
and activated carbon [1,17-23], there are relatively few concerning sample
storage. The US EPA Method TO-17 recommends analysis of adsorbent samples
within two weeks of sampling.
Rudling et al. [24] investigated the stability of organic solvents on two
activated carbons, SKC and Merck, during storage times of up to 50 days.
Ketones showed large losses, which the authors attributed to reaction with the
adsorbent material. Compounds which were found stable (less than 5% loss) on
both carbons included toluene, styrene, and 1,1,1-trichloroethane.
Schrader et al. [25] studied the degradation of a-pinene on Tenax-TA under
different storage conditions, including samples placed in sunlight, samples
wrapped in aluminum foil, storage in a refrigerator at 4C, and storage in a
freezer at -18°C. After one month of storage, wrapped samples showed over 99%
recovery. Analysis of samples exposed to daylight showed evidence of degrada-
tion after one day. At one month, 12% of the o-pinene had reacted. The authors
found that storage in a freezer to be most effective in reducing the degradation of
c-pinene.
Zielinska et al. [9] investigated storage stability of hydrocarbons from C to
C1 9 in Tenax samples over a storage period of 7 to 8 months. They found that the
difference in total concentrations was only +5% to -12% (the negative value
indicating a decrease in concentration), but the differences in the concentrations
of the less abundant hydrocarbons were much higher: up to -50%.
There have been some preliminary studies performed by Goliff and Zielinska
[26] looking at sample stability of Tenax-TA stored for more than six months.
Storage conditions were as follows. Sampling tubes were capped with
Swagelok® fittings and then wrapped in aluminum foil. The tubes were then
placed in containers along with activated charcoal. The containers of sampling
tubes were stored in freezers at -20°C. Initial findings suggest good sample
integrity under these storage conditions, including the lightest compound: ethyl
benzene. There was one anomalous sample that showed large losses, which the
authors attribute to sampling error.
90
360-375 nm range and can be analyzed by the high performance liquid
chromatography (HPLC) method coupled with UV detection; this method offers
very high selectivity and sensitivity of analysis. The US EPA Method TO-11A
[27] and TO-5 [28] describe the collection and analysis of carbonyl compounds in
ambient air, using either DNPH impregnated cartridges (silica gel or C,,), or
impinger, respectively.
Hydrazones of stable carbonyls, such as formaldehyde, acetaldehyde, pro-
pionaldehyde, acetone, butanal and butanone, maintain their integrity on silica
gel and C18 cartridges, as well as in impinger reagent solutions, for over a month
under refrigerated storage [29,30]. The stability of these hydrazones in the
cartridges at ambient temperature has been also demonstrated [31] by the
collection of duplicate samples in the field. One sample from each duplicate pair
was retrieved and analyzed immediately after collection, while the other
remained in the sampler until it was retrieved during the normal weekly visit by
the field technician. The excellent agreement between the pairs confirmed that
after collection the samples were stable during the seven days of holding time
encountered in the study.
However, it has been reported [30] that the olefinic aldehydes such as
acrolein and crotonaldehyde degrade partially on the cartridges, either during
sampling or storage, and form unknown products. Also, the acrolein-DNPH
derivative is not stable in the strongly acidic DNPH solutions used in the
impinger method. In order to preserve its integrity it has to be extracted from
this solution immediately after sampling [32]. The EPA Method TO-11A [27]
recommends a maximum storage time of 2 weeks at or below 4°C for samples
collected on silica gel or C cartridges. However, the extracts may be stored
refrigerated for a longer period, up to approximately one month.
A study was performed by Fung et al. [33] to assess the effectiveness of vari-
ous types of packing used to protect the DNPH cartridges from contamination
during storage and shipment. In general, the amount of contamination was
reduced significantly by using a protective closure as opposed to an open car-
tridge. Even the PVC caps, the worst performers of the group, reduced formalde-
hyde exposure to approximately 2% of an open cartridge. For formaldehyde, a
polyethylene cap provided ten times more protection than PVC caps and about
twice as much protection as plugs. Screw-capped vials offer additional protec-
tion. Samples should be shipped from the field tightly capped, packed in screw-
capped vials, and chilled with blue ice (or equivalent).
Exposure to sunlight causes significant production of carbonyls, which can
be eliminated by wrapping the cartridges in aluminum foil during sampling and
storage [34].
91
adsorbent, such as polyurethane foam (PUF) plugs, or XAD-2 or 4 resins, or
"sandwich" type PUF/XAD/PUF cartridges. This method of collection is used for
organic compounds, which are distributed between gas and particle phases, such
as polycyclic aromatic hydrocarbons (PAH), polychlorinated dibenzo-p-dioxin
and furans, pesticides, polychlorinated biphenyls, etc. The US EPA compendium
[35] of methods for the determination of toxic organic compounds in ambient air
describes the standardized procedures for the collection and analysis of some of
these SVOC (Methods TO-4A [36], TO-9A [371, TO-10A [38], TO-13A [39]).
These methods also specify the holding time and the storage requirements for
the specific samples. For example, the U.S. EPA Method TO-13A requires the
extraction of samples collected for PAH analysis within seven days from sample
collection if stored in a refrigerator at or below 4°C. The extracts should be
analyzed within 40 days of extraction. In addition, since PAH are chemically
reactive, and heat, ozone, NO2 , and UV light may cause sample degradation
during transport and analysis, it is recommended that samples are shipped
chilled (in blue ice) overnight from the field to the analytical laboratory and
incandescent or UV-shielded fluorescent lighting should be used in the
laboratory during sample preparation and analysis.
The maximum seven-day requirement for sample holding time is very
stringent for most analytical laboratories. Some studies have suggested that the
holding time can be extended to a month or more if samples are stored in much
lower temperatures in proper containers. Svaderup et al. [40] investigated
several different sample packing techniques and the subsequent stability of the
surrogate samples under different storage conditions. Surrogate samples for
power plant plume fly ash were created by using electrostatic precipitator
hopper ash. Particles smaller than 10 Am were suspended and mixed with three
PAH, one nitro-PAH, salt, and sulfuric acid. The particle mixture was then
collected on quartz filters to create surrogate samples. Samples were stored for
periods of 30 and 120 days at either 20 or -79°C, as well as under a set of variable
conditions of temperature, light, and humidity. Samples were then analyzed by
gas chromatography/mass spectrometry (GC/MS) and scanning electron
microscopy and compared with samples analyzed right after collection (day 0).
The authors concluded that preservation of organic compounds in the samples
was best achieved when the samples were packed in Teflon-wrapped glass dishes
and stored at -79°C in the dark. Under these conditions, loss of the lightest PAH
(fluorene) was minimized and the heaviest PAH was stable for at least 120 days.
However, sulfuric acid had a deleterious effect on two of the spiked PAI. For
nitro-PAH (2-nitrofluorene) the results were inconclusive due to the unequal
loadings on the filters.
Chuang at al. [41] reported the results of storage stability study of PAH on
quartz filters and two different sorbents: PUF and XAD-2. Air samples were col-
lected on quartz filters followed by PUF cartridges or XAD-2 resin. Prior to sam-
pling, the sorbent was spiked with known amount of perdeuterated PAH. After
sampling, the samples were stored in the dark at room temperature (--20"C) for
92
0, 10-, 20-, and 30-day intervals, and at -20°C for 10- and 20-day intervals. The
results showed no significant loss of PAH on XAD-2 after 30 days of storage in
the dark at room temperature, but decreasing concentrations of naphthalene,
anthracene, and benzo(a)pyrene on PUF. Storage losses of PAH collected on
quartz filters were insignificant under these conditions, with the exception of
cyclopenta[c,d]pyrene, which decreased to approximately half of the initial
amount. Levels of its degradation product, pyrene dicarboxylic acid anhydride,
increased more than 1.5 times after 30 days of storage. In addition, levels of 2-
and 3-nitrofluoranthene decreased more than 30% of the initial amount. How-
ever, when samples were stored at -20°C in the dark, all target PAH and nitro-
PAH were stable on quartz filter and XAD-2 cartridges when analyzed after 10
and 20 days of storage.
For analysis of polychlorinated/polybrominated dibenzo-p-dioxins and
furans, Method TO-9A specifies a maximum holding time of 7 days from sample
collection to extraction when stored at or below 4C. The extract should be
analyzed within 40 days of extraction. However, since dioxins and furan are
chemically stable, it is possible that storage in much lower temperatures to
prevent losses due to evaporation would result in much longer safe sample
holding times. EPA Methods TO-4A and 10A, describing the determination of
the broad range of pesticides and polychlorinated biphenyls by high- and low
volume sampling, respectively, requires sample extraction within seven days of
collection when stored at or below 4°C, and analysis of the extract within 40 days
of extraction. The ASTM document D-4861-00, Appendix X3 [42] gives data on
the storage stability of various common pesticides on PUF cartridges, when
stored at room temperature (24°C) for 30 days. With the exception of several
pesticides (for example, Ronnel, Malathion, Chlorothalonil) the losses are in the
range of 0-40%. The ASTM document recommends that the cartridges be stored
at water ice or dry ice temperatures immediately after collection in the field and
at -10°C or lower temperature in the laboratory freezer for no longer than 30
days prior to extraction.
In addition to the US EPA Methods for ambient air monitoring, the National
Institute for Occupational Safety and Health (NIOSH) publishes the Manual of
Analytical Methods [43], suitable for the determination of various toxic
chemicals concentrations in occupational environment. In general, these
methods use solid adsorbents such as PUF, XAD, activated charcoal or Tenax,
and sometimes, filters followed by solid adsorbents. Thus, the same sample
holding time considerations would apply to these samples. It has been reported
[44,45] that organophosphorus and organonitrogen pesticides collected on
OSHA versatile sampler tubes (OVS, consisted of XAD-2 resins) were stable
when stored for up to 30 days under either ambient or refrigerated conditions.
Although the authors did not indicate the necessity for sample refrigeration,
they noted minor improvements in sample recovery at the 21- and 31-day
storage time in some instances. Therefore, refrigeration of samples prior to
analysis was recommended.
93
TABLE 4.1
Recommended storage conditions for sampling media
Sampling media Temperature (EPA) Duration of storage (EPA) Lighting
requirements
Canisters At room temperature 6 months (1 month) Not applicable
Solid adsorbents -20°C (-20°C) 8 months (2 weeks) In the dark
DNPH cartridges (<4°C) (2 weeks) In the dark
Filter-PUF-XAD (<4'C) 7 days In the dark
4.6 SUMMARY
REFERENCES
1 B. Zielinska and E. Fujita, In: B. Market (Ed.), Environmental Sampling for Trace
Analysis. WCH, Weinheim, 1994, Chapter 7.
2 US EPA, Method TO-14A, In: Compendium of Methods for the Determination of
Toxic Organic Compounds in Ambient Air, 2nd Edn. EPA/625/R-96/010b, January
1999.
3 R.A. Rasmussen and J.E. Lovelock, J. Geophys. Res., 88 (1983) 8369-8378.
4 H. Westberg, W. Lonneman and M. Holdren, In: L.H. Keith (Ed.), Identification and
Analysis of OrganicPollutantsin Air. Butterworth, Woburn, MA, 1984, pp. 323-337.
5 K.D. Oliver, J.D. Pleil and W.A. McClenny, Atmos. Environ., 20 (1986) 1403-1411.
6 D.A. Brymer, L.D. Ogle, C.J. Jones and D.L. Lewis, Environ. Sci. Technol., 30 (1996)
188-195.
7 J.C. Evans, J.L. Huckaby, A.V. Mitroshkov, J.L. Julya, J.C. Hayes and J.A. Edwards,
Environ. Sci. Technol., 32 (1998) 3410-3417.
8 W.S. Goliff and B. Zielinska, Paper 1122, Proceedings of the Air &Waste Manage-
ment Association's 93rd Annual Conference &Exhibition, Salt Lake City, Utah, June,
2000.
9 B. Zielinska, J.C. Sagebiel, G. Harshfield, A.W. Gertler and W.R. Pierson, Atmos.
Environ., 30 (1996) 2269-2286.
10 B. Pate, R.K.M. Jayanty, M.R. Peterson and G.F. Evans, J. Air Waste Manage.
Assoc., 42 (1992) 460-462.
11 T.J. Kelly, P.J. Callahan, J. Plell and G.F. Evans, Environ. Sci. Technol., 27 (1993)
1146-1153.
12 S.A. Batterman, G.-Z. Zhang and M. Baumann, Atmos. Environ., 32 (1998)
1647-1655.
13 W.S. Goliff and B. Zielinska, Paper 403, Proceedings of the Air &Waste Management
94
Association's 9 4 th Annual Conference &Exhibition, Orlando, Florida, June 2001.
14 A.R. Gholson, J.F. Storm, R.K.M. Jayanty, R.G. Fuerst, T.J. Logan and M.R.
Midgett, JAirPollut. Contr. Assoc., 39 (1989) 1210-1217.
15 US EPA, Method TO-1, In: Compendium of Methods for the Determinationof Toxic
Organic Compounds in Ambient Air, 2nd Edn. EPA/625/R-96/0
10b, January 1999.
16 US EPA Method TO-17. In: Compendium of Methods for the Determinationof Toxic
Organic Compounds in Ambient Air, 2nd Edn. EPA/625/R-96/010b, January 1999.
17 G. MacLeod and J.M. Ames, J. Chromotogr., 355 (1986) 393-398.
18 B. Zielinska, J. Arey, T. Randahl, R. Atkinson and A.M. Winer, J. Chromatogr.,363
(1986) 382-386.
19 H. Rothweiler, P.A. Wager and C. Schlatter, Atmos. Environ., 25B (1991) 231-235.
20 X.-L. Cao and C.N. Hewitt, Chemosphere, 27 (1993) 695-705.
21 X.-L. Cao and C.N. Hewitt, Atmos. Environ., 27A (1993) 1865-1872.
22 A. Calogiroou, B.R. Larsen, C. Brussol, M. Duane and D. Kotzias, Anal. Chem., 68
(1996) 1499-1506.
23 C. Coeur, V. Jacob, I. Denis and P. Foster, J. Chromatogr. A, 786 (1997) 185-187.
24 J. Rudling, E. Bjorkholm and B.-O. Lundmark, Ann. Occup. Hyg., 30 (1986) 319-327.
25 W. Schrader, J. Geiger, D. Klockow and E.-H. Korte, Environ. Sci. Technol., 35
(2001) 2717-2720.
26 W.S. Goliff and B. Zielinska, private communication.
27 US EPA Method TO-11A, In: Compendium of Methods for the Determinationof Toxic
Organic Compounds in Ambient Air, 2nd Edn. EPA/625/R-96/010Ob, January 1999.
28 US EPA Method TO-5, In: Compendium of Methods for the Determinationof Toxic
Organic Compounds in Ambient Air, 2nd Edn. EPA/625/R-96/10Ob, January 1999.
29 K. Fung, M. Porter, D. Fitz and E.M. Fujita, PlanningAir QualityMeasurement and
Modeling Studies: A Perspective Through SJVAQS/AUSPEX, P. Solomon (Ed.), Air
&Waste Management Association, Pittsburgh, PA, 1993.
30 A. Vairavamurthy, J.M. Roberts and L. Newman, In: E.D. Winegar and L.H. Keith
(Eds.), Sampling andAnalysis of AirbornePollutants. Lewis Publishers, Boca Raton,
FL, 1993, Chapter 10, pp. 149-174.
31 K. Fung and B. Wright, Aerosol Sci. Technol., 12 (1990) 44.
32 R.R. Freeman, Proceedingsof the Third Annual West Coast Regional Air & Waste
Management Association Conference On Current Issues in Air Toxics, L. Edwards
(Ed.), Air &Waste Management Association, Pittsburgh, PA, 1992, pp. 48-53.
33 K. Fung, M. Porter, D. Fitz and E. Fujita, Planning and Managing Regional Air
Quality Modeling and Measurement Studies-A Perspectivethrough the San Joaquin
Valley Air Quality Study and AUSPEX, P.A. Solomon and T.A. Silver (Eds.), Lewis
Publishers with Pacific Gas and Electric Company, 1995.
34 X. Zhou and K. Mopper, Environ. Sci. Technol., 24 (1990) 1482-1485.
35 US EPA, Compendium of Methods for the Determination of Toxic Organic Com-
pounds in Ambient Air, 2nd Edn. EPA/625/R-96/010Ob, January 1999.
36 US EPA Method TO-4A, In: Compendium of Methods for the Determinationof Toxic
Organic Compounds in Ambient Air, 2nd Edn. EPA/625/R-96/O10b, January 1999.
37 US EPA Method TO-9A, In: Compendium of Methods for the Determinationof Toxic
Organic Compounds in Ambient Air, 2nd Edn. EPA/625/R-96/010b, January 1999.
38 US EPA Method TO-10A, In: Compendium of Methods for the Determinationof Toxic
Organic Compounds in Ambient Air, 2nd Edn. EPA/625/R-96/01Ob, January 1999.
39 US EPA Method TO-13A, In: Compendium of Methods for the Determinationof Toxic
Organic Compounds in Ambient Air, 2nd Edn. EPA625/R-96/010OlOb, January 1999.
40 G.M. Svaderup, B.E. Buxton and J.C. Chuang, Environ. Sci. Technol., 24 (1990)
1186-1195.
41 J.C. Chuang, N.K. Wilson and R.G. Lewis, FreseniusEnvir. Bull., 8 (1999) 547-556.
95
42 American Society for Testing and Materials (ASTM), Standard practice for sampling
and selection of analytical techniques for pesticides and polychlorinated biphenyls in
air. In: Annual Book of ASTM Standards, D-4861-00, 2000.
43 M.E. Cassinelli and P.F. O'Connor (Eds.), NIOSH Manual of Analytical Methods
(NMAM), 4th Edn. DHHS (NIOSH) Publication 94-113, August 1994.
44 E.R. Kennedy, M.T. Abell, J. Reynolds and D. Wickman, Am. Ind. Hyg. Assoc. J., 55
(1994) 1172-1177.
45 E.R. Kennedy, J. Lin, J.M. Reynolds and J.B. Perkins, Am. Ind. Hyg. Assoc. J., 58
(1997) 720-725.
96
Chapter 5
5.1 INTRODUCTION
During the last two decades, there has been an increasing realization of the
importance of trace gases and aerosols: they play a significant role in almost all
areas of air quality related concerns, including photochemical smog, strato-
spheric ozone depletion, global climate change, indoor air quality, public and
occupational health, and the genesis of atmospheric acidity and its effects. The
implications of trace gases and aerosol particles on atmospheric visibility span
the gamut from esthetic values of clear sight across the Grand Canyon to the
successful mission of an optically guided armament.
Both gases and particles can be organic or inorganic. In the gas phase, the
dominant components of interest in ambient air are typically simple, small
molecules, with molecular weight (MW) usually <100. Ambient particulate
matter is a more complex mixture than the air it is suspended in. Although
atmospheric fine particulate matter is also dominantly inorganic (urban fine
particles are typically composed of ammonium sulfate and/or nitrate), signifi-
cant quantities of organics, including large and complex macromolecules, are
often present in ambient aerosols in readily detectable concentrations. The
automated measurement of such compounds is a daunting task and not much
progress has yet been made.
The focus of this book is on sample preparation. The best sample preparation
technique is one in which sampling, sample preparation, and analysis are
coupled in an integral and automated manner. This reduces contamination
problems, labor, analysis cost per sample, and improves data quality, system
reliability and throughput.
5.1.1 Gases
Air samples are not spatially and temporally homogeneous. Moreover, the
aerosol may or may not be in equilibrium with the gas phase it is suspended in. If
a sample is collected in a composite manner, often the phase-specific origin of an
5.1.2 Aerosols
98
respectively. This situation largely reflects the capability of extant commercial
instrumentation. While it is obvious that chemical composition of the aerosols
must significantly influence any health effects they may elicit, there are still no
chemically specific regulations. In contrast, the NAAQS has been revised several
times from the original size-indifferent mass concentration parameter (referred
to as total suspended particulate matter or TSP) to a mass concentration
parameter of 50 g/m3 annual average and 150 Ag/m 3 24 h average for PM,, in
1987 and additionally, 15 Ag/m 3 annual average and 65 Ag/m 3 24 h average for
PM2 .5 in 1997 [4]. This has been possible due to the availability of automated
measuring instruments capable of providing size-fractionated mass measure-
ments, some on a near real time basis. A wealth of information on mass
concentration as a function of particle size is available (albeit to date, such data
primarily concerns the PM,, fraction but this is changing). The need for and the
approaches to automated aerosol composition measurement are addressed in a
separate review.
99
should also be borne in mind that the collection efficiency will decrease further
when every collision is not effective in uptake. This aspect has been quantita-
tively examined in many studies and has been adequately reviewed previously
[1]. The salient point is that a simple single tube denuder generally is a
quantitative collector only at relatively low flow rates, this may not be acceptable
in all applications.
100
where A and a are dependent on the effectiveness of wall collisions and are
generally determined empirically. In the case of perfect sink efficiency, the
values of 0.91 and 3.77 forA and a are respectively expected [20].
Glass annular denuders are commercially available from several vendors
[21]. One interesting addition to the annular denuder family is a carbon-coated
version; this is useful for the collection of NO 2 and peroxyacetyl nitrate [22]. The
gas collection efficiency and entrance flow effects in the annular denuder design
have been theoretically modeled and compared with experimental data [23].
Multiple concentric tubes can be thought of as the same geometric analog to
an annular denuder that a Gatling gun design is to a single tube denuder.
Coutant et al. [24] designed a 12-element concentric denuder for high volume
collection of organics. A program is available from the authors to compute the
collection efficiencies of such denuders (containing any number of elements)
where the wall reaction probability can be input as a separate parameter.
Mathematically, the annular denuder geometry assumes a parallel plate
configuration as the diameter of the tube(s) approaches infinity. Simon and
Dasgupta first reported this parallel plate denuder (PPD) geometry in a
continuously operating configuration [25]. For two parallel plates of active width
w separated by a distance s, the collection efficiency is given by [26]:
f = 1- 0.91 exp (-2.4wA/s)) (5.4)
Such a denuder is much more efficient than its circular counterpart. A denuder
with L = 30 cm, w = 5 cm and s = 0.3 cm will remove SO2 with an efficiency of
99.8% and 95. 2% at flow rates of 5 and 10 LPM, respectively.
To avoid the oddity of a rectangular aperture in a world of circular connect-
ing tubes, Staffelbach et al. [27] described a circular end parallel plate denuder
fabricated by squashing the middle, active, section of a circular tube into a
flattened, rectangular geometry while inlet/outlet connections were made to the
circular termini.
Multiple parallel plates instead of a single pair of active plates provide for very
high flow rates. Eatough et al. [28-33] have pioneered such denuders and incorpo-
rated them in integrated gas-particle differentiating systems and provided them
with memorable acronyms: The BOSS (Brigham Young University Organic
Sampling System) [28], the BIG BOSS (the BOSS capable of sampling at 250
LPM) [32], the PC (Particle Concentrator)-BOSS [33] and the RAMS (Real-Time
Ambient Mass Sampler) [33]. Except for the last device, measurement strategies
are off-line. The multiple PPD in BOSS was composed of seventeen 4.5x58 cm
strips of charcoal impregnated filter (CIF) strips, separated by 2 mm dia. glass
rods at the long edges and contained in a 5 x 5 cm square cross section Al tube. The
denuder is designed to remove vapor phase organic material and is followed by
quartz and CIF filters. Particle losses in the denuder are - <10%. After collection,
CIF and quartz filters are heated in a temperature-programmed manner in an N2
atmosphere. The evolved gases are catalytically oxidized to CO 2 , which is then
measured by a non-dispersive infrared analyzer.
101
The PC-RAMS consists of a particle concentrator inlet originally developed
by Sioutas et al. [34-36]. The PC is a slotted virtual impactor with a 50% cut
point at a mass median aerodynamic diameter (MMAD) of -0.1 m. Ahead of the
PC is an inlet that rejects large particles, with 50% rejection at an MMAD of -2.5
m. As the aerosol stream (14 LPM) enters the PC, 75% of the flow is drawn from
a direction perpendicular to the entrance. Particles - >0.1 ,tm cannot make the
sharp turn and proceed with the minor flow in the original flow direction
through a slot. Then the stream, four-fold enriched in 0.1- 2.5 ium particles,
flows sequentially through appropriately coated glass annular denuders which
remove NH3 , NO2, and 03, a multiple PPD with CIF strips, a Nafion dryer (that
operates with a countercurrent dry air flow) to dry the aerosol and finally a
sandwich filter (containing a Teflon coated filter followed by a CIF) mounted on
a tapered element oscillating microbalance (TEOM [37,38]) which thus
continuously monitors the dry aerosol mass.
Dasgupta et al. [39] described a multiple PPD housed within a 5 cm i.d. x 30
cm length shell. The plates of the 11-element denuder are composed of twin-ply
polyester fabric. The plates are unequal in width. A computational approach to
the collection efficiency expected from such a design was developed. Based on
this, the spacing between the plates was adjusted according to the widths such
that each individual flow channel was isoefficient. Wetting liquid was pumped
into each element resulting in a wetted area in excess of 0.1 m2. The liquid was
withdrawn from 2 V-grooves at the bottom of the denuder. The device closely
followed theoretical expectations and removed >95% of influent SO2 at a
sampling rate of 50 standard liters per minute (SLPM) at an ambient pressure of
680 mm Hg. Loss of 0.4 Jm MMAD particles was only -0.2%.
It is worth noting here that diffusion denuders are increasingly being used
for purposes other than atmospheric analysis. Lewis et al. used a diffusion
denuder technique to determine that 99% of the nicotine in cigarette smoke is in
the aerosol, rather than the gas phase [40]. Vassilev et al. have used diffusion
denuders to determine reaction isotherms [411.
5.2.2.1 Thermodenuders
The ability to perform collection and analysis in a coherent integrated auto-
mated manner is the hallmark of a technique that others want to use. The long
history of the use of sampling for organics through a sorbent bed and subsequent
thermal desorption into a gas chromatograph has logically led to thermo-
denuders-sorbent-coated denuders that are thermally desorbed and reused.
Indeed, as discussed elsewhere in this volume, this is the mainstay of SPME as
well. Thermodenuders were originally developed for the thermally reversible
sorption of gases. Subsequently, speciation between aerosol H2 SO4 and various
sulfate salts was made possible by passing the sample through a controlled tem-
perature tube prior to a denuder. The transit temperature was so adjusted as to
102
volatilize sulfuric acid to the vapor phase and thus remove it by the denuder
while the other aerosol sulfur compounds passed through. The previously men-
tioned Krieger-Hites Gatling Gun design [10,11] is configured as a thermo-
denuder.
However, for most small inorganic gases, the thermodenuder approach is not
very attractive because of (a) the difficulty of recovering the collected gas in an
unambiguously identifiable form, (b) interferences from other analytes, (c) large
power requirements in the field, and (d) significant cool down periods before the
denuder can be reused. Since the time of the last review [1], only a handful of rele-
vant articles have appeared: a review [42], a comparative evaluation of different
NH3 monitors [43] (including a thermodenuder based instrument; only a continu-
ous wet effluent denuder based instrument was found suitable for the intended
task), the use of thermodenuders in studying particles present in internal combus-
tion engine exhaust [44,45], a study involving continuous monitoring of sulfate
aerosols and gaseous SO, (largely using established principles) [46], and a tech-
nique for collecting HCHO [47].
The last named technique collects HCHO as a derivative of 2-hydroxy-
methylpiperidine (coated on the walls of a single tube denuder) followed by
off-line thermal desorption and gas chromatography- mass spectrometry (GC-
MS) analysis. The basic problem with most approaches that uses a derivatization
agent followed by thermal desorption is that the derivatization agent does not
generally survive the thermal treatment. Consequently, automated reuse is dif-
ficult. This trait limits the use of derivatization-based thermodenuders and is
not likely to change in the future.
103
put in line, it could be used for much longer periods. Conductivity detection is
not specific, and it is to this end that the authors have shown considerable
chemical ingenuity. The air sample is drawn through a 0.2 M sulfamic acid
solution that removes any NH3 or HCl. Nitrogen dioxide is removed by the
sulfamic acid, forming nitrogen. Weaker acid gases like CO2 are not absorbed by
the HO,. I estimate the limit of detection (LOD) to be - 1.5 parts per billion by
volume (ppbv) SO2 . The effective integration time in such a system will be
dictated by the equilibration time of the sulfamic acid solutions to the changing
SO2 concentrations. The choice of this pretreatment solution volume (which is
obviously subject to evaporation from the passage of the sample air) represents a
compromise between the continuous operational period and the response time.
More recently, a wetted single-tube glass denuder arrangement was reported
by Luo et al. [50] which is vaguely similar to the Gacs-Ferraroli approach. A
glass tube, 1 mm i.d., was coated in situ with a film of glycerol/phosphoric acid,
then air was sampled to collect NH3. The collected NH3 was next washed off
directly with reagents suitable for initiating the indophenol blue reaction.
Finally, the blue product was measured; all this was carried out with a single-
pump, sequential injection analysis system. Although the particular exposition
may not have been exemplary in all respects (e.g., collection of NH3 was less than
that would theoretically be expected, probably because the coating solution did
not form an uniform film on the untreated glass walls-less than theoretical
collection efficiency was also the case with the GAcs- Ferraroli approach), the
concept holds promise for a variety of applications and a variety of detection
systems.
104
DS devices are easily subdivided into different classes depending on the
nature of the membrane: (a) hydrophilic membranes such as Nafion® present a
wet surface to the sampled gas that readily captures water soluble gases, (b)
porous hydrophobic membranes such as porous poly(tetrafluoroethylene)
(PTFE, Gore-Tex, expanded PTFE (ePTFE), etc.), porous polypropylene
(CelgardT ', Accurel®, Metricel M etc.) in which the analyte gas molecules diffuse
through the pore space into the absorbing liquid on the other side, (c) nonporous
thin-walled membranes such as silicone or Teflon AF in which the analyte gas
molecules permeate through the membrane to be taken up by a suitable
absorber liquid on the other side. Asymmetric membranes, where very thin
nonporous skin is present on a porous support structure, constitute a subclass of
the last group. Although there are many industrial applications of asymmetric
membranes, I know of only one application of such a membrane for trace gas
collection and analysis. This, too, was a custom-made membrane [53].
With appropriate choice of sampling rate, dimensions and scrubber liquid, it
is possible to attain quantitative or near-quantitative collection of desired
analyte gases above for a and b type devices. In general, membrane transfer rate
for type c devices is much lower and quantitative collection is not attained at
practical flow rates. The intrinsic response time of these devices of the different
types also vary (where chemical kinetics to form the product that is detected and
liquid transport time to the detector are not limiting factors). Whereas type a
and c devices involve transport through the membrane matrix, type b devices
involve gas phase diffusion through pore space and this latter process is much
faster. Of course, if it is possible to build membrane devices with very thin
membranes, then type a and c devices can also have very fast response.
Collection efficiencies of DS based devices as a function of sampling rate
show an exponential dependence on the reciprocal of the sampling rate as in the
Gormley-Kennedy equation (Eq. (5.1)). Methods of numerical evaluation are
available for different geometries [1]. In essence, the analyte mass collected
initially increases linearly with sampling rate because collection efficiency is
(nearly) quantitative. Then the slope begins to decrease. At very high flow rates,
the analyte mass collected becomes virtually constant, independent of flow rate.
Obviously, if possible, there are particular advantages to operating in two
different flow regimes: (a) where collection is quantitative and (b) where the
total amount collected is flow-independent. Actual practice is typically dictated
by sensitivity needs.
Much of the initial work with diffusion scrubbers were conducted with ion
exchange membrane DS devices [1]. Of ion change membranes, only a cation
exchanger, Nafion®, is commercially available in tubular form in various
diameters. Both anion and cation exchange membranes are available in planar
format from various vendors. Early work on gas collection was carried out with
105
ion exchange membranes for collecting characteristic ionogenic gases (e.g.,
Nafion® was used with dilute H 2SO4 as the DS liquid for collecting gaseous NH 3
as NH4 , or a radiation-grafted custom anion exchange membrane tube with a
K2 SO4 solution as the DS liquid was used for collecting HNO, as NO,-) [51,54].
However, this mode of operation can lead to large integration times because the
analytes occupy ion exchange sites and must be displaced by similarly charged
ions in the DS liquid. As a result, the collected analyte ions move essentially in a
chromatographic fashion along the membrane. Clearly, the greater the affinity
of the analyte ion for the ion exchange sites, greater is the effective integration
time. It is possible to ameliorate this to a degree by using DS liquids that also
contain similarly charged ions with high affinities for the ion exchange sites
(e.g., C104- or Ca 2+) and at high concentrations. Blank contamination, an issue
with any reagent in high concentration, then becomes the limiting factor. Note
that Donnan exclusion prevents like-charged ions from penetrating a membrane
matrix; NO,- is not, for example, transferred through a cation exchange
membrane. For very weakly ionized species, this limitation may not hold; H2S,
for example, appears to be transported through base-treated Nafion®
membranes, but the reproducibility of such behavior has not been clearly
established [55].
As a result, the principal use of ion exchange membrane DS devices has thus
far been in the collection of small polar nonionic molecules, like HCHO or H2 02
as described in the next section.
106
6.00-
4.00-
3.00-
U 2.00-
1.00-
0.00-
I 3I 4I
10.00 20.00 30.00 40.00
Hours from 0000, June 1, 1995; Boulder, CO
Fig. 5.1. Formaldehyde data obtained by a Nafion membrane diffusion scrubber collector coupled
to CHD derivatization and fluorometry (NMDS-CHD TTU) and a tunable diode laser absorption
spectrometer operated by the National Center for Atmospheric Research (NCAR).
reaction is carried out for only a short period of time, the solution phase reaction
is > 300 times more selective for HCHO than the nearest congener CH3 CHO. For
a gaseous sample, CH3 CHO will be less efficiently collected than HCHO based
both on diffusion and solubility issues. In the majority of ambient applications,
HCHO concentrations far exceed that of CH 3 CHO (except in countries like
Brazil where ethanol is used as a major fuel for motor vehicles), this approach is
then nearly specific for HCHO. The instrument has a reported (S/N = 3) LOD of
8 parts per trillion by volume (pptv, 10 -12 atm). In a blind intercomparison study
conducted by the National Center for Atmospheric Research (NCAR) for various
HCHO measurement techniques, the instrument produced ambient data
essentially identical to the referee instrument, a tunable diode laser absorption
spectrometer (TDLAS) operated by NCAR [59]. Figure 5.1 shows the correspon-
dence of ambient HCHO measurement results. (We note in passing that more
often than not, this degree of agreement is not observed in intercomparison
studies between two otherwise perfectly good instruments, due to a variety of
reasons.) A similar instrument was built by researchers at Purdue University
and used to measure formaldehyde concentrations during the 1998 polar sunrise
experiment [60]. A more detailed description of this instrument appears
elsewhere [61], this latter paper also describes the use of the instrument to study
HCHO chemistry above a forest canopy. More recently, the instrument has been
used to measure HCHO coming from the snowpack during the ALERT 2000
experiment (Alert, Nunavut, Canada) [62].
A newly designed instrument using the same basic collection principle and
reaction chemistry has been more recently reported [63]. A Nafion tubular
membrane (0.75 mm i.d., 0.90 mm o.d.) is concentrically disposed within a 2.7
mm i.d. PTFE tube in a thermostated enclosure (Fig. 5.2). The complete instru-
107
vV
Fig. 5.2. Diffusion scrubber and housing. U1, U2, U3: PVDF 1/4-/ tubing union fittings; SI:
Sample inlet; ZA: zero air inlet; AP: to air pump; T1, T2, T3: PVDF tees, E: PVDF elbow; N:
Nation tube; J: PTFE jacket, SS: stainless steel tubes; WI/WO: water inlet/outlet (PTFE
capillaries); RTD: platinum resistance thermometer connected to miniature 3-wire jack M; PF:
Plexiglas frame; C; filler cap; H: heater (top and bottom), HL: heater leads. Reproduced with
permission from Ref. [63] (copyright 2001 John Wiley and Sons).
Peltie,Cooler pL/min
Fig. 5.3. Instrument schematic. LP: peristaltic pump; DS: diffusion scrubber, L: mixing and
reaction coil thermostated at 95°C; AP: air pump; FC: flow control needle valve, V: three-way
solenoid valve controlled by timer; ZAC: cartridge to remove HCHO from pump exhaust, I:
six-way liquid injection valve; FM: flow meter; MTB: moisture trap bottle; DP: debubble port
(normally closed); N: 23 ga. hypodermic needle (supplementary flow); T: HCHO removal trap.
Reproduced with permission from Ref. [63] (copyright 2001 John Wiley and Sons).
108
CHD reagent in a Serpentine-II knit reactor (residence time -150 s, potted in a
low-melting bismuth-tin alloy containing a Pt resistance thermometer and
surrounded by two flexible heaters and thus maintained at 95°C) and then flows
through a liquid core waveguide (LCW) based fluorescence detector [64]. The
CHD reagent solution is kept cold (-7°C) by a small Peltier cooler with water as
the heat transfer agent. This reagent "refrigerator" was incorporated within the
instrument enclosure. The debubble port, normally closed, is provided via a
tee-arm to remove recalcitrant bubbles by flushing with methanol or to wash
and clean the detector. Both the water bottle that supplies the DS and the CHD
reagent bottle are protected from ambient HCHO intrusion by traps that
contain sequentially chromic acid-impregnated silica gel and soda lime. The
exhaust of the miniature air pump used for air sampling was connected to a
timer-controlled three-way valve. When the valve is energized, the pump
exhaust is vented. When it is off, the pump exhaust is passed through the HCHO
removal cartridge to feed this HCHO-free gas to the DS. The total aspiration by
AP (controlled by flow control knob FC typically to 1.1 SLPM) constituted the
airflow through the DS, plus a small amount (-2% of the total) through a 23 ga.
hypodermic needle to the AP inlet by a tee. This design ensures that when valve
provides zero gas to the DS inlet, the airflow is sufficient and no ambient air is
drawn through the sample inlet; rather, a small amount of zero air back-flushes
the sample inlet line. In this instrument, the valve is programmed to sample
ambient HCHO for 3 min and zero gas for 7 min. At the stated sampling rate of
1.1 SLPM (p = 680 mm Hg), the DS collects 70% of the HCHO (at 0.76-2.94
SLPM, -81-50% of the HCHO is collected). At the same mass flow rate but at a
higher ambient pressure, sample flow velocity will decrease and collection
efficiency will increase. The liquid core waveguide (LCW) fluorescence detector
design utilizes a GaN based light emitting diode (LED) emitting in the near UV
and provides a very stable, inexpensive, long-life excitation source that produces
no heat. Two disks of blue plastic, placed on top of the photomultiplier tube
(PMT) window, serve as the emission filter.
The entire instrument is housed in a tower-style personal computer (PC)
chassis, and uses the PC power supply. Figure 5.4 shows the front and side views
of the instrument; Fig. 5.5 shows calibration stability and response to sub-ppb
levels of HCHO measured in a 1999 field study [63]. A similar instrument based
on pentanedione chemistry (instead of CHD) has become commercially available
in a rack-mountable version [65].
109
Fig. 5.4. Front and side views of HCHO measurement instrument. Thermostated DS unit is
mounted on the side.
.2)
0)
DS liquid can collect H20 2. Typically, the collected H202 is determined fluoro-
metrically. A nonfluorescent substrate (typically a p-substituted phenol) is
oxidized by H20 2 in a reaction mediated by a peroxidase enzyme (typically
horseradish peroxidase, HRP) to form a fluorescent product [69]. Genfa and
Dasgupta [70] showed that hematin, an inexpensive product from animal blood,
110
RC
AAW ?e_: 325
2000x0.3 ?,:m >410
mm
MDS
Fig. 5.6. Single line hydrogen peroxide measurement instrument. H: Ammoniacal hematin
solution; P: pump; R: membrane reactor containing porous polypropylene membrane, M,
suspended over p-cresol, C; NMDS: Nafion membrane diffusion scrubber; RC: reaction coil, D:
fluorescence detector. Reproduced with permission from Ref. [71] (copyright 1992 Elsevier
Science).
can be used as a very effective peroxidase substitute. This led to a simple single
liquid flow channel system for the measurement of gaseous H2 02 [71]. The
instrument is schematically shown in Fig. 5.6. Ammoniacal hematin solution is
pumped into the NMDS via a porous membrane device contained in an enclosure
saturated with p-cresol vapor. The p-cresol is thus dynamically introduced into
the stream (a premixed reagent containing hematin and p-cresol shows a rapid
increase in blank). The collected H2 0 2 reacts with the mixed reagent in a
reaction coil and the fluorescence is then detected with a commercial filter
fluorometer equipped with a Cd-lamp source emitting at 325 nm. Illustrative
performance appears in Fig. 5.7. A similar instrument was flown on an airship
during the summer of 1994 and an extensive account of the results was given
[72]. A fully self contained, self-calibrating H2 02 instrument based on HRP
mediated oxidation of 4-ethylphenol has been extensively used and flown aboard
motor gliders and balloons in exemplary expositions of optimum instrument
design for atmospheric chemistry studies [69,73].
In further work conducted on hydrogen peroxide and HMHP (the adduct of
HCHO and H2 02 , which the authors contend cannot be distinguished from
H2,0), Li and Dasgupta [74] designed a compact field deployable instrument.
The oxidation of phenol derivatives as used in previous work require excitation
in deeper UV than currently available solid-state sources can provide. The new
Fig. 5.7. Diagram for gas phase H20, for the system in Fig. 5.6.
The inset shows S/N performance for 36 pptv H202 . Reproduced
with permission from Ref. [71] (copyright 1992 Elsevier
Science). -
111
Fig. 5.8. (Top) Spectroscopic details relevant to
the HO 2 , measurement system. (a) Excitation 0
spectrum of thiochrome (,, 440 nm); (b) Emis- C
sion spectrum of GaN LED; (c) Emission spect-
rum of thiochrome (ex 370 nm); (d) Absorbance
due to the reaction medium, 5 cm pathlength, E B
right ordinate; (e) Absorbance of 1 disk of m ·f;
colored plastic used as emission filter. (Bottom) L
Transversely illuminated fluorescence detector
schematic. T: Entrance PEEK tee; LI: liquid
inlet; FO: silica fiber optic; OF: colored plastic
optical filter; PMT: miniature photomultiplier
tube; AF: Teflon AF 2400 liquid core wave-
guide; SSJ: stainless steel jacket, LED: UVW Li
emitting light emitting diode; PD: photodiode u
for monitoring source light intensity; U: outlet LED r f
5SJ OF
union; W: waste. Reproduced with permission
from Ref. [74] (copyright 2000 American
PD AF T PO
Chemical Society).
112
intervals) and produces reading for each channel in turn. Software sorts and
processes the signals. The method has been extensively field tested during the
NEO3 PS study in Philadelphia, PA in the summer of 2001 [75].
The advent of LCW based detectors not only makes possible simple fluores-
cence detectors, they allow the ready fabrication of even simpler very sensitive
chemiluminescence (CL) detectors using inexpensive small area photomultiplier
tubes [64,76]. A recent paper [77] makes use of NMDS collection of H2 02 .
Luminol and Co(II) are added directly inside a LCW CL detector. The linear
range is at least up to 100 ppbv H202; an S/N = 3 LOD of 25 pptv H,,202 is
reported. The interference equivalent from ozone is below 0.1% and that from
NO2 is negligible.
the hollow fiber class (much work have been done earlier with CelgardT' mem-
branes that range from 0.1 to 0.4 mm i.d., 25 Am wall, 0.02 um pores, no
tortuosity, 40% surface porosity; currently these membranes are available only
in 200-240 Am i.d., elongated pores with 0.03-0.04 x 0.10 gm pore size, 30-50
fm wall, 25-40% surface porosity [78]) and offer significant resistance to wall
leakage.
Except as detailed below, none of the above membranes permits the use of
DS liquids containing dissolved solids. As the solution evaporates, the solids are
left in the pores and cause pore blockage. Even with pure water, nanopore
membranes like Celgard"M show evidence of pore blockage over a period of use.
However, weekly cleaning with a wetting solvent, e.g., methanol, can restore the
membrane. By the same token, porous hydrophobic membranes are generally
incompatible with organic or mixed aqueous solvents as DS liquids. The
membrane generally leaks if wetted by the solvent.
The susceptibility to blockage depends on the pore size, pore tortuosity and
wall thickness. Not all porous membranes may be subject to such blockage. An
113
Fig. 5.9. Electron micrograph of an ePTFE membrane. Reproduced with permission from Ref.
[55] (copyright 2002 American Chemical Society).
114
An 8 mm active length, 400 Am i.d. Celgard" fiber, jacketed in a -2 mm i.d.
PTFE tube was used for sampling ammonia in conjunction with a capillary scale,
electroosmotically pumped, sequential injection analysis system using the
indophenol blue chemistry and absorbance detection. Even with the modest 250
Atm optical path length, it was possible to attain an LOD of low ppb levels for a 4
min sample drawn at 120 ml/min [85].
115
volume 2 ml), amines at very low levels can also be detected [94,95]. Methyl-
amine, dimethylamine and trimethylamine and ethylamine exhibited LODs of
1.0, 0.5, 5.3, 0.7 pptv, respectively. Ambient concentrations of these amines in
Seoul, Korea ranged from less than LOD to tens of pptv: In this latter work, the
authors used a sheet membrane-based planar DS unit with the sample air
flowing on one side of a membrane and liquid on the other.
Researchers at NEC Inc. have been interested for some time in applying
PMDS devices coupled to an IC for the measurement of trace NH3 . Air samples
ranging in concentration from <1 to 10 ppbv could be measured at 10-min
intervals (minimum) with good reproducibility. A multiple-point monitor was
also developed with two sampling arms. Each arm consisted of a DS, an air
pump, and a valve for choosing between one of two sampling ports. When one DS
was measuring the air sample, the other was on standby and was conditioned
with the next air sample for 25 min. Since the air sample measurement and the
conditioning with the next air sample were carried out concurrently, it took 1 h
to measure at four different points [96]. This was later expanded to 16 measure-
ment ports, four DS units, still coupled to one IC and provided the capability of
making nine measurements per day per sampling point. The setup readily
detected an increasing number of people in a cleanroom, and was able to monitor
an area 40 m in radius [971. Ethanolamine is also a critical component for
semiconductor production processes. When an identical approach to that used
for NH3 was tried to measure ethanolamine using multipoint sampling, the
results were poor. This was found to be largely because of strong adsorption of
the analyte on the long sample conduits. Thereafter, individual DS units were
allocated to each measurement point. Satisfactory results were obtained [98].
116
FIA analysis of the trapped ammonium into an acid-base indicator stream
receptor. Despite the complexity of the flow scheme, the system was found
reliable with sub-ppbv LODs. For SO2 , the method of choice involved an alka-
line-ethylenediaminetetraacetic acid (EDTA) absorber in a planar DS and
reacting the collected analyte with pararosaniline-formaldehyde with colori-
metric detection. The LOD was -10 ppbv.
Frenzel refers to the above devices as "Permeation Denuders" even though
the paper deals almost exclusively with porous membranes. I do not believe that
this terminology is appropriate. The dominant transfer process in a porous
membrane is diffusion through the pore space, which is porous on an observable,
microscopic scale. In contrast, permeation is a process that is dominant only
with nonporous membranes (more accurately, membranes which are porous
only on a molecular scale) and which occurs by (i) dissolution in the membrane
matrix, (ii) diffusion across the polymer and (iii) evaporation on the other side
accompanied or followed by uptake by a suitable absorber. The terminology is
long established [103,104]. In a subsequent paper dealing with the use of
Griess-Saltzman reagent for determining NO2 by a planar DS utilizing a sheet
Accurel® membrane, to my considerable happiness, Scheppers et al. [105]
dropped the term "permeation". Using a short (20 mm) receiver channel filled
with the chromogenic reagent that contained optical fibers for direct measure-
ment of the reaction product (as gas was sampled in a stopped liquid flow mode),
the Gas-Diffusion Detection Device allowed a working range of -25-125 ppbv
NO 2 . These devices are the forerunners of true permeation based scrubbers
made from Teflon AF which are able to function as long path optical cells
because of their ability to act as LCWs. These are discussed in a subsequent
section.
Toda et al. [106] described a miniaturized detector for SO 2 based on con-
ductometry, using a configuration akin to a planar PMDS. Two pairs of Pt
electrodes were fabricated on a glass-epoxy substrate through conventional
photolithography and sputtering. On this substrate, a cavity for the flow of an
absorbing solution was formed with a Teflon sheet and a porous membrane. The
absorbing solution, H2 SO4 -H 2 02 , was made to flow through this cavity (2 mm
width and 0.5 mm depth) while the sample gas flowed on the other side of the
membrane. Gaseous SO 2 permeated through the membrane and dissolved into
the absorbing solution. The difference in the conductivity, as measured by the
upstream and downstream Pt electrodes, was used to quantitate the SO2 concen-
trations. Levels as low as 10 ppbv SO 2 could be measured, with a dynamic range
up to 1 part per million by volume (ppmv). The 90% response time was 82 s
(liquid flow rate 10 Al/min). Sensitivity to CO 2 was 3 x 10 -5 of that of SO 2. The
effect of RH changes was not addressed.
117
There are several selective and sensitive methods for the determination of
cyanide. However, few can be made to work in the presence of large amounts of
H2S. In coke oven gas, H2 S is typically present at concentrations an order of
magnitude greater than that of HCN. In the desulfurization of petroleum, the N
content is relatively much lower and the H2 S:HCN ratio can approach 105. It is
nevertheless necessary to keep a close tab on HCN because of its detrimental
effects on some key processes and equipment - it is also believed to be a key agent
responsible for the corrosion of stainless steel making such a measurement of
considerable practical importance. In the Kuban-Dasgupta approach, the test
gas is sampled by a mini-DS composed of a 10 cm active length of Accurel®
PVDF membrane (1.0 mm i.d., 1.5 mm o.d.), disposed within a -1.6 mm i.d.
PTFE jacket, with the gas flowing at 0.2 1/min through the interannular space.
The internal liquid is NaOH (100 gl/min), which captures both the H2 S and
HCN. The pH of the receptor is then adjusted downstream to a pH between 9
and 10. At this pH, a significant amount of the HCN (pKa 9.21) is protonated but
leaves the H2S almost quantitatively ionized (pKa 6.97). This mixture flows
across a silicone rubber membrane device where the HCN selectively permeates
in to an alkaline absorber maintained in stopped flow condition. After a 10 min
preconcentration, it is injected into a colorimetric determination system for
cyanide based on the Kinig reaction. It was possible to achieve an LOD of 35
ppbv HCN in a matrix containing several thousand ppmv H2 S. Just using a
single stage DS device, whether made from porous PVDF or silicone, one could
successfully analyze HCN in presence of H2 S up to about 2-3 orders of magni-
tude greater in concentration, but no higher. Interestingly, when the super-
incumbent pressure was increased on the DS device, the uptake increased for the
silicone rubber DS (permeation rate is a direct function of pressure) while the
uptake decreased for the porous membrane DS. Presumably, application of
external pressure causes the gas liquid interface in a pore to move further to the
bulk liquid side, thus increasing necessary diffusion distance and reducing
collection efficiency.
The same principle of collecting an acid gas using a porous membrane DS
with an alkaline absorber and then acidifying the absorber liquid to release the
collected gas back to the gas phase in more concentrated form was used by
Tarver and Dasgupta [108] to construct a H2 S measurement instrument. The
instrument was capable of measuring sub-ppbv concentrations of H2 S and
deployed off a mobile platform to monitor H2 S fluxes in West Texas oilfields
[109]. The instrument consists of a 20 cm active length Accurel® PVDF mem-
brane (1.8 mm i.d., 2.6 mm o.d.) in a 3 mm i.d. PTFE shell tube. In the annular
space, 0.1 M NaOH is pumped at 110 ,l/min while air is sampled countercurrent
at 1 SLPM. Acidic gases, including H2 S and mercaptans, are captured. The DS
effluent liquid is acidified by merging 110 l/min 0.1 M H3PO4 to re-liberate the
acid gases-the acidic liquid is pumped through a DS operating in reverse (17 cm
active length of Accurel® PVDF membrane 1.2 mm i.d., 1.8 mm o.d., filled with
1.0 mm dia. PTFE filament to reduce internal volume and put in a 2.1 mm i.d.
118
B
I
Fig. 5.10. Instrument response to ambient air in the vicinity of a natural gas processing plant.
Peaks B and C are due to H2S and CH 3 SH, respectively, the artifact peak A is much smaller
relative to these levels. The highest B and C peaks correspond respectively to 1.5 ppbv H2S and 7
ppbv CH 3SH. Reproduced with permission from Ref. [108] (copyright 1995 Pergamon Press).
jacket, the acidic liquid flows in the membrane tube, N2 gas flows at 3.6 ml/min
on the outside). The liberated gas fills one of the two 2-ml loops of a dual loop
injection valve connected to a packed column gas chromatograph equipped with
a sulfur-selective flame photometric detector. On a Chromasil 310 column, the
dominant compounds, H2 S and CH3 SH, are easily separated and an injection is
made every 2.5 min. The LOD for H2 S is less than 200 pptv. Ambient mercaptan
concentrations are generally lower than this even in the oilfields but are signifi-
cant near natural gas processing plants, as shown in Fig. 5.10.
A more recent DS-based entry to the measurement of H2S is a small portable
instrument (4.5 kg) that is operated in the field from a 12 V battery. The instru-
ment can produce one measurement every 30 s and boasts of an LOD better than
100 pptv [55]. Air is sampled by an ePTFE membrane based diffusion scrubber
and collected into an alkaline fluorescein mercuric acetate (FMA) solution flow-
ing under a controlled and constant pneumatic pressure. The collected sulfide
quenches the fluorescence, which is measured with a miniature blue LED-
photodiode based fluorescence detector. The reaction is nearly selective for H2 S.
All flow control, signal acquisition and interpretation is carried out via a mini-
notebook PC. The instrument provides for self-calibration and zero functions
using an on-board permeation tube enclosed in a thermostated block, at any pre-
programmed desired interval. The system was tested near oilfield operations in
West Texas and a wastewater plant in Louisville, KY. The data showed good cor-
relations with a concurrently operated lead acetate tape based commercial sam-
pler, with a response speed and time resolution much better than that of the
latter instrument. Figure 5.11 shows the schematic layout of the instrument.
The instrument operates in three modes, zero, calibrate and sample. The status
of valves V1 and V2 govern these modes. Only a maintenance flow through the
119
Left: Fig. 5.11. H2S measurement system schematic. P1, P2: air pumps; R: relay; IC: iodized
charcoal H2S removal columns; MFS1, MFS2: mass flow sensors; TE: thermostated enclosure;
TEC: thermal equilibration channel; PT: H2S permeation tube; F: inlet filter; V1, V2: three-way
solenoid valves; WS: water saturator constituting ofjacketed Accurel porous membrane PP fed
on the outside by water bottle WB; DS: ePTFE diffusion scrubber; T: mist trap; H: in-line heater
(prevents condensation); C: compressor; PS: pressure sensor; PC: pressure controller; RB:
reagent bottle; V3: manual on/off valve; LFR: liquid flow restrictor; D: fluorescence detector;
WC: waste container. Reproduced with permission from Ref. [55] (copyright 2002 American
Chemical Society).
Right: Fig. 5.12. Effect of sample gas flow rate on the response. The ordinate is the ratio of the
FMA decomposition in AM per ppb H2S (a) silicone DS (0.41 mm i.d. 0.61 mm o.d., 870 mm long,
2.69 mm i.d. jacket), 82-ppbv H2S, liquid flow 100 Al/min, right hand ordinate scale applies; (b)
NaOH-treated Nafion DS (0.41 mm i.d. 0.58 mm o.d., 700 mm long, 2.69 mm i.d. jacket), 11 ppbv
H2S, liquid flow 120 l/min; (c) ePTFE DS (1.02 mm i.d., 1.04 mm o.d., 600 mm long, filled with
0.53 mm dia. monofilament, 2.69 mm i.d. jacket), 5.5 ppbv H2 S, liquid flow rate 200 L/min; (d)
Teflon AF DS (1.07 mm i.d., 1.27 mm o.d., 16 mm long, 250 ppbv H2S, 10 min stopped flow.
permeation tube chamber is applied during sample and zero periods. The liquid
reagent container has a pressure sensor and this controls the attached compres-
sor to maintain a constant pressure. Several different types of membranes were
tried as DS units and the relative response of three membrane types are shown
in Fig. 5.12. Note that there is no flow rate dependence for a poor collector. For
an active membrane length of 70 cm, the equivalent quantitative sampling rates
for Teflon AF, silicone, Nafion and DS units were 0.5, 2.7, 135 and >1200
ml/min. Figure 5.13 shows how the upper applicable limit varies with the
reagent concentration and typical instrument output.
120
4
a
0
8
8
0s
2.05
F ' I r I '
0 50 100 150 0 2000 4000
Hydrogen Sulfide Concentration, ppbv Time, s
Fig. 5.13. (a) Calibration curves obtained with three different FMA concentrations, 35 cm
filament filled ePTFE DS; (b) multipoint calibration with a 5 AM FMA absorber, 31 cm
filament-filled ePTFE DS; concentrations in ppbv H2S are indicated; liquid flow rate 200 Al/min
in this and (a); (c) low level measurement with a 60 cm filament-filled ePTFE DS, 1 AM FMA;
liquid flow rate 130 l/min. Reproduced with permission from Ref. [55] (copyright 2002
American Chemical Society).
A PMDS-IC was used for the collection and determination of gaseous NH3
and HNO, at ppbv levels using purified water as the DS liquid [110]. For
validation, the gaseous HNO 8 concentration was measured by the impinger
trap-IC method and/or by a reduction (to NO)-CL method. The gaseous NH,
concentration was measured by the impinger trap (boric acid solution)-IC
method. The DS collection efficiencies were more than 97.4% for HNO3 and
>92.3% for NH3 for a relative humidity of 0-70%. Remarkably, the relative
standard deviation (RSD) values were less than 1.1% (n = 5) for successive
collection and analysis. It was confirmed that this system determines gaseous
HNO3 and NH3 correctly in the presence of NH4 NO3 particles.
Recently, Suzuki [111] gave an excellent account of determining low-MW
organic acid gases in air using IC. A particularly long DS (2.2 m porous PTFE, 1
mm i.d., 2 mm o.d., 0.1 Am pores, 60% surface porosity, put in a 3 mm i.d. PTFE
jacket tube and configured as a large diameter coil) was used to attain 92-98%
collection of formic, acetic, propionic, vinylacetic and methacrylic acids at a flow
rate of 1 LPM. The collection efficiencies for isobutyric and n-butyric acids
approached 90%. A clever valve arrangement switched the organic acids from
the much more strongly held inorganic ions which were made to migrate in the
opposite direction in the precolumn after valve switching. With a DS liquid flow
rate of 100 gl/min and a 500 Al sample loop, the LODs of the straight chain
C1-C3 acids ranged from 200-300 pptv. The analytical cycle was 20 min. Some
organic acids were unresolved under the conditions utilized. It is likely that with
121
Fig. 5.14. Automated continuous simultaneous measurement system for monitoring trace acidic
and basic gases in the atmosphere by using a diffusion scrubber coupled to an ion chroma-
tograph. AF, air purifier; F, filter; LT, liquid trap; AD, air drier; MFC, mass flow controller; AP,
air pump; V1, two-port valve; V2, V3 and V4, three-port valves; STD, standard solution; PP,
plunger pump; CC, concentrator column; SC, separator column; CD, conductivity detector.
Reproduced with permission from Ref. [80] (copyright 1999 Royal Society of Chemistry).
a little longer analytical cycle, gradient elution and proper choice of a separation
column, it will be possible to achieve a more complete separation; the method
holds much promise. Evaporative loss from the DS, effects of variations in
sample RH and jacket adsorptive losses were found to be very minor.
Komazaki et al. [80] have also developed a generally applicable ionogenic
trace gas analyzer using a PMDS-IC. This instrument simultaneously collects
the major inorganic trace gases (HCl, HNO3 , SO2 , and NH3 ) using a Poreflon
tube based DS (50 cm active length, vide supra) and an air sampling rate of 1
LPM. The instrument is schematically shown in Fig. 5.14. The liquid flow
through the DS does not occur continuously. The liquid holdup volume of the DS
is about 4 ml and it is gravity-fed from a water reservoir itself fed from an
ultrapure water (UPW) generator. At the end of a 52 min sampling cycle, first
the anion preconcentrator and then the cation preconcentrator are cleaned with
UPW. Then the plunger pump aspirates a total of 12 ml through the DS, taking
in all the collected analyte as well as filling it with fresh UPW, and then loads the
preconcentration columns. The anion concentrator column is followed by the
cation preconcentrator column (both 10x4 mm i.d.) in series; they are loaded at
2.0 ml/min over 6 min. Chromatography is carried out on 2 mm i.d. columns
using a 0.1 ml/min eluent flow rate in each channel, with a total chromato-
graphic cycle time of 1 h. Under these sampling conditions, the collection of the
target gases is essentially quantitative and allows -10 pptv LOD for each gas
even with nonsuppressed IC. No significant NO 2 interference was found in
measuring HNO3 . The instrument was extensively field-tested; the cross corre-
122
lation of the SO 2 data between the described instrument and a pulsed
fluorescence based SO2 analyzer was excellent, with a slope of 1.02 and a linear r2
value >0.98. Collected SO 2 produces both sulfite and sulfate peaks, which are
summed. However, at low SO2 concentrations, virtually all the collected S is
oxidized to sulfate. The instrument provides for on-board calibrations, over a 50
d field measurement effort, the RSD of the individual peaks from the standards
were <5%. A more recent paper reports the placement of the DS outside,
without inlet conduits, to prevent loss of gases such as HNO, or HCl [82]. The
instrument has become commercially available.
123
the HPLC eluent, it may be possible to elute the PAR quicker and reduce the
necessary cycle time. Moreover, this will also affect the analyte peak beneficially
(assuming the complex is not destroyed by an organic solvent) in improving its
efficiency. However, the method is not without problems. The scrubber liquid is
held at a low pH supposedly to decrease interference from 03 and SO2 . Regarding
ozone, ozone forms H0 by surface reactions as well as bulk reactions.
Deposited carbonaceous particles in an inlet line can be a large source of artifact
H202 upon reaction with 03 [114]. Nevertheless, barring inlet artifacts, low pH
can understandably help reduce 03 interference by limiting the 03-OH- reaction
to form HO2- and thence HO,. The authors report that they do see a signi-
ficantly lower 03 effect than that observed by de Serves and Ross [114] when
tests are conducted at realistically high levels of 03. However, they also argue
that SO2 interference should decrease with decreasing pH because SO2 solubility
in water decreases with increasing acidity. This is fallacious. The kinetics of the
H202-S(IV) reaction increases with decreasing pH. The problem is compounded
by the fact that the maximum SO2 concentration tested by the authors was only
14 ppbv. At this concentration, an equivalent of 0.4 ppbv of H202 was destroyed;
this would represent a very large negative error in measuring typical ambient
levels of H202. To be fair, the authors use DS liquid pre-doped with H,20 in their
experiments, simultaneous sampling of gaseous H202 and SO, will likely show
less interference.
Some time ago, Matuszewski and Meyerhoff described a simple detection
system suitable for the continuous measurement of atmospheric H202 at the
pptv levels [115]. The method involves the use of a PMDS as a collector, with the
effluent being sensed by a Pt electrode detector polarized at +0.4 V vs. the
saturated calomel electrode (SCE). The anodic oxidation current is directly
proportional to the H202 in the gas phase. In addition, it is well known that many
oxidase enzymes will produce an equivalent amount of HO, when they oxidize
their substrates (e.g., glucose, cholesterol, sulfite, alcohol, etc.). Thus, when
appropriate immobilized enzymes are incorporated into the flow arrangement
prior to the amperometric detector (e.g., sulfite oxidase or alcohol oxidase), the
same measurement scheme can be used to monitor volatile substrates (e.g.; SO2
and low-MW alcohols) of these enzymes, continuously at ppbv levels. The
authors also describe the stopped-flow configuration of the DS for
preconcentration.
arctic [116]. Trapp and de Serves [117] compared the collection and measure-
ment of HCHO by a PMDS followed by on-line fluorometry, and 2,4-dinitro-
phenylhydrazine (DNPH)-traps followed by off-line high performance liquid
chromatography-UV detection. The time resolution for the DS instrument was
124
5 min while the DNPH-trap samples were collected over 30-60 min. The
measured concentrations ranged from the LODs (0.045 ppbv for the DS, 0.1
ppbv for the DNPH-traps) up to 2 ppbv. The two techniques were well correlated
(r2 = 0.80, n = 48) with a slope indistinguishable from unity (1.02 + 0.03).
Komazaki et al. [118] have also used their DS-Liquid chromatograph (LC)
set up described above for H202 for the measurement of HCHO and CH 3 CHO.
2,4-dinitrophenylhydrazine (60.6 pg/ml) and 3% v/v H3PO4 in ACN constituted
the DS liquid (note the lack of obvious problems with the use of such an organic
solvent) and after collection (typ. at 0.2 LPM, which results in essentially
quantitative collection of both the target gases), an aliquot of the DS liquid is
injected into a conventional format reversed phase column with 40% aqueous
ACN as eluent and detection at 360 nm. All the operations are sequenced by a
programmable controller, and automated continuous measurements are per-
formed with a typical temporal resolution of 1 h. The LOD, based on three
standard deviations of the DS liquid blank were respectively, 0.05 ppbv HCHO
and 0.10 ppbv CH3 CHO. Analytical response from the HPLC during a ten-day
continuous measurement was unchanged (RSD < 1.0%). Interferences from 0O
and NO2 were both insignificant. For both HCHO and CH 3 CHO measurements,
concentrations from this system agreed well with those measured by a
DNPH-impregnated silica cartridge method.
The authors applied this method to diluted exhaust from a methanol-fueled
vehicle (MFV) [79]. Higher sampling frequency compared to the ambient
method above was possible due to higher analyte concentrations. HCHO was the
dominant carbonyl in MFV exhaust and aside from CH 3 CHO, some acetone, at
levels equal to or higher than that of CH 3 CHO was also noted. The emitted levels
were much higher during initial start up (cold engine) than normal running
conditions after warm up. The method showed no response to the relatively high
NOx levels that are typically present in auto exhaust. For HCHO, the results
produced by this instrument were almost identical to those produced by a
method involving impinger-based collection followed by manual HPLC analysis.
125
within the DS or in the inlet lines). Evaporative loss of liquids is much lower
than with other membranes.
Van Nugteren-Osinga et al. [53] were the first to evaluate the behavior of
such a system. They used a bundle of three asymmetric membrane fibers (200
mm length x 0.65 mm o.d. x 0.30 mm i.d.) that were exposed in a test chamber
containing CS2 vapor. The internal liquid was ethanol. After stopped flow
sampling for 15 or 20 s, the internal liquid was pumped out and reacted with a
reagent containing copper acetate and diethylamine in triethanolamine. Yellow
copper diethyldithiocarbamate was formed and was measured at 422 nm. With a
20 s stopped flow, the LOD was -500 ppmv.
As outlined earlier, Toda et al. [55,119] have examined different membranes
for the measurement of H2S, including nonporous hydrophobic membranes such
as silicone rubber and Teflon AF. Although not competitive with ePTFE for high
sensitivity and rapid response applications, in the stopped flow mode, low ppb
LODs are readily attainable with silicone rubber DS devices with the same FMA
chemistry [551]. The FMA chemistry is not completely selective for H2 S; a smaller
response is elicited, e.g., by CH 3 SH. Relative transfer rates of different analytes
through different membranes can vary considerably and by using a multichan-
nel membrane collector with different membranes, it seems possible to obtain
measurements for different reduced sulfur gases by matrix deconvolution of the
multidimensional signal [120].
Olson et al. [121] have described the use of a silicone membrane DS for the
real-world determination of HCN in process streams.
126
Reagent inlet Air in
Optical fiber Glass jacket Optical fiber
to photodiode L LED
Fig. 5.15. Basic construction of an absorptiometric liquid core waveguide based gas sensor.
Reproduced with permission from Ref. [125] (copyright 1998 American Chemical Society).
i
0.03- r;
i
Fig. 5.16. Output from a Teflon AF 2400
sensor (15 cm x 279 gm i.d., 533gim o.d.) a .2
I-/
periodically exposed to pulses of chlorine 0.01-
(1.4 ppm, 1 min duration of each expo-
sure). Reproduced with permission from
Ref. [125] (copyright 1998 American 0.0 10.0 200 30.0 40.0 500 60.0
Chemical Society). Time (min)
0.02-
127
E
0
B
Time (sec)
Fig. 5.18. System response to NO2 , 14 cm x 78 7 Am i.d., 1041 Am o.d. device. All flow rates were
200 seem except one blank run, as indicated. The linear r2 values of each temporal response are
indicated and the inset shows that these slopes exhibit excellent linearity vs. concentration.
Reproduced with permission from Ref. [125] (copyright 1998 American Chemical Society).
128
5.4 DROPS AND FILMS AS COLLECTORS AND REACTION MEDIA
FOR DETERMINING TRACE GASES
Between liquids fully contained within membrane enclosures and liquid flowing
through wetted denuders, drops and static films constitute interesting interme-
diates as gas collection devices. The use of drops for gas collection and analysis
(and other applications) has been reviewed [130,131].
5.4.1 Static drops
Liu and Dasgupta [132] suggested the first application for sampling gases with a
drop. They used an electroosmotically pumped capillary scale sequential injec-
tion analyzer to form a 6-18 1ldrop at the tip of a capillary. In one application,
dilute H202 was used as the drop liquid, SO2 was sampled and determination was
accomplished via conductivity measurement. In another application, NH3 was
sampled into a dilute H2SO4 drop and the drop contents were withdrawn and
analyzed by the Berthelot reaction involving the formation of indophenol blue.
The sampling rate was 100-400 ml/min. Only 1-1l aliquots are necessary for
analysis in this system. It was possible to take 17 successive aliquots from an
18-Al droplet and analyze them individually. This allowed mapping of the radial
analyte distribution within the drop. The results showed reasonable agreement
with the diffusive model of analyte collection. The authors also incorporated a
liquid-air meniscus detector between the drop and the analysis system. In this
way, they probed the volume of the liquid evaporated as a function of the relative
humidity (RH) and found that Frossling's equation below is applicable.
rf = ro2 -k (1-RH)(NRe) 0 5 -t (5.5)
where r and rf are the initial and final droplet radii, k is a constant, NRe is
Reynold's number and t is the time of exposure. With a constant initial droplet
size and the same sampling rate, r2 was linearly related to the sample RH. For a
diffusion-based collector, the collection efficiency f exhibits the general form
ln(1-f) = _b + lna (5.6)
Q
where a and b are constants, Q is the sampling flow rate. This was also found to
be applicable. In a very recent exposition, Felix et al. [133] carried out not only
the collection of ammonia but also the indophenol blue reaction directly in a
suspended drop; the attainable LOD would, however, be insufficient to make
ambient measurements.
Pereira and Dasgupta [134] showed how to carry out sequential reaction in a
drop. A 1/4-28 threaded chromatographic union was configured both as the
optical detector and the "drop head". The reaction of 3-methyl-2-benzothia-
zoline hydrazone (MBTH) with formaldehyde followed by subsequent oxidation
by FeCl3 to form a blue cationic dye is well known. An azine is first formed by the
reaction of MBTH with HCHO and this then oxidatively couples to a second
129
molecule of MBTH with the FeCl3 . Reagents must be added sequentially. In
practice, a 100-1l drop was formed at the drop head through an inlet mounted on
the side (liquid flows down the thread to form a hemispheroid drop). Two 1.5 mm
dia. optical fibers, respectively connected to a 660 nm LED and a photodiode, are
diametrically placed across the bottom of the fitting and can measure light
transmission through the drop. Air is sampled at 0.1 LPM for -5 min around the
drop, then flushed with clean air for 2 min. After allowing some more time to
allow the azine formation to be completed, 40 pl of the FeCl 3 -sulfamic acid
reagent (the latter removes any interference from NO2 ) is added through a
second inlet. The absorbance is noted after some further reaction time. An LOD
of -6.5 ppb was attainable in a 5 min sample. There are many means of
improving this LOD, e.g., by increasing the sample time and sampling rate,
kinetic extrapolation of data, etc. Further work has been subsequently con-
ducted for the determination of HCHO both in indoor and outdoor air [135].
Milani et al. [136] used a drop containing acidic malachite green and probed
it at 625 nm as decolorization occurred by sampled SO 2 . Parts per billion levels of
the gas was detectable. Cardoso et al. [137] have described a fluorornetric drop
sensor for H2 S utilizing previously mentioned FMA chemistry. An optical fiber
brings in blue light directly inside a drop of FMA solution that is created by a
PTFE tube as a drop head. An interference filter equipped photodiode is used to
monitor the fluorescence signal. Even with a small drop (-50 gl), diffusive
mixing within the liquid phase is slow. In a typical protocol, air sample con-
taining H2 S is sampled past the drop for 2 min at 0.16 LPM and following this,
the system is flushed with pure air for 10 s. Mixing brought in by frictional
circulation induced by the air flowing rapidly past the drop is clearly noticeable
and beneficial. This is in agreement with other observations made previously
[132]. All gas flow is then ceased. The current initially increases after the
cessation of all flow and decreases again to a minimum and then increases slowly
to a stable plateau. The analytical signal is taken to be the difference between
the photocurrent at the beginning of the experiment and that at the minimum
occurring between 2.3 and 3.5 min, leading to a linear relationship within the
H2 S concentration range of 0-100 ppb H2S.
More recently, Toda et al. [138] developed a small gas detector in which an
electrode and a gas absorber were combined in a small droplet. A gold film was
coated on the outside of a fused-silica capillary, and a silicone rubber film then
covered the Au film for insulation. When the capillary is cut, a micro-ring
electrode is present on the tip. A small droplet is formed at the tip and gas is
absorbed in it. The absorbed component was measured amperometrically by the
microring electrode in contact with the droplet solution (1 pl). The system was
characterized using SO2 as the test gas.
5.4.2 Falling drops
Liu and Dasgupta [139] carried out several novel analytical experiments
centered around a liquid drop. Animated examples are illustrated in Liu's
130
Fig. 5.19. Schematic diagram of the falling drop
t
chlorine detection unit: h is the vertical distance
between the tip of the droplet forming capillary
and the centerline of the optical fiber. Repro-
duced with permission from Ref. [139] (copyright
1995 American Chemical Society).
131
6 N C12
-
D
Q - 440 ppb C2
(I
C 880 ppb Ci2
M
in
Left: Fig. 5.20. Temporal signal profile during the lifetime of a drop.
Right: Fig. 5.21. Drop traces from three sets of four drops each, from three separate exposure
regimens. The optimum point of measurement (where the drop traces differ most) is just before
the drop falls and a new drop begins.
very interesting aspect of falling drops for gas sampling is the effect of the
sample RH. Different sample RH cause different degrees of drop
evaporation-at any given sample RH, there is a linear relationship between the
minima absorbance and the analyte concentration. However, the slopes of the
individual calibration plots are a function of RH. This may initially seem to be a
major obstacle. However, the drop period is a direct function of sample RH. The
drop dislodges when gravitational force related to the mass of the drop exceeds
the adhesional surface forces that hold the drop to the drophead. At the same
input flow rate, when the sample RH decreases, more liquid evaporates from the
drop and the drop period increases. It is easy to determine the drop period, as
evidenced from Fig. 5.20. The falling drop arrangement thus has a built in RH
sensor and this information can be used to make measurements without the
influence of RH (Fig. 5.22). Dual wavelength absorbance detection [141] should
make measurement with falling drops an even more robust technique and
increase the range of applicability.
Sampling by a drop where the drop falls through a static or moving sample
(this is the way not-so-pure drops of rain fall through the atmosphere and get
even less pure) has not as yet been used for routine analysis. Theoretical
treatment of collection of a gas by a moving drop can be formidably complex.
However, a wealth of information is already available, see, e.g., Amokrane and
Caussade [142].
5.4.3 Films
Compared with ovoid or spheroid drops, suspended films may offer greater
surface/volume ratio and hence greater sensitivity. Cardoso and Dasgupta [143]
132
a
53
2o
a0
0
.0
D
Chlorine concentration, ppmv Observed response, mAU
Fig. 5.22. Left: Variation in response as a function of RH. Right: When RH is computed from drop
period, a single equation can be used for all data regardless of RH.
133
3VO.60 01 (20.
oo i .
0.00 0.00 -r-pr-r-r-y-'--r0
0.0 4.0 8.0 12.0 ao0 4.0 8.0 f2.
Time, min Time, min
3 (b 3 01 (d
Fig. 5.23. Electropherograms of (a) outside air of TTU Chemistry building, (b) vapors from a cut
onion, (c) vapors from a concentrated HC104 (15 s sampling), and (d) vapors from a half-filled
diet cola can (half-full and mostly defizzed). Migration-based identifications: 1: acetate, 2:
carbonate, 3: formate, 4: nitrate, : sulfate (probably originally sulfite), 6: perchlorate and
chloride overlapped (can be resolved at lower concentrations), 7: benzoate, 8: phosphate.
Unlabelled peaks could not be identified. Reproduced with permission from Ref. [145] (copyright
1998 American Chemical Society).
134
Fig. 5.24. Stopped flow, especially at high gas .
sampling rates, remarkably improves sensitivity.
Circles: response under conventional conditions,
gas sampling rate 1 1min'; squares: response with
the liquid flow stopped; triangles: response under
stopped flow conditions with the gas sampling
rate 2 1min'. Reproduced with permission from D
Ref. [148] (copyright 1997 Pergamon Press). H,2 , ppbv
system. Huang and Dasgupta [148] examined electrochemical sensors for hydro-
peroxides based on thin flowing or stationary films. The sensor was composed of
two segments of a Nafion tubing put on a silver wire. A small portion of the silver
wire was exposed and was chloridized to function as the reference electrode. One
Nation segment had a Pt-wire coil wrapped on it to function as the counter elec-
trode and the other had a similar Pt/Rh wire coil, which functions as the working
electrode. A collection solution flows as a thin film on the sensor surface and also
functions as the collection medium. Hydrogen peroxide and cumene hydro-
peroxide were examined as test compounds. The former could be oxidatively
determined with a Pt/Rh electrode over a large range (ppbv-ppmv) without any
significant influence of relative humidity. In the stopped flow mode, the sensitiv-
ity increased further (Fig. 5.24). Cumene hydroperoxide, an industrially import-
ant hydroperoxide, could be determined easily with a relative precision better
than 10% in the vapor phase over simulated process reaction mixtures contain-
ing per cent levels of the analyte by reduction on a Pd electrode. The sensor is
simple and inexpensive to fabricate and requires only a suitably equipped per-
sonal computer for operation. These authors have also reviewed the general area
of electrochemical sensing of gases based on liquid collection interfaces [149].
I conclude this section with an unique continuously flowing, recirculable,
film/drop sampler that exhibits the attributes of direct gas-liquid contact, a
continuous steady state flow of liquid maintaining a stable film and the ability to
recirculate the liquid [150]. This is perhaps the most appropriate transition
device to wetted denuders. This gas-liquid equilibrator or a gas collector is
intended as a low-volume interface between a soluble gaseous sample and a
liquid phase analyzer. It can also be used between a liquid phase sample and a
detector designed for use with gas samples. This paper addresses the application
of the device for the measurement of trace atmospheric ammonia. Gas collection
occurs solely by diffusive sampling such that aerosol particles are not collected.
The device, shown in Fig. 5.25, consists of a tube surrounded externally by a
jacket. Gas flows through the jacket and contacts a liquid film flowing on the
135
- LO
N1 -
- HS
T -
_ _ N2
I ''
I Vt
Ss
Gas out -
11 (
ili
"
c___-
Left: Fig. 5.25. Continuous film gas-liquid collector/ equilibrator. N1, N2: 1/16-in nuts; N3:
1/8-in nut; T: PEEK tee; LI/ LO: Liquid inlet/outlet lines; S: PEEK sleeve; HS: heat shrink tube;
J: Jacket; C: bottom cap. Inset shows details of the stainless steel SS tube tip. Reproduced with
permission from Ref. [150] (copyright 2000 American Chemical Society).
Right: Fig. 5.26. Calibration behavior of the continuously flowing film collector (fluorometric
analysis in the 0-50 pptv region. Note that the first peak in a series always tends to read higher
due to flow disruption when concentration is changed. Reproduced with permission from Ref.
[150] (copyright 2000 American Chemical Society).
outer surface of the tube. The flowing film forms a drop at the tube terminus and
is aspirated off via the inner bore of the tube. The collected analyte can be (a)
directly sent to an analysis system, or (b) preconcentrated on a suitable station-
ary phase, or (c) the device effluent recycled. With a fluorometric flow injection
analysis system harnessed to measure NH 3 with such a collector, the limit of
detection (LOD, S/N=3) for a sample drawn for 18 min at 200 ml/min was 4.5
pptv, with the linear range extending up to 30 ppbv. Establishing true zero or
removing ammonia from air to prepare a zero level sample for the measurement
of pptv levels of ammonia was shown to be a formidable task. Figure 5.26 shows
system behavior at low pptv levels of ammonia.
136
5.5.1 Single tube wet denuders
The original development of simple single tube wetted denuders [151,152] was
discussed in the previous review [1]. Using such a wet denuder and dedicated
colorimetric detection, Vecera and Dasgupta [153] explored the production of
indoor HONO from open flame sources. Sources such as gas heaters can produce
alarming levels of HONO.
Taira and Kanda [154] described a similar glass WEDD, that is, however,
much smaller in i.d. (1.6 mm vs. 8 mm). With this much smaller radius of curva-
ture, the authors found that surface treatment with conc. H2SO4 was sufficient
to render the surface wettable. While they demonstrated its successful applica-
tion for the collection and analysis of HNO,, it is noteworthy that at the recom-
mended sampling rate of 4 LPM, the flow is not laminar. The deposition of very
small particles (30-110 nm) was measured in the denuder at a flow rate of 2 LPM
(NRe 1777). The loss was significant (3-11% with NaOH flowing down the walls)
but the authors did not measure the aerosol in the denuder effluent. Rather,
they measured the particle counts before and after the denuder and this can be
misleading because deposition in the nonwetted areas affects the measurement.
Neftel et al. [155,156] used a WEDD as a collector for gaseous N species. The
artifact production of HONO from NO2 and peroxyacetylnitrate was determined
to be 0.1 and 0.4%, respectively. (A cautionary note is necessary regarding the
measurement of outdoor HONO levels with a wet denuder. Although NO2 and/or
NO do not seem to lead to significant levels of artifact HONO in a wet denuder, a
thorough study that includes a more complex mixture (e.g., SO2 ), has never been
rigorously conducted. Ambient HONO levels measured by direct spectroscopic
techniques tend to be lower than those measured by wet denuders.) Neftel et al.
[157] also constructed a miniaturized WEDD (12.5 cm long, 0.5 cm i.d.) for
studying ammonia equilibration in soil [157].
Vecera et al. [158] reviewed the general possibility of using diffusion de-
nuders to sample organic compounds and pointed out the potential of WEDD
devices in particular in such applications. They also followed up on this. Zdrahal
and Vecera [159] used an WEDD for the preconcentration of gaseous 2,4-trichlo-
rophenol (TCP) from air. A mildly alkaline scrubber liquid (water adjusted to pH
8.5) was used as the WEDD liquid. The concentration of TCP was determined
using reversed-phase liquid chromatography with a C-18 stationary phase and
UV detection at 218 nm. The system allowed results to be obtained at 7 min.
intervals with an LOD of 685 ng/m 3. Collection of organic vapors is better
attained with organic solvents as WEDD liquids. Zdrahal et al. [160] studied the
collection of 1,4-dichlorobenzene (DCB) as a model compound using alcohols,
glycols, and their aqueous mixtures as WEDD liquids. 1-Propanol was found to
be the most suitable. With sufficient flow of the WEDD liquid (> 150 tl/min), the
experimental collection efficiencies agreed with theoretical values for air flow
rates from 0.5 to 4.3 LPM at DCB concentrations of 48 and 720 ,Ag/m3. The
denuder operated without difficulties at temperatures from 18-33°C and over an
137
RH range of 10-100%.
Peskova et al. [161] recently studied WEDD preconcentration for the deter-
mination of methanol, ethanol, -propanol, 2-propanol, 1-butanol, 1-pentanol,
acetone, methyl ethyl ketone, diethyl ketone and methyl n-propyl ketone in air,
with gas chromatography- flame ionization detection (GC-FID)or GC-MS
analysis. The compounds are continuously collected into water. The LOD of
alcohols and ketones in the effluent are as low as 1 gg/l by GC-FID and 1 ng/l by
GC-MS. The authors state that the technique should be applicable for the
continuous monitoring of ppbv levels of both alcohols and ketones in the air.
Jaeschke et al. [162] described their static wet denuder system for the collec-
tion of ionogenic gases. Two parallel denuder systems are used, one with an
acidic absorber (pH 2.7 with HC1) for the collection of ammonia, and the other
with an alkaline absorber (pH 10 with NaOH) for the collection of acid gases.
The authors characterize their denuder system as "vertical", a distinction lost
on this author because all the previous denuders described in this section also
operate vertically. Perhaps the intent is to distinguish it from the Netherlands
Energy Research Foundation (ECN) annular WEDDs videe infra) which operate
horizontally. Regardless of the terminology, the denuder system does have char-
acteristics that are unique. The wetted surface is made from fine nylon stock-
ings. (In screening various material as denuder liners [151], this material was
also studied in our laboratory. Similar to the use of polycarbonate liners, we
observed a small but discernible loss of HNQO. However, a thorough study of all
types of available nylon stockings was not conducted!). The denuder is fed from a
constant pressure head using a pressure sensor that controls the feed pump. The
denuder effluent (9 ml recirculated volume) was withdrawn periodically and
analyzed batch mode by IC. It is interesting to note that given the dimensions of
the denuder (25 mm i.d., 350 mm long) the stated collection efficiencies for NH,
(96, 87 and 79% at flow rates of 2, 4, and 6 LPM) are significantly super-
theoretical (at 6 LPM, the computed collection efficiency is 68%; see Eq. (5.1)
and Section 5.2.1). The merit of batch mode operation is not clear. Recently
Costa et al. [1631 have described the use of such denuders with NaF and Na2 CO3
solutions as DS liquids in individual DS units to collect various gases and deter-
mine them by IC analysis.
(Other "fabric denuders", consisting of a piece of porous fabric, with or
without impregnation, stretched across an opening, e.g., a filter holder, have
been reported [164]. These are very different from any other devices described
herein. Similar denuders based on porous metal disks were reported earlier.
These authors modeled the collection with membrane filtration theories [165]. It
has been reported that particle losses can be low. Perhaps this type of denuder
will find increased use, especially in personal monitors [166].)
Buhr et al. [167] have described a single tube wet denuder-fritted filter
combination coupled to an IC for the measurement of gaseous nitric acid and
particulate nitrate and sulfate. Reasonable agreement was obtained for nitric
acid values with collocated filter pack measurements of nitric acid.
138
Extraction
solution
One very interesting single tube denuder designed by Komazaki et al. [168]
is of annular construction but with only the inside of the outer tube as the active
surface. NOx (both NO and NO2 ) are photocatalytically collected on a
TiO 2 /hydroxyapatite surface. Figure 5.27 shows the device. A tubular fluores-
cent lamp (Hg "black light", 365 nm) surrounded by a quartz tube constitutes
the inner element and the outer tube (glass) has a mixed coating of TiO 2 and
hydroxyapatite. The denuder operates in a batch mode; water washes the walls
after a preset period of illuminated sampling. The effluent is sent to an IC for
analysis and total NOx is measured from the sum of nitrite and nitrate. The cata-
lytic activity is completely recovered on washing and drying. Presumably it does
not work if the liquid is continuously flowing but this is not explicitly addressed.
Figure 5.28 shows the need for both constituents of the mixed coating and UV
illumination to successfully collect NO.
Chang et al. has advanced an equally interesting concept in wet denuders
[169]. The effluent liquid in the cylindrical denuder is not directly added but gen-
1
II ,
-UV illumination-----*
___
Fig. 5.28. Collection of NO by the photocatalytic WEDD
(flow rate 0.2 1/min). Note that NO passes through com- .o
pletely with a coating of only hydroxyapatite (HAP) in the
/ --- no
dark; some is collected upon turning on the light but bulk / .... TO
still passes through. With only a TiO2 coating, nothing is
collected in the dark; with UV illumination some analyte Dor
is initially collected but things rapidly return to the no- S2 I- -···H~
TTi2+HAp
collection condition. Only with both TiO2 and HAP
present and in the presence of UV illumination is NO a-
efficiently collected. Reproduced with permission from 10 20 30
Ref. [168] (copyright 1999 Elsevier Science). Time (min)
139
erated by cooling so that condensation occurs. The temperature and RH of the
sample is controlled by passing the sample through a thermostated Nafion
humidifier (to give 85% RH at 40°C) and the analyte gases are then collected by a
cooled (2°C) cylindrical denuder where condensation occurs on the walls. An IC
analyzes the liquid effluent. The LODs for CH3 COOH, HONO and SO2 are 22, 19
and 9 pptv, respectively. The precision ranges from 0.3 to 3.0% RSD. The method
was successfully applied to urban air analysis in Seoul, Korea. The temporal
results closely followed those produced by a PMDS-IC system operated in paral-
lel. The use of a cation exchange membrane for humidification precludes the
measurement of ammonia or other soluble gases that are not acidic. Although
the method was said to be applicable to weak acid gases such as CHCOOH, no
clear recovery data were presented to indicate whether or not such a weak acid
gas is quantitatively transmitted through the humidifier. The method has been
patented [170] and is discussed in detail in a PhD thesis along with several other
(parallel plate, annular) condensation denuder and DS designs [171]. Particle
deposition in such denuders need to be examined in greater detail since thermo-
phoresis can attract particles to cold surfaces [172].
The wet annular denuder was actually the first wet denuder to be introduced
and predates the less sophisticated single tube wet denuders. The wet annular
denuder was invented by the ECN group who have made many other contribu-
tions to atmospheric analysis [173]. In the originally introduced device, opera-
tion was in the batch mode. It was a wetted annular denuder (do = 45 mm, di =
42 mm, L = 30 cm) held together by PTFE spacers. Twenty-five milliliters of an
aqueous absorber was charged into it and the outer tube was rotated by a
motor-belt-pulley arrangement at 40 revolutions per minute (rpm)-this main-
tained a uniform film on the walls of the device. Near-quantitative collection of
several gases of interest was achieved at flow rates as high as 32 LPM.
Subsequently, Wyers et al. [174] from the same ECN group reported a
rotating wet annular denuder (RWAD) from which solution is continuously
pumped in and out such that analyzers can be continuously coupled. Further
refinements have been reported since [175-177]. The authors strived from early
on to reach excellent analytical precision, to allow these methods to measure
fluxes of different pollutants aided by simultaneous micrometeorological mea-
surements. Several field studies have been conducted with these devices, partic-
ularly with a view to measure the flux of nitrogen compounds [178]. In most
cases, few other alternatives have been available. In one study for example, the
only competition for a continuously operating RWAD was another RWAD oper-
ating in batch mode [179]! The results of these measurements have made a
significant contribution to better descriptions regarding surface exchange of
important pollutants. In a recent review on the advances in micrometeorological
methods to measure gas and particulate N fluxes, Fowler et al. [180] conclude
140
Fig. 5.29. Schematic diagram of the continuous-flow denuder.
that relative to eddy accumulation techniques, wet denuders can provide the
same value (within 10%) at less than 10% of the cost. In another recent review on
methods for measuring emission rates of ammonia from livestock buildings and
from slurry or manure storage, Phillips et al. [181] also professed their love for
AMANDA (Ammonia Measurement by ANnular Denuder sampling with on-line
Analysis).
The ECN RAWD units operate at a high sampling rate of 25-30 LPM. All
components are packaged into a transportable box, to facilitate field deploy-
ment. Current versions use liquid mass flow meters to monitor peristaltic pump
flows exactly, an IR sensor is used to maintain constant liquid level in the
denuder. Bromide, typically not present in atmospheric samples in discernible
amounts, is added as an internal standard. Anions are measured by carbonate
eluent anion chromatography, ammonia is measured in a portion of the stream
by adding strong NaOH, diffusing the liberated NH3 across a membrane to a
water receptor and measuring conductivity following a procedure originally
outlined by Carlson [182].
Figure 5.29 shows the ECN RAWD. The LOD for the acid gases, as measured
by IC with 20 min time resolution, is easily below 100 ng/m3 for HC1 and HONO
and below 200 ng/m 3 for HNO3 and SO2 [177]. In early work [175], the relative
standard deviation was -4% at gas concentrations of 1-2 g/m 3 , this has been
further improved significantly with the incorporation of internal standards. The
ammonia measurement system has been widely used and cross comparison
against a laser based detection scheme (which has inferior LODs) has provided
favorable results [183].
5.5.3. TTU Gravity flow wet annular and parallel plate denuders
Many contributions from the ECN group and the present author's laboratory at
Texas Tech University (TTU) have proceeded in parallel. Thus, while the ECN
group contributed the first wet denuder, TTU contributed the first continuously
operating wet denuder [151]. The first continuously operating annular denuders
141
OT- -
LII IT
SN
Left: Fig. 5.30. TTU Wet Annular Denuder. OT: outer tube, LII: liquid inlet to inner tube, LIO:
liquid inlet to outer tube, IT: inner tube, OR: O-ring, C: cavity, LC: Plexiglas liquid collector cup,
LOI: liquid outlet from inner tube, OF: outer fitting, LOO: liquid outlet from outer tube, SN:
supporting syringe needle, PF: Kynar® porous frit, PG: top Plexiglas insert, SA: silicone
adhesive plug. Reproduced with permission from Ref. [25] (copyright 1993 American Chemical
Society).
Right: Fig. 5.31. Wet parallel plate denuder: (a) overall appearance, (b) view of one plate, (c): top
cross section. GP: Glass plate, LI: hypodermic needle liquid inlets, PG: Plexiglas spacer, TF:
Teflon film, SC: silica coating, LO: liquid outlet, Al: air inlet (the outlet, not shown, is similar.
Reproduced with permission from Ref. [25] (copyright 1993 American Chemical Society).
were described by the ECN and TTU groups in the same year [174,25]. The
design of the TTU wet annular denuder (WAD) is shown in Fig. 5.30. The inner
tube is supported from the outer tube by three symmetrically placed hypodermic
needles, one of which serves as the liquid inlet for the inner member. It fills a cup
shaped cavity and overflows along the outer walls of the inner tube, coated with
easily wettable porous glass [151]. The water flows down into a cup-shaped
collector from which it is withdrawn. Flow in and out along the inner wall of the
outer tube (bearing a similar wettable surface) is very much along the lines of
previous single tube WEDD units. The TTU WAD (7 mm od 300 mm long tube
inserted in 10 mm id, 400 mm length tube) is much smaller than the ECN WAD.
It does not require a mechanical rotating arrangement but quantitative
sampling capability is also far less. The device quantitatively captures SO2 at a
sampling rate of 3 SLPM and efficiency drops to -90% by 5 SLPM (WAD liquid
0.5 mM H, 2,, p = 680 mm Hg).
In the same paper [25], the first parallel plate wet denuder (PPWD) was also
described (Fig. 5.31). Typically each glass plate measured 600 x50 mm (active
coated area width 50 mm) The silica-coated area is represented by the shaded
region. The two plates are separated by a 3 mm thick Plexiglas spacer covered
with PTFE tape. In each plate, ca. 12 cm are left uncoated at the bottom to allow
142
the development of laminar flow before the active collection region begins. At the
top of the coated region, three evenly spaced holes (0.75 mm) are drilled. At the
bottom, one hole is provided just above the point of the vee. Liquid is pumped in
parallel through all three needles, typically for a total flow rate of 200-300
/l/min. It is aspirated at the same rate at the bottom port and the opposing plate
has an identical arrangement. Air inlet or outlet connections to the denuder
were made by thermally deforming and stretching thin-walled PTFE tubes of
the appropriate diameter (ca. 10 mm for the PPWD described) to a rectangular
cross section at one end. The PPWD was substantially more efficient than the
TTU WAD, removing SO 2 quantitatively at 8 SLPM and removing -95% at 10
SLPM. Several other characteristics are noteworthy. At the stated liquid flow
rate, the mean film thickness of the falling water film was estimated to be 25-50
Am. Less water evaporates from the PPWD (which has a greater hydraulic cross
section. At 23°C, no dry spots develop in the PPWD even with a liquid flow rate
as low as 100 bl/min and an airflow rate as high as 16 SLPM (dry air). This is,
however, very dependent on the water affinity of the wetted surface, with
polymer surfaces (vide infra) much more water is seen to evaporate. Significant
amounts of the liquid input are evaporated during transit through the denuder
but the exit air does not reach saturation RH (referring to saturation at ambient
temperature). Due to the significant latent heat of evaporation of water and
near-adiabatic cooling, the temperature of the wetted surface is substantially
below ambient. For example, with the PPWD operating at an air sampling rate
of 15 SLPM and an inlet humidity and temperature of <-30% and 23°C
respectively, within a few cm of the inlet, the temperature of the PPWD, as
measured from the external glass surface, drops to 11.8°C. Nevertheless,
observed collection efficiency seems to be well predicted by extant theoretical
equations for such denuders, using the diffusion coefficient of the gas at ambient
temperature [26].
Figure 5.32 shows the coupling of the PPWD to an IC (the lowest rung IC
then available, a "QIC" from Dionex Corp. was used). The aspirated effluent
from both plates were combined and preconcentrated on a preconcentrator
column. The dual loop injection valve is configured with two such precon-
centrator columns; while one is being loaded with the sample, the other is being
eluted. Typically, the valve was switched every 8 min, a new chromatogram is
thus generated with this frequency.
The lower sampling rate of the TTU PPWD relative to the ECN RAWD is
compensated for by the superior IC LODs afforded by a hydroxide (rather than a
carbonate based) eluent. To date, this difference in eluent use in IC between the
two groups persists. Figure 5.33 shows the triplicate chromatograms resulting
from sampling clean air and 19 pptv SO 2. Reproducibility of the sulfate peak at
this level of SO2 was 0.8% in RSD. The precision level of the blank (clean air)
signal was equal or better. These data lead to a computed limit of detection of 0.5
pptv (1.3 ng/m3), well below any other reported technique. In a similar study to
measure HONO and HNO, by PPD-IC, the method LODs were found to be 0.11
143
Fig. 5.32. Liquid phase analytical system schem-
atic for a PPD coupled to an IC. CC: preconcen-
tration columns, GC: guard column, AC: analytical
column, S: membrane suppressor, D: conductivity
detector. Reproduced with permission from Ref.
[25] (copyright 1993 American Chemical Society).
Fig. 5.33. Triplicate chromatograms resulting from sampling: (left) clean air, (right) 0.019 ppbv
(50 ng/m3 ) SO2 . Peak A is sulfate, B is from CO 2 and C is due to an unknown analyte that is
leached from poly(vinyl chloride) pump tubing. Reproduced with permission from Ref. [25]
(copyright 1993 American Chemical Society).
and 0.23 pptv for these gases [184]. However, gaseous NO2 produces some NO2-
and NO- by hydrolytic reactions (-10 times as much NO2 as NO,-). Although a
very small extent (0.06%, see also Ref. [154]) of the gaseous NO2 is actually con-
verted, this is sufficient to significantly alter practical LODs in ambient air.
Moreover, there is some evidence that N2 0, and other, as yet unidentified nitro-
genous species in ambient air, may also generate such artifacts upon reaction. In
addition, NO2 uptake may be higher in the presence of significant amounts of
SO2 , this has never been directly examined.
Both the WAD and the PPWD exhibited very low particle deposition, as
shown in Fig. 5.34. In this context, it is interesting to note that in all wetted
denuders, particle loss increases when the same denuder is operated dry. When
wet, evaporation of the liquid from the denuder surface creates a flux that
strongly inhibits particle deposition on that surface due to diffusiophoresis
144
200 ....
0
'o
° 160
I A
C
'C
Mr
0
U)
a
t
a
8-
[172]. Water evaporating from the wet surfaces set up a flux of molecules that
bombard incoming particles numerous times and make it more difficult for
particles to be deposited. Others have also noted decreased particle deposition in
wetted denuders when the latter are in normal operation compared to when they
are operated dry [154].
Between the WAD and the PPWD, the PPWD is more efficient and is
significantly easier to construct. For this reason, in our laboratory, we have
chosen to work with the PPWD. The present design uses Plexiglas plates with
PTFE spacers, rather than glass plates. This eliminates fragility problems. The
desired area of each Plexiglas plate is microstructured with a grinding tool to
render it more easily wettable (nevertheless, it is not as wettable as their porous
glass counterparts and must initially be filled with water to ensure even
wetting). The denuder stand is conveniently made from an wooden support base
to which threaded pipe flanges are secured by screws and the threaded end of a
3/8 in i.d. galvanized steel piping is used as the support stands. The denuder is
simply bolted to these. If the denuder is used for gas analysis and deployed
without a cyclone with a very short inlet (<5 cm, Pefluoroalkoxy fluorocarbon
(PFA Teflon)), the sample air must enter the denuder at the bottom. As such, the
denuder needs to be on a tall stand that allows the inlet to be -60 cm off the
support base. To minimize interaction of the inlet air sample with the stand
components, especially in still and stagnant air, we also recommend that the
lengths of the iron support stand from the base to the bottom of the denuder be
wrapped with PTFE tape. Each denuder plate is 10.0x55 cm (3/8 in thick) with
the active wettable area of 6.5x42 cm, starting 7.5 cm from the top and 1.75 cm
145
L
Ai
Fig. 5.35. Left: TTU PPWD shown schematically; AI/AO: air inlet/outlet, LI/LO: liquid
inlet/outlet, LR: Liquid reservoir WA: wetted area, S Spacer between plates, SH Screw holes to
hold plates and spacer together, also affixes denuder to stand. Middle: Complete denuder on
stand. Right: Top and bottom detail of PPWD.
from each edge. The denuder liquid is forced through a fritted PVDF barrier to
allow even flow down the plate and is aspirated from the V-groove at the bottom,
4.5 cm from the bottom. The two plates are spaced by a 3 mm PTFE spacer. The
air inlet/outlet holes, circular at the termini, are machined with a contour that
becomes elliptical as they approach the interior of the denuder to allow a smooth
entrance/exit of the airflow. Pefluoroalkoxy fluorocarbon (PFA Teflon) tubing (1
ga., 8.3 mm o.d., 7.5 mm i.d.) fit tightly into these apertures. The denuder is
shown in detail in Fig. 5.35. Note that serious analyte losses and/or artifact
reactions can occur in a cyclone inlet placed ahead of a denuder [185]; this should
be avoided.
In this configuration, these PPWD units have been used in several USEPA
sponsored field studies, both for trace gas analysis and for removing gases prior
to particle collection, in Atlanta (1999) [186,187], Houston (2000) [188], and
Philadelphia (2001) [75] as well as to remove acid/basic gases before a system
that determines the acidity of the aerosol fraction [189].
A very interesting paper by Rosman et al. [190] appeared during the prepara-
tion of this chapter, in which the authors describe a PPWD constructed of PFA
teflon with an inert screen material as the wetted element. Detailed character-
ization, performance and field evaluation data were presented.
Wet denuders based on the circular-end, parallel plate geometry, which facilitate
air inlet/outlet connections, were first used by Staffelbach et al. [27]; few
operational details were given. The same geometry has been subsequently used
146
by Hesterberg et al. [191], Ammann et al. [192] and described in more detail by
Zellweger et al. [193]. Overall, the deployment is similar to that used by the TTU
group, except that -50 M NaHCO 3 was used as the denuder liquid (vide infra)
and a carbonate- bicarbonate eluent was used for IC.
Loflund et al. [194] described the most sophisticated CEPPWD-based analyt-
ical system to date, with very impressive capabilities. The collection device is a
quartz CEPPWD with a silica/silicate coated active area. The air sampling rate is
5 LPM and the liquid flow rate is 167 pl/min. A gradient IC system is used with
an electrodialytic eluent generator for analysis. The system has the capability to
measure the major organic (formic, acetic, propionic, oxalic, malonic and suc-
cinic) as well as inorganic (SO2 , HONO, HNO3 , HCl and HF) acidic gases as well
as ammonia by cation chromatography on a second system. A pair of cation and
anion preconcentrator columns are respectively incorporated in two 10-port
valves. The cation and anion systems are alternately loaded and analyzed (30
min each). The LODs (pptv) stated for the acid gases were as follows: formic (5),
acetic (4), propionic (30) butyric (19), valeric (21), pyruvic (45), oxalic (22),
malonic (6), succinic (42), o-phthalic (10), citric (2), maleic (8), HF (56), HC1 (74),
HONO (10), HNO3 (11), SO 2 (9), H3 PO4 (22). At 400 pptv, the LOD for NH 3 was
significantly higher. The device was successfully used for field measurements at
an alpine station.
147
TABLE 5.1
The tale of the four denuders. Isoefficient (95% efficient in collecting SO2 ) at stated flow rate.
Single tube Annular thin and short Annular fat and long Parallel plate
d = 3 mm di = 1.0 cm di = 3.0 cm w = 5 mm
do = 1.3 cm d = 3.3 cm s= 1.5 mm
L = 50 cm L= 10 cm L = 20 cm L =30 cm
Q = 1 LPM Q= 3.8 LPM Q = 20LPM Q = 20.3 LPM
Relative mass collected
1.00 2.23 11.8 11.9
Active surface area, cm 2 (liq. flow in ul/min to maintain same velocity)
47.1 (283) 72.3 (433.8) 396 (2376) 300 (1800)
Relative analyte concentration in effluent
1.00 1.46 1.41 1.87
The present TTU configuration used for this analysis is shown in Fig. 5.36.
Typically the air sampling rate is 5 SLPM and 0.5 mM H2 02 is used as the
denuder liquid at 0.5 ml/min on each plate, each stream pumped through a
disposable mixed bed ion exchange resin column MB (0.67 cm i.d. x 15 cm,
PTFE column filled with Dowex MR-3 resin) located immediately before the
PPWD liquid entrance ports. The effluent streams are aspirated at 1.1 ml/min
from each plate (using same peristaltic pump but larger tubing, 1.52 mm vs. 0.89
mm i.d. Pharmed® tubes are used for input vs. aspiration) to ensure all liquid is
aspirated from the bottom of the PPWD. The aspirated flow streams are
combined and sent to the analysis system. An 8-channel peristaltic pump is used
for all low pressure pumping, 7 of 8 channels are used.
A significant difference between the TTU system and other similar systems
such as that of Loflund et al. [194] and the ECN RWAD [174-177] is in feeding
the same sample serially to cation exchange and anion exchange preconcentra-
tors (in that order, which is important because anion preconcentrators used by
us contain accessible cation exchange sites). This increases sensitivity and/or
sample throughput, compared to parallel or temporally sequential arrange-
ments. Anions pass unretained through a cation preconcentrator CC (Na+-form
strongly acidic cation exchanger Dowex 50Wx8, 100 mesh, 4x25 mm) on to an
anion preconcentrator TACLP1, each paired with an identical twin preconcent-
rator in 10-port dual loop injector valves V1 and V2. At the end of a typical 15
min sampling cycle, both of these valves switch, and the other set of precon-
centrators are brought on-line. In the anion analysis system, electrodialytically
generated KOH (15.5 mM) elutes the collected anions at 1.5 ml/min at 29°C
through high capacity AG11HC (4x50 mm) and AS11HC (4x250 mm) guard
and separator columns, a water-regenerated electrodialytic suppressor and a
148
p MFC )
PPWD
D
LC3S
Fig. 5.36. P: Air aspiration pump, MFC: mass flow controller (5 I/min), PPWD: parallel plate wet
denuder, EG40: electrodialytic eluent generator, IS25: eluent pump (1 ml/min), TAC-LP1: 4 mm
anion preconcentrators, AG11HC: guard column, AS11HC: separator Column, SRS: electro-
dialytic suppressor, CD25: conductivity detector, LC30: Column oven, MB: mixed bed resin
columns, PP: peristaltic pump, PMD: porous membrane device, CC: cation preconcentrators, V1,
V2: 10-port valves, V3: 4-port valve, V4: 3-way valve, R: 0.25x70 mm restrictor tubing to
eliminate formation of air bubbles, W: waste.
149
1.20-
Houston HRM-3
('K 1:38 to 2:08 09/12/00
E
E
'8 0.80
.2
E
a; S
\ Ammonia 1.70 ppb
C
.f 0.40-
E
.2
.5e i
t
0.0.
16.00 18.00 20.00 22.00 24.00
Time (min) Time (min)
Fig. 5.37. Top: Gas phase acids. Bottom: Gas phase ammonia from TTU system.
on a Na+-form column. The output from the detectors during a field campaign in
Houston, TX is shown in Fig. 5.37.
We have also experimented with a simple setup where the same conductivity
detector and the same data processing channel can be used to measure both the
acid gases and ammonia. The arrangement is particularly attractive if one is
primarily interested in the principal acid gases, SO2 and HNO,, corresponding to
the sulfate and nitrate peaks (this is specially true for aerosol phase analysis
where sulfate and nitrate are frequently the focal points), such that only a
rudimentary separation between chloride (plus organic acids), nitrate and
sulfate is needed. This can frequently be accomplished with a short separation
column. The arrangement is shown in Fig. 5.38; the principal difference from
Fig. 5.36 is that the suppressor effluent is routed through a four-port valve V5.
In one position, the suppressed effluent proceeds directly to the detector. Only
when the anion chromatogram has finished (sulfate has eluted), the valve is
switched, directing the suppressor effluent (essentially water) through a mixed
bed resin to the receiver side of the PMD. At this point, V1 is switched to elute
the collected ammonia, the PMD receiver effluent proceeds to the detector,
measuring ammonia. The right panel in Fig. 5.38 shows the output of such a
system, showing peaks due to 10 nmol amounts of chloride, nitrate and sulfate
and 40 nmol NH3 . (Note that these experiments were carried out with bottled,
rather than an electrogenerated hydroxide eluent. The chromatographic back-
ground is high due to the presence of CO,, and the baseline goes down when V5
effluent is switched for ammonia measurement because the background
becomes essentially freshly deionized water.) It should be obvious that ammonia
measurement can be carried out before or after anion elution and the PMD
receptor can be maintained in a stopped flow mode to increase sensitivity. In the
example shown, we have eluted the ammonia where the void volume peak will
normally elute, making efficient use of time.
150
2-
npnl ,2s,
H4
Time (min)
Fig. 5.38. Left: One detector anions-ammonium measurement system. Legend as Fig. 5.35. V5 is
an additional 4-port valve. Right: First set of peaks in "normal" anion IC, second set of peaks
shows ammonium detection system switched in where void peak normally elutes.
151
effluent flow rate of i ml/min will lead to 20 ,M H2SO4 in the liquid effluent. This
represents a pH of -4.4. In such a solution, many of the weaker acids will not be
fully ionized and therefore may have solubility limitations. Zellweger et al. were
particularly concerned with nitrous acid (pKa 3.1-3.2), although the problem
may obviously be more of an issue with analytes such as acetic acid (pK 4.75).
Zellweger et al. proposed a dilute solution of the chromatographic eluent,
alkaline in nature (-50 JM NaHCO3), as the denuder scrubbing liquid.
For several reasons, this solution may not be ideal. It is not clear in the
presence of large amounts of SO,, how far along the denuder the influent
NaHCO3 solution will maintain an acceptable pH. Second, weak acids dissolve in
aqueous solution both by their ionization and through their Henry's law parti-
tion (intrinsic solubility). The latter can be high, irrespective of the pH of the
collector-HCN, a very weak acid, has a very high intrinsic solubility, for
example [107]. Also, 100 ppbv or higher levels of SO, are found sporadicallyas a
plume impacts a sampling location but such levels on a sustainedbasis are not
common, at least in the USA. The suggested approach may provide a solution for
an exceptional case but generates problems for other, more common situations.
In a study conducted in Vienna in the winter of 2000 to determine penetration
into a second denuder, Loflund et al. [194] found 97% of the HONO was collected
in the first denuder.
These problems include the incompatibility of large amounts of carbonate in
the sample with hydroxide eluent based anion chromatography, presently the
preferred practice. Use of a carbonate containing denuder liquid generates a
substantial amount of carbonate in the effluent, a broad tailing carbonate peak
can obscure smaller analyte peaks in that region.
An alternative solution to the problem is to buffer the denuder liquid (prefer-
ably at a pH of 5.5-7). This is easily practised if the denuder liquid is not ana-
lyzed for collected gases but the interest is in removing the gases to perform
aerosol analysis. A zwitterionic buffer was used to remove high levels of acidic or
basic gases (as may be present in indoor environments when a kerosene-fueled
heater is operated) or high levels of ammonia (encountered in homes with live-in
pets) before aerosol analysis [189]. However, these approaches have not been
used when the denuder effluent, i.e., the collected gases, are of interest and
moreover, when it is desired to preconcentrate the sample. Zwitterionic buffer-
ing material may still be useful. Glycine, for example, has the correct pKa to be
useful and is known to be suppressible. Morpholinoethanesulfonic acid and
Bis-tris are other potentially useful zwitterionic buffers. It is known that at least
under some conditions it is possible to suppress zwitterionic buffers to provide a
low conductivity background [195]. If an external base must be added, LiOH will
be preferred since Li+ has such weak affinity for other cation exchangers.
(Indeed, it will be better to use LiOH rather than NaOH in ammonia elution and
measurement schemes as well). However, it is a long way from this knowledge to
establish whether any of the available buffer material can be purified to a suffi-
cient degree to make such trace measurements practical. Further, the optimum
152
concentration of such a buffer, as a compromise between maintaining pH and
achieving good preconcentration, will also need to be determined.
153
5.6 CONCLUSIONS
Diffusion based collection of trace gases into a liquid not only offers a means of
discrimination from concurrently present particles, the integration of continu-
ous or batch mode analytical systems for the analysis of the liquid effluent
provides unprecedented sensitivity and accuracy. Based on these principles,
many gases can be routinely measured at low to sub-ppb levels. DS devices
provide a contained liquid volume and can be operated in a stopped liquid flow
mode where they can themselves provide further preconcentration. They are
attitude insensitive and can be (and has been) used in airborne applications.
LCW based DS devices constitute a unique new entry and may offer very
inexpensive alternatives for measuring gases at typical urban levels. Selective
luminogenic reactions for gas measurement using LCW DS devices have not
been reported and may be uniquely attractive and sensitive. Wetted denuders
exhibit direct gas liquid contact. Especially in the annular and parallel plate
design, they offer high efficiency removal of gases at high flow rates. They are
however, attitude sensitive and have not been used in airborne applications. In
vertical WADs and PPWDs, the two liquid streams are physically distinct. Is it
possible to use different scrubber liquids with different selectivities (e.g., one
acid, one base) to capture different analytes? DS analogs to the more efficient
wet denuder designs, especially the PPWD, should also be particularly
interesting. Much has transpired in this field since the last review [1] and it is an
exciting time to look towards the future.
Acknowledgements
I thank Kajori Dasgupta for making it possible for me to write this review. Genfa
Zhang is thanked for his assistance to procure many not-so-easy-to-find refer-
ences, often several times a day. I thank USEPA in general and Bill McClenny in
particular, for continued support of our efforts in this area. This review was
made possible in part by the Tropospheric Ozone Research Program of the
National Exposure Research Laboratory, U.S. EPA. In the draft stage, I received
the benefit of meticulous comments from Albrecht Neftel, Don Olson, Kei Toda,
and Zbynek Vecera. I thank them in particular and also many others who read it
and provided positive reinforcement. Finally I thank my Friday, Susie Gonzalez,
for making all the corrections and reformatting and renumbering the citations.
REFERENCES
154
5 J.S. Townsend, Philos. Trans. R. Soc. London, 193 (1900) 129-158.
6 W.L. Crider, N.O. Barkley, M.L. Knott and R. Slater, Anal. Chim. Acta, 47 (1969)
237-241.
7 P.G. Gormley and M. Kennedy, Proc. Roy. Soc. Ir. Acad. Sci., 52A (1949) 163-169.
8 D.B. Ingham, J. Aerosol Sci., 6 (1975) 125.
9 D.Y.H. Pui, C.W. Lewis, C.-J. Tsai and B.Y.H. Liu, Environ. Sci. Technol., 24 (1990)
307-312.
10 M. Krieger and R.A. Hites, Enuiron. Sci. Technol., 26 (1992) 1551-1555.
11 M. Krieger and R.A. Hites, Enuiron. Sci. Technol., 28 (1994) 1129-1133.
12 M. Dudek, L. Wolska, M. Pilarczyk and J. Namiesnik, Chemia Analyticzna, 45 (2000)
781-794.
13 M. Dudek, L. Wolska, M. Pilarczyk, B. Zygmunt and J. Namiesnik, J. High Resolut.
Chromatogr., 23 (2000) 449-454.
14 P. Koutrakis, C. Sioutas, S.T. Ferguson, J.M. Wolfson, J.D. Mulik and R.M. Burton,
Environ. Sci. Technol., 27 (1993) 2497-2501.
15 C. Sioutas, P.Y. Wang, S.T. Ferguson, P. Koutrakis and J.D. Mulik, Atmos. Environ.,
30 (1996) 885-895.
16 P. Koutrakis, P.-Y. Wang, J.M. Wolfson and C. Sioutas, US Patent 5,932,795, August
1999.
17 http://www.rpco.com/products/ambprod/amb3500/index.htm
18 C. Sioutas, P. Koutrakis and J.M. Wolfson, Aerosol Sci. Technol., 21 (1994) 137-148.
19 M. Possanzini, A. Febo and A. Liberti, Atmos. Environ., 17 (1983) 2605-2610.
20 Z. Ali, C.L. Paul Thomas and G.F. Alder, Analyst (London), 114 (1989) 759-769.
21 http: // www.rpco.com / products / ambprod / amb2500 / index.htm http:// snobear.
colorado.edu/ cgi-bin/Kiowa/Kiowa.con.pl?hsdenuder.doc. html
22 F. De Santis, I. Allegrini, P. Di Filippo and D. Pasela, Atmos. Environ., 30 (1996)
2637-2645.
23 B.J. Fan, Y.S. Cheng and H.C. Yeh, Aerosol Sci. Technol., 25 (1996) 113-120.
24 R.W. Coutant, P.J. Callahan, M.R. Kuhlman and R.G. Lewis, Atmos. Environ., 23
(1989) 2205-2211.
25 P.K. Simon and P.K. Dasgupta, Anal. Chem., 65 (1993) 1134-1139.
26 F. De Santis, Anal. Chem., 66 (1994) 3503-3504.
27 T. Staffelbach, A. Neftel, A. Blatter, A. Gut, M. Fahrni, J. Stahelin, A. Prevot, A.
Hering, M. Lehning, B. Neininger, M. Baumle, G.L. Kok, J. Dommen, M. Hutterli
and M. Anklin, J. Geophys. Res., 102 (1997) 23,345-23,362.
28 D.J. Eatough, A. Wadsworth, D.A. Eatough, J.W. Crawford, L.D. Hansen and E.A.
Lewis, Atmos. Environ., 27A (1993) 1213-1218.
29 H. Tang, E.A. Lewis, D.J. Eatough, R.M. Burton and R.J. Farber, Atmos. Environ., 28
(1994) 939-947.
30 D.J. Eatough, H. Tang and W. Cui, J. Machir. Inhal. Toxicol., 7 (1995) 691-710.
31 D.J. Eatough, D.A. Eatough, L. Lewis and E.A. Lewis, J. Geophys. Res., 101 (1996)
19,515-19,532.
32 W. Cui, D.J. Eatough and N. Eatough, J. Air Waste Manage. Assoc., 49 (1998)
1024-1037.
33 D.J. Eatough, F. Obeidi, Y. Pang, Y. Ding, N.L. Eatough and W.E. Wilson, Atmos.
Environ., 33 (1999) 2835-2844.
34 C. Sioutas, P. Koutrakis and B.A. Olson, Aerosol Sci. Technol., 21 (1994) 223-235.
35 C. Sioutas, P. Koutrakis and R.M. Burton, J. Aerosol Sci., 25 (1994) 1321-1330.
36 C. Sioutas, P. Koutrakis, R.M. Burton, Particul.Sci. Technol., 12 (1994) 207-221.
37 H. Patashnick and E.G. Rupprecht, J. Air Waste Manage. Assoc., 41 (1991)
1079-1083.
38 E.G. Rupprecht, M. Meyers and H. Patashnick, J. Aerosol Sci., 23 (1992) S635-S648.
155
39 P.K. Dasgupta, L. Ni, S.K. Poruthoor and D.C. Hindes, Anal. Chem., 69 (1997)
5018-5023.
40 D.A. Lewis, I. Colbeck and D.C. Mariner, Anal. Chem., 66 (1994) 3525-3527.
41 G.P. Vassilev and F. Roubani-Kalantzopolou, J. Chem. Soc. Faraday Trans., 91
(1995) 485-494.
42 J. Slanina and G.P. Wyers, Fres. Z. Anal. Chem., 350 (1994) 467-473.
43 M.G. Mennen, B.G. Van Elzakker, E.M. Van Putten, J.W. Uiterwijk, T.A. Regts, J.
Van Hellemond, G.P. Wyers, R.P. Otjes, A.J.L. Verhage, L.W. Wouters, C.J.G.
Heffels, F.G. Rimer, L. Van Den Beld and J.E.H. Tetteroo, Atmos. Environ., 30
(1996) 3239-3256.
44 G. Skillas and S. Kuenzel, J. Aerosol. Sci., 28 (1997) S43-S44.
45 H. Burtscher and S. Kuenzel, J. Aerosol. Sci., 26 (1996) S129-S130.
46 A.S. Kozlov and A.N. Ankilov, J. Aerosol. Sci., 28 (1997) S131-S132.
47 C.L. Paul Thomas, C.D. McGill and R. Towill, Analyst (London), 122 (1997)
1471-1476.
48 R. Bos, J. Air Pollut. Contr. Assoc., 30 (1980) 1222-1224.
49 I. Gacs and R. Ferraroli, Anal. Chim. Acta, 269 (1992) 177-185.
50 Y. Luo, R. Al-Othman, G.D. Christian and J. Ruzicka, Talanta,42 (1995) 1545-1551.
51 P.K. Dasgupta, Atmos. Environ., 18 (1984) 1593-1599.
52 R.L. Tanner, J. Shen, Aerosol. Sci. Technol., 12 (1990) 86-97.
53 I.C. Van Nugteren-Osinga, M. Bos and W.E. Van der Linden, Anal. Chim. Acta, 226
(1989) 171-175.
54 D.A. Phillips and P.K. Dasgupta, Sep. Sci. Technol., 22 (1987) 1255-1267.
55 K. Toda, P.K. Dasgupta, J. Li, G.A. Tarver and G. Zarus, Anal. Chem. (in press).
56 Q. Fan and P.K. Dasgupta, Anal. Chem., 66 (1994) 551-556
57 P.K. Dasgupta, S. Dong, H. Hwang, H.-C. Yang and Z. Genfa, Atmos. Environ., 22
(1988) 949-963.
58 S. Waiz, B.M. Cedillo, S. Jambunathan, S.G. Hohnholt, P.K. Dasgupta and D.K.
Wolcott, Anal. Chim. Acta, 428 (2001) 163-171.
59 T. Gilpin, E. Apel, A. Fried, B. Wert, J. Calvert, Z. Genfa, P.K. Dasgupta, J.W.
Harder, B.G. Heikes, B. Hopkins, H. Westberg, T. Kleindienst, Y.-N. Lee, X. Zhou, W.
Lonneman and S. Sewell, J. Geophys. Res., 102 (1997) 21,161-21,188.
60 A.L. Sumner and P.B. Shepson, Nature, 398 (1999) 6724-6726.
61 A.L. Sumner, P.B. Shepson, T.L. Couch, T. Thornberry, M.A. Carroll, S. Sillman, M.
Pippin, S. Bertman, D. Tan, I. Faloona, W. Brune, V. Young, O. Cooper, J. Moody and
W. Stockwell, J. Geophys. Res., (2001) in press.
62 A.L. Sumner, P.B. Shepson, A.M. Grannas, J.W. Bottenheim, K.G. Anlauf, D.
Worthy, W.H. Schroeder, A. Steffen, F. Domine, S. Perrier and S. Houdier, Atmos.
Environ. (in press).
63 J. Li, P.K. Dasgupta, Z.Genfa and M.A. Hutterli, Field Anal. Chem. Technol., 5
(2001) 2-11.
64 P.K. Dasgupta, Z. Genfa, J. Li, C.B. Boring, S. Jambunathan and R. Al-Horr, Anal.
Chem., 71 (1999) 1400-1407.
65 http://www.alphaomegapt.com/methanalyzer.htm
66 M.J. Navas, A.M. Jimenez and G. Galan, Atmos. Environ., 33 (1999) 2279-2283.
67 W.G. Gunz and M.R. Hoffman, Atmos. Environ., 24A (1990) 1601-1634.
68 A.V. Jackson, Crit. Rev. Env. Sci. Technol., 29 (1999) 175-228.
69 A. Sigg, A. Neftel and P.K. Dasgupta, in: Proceedings of the Workshop on the
Development of Analytical Techniques for Atmospheric Pollutants, Rome, April
13-15, 1992. Air PollutionResearch Report41 Commission of the European Commu-
nities, I. Allegrini (Ed.), pp. 111-117. Dordrecht, Boston, 1993.
70 Z. Genfa and P.K. Dasgupta, Anal. Chem., 64 (1992) 517-522.
156
71 G. Zhang, P.K. Dasgupta and A. Sigg, Anal. Chim. Acta, 260 (1992) 57-64.
72 Z. Genfa, P.K. Dasgupta, G.M. Frick and W.A. Hoppel, Microchem. J., 62 (1999)
99-113.
73 J. Dommen, A. Neftel, A. Sigg and D.J. Jacob, J. Geophys. Res., 100 (1995)
8953-8966.
74 J. Li and P.K. Dasgupta, Anal. Chem., 72 (2000) 5338-5347.
75 http://www.cgenv.com/Narsto/
76 www.globalfia.com
77 J. Li and P.K. Dasgupta, Anal. Chim. Acta, 442 (2000) 63-70.
78 http://www.celgard.com/products/hfmproductinfo.cfm
79 Y. Komazaki, Y. Narita and S. Tanaka, Analyst, 123 (1998) 2343-2349.
80 Y. Komazaki, S. Hashimoto, T. Inoue, S. Tanaka, Atmos. Environ. (2002) in press.
81 Y. Komazaki, T. Inouie and S. Tanaka, Analyst, 126 (2001) 587-593.
82 Y. Komazaki, S. Hashimoto, T. Inouie and S. Tanaka, Atmos. Environ., (2002) in
press.
83 L. Bao and P.K. Dasgupta, Anal. Chem., 64 (1992) 991-996.
84 F.R. Rocha, J.A.F, da Silva, C.L. do Lago and I.G.R. Gutz, Abstract SE-79, Proceed-
ings, 11 Encontro Nacional de QuimicaAnalitica, Campinas, Brazil, September 2001.
85 S. Liu and P.K. Dasgupta, Anal. Chim. Acta, 308 (1995) 281-285.
86 Z. Genfa, P.K. Dasgupta and S. Dong, Environ. Sci. Technol., 23 (1989) 1467-1474.
87 R.M. Harrison and I.M. Msibi, Atmos. Environ., 28 (1994) 247-255.
88 L.L. S0rensen, K. Granby, H. Nilesen and W.A.H. Asman, Atmos. Environ., 28 (1994)
3637-3645.
89 C. Milford, M.A. Sutton, A.G. Allen, A. Karlsson, B.M. Davison, J.D. James, K.
Rosman, R.M. Harrison and J.N. Cape, Tellus, 52B (1990) 273-289.
90 J.C. Park, E.Y. Bae, C.G. Park, W.S. Han, Y.B. Koh, M.Y. Lee and J.G. Lee,
Proceedings of the Symposium on Ultrasonic Electronics. Jpn. J. Appl. Phys., 34
(1995) 6770-6773.
91 I.-H. Chang, D.-S. Lee, Y.-K. Lee and K.-J. Whang, Anal. Sci., 13 suppl. (1997)
267-272.
92 P.F. Lindgren and P.K. Dasgupta, Anal. Chem., 61 (1989) 19-24.
93 D.S. Lee, Yonsei University, Seoul, Korea. Personal communication, 2001.
94 D.S. Lee and I.H. Chang, Proceedings of the Insung Chromatech Conference on Ion
Chromatography, Seoul, Korea, August 13, 2001.
95 D.S. Lee, Abstract No. 94, Abstracts, 2001 International Ion Chromatography
Symposium, Oakbrook, IL.
96 U. Hase and Y. Matsuyoshi, NEC Res. Dev., 39 (1998) 95-100.
97 Y. Matsuyoshi, Y. Satoh, T. Shinozaki, E. Suzuki and N. Nagata, NEC Res. Dev., 41
(2000) 93-97.
98 Y. Satoh and Y. Matsuyoshi, NEC Res. Dev., 42 (2001) 314-318.
99 W. Frenzel, Fres. J. Anal. Chem., 342 (1992) 817-821.
100 W. Frenzel, Anal. Chim. Acta, 291 (1994) 305-320.
101 C.J. Patton and A.P. Wade, Continuous flow analyzers. In: G.W. Ewing (Ed.),
Analytical InstrumentationHandbook, 2nd Edn. Marcel Dekker, New York, 1997,
pp. 125-220.
102 J. Ruzicka and E.H. Hansen, Flow Injection Analysis, 2nd Edn. Wiley, 1988.
103 J. Crank and G.S. Park, Diffusion in Polymers. Academic Press, London, 1968.
104 J.E. Lundstrom and R.J. Bearman, J. Polym. Sci., 12 (1974) 97-114.
105 D. Scheppers, G. Schulze and W. Frenzel, Anal. Chim. Acta, 308 (1995) 109-114.
106 K. Toda, H. Inoue and I. Sanemasa, Bunseki Kagaku, 47 (1998) 727-734.
107 V. Kuban and P.K. Dasgupta, Anal. Chem., 64 (1992) 1106-1112.
108 G.A. Tarver and P.K. Dasgupta, Atmos. Environ., 29 (1995) 1291-1298.
157
109 G.A. Tarver and P.K. Dasgupta, Environ. Sci. Technol., 31 (1997) 3669-3676.
110 H. Kobara, K. Takeuchi and T. Ibusuki, Bunseki Kagaku, 46 (1997) 881-886.
111 Y. Suzuki, Anal. Chim. Acta, 353 (1997) 227-237.
112 M. Stigbrand, A. Karlsson and K. Irgum, Anal. Chem., 66 (1996) 3945-3950.
113 H.B. Ross, C. Johansson and C. de Serves, J. Atmos. Chem., 14 (1992) 411-423.
114 C. de Serves and H.B. Ross, Environ. Sci. Technol., 27 (1993) 2712-2718.
115 W. Matuszewski and M,E. Meyerhoff, Anal. Chim. Acta, 248 (1991) 379-389.
116 C. de Serves, J. Geophys. Res., 99 (1994) 25391-25398.
117 D. Trapp and C. de Serves, Atmos. Environ., 29 (1995) 3239-3243.
118 Y. Komazaki, M. Hiratsuka, Y. Narita, S. Tanaka and T. Fujita, Fres.J. Anal. Chem.,
363 (1999) 686-695.
119 K. Toda, P.K. Dasgupta, J. Li, G.A. Tarver, G. Zarus, S. Ohira and E.L. Loree, Anal.
Sci. (in press).
120 K. Toda, Kumamoto University, Japan. Personal communication, August, 2001.
121 D.C. Olson, S.R. Bysouth, P.K. Dasgupta and V. Kuban, Process Control Qual., 5
(2000) 259-265.
122 A.Yu. Alentiev, Yu. P. Yampolskii, V.P. Shantarovich, S.M. Nemser and N.A. Plat,
J. Membr. Sci., 126 (1997) 123-132.
123 I. Pinnau and L.G. Toy, J. Membr. Sci., 109 (1996) 125-133.
124 K.E. Hong and L.W. Burgess, Proc. SPIE, 2293 (1994) 71-74.
125 P.K. Dasgupta, Z. Genfa, S.K. Poruthoor, S. Caldwell, S. Dong and S.-Y. Liu, Anal.
Chem., 70 (1998) 4661-4669.
126 P,K. Dasgupta and S.-Y. Liu, US Patent 6,011,882, January 4, 2000.
127 H. Fein and S.-Y. Liu, US Patent 6,016,372, January 18, 2000.
128 M.R. Milani and P.K. Dasgupta, Anal. Chim. Acta, 431 (2001) 169-180.
129 0. Inya-Agha, S. Stewart, T. Veriotti, M.L. Bruening and M.D. Morris, Appl.
Spectrosc. (2002) in press.
130 H. Liu and P.K. Dasgupta, TrAC, 15 (1996) 468-475.
131 H. Liu and P.K. Dasgupta, Microchem. J., 57 (1997) 127-136.
132 S. Liu and P.K. Dasgupta, Anal. Chem., 67 (1995) 2042-2049.
133 E.P. Felix, C. Ugucione and A.A. Cardoso, Abstract EM-22, Proceedings, 11 Encontro
Nacional de Quimica Analitica, Campinas, Brazil, September 2001.
134 E.A. Pereira and P.K. Dasgupta, Int. J. Environ. Anal. Chem., 66 (1997) 201-213.
135 E.A. Pereira, A.A. Cardoso and P.K. Dasgupta, Quim. Nova, 24 (2001) 443-448.
136 M.R. Milani, J.A.C. Neto and A.A. Cardoso, Microchem. J., 62 (1999) 273-281.
137 A.A. Cardoso, H. Liu and P.K. Dasgupta, Talanta 44 (1997) 1099-1106.
138 K. Toda, M. Kitazaki and I. Sanemasa, Bunseki Kagaku, 49 (2000) 989-995.
139 H. Liu and P.K. Dasgupta, Anal. Chem., 67 (1995) 4221-4228.
140 http://www.geocities.com/ResearchTriangle/Lab/4688/
141 H. Liu and P.K. Dasgupta, Anal. Chim. Acta, 289 (1994) 347-353.
142 H. Amokrane and B. Caussade, J. Atmos. Sci., 56 (1999) 1808-1829.
143 A. Cardoso and P.K. Dasgupta, Anal. Chem., 67 (1995) 2562-2566.
144 S. Kar and P.K. Dasgupta, Am. Lab., 29 (1997) 17C-17M.
145 P.K. Dasgupta and S. Kar, Anal. Chem., 67 (1995) 3853-3860.
146 S. Kar and P.K. Dasgupta, J. Chromatogr., 739 (1996) 379-387.
147 K. Surowiec and P.K. Dasgupta, J. Microcol., 10 (1998) 265-271.
148 H. Huang and P.K. Dasgupta, Talanta, 44 (1997) 605-615.
149 H. Huang and P.K. Dasgupta, Electroanalysis, 9 (1997) 585-591.
150 Z. Genfa and P.K. Dasgupta, Anal. Chem., 72 (2000) 3165-3170.
151 P.K. Simon, P.K. Dasgupta and Z. Vecera, Anal. Chem., 63 (1991) 1237-1242.
152 Z. Vecera and P.K. Dasgupta, Anal. Chem., 63 (1991) 2210-2216.
153 Z. Vecera and P.K. Dasgupta, Int. J. Anal. Chem., 4 (1994) 311-316.
158
154 M. Taira and Y. Kanda, Anal. Chem., 65 (1993) 3171-3173.
155 A. Neftel, A. Blatter, R. Hesterberg and T. Staffelbach, Atmos. Environ., 30 (1996)
3017-3025.
156 A. Blatter, A. Neftel, P.K. Dasgupta and P.K. Simon, in: G. Angletti and G. Restelli
(Eds.), Physico-Chemical Behavior of Atmospheric Pollutants, Proc. 6th European
Symposium, Report EUR 15609/2 EN, Luxembourg, 1994. pp. 767-772.
157 A. Neftel, A. Blatter, A. Gut, D. Hogger, F. Meixner, C. Ammann and F.J. Nathaus,
Atmos. Environ., 32 (1998) 499-505.
158 Z. Zdrahal, P. Mikuska and Z. Vecera, Chem. Listy, 88 (1994) 353-359.
159 Z. Zdrahal and Z. Vecera, J. Chromatogr., 668 (1994) 371-374.
160 Z. Zdrahal, P. Mikuska and Z. Vecera, Anal. Chem., 67 (1995) 2763-2766.
161 J. Peskova, P. Parizek and Z. Vecera, J. Chromatogr., 918 (2001) 153-158.
162 W. Jaeschke, J.P. Dierssen, A. Gunther and M. Schumann, Atmos. Environ., 32
(1998) 365-371.
163 A.M. Costa, A.P.S. Novato, V.P. Campos and T.M. Tavares, Abstract AB-70, Proc., 11
Encontro Nacional de Quimica Analitica, Campinas, Brazil, September 2001.
164 D.R. Fitz and N. Motallebi, J. Air Waste Manage. Assoc., 50 (2000) 981-992.
165 D.S. Poon, D.Y.H. Pui, C.T. Lee and B.Y.H. Liu, J. Aerosol Sci., 25 (1994) 923 -934.
166 C.J. Tsai, C.H. Huang, S.H. Wang and T.S. Shih, Aerosol Sci. Technol., 35 (2001)
611-616.
167 S.M. Buhr, M.P. Buhr, F.C. Fehsenfeld, J.S. Holloway, U. Karst, R.B. Norton, D.P.
Parrish and R.E. Sievers, Atmos. Environ., 26 (1995) 2609-2624.
168 Y. Komazaki, H. Shimizu and S. Tanaka, Atmos. Environ., 33 (1999) 4363-4371.
169 I.H. Chang, N.H. Choi, B.K. Lee and D.S. Lee, Bull. Kor. Chem. Soc., 20 (1999)
329-332.
170 D.S. Lee, N.H. Yu, S.Y. Lee and J.S. Whang, Korean Patent 97-21338.
171 I.H.Chang, Ph.D. Dissertation, Yonsei University, Korea, August 2001.
172 W.C. Hinds, Aerosol Technology. Wiley, New York, 1982. p 161.
173 M. Keuken, C.A.M. Schoonebeek, A. Wensveen-Louter and J. Slanina, Atmos.
Environ., 22 (1988) 2541-2548.
174 G.P. Wyers, R.P. Otjes and J. Slanina, Atmos. Environ., 27A (1993) 2085- 2090.
175 J. Slanina and G.P. Wyers, Fres. J. Anal. Chem., 350 (1994) 467-473.
176 M.T. Oms, P.A.C. Jongejan, A.C. Veltkamp, G.P. Wyers and J. Slanina, Int. J.
Environ. Anal. Chem., 62 (1996) 207-218.
177 P.A.C. Jongejan, Y. Bai, A.C. Veltkamp, G.P. Wyers and J. Slanina, Int. J. Environ.
Anal. Chem., 66 (1997) 241-251.
178 M.A. Sutton, C. Milford, U. Dragosits, C.J. Place, R.J. Singles, R.I. Smith, C.E.R.
Pitcairn, D. Fowler, J. Hill, H.M. ApSimon, C. Ross, R. Hill, S.C. Jarvis, B.F. Pain,
V.C. Phillips, R. Harrison, D. Moss, J. Webb, S.E. Espenhahn, D.S. Lee, M. Hornung,
J. Ullyett, K.R. Bull, B.A. Emmett, J. Lowe and G.P. Wyers, Environ. Pollut., 102S
(1998) 349-361.
179 M.A. Sutton, E. Perthue, D. Fowler, R.L. Storeton-West, J.N. Cape, B.G. Arends and
J.J. Mols, Atmos. Environ., 31 (1997) 2615-2624.
180 D. Fowler, M. Coyle, C. Flechard, K. Hargreaves, E. Nemitz, R. Storeton-West, M.
Sutton and J.W. Erisman, Plant Soil, 228 (2001) 117-129.
181 V.R. Phillips, D.S. Lee, R. Scholtens, J.A. Garland and R.W. Sneath, J. Agric. Eng.
Res., 78 (2001) 1-14.
182 R.M. Carlson, Anal. Chem., 50 (1978) 1528-1531; US Patent 4,209,299, June 24,
1980.
183 H.S.M. Devries, F.J.M. Harren, G.P. Wyers, R.P. Otjes, J. Slanina and J. Reuss,
Atmos. Environ., 29 (1995) 1069-1074.
184 P.K. Simon and P.K. Dasgupta, Environ. Sci. Technol., 29 (1995) 1534-1541.
159
185 X. Li-Jones, D.L. Savoie and J.M. Prospero, Atmos. Environ., 35 (2001) 985-992.
186 http://www-wlc.eas.gatech.edu/supersite/
187 Z. Genfa, B. Hartsell, J. Slanina, K. Baumann, C.B. Boring, P.A.C. Jongejan, E.
Edgerton and P.K. Dasgupta, J. Geophys. Res. (submitted)
188 http://www.utexas.edu/research/ceer/texaqs/
189 K. Ito, C.C. Chasteen, H.-K. Chung, S.K. Poruthoor, Z. Genfa and P.K. Dasgupta,
Anal. Chem., 70 (1998) 2839-2847.
190 K. Rosman, M. Shimmo, A. Karlsson, H.-C. Hansson, P. Keronen, A. Allen and G.
Hoenninger, Atmos. Environ., 35 (2001) 5301-5310.
191 R. Hesterberg, A. Blatter, M. Fahrni, M. Rosset, A. Neftel, W. Eugster and H.
Wanner, Environ. Pollut., 91 (1996) 21-34.
192 M. Ammann, M. Kalberer, DT. Jost, L. Tobler, E. Rossler, D. Piguet, H.W. Gaggeler
and U. Baltensperger, Nature, 395 (1998) 157-160.
193 C. Zellweger, M. Ammann, P. Hofer and U. Baltensperger, Atmos. Enuiron., 33
(1999) 1131-1140.
194 M. Loflund, A. Kasper-Giebl, W. Tscherwenka, M. Schmid, H. Giebl, R.
Hitzenberger, G. Reischl and H. Puxbaum, Atmos. Environ., 35 (2001) 2861-2869.
195 J.P. Ivey, J. Chromatogr., 287 (1984) 128-132.
196 Z. Genfa, T. Uehara, P.K. Dasgupta, T. Clarke and W. Winiwarter, Anal. Chem., 70
(1998) 3656-3666.
160
Chapter 6
6.1 INTRODUCTION
162
ideally provide size discriminated collection and analysis in all size ranges of
interest. If only a composite analysis can be performed, it is vital that the aerosol
is quantitatively collected in all the size ranges of interest. Even a small bias
towards collection of larger particles can affect the overall composition result
significantly because larger particles can represent a much larger mass fraction.
In filter collection methods, it is not difficult to select suitable filters to achieve
quantitative collection across the entire size ranges of interest. Not all com-
monly used filter media are actually quantitatively retentive of fine particles
[42,43].
There are two general approaches to aerosol composition measurements.
Traditionally, particles are sampled on some media and later taken to the labora-
tory and analyzed. Many of the instruments available for National Ambient Air
Quality Standard (NAAQS) compliance measurements were compatible with
this approach and the bulk of the extant data on chemical composition informa-
tion of ambient aerosols has been obtained by this technique. Atmospheric aero-
sols can be collected by inertial methods, gravitational settling, centrifugation,
electrostatic and thermal precipitation, and filtration. Of these, filtration, with
or without inertial classification, is the most commonly used approach [27,
44-53]. Collection efficiencies and other characteristics of various filters have
been evaluated and are available in the literature [54]. More often than not,
aerosol samples are collected for an extended period of time, usually many hours.
The collected samples are then stored and later analyzed. It is generally not prac-
tical to complete the analysis step in the field: only the sample collection step is
carried out in the field [18]. The media containing the collected aerosol are
placed in envelopes or sealed containers and transported to a regular laboratory
for analysis. The total elapsed period may range from days to months. Obviously,
such an arrangement can neither provide good temporal resolution nor measure
unstable species. The integrity of the process may be compromised when sam-
ples are stored for long periods of time: collected species may undergo various
chemical and physical changes and therefore the analytical results may not be
meaningful. There are also problems particular to specific analytes and specific
collection/analysis methods (e.g., sampling surfaces absorb or react with gases
and particles leading to positive or negative artifacts) [28,53, 55-62]. As dis-
cussed in a previous chapter, most often it is necessary to remove potentially
interfering gases with a diffusion denuder before the aerosol is collected or oth-
erwise processed. Discrete sampling and analysis is very difficult to automate
and is consequently labor-intensive. In the long run, large-scale measurements
by such approaches are not economically attractive.
In the alternative approach, airborne particles are processed directly
through a real-time or near real-time dynamic measurement system. The collec-
tion (if there is a discrete collection step) and analysis systems are integrated and
the entire system is field deployed, permitting rapid data output and signific-
antly lower cost per analysis. Methods of different types have been reported for
the automated determination of several individual species in aerosols.
163
6.2.1 Thermal decomposition based methods
164
reduces bounce-off from the impaction surface; laboratory generated (NH4 )2 SO4
aerosols in the 100-800 nm size range were collected with 95-99.8% efficiency
[77]. A commercial version of this instrument is now available with a time
resolution of 10 min and a concentration resolution of 200 ng/m3 [78].
Whereas the Hering and Stolzenburg approach is based on actual collection
by impaction, a recent strategy developed by Harvard School of Public Health
researchers [79] is more reminiscent of previous thermodenuder based sulfate
measurement/speciation approaches. Referring to Fig. 6.1, the instrument is
zeroed by allowing it to sample particle-free air that is made to pass by the
sample inlet. A small amount of ammonia is continuously added to the sampled
air stream from an on-board permeation source. The amount is not critical as
long as it is sufficient to neutralize the aerosol sulfate completely to (NH4 )2 SO4
and thus avoid any difference in thermal behavior of (NH4 )2 SO4 from more acidic
sulfates, particularly H2 SO4 . However, large excess should be avoided since NH3
is oxidized to NO and contributes to the baseline noise. In urban areas, where
the aerosol is generally not present as H2 SO4 , it is not necessary to add any NH3
at all. During sampling, the instrument draws ambient air through an inlet
system that removes particles larger than 2.5 gm. Ammonia (few hundred ppb at
10-15 cc/min) is then added to the 0.65 1/min sample stream (the ammonia flow
rate is kept small relative to the main stream so dilution is insignificant). The
mixture proceeds through an alkaline denuder that removes acidic sulfur gases.
The sulfur-gas free aerosol stream then proceeds through a thermal converter (6
ft long stainless steel coil maintained at 800°C), which converts all the sulfate to
SO,. The conversion of sulfate to SO2 is highly efficient. The carbon present in
the steel may play a role in this conversion, but this has not yet been confirmed.
Ambient Air In Systiem Inlet
+ PM,,-lmpacto'r or Cyclone Inlet
I Particle Filter D 7
165
Any residual particles are removed from the gas stream prior to measurement of
SO2 concentration with a modified high sensitivity pulsed fluorescence based
analyzer. The system differs from previous thermodenuder-based approaches in
several ways. The most important difference is that the approach is truly
continuous; there are no discrete collection and analysis steps. From a practical
point of view, it also does not rely on the FPD and is therefore intrinsically safe
and does not need compressed gases. Although it is not a factor in urban
measurements, the sensitivity of current SO 2 analyzers is such, however, that
long integration times are needed to measure low sulfate concentrations. The
LOD is 0.5 g/m3 for a 1-h mean, which is equivalent to twice the standard
deviation of clean ambient air. The precision of collocated prototype instruments
for 1-h mean values in urban areas is -7%. It may be better in eventual
commercial versions [80]. A patent has been applied for. The general agreement
with other methods, e.g., ion chromatography (IC) is reportedly excellent.
Similarly, chemiluminescence (CL) detection of liberated NOx has been used
for some time to determine nanogram quantities of nitrate after the thermal
decomposition of nitrate particles [81]. Cox 82] measured nitrite and nitrate in
solution to liberate NO, reducing nitrite with acidic iodide and reducing total
nitrite/nitrate with FeSO4 and ammonium molybdate in conc. H2SO4. The NO
was detected by CL. Klockow et al. [83] described two different approaches for
the measurement of NH4 NO3. In both, all other N species are removed with two
predenuders. NH4NO3 is decomposed and volatilized and collected using either
an MgSO 4 coated denuder at 150°C or a NaF-coated denuder at 140°C. After
collection, the denuder is heated to 7000C and the liberated NOx is detected with
a CL-based monitor. LODs are 60-100 ng/m3 NH4 NO3 , depending on sampling
time and sampling rate.
In coastal areas, significant concentrations of aerosol NaNO3 may be present
from the reaction of gaseous HNO3 with NaCl. It was suggested that NaNO3 and
NH 4NO 3 could be differentiated by their different temperatures of volatilization
[84]. Subsequent work by Sturges and Harrison [85] has shown, however, that
such differentiation is in fact fraught with difficulty; they show that it does not
work for the corresponding chloride case either. More recently, Ten Brink et al.
[86] have reevaluated the thermodenuder based determination of NH4NO 3.
Aside from some problems encountered at high humidities, this instrument is
reported to work well.
Yamamoto and Kosaka [87] collected aerosols on type 304 stainless steel
strips and applied the FV approach with CL-based NO detection. To assure
conversion of NO2 to NO, they added CrO3. This resulted in quantitative
production of NO from NaNO. With NH4 NO,, the NO recovery was only
30-50% and artifact NO production was seen from (NH4)2SO 4 and NH4C1.
Addition of NaOIH prior to heating removed the NH4+ as NH3 and eliminated
this interference.
The recent work of Stolzenburg and Hering [76,77] was mentioned earlier in
connection with a continuous sulfate monitor. Indeed, the same apparatus can
166
Inlet
Precut
-and
Denuder
Ventilated
Enclosure
p
Fig. 6.2. Stolzenburg-Hering apparatus schematic
for the impaction-flash volatilization measurement
of sulfate and nitrate. Reproduced with permission
from Ref. [77] (copyright 2000 American Chemical
Sointv).
be used to measure nitrate with a CL-based NOx monitor; the system setup is
shown schematically in Fig. 6.2. Laboratory generated NH 4 NO3 particles in the
200-800 nm size range were collected with 97-99.8% collection efficiency. The
FV conditions were optimized to elicit minimum response to ammonium, the
response to (NH4 )2SO4 was only 4% of that of a corresponding amount of nitrate.
The possible vaporization of NH4 NO, during collection was investigated. No
significant losses (2±4%) were found. The LOD is primarily governed by the
field blank, which arises due to the adsorption of nitrogen bearing gaseous
material (that are not removed by the denuder) on the impactor surface. This
could be as high as 0.4-0.7 ptg/m3 in high NOx environments. Nevertheless,
excellent agreement with traditional measurement methods has been generally
observed. In intercomparison studies with more traditional denuder-filter mea-
surement methods in three cities, correlation coefficients were greater than 0.97
and regression slopes ranged from 0.96 to 1.06. Over prolonged field experi-
ments, the system exhibited data recovery rates as high as 97%. A commercial
version of this instrument is now available with a time resolution of 10 min and a
concentration resolution of 200 ng/m 3 [88]. Inter-instrument precision is excel-
lent, the output of two collocated instruments are shown in Fig. 6.3.
Spectroscopic methods for aerosol analysis have generally been used only in a
post-sampling off-line mode. The determination of aerosol carbon is an excep-
tion; this is discussed in a later section. Techniques available for analysis of
aerosol particles, once they have been collected, are of course, quite numerous.
An exhaustive review of these approaches is not appropriate here; the present
167
or n
l JUU ?
A -Nitrate, Instrument 1 ONitrate, instrument 2
O 10.0
E
2
. -
1
20.0- i i i
E0
4~~~~~~~~~~~~~~~~~~~
c:
z
o.1 0.I4Ak,;k&A4*V _1"
0.10 1.00 10.00 8114100 8115/00 8116100 8117/00
Nitrate, pg/m3, Instrument 1 Date
Fig. 6.3. Precision of two collocated nitrate measurement instruments based on impaction-flash
volatilization. Data Courtesy of Susanne V. Hering, Aerosol Dynamics Inc.
168
gaseous and particulate nitrate, can cause partial retention of SO 2 and are not
reliable for particulate sulfate measurement. The positive interference from SO2
is variable and can be as high as 70% [28,100-102]. Sulfate measurements with
filters of relatively low alkalinity (Teflon and quartz fiber) tend to be more
reliable. Similar interferences from the adsorption of organic vapors on the
collection media occur for organic and elemental carbon measurement as well.
The use of diffusion denuders for removal of gases prior to collection of
particles, discussed in detail the foregoing chapter, is by far the best way to
prevent artifacts due to reactions with gases. Many aerosol chemical speciation
studies as well as many atmospheric gas measurement efforts have utilized
diffusion denuders [103-107].
The loss of particles by volatilization or artifact generation due to
particle-particle reactions are not, however, eliminated by the use of diffusion
denuders. The only way to ameliorate this is to use a small collection period.
The application of mass spectrometry (MS) to aerosol analysis has had a long
and illustrious history. Real-time measurement via ionization of collected parti-
cles followed by MS analysis was demonstrated more than two decades ago [108].
Mansoori et al. [109] demonstrated the quantitation of ionic species in single
microdroplets by on-line laser desorption/ionization. For real-time aerosol sam-
pling and analysis, if size information is obtained for the same particle for which
chemical composition data are generated, the resulting size-dependent compo-
sition information will be particularly valuable. As the sample is drawn through
the system across a laser beam path, the size information is deduced from the
magnitude of the scattered light registered on a photodetector placed at right
angles to the beam path. A second high-power pulsed laser is then used for the
photoionization of the constituents of the aerosol particle and this is followed by
quadrupole MS [110]. This arrangement provides a brief ion burst and does not
allow m/e scanning of the quadrupole within the time available for a single
particle. This limitation was removed by using time-of-flight mass spectrometry
(TOFMS) instead of a quadrupole MS and then further developed to permit
simultaneous analysis of both positive and negative ions from a single particle,
by adding a second time-of-flight tube [111,112]. While such an arrangement
provides satisfactory results with monodisperse aerosols, difficulties are still
encountered when particles of different size are present. This is because the
second ionizing laser pulse illuminates the particle at a different spatiotemporal
coordinate, after it has passed through the sizing beam traverse. The travel time
between the two illumination zones is size dependent. Theoretically it is possible
to scan for different size particles, but it has not been reduced to practice. This
difficulty is eliminated by firing the high energy ionizing pulse immediately after
the sizing pulse, so that the particle effectively does not travel any appreciable
distance within this time [113-116]. Actual applications for the measurement of
169
atmospheric aerosol particle composition have been demonstrated 117,118].
Mobile versions of aerosol mass spectrometers have been reported [119-121].
In the last few years, single particle mass spectrometry has blossomed; for a
more detailed account the reader is referred to several recent reviews [122-124].
Maynard [125] has examined the various methods available for the characteriza-
tion of single ultrafine (<100 nm) particles to see how different methods can
determine features in terms of size, morphology, topology, composition, structure
and physicochemical properties. Size, morphology and surface properties are best
characterized by scanning transmission electron microscopy (STEM), while high-
resolution transmission electron microscopy allows structural information on
particles and atomic clusters to sub-0.2 nm resolution. Electron energy loss spec-
troscopy and X-ray emission in the STEM allow the chemical analysis of particles
and particle regions down to nanometer diameters. Scanning probe microscopy
offers the possibility of analyzing nanometer-diameter particles under ambient
conditions. Developments in aerosol mass spectrometry are providing the means
for chemically characterizing size-segregated ultra-fine particles down to 10 nm in
diameter.
Lazar et al. [126] have shown that it is possible to determine just the surface
composition of individual airborne particles by separating desorption and ioniza-
tion steps using a two-laser real-time aerosol mass spectrometry technique. A
weak excimer laser pulse is first used to desorb the semivolatile components
from the particle surface when the particle is in the center of an ion-trap MS.
After a short delay, another excimer laser pulse is used to ionize the semivolatile
surface components in the gas phase and these ions are then subjected to mass
analysis. The conventional one-laser and the two-laser technique above produce
complementary results. The two-laser technique was found to be particularly
suitable for determining polynuclear aromatic hydrocarbons (PAHs) adsorbed
on particle surfaces. PAH compounds adsorbed on contaminated soil particles
have been studied by this group [127]. These authors later reported the first
direct observation of the evolution of the carbonization process in a soot-forming
flame [128,129]. The degree of carbonization of each measured particle, and the
size distribution of both PAH-containing and mature soot particles were deter-
mined. The carbonization process is characterized by rapid hydrogen exchange
between the PAH and pyrolytic addition of small hydrocarbons to form larger
PAH molecules. The hydrogen exchange rate builds until carbon- carbon bond
rearrangement becomes facile. The same authors have shown elsewhere that
even a submonolayer amount of tri(n-)butylphosphate can be monitored in real
time on the surface of micron-sized particles by forming alkali metal ion adducts,
and this can be done with relatively low power lasers because of the ease of pro-
ducing alkali metal cations [130]. Paraquat on the surface of particles has also
been monitored [131].
Real-time aerosol mass spectrometry provides unparalleled power and versa-
tility: the current generation of instruments provide rapid, high throughput
data on both size and chemical composition. Nevertheless, the method is not per-
170
fect: the cost, size, power and operator skill requirements are all substantial.
Quantitative analysis can be matrix and size dependent [132-134]. Kane and
Johnston [133,134] have shown that these dependencies arise from the trans-
mission characteristics of the particle inlet and the intrinsic ability of a particle
to be vaporized and ionized. Small particles are generally more difficult to detect
and analyze than larger particles because of the difficulty of focusing through
the inlet into a tight beam as well as the difficulty of ablation. Particles composed
of PAHs, NH4 NO3 , and alkali metal salts are efficiently ablated by a laser pulse.
Aliphatic organics are less efficiently ablated and (NH 4)2SO4 is very difficult to
detect in a positive ion spectrum. Ultrafine particles fragment so extensively
into small mass fragments that it becomes difficult to distinguish between
aliphatic and aromatic components.
The massive amount of raw data generated by real-time single particle mass
spectrometry is both a blessing and a curse. A significant amount of effort must
be spent to distil the desired information. Despite innovative efforts towards
adapting new and old algorithms to provide artificial intelligence for classifying
the spectra and deriving information from it [135], it is clear that this is not yet a
routine operation. While an aerosol mass spectrometer represents a first rate
research tool, the time for routine ambient monitoring by this technique has not
yet come. It should also be noted that aerosol mass spectrometers differ con-
siderably in their design, performance, and nature of data output. In the Atlanta
Supersite study of 1999 [136], four different aerosol mass spectrometers were
operated at the same site for the first time and the data obtained was both
complementary and overlapping. For an account of this study, see Middlebrook
et al. [137].
Trace metals in aerosols can often be used as unique signatures for a specific
source. This is particularly valuable for source apportionment and other issues
that are otherwise very difficult to address. If such detection/measurement can
be done in real time, there is the exciting possibility of performing eddy
correlations for flux measurement and source apportionment. In favorable
cases, even simple flame emission spectroscopy can be very sensitive. Clark et al.
[138] report that a single 100 nm diameter NaCl particle can be detected with a
modified flame photometer, and this is free from matrix interferences.
A broad multi-element detection scheme in which an air sample is directly
aspirated into a plasma source will seem to be of general use and wide interest.
Unfortunately, most inductively coupled argon plasmas tend to become unstable
when significant amounts of air are introduced. To this end, Bacri and Gomes
initiated studies with air and air-argon plasmas [139]. A preliminary apparatus
for direct analysis of atmospheric aerosols was then introduced [140]; a more
developed version was shown to be adequate for workplace monitoring [141].
171
With an air plasma system, the sample to be analyzed is used directly as the
plasma gas. If an argon plasma is used, the sample to be monitored is injected at
a low flow rate (0.5 l/min) into the plasma. LODs ranged from 1-10 Ag/m 3 with a
response time of less than 1 min [142]. The authors have modeled relevant
relaxation processes in an air plasma [143]. It has also been subsequently shown
that a 50:50 argon:air plasma (total flow 15 l/min) gives a sufficient sample of
airflow for analysis and appreciably lowers the LOD [144].
None of the above instruments are readily transportable. More recently,
Duan et al. [145,146] have described a field-portable low-power instrument
which uses a microwave induced argon plasma that can tolerate up to 20% air.
Although the limits of detection (LODs) are not as yet good enough for use in
ambient air in any of the above instruments, coupled to an inertial particle
concentrators [147-149] such an approach may be practical.
Having examined 1300 samples from 66 urban and rural sites in the US, Shah et
al. [150] reported that carbonaceous particles constitute an average of 13% of the
total suspended particulate mass. About two-thirds of the carbon may be
organic, derived mostly from combustion. Gray et al. [151] found 40% of the fine
particle mass (2.1 Mm) in Los Angeles to be carbonaceous. Offenberg and Baker
[152] found 49% of summertime fine particle (0.15-0.45 Am) mass to be organic.
On a molar basis, carbon is the most common element in atmospheric fine parti-
cles [153]. Organic and elemental carbon (OC and EC) in such particles are the
primary contributors to visibility degradation. Light extinction by OC occurs
primarily by scattering (with an efficiency of 2-5.5 m2 /g) and EC (this is gener-
ally equivalent to the terms black carbon (BC) or graphitic carbon (GC)) is the
major contributor to light absorption (with an extinction efficiency of -10 m2/g
at 515 nm and varying as 1/) [154]. EC has also been suspected to be a catalytic
agent in the conversion of SO2 to acid sulfates [155] and NO2 to HONO [156,
157]. Inhalation of soot particles from diesel-powered buses can be a threat not
only for those outside, but also those inside a bus [158].
Naturally, the measurement of carbonaceous particles has long been
regarded as an important challenge. The first near real-time measurement was
accomplished by Hansen et al. [159,160]. This instrument, named the aethal-
ometer, measures the reduction in light transmission through a filter as air is
sampled through it. The measurement is generally referred to that of black car-
bon. The original instrument used laser illumination. For a time, commercial
versions of the aethalometer utilized incandescent sources. Currently three fully
automated versions are available with self-advancing quartz-fiber filter tapes,
adjustable sampling rates in the 2-6 1/min range and light emitting Diode (LED)
based light sources at (a) 880 nm, (b) simultaneously at 880 and 370 nm and (c)
simultaneously at 370, 450, 571, 615, 660, 880 and 950 nm [161]. The aethal-
ometer has been used from experiments in the poles [162] to airborne experi-
172
ments [163] including measurements over the Siberian Sea. With high flow
rates, a 10 s temporal resolution is possible with an LOD of 10 ng/m3 [164]. EC is
by far the most important contributor to light absorption by atmospheric aero-
sols; OC, aerosols containing Pb or Fe contribute insignificantly. For these
reasons, the aethalometer is the most popular instrument used for BC measure-
ment [165]. However, to obtain EC concentrations in 'Lg/m 3, it is necessary to use
a value for the specific attenuation cross section. Most investigators simply use
the value supplied by the manufacturer [166]. Although this value is stated to be
10 m2/g, in practice this can vary by a factor of two either way depending on the
aerosol matrix, whether the aerosol is an internal or external mixture and the
aerosol age [167]. It is, of course, also dependent on the nature of the illumina-
tion source and the filter type [168]. Comparison between different aethalo-
meters using different wavelengths and known test aerosols composed of salt,
graphite, and iron oxide has shown that the aethalometer response depends not
only on the absolute amount of the absorbing species on the filter, but also on the
ratio of the absorbing to the nonabsorbing component [169]. Petzold et al. [170]
have also reported on the dependence of the specific attenuation cross section on
the mass fraction and particle size. The fact that the measurement is made on a
scattering matrix (a fiber filter) complicates the measurement. Kopp et al. [171]
carried out an investigation on the specific attenuation cross section of aerosols
deposited on fiber filters with a polar photometer. Interestingly, they found that
at a specific angle (reflected hemisphere 0 = 165°), the specific attenuation cross
section of EC is less dependent of the aerosol matrix than at any other angle.
Ballach et al. [172] have immersed sampled filters in a matching refractive index
oil to eliminate many of the problems from the filter matrix, and report good
agreement with the polar photometer/integrating sphere techniques and Mie
scattering calculations.
Other spectroscopic techniques have been developed that provide a more
direct measurement of BC. Direct photoacoustic spectroscopy using the 514.5
nm line from an Argon ion laser was first validated against a filter collection
technique by Adams et al. [173]. This method permits in situ measurement
without filter collection, but one still must assume a value of the specific
attenuation coefficient for BC. The instrument signal is also affected by humid-
ity and temperature variations and needs to be frequently recalibrated. These
authors subsequently applied the technique to the real-time measurement of BC
in Los Angeles [174]. This was one of the first methods without filter collection
to show a diurnal pattern in BC with low values in night with concentrations
increasing around sunrise and with persistent higher values through mid-
afternoon [175]. They found similar patterns in other urban areas as well as in
rural locations. They noted that visibility degradation due to aerosol light
absorption assumes a relatively greater importance under less polluted condi-
tions. For more recent work on photoacoustic sensing of BC, see Petzold and
Niessner [176] and Arnott et al. [177]. Niessner [178] also developed a photoelec-
tric aerosol sensor which can be used for BC measurements [179].
173
Other spectroscopic techniques, notably Raman and Fourier Transform
Infra Red (FTIR) spectroscopies, have both been used for the determination of
BC on filer substrates. Keller and Heintzenberg [180] used a diode pumped
Nd3+-YAG laser at 1064 nm in a backscattering mode with a defocused beam
diameter of -1 mm on a polycarbonate filter with an incident power of 20-270
mW to carry out Raman spectroscopy with a liquid nitrogen cooled Ge detector.
Pollard et al. [181] used infrared transmission in the 650-666 cm-' region to
measure BC on a PTFE filter by FTIR spectroscopy. One clear attraction of this
method is that it allows the simultaneous determination of ammonium, nitrate,
sulfate and several organic compounds as general classes (see Section 6.2.2).
Seasonal measurements of BC generally show lower values in the summer
and higher values in the winter [182,183]. In pristine mountain tops where
precipitation removes suspended matter, e.g., atop Mt. Jungfraujoch in
Switzerland, the opposite behavior is observed [184]. Pakkanen et al. [185]
report little seasonal variation in Helsinki, Finland, although they do observe
daily peaks around 8 a.m. except on weekends and holidays, when BC values are
much lower overall. Similar (weekend vs. weekday) behavior has been observed
in Saint-Denis, the largest urban site in la Reunion island in the Southern
Indian Ocean [186]. Here the BC levels are as high as those in any large
European city, and there is no seasonal dependence. It is also interesting to note
that secondary OC in the aerosol, which probably has a photochemical origin, is
produced a lot less in the wintertime [1871. BC and OC particle size distributions
in urban areas and over large water bodies have been determined [152]. In urban
areas, weekday morning peaks are typical and results can be greatly influenced
by local vehicular sources [188].
174
presumed to extract the OC fraction only. Extracted and unextracted filters are
combusted in oxygen to form CO2 (that is typically determined by micro-
coulometry, by a flame ionization detector (FID) after conversion to methane, or
non-dispersive infrared spectroscopy [191]) and thus respectively determine OC
and TC.
In the other, thermal/optical approach, there are many subtle variations. In
general, the sample loaded quartz fiber filter is heated while the transmittance
and/or the reflectance of the filter is monitored. In addition, the evolved gases
are oxidized to CO2, typically passing over an MnO2 bed at a high temperature.
The CO2 is then put through a methanator and measured by an FID. Finally, the
filter is heated at a high temperature in the presence of oxygen to oxidize the
remaining material, taken to be EC, to CO2. The classic paper by Turpin et al.
[192] contains all of these features and is recommended reading. Details of
different methods of this genre differ in: (1) the exact temperatures to which the
sample is subjected in the different steps, (2) the length of treatment at any
temperature, (3) the rate of temperature change during a thermal step, (4) the
composition of the atmosphere surrounding the sample at any step, (5) the
nature of optical monitoring of pyrolysis/volatilization/combustion, and (6) the
nature of calibration standards. Chow et al. [154] provides a most readable
description of the approaches available until that time and also specifically their
own instrument, which subdivides OC into several different subclasses. This
instrumental protocol became the basis of the Interagency Monitoring of
Protected Visual Environments (IMPROVE). More than 100,000 PM,, and PM 2 .5
samples have been analyzed with this protocol. It has been shown that if
allocation of subclasses arising at different temperatures are changed, the
IMPROVE protocol and the National Institute of Occupational Safety and
Health protocol produces basically the same values for EC and OC [193].
Simpler methods of EC/OC speciation have been advocated. Fung [194]
described a procedure in which the sampled filter is heated in intimate contact
with MnO2 in two steps. Carbon dioxide evolving at 525°C is taken to be due to
OC, while that subsequently evolving at 850°C is taken to be due to EC.
However, judging from the trend and given that the line of differentiation is
subjective, it does not appear likely that there will soon be general acceptance of
any EC/OC speciation approach that relies on such a simple step.
Numerous intercomparison studies have been performed [179, 195-197]: a
strong correlation between aethalometric BC measurements and thermal/ opti-
cal EC measurements has almost always been found.
Many practitioners of atmospheric analysis shy away from wet chemical analysis
because of the perception that gallons of reagents must be routinely made and
disposed of. The argument is routinely made that wet instruments are somehow
more complex [76] and/or less elegant. Most such feelings are archaic and do not
175
reflect the current practice of wet analysis. At best, the arguments are subjec-
tive: clearly, the beauty is in the eye of the beholder. Current state-of-the-art IC
systems require only water to operate and exhibit reliabilities similar to gas
phase instruments. Present continuous flow analysis systems can operate at low
enough liquid flow rates such that reagent replacement and waste disposal
issues have markedly changed compared to yesteryears. Doubtless, the present
author's view is biased. Nevertheless, it seems oxymoronic for Homo sapiens to
hold the view that a wet instrument is somehow less elegant. One outstanding
advantage that liquid phase analysis has over gas phase techniques is the facility
of preconcentration. This is particularly simple with ionic analytes where ion
exchange preconcentration is applicable. As a result, while the current state-
of-the-art thermal decomposition based sulfate measurement methods are taxed
to attain LODs of the order of a ~jg/m3, orders of magnitude lower LODs are
possible by wet analysis techniques with very comparable integration times
(vide infra).
176
In the very moist environment inside a PPWD, the maximum voltage that
could be applied before arcing occurred was 2.5 kV (corresponding to a field
strength of 8.3 kV/cm). Under these conditions, with -1 Avm mass median
aerodynamic diameter (MMAD) (NH4 )2 SO4 as the test aerosol, the highest
collection efficiency obtained was <50% (for a 40 cm x 36 mm active area on each
plate and a sample flow of 5 standard liters/min (SLPM). Subsequent work has
shown that smaller particles are collected with much better efficiency. Also,
having a dry region prior to the wet area where particles are charged by applying
voltage to the outer glass surface dramatically improves collection. There are
some intrinsic safety questions in such systems, however, and this approach was
not pursued further.
Liu and Dasgupta [202] subsequently described an aerosol collection inter-
face that is similar to a wire-in-tube electrostatic precipitator [198]. The
collection device consists of two concentric stainless steel tubes of 30x2.7x3.4
mm and 80x0.2x0.4 mm (length x o.d. x i.d.), functioning as the outer and the
center electrode, respectively. The tubes are fixed in position through polypropy-
lene tees (Fig. 6.4). The top of the center electrode extends through the tee and
connects to the preconcentration column of an IC system via a peristaltic pump.
ssrosal I
Fig. 6.4. Corona discharge based aerosol collection apparatus for coupling to an ion chroma-
tograph. Reproduced with permission from Ref. [202] (copyright 1996, Elsevier Science).
177
The exposed part of this metal tube is connected to the HV supply. The outer
electrode is electrically grounded. The bottom end of the center electrode pro-
trudes ca. 2 mm beyond the outer electrode end. This avoids arcing between the
ends of the electrodes. The aerosol sample is isokinetically aspirated into the
collection chamber and after it is collected for a period of time (typ. 20 min), HV
is turned off. Deionized water is pumped into the annular space and aspirated
back through the center tube. The washings are preconcentrated on the IC
system. To ensure that all of the input liquid wash solution is aspirated back, the
aspiration flow rate is set at nearly twice the input flow rate. Typically, -2 ml is
sufficient for complete washing. Before aerosol sampling is begun again, the
system is cleaned and dried by admitting compressed N2 (8 psi) into the system.
A stable corona discharge could not be generated if the center electrode was
made significantly larger to improve rigidity. On the other hand, too small a cen-
ter electrode made it difficult to be fixed rigidly in position, Generation of corona
discharge inside the collection chamber results in charging of the particles. In
industrial precipitators, negative corona discharge is almost exclusively used. In
the Liu-Dasgupta arrangement, it was easy to obtain a stable positive corona in
the current range of -6 to -70 AMA but a stable negative corona was elusive. A
voltage of -2.7 kV was minimally required to generate a corona. This threshold
voltage varied with experimental parameters such as aerosol particle composi-
tion and concentration, gas phase composition, pressure, temperature, sampling
flow rate, etc. No corona discharge was ever observed below 2.5 kV.
Before the presence of any corona discharge, the collection efficiency was
low. But as the discharge voltage is approached, the collection efficiency inc-
reased linearly with increasing voltage. Aerosol collection occurs in this regime
because even in the absence of charging by a corona, there is equilibrium
Boltzmann charge on the particles [200]. Similar to the situation with an
Electrostatic Aerosol Analyzer [203], more particles deposit on the collection
electrode as a higher voltage is applied. After corona discharge initially starts at
-2.5 kV, the current increased very quickly with the voltage up to 2.8 kV,
reaching a stable current value of -70 MA. The collection efficiency, like the
corona current, increased exponentially with the applied voltage in this range.
The collection efficiency increased essentially linearly with the corona current
up to -20 AMA at which point collection efficiency was quantitative (0.5 SLPM)
for -1.15 and 2.14 d/m MMAD Na 2SO4 particles. Collection efficiency decreased
with increasing sampling rate. Similar results were obtained for 0.5 /xm
particles.
The major problem with the above system was that, much as happens during
an electrical thunderstorm [204], NO, was generated in corona discharge. This
showed up as artifact nitrate (5-10% was present as nitrite). The relationship
between artifact nitrate and corona current was linear, -2 ng//iA of nitrate was
produced. There is little doubt that the nitrate originates from the reaction of
excited N and O atoms in the corona followed by a reaction with water vapor.
Substitution of cylinder N2 for air as the carrier gas virtually eliminated nitrate
178
formation. Nevertheless, the proposed arrangement will only be applicable
where nitrate and nitrite in the aerosol are not of interest.
In a subsequent paper [205], the same authors describe a system that does
not use corona discharge. It uses only field-charging and NOx production is
eliminated or reduced to very low values. The authors also provide a detailed
theoretical description of the field-charging method and show that the experi-
mental results agree closely with theoretical expectations. Basically, an arrange-
ment identical to Fig. 6.4, except for different collector dimensions (150x2.4x
3.0 mm (length x i.d. x o.d.) tube inserted inside a 100 x2.4x3.0 mm tube), was
used. Compared to the previously used dimensions, the annular gap is smaller
but the residence time is nearly five times longer at the same flow rate. At a
voltage of 2 kV applied to the center electrode, the field strength is -28 kV/cm
and corona discharge is not observed. At a sampling rate of 300 ml/min, the
collection efficiency is 85% and is nearly independent of the particle size in the
0.86-2.14 Am range. The collection efficiency decreases with increasing flow rate
and is -60% at 1 /min. The system was applied to ambient sulfate and nitrate
measurement using integrally coupled ion chromatography. The amount of
artifact nitrate formed was well below an equivalent ambient concentration of
0.1 gug/m 3 .
6.3.1.2 Impaction
Impaction on a solid surface is the process most commonly used for particle
collection. The same approach can be utilized for continuous transfer of particles
into a liquid surface. If the liquid constitutes a moving film, then the liquid
containing dissolved and undissolved particles can be directly sent to an analysis
system. Quantitative collection of small particles may necessitate serial,
multiple impaction, which can be achieved by multiple and frequent changes in
flow direction. One way of maintaining a thin liquid film under such a condition
is to introduce the liquid at the inlet along with the aerosol to be sampled. Also, if
the liquid enters the impactor through a narrow orifice, it will itself be aerosol-
ized. For an instrument of such a design, in addition to the impaction of particles
on to the wetted walls, there will be particle growth through particle coagulation
and thus increase the particle collection efficiency of the impaction system. On
the other hand, if the fluid velocity at the impactor exit is reduced by making the
cross-sectional area large, the liquid containing the aerosol particles should
come out as larger droplets, which can be separated from gas phase by an inertial
separator and sent to an analysis system.
This principle was tested with a maze type impactor as represented in Fig.
6.5. As shown, a perpendicular maze was cut into the bottom section; the top
plate is a flat piece of same length and width as the bottom plate. When
assembled, it forms a sealed enclosure. Water (0.5-1 ml/min) was introduced at
the inlet along with the sample aerosol (5 1/min). At the exit, liquid droplets were
collected with an inertial air/ liquid separator represented in Fig. 6.6. Collection
efficiency for 2.2 m MMAD Na2 SO4 aerosol was essentially quantitative but this
179
^.
..... o nn - Outlet
- m 38mm
mm
Inlet 3 mm
m Air/Liquid
Inlet
3 mnmi mm
69 rrn I
The cioerel is 5 amdeep LiquidOutlet
dropped to < 70% for 0.7 Avm MMAD aerosol. Further experiments clearly estab-
lished that such an arrangement could not collect submicron particles with
acceptable efficiency. Similar results were obtained with an impactor consisting
of a flattened glass tube (elliptical cross section) with 22 U-turns. Collection of
large particles is efficient but submicron particles are difficult to capture. Of
course, this is known to be the case for impaction in general. The conclusion
drawn from these experiments was that if small particles could somehow be
increased in size prior to impaction, without altering their chemical composi-
tion, then this would be an attractive approach.
Nevertheless, direct impaction into a liquid has its attractions. Karlsson et
al. [206] reported a novel single stage flowing liquid film impactor for continuous
on-line particle composition measurement. A continuous liquid flow is provided
to a support plate made of etched glass that constitutes the impaction area.
Continuous transport of the impaction liquid makes on-line analysis by an
integrally coupled IC possible. With multiple nozzles (74 holes, 0.3 mm i.d.) and
an air sampling rate flow of 10 l/min, the 50% cut-off was 0.41 ± 0.02 Alm. The
design is shown in Fig. 6.7. The actual performance in terms of particle size cut
off is shown in Fig. 6.8. While this is an interesting design, for the collection and
analysis of ambient aerosols, a significant mass of particles exist below the cutoff
size of this impactor, and it is not practical for that purpose.
6.3.1.3 Filtration
Filtration is one of the most common methods of aerosol sampling. Although the
general principles are well known, there are still significant gaps between theory
and experiment. A common misconception is that aerosol filters work like
macroscopic sieves through which only particles smaller than the holes can pass.
This is not how aerosol filtration works [207]. All of the following mechanisms
lead to deposition of aerosol particles on or within a filter: interception, inertial
180
G
C X
D
Fig. 6.7. Assembly drawing of the wet impactor. The supporting base A is equipped with three
screws B for vertical and horizontal adjustment of the impaction liquid support plate C. Also
shown are the impaction liquid drain spout D, inlet fitting for impaction liquid E, nozzle plate F
and the impaction liquid film G. Liquid I/O tubing H were connected through bulkhead fittings.
Reproduced with permission from Ref. [206] (copyright 1997, Elsevier Science).
e
90-
a'
a'
.2 60 - A
40 - __C]-_
C
--a --
#2
U
e 20
a'
10
0 I I
0 0.2 0.4 0.6 0.8 1
Dp/pm
Fig. 6.8. Impactor performance (Karlsson et al.). Reproduced with permission from Ref. [206]
(copyright 1997, Elsevier Science).
181
Entrance length
Water in
Denuder I
Wet Effluent
Denuder
3as/Liquid
ri
L
i
r
Separator ? Water in
Frit
Pyrex Frit E
Fig. 6.9. Single tube wet denuder-wetted glass frit filter for
simultaneous gas and aerosol sampling. Reproduced with permission
Gas/Liquid
from Ref. [210] (copyright 1995, Elsevier Science). Separator
Fig. 6.10. Analytical system schematic for gas and aerosol analysis according to Buhr et al. [208].
Note provisions for liquid phase calibration and measurement of blank. Reproduced with
permission from Ref. [208] (copyright 1995, Elsevier Science).
particles, an average retention efficiency of 79 ± 8% was found for the glass frit.
Field measurements and comparisons with 1-h filter pack (FP) samples were
consistent with this: the data showed close agreement for sulfate in terms of
temporal variations, with the frit system measuring 12% lower sulfate on the
average. The data for aerosol nitrate as measured by the frit system was not so
straightforward. The measured nitrate value was much higher than that
182
measured by the FP and showed a diurnal variation not exhibited by the FP. The
artifact nitrate generated appears not to be related to NO2 or NOx but to some
photochemically generated intermediate that produces nitrate at the wet frit.
This general issue of intermediates that react with water to produce some key
nitrogen species appears in many places and has not been resolved. There is
accumulating evidence that direct liquid contact based instrumentation (e.g.,
wet denuders) tend to produce significantly higher values for HONO than
collocated direct spectroscopic instruments [209]. It is vital to resolve these
differences and get a better understanding of the responsible species before an
accurate picture of nitrogen deposition in ecosystems can be obtained.
An aspect of wet effluent systems that are coupled to solid phase pre-
concentrators is the necessity to aspirate all of the effluent liquid so that liquid
does not enter the sample inlet region. This is difficult to do without aspirating
bubbles in the process. A very interesting finding of Buhr et al. was that solid
phase preconcentration columns exhibit greater analyte capture efficiency for a
segmented flow stream (i.e., one containing air bubbles), rather than an
unsegmented one.
Samanta et al. [210] described a glass-fiber filter-based instrument for the
automated measurement of aerosol Cr(VI). Hexavalent Cr is a notorious inhala-
tion carcinogen that is produced in the aerosol form in several workplace situa-
tions. They solved the automation problem by using a system that alternately
collects the sample on one of two glass-fiber filters. After 15 min sample
collection on one filter, the sampling switches to the second filter. The freshly
sampled filter is washed for 8.5 min and the washings are preconcentrated on a
minicolumn packed with anion exchange resin. The washed filter is dried with
filtered hot air for the next 6.5 min so that it is ready for sampling at the end of
the 15 min cycle. The preconcentrated Cr(VI) on the column is eluted with 0.1 M
sodium perchlorate and then reacted with sym-1, 5 diphenylcarbazide prior to
absorbance detection with a light emitting diode (LED)-based dedicated flow-
through absorbance detector. The detection limit (S/N = 3) is 5 ng Cr(VI)/m 3,
orders of magnitude lower than current regulatory requirements. The instru-
ment operates unattended over long periods. In continuous round-the-clock
operation, filter replacement frequency is every 24-72 hours depending on dust
loading. The system is portable for facile field deployment and permits fully
automated rapid determinations at a very low analysis cost per sample.
183
Drying Air Drying Air
Sampling
Air
(b)
FA
FA FB
ON OFF
Fig. 6.11. (a) Total system airflow schematic. FA, FB: Glass fiber filters, T: trap bottles,
MFC-A,B,C,D: Mass flow controllers, C; cyclone, FC: 47 mm filter for MS analysis, P1,2: air
sampling pumps, PP: peristaltic pump, F: filter, P: purifer, H: heater. The dotted section
including the denuder is on the roof and the air pumps are either below the instrument shelter or
in a modified doghouse with forced air ventilation. Vi: Aerosol switching valve, shown in detail
in (b).
184
(adjusted to pH 7) containing 0.5 mM H2 02 , on each plate. This solution serves to
remove both acidic and basic gases; the denuder effluent (aspirated at -1.5
ml/min per plate) is sent to waste. The gaseous effluent from the denuder
bearing the aerosol proceeds to the PCS.
The first element of the PCS is a specially constructed rotary valve V1 that
directs the ambient air stream to either filter A or Filter B. This valve must
provide a straight passageway for the sample stream to one of the two sample
filters without aerosol loss. The valve is shown in functional detail in Fig. 6.11lb.
The stator plate has three holes, the central port is connected to the sample air
stream (from the PPWD) while the two other ports are connected in common
through a Y-connector to a sequential trap containing a particle filter (F), acid-
washed silica gel (T1, this removes NH3 ) followed by a soda-lime trap (T2, this
removes acid gases) and a heater (H) that thus provides a hot dry clean air
source. The rotor plate has two holes, connected to Filter A (FA) and Filter B
(FB), respectively, and is rotated by a spring-return rotary solenoid. With the
solenoid un-energized, ambient air is sampled on Filter A and with the solenoid
energized, ambient air is sampled on Filter B; flow is thus switched without
aerosol loss. Other air valves (V2-V4) govern airflow in the PCS.
Purpose-machined holders hold 25 mm diameter filters. A glass fiber filter
with a paper filter underneath rest on a stainless screen; ports on opposite sides
of the top half of the filter holder provide entry of wash liquids. The bottom half
of the filter holder is designed as a shallow cone with the air outlet at the center.
The liquid exit port is located equidistant from the inlet apertures such that the
inlet/outlet apertures constitute an equilateral triangle in top view. Air/liquid
separators are incorporated below each filter holder in the air exit path. These
contain air in/out ports, as well as a port to remove accumulated water (period-
ically, e.g., every 24 h) using a syringe. These separators serve to keep any wash
liquid from entering the respective mass flow controllers (MFC-A, B). The
diaphragm pump (P2, same as P1) used for sampling is capable of aspirating at
>8 1/min through each filter holder simultaneously. Perfluoroalkoxy (PFA)
Teflon tubes, used for connecting PCS components upstream of the filter
holders, are externally wrapped with electrically grounded Al tape and then with
bare Cu wire. This served the dual purpose of improving its structural strength
and reducing electrostatically induced aerosol loss.
A six-channel peristaltic pump provides liquid pumping. In Fig. 6.12, valves
V5-V8 are liquid-handling solenoid valves. Valves V5 and V6 direct the flow of
deionized water to the filter holders. Valves V7 and V8 (pinch type valves)
handle filter extract in which stray glass fibers may be present. A low volume
fiber-trap-filter (FTF) placed prior to the injection valve prevents glass fiber
intrusion to the preconcentration columns. Injection valve IV contains two
low-pressure drop anion preconcentration columns PCi and PC2 that alternate
in function.
Table 6.1 shows the air and liquid valves and their respective on/off status.
Figures 6.12a and b illustrate the four states of the instrument cycle. The first
185
(a) (b) 1FA , Liquid
Phase
PPWD
Na2HPo X X
H202
Ambien
Air In Air In V2 Gas
TI T i lPPWD Phase
F
FA
Ff FA --p Ambient
Air In
MFC-I MFC-A
V3) (VS
vP
P1 1-
Fig. 6.12. PCS set up. (a) State 1, (b) State 2. V5,6: All-PTFE solenoid valves, V7,8: Pinch-type
solenoid valves, FTF: fiber trap filter, IV: 10-port injection valve bearing preconcentration
columns PC1 and PC2, IC2: PCS Ion Chromatograph, V2-4: on/off solenoid valve for air flow, T1:
acid washed silica gel, T2: soda-lime (T1 and T2 together constitute the purifier P in Fig. 6.2),
PPWD: PCS gas denuder, WB: water bottle, W: waste. The top and bottom portions respectively
indicate liquid and gas flow.
state (Fig. 6.12a) is 8.5 min in duration. In the particle collection system, the
soluble gas denuded aerosol flow stream is directed to Filter A by valve V1. Air
passes through Filter A though MFC-A, which regulates the airflow to 5 SLPM,
and finally through valve V4, which is on during state 1. Valves V2 and V3 are
off, and filter holder B (FB) is under airlock.
TABLE 6.1
Four states of the instrument
Status IC using V1 V2 V3 V4 V5 V6 V7 V8
Begin sampling on FA, start washing PC2 OFF OFF OFF ON OFF OFF OFF OFF
FB, water being aspirated through FB
and being loaded on to PC1
186
In the liquid extraction portion of the instrument, deionized water is
contained in a bottle WB. The air entrance to the water bottle is equipped with a
soda-lime trap to minimize acid gas intrusion into the bottle. Water from WB is
aspirated and then pumped at 1 ml/min by the peristaltic pump PP through a
mixed bed ion exchange column MB1 to remove any trace impurities present in
the deionized water. Valve V5 directs flow to valve V6, which in turn directs the
water to filter FB. The water enters FB through the two ports in the top of the
holder and is simultaneously aspirated from the bottom of FB through valves V7
and V8 by the peristaltic pump. Since FB is under airlock, water does not enter
the air outlet tubing at the bottom of the filter holder. The extracted material
from the filter is pumped through the fiber trap filter FTF to remove glass fibers
from the flow stream before passing to the appropriate preconcentration
column. Valve IV is configured such that while one preconcentration column is
chromatographed, the other preconcentration column is loaded with sample or
washed with water. In the present case, preconcentration column PC1 is loaded
with sample. After 8.5 min, state 2 begins (Fig. 6.12b).
During state 2 in the PCS, ambient air continues to be sampled on FA, just as
in state 1. Valves V2 and V3 are activated in state 2, allowing clean hot air to pass
through Filter FB for the duration of this state. Clean (ammonia/acid gas- and
particle-free) air, produced by passing ambient air through F, T1 and T2, is
heated to -75°C by passing it over a heater switched in parallel with valve V2.
This clean, hot air is aspirated through the previously extracted filter FB, to dry
it prior to state 3.
Note that at the time the instrument enters state 2 from state 1, although all
the analyte has been extracted from filter FB and preconcentrated, the last
portion of the wash water is still contained in the filter housing. This water is
aspirated into the trap bottle ahead of MFC-B. Water that enters into the trap
bottle is generally of the order of 1 ml/cycle. This volume may be used to monitor
the filter extraction process; excessive water accumulation in the water trap
bottle indicates flow problems through the filter or through the relevant pre-
concentration column.
In the liquid extraction system, valves V5 and V8 are activated. Valve V5 now
directs water used to wash Filter FB in state 1, back into the water bottle. This
recycling procedure helps maintain the purity of the water in WB. As a result of
liquid being aspirated faster from the filter housing than it is pumped in, air
bubbles inevitably enter into the preconcentration column. To remove the air
bubbles before the sample is injected, valve V8 is activated and water is aspirated
by the pump through a mixed bed ion exchange column MB2 through V8 and
pumped through the preconcentration column PC1. The duration of state 2 is
6.5 min.
After state 2 ends, state 3 (8.5 min) and state 4 (6.5 min) follows. States 3 and
4 are identical to states 1 and 2, respectively, except that the roles of filters A and
B are interchanged relative to those in states 1 and 2. States 1-4 constitute an
instrument cycle and the instrument starts anew with state 1 at the end of state
187
Ir - _;Z . -;
..t' , ij
- :I: r
Nit rous acid
A Nitrit
10
45-
Sulfur Dioxide
A Sulfate
a 30
15-
0-
8116f99 8118199 8120199
Fig. 6.13. HNO 3 /Nitrate, HONO/Nitrite and SO2 /Sulfate patterns at a Midtown location in
Atlanta, GA. Note nocturnal maxima in the middle panel and opposite behavior in others.
188
2.5-
Oxalic acid
A Oxalate
2.0-
1.0 - Q.
ii
0.5 -
'"-"'
.
VIANSWUNL IDA
0.0T- I 7- - IWIIE7 -
8119/00 8121100 8123100 8125100 8/31/00 9/2/00 9/4100 916/00
Fig. 6.14. HC1/Chloride, Oxalic acid/Oxalate levels at a heavily industrialized site close to the
shipping channel in Houston, TX. This site is known to be close to one of the largest chlorine
emission sources in the region.
8.00
6.00
.P
4.00
g 2.00
0.00
void volume corrected retention times (tR) under isocratic elution conditions at
three or more different eluent concentrations. Under these conditions, it is well
known that the slope of a log t vs. log [eluent] plot is equal to the ratio of the
charge on the analyte ion to that on the eluent ion (unity for hydroxide) [212];
189
I
,
9
Fig. 6.16. Gradient ion chromatogram of an aerosol collected during the Atlanta experiment.
Readily identifiable peaks: (1) fluoride, formate, acetate, (2) chloride, (3) nitrite, (4) carbonate,
(5) nitrate, (6) sulfate, (7) oxalate.
-1.00- I ' I I
1.10 1.20 1.30 1.40 1.50
log Hydroxide Eluent Concentration, mM]
Fig. 6.17. Log t vs log [eluent] plots reveal charge on analytes, aiding search for a confirmatory
standard.
this is shown in Fig. 6.17. While MS is ultimately the only completely unambigu-
ous means of identification when confirmed by a matching standard, IC informa-
tion above and the nature of UV response of the analyte often make possible to
determine the identity of the analyte.
Table 6.2 lists average daytime and nighttime aerosol composition for a rela-
tively polluted period during the Atlanta measurement campaign. The analysis
was conducted by IC-CD-UV-MS, with identification confirmed by MS and con-
190
TABLE 6.2
Average anion composition of day and night-time aerosol in Midtown Atlanta, August, 1999a
Retention time (min) Concentration (g/m 3 )
Conductivity detector UV detector Analyte Samples (day) Samples (night)
8.34 Fluoride 1.1 0.58
8.95 Glycolate 0.28 0.19
9.37 Acetate 0.58 0.25
9.56 Lactate 0.81 0.32
9.83 Formate 0.91 0.71
10.96 a-Hydroxyisobutyrate 0.02 0.03
11.23 Unknown [0.015] [0.02]
11.87 Methanesulfonate 0.05 0.04
13.04 Chloride 9.8 5.5
13.27 Pyruvate tr tr
14.93 Unknown [0.004] [0.01]
15.52 Nitrite 0.11 0.15
15.60 Carbonate nd nd
16.23 Malate 0.30 0.24
16.57 Malonate 0.36 0.26
17.23 Sulfate 16 11
18.13 18.34 Oxalate 0.34 0.27
20.46 Unknown [0.01] [0.02]
21.58 Phosphate 0.03 0.03
23.28 23.52 Nitrate 1.9 1.7
24.33 24.66 Unknown [0.02] [0.03]
24.87 Unknown [0.03] [0.03]
25.87 26.06 Unknown [0.004] nd
26.72 Unknown [0.003] [0.007]
28.50 28.83 o-Phthalate tr tr
29.10 Unknown [0.004] [0.072]
aRetention times using a hydroxide gradient chromatographic protocol. Numbers in parentheses
provided for unknown peaks are calculated from the conductivity peak areas based on the
average response. These likely the lower limits. tr: trace.
191
Air suction
Hydrophobic
i Bonnet
Water out
Fig. 6.18. Particle collector, incorporating integral cyclone, _I Water in
liquid spray and hydrophobic refluxing filter.
system [213]. Perhaps many of the low molecular organic analytes we find in
ambient aerosol are biogenic.
192
into larger particles without a change in their composition, then these methods
become applicable. Fortunately, fine particles are excellent nucleation sites for
condensation. As early as 1888, Aitken [219] used vapor condensation to make
very small particles grow so that they are large enough to be optically counted.
Even today, available condensation nuclei counters rely on this principle [220].
Typically, the aerosol is equilibrated with a liquid such as 2-propanol or water.
The vapor saturated aerosol is cooled, often by adiabatic expansion, to induce
supersaturation.
Water is more compatible with most analysis systems than 2-propanol. One
can deliberately introduce steam to the sample aerosol and then cool the
mixture. Cooling causes a large supersaturation, causing the steam to condense
on the available particles to form liquid droplets. These droplets have much
larger size than the original aerosol present and can be collected by various
means, including impaction. A flow-through chamber to provide sufficient
residence time allows the growth to occur. If the impactor itself is cooled,
collection on the cold impactor surface is further aided by thermophoresis
(aerosol particles are attracted to a cold surface) [221].
It is instructive to compute the degree to which such growth occurs. If we
assume an idealized case of a monodisperse aerosol of density 2.0 g/cm3 , a
sampling rate of 10 /min, an aerosol concentration of 100 tg/m 3 (this is much
higher than typical ambient fine particle concentration), a steam injection rate
of 500 mg/min, and homogeneous growth, it is readily computed that a particle
will grow in the radial dimension by 100-fold (final particle density is taken to be
the same as that of liquid water), i.e., a particle with an original diameter of 10
nm will grow to 1 Am. Simple considerations will also indicate that the factor by
which each particle grows increases with higher steam injection rates, lower
sampling rates, and lower aerosol concentrations. The process of particle growth
is vital to impaction-based collection processes and the importance of growth in
size due to vapor condensation cannot be over-emphasized. The mixing chamber
thus serves the function of a cloud chamber.
Blatter et al. [222] and Simon and Dasgupta [223,224] first described a particle
collection and analysis system based on this principle; their arrangement is shown
in Fig. 6.19. The soluble gas-denuded aerosol stream (5 1/min) proceeds to the mix-
ing chamber where steam is introduced at a rate of 800-900 mg/min. Then it
enters a thermoelectrically cooled maze (see Fig. 6.5), which largely functions as a
rain chamber. Agglomeration and impaction occurs, and the process of vapor con-
densation is completed by cooling. Note that the moisture content of the sample
stream exiting the wet denuder is high and cooling to 2°C can add almost another
- 100 mg/min of liquid water to the system effluent. An inertial gas-liquid separa-
tor is added at the exit. The liquid is collected from here and pumped to the analy-
sis system. The overall collection efficiency for the system is shown in Table 6.3,
for all practical purposes, the collection is quantitative.
A number of other researchers have also used steam condensation prior to
aerosol growth. The apparatus of Khlystov et al. [225] developed at the Nether-
193
Fig. 6.19. Vapor condensation particle collection system. Reproduced with permission from Ref.
[223] (copyright 1995, American Chemical Society).
lands Energy Research center (ECN) is shown in Fig. 6.20; the device has been
christened the Steam Jet Aerosol Collector (SJAC) by its developers. In this
arrangement, the soluble gas denuded sample air mixes with steam in the
mixing reservoir. Then two sequential glass cyclones are used to collect the
condensed droplets bearing the aerosol. The liquid effluent from the cyclones are
combined and sent to the analysis system. The gaseous effluent is sent through a
cooler to remove residual moisture prior to passage through flow controllers.
TABLE 6.3
Efficiency of aerosol collection in the Simon-Dasgupta [225] vapor condensation-PCS
Composition of aerosol MMAD (m) Concentration Percent collected by
O(g/m 3) system
Sulfuric acid 0.5475 + 0.0034 1.63 96.97a
Sulfuric acid 0.7139 + 0.0068 9.85 - 2.06 99.95±0.06
Ammonium sulfate 0.7175 + 0.0054 6.56 + 0.17 99.23+1.08
Ammonium sulfate 1.0057 + 0.0065 27.96 ± 0.53 99.21±0.14
Ammonium nitrate 0.6134 - .0065 6.13 ± 0.36 99.88-+0.04
Ammonium nitrate 0.840 ± 0.01 13.27 + 0.51 99.31±0.66
Sodium nitrate 0.6038 ± 0.0097 72.11 + 3.5 100.27 ± 0.52
Sodium nitrate 0.86 b 307.4 + 2.2 99.46 + 0.09
Sodium sulfate 0.4012 ± 0.0021 64.46 + 1.4 99.61 + 0.10
Sodium sulfate 0.67 b 302.4 + 3.6 99.44 ± 0.31
aUncertainty was not determined. At these low levels, the finite filter blank for the post-system
filter is the source of a significant negative bias for the computation of collection efficiency.
Reproduced with permission from Ref. [2251 (copyright 1995, American Chemical Society).
194
steam
In condenser
airin air out
mixin
reservoir /
Fig. 6.20. ECN Steam Jet Aerosol Collector cyclnes condensed water out
(SJAC). Reproduced with permission from Ref. sampled aerosol solution
[225] (copyright 1995, Elsevier Science). out (for analysis)
The ECN apparatus can sample a much larger volume of air, typically 20-30
1/min, compared to the 5-10 /min flow rates used in the apparatus described in
the foregoing section. When the atmosphere is supersaturated with water vapor,
the growth of the particles is very rapid. Figure 6.21 shows the rapid growth of
100 nm MMAD particles in a supersaturated atmosphere. Figure 6.22 shows the
effect of the extent of steam injection on the particle size distribution.
More recently, the ECN group has elaborated [226] on the refinements made
to their original instrument over the 6 years since its original introduction.
Current practice includes the use of a 10 /M sodium carbonate solution in the
denuder, which is said to effectively scrub out all gases of interest, including SO 2,
HONO, HNO3 , and NH3 . Air sampling is conducted with a pump equipped with a
critical orifice. After the introduction of steam (2.5 g/min), all particles over 10
nm initially in size grow to at least 2 CLm. More than 99.9% of these particles are
collected in the system. Indeed, even at a flow rate of 60 1/min, the collection
efficiency is >99%. The system contains liquid mass flow controllers on all
critical liquid flow lines for quality assurance. The combined sample from the
cyclones is split into two streams. To one stream, NaOH is added and the mixed
stream passes across a porous PTFE membrane based gas diffusion device [227].
Under the test conditions, -30% of the ammonia transfers to the receptor side,
which is pure water. This is measured by conductometry. This principle for
ammonia measurement was originally described by Carlson [228] and adapted
by the ECN group [229,230]; it is commercially available as well [231]. If 50-60
/g/l ammonium is added to the NaOH, the calibration curve is linear. The
Dry
-J
zi I---- 20%
195
100000
100001
a 100
10
196
80000-
60000-
x 0
* I 10
40000-
*·
20000-
·
, ,*,0,t
, t I , , ,l
0-
I . . . I II ' ' ' I' 7 -I
10.00 100.00 1000.00
d,(nm)
Fig. 6.23. Efficiency of the GAMS [233] to remove (NH4 )2 SO4 aerosols. Data courtesy of Maria
Loflund, Vienna University of Technology, 2001. Solid and open circles respectively represent
quantitative size distribution of the aerosol before and after the GAMS.
an
- ---
10000-
1000-,
100-
A A
A
10-
A oA A A
A Primary size distribution
O Size distribution after denuder
A4 Size distribution after GAMS
AAAAAAAAAAAAA
0-
Fig. 6.24. Size distribution of a non hygroscopic test aerosol, bis(2-ethyl-hexyl)sebacate before
and after the wet denuder and after the GAMS [233]. Data courtesy of Maria Loflund, Vienna
University of Technology, 2001.
197
TABLE 6.4
Analytical figures of merit for GAMSa
Species. LOD (ng) Limit of determination Repeatability
Aerosol (g/m 3) CV (%)
Formate 3.6 0.02 7
Acetate 2.6 0.09 6
Propionate 1.4 0.04 7
Butyrate 10 0.07 9
Valerate 13 0.09 9
Pyruvate 5.6 0.04 5
Flouride 5.0 0.03 9
Chloride 0.45 0.03 8
Nitrite 3.2 0.05 9
Nitrate 1.1 0.07 5
Sulphate 5.3 0.08 4
Phosphate 1.4 0.01 11
Oxalate 0.75 0.02 1
Malonateb 3.8 0.03 12
Succinate 7.8 0.05 10
Phthalate 10 0.07 2
Citrate 0.75 0.01 13
Maleate 2.0 0.20 3
Sodium 29 0.19 3
Ammonium 26 0.17 12
Potassium 4.8 0.05 5
Calcium 120 0.79 9
Magnesium 28 0.33 4
aLimits of determination are calculated for 150 1 air samples, 30 min samples at 5 I/min.
bIs dependent on carbonate concentrations.
Reproduced with permission from Ref. [233] (copyright 2001, Elsevier Science).
aerosol collection and analysis. This instrument also uses steam introduction for
particle growth followed by impaction and dual IC analysis for rapid measure-
ment of key species of interest, most notably nitrate, sulfate and ammonium.
Conventional dry denuders, replaced periodically, are used with this instrument.
In this instrument, the particle growth device is modeled after a design by
Okuyama [235]. A single jet inertial impactor is used to collect the droplets onto
a vertical glass plate that is continually washed with a constant water flow. The
device is shown in Fig. 6.25. The exit liquid flow is divided and then analyzed by a
dual channel ion chromatograph. In this initial implementation, integrated
samples were measured every 7 min. The instrument provides detection limits of
approximately 100 ng/m3 for chloride, nitrate, sulfate, sodium, ammonium,
calcium and potassium. The authors also experimented with butanol as a solvent
for vapor condensation on the aerosol using a commercial condensation particle
198
5Umi.n Vacuum
Pump
Mixing Cooling Glass
Coils from Impactor 0.ImLJmin.
Chamber H20
Water Jacket \
Cartridge
Heater
\ from .. S GrowthChamber
Sample
A~ Drai~/ YTo IC
H20 from
Penstaltic Am-Ienr
Pump Aerosol
Fig. 6.25. Vapor condensation-impaction apparatus of Weber et al. Reproduced with permission
from Ref. [234] (copyright 2001, American Association of Aerosol Research).
3a5 a: I-MM
S, Eastern Standard Time
0_
f00:00 2 exO
0:00 22. :c00:00 24:t02 0 2$:0 :000 26o : 2t e
Day: HH: MM: SS, Eastem Standard Time
Fig. 6.26. Intercomparison of the SJAC [226] and the GT/BNL [234] nitrate data from the
Atlanta Supersite 1999 experiment. Reproduced with permission from Ref. [226] (copyright
2001, Elsevier Science).
199
+ SJAC ,t Filterpack O Thermoden.
25
E 20
0)
E
a_ 15
E
.g 10
0
.> 5
S
0 5 10 15 20 25
"3)
Average of 3 methods (pglm
Numerous examples of ambient air measurement in outdoor ambient air for the
measurement of aerosol ionic composition have already been given. Here we look
at various other applications.
200
6.3.2.1 Indoor air quality: kerosene-fueled space heaters
In most societies, people are spending more and more time indoors, especially in
the winter. Assessment of total human exposure to pollutants obviously requires
measurement of indoor pollutants, which can be qualitatively and quantitatively
different from those outdoors. Residential indoor environments are significant
contributors to human exposure to airborne pollutants, including inhalable
particulate matter. Residential wood combustion [239] and environmental
tobacco smoke [240] have long been known to be major contributors to indoor
aerosols; it has more recently been shown that improper use of ultrasonic
humidifiers can lead to significant contributions to inhalable particulate matter
concentrations indoors [241]. Meat cooking has been shown to contribute to
organic aerosols [242].
Unvented kerosene space heaters have become a popular way to provide
localized supplementary domestic heating. The annual sale of such heaters
approaches a million units [243]. These are a potential indoor source of fine
particles, especially sulfate and acid aerosols, in many homes. Sulfate and acidic
aerosol levels in such homes can substantially exceed peak outdoor concentra-
tions [244]. Kerosene heaters can also release various organic products including
some mutagenic compounds [245,246].
The steam condensation-PCS instrument was utilized to measure the partic-
ulate anions and acid gases emitted by a kerosene-fueled space heater in a con-
trolled laboratory study; the results are shown in Fig. 6.28 [224]. To determine
worst-case conditions, air ventilation systems were deliberately shut off. Two
different heaters, rated at 9000 and 89000 BTU/h, were sequentially tested. Very
substantial concentrations of gaseous SO2 and HONO, particulate nitrite and
slightly lower concentrations of particulate sulfate were measured. Note that
the formation of particulate nitrite lags behind that of gaseous HONO and
HONO concentration decreases at a rate much slower than does SO2 after the
heaters are turned off. Further, the ratio of particulate nitrite over gaseous
201
HONO was lower while the heater was lighted than the ratio observed an hour
after the heaters were switched off. It is likely that the primary emission consists
of HONO, NO, NO2 and SO 2 along with other gaseous and particulate species.
Particulate nitrite in these cases does not appear to be an artifact. What is being
measured as particulate nitrite can have two sources: the generation of nitrite
on the particle surface by the reaction of NOx with moisture, with or without
co-adsorbed SO 2 , and the adsorption of HONO already formed on to the parti-
cles. Both of these processes can cause the particulate nitrite concentration to
lag behind that of HONO. Particulate nitrite concentration predictably in-
creases with time. Production of indoor HONO by reaction of NO2 with surface
has been suggested earlier to account for the higher HONO concentrations
observed in homes that do not have unvented combustion appliances [247]. Sim-
ilar processes have been suggested to be responsible for at least part of outdoor
HONO [248,249].
Once the nitrite concentration on particles is sufficiently high, there will be a
net desorption of HONO after the heaters are turned off and primary emission of
HONO ceases. This likely explains the change in HONO to nitrite ratio as well as
the slower fall of HONO concentration after the heaters are turned off. When
significant amounts of NO2 are present, the reaction between sulfite and NO is
more important in the generation of nitrite [250,251]. It is clear that very
significant amounts of SO2 are present in kerosene heater emissions. While no
significant amounts of particulate S(IV) were seen, it would have been rapidly
oxidized to sulfate in the alkaline elution conditions in the IC (the IC eluent was
not deoxygenated). Large quantities of particulate sulfate were indeed observed;
this is not only due to the formation from the oxidation of sulfite, but also due to
primary emission of H2SO4 . Evidence of direct emission of H2 SO4 aerosols has
been acquired by the particle acidity measurement system described in a subse-
quent section, It is clear that the simple hydrolytic reaction of NO2 plays a
relatively minor role in these situations since such a reaction must produce
equal amounts of nitrite and nitrate. Significant amounts of aerosol nitrate were
not observed. It is unlikely that nitrate was not observed due to loss of nitric acid
to the walls as the measurement was continuous and the sampling point was
closer to the heater than to the walls.
Measurement in actual home environments with modest kerosene heaters
showed relatively low concentrations of particulate nitrite, <1 /g/m3. The con-
centration of HNO, and particulate nitrate was even lower. Maximum HONO
concentration was in the range of 10 to 20 g/m 3 (ca. 5-10 ppb) and is similar to
the values reported with open flame combustion sources in other studies
[247,252].
202
market for diagnosing and treating indoor air pollution problems is estimated to
be several billion dollars, of which a tenth is likely devoted to analytical and
consulting services [254]. Sick building syndrome represents the tip of the ice-
berg: ten times as many buildings that have not been designated "sick" are still
operating with an unhealthy indoor air environment, their occupants function-
ing sub-optimally. Public awareness of health effects of indoor air pollution is
presently at a peak. New standards for indoor air quality have been proposed.
The biggest concern lies with the children: the General Accounting Office
reports that severe indoor air quality problems exist in 15,000 US schools,
affecting nearly nine million children [255]. In the wake of today's bioterrorism,
these threats and concerns are likely to remain for a very long time to come.
To enable rapid treatment of indoor air pollution problems in a technically
sound manner, it must first be determined whether the problem is primarily
chemical or biological in nature. If, for example, it is diagnosed to be chemical,
portable gas chromatograph-mass spectrometers can be brought on site to
unambiguously characterize unusual amounts of any species present. However,
it is not completely straightforward-unusual concentrations of certain volatile
organic compounds may originate also from fungal or microbiological sources
[256]. It would be better to first decide if a biological problem exists.
Bioaerosols include airborne bacteria, vira, spores, pollens, plant/insect frag-
ments, etc. Above a threshold concentration, many of these species can elicit
pathogenic and allergenic effects. Indoor bioaerosols are currently being investi-
gated with the goal of discovering a cause for ongoing disease or discomfort [257]
to connect specific health effects with specific biocomponents, and to increase
understanding the ecology and physiology of the source organisms and the fate
of their airborne effluents [258]. Whether it is dead tissue or a living organism,
bioaerosols have a greater protein content than the ambient background of
dominantly inorganic aerosol. The vapor condensation based PCS coupled to an
online protein measurement system was used by Poruthoor et al. [259] to
measure protein concentrations in ambient air, in home environment as well as
a biologically contaminated hotel room.
Several modifications to the original steam condensation-PCS were made.
The new design is shown in Fig. 6.29. One limitation of the original PCS is
Peltier cooling, which requires relatively large current. Instead of the Peltier
cooled maze, a flattened glass coil (Fig. 6.29, inset b) made by coiling a glass tube
(120 cm x 0.5 cm i.d.) was used as the impactor. These authors found that
increasing the residence time after steam introduction facilitated the collection
of the particles. In the final arrangement, the mixing chamber, impactor, the
air/liquid separator and other downstream components are all mounted inside a
circular acrylic jacket and air cooled by a fan placed at the bottom of the acrylic
housing. (In this context, glass bead packed column impactors were also tested
for aerosol collection after steam introduction, with or without further water
addition [260]. It is doubtful that the results are competitive with the other
alternatives already described.)
203
Aerosol
Air (b)
Fig. 6.29. Simplified vapor condensation particle collection system [259]. Inset: the details of (a)
the air/liquid separator and (b) the flattened glass coil impactor. Reproduced with permission
from Ref. [259] (copyright 1998, American Chemical Society).
204
nnl;. 7 ^
u.ul~
0 4 1 9 I
lo
o 0.010
o0
o
<0.005
1 2 3
n nn 11<k 4I I
u.uu
U (a) (b) (c) (d)
Fig. 6.30. Representative system output. (a) Blank; peaks 1-3 are from loading 8 ml blank each.
(b) 50 ng/ml BSA standard; peaks 4-6 are from loading 8 ml solution each. (c) System output in
an apartment dwelling, during cooking; peaks 7 and 8 are from 700 and 650 1 air sample. (d)
System output in a hotel room known to have fungal contamination; peaks 9 and 10 are from 450
and 420 1 air sampled inside the contaminated room. Reproduced with permission from Ref.
[259] (copyright 1998, American Chemical Society).
205
H20
I Separator
W
Air '
Anion
0.50
2
mL/min
Peristaltic Pump
Fig. 6.31. Aerosol acidity instrument schematic. See text for details. Reproduced with permission
from Ref. [267] (copyright 1998, American Chemical Society).
206
TABLE 6.5
Aerosol acidity limits of detection as a function of background aerosol composition [269]
H2SO4 :(NH4 )2 SO4 Ratio Background No. data H+ LOD (ng/m3 )
(NH 4 )2SO4 points
Nominal in Found in coneh. g/im3) Experimental Projected
nebulizer feed aerosol 5SLPM, 5 min 5SLPM, 60 min
4:1 3.7:1 2.35 16 8.3 0.7
1:1 1:0.98 5.1 19 7.0 0.6
1:5 1:5.4 16.5 28 21.6 1.8
1:7.5 1:8.5 17.4 16 31.8 2.7
1:10 1:11 15.1 26 38.2 3.2
1:15 1:17 8.25 12 9.7 1.6
Reproduced with permission from Ref. [269] (copyright 1998, American Chemical Society).
I He.ater Off
-- r---l
14.00 16.00 18.00 20.00 22.00 24.00
Time, hours, March 2, 1996
Fig. 6.32. Kerosene heaters cause readily measurable acid aerosol emission. Reproduced with
permission from Ref. [267] (copyright 1998, American Chemical Society).
that the acidity is clearly traceable to the production of H2SO4 aerosol. The
application of this instrument during inhalation toxicology studies concerned
with the effect of low levels of H2 SO4 aerosols on humans in test chambers clearly
showed that almost all of the acid aerosol is actually neutralized by expired
ammonia in the chambers (Fig. 6.33).
207
3.0 -
1.0 -
0.0 -
I I I I 1 1 I I I I I
14.00 1450 15.00 15.50 15.50 16.50 17.00 17.5
Diumal inme, h
Fig. 6.33. It is difficult to perform human toxicology studies to low levels of acid aerosol because
expired ammonia neutralizes the aerosol. When three people walk into an exposure chamber
3
containing 200 Ag/m H2 SO,, the acid is rapidly neutralized. Reproduced with permission from
Ref. [267] (copyright 1998, American Chemical Society).
6.3.2.5 Vapor condensation based systems as collector front ends for the
measurement of trace metals.
A unique concern of relevance to aerosol composition measurements relates to
the monitoring of hazardous nuclear material that may be released in particu-
late form. Since the end of the cold war, large quantities of used nuclear
materials are being brought to the US for destruction and reprocessing. Acciden-
tal radioactive exposures due to human error can occur in all facilities handling
radioactive material. In the case of a major disaster, the threat of nuclear
208
20-
i '"J -S02 [
15 - * Sulfate
[a
f :~''…-^
E 10-
a -
0-
11 12 13 14 15
Diurnal time (hours)
Fig. 6.34. Sulfur dioxide and sulfate concentrations after an accidental release of H2S in a West
Texas oilfield, release occurred at ca. 10:45 am. Oxidation proceeds directly to sulfate, rather
than via SO 2 . Reproduced with permission from Ref. [223] (copyright 1995, American Chemical
Society).
209
6.4 CONCLUSIONS
The demand for (near) real-time aerosol composition measurement has never
been greater. While the focus of this review has been primarily on ambient
outdoor air measurement, there is already worldwide desire to find an effective
means of infective bioaerosol detection. Major inroads have been made in mass
spectrometric identification of intact bacteria [273-275] but the same concerns
that exist with single particle MS studies of ambient air chemistry also exist
here-how practical will it be to deploy such instruments widely? While this
author does not claim to have a crystal ball, it seems likely in the near future that
immunosensing approaches coupled to wet effluent aerosol collectors hold much
promise [276]. Fundamental investigations in both stratospheric and tropo-
spheric chemistry will likely be dominated in the future by particle MS tech-
niques. On the other hand, near real-time systems that integrate collection and
can measure a broad suite of substances at one time (by chromatography or
chromatography-MS techniques) hold much promise in regulatory, epidemiolog-
ical and toxicological applications and more routine measurements.
Acknowledgement
This manuscript would have been impossible without the help of Kajori
Dasgupta and Susie Gonzalez. PKD gratefully acknowledges comments by Don
C. Olson and Zbynek Vecera on the draft manuscript. We also thank Janusz
Pawliszyn for his patience. This review was made possible in part by the
Tropospheric Ozone Research Program of the National Exposure Research
Laboratory, U.S. EPA.
REFERENCES
210
13 D.V. Bates, M. Baker-Anderson and R. Sizto, Environ. Res., 51 (1990) 51-70.
14 G.D. Thurston, K. Ito, M. Lippmann and C. Hayes, Envuiron. Health Persp., 79 (1989)
73-82.
15 M.E. Raizenne, R.T. Burnett, B. Stern, C.A. Franklin and J.D. Spengler, Environ.
Health Persp., 79 (1989) 179-185.
16 T.A. Kimmel, L.C. Chen and C. Nadziejkp, Toxicol. Appl. Pharm., 144 (1997)
348-355.
17 R.F. Phalen and D.V. Bates, Inhal. Toxicol., 7 (1995) 1-163.
18 J.C. Chow, J. Air Pollut. Contr. Assoc., 45 (1995) 320-382.
19 R.B. Schlesinger, Inhal. Toxicol., 7 (1995) 99-110.
20 A.L. Hines, T.K. Ghosh, S.K. Loyalka and R.C. Warder Jr. IndoorAir Quality and
Control. PTR Prentice Hall, New Jersey, 1995.
21 H. Klingenberg and H. Winneke, Sci. Total Environ., 93 (1990) 95-105.
22 I.L. Mauderly, in: M. Lippmann (Ed.), Environmental Toxicants: Human Exposures
and Their Health Effects. Van Nostrand Reinhold, New York, 1992, pp. 119-162.
23 R.A. Goyer, in: C.D. Klaassen, M.O. Amdur and J. Doull (Eds.), Toxicology.
Macmillan, New York, 1986, pp 582-635.
24 D.V. Bates, in: D.E. Gardner (Ed.), Proceedings of the Colloquium on Particulate Air
Pollution and Human Mortality and Morbidity. Inhal. Toxicol., 7(1) (1995) ix-xiii.
25 A. Gupta, D. Tang and P.H. McMurry, J. Atmos. Chem., 20 (1995) 117-139.
26 R.W. Talbot, A.S. Vijgen and R.C. Harris, J. Geophys. Res., 95 (1990) 7553-7561.
27 M. Ferm, Atmos. Environ., 20 (1986) 1193-1201.
28 J.L. Durham, L.L. Spiller and T.G. Ellestad, Atmos. Environ., 21 (1987) 589-598.
29 W. John, S.M. Wall and J.L. Ondo, Atmos. Environ., 22 (1988) 1627-1635.
30 B.R. Appel, V. Povard and E.L. Kothny, Atmos. Environ., 22 (1988) 2535-2540.
31 A.M.N. Kitto and R.M. Harrison, Atmos. Environ., 26A (1992) 235-241.
32 Air quality criteria for particulate matter. US. EPA. Research Triangle Park, NC;
EPA/600-AP-95-1001A.
33 J. Samet and S. Jager, as cited in ref. 34.
34 T. Reichhardt, Environ. Sci. Technol., 29 (1995) 360A-364A.
35 G. Oberdorster, R.M. Gelein, J. Ferin and B. Weiss, Inhal. Toxicol., 7 (1995) 111-124.
36 D.A. Lundgren and R.M. Burton, Inhal. Toxicol., 7 (1995) 131-148.
37 D.J. Jacob, J.M. Waldman, J.W. Munger and M. Hoffman, J. Geophys. Res., 91 (1986)
1089-1096.
38 J.G. Watson, J.C. Chow, J.J. Shah and T. Pace, JAPCA, 33 (1983) 114-119.
39 W.A. Hoppel, J.W. Fitzgerald, G.M. Frick and R.E. Larson, J. Geophys. Res., D4
(1990) 3659-3686.
40 W. John, S.M. Wall, J.L. Ondo and W. Winklmayr, Atmos. Environ., 24A (1990)
2349-2359.
41 C.S. Sloane, J.G. Watson, J.C. Chow, L.C. Pritchett and L.W. Richards, Atmos.
Environ., 25A (1991) 1013-1024.
42 M.E. Kitto and D.L. Anderson, Atmos. Environ., 22 (1988) 2629-2630.
43 J.P. Lodge, Jr., Atmos. Environ., 20 (1986) 1657.
44 C.V. Mathai, J.G. Watson, Jr., C.F. Rogers, J.C. Chow, I. Thombach, J.O. Zwicker, T.
Cahill, P. Feeney, R. Eldred, M. Pitchford and P.K. Mueller, Environ. Sci. Technol.,
24 (1990) 1090-1099.
45 D.A. Lane, N.D. Johnson, S.C. Barton, G.H.S. Thomas and W.H. Schroeder, Environ.
Sci. Technol., 22 (1988) 941-947.
46 S. Rapsomanikis, M. Wake, A.M.N. Kitto and R.M. Harrison, Environ. Sci. Technol.,
22 (1988) 948-952.
47 K.T. Knapp, J.L. Durham and T.G. Ellestad, Environ. Sci. Technol., 20 (1986)
633-637.
211
48 W.T. Sturges and R.M. Harrison, Atmos. Environ., 23 (1989) 1987-1996.
49 W. John, S.M. Wall and J.L. Ondo, Atmos. Environ., 22 (1988) 1627-1635.
50 J.F. Pankow, Atmos. Environ., 2 (1987) 2275-2283.
51 P. Koutraldkis, J.M. Wolfson and J.D. Spengler, Atmos. Environ., 22 (1988) 157-162.
52 W.T.Sturges and G.E. Shaw, Atmos. Environ., 27A (1993) 2969-2977.
53 H.V. Andersen and M.F. Hovmand, Atmos. Environ., 28 (1994) 3495-3512.
54 B.Y.H. Liu, D.Y.H. Pui and K.L. Rubo, in: Aerosols in the Mining and Industrial
Work Environments. Ann Arbor Science, Ann Arbor, MI, 1983.
55 W.H. Chan, D.B. Orr and D.H.S. Chung, Atmos. Environ., 20 (1986) 2397-2401.
56 T.R. Fogg, Atmos. Environ., 20 (1986) 1633-1634.
57 V.R. Highsmith, A.E. Bond and J.E. Howes, Jr., Atmos. Environ., 20 (1986)
1413-1417.
58 X. Zhang and P.H. McMurry, Atmos. Environ., 26A (1992) 3305-3312.
59 S.V. Hering, D.R. Lawson, I. Allegrini, A. Febo, C. Perrino, M. Possanzini, J.E.
Sickles, II, K.G. Anlauf, A. Wiebe, B.R. Appel, W. John, J. Ondo, S. Wall, R.S.
Braman, R. Sutton, G.R. Cass, P.A. Solomon, D.J. Eatough, N.L. Eatough, E.C. Ellis,
D. Grosjean, B.B. Hicks, J.D. Womack, J. Horrocks, K.T. Knapp, T.G. Ellestad, R.J.
Paur, W.J. Mitchell, M. Pleasant, E. Peake, A. MacLean, W.R. Pierson, W.
Brachaczek, H.I. Schiff, G.I. Mackay, C.W. Spicer, D.H. Stedman, A.M. Winer, H.W.
Biermann and E.C. Tuazon, Atmos. Environ., 22 (1988) 1519-1539.
60 P. Masia, V. Di Palo and M. Possanzini, Atmos. Environ.,28 (1994) 365-366.
61 F.W. Lipfert, Atmos. Environ., 28 (1994) 3233-3249.
62 B.L. Davis, Y. Deng, D.J. Anderson, L.R. Johnson, A.G. Detwiler, L.L. Hodson and
J.E. Sickles, Atmos. Environ., 28 (1994) 2485-2491.
63 W.L. Crider, Anal. Chem., 37 (1965) 1770-1773.
64 J.D. Husar, R.B. Husar and P.K. Stubits, Anal. Chem., 47 (1975) 2062-2065.
65 P.T. Roberts and S.K. Friedlander, Atmos. Environ., 10 (1976) 403-408.
66 J.J. Huntzicker, R.S. Hoffman and R.A. Cary, Atmos. Environ., 12 (1978) 83-88.
67 J. Coburn, R.B. Husar and J.D. Husar, Atmos. Environ., 12 (1978) 89-98.
68 R.L. Tanner, T. D'Ottavio, R. Garber and L. Newman, Atmos. Environ., 14 (1980)
121-127.
69 T. D'Ottavio, R.L. Garber, R.L. Tanner and L. Newman, Atmos. Environ., 15 (1981)
197-203.
70 J. Slanina, L.V. Lamoen-Dorrnenbal, W.A. Lingera, W. Meilof, D. Klockow and R.
Niessner, Int. J. Environ. Anal. Chem., 9 (1981) 59-70.
71 R.W. Garber, P.H. Daum, R.F. Doering, T. D'Ottavio and R.L. Tanner, Atmos.
Environ., 17 (1983) 1381-1385.
72 B.R. Appel, R.L. Tanner, D.F. Adams, P.K. Dasgupta, K.T. Knapp, G.L. Kok, W.R.
Pierson and K.D. Reiszner, in: J.P. Lodge (Ed.), Methods of Air Sampling and
Analysis, 3rd edn. Lewis, Chelsea, MI, 1988, pp. 523-532.
73 J. Slanina, C.A.M. Schoonebeek, D. Klockow and R. Niessner, Anal. Chem., 57 (1985)
1955-1960.
74 J.J. Huntzicker, Anal. Chem., 58 (1986) 653-654.
75 F. Lindqvist, Atmos. Environ., 19 (1985) 1671-1680.
76 S. Hering and M.R. Stolzenburg, Integrated collection and vaporization particle
chemistry monitoring, US Patent, 5,983,732, November, 1999.
77 M.R. Stolzenburg and S.V. Hering, Environ. Sci. Technol., 34 (2000) 907-914.
78 http://www.rpco.com/products /ambprod/amb400sindex.htm
79 G.A Allen,. Personal communication, October, 2001.
80 http://www.thermoei.com
81 A.H. Moscowitz, Particle size distribution of nitrate aerosols in the Los Angeles air
basin. EPA-600/3-77-053, 1977.
212
82 R.D. Cox, Anal. Chem., 52 (1980) 332-335.
83 D. Klockow, R. Niessner, M. Malejczyk, H. Kiendl, B. Vomberg, M.P. Keuken, A.
Wayers-Ypelaan and J. Slanina, Atmos. Environ., 23 (1989) 1131-1138.
84 K. Yoshizumi and A. Hoshi, Environ. Sci. Technol., 19 (1985) 258-261.
85 W.T. Sturges and R.M. Harrison, Environ. Sci. Technol., 22 (1988) 1305-1311.
86 H.M. Ten Brink, A. Wayers-Ypellan and S. Slanina, J. Aerosol Sci., 29 (1998)
S633-S634.
87 M. Yamamoto and H. Kosaka, Anal. Chem., 66 (1994) 362-367.
88 http://www.rpco.com/products/ambprod/amb8400n/index.htm
89 B.R. Appel, in: K. Willeke and P.A. Baron (Eds.), Aerosol Measurement-Principles,
Techniques and Applications. Van Nostrand Reinhold, New York, 1993, pp. 233-259.
90 R. Klockenkimper, H. Bayer, A. von Bohlen, M. Schmeling and D. Klockow, Anal.
Sci., 11 (1995) 495-501.
91 D. Pardess, Z. Levin and E. Ganor, Atmos. Environ., 26A (1992) 675-680.
92 P.J. Sheridan, R.C. Schnell, J.D. Kahl, J.F. Boatman and D.M. Garvey, Atmos.
Environ., 27A (1993) 1169-1183.
93 S. Brown, M.C. Dangler, S.R. Burke, S.V. Hering and D.T. Allen, Aerosol Sci.
Technol., 12 (1990) 172-181.
94 W.L. McClenny, K.J. Krost, E.H. Daughtrey, D.D. Williams and G.A. Allen, Appl.
Spectrosc., 48 (1984) 706-712.
95 K.J. Krost and W.L. McClenny, Appl. Spectrosc., 48 (1984) 702-705.
96 B.R. Appel, Y. Tokiwa and M. Haik, Atmos. Environ., 18 (1984) 409-416.
97 E. Bondietty and C. Papastefanou, J. Aerosol Sci., 20 (1989) 667-670.
98 X. Zhang and P.H. McMurry, Atmos. Environ., 26A (1992) 3305-3312.
99 X. Li-Jones, D.L. Savoie and J.M. Prospero, Atmos. Environ., 35 (2001) 985-992.
100 W.H. Chan, D.B. Orr and D.H.S. Chung, Atmos. Environ., 20 (1986) 2397-2401.
101 F.W. Lipfert, Atmos. Environ., 28 (1994) 3233-3249.
102 R.W. Coutant, Environ. Sci. Technol., 11 (1977) 873-878.
103 Z. Ali, C.L. Paul and J.F. Alder, Analyst, 114 (1989) 759-769.
104 G. Lammel, Fres. Z. Anal. Chem., 340 (1991) 684-986.
105 C.L. Benner and D.J. Eatough, Atmos. Environ., 25A (1991) 1537-1545.
106 C. Rosenberg, W. Winiwarter, M. Gregori, G. Pech, V. Casensky and H. Puxbaum,
Fres. Z. Anal. Chem., 331 (1988) 1-7.
107 B.J. Turpin, S.-P. Liu, K.S. Podolske, M.S.P. Gomes, S.J. Eisenreich and P.H.
McMurry, Environ. Sci. Technol., 27 (1993) 2441-2449.
108 R.L. Myers and W.L. Fite, Environ. Sci. Technol., 9 (1975) 334-336.
109 B.A. Mansoori, M.V. Johnston and A.S. Wexler, Anal. Chem., 66 (1994) 3681-3687.
110 M.P. Sinha, C.E. Giffin, D.D. Norris, T.J. Estes, V.L. Vilker and S.K.J. Friedlander,
J. Colloid Interface Sci., 87 (1982) 140-153.
111 K.P. Hinz, R. Kaufmann and B. Spengler, Anal. Chem., 66 (1994) 2071-2076.
112 K.P. Hinz, R. Kaufmann and B. Spengler, Aerosol Sci. Technol., 24 (1996) 233-242.
113 O. Kievit, J.C.M. Marijinissen, P.J.T. Verheijen and B. Scarlett, J. Aerosol Sci., 23
(1992) S301-S304.
114 P.J. McKeown, M.V. Johnson and D.M. Murphy, Anal. Chem., 63 (1991) 2069-2073.
115 D.M. Murphy and D.S. Thomson, Aerosol Sci. Technol., 22 (1995) 237-249.
116 P.G. Carson, K.R. Neubauer, M.V. Johnson and A.S. Wexler, J. Aerosol Sci., 26
(1995) 535-545.
117 K.A. Prather, T. Nordmeyer and K. Salt, Anal. Chem., 66 (1994) 3540-3542.
118 D.-Y. Liu, D. Rutherford, M. Kinsey and K.A. Prather, Anal. Chem., 69 (1997)
1808-1814.
119 E. Gard, J.E. Mayer, B.D. Morrical, T. Dienes, D.P. Fergenson and K.A. Prather,
Anal. Chem., 69 (1997) 4083-4091.
213
120 D.M. Murphy and D.S. Thomson, J. Geophys. Res., 102 (1997) 6341-6352.
121 D.M. Murphy and D.S. Thomson, J. Geophys. Res., 102 (1997) 6353-6368.
122 D.T. Suess and K.A. Prather, Chem. Rev., 99 (1999) 3007-3035.
123 M.V. Johnston, J. Mass Spectrom., 35 (2000) 585-595.
124 C.A. Noble and K.A. Prather, Mass Spectrom. Rev., 19 (2000) 248-274.
125 A.D. Maynard, Philos. Trans. Roy. Soc. A, 358 (2000) 2593-2609.
126 A. Lazar, P.T.A. Reilly, W.B. Whitten and J.M. Ramsey, Environ. Sci. Technol., 33
(1999) 3993-4001.
127 R.P. Rodgers, A.C. Lazar, P.T.A. Reilly, W.B. Whitten and J.M. Ramsey, Anal.
Chem., 72 (2000) 5040-5046.
128 A. Lazar, R.A. Gieray, W.B. Whitten and J.M. Ramsey, Combust. Flame, 122 (2000)
90-104.
129 P.T.A. Reilly, R.A. Gieray, W.B. Whitten and J.M. Ramsey, J. Am. Chem. Soc., 122
(2000) 11596-11601.
130 A. Lazar, P.T.A. Reilly, W.B. Whitten and J.M. Ramsey, Environ. Sci. Technol., 72
(2000) 2142-2147.
131 A. Lazar, P.T.A. Reilly, W.B. Whitten and J.M. Ramsey, Rapid Commun. Mass
Spect., 14 (2000) 1523-1529.
132 P.T.A. Reilly, A.C. Lazar, R.A. Gieray, W.B. Whitten and J.M. Ramsey, Aerosol Sci.
Technol., 33 (2000) 135-152.
133 D.B. Kane and M.V. Johnston, Environ. Sci. Technol., 34 (2000) 4887-4893.
134 D.B. Kane, B. Oktem and M.V. Johnston, Aerosol Sci. Technol., 34 (2000) 520-527.
135 D.J. Phares, K.P. Rhoads, A.S. Wexler, D.B. Kane and M.V. Johnston, Anal. Chem.,
73 (2001) 2338-2344.
136 http://www-wlc.eas.gatech.edu/supersite/
137 A.M. Middlebrook, D.M. Murphy, S.-H. Lee, D.S. Thomson, K.A. Prather, R.J.
Wenzel, D.-Y. Liu, D.J. Phares, K.P. Rhoads, A.S. Wexler, M.V. Johnston, J.L.
Jimenez, J.T. Jayne, D.R. Worsnop, I. Yourshaw, J.H. Seinfeld, R.C. Flagan, S.V.
Hering, R.J. Weber, P. Jongejan, J. Slanina and P.K. Dasgupta, J. Geophys. Res. (in
press).
138 C.D. Clark, P. Campuzano-Jost, D.S. Covert, R.C. Richter, H. Maring, A.J. Hynes and
E.S. Saltzman, J. Aerosol Sci., 32 (2001) 765-778.
139 J. Bacri, A.M. Gomes, J.M. Fieni, F. Thouzeau and J.C. Birolleau, Spectrochim. Acta,
44B (1989) 887-895.
140 D. Nore, A.M. Gomes, J. Bacri and J. Cabe, Spectrochim. Acta, 48B (1993) 1411-1419.
141 A.M. Gomes, J.-P. Sarrette, L. Madon and A. Almi, Spectrochim. Acta, 51B (1996)
1695-1705.
142 A.M. Gomes, C. Trassy, A. Almi, F. Seddiki and S. Hassaine, High Temp. Mater.
Proc., 1 (1997) 461-472.
143 A. Epifanie, J. Bacri, J.P. Sarrete and A.M. Gomes, High Temp. Mater. Proc., 1 (1997)
83-95.
144 A.M. Gomes, A. Almi, P. Teulet and J.P. Sarrette, Spectrochim. Acta, 53B (1998)
1567-1582.
145 Y. Duan, Y. Su, Z. Jin and S. Abeln. Anal. Chem., 72 (2000) 1672-1679.
146 Y. Duan, Y. Su, Z. Jin and S. Abeln, A/P, 71 (2000) 1557-1563.
147 C. Sioutas, P. Koutrakis and B.A Olson, Aerosol Sci. Technol., 21(1994) 223-235;
Particul. Sci. Technol., 12 (1994) 207-221.
148 C. Sioutas, P. Koutrakis and R.M. Burton, J. Aerosol Sci., 25 (1994) 1321-1330.
149 C. Sioutas, P. Koutrakis and R.M. Burton, Environ. Health Persp., 103 (1995)
172-177.
150 J.J. Shah, J.G. Watson, J.A. Cooper and J.J. Huntzicker, J. Air Poll. Contr.Assoc., 36
(1986) 254-257.
214
151 H.A. Gray, G.R. Cass, J.J. Huntzicker, E.K. Heyerdahl and J.A. Rau, Environ. Sci.
Technol., 20 (1986) 580-589.
152 J.H. Offenberg and J.E. Baker, Atmos. Environ., 34 (2000) 1509-1517.
153 A. Molndr, E. M6szdros, H.C. Hansson, H. Karlsson, A. Gelencs6r, G. Kiss and Z.
Krivdcsy, Atmos. Environ., 33 (1999) 2745-2750.
154 J.C. Chow, J.G. Watson, L.C. Pritchett, W.R. Pierson, C.A. Frazier and R.G. Purcell,
Atmos. Environ., 27A (1993) 1185-1201.
155 M. Bizjak, R. Cigler, A.D.A. Hansen and V. Hudnik, Atmos. Environ., 27A (1993)
1347-1350.
156 M. Ammann, M. Kalberer, D.T. Jost, L. Tobler, E. Rossler, D. Piguet, H.W. Gaggeler
and U. Baltensperger, Nature, 395 (1998) 157-160.
157 M. Kalberer, M. Ammann, H.W. Gaggeler and U. Baltensperger, Atmos. Environ., 33
(1999) 2815-2822.
158 M. Bizjak and J. Tursic, J. Aerosol Sci., 29 (1998) S291-S292.
159 A.D.A. Hansen, H. Rosen and T. Novakov, Appl. Opt., 21 (1982) 3060-3062,
160 A.D.A. Hansen, H. Rosen and T. Novakov, Sci. Total. Environ., 36 (1984) 191-196.
161 www.mageesci.com
162 J.F. Hopper, D.E.J. Worthy, L.A. Barrie and N.B.A. Trivett, Atmos. Environ., 28
(1994) 3047-3054.
163 R.L. Gunter, A.D.A. Hansen, J.F. Boatman, B.A. Bodhaine, R.C. Schnell and D.M.
Garvey, Atmos. Environ., 27A (1993) 1363-1368.
164 A.D.A. Hansen, A.V. Pollissar and R.C. Schnell, Atmos. Res., 44 (1997) 153-165.
165 M. Bizjak, J. Tursic, M. Lesnjak and T. Cegnar, Atmos. Environ., 33 (1999) 2783-
2787.
166 S.G. Jennings, F.M. McGovern and W.F. Cooke, Atmos. Environ., 27A (1993)
1229-1239.
167 C. Liousse, H. Cachier and S.G. Jennings, Atmos. Environ., 27A (1993) 1203-1211.
168 L.A. Gundel, R.L. Dod, H. Rosen and T. Novakov, Sci. Total Environ., 36 (1984)
197-202.
169 K. Ruoss, R. Dlugi, C. Weigl and G. Hanel, Atmos. Environ., 27A (1993) 1221-1228.
170 A. Petzold, C. Kopp and R. Niessner, Atmos. Environ., 31 (1997) 661-672.
171 C. Kopp, A. Petzold and R. Niessner, J. Aerosol Sci., 30 (1999) 1153-1163.
172 J. Ballach, R. Hitzenberger, A.E. Schultz and W. Jaeschke, Atmos. Environ., 35
(2001) 2089-2100.
173 K.M. Adams, L.I. Davis, Jr., S.M. Japar and W.R. Pierson, Atmos. Environ., 23 (1989)
639-700.
174 K.M. Adams, L.I. Davis, Jr., S.M. Japar, D.R. Finley and R.A. Cary, Atmos. Environ.,
24 (1990) 597-604.
175 K.M. Adams, L.I. Davis, Jr., S.M. Japar and D.R. Finley, Atmos. Environ., 24 (1990)
605-610.
176 A. Petzold and R. Niessner, Appl. Phys., B63 (1996) 191-197.
177 W.P. Arnott, H. Moosmiiller, C.F. Rogers, T. Jin and R. Bruch, Atmos. Environ., 33
(1999) 2845-2852.
178 A. Petzold and R. Niessner, J. Aerosol Sci., 17 (1986) 705-714.
179 A. Petzold and R. Niessner, J. Aerosol Sci., 24 (1993) S195-S196.
180 S. Keller and J. Heintzenberg, J. Aerosol Sci., 28 (1997) S609-S610.
181 M. Pollard, J. Jaklevic and J. Howes, Aerosol Sci. Technol., 12 (1990) 182-193.
182 E. Schultz, 27A (1993) 1241-1249.
183 M. Bizjak, J. Tursic, V. Hudnik, P. Pavli, R. Cigler, B. Paradiz and B. Paradiz, J.
Aerosol Sci., 28 (1997) S207-S208.
184 V.M.H. Lavanchy, H.W. Gaggeler, S. Nyeki and U. Baltensperger, Atmos. Environ.,
33 (1999) 2759-2769.
215
185 T.A. Pakkanen, C.H. Ojanen, R.E. Hillamo and T. Mkela, J. Aerosol Sci., 29 (1997)
S243-S244.
186 C. Bhugwant, H. Cachier, M. Bessafi and J. Leveau, Atmos. Environ., 34 (2000)
3463-3473.
187 L.M. Castro, C.A. Pio, R.M. Harrison and D.J.T. Smith, Atmos. Environ., 33 (1999)
2771-2781.
188 A.D.A. Hansen and T. Novakov, Aerosol Sci. Technol., 12 (1990) 194-199.
189 G.A. Allen, J. Lawrence and P. Koutrakis, Atmos. Environ., 33 (1999) 817-823.
190 H. Schmid, L. Laskus, H.J. Abraham, U. Baltensparger, V. Lavanchy, M. Bizjak, P.
Burba, H. Cachier, D. Crow, J. Chow, T. Gnauk, A. Even, H.M. Ten Brink, K.-P.
Giesen, R. Hitzenberger, C. Hueglin, W. Maenhaut, C. Pio, A. Carvalho, J.-P. Putaud,
D. Toom-Sauntry and H. Puxbaum, Atmos. Environ., 35 (2001) 2111-2121.
191 A. Even, A. Khlystov and H.M. Ten Brink, J. Aerosol Sci., 29 (1998) S873-S874.
192 B.J. Turpin, R.A. Cary and J.J. Huntzicker, Aerosol Sci. Technol., 12 (1990) 161-171.
193 J.C. Chow, J.G. Watson, D. Crow, G.H. Lowenthal and T. Merrifield, Aerosol Sci.
Technol., 34 (2001) 23-34.
194 K. Fung, Aerosol Sci. Technol., 12 (1990) 122-127.
195 R. Hitzenberger, S.G. Jennings, S.M. Larsson, A. Dillner, H. Cachier, Z. Galambos,
A. Rouc and T.G. Spain, J. Aerosol Sci., 29 (1998) S751-S752.
196 A. Petzold and R. Niessner, Mikrochim. Acta, 117 (1995) 215-237.
197 A.D.A. Hansen and P.H. McMurry, 40 (1990) 894-895.
198 T.T. Mercer, Aerosol Technology in HazardEvaluation. Academic Press, New York,
1973, p 146.
199 B.J. Finlayson-Pitts and J.N. Pitts, Jr., Atmospheric Chemistry: Fundamentalsand
Experimental Techniques. Wiley, New York, 1986, p. 823.
200 K.T. Whitby and W.E. Clark, Tellus, 18 (1996) 573-586.
201 H-C. Yeh, in: K. Willeke and P.A. Baron (Eds.), Aerosol Measurement Principles,
Techniques andApplications. Van Nostrand Reinhold, New York, 1993, pp. 410-425.
202 S. Liu and P.K. Dasgupta, Talanta, 43 (1996) 1681-1688.
203 R. Gunn, J. Colloid Sci., 10 (1955) 107-119.
204 D.K. Brandvold, P. Martinez and R. Hipsch, Atmos. Environ., 30 (1996) 973-976.
205 S. Liu and P.K. Dasgupta, Anal. Chem., 68 (1996) 3638-3644.
206 A. Karlsson, K. Irgum and H.-C. Hansson, J. Aerosol Sci., 28 (1997) 1539-1551.
207 W.C. Hinds. Aerosol Technology. 1982, Chapter 9, pp. 164-186.
208 S.M. Buhr, M.P. Buhr, F.C. Fehsenfeld, J.S. Holloway, U. Karst, R.B. Norton, D.P.
Parrish and R.E. Sievers, Atmos. Environ., 26 (1995) 2609-2624.
209 A. Neftel, Atmospheric Research Laboratory, Swiss Department of Agriculture,
Liebfeld, Switzerland, personal communication, 2001.
210 G. Samanta, C.B. Boring and P.K. Dasgupta, Anal. Chem., 73 (2001) 2034-2040.
211 C.B. Boring, R. Al-Horr, G. Zhang, P.K. Dasgupta, M.W. Martin and W.F. Smith,
Anal. Chem., (submitted).
212 H. Small, Ion Chromatography.Plenum, New York, 1989, pp. 68-69.
213 http://oxmedinfo.jr2.ox.ac.uk/Pathway/Miscell/24028.htm
214 W.R. Cofer, V.G. Collins and R.W. Talbot, Environ. Sci. Technol., 19 (1985) 557.
215 W.R. Cofer, III and R.A. Edahl, Jr. Atmos. Environ., 20 (1986) 979.
216 J. Janak and Z. Vecera, Anal. Chem., 59 (1987) 1494-1498.
217 J. Janak and Z. Vecera, Mikrochim. Acta, 3 (1990) 29-34.
218 R. Al-Horr. G. Samanta and P.K. Dasgupta, to be published, 2002.
219 J. Aitken, Proc. Roy. Soc. Edinburgh(1888) 35-40.
220 Y.S. Cheng, in: K. Willeke and P.A. Baron (Eds.), Aerosol Measurement Principles,
Techniques andApplications.Van Nostrand Reinhold, New York, 1993, pp 427-448.
221 L. Waldmann and K.H. Schmitt, in: C.N. Davies (Ed.), Aerosol Science. Academic
216
Press, New York, 1966, pp 137-162.
222 A. Blatter, A. Neftel, P.K. Dasgupta and P.K. Simon, in: G. Angletti and G. Restelli
(Eds.), Physico-Chemical Behavior of Atmospheric Pollutants, Proc. 6th European
Symposium, Report EUR 15609/2 EN, Luxembourg, 1994, pp. 767-772.
223 P.K. Simon and P.K. Dasgupta, Anal. Chem., 67 (1995) 71-78.
224 P.K. Simon and P.K. Dasgupta, Environ. Sci. Technol., 29 (1995)1534-1541.
225 A. Khlystov, G.P. Wyers and J. Slanina, Atmos. Environ., 29 (1995) 2229-2234.
226 J. Slanina, H.M. ten Brink, R.P. Otjes, A. Even, P. Jongejan, A. Khlystov, A.
Waijers-Ypellan, M. Hu and Y. Lu, Atmos. Environ., 35 (2001) 2319-2330.
227 J. Ruzicka and E.H. Hansen, Flow Injection Analysis, 2nd edn., 1981.
228 R.M. Carlson, Anal. Chem., 50 (1978) 1528-1531; US Patent 4,209,299, June 24,
1980.
229 G.P. Wyers, R.P. Otjes and J. Slanina, Atmos. Environ., 27A (1993) 2085-2090.
230 J. Slanina and G.P. Wyers, Fres. J. Anal. Chem., 350 (1994) 467-473.
231 http://www.timberlineinstruments.com
232 C. Zellweger, M. Ammann, P. Hofer and U. Baltensperger, Atmos. Environ., 33
(1999) 1131-1140.
233 M. Loflund, A. Kasper-Giebl, W. Tscherwenka, M. Schmid, H. Giebl, R.
Hitzenberger, G. Reischl and H. Puxbaum, Atmos. Environ., 35 (2001) 2861-2869.
234 R.J. Weber, D. Orsini, Y. Daun, Y.N. Lee, P.J. Klotz and F. Brechtel, Aerosol Sci.
Technol., 35 (2001) 718-727.
235 K. Okuyama, Y. Kousaka and T. Motouchi, Aerosol Sci. Technol., 3 (1984) 353-366.
236 J.H. Seinfeld, Atmospheric Chemistry andPhysics ofAir Pollution. Wiley, New York,
1986, pp 195-246.
237 R. Atkinson, W.P.L. Carter, J.N. Pitts, Jr. and A.M. Winer, Atmos. Environ., 20
(1986) 408-410.
238 M. Ferm and A. Sj6din, Atmos. Environ., 19 (1985) 979-983.
239 K. Sexton, J.D. Spengler and R.D. Treitman, Atmos. Environ., 18 (1984) 1371-1383.
240 J.L. Repace, in: I.K. O'Neill, K.D. Brunnemann, B. Dodet and D. Hoffmann (Eds.),
Environmental Carcinogens. Passive Smoking. IARC Scientific publications No. 81.
International Agency for Research on Cancer, Lyon, France, 1987, Vol. 9, pp.
141-162.
241 V.R. Highsmith, R.J. Hardy, D.L. Costa and M.S. Germani, Environ. Sci. Technol.,
26 (1992) 673-680.
242 W.F. Rogge, L.M. Hildemann, M.A. Mazurek and G.R. Cass, Environ. Sci. Technol.,
25 (1991) 1112-1125.
243 J. Barnes, P. Holland and P. Mihlmester, Characterization of population and usage
of unvented kerosene space heaters. EPA-600/7-90-004. US EPA, Office of Research
and Development, Washington, DC, 1990, pp. 1-37
244 B.P. Leaderer and P.M. Boone, Environ. Sci. Technol., 24 (1990) 908-912.
245 G.W. Traynor, M.G. Apte, H.A. Sokol, J.C. Chuang, W.G. Tucker and J.L. Mumford,
Environ. Sci. Technol., 24 (1990) 1265-1270.
246 J.L. Mumford, R.W. Williams, D.B. Walsh, R.M. Burton, D.J. Svendsgaard, J.C.
Chuang, V.S. Houk and J. Lewtas, Environ. Sci. Technol., 25 (1991) 1732-1738.
247 J.D. Spengler, M. Brauer, J.M. Samet and W.E. Lambert, Environ. Sci. Technol., 27
(1993) 841-845.
248 J. Notholt, J. Hjorth and F. Raes, Atmos. Environ., 26A (1992) 211-217.
249 W. Junkermann and T. Ibusuki, Atmos. Environ., 26A (1992) 3099-3103.
250 L.R. Martin, D.E. Damschen and H.S. Judeikis, Atmos. Environ., 15 (1981)
1615-1621.
251 S.E. Schwartz, in: J.G. Calvert (Ed.), SO , , NO and NO 2 Oxidation Mechanisms:
Atmospheric Considerations.Butterworth Publishers, Boston, 1984, pp. 173-208.
217
252 Z. Vecera and P.K. Dasgupta, Int. J. Environ. Anal. Chem., 4 (1994) 311-316.
253 B. Lighthart and A.J. Mohr (Eds.), Atmospheric Microbial Aerosols. Chapman &
Hall, New York, 1994.
254 Indoor Pollution News, July 25 (1994) 199.
255 Federal Register, 29 CFR parts 1910, 1915, 1926 and 1928, April 5, 1994; U.S.
General Accounting Office: School Facilities. Condition of America's Schools,
GAO/HEHS-95-61, February, 1995.
256 R. Kostiainen, Atmos. Environ., 29 (1995) 693-702.
257 H.A. Burge, Bioaerosols. Lewis, Ann Arbor, 1995, pp. 1-23.
258 R.A. Wood, A.N. Laheri and P.A. Eggleston, Clin. Exp. Allergy, 23 (1993) 733-739.
259 S.K. Poruthoor, P.K. Dasgupta and Z. Genfa, Environ. Sci. Technol., 32 (1998)
1147-1152.
260 S. Liu and P.K. Dasgupta, Microchem. J., 62 (1999) 50-57.
261 B.M. Loffler and H. Kunze, Anal. Biochem., 177 (1989) 100-102.
262 P.K. Dasgupta, Z. Genfa, S.K. Poruthoor, S. Caldwell, S. Dong and S.-Y. Liu. Anal.
Chem., 70 (1998) 4661-4669.
263 E.L. Avol, W.S. Linn, D.A. Shamoo, K.R. Anderson, R.C. Peng and J.D. Hackney, Am.
Rev. Respir. Dis., 142 (1990) 343-348.
264 K.W. Clark, K.R. Anderson, W.S. Linn and H. Gong, Jr., J. Air Waste Manage. Assoc.,
45 (1995) 923-926.
265 J.A. Last and K.E, Pinkerton, Toxicology, 116 (1997) 133-146.
266 H.Small, T.S. Stevens and W.C. Bauman, Anal. Chem., 47 (1975) 1801-1809.
267 K. Ito, C.C. Chasteen, H.-K. Chung, S.K. Poruthoor, Z. Genfa and P.K. Dasgupta,
Anal. Chem., 70 (1998) 2839-2847.
268 S. Gupta and P.K. Dasgupta, J. Chromatogr.Sci., 26 (1988) 34-38.
269 G.A. Tarver and P.K. Dasgupta, Atmos. Environ., 29 (1995) 1291-1298.
270 G. Andrew, Prospects for environmental monitoring in international safeguards.
IAEA-SM-333/213. International Nuclear safeguards 1994: Vision for the Future,
Vol. 1. IAEA, Vienna, 1994.
271 NA. Wogman, R.W. Perkins and G.R. Holdren, Environmental sampling and
analysis as safeguards tool. IAEA-SM-333/98. International Nuclear Safeguards
1994: Vision for the Future. Vol. 1. IAEA, Vienna, 1994.
272 S.K. Poruthoor and P.K. Dasgupta, Anal. Chim. Acta, 361 (1998) 151-159.
273 S. Vaidyanathan, J.J. Rowland, D.B. Kell and R. Goodacre, Anal. Chem., 73 (2001)
4134-4144.
274 P.A. Demirev, J.S. Lin, F.J. Pineda and C. Fenselau, Anal. Chem., 73 (2001) 4566-
4573.
275 F.Y. Meng, B.J. Cargile, L.M. Miller, A.J. Forbes, J.R. Johnson and N.L. Kelleher,
Nat. Biotechnol., 19 (2001) 952-957.
276 See, for example, Autonomous Pathogen Detection System (APDS), www.globalfia.
com
218
Chapter 7
7.1 INTRODUCTION
TABLE 7.1
Important physicochemical properties of aqueous and solid environmental matrices with
reference to worldwide (ISO), European (EN), or German (DIN) standardized determination
methods. The properties in bold type should be measured during field sampling
Aqueous matrices Solid matricesa
Temperature DIN 38404-C4 Temperature
Density DIN 38404-C9 / Water content ISO 11465: 1993 /
ISO 15212-1: 1988 EN 12880: 2000
Colouring / ISO 7887: 1994 Density ISO 11508:1998/
UV-Vis absorption ISO 11272: 1998
Turbidity ISO 7027:1999 Cation exchange capacity ISO 11260: 1994/
(CEC)b ISO 13536: 1995
pH value ISO 10523:1994 pH valueb ISO 10390: 1994
EN 12178: 1998
Redox potential DIN 38404-C6 Redox potential ISO 11271: 200?c
Conductivity ISO 7888:1985 Conductivityb ISO 11265:1994
Surface tension ISO 4311: 1979 / Particle size distribution ISO 11277: 1 9 98 d
ISO 9101: 1987 (and soil types)
Suspended solids ISO 11923: 1997 Porosity and surface area ISO 9277:1995
Total / Dissolved ISO 8245: 1999 Total Organic Carbon (TOC) ISO 10694: 1995 /
Organic Carbon ISO 14235: 1998
(TOC/DOC)
aAdditional recommendation: ISO 11464: 1994 (soil quality-pretreatment of samples for
physico-chemical analyses).
bThis parameter is measurable only when the solid material is suspended in water.
CFinal draft (2002-03-27) in approval stage.
dSee also the other ISO standards on particle size analysis in ICS field 19.120 (www.iso.ch).
220
7.2 PHYSICOCHEMICAL PROPERTIES OF AQUEOUS MATRICES
221
7.2.3 Colouring [7-9]
The pH, defined as the negative decadic logarithm of the proton (hydronium ion)
activity, is one of the most important parameters in water chemistry. Together
with redox potential it strongly influences the stability, reactivity and mobility
of elements, inorganic and polar organic compounds in aquatic compartments of
the environment and of biological systems. Pure water saturated with carbonic
acid is not pH-neutral but has a pH value of 5.5-5.8. In natural waters the pH
222
I
Fig. 7.1. Selected trace metal concentration in leachates of a topsoil after 24 h shaking at
different adjusted pH (expect pH = 6.8 which is the unaffected value; see text and Ref. [13] for
more details).
spans a range between 9 and 4. Even much lower pH values have been measured
with samples of acid rain and with soil solutions. But often the carbon dioxide/
hydrogen carbonate buffer system determines the pH of water and thus it varies
only between 6.8 and 7.5. Leachates can have very much lower or higher pH
compared to natural waters. A first pH test (e.g. during sampling in the field) is
possible using so-called universal indicators (as paper strips or slabs). Accurate
pH values require an electrometrical determination of the potential difference
between a glass electrode and a reference electrode (e.g. a saturated calomel
electrode) by the use of a commercial pH meter.
The role of pH in aqueous samples is illustrated in the following examples.
The first is taken from Ref. [13]. Subsamples of a topsoil (Greppin, Germany,
0-10 cm) were shaken with water for 24 h in a batch apparatus with a solid: liq-
uid ratio of 1:10. The pH value of the leachates were adjusted by continuous
titration with dilute nitric acid in order to simulate a successive deposition of
acid rain. Figure 7.1 shows how the pH of a leachate controls the mobilization of
trace metals from this soil material and thus the concentration levels to
quantify. With the exception of Fe, the metals show, in most cases, a drastic
increase in solubility at lower pH values. But iron and aluminium concentration
in leachates also increase at higher pH values. These latter effects further
increase in both directions of the pH scale (outside the shown range) and are
responsible for the mobility of Fe and Al in soil systems [4,6,13].
The second example, according to Ref. [14], demonstrates the pH influence
on the partitioning of three model compounds for the 4-quinolone group of
antimicrobials between SPME fibre coating and water sample. Whereas the
223
TABLE 7.2
pH Dependence of the apparent distribution coefficient for three 4-quinolones between the
SPME fibre (type CW/TPR from Supelco, Bellefonte, PA, USA) and buffer solution (calculated
from the regression models given in Ref. [14])
Substance pH
3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0
Flumequine 49 49/ 49, 48 47 45 38 26 14 8 5 4 4
Oxolinic acid 22 22 22 22 22 21 19 16 10 4 2 1 0.2
Sarafloxacin 15 27 49 75 90 94 88 71 45 25 15 12 10
224
free samples (groundwater). Many other instrumental details, for example
impurified electrode surface (i.e. coated with PtO/PtO 2 or PtS) and very small
exchange currents, as well as objective facts, such as interfering reactions at the
electrode surface or the presence of inert or multiple redox couples in the
sample, can limit the interpretation of measured Eh values. Therefore, meaning-
ful determination and interpretation ofEh in natural water is a challenging task.
For more information the interested reader should consult special reviews on
redox potential measurements in the laboratory and in the field [1, p. 495] or [9,
pp. 2-60].
An interesting example of the combined effect of pH and redox potential on
the mobilization of the heavy metals Pb, Cd, and Zn from a contaminated soil can
be found in Ref. [15]. Stirring experiments with soil suspensions were conducted
under three different pH values and three redox potentials (325, 0, -100 mV).
The results show that the metals were sparingly soluble under alkaline condi-
tions (pH = 8.0). Metal solubilities were higher then under slightly acidic condi-
tions (pH = 5), and increased drastically when pH was kept at 3.3. When
solubilities were compared under the same pH values, it was observed that they
increased as redox potential decreased. The authors summarized that acidic and
reducing conditions were most favourable for the mobilization of considered
metals, and that the effect of pH was more significant than that of the redox
potential. However, due to the limited number of tests performed and, above all,
when considering the wide variety of possible binding forms and release/fixation
processes with heavy metals in soil and sedimentary matter, this general conclu-
sion must be looked at with reservation. Investigations by others [16] with sedi-
ment suspensions also show the dominating influence of pH, but at identical pH
values the mobilized portions of Cu and Pb from oxic sediment are ten-fold
higher compared to anoxic conditions. For Cd and Zn two-fold higher concentra-
tions are found in oxic states. Altogether, these converse results show that both
factors, pH and redox potential, can strongly affect the mobilization of heavy
metals.
225
conductivity values of only between 10 and 100 AS cm -1 , but springs from
limestone show values of 1000 AS cm-l and above. Compared with that, sea water
has a fifty times higher conductivity. In soil solutions the conductivity can range
from < 1 mS cm -1 to > 15 mS cm 1'. The conductivity meter is a resistance meter
with bridge fed with high-frequency alternating current which has to be applied
to prevent the polarization of electrodes and additional resistances thus arising.
The measuring cell contains electrodes, platinized or of stainless steel. It is
necessary to read off the temperature during the measurement, i.e. after reach-
ing thermal equilibrium in the measuring vessel. The measured conductivity
value has to be converted by a tabulated correction factor to the reference
temperature of 25°C.
The intermolecular attractive forces in liquid states cause the macroscopic prop-
erty of surface tension which has an extremely high value in pure water: 72.8
mN/m at 20°C [18], exceeded only by mercury. This water property is very sensi-
tive to impurities. Not only the so-called surface-active agents (detergents), but
most added organic substances reduce the surface tension of water considerably.
(Only dissolution of inorganic salts, mineral acids, and some sugars increase the
surface tension to higher values than that of pure water.) For example, addition
of only 0.1% of n-propanol to water lowers the surface tension to 67 mN/m [18];
the same reduction results if only 30 ng of di-2-ethyl-hexyl phthalate, a ubiqui-
tous plasticizer, are present per litre water [19]. The existing techniques for
measuring surface tension have self-explanatory names such as capillary height,
drop weight, ring detachment method, etc. In all such methods the hydrostatic
pressure drop against the surface and radii of curvature are implicated in differ-
ent ways. Laboratory equipment is available commercially at different levels of
accuracy. Practical problems with reproducibility can arise because the surface
tension depends on the kinetics of diffusion of surface-active impurities between
the bulk phase and the surface region. Surface tension measurements of real
water samples (fog, rain and wastewater) using the drop weight method have
been reported [20,21] and can serve as an overall pollution indicator. But mea-
surements of surface tension of water samples can also, at least in some cases,
replace expensive isolation and chemical analysis of detergents. Moreover, a rel-
atively small surface tension (< 65 Nm/m) indicates the possible presence of col-
loidal structures (micelles) in the water sample (precisely when the critical
micelle concentration is exceeded) and can cause the addition of a micro-
emulsion-breaking agent before continuing with sample preparation.
226
and animal tissue (e.g. polysaccharides and proteins), humic substances,
residues from coal and oil processing and fuel combustion (hydrocarbons),
detergents, pesticides, etc. One can subdivide this organically bound carbon
more or less arbitrarily by filtration with, for example, a membrane of 0.45 Am
pore diameter into a particulate and dissolved fraction. The latter fraction could
be further divided into a colloidal and a truly dissolved part. In many cases,
however, only the total of not-settleable organic material in a water sample is of
interest for sample preparation and analysis. In order to determine TOC/DOC,
the organic substances in the sample are completely oxidized by wet-chemical
means or UV rays, or burnt at high temperature. The amount of CO2 released
can then be measured. Commercial apparatus often use non-dispersive infrared
analysis devices. The measurement requires calibration using standard solu-
tions. Since water samples contain also inorganic carbon (CO2, HCO,-, CO3 2 -),
this must be removed by adding HC1 before TOC/DOC determination. Bidistilled
water has a TOC value < 0.1 mg/l (given as elementary C), deionized water due
to traces from ion-exchange resin already up to 1 mg/l. Untouched natural
waters and drinking water normally show a TOC between 1 and 2 mg/l; in
polluted rivers it is ten times higher. Wastewaters and leachates are often far
above that. The presence and quality of TOC/DOC in aqueous media are import-
ant parameters for the distribution and mobility of trace metal ions as well as
organic substances, both nonpolar and partly dissociating. As mentioned above
in connection with turbidity, it is a challenging task to clear up size distribution
and constitution of TOC/DOC determined summarily in a water sample.
Depending on the sorption behaviour of analytes, TOC/DOC acts as an
additional sink or source for them and modifies their available water concentra-
tion as well as their mass-transfer characteristics during extraction steps, etc.
Such processes can strongly influence analytical sample preparation and calibra-
tion and are often summarized under the term matrix effects. The following
examples illustrate this for both organic analytes and metal ions.
The influence of DOC of different origins (and of pH) on the direct SPME of
halogenated phenols is described in [22]. The authors have tested the natural
organic matter of a brown water lake, of a compost extract, of commercial
Aldrich humic acid, and of the protein bovine serum albumine. The distribution
coefficients between SPME fibre coating and sample were calculated according
to a first-order extraction kinetics. In general, the distribution equilibrium was
established faster in the presence of dissolved organic matter. An additional
transport process of DOC-sorbed analytes to the fibre was assumed (carrier
function of DOC molecules) and also a change of the sorption characteristics of
the SPME fibre itself (partial coating with DOC) could not be excluded. A similar
accelerating effect on the SPME kinetics was observed with very hydrophobic
polyaromatic hydrocarbons (PAHs) with four or more rings in the presence of a
Roth humic acid [23]. Interestingly, for the medium hydrophobic three-ring
PAHs the extraction kinetics is not significantly affected and for less hydropho-
bic compounds (naphthalene, biphenyl) even retarded. Leaving all doubtful
227
mechanistic interpretation aside, these examples lead us to the conclusion that
such a strong influence of DOC on this special preconcentration technique can
only be taken into account by applying matrix-related calibration.
The investigation of particle distribution of the DOC (and of pH) and its
effect on measured Cu concentrations is reported in Ref. [24]. Soil solutions were
filtered through different pore size membrane filters (0.45, 0.22, and 0.1 /tm).
There was no difference for DOC concentrations in the three pore size differenti-
ated filtrates at low pH values where fulvic acid is the predominated dissolved
organic matter. At higher pH values, a reduction in membrane filter pore size
decreased the DOC concentration. This indicates an increasing presence of
humic acid with higher molecular weight. With Cu it was further shown that, by
using different sizes of filters, the evaluation of mobile pollutants could be quite
different. For example, the DOC concentration determined in 0.22 and 0.45 ,-m
filtrates differ by 35%, which resulted in 54% soluble Cu concentration differ-
ence due to the strong binding affinity of Cu to the higher molecular weight
organic matter. The effect of DOC on Cu partitioning was further investigated in
river water samples [25].
To end this section, a few more papers can be recommended which further
illustrate the role of selected physicochemical parameters in preparation and
analytical processing of aqueous samples, especially of landfill leachates [26], or
critically review existing knowledge on the characterization of aquatic colloids
and macromolecules [27] and on estimating dissolved organic carbon partition
coefficients for nonionic organic chemicals [28].
Water is present in almost all solid matrices. It is necessary to know the water
228
content of a sample for weighing definite amounts of solid material required in
further sample processing (extraction, leaching, mixing) and for relating analyti-
cal results to dry weight. Some other analytical methods take advantage of the
fact that the presence of some percentage of water can enhance volatilization of
nonionic or weakly polar organic analytes (BTEX, pesticides, etc.) from solid
samples by displacing them from adsorption sites. Reproducible work also
requires here the determination of initial water content. Moreover, it is partly
necessary to remove water from the sample before subdivision, extraction, or
conservation of the sample. The major problem in water-content determination
is to define at what point the sample is "dry", especially for porous materials
(soils, sediments, etc.) where water can be present in different forms. Colloi-
dal-bound water and portions adsorbed at the mineral surface are more easily
removable than water of crystallization ("structural water") which only comes
off at elevated temperatures. Thermal dehydration curves can be very specific
for the different materials and so exact reporting of the temperature of determi-
nation is important. Additional critical parameters for reproducible and compa-
rable data are accurate temperature measurement and control of the laboratory
oven used. The traditional procedures for most purposes, where the highest
precision in water content determination is not required, involves weighing a
representative part of a moist sample, oven drying at 105°C, reweighing, and cal-
culating the mass of water lost as a percentage of the mass dried material. The
reader should follow the standard procedures closely because, in addition to the
problem of temperature control, drying duration and cooling the sample in a des-
iccator are also important steps. Usually, one continues drying to constant
weight and this can take up to three days depending on the amount of sample.
Special problems can arise if the sample contains solvents or oil/grease in more
then trace quantities. In that case, it is either possible to distil the sample with
toluene in a special apparatus or to use Karl Fischer titration for water-content
determination.
229
collectively and thus means the overall solid density in the sense described
above. Further on, soil scientists have defined a so-called "bulk density" which is
the ratio of the mass of oven-dried solid material to the bulk volume of the solids
plus the pore space at some specified water content, usually that at sampling.
Due to soil shrinking/swelling on desiccation/moisturisation, this "bulk density"
is a dynamic property. When measured, e.g. by the core method, it can be used to
estimate the total porosity of a soil under defined environmental conditions.
Mainly due to their content of clay and amorphous minerals and of organic
matter, solid environmental matrices possess a negatively charged surface that
results from atomic substitution in the lattice of the minerals or from (pH-
dependent) deprotonation of functional groups. This negative charge is balanced
by positively charged cations and as a result the sorbed "exchange complex" is
formed. The cation exchange capacity (CEC) is an important property of the
considered solid materials because it controls, to a large extent, the retention of
inorganic and organic species. The CEC measuring principle is to saturate the
material with some "index" cation (Ba2 + ) that forces other cations, already
present on the charged surfaces (mainly Mg + , Ca2+, K+ , Na + , NH4+ , Al3+), into
solution, followed by quantification of these displaced cations. The measure-
ments can be complicated (erroneous) due to additional dissolution of soluble
salts (calcite and gypsum), specific adsorption of the small cations (K+ , NH4+ ) in
the interlayer position of particular clay minerals, and of the specific adsorption
of trivalent cations (Al3+, Fe 3+ ) at the material surface. The so-called "potential"
CEC is the value measured at the pH of 8.2 where complete deprotonation of the
carboxyl groups of humic matter is assumed. In contrast, the "effective" CEC is
measured at the native pH of the soil/sediment (solution). To give examples of
the occurring CEC values, we take the clay minerals kaolinitite and montmoril-
lonite as two extremes. The former shows a low CEC from 2-15 cmol/kg. For the
latter, one can measure CEC values from 80-150 cmol/kg. The differences are
caused by the different structures of these minerals, especially the accessibility
of interlayer spaces between silica tetrahedral sheets and aluminium octahedral
sheets. Additional large contributions to the CEC (up to 80%) of soils are
provided by the organic matter present. These contributions are variable (pH-
dependent). Knowing the CEC of the material is advantageous for sample
preparation to prevent (or take into account) possible losses of analytes or
standards during operation steps. One should consider that not only inorganic
species but also organic cations (n-alkyl ammonium ions, etc.) are affected by
such exchange processes. The maximum adsorbed amount of the herbicide
Paraquat (1,1'-dimethyl-4,4'-dipyridylium-dichloride), for example, was deter-
mined to 4.0 cmol/kg on kaolinite and to 78 cmol/kg on montmorillonite [41] and
hence reaching nearly the CEC value of these clay minerals.
230
7.3.5 pH Value [4,5,42]
Just as in the aqueous phase, the measurement of redox potentials in soils and
sediments is usually done with a Pt electrode in combination with a reference
electrode as described above. But, as already mentioned, gases present in large
quantities can make Eh measurements questionable, especially in aerated soils
(02) and sulphur-rich sediments (H2S), because the electrodes could react with
these gases and form coatings of oxides or sulphides on the electrode surface.
Moreover, a measured Eh value mostly represents mixed potentials due to the
mineralogical heterogeneity of a soil/sediment system compared to the dimen-
sion of the electrode. The degree and variability of soil moisture also influence Eh
as well as the predominant pH value (about -59 mV per pH unit). Thus, mea-
surements of redox potential are most reliable, i.e. unambiguous to interpret, for
231
flooded soils and recent sediments, if the layer structure is not disturbed too
much by inserting the electrodes and any "contamination" with air is avoided.
Furthermore, Eh should be recorded not before the signal becomes stabilized. In
soils and sediments this takes ten minutes and more; sometimes it is recom-
mended to wait at least 30 min before recording and to repeat the measurement
in periods of 5-10 min until potential drift stops. For oxidized soils redox poten-
tials of +400 to + 700 mV can be measured. Seasonally water-saturated soils can
change to a highly reduced status (below -250 mV). The overall redox status is
determined by a number of redox reactions mainly under participation of
iron(III) and manganese(III/IV) oxides/hydroxides, which are quite common in
soils and sediments as suspended particles and as coatings on clay mineral sur-
faces, and of Fe(II) minerals and a number of natural organic substances with
(carboxyl, phenolic, etc.) functional groups as reducers. Although according to
the above-mentioned limitations, Eh is no quantitative measure of redox status,
it provides important information on conditions that are favourable for
increased mobility, bioavailability, etc., of nutrients (e.g. nitrate-nitrite) and
trace elements in soils or sediments. Chromium and arsenic are well-known
examples of toxic trace elements which can exist in several oxidation states and
thus are strongly affected by redox potential changes. This has to be taken into
account when choosing sample preparation and analytical methods.
Masscheleyn et al. [44] describe a laboratory study undertaken to show the
influence of redox potential and pH on the arsenic speciation and solubility in
subsamples of a contaminated soil which were incubated in canisters under
flooded conditions for different periods of time (from 1 to 105 days). At higher
soil redox levels (500-200 mV) arsenic solubility was low and the major part of
arsenic in solution was present as As(V). An alkaline pH or the reduction of
As(V) to As(III) released substantial proportions of arsenic into solution. Under
moderately reduced soil conditions (100-0 mV), arsenic solubility was controlled
by the dissolution of iron oxyhydroxides. Arsenic coprecipitated before as As(V)
together with iron oxyhydroxides, was released upon their solubilization. Upon
reduction to -200 mV, the soluble arsenic content increased 13-fold as compared
to 500 mV. Interestingly, these authors also report that an equilibration period
of 4 h was necessary for the microelectrodes to obtain a constant redox reading.
Particle size (and shape) controls sedimentation rate and packing density of
solid material. Usually, we distinguish only between three fractions: first the
coarse particles (2 m and above) which are observable by the naked eye; second
the fine particles (2-0.5 Am) which settle in a limited period of time; and third
the ultrafine particles (<0.5 Abm) which normally stay dispersed for a long time
and cause the turbidity of a suspension. But the particles contained in a given
solid sample can also be sorted into narrower size classes to obtain a distribution
function. Currently, there exist several classification schemes which set slightly
232
different but comparable limits for the particle classes. These classes are gravel
(8-1 cm), sand (<1 cm to 63 (or 50) Am), silt (<63 (or 50) Am to 2 ptm), and clay
(<2 m). Additionally, each class is subdivided in a fine, medium, and coarse
fraction. Particle-size distribution in itself is not a particularly interesting
function: rather it is the influence of size on the macroscopic properties of a
dispersed system that is important. In soil science, for instance, particle-size
analysis is used to evaluate the soil texture which is based on different combi-
nations of sand, silt, and clay and can be systematized by means of a so-called
textural triangle. The texture has a determining influence on soil properties
such as water content and retention, hydraulic conductivity, etc., which in turn
have great significance for soil fertility. Another aspect, more relevant to
analytical-chemical investigations, is that particle size determines the available
outer particle surface area which gives sorption capacity for inorganic coatings,
biofilms, trace elements, and organic compounds. A simple calculation of surface
area per unit mass of uniform cubes or spheres according to the equation
Area/Mass = 6 /(p x d),
where p is the density of particles and d either the edge of the cube or the
diameter of the sphere, illustrates the increase in specific area with decreasing
particle size [17]. Assuming a density of 2.00 g/m 3 , the specific area of a particle
with d of 1 cm, 1 mm, and 1 im is 3, 30, and 30,000 cm2 /g, respectively. A
knowledge of the particle-size distribution of a solid sample and the separation
and investigation of special size classes can bring more insight into distribution
and transport behaviour of contaminants in soils, sediments, and waste
materials. It is also well known from practice that working with (sub)samples
which have a definite size range gives better reproducibility in leaching tests and
also better comparability of results from different samples.
A number of particle-sizing methods have been developed; in general, they
do not cover the same size ranges, require different amounts of sample material
and different methods of sample pretreatment (removal of carbonates, iron
oxide, organic matter, etc.) and dispersion techniques (chemically by adding
phosphates and detergents or physically by stirring or ultrasonication). Not all
of these methods are useful for preparative purposes. They can be classified
according to the measuring principle used or to their applicability for special size
ranges. Sieving is by far the most common procedure; it can cover a wide range of
sizes (down to 5 Atm) and yields also larger, i.e. analytically useful, quantities of
particular size classes. Sieving can be "wet" or "dry" according to the dispersion
medium. However, particle sizing by sieving is affected by several variables such
as particle shape, cohesiveness of particles, brittleness of particles, shape of the
sieve openings, sieve loading, agitation, and duration of sieving. But most of
these possible pitfalls can be prevented by following the standard procedures
and the descriptions given by the manufacturers of sieving equipment.
Ackermann et al. [45] found a simple procedure to separate the < 60-/vm and the
<20-tim size fraction of sediments for heavy metal analysis which is in routine
233
use now (e.g. regulated by the environmental authority of Hamburg, Germany):
a sediment suspension (siftings of the greater sieve) is transferred to a plastic
sieve which is placed in a beaker in an ultrasonic bath together with some
volume of distilled water, followed by 5-15 min of ultrasonic treatment (see
original publication for further details).
Particle-size distribution may also be determined by measuring sediment-
ation velocity of particle settling under gravitational forces. The general method
is to sample the mass of particles in an initially homogeneous dispersion at
different heights at different times (e.g with a special pipette). It is assumed that
settling of particles obeys Stokes' law, which relates the size of free falling
spherical particles (i.e. at high dilution) to the sedimentation velocity in a
laminar flow. Usually some of these assumptions are not exactly fulfilled and
thus the calculated size may differ from the real particle size. Practically,
sedimentation methods are limited to size ranges larger than 5 m and to
dispersions settling within 30 min. These limitations can be overcome by
applying an elutriator, a cyclone, or a centrifuge and thus the measuring range
can be extended to submicron particles. During centrifuging the particles do not
sediment at a constant velocity, but they accelerate as they move farther and
farther from the axis of rotation. The time required for sampling (i.e. the time for
particle settling up to a definite cut-off size) can be adjusted by varying the
angular velocity (calculable for the given geometry of rotor, viscosity of
dispersion fluid, and density difference between particle and fluid). Optical
microscopy is a particle-sizing method with broad applicability and the only
technique that allows direct observation and measurement of fine particles. A
serious drawback of microscopy is that the specimen examined includes only a
minute fraction of the whole sample and therefore representative sample
splitting is a critical point. Additionally, dispersions cannot be examined directly
but the subsample has to be dried and mounted prior to analysis. Analogous
problems arise with the more advanced microscopic techniques-transmission
and scanning electron microscopy. Radiation scattering of particles is the basic
principle of another common type of method and includes static and dynamic
(laser) light scattering, small-angle neutron scattering or X-ray scattering, and
others. The useful range of these methods is limited by the size of the particles
relative to the wavelength of the radiation. One of the chromatographic tech-
niques useful for particle-sizing is the so-called hydrodynamic chromatography.
Particles in the 30-1500 nm size range can be separated by using packed or
capillary columns. Size-exclusion chromatography uses a column packed with
porous particles providing an additional separation effect and is applicable to
particles smaller than 400 nm. Field-flow fractionation employs an external field
(a centrifugal force, a cross flow, an electrical potential, or a temperature
difference) to induce migration of particles in a direction perpendicular to the
laminar carrier liquid flow in a narrow separation channel. The particle size
amenable to separation may range from 10 nm to 100 um, depending on the field
strength and the method used. More details of these chromatographic methods
234
8
A
Zn
Fig. 7.2. Metal concentration in the size fractions <20 Am and 63-20g/m of a brook sediment (BS)
and a lake sediment (LS) separated by dry sieving (DS) and wet sieving (WS), respectively (see
text for more details).
as well as other separation techniques not described here (e.g. electric zone
sensing) can be found in the literature. However, it should be mentioned that
sequential filtration of a dispersed sediment sample seems to be, against widely
held opinion, unsuitable for particle-sizing and separating of definite fractions,
irrespective of the filter membrane types used [46].
An example should demonstrate the uneven heavy metal burden of different
particle size fractions of two contaminated sediment samples [47]. Bottom
sediment samples were taken from a stream (Boese Sieben) and a lake (Suesser
See) in the Mansfeld region in Saxony-Anhalt, Germany, where 800 years of
black shale mining and copper smelting generated large quantities of mining
debris and processing waste (slag heaps, tailings, etc.). Wet sieving (with water
jet siever WS1 from Retsch, Haan, Germany) and dry sieving (with sonic sifter
Analysette 28 from Fritsch, Idar-Oberstein, Germany) were used to obtain two
different size classes of fine particles in sufficient quantity for extraction with
aqua regia and following element determination using ICP-AES. It can be seen
from Fig. 7.2 that the < 2 0 -,m particle fractions contain twice as much Zn, Pb,
and Cu than the 63-20-utm fractions. Both sieving methods (including the
different prior conservation/preparation techniques) yield comparable results.
Dependencies of pollutant concentration on particle size have also been
identified with hydrophobic chemicals (e.g. PAHs, PCBs, chlorinated hydrocar-
bons) in natural sediments [48], whereby it is not necessarily the finest fraction
(clay) where the major part of these compounds are concentrated [49,50]. The
organic carbon content of soils/sediments seems to be the primary factor control-
ling the sorption of hydrophobic organic substances (see Section 7.3.9).
235
7.3.8 Surface area and porosity [6,12,17,33,34,391
236
Section 7.3.3). It gives an estimate of the fraction of total sample volume
occupied by pores. More advanced methods (porosimetry) examine porous
structures often indirectly by filling the pores with a gas (helium) or liquid
(mercury) to measure the pore volume. As with the surface area measurements,
assumptions have to be made about the accessibility of pores of different size by
the measuring fluid and thus the porosity values depend on the method of deter-
mination. This limited comparability applies in particular to environmental
solid materials due to their complex nature and has to be taken into account in
evaluating data from literature. The presence of pores in a solid material can
considerably increase the diffusion path length of atoms/molecules, resulting in
a slowing down of adsorption/desorption kinetics. Hence, porosity can largely
influence sample preparation techniques (possible extraction yields, limits of
quantification), particularly when short-term methods are performed.
Plant debris and degradation products of animals form the organic matter of
soils and sediments. The contents range from 0.5 to 5% (on a weight basis) in the
surface horizon of mineral soils and can be much higher in sediments, peat mate-
rial, and compost. The main elementary constituents of this organic matter are
C (>50%), O (30-40%), H (3-5%), and N (4%). One can distinguish chemically
between persistent humic matter and easily degradable nonhumic substances
(carbohydrates, proteins, amino acids, fats, waxes, etc.). Humic matter can be
subdivided into fulvic acids, humic acids and humin, based on their solubility in
acid or base. This highly condensed organic matter has a large specific surface
(800-900 m2/g) and CEC (150-300 cmol/kg). Due to these properties, it is an
important adsorbent of plant nutrients, heavy metals, and organic compounds
such as pesticides and PAHs. The linear sorption coefficients of a nonpolar
chemical with different soils/sediments were found directly to be proportional to
the organic carbon content. Based on this fact, an organic-carbon normalized
partition coefficient, K,,, can be defined [48]. Numerous studies have shown
that Koc values for a nonpolar organic compound of limited water solubility
(<10-3 M) are relatively constant and, within the experimental errors, independ-
ent of the particular soil or sediment investigated. Similarly to aqueous-sample
processing, the methods for determining TOC are based either on wet oxidation
of a specimen in an acid dichromate solution, followed by determination of the
remaining dichromate or collection and quantification of evolved CO2 , or alter-
natively on dry oxidation of a sample in a furnace followed by CO 2 determina-
tion. It is necessary to differentiate organic carbon from the inorganic part
either by sample pretreatment with a mineral acid to decompose carbonates or
by employing different furnace temperatures. Commercial instruments are
available for the combustion method, most of them work with temperature
ramping and determine the evolved CO2 by infrared absorption measurement.
Knowledge of TOC of material under investigation is important not only for a
237
reasonable comparison of results (OC-normalization) but also for sample prepa-
ration prior to the analysis of hydrophobic organic substances to account for
incomplete and varying recovery of standards (formation of bound residues).
Another aspect is the potential dissolution of organic carbon constituents which
can lead to undesirable effects during sample processing and analysis (masking
of analytes, cf. Section 7.2.9).
Finally, some additional articles can be recommended here which summarize
existing knowledge on solid-solution partitioning of metals in contaminated
soils, especially the dependence on pH, total metal burden, and organic matter
[51], or review the factors affecting sorption of organic compounds in natural
sorbent/water systems [52,531.
REFERENCES
1 W. Stumm and J.J. Morgan, Aquatic Chemistry. 3rd edn. Wiley, New York, 1996.
2 F.J. Millero, The Physical Chemistry of Natural Waters. Wiley, New York, 2001.
3 K.B. Krauskopf and D.K. Bird, Introductionto Geochemistry, 3rd edn. McGraw-Hill,
New York, 1995.
4 P. Schachtschabel, H.P. Blume, G. Briimmer, K.H. Hartge and U. Schwertmann,
Scheffer/Schachtschabel-Lehrbuch der Bodenkunde, 13th edn. F. Enke Verlag,
Stuttgart, 1992.
5 D.L. Rowell, Soil Science: Methods and Applications. Addison-Wesley Longman,
London, 1994.
6 D.L. Sparks, Environmetal Soil Chemistry. Academic Press, San Diego, CA, 1995.
7 L. Hiitter, Wasser und Wasseruntersuchung, 5. Aufl., Salle-Verlag, Frankfurt a.M.,
Germany &Verlag Sauerldnder, Aarau, Switzerland, 1992.
8 W. Fresenius, K.E. Quentin and W. Schneider (eds.) Water Analysis-A Practical
Guide to Physicochemical, Chemical and Microbiological Water Examination and
Quality Assurance. Springer-Verlag, Berlin, 1988.
9 A.E. Greenberg, L.S. Clesceri and A.D. Eaton (eds.), Standard Methods for the
Examination of Water and Wastewater, 18th edn. American Public Health
Association/American Water Works Association/Water Environment Federation,
Washington, DC, 1992.
10 B.G. Bierwagen and A.A. Keller, Environ. Toxicol. Chem., 20 (2001) 1625-1629.
11 I.D. Wilson, E.A. Adlard, M. Cooke and C.F. Poole (eds.), Encyclopedia of Separation
Science. Academic Press, San Diego, CA, 2000.
12 E. Kissa, Dispersions-Characterization,Testing, and Mesurement. Marcel Dekker,
New York, 1999.
13 A. Kruiger, A. Paschke, B. Schneider, E. Biittner and H. Neumeister, Petermanns
Geograph. Mitteil., 142 (1998) 375-392.
14 H.C. Holten Liitzh0ft, W.H.J. Vaes, A.P. Freidig, B. Halling-S0rensen and J.L.M.
Hermens, Environ. Sci. Technol., 34 (2000) 4989-4994.
15 M.C. Chuan, G.Y. Shu and J.C. Liu, Water Air Soil Pollut., 90 (1996) 543-556.
16 W. Calmano, J. Hong and U. F6rstner, Vom Wasser, 78 (1992) 245-257.
17 S. Ross and I.D. Morrison, Colloidal Systems and Interfaces. Wiley, New York, 1988.
18 D.R. Lide (ed.), Handbook of Chemistry and Physics, 73rd edn. CRC Press, Boca
Raton, FL, 1992.
19. M. Thomsen, L. Carlsen and S. Hvidt, Environ. Toxicol. Chem., 20 (2001) 127-132.
20 P.D. Capel, R. Gunde, F. Zfircher and W. Giger, Environ. Sci. Technol., 24 (1990)
722-727.
238
21 R. Gunde, M. Dawes, S. Hartland and M. Koch, Environ. Sci. Technol., 26 (1992)
1036-1040.
22 G. Ohlenbusch, M.U. Kumke and F.H. Frimmel, Sci. Total Environ., 253 (2000)
63-74.
23 F.D. Kopinke, J. Prschmann, and A. Georgi, Chapter 8 in: J. Pawliszyn (ed.),
Applications of Solid Phase Microextraction. Royal Society of Chemistry, Cambridge,
1999, pp. 111-128.
24 S.J. You, Y. Yin. and H.E. Allen, Sci. Total Environ., 227 (1999) 155-160.
25 Y. Lu and H.E. Allen, Sci. Total Environ., 277 (2001) 119-132.
26 V. Goundaris, P.R. Anderson and T.M. Holsten, Environ. Sci. Technol., 27 (1993)
1381-1387.
27 J. Buffle and G.G. Leppard, Environ. Sci. Technol., 29 (1995) 2169-2184.
28 L.P. Burkhard, Enuiron. Sci. Technol., 34 (2000) 4663-4668.
29 S.A. Taylor and R.D. Jackson, Chapter 37 in: A. Klute (ed.), Methods of SoilAnalysis.
Part 1-Physical and MineralogicalMethods, 2nd edn. American Society of Agron-
omy & Soil Science., Madison, WI, 1986, pp. 927-940.
30 W.H. Gardner, Chapter 21, ibid., pp. 491-544.
31 G.R. Blake and K.H. Hartge, Chapter 13 and 14, ibid., pp. 363-382.
32 G.W. Gee and J.W. Bauder, Chapter 15, ibid., pp. 383-412.
33 D.L. Carter, M.M. Mortland, and W.D. Kemper, Chapter 16, ibid., pp. 413-424.
34 R.E. Danielson and P.L. Sutherland, Chapter 18, ibid., pp. 443-462.
35 G.C. Topp, Chapter 51 in: M.R. Carter (ed.), Soil Sampling and Methods ofAnalysis.
Lewis Publishers, Boca Raton, FL, 1993, pp. 541-557.
36 J.L.B. Culley, Chapter 50, ibid., pp. 529-540.
37 W.H. Hendershot, H. Lalande and M. Duquette, Chapter 19, ibid., pp. 167-176.
38 B.H. Sheldrick and C. Wang, Chapter 47, ibid., pp. 499-511.
39 M.R. Carter and B.C. Ball, Chapter 54, ibid., pp. 581-588.
40 H. Tiessen and J.O. Moir, Chapter 21, ibid., pp.187-200.
41 B.A.G. Knight and T.E. Tomlinson, J. Soil Sci., 18 (1967) 233-243.
42 A. Murdoch and J.M. Azcue, Manual of Aquatic Sediment Sampling. Lewis
Publishers, Boca Raton, FL, 1995.
43 V.P. Evangelou, Environmental Soil and Water Chemistry. Wiley, New York, 1998.
44 P. Masscheleyn, R.D. Delaune and W.H. Patrick, Jr., Environ. Sci. Technol., 25
(1991) 1414-1419.
45 F. Ackermann, H. Bergmann and U. Schleichert, Environ. Technol. Lett., 4 (1983)
317.
46 I.G. Droppo, B.G. Krishnappan, S.S. Rao and E.D. Ongley, Environ. Sci. Technol., 29
(1995) 546-550.
47 A. Paschke and R. Wennrich, unpublished results, UFZ Centre for Environmental
Research Leipzig, Germany, 2000.
48 S.W. Karickhoff, D.S. Brown and T.A. Scott, Water Res., 13 (1979) 241-248.
49 G. Umlauf and R. Bierl, Z. Wasser-Abwasser-Forsch., 20 (1987) 203-209.
50 I.A.Razak, A. Li and E.R. Christensen, Water Sci. Tech., 34 (1996) 29-35.
51 S. Sauv, W. Hendershot and H.E. Allen, Environ. Sci. Technol., 34 (2000)
1125-1131.
52 R.G. Luthy, G.R. Aiken, M.L. Brusseau, S.D. Cunningham, P.M. Gschwend, J.J.
Pignatello, M. Reinhard, S.J. Traina and J.C. Westall, Environ. Sci. Technol., 31
(1997) 3341-3347.
53 A. Delle Site, J. Phys. Chem. Ref. Data, 30 (2001) 187-439.
239
Chapter8
8.1 INTRODUCTION
Xphl Xh2
C~ = f -Ct (8.4)
242
with fu (fraction unbound):
fu = 1/(1 +K fup Pt) (8.5)
in which: Cbd is the concentration of bound analyte; C, is the concentration of
unbound analyte; Ct is the total concentration of analyte; P is the concentration
of unoccupied protein; Pt is the total protein concentration; fu is the fraction of
the analyte that is unbound; f is the fraction of the total number of binding
sites unoccupied; and K a is the association constant.
These equations and abbreviations, taken from Ref. [28], assume that there
is only one binding site, while in practice a protein molecule can have more
binding sites with different affinities for the analyte [29]. Information about the
unbound analyte is particularly important in understanding and modelling
kinetic behaviour, while information on the bound fraction is relevant in quanti-
fying drug-receptor interactions. The analysis of binding experiments often
assume that the free concentration of analyte is equal to the amount added [30]
and the affinity constant K. is then derived from a dose response curve with the
total concentration on the x-axis. However, in the analysis of drug-receptor
binding, information on the freely dissolved concentration is needed in those
cases where the unbound concentration of the agonist decreases due to binding
and the discrepancy between the total and free concentration is not the same in
each concentration.
So, binding affinities can be calculated by measuring the bound fraction and
in some cases, information on the free fraction is needed. As an alternative,
binding affinities or partition coefficients can also be derived from measuring
only the unbound concentration as a function of the increasing concentration of
the second phase (ph 2) or protein (Pt). Assuming a 100% mass balance, Kph2,phl
from Eq. (8.1) and Ka from Eq. (8.5) can be calculated from:
Cu (Cfie0) C=
t (8.6)
with fu = 1/(1 + Kph2,,h' Mph2 /V)
243
8.2 TECHNIQUES FOR MEASURING FREELY DISSOLVED
CONCENTRATION AND ANALYTE BINDING
TABLE 8.1
Methods used to measure freely dissolved concentrations and or partition/binding phenomena
Method General reference
Methods used in drug research
Equilibrium dialysis [29]
Ultrafiltration [29]
Ultracentrifugation [29]
Affinity chromatography (protein stationary phases) [29]
High performance size exclusion chromatography [29]
Capillary electrophoresis [29]
Fluorescence spectroscopy [29]
Methods used in environmentalresearch
Dynamic headspace analysis [43-45]
Static headspace analysis [46]
Reversed phase cartridge [54]
Equilibrium dialysis [9]
Semipermeable membrane devices (SPMD) [55,56]
Solid phase partitioning (empore disk) [57]
Solid phase micro-extraction (SPME) [39,41,49,50]
Thin-film solid-phase extraction [58]
244
for the determination of freely dissolved concentrations of drugs and serum-
protein binding include equilibrium dialysis, ultra-filtration and ultra-centri-
fugation and gel filtration [29]. Other techniques include chromatographic
methods, solid phase extractions and headspace analyses (see Table 8.1). Each of
these methods has its own advantages, disadvantages and limitations. It is not
the aim of this chapter to give a detailed analysis and evaluation of all available
methods, so the overview is kept short. More detailed information can be found
in review papers, e.g. [29].
Many methods are available to make a physical separation between the two
phases and thereby are very suitable to determine binding affinities. Neverthe-
less, only a few can also be used to determine the free concentration in plasma,
for example. The major problem of a technique like equilibrium dialysis is that
the plasma should equilibrate with the aqueous phase on the other side of the
membrane. Besides the kinetic drawbacks, there still remains the issue of
increased volume. The addition of a volume of competing phase on the other side
of the membrane can influence the equilibrium between the aqueous phase and
the protein bound fraction in the serum sample. Thus, the measured free
concentration is not representative of the free concentration of the non-dialysed
serum sample. Similar drawbacks exist for most other phase-separation tech-
niques, although it must be mentioned that the extent to which bias is
introduced by the physical separation greatly depends on the affinity constants
and volumes used. It should be emphasised that this experimental bias is only of
influence for the free concentration measurements, but not for binding
coefficients for which the analyte is determined in both phases.
For any method of measuring the true freely dissolved concentration, the key
is not to perturb in any way the equilibrium that exists between the aqueous
dissolved fraction and the bound fraction. In principal, this is an impossible
assignment, since almost every measurement either extracts, concentrates or
dilutes the sample during its treatment. Nevertheless, minimising this effect of
the sample treatment by, e.g., extracting a negligible amount, diluting with a
negligible volume (dialysis), or concentrating only a negligible volume (ultra-
filtration), will have only a negligible effect on the measurement of the free
concentration.
In this chapter we focus on the application of negligible depletion solid phase
microextraction (nd-SPME), which is an example of an extraction technique
based on earlier described negligible influence measurements. The nd-SPME
technology, which we introduced in 1996, is rather new and therefore the
mechanisms, advantages and drawbacks will be discussed.
245
oped by Pawliszyn and co-workers [31] as a simple extraction method with many
advantages over classical solid phase extractions. The principle, theoretical
aspects and experimental details of the method are described in several review
papers [32-35] and others discuss applications within different fields such as in
environmental analyses [32] and analyses of drugs in biological samples [36-38].
In the early stages, the development and application of SPME was mainly
focused on its capacity as an extraction technique. Much attention was given to
optimising recoveries or extraction efficiencies. Later on, it was realized that the
technique also represented a simple method for measuring freely dissolved
concentrations in complex matrices [16,27,39-41]. Based on the assumption that
only the freely dissolved fraction of a chemical partitions to the fiber and under
conditions where the extracted amount is negligibly small in comparison to the
total freely dissolved amount in a sample, the concentration on the fibre is in
principle directly proportional to the freely dissolved concentration. This specific
application of SPME is also referred to by Vaes et al. [39] as "negligible
depletion" SPME; the principle of nd-SPME is illustrated in Fig. 8.1.
The basic condition that needs to be met for nd-SPME is that the sampling
does not (or hardly) affect the amount present in the aqueous phase and,
therefore, the equilibrium between the bound and unbound fraction remains
unperturbed. The concentration on the fiber (Cf) is then directly proportional to
Cu. Additional conditions that need to be met are [39]:
- there is equilibrium between C. and Cb;
- the uptake kinetics is not influenced by the matrix;
- the matrix does not bind so substantially to the fiber that the concentration
becomes higher than expected based on Cu or that the characteristics of the
fiber are affected;
- the depletion of the amount present in the aqueous phase should be neglig-
ible; Vaes at al. [39] used a 5% limit for the depletion, which is represented by
Eq. (8.8). These conditions are discussed in more detail in Section 8.3.
Cf Vf < 0.05 C V (8.8)
with
Cf /C = Kfw
where Cf is the concentration in the fiber; Vf is the volume of the fiber; Cu is the
unbound concentration at equilibrium; V is the volume of the aqueous phase;
and KfW is the fiber-water partition coefficient.
The advantage of nd-SPME is that the sampling is relatively simple and that
no equilibrium is needed, assuming that the kinetics are independent of the test
conditions. The approach became possible due to the very small volume of the
SPME fiber as sampling phase.
SPME is also applied in more complex systems including other phases (see
Fig. 8.2). Headspace SPME is well known and applied in several studies [32] with
the advantage of clean samples. In relation to free concentration measurements,
246
., .
IeorlSPME fiber
Left: Fig. 8.1. Principle of negligible solid phase micro-extraction (nd-SPME) for measuring
freely dissolved concentrations of analytes [39].
Right: Fig. 8.2. Solid phase micro-extraction in a multi compartment system.
Although there are many studies and review publications [36,37] on the applica-
tion of SPME in biomedical analyses, only a few studies focus on the measure-
ment of freely dissolved concentrations in biological samples. The negligible
247
depletion SPME method (nd-SPME) was developed specifically to measure
freely dissolved concentrations [16,39]. The first study presented data for serum
protein binding of aniline, nitrobenzene, 4-chloro-3-methylphenol and 4-n-
pentylphenol [39]. Fractions freely dissolved were calculated from plots of total
versus free concentrations of the test chemical at one fixed protein concentra-
tion. A dialysis experiment was performed to check for a disturbance of the
analysis by direct binding of protein to the fiber. Peak areas of the SPME fibers
exposed on both sides from the dialysis membrane that separated the protein
solution from the buffer solution were the same, which shows that under these
test conditions no influence of binding of protein on the fiber is observed. The
same approach was applied to measure freely dissolved concentrations in
suspensions that contain microsomes, cells and membrane vesicles [16]. Koster
et al. [47] applied direct immersion SPME to extract the local anaesthetic
lidocaine from human plasma. The protein binding affinity was calculated from
a plot of extraction yield against the dilution factor of the protein solution using
equilibrium partitioning according to Eq. (8.9).
nf = KfVfno/(KfVf + KpNp + Vpl) (8.9)
where nf is the amount of drug in the fiber coating, Kf is the fiber-matrix
partition constant, n0 is the total amount of drug in the sample, Kp is the protein
association equilibrium constant, N is the number of available sites for protein-
drug binding and Vf and Vp; are the volume of the fiber coating and the plasma
sample, respectively. In this case, depletion of the bound and free fraction in the
sample is allowed because the analysis is performed at equilibrium. If Kf is
known, Kp N can be derived from Eq. (8.9) by varying Vp1.
The mathematical treatment of the analysis results by Koster et al. [47]
immediately points out that total concentration measurements with SPME in
samples where protein binding can play an important role, are doomed to fail.
High binding affinities will decrease nf and thereby negatively influence extrac-
tion recovery (ratio of nf and no). Moreover, interindividual variation in concen-
trations of plasma proteins can give interindividual variations in measured
plasma concentrations of the analyte, that are only a consequence of differences
in protein concentration instead of analyte concentrations. One elegant way to
circumvent this problem by GC-MS and isotopically labelled chemicals has been
described by Poerschmann et al. [40]. They used a deuterated analyte as internal
standard, which behaves according to the same affinity constant or partition
coefficient as the analyte of interest. From the ratio of the abundance of the ions
of both the analyte as the deuterated internal standard, the total concentration
can be determined.
For free concentrations, the advantage of the negligible depletion method is
that one calibration curve can be used to directly derive the freely dissolved
concentration and that, in principle, measurements can be performed under
non-equilibrium conditions. This gets even more important in headspace SPME
in which four phases are involved (fiber, headspace, protein and the aqueous
248
phase with the freely dissolved concentration). Yuan et al. [48] discussed the
measurement of protein binding via headspace SPME and pointed to the calibra-
tion problem and the effects of depletion on the calibration curve of concentra-
tion on the fiber versus the freely dissolved concentration. They demonstrated
that "for the three-phase system the amount of the analyte partitioned in the
headspace could be ignored only in certain circumstances, where the Henry's
law constant and the ratio between headspace volume and sample volume are
sufficiently small" [48].
Kopinke et al. [41] pointed at the possibility of applying SPME in the
measurement of sorption coefficients to humic acids because only the free solute
can interact with the fiber coating. This same group measured sorption coeffi-
cients to dissolved organic matter (DOM) of a series of organotin compounds [49]
using headspace as well as direct immersion SPME. Although some fouling of
the fibers occurred when applying them in a solution, these phenomena did not
affect the extraction. Sorption coefficients measured in headspace and solution
SPME did not show significant differences, although the headspace SPME was
preferred because fiber fouling can be completely excluded. In this same study,
the influence of DOM on the extraction kinetics to the fiber was investigated
with the idea that kinetic measurements would be less time-consuming than
equilibrium measurements. The addition of DOM did not affect the kinetics of
the partition process.
Using the negligible depletion SPME method, Urrestarazu-Ramos et al. [50]
measured sorption coefficients to Aldrich humic acids of penta and hexachloro-
benzene, PCB 77 and DDT. The sorption coefficients were derived from mea-
surements of the freely dissolved concentration at different humic acid
concentration using Eq. (8.6). PDMS fibers were used and the measurements
were performed in the aqueous solution and in the kinetic uptake phase. Figure
8.3 gives plots of the decrease in free concentration of PCB77 with increasing
0
05
Z
_u
1B_
0 5 10 15 20 25 30
concentration of humic acids (mg/l) CDOC
Fig. 8.3. Decrease in free concentration (plotted as freely dissolved fraction) as a function of the
humic acid concentration (mg/1) for PCB#77 (from [50]).
249
humic acid concentration. The advantage of this procedure is that no absolute
calibration of the concentration is needed. In principle, the partition coefficient
can be calculated based on the response of the analytical detection method. The
sorption coefficients were very close to reported literature values determined
with other techniques [50]. In a study of freely dissolved concentrations of
hydrophobic chemicals in soil and in artificial human digestive mixtures, Oomen
et al. [51] observed a slight increase in free concentrations of PCB52 with
increasing protein concentrations. Although the conditions in this study were
rather extreme (very low free fractions of less than 1%), these observed
phenomena may point, as suggested by the authors, to an increased flux in the
diffusion layer due to desorption from the matrix.
In a recent review Ulrich et al. [38] discussed the possibility of using SPME
as a method for measuring freely dissolved concentrations in biomedical analy-
sis. He addressed the topics of fouling of the fiber with proteins and the effects of
the matrix on the kinetics of uptake in the fiber. The problem of fouling of the
fiber has also been discussed in a recent review from Lord and Pawliszyn [52].
Ulrich [38] concluded that headspace SPME should be applied whenever
possible in SPME of body fluids. He mentioned that "the burden of the fiber with
proteins, for instance is considerable decreased. The lifetime of the fiber is
increased because irreversible damage is delayed. A reversible change of
extraction properties of the coating is also avoided and, therefore, the precision
of the method is improved". The introduction of disposable fiber material by
Mayer et al. [42,53] may partly reduce potential fouling problems, because the
fiber is used only once.
250
For the majority of applications, direct immersion nd-SPME continues to be
a good option for measuring freely dissolved concentrations. The conditions of
nd-SPME should, however, be carefully examined for each application to ensure
that the necessary requirements are met. Information about the freely dissolved
concentrations of chemicals is still of utmost importance in understanding the
fate of chemicals in the environment and exposure in toxicology. At present, the
availability of measurement techniques for the determination of freely dissolved
concentration limits its use. The methods that are now in use each have their
advantages, disadvantages [29] and limitations. As with all new methods,
further development to enable the routine measurement of free concentrations
will take time.
REFERENCES
251
26 E.M.J. Verbruggen, W.H.J. Vaes, T.F. Parkerton and J.L.M. Hermens, Environ. Sci.
Technol., 34 (2000) 324.
27 J.L.M. Hermens, A.P. Freidig, E. Urrestarazu-Ramos, W.H.J. Vaes, W.M.G.M. van
Loon, E.M.J. Verbruggen and H.J.M. Verhaar, in: PersistentBioaccumulative Toxic
Chemicals. ACS Symposium Series, Vol. 773, American Chemical Society,
Washington, DC, 2000.
28 M. Rowland and T.N. Tozer, Clinical Pharmacokinetics:Concepts and Applications.
Williams and Wilkens, Baltimore, MD, 1998.
29 J. Oravcova, B. Bohs and W. Lindner, J. Chromatogr. B, 677 (1996) 1.
30 H.J. Motulsky, Analyzing Data with GraphPad Prism, GraphPad Software Inc., San
Diego, CA, 1999.
31 J. Pawliszyn, Trends Anal. Chem., 14 (1995) 113.
32 J. Pawliszyn, Solid Phase Microextraction: Theory and Practice. Wiley-VCH, New
York, 1997.
33 D. Louch, S. Motlagh and J. Pawliszyn, Anal. Chem., 64 (1992) 1187.
34 H. Lord and J. Pawliszyn, J. Chromatogr.A, 885 (2000) 153.
35 J. Pawliszyn, J. Chromatogr. Sci., 38 (2000) 270.
36 N.H. Snow, J. Chromatogr.A, 885 (2000) 445.
37 G. Theodoridis, E.H. Koster and G.J. de Jong, J. Chromatogr.B, 745 (2000) 49.
38 S. Ulrich, J. Chromatogr.A, 902 (2000) 167.
39 W.H.J. Vaes, E. Urrestarazu Ramos, H.J.M. Verhaar, W. Seinen and J.L.M.
Hermens, Anal. Chem., 68 (1996) 4463.
40 J. Poerschmann, Z. Zhangh, F.-D. Kopinke and J. Pawliszyn, Anal. Chem., 69 (1997)
579.
41 F.-D. Kopinke, J. Porschmann and M. Remmler, Naturwissenschaften, 82 (1995) 28.
42 P. Mayer, W.H.J. Vaes, F. Wijnker, K.C.H.M. Legierse, R. Kraaij, J. Tolls and J.L.M.
Hermens, Environ. Sci. Technol., 34 (2000) 5177.
43 C. Yin and J.P. Hassett, Environ. Sci. Technol., 20 (1986) 1213.
44 J.W. Sproule, W.Y. Shiu, D. Mackay, W.H. Schroeder, R.W. Russell and F.A.P.C.
Gobas, Environ. Toxicol. Chem., 10 (1991) 9.
45 M. Horstmann and M.S. McLachlan, Environ. Sci. Technol., 26 (1992) 1643.
46 J. Resendes, W.Y. Shiu and D. Mackay, Environ. Sci. Technol., 26 (1992) 2381.
47 E.H. Koster, C. Wemes, J.B. Morsink and G.J. de Jong, J. Chromatogr.B, 739 (2000)
175.
48 H. Yuan, R. Ranatunga, P.W. Carr and J. Pawliszyn, Analyst, 124 (1999) 1443.
49 J. Poerschmann, F.-D. Kopinke and J. Pawliszyn, Environ. Sci. Technol., 31 (1997)
3629.
50 E. Urrestarazu-Ramos, S.N. Meijer, W.H.J. Vaes, H.J.M. Verhaar and J.L.M.
Hermens, Environ. Sci. Technol., 32 (1998) 3430.
51 A. Oomen, J.P. Groten, D.T.H.M. Sijm, J. Tolls and A.J.A.M. Sips, Environ. Sci.
Technol., 34 (2000) 297.
52 H. Lord and J. Pawliszyn, J. Chromatogr.A, 902 (2000) 17.
53 P. Mayer, W.H.J. Vaes and J.L.M. Hermens, Anal. Chem., 72 (2000) 459.
54 P.F. Landrum, S.R. Nihart, B.J. Eadis and W.S. Gardner, Environ. Sci. Technol., 18
(1984) 187.
55 J.N. Huckins, M.W. Tubergen and G.K. Manuweera, Chemosphere, 20 (1990) 533.
56 G.S. Ellis, J.N. Huckins, C.E. Rostad, C.J. Smitt, D.D. Petty and P. MacCarthy,
Environ. Toxicol. Chem., 14 (1995) 875.
57 A.P. Freidig, E. Artola Garicano, F.J.M. Busser and J.L.M. Hermens, Environ.
Toxicol. Chem., 17 (1998) 998.
58 J.B. Wilcockson and F.A.P.C. Gobas, Environ. Sci. Technol., 35 (2001) 1425.
252
Chapter 9
9.1 INTRODUCTION
The analytical procedure for complex samples consists of several steps typically
including sampling, sample preparation, separation, quantification, statistical
evaluation, and decision making (see Fig. 9.1). Each step is critical for obtaining
correct and informative results. The sampling step includes deciding where to
get samples that properly define the object or problem being characterized, and
choosing a method to obtain samples in the right amounts. The objective of the
sample preparation step is to isolate the components of interest from a sample
matrix, because most analytical instruments cannot handle the matrix directly.
Sample preparation involves extraction procedures and can also include "clean
up" procedures for very complex "dirty" samples. This step must also bring the
analytes to a suitable concentration level for detection and therefore sample
preparation methods typically include enrichment. During the separation step
of the analytical process, the isolated complex mixture containing target
analytes is divided into its constituents, typically by means of a chromatographic
ComprehensiveAnalytical Chemisty XXXVII
J. Pawliszyn (Ed.) 253
© 2002 Elsevier Science B.V. All rights reserved
-
254
much more difficult to copuple sampling and sample preparation steps,
primarily because the current state of the art in sample preparation techniques
employs multi-step procedures involving organic solvents. These characteristics
make it difficult to develop a method that integrates sampling and sample
preparation with separation methods, for the purposes of automation. The
result is that over 80% of analysis time is currently spent on sampling and
sample preparation steps for complex samples.
One of the reasons for the slow progress in the area of sample preparation is
that the fundamentals of extraction involving natural, frequently complex
samples are much less well developed and understood compared to more
physicochemically simpler instrumental systems used in separation and
quantification steps such as chromatography and mass spectrometry. This
situation creates a feeling that rational design and optimisation of extraction
systems is not possible. Therefore, the development of sample preparation
procedures is frequently considered to be an "art" not a "science".
BatchEquilibrium
and Pre-equilibrium
255
supporting material and when the sample is passed through, analytes are
retained on the bed. Large volumes of sample can be passed through a small
cartridge and the flow through a well-packed bed facilitates rapid mass transfer.
The extraction procedure is followed by re-extraction of analytes into a small
volume of solvent resulting in high enrichment. This strategy is used in sorbent
trap techniques and Solid Phase Extraction (SPE) [1]. Alternatively, a sample
(typically a solid sample) can be packed in the bed and the extraction phase can
be used to remove and transport the analytes to the collection point. In super-
critical fluid extraction (SFE) compressed gas is used to wash analytes from the
sample matrix, while in purge and trap atmospheric pressure inert gas performs
the same function. In dynamic solvent extraction, such as Soxhlet, the solvent
continuously removes the analytes from the matrix at the boiling point of the
solvent. In more recent hot solvent extraction techniques smaller volumes of
organic solvent are used to accomplish higher enrichment at the same time,
because of increased solvent capacity and eluting capability at high tempera-
tures and pressures [2].
Alternatively, non-exhaustive approaches can be designed based on equilib-
rium, pre-equilibrium and permeation principles [3]. Equilibrium non-exhaust-
ive techniques are fundamentally analogues to equilibrium exhaustive tech-
niques; however the capacity of the extraction phase is smaller, and in most
cases is not sufficient to remove the majority of analytes from the sample matrix.
This is caused by use of a small volume of the extracting phase relative to sample
volume (microextraction, such as solvent microextraction or solid phase micro-
extraction (SPME) [41 or a low sample matrix-extraction phase distribution con-
stant, such as is typically encountered in gaseous headspace techniques [5].
Pre-equilibrium conditions are accomplished by breaking the contact
between the extraction phase and the sample matrix before the point of equilib-
rium with the extracting phase has been reached. The devices frequently used
are identical to the microextraction systems, but extraction times are shorter. In
permeation techniques, such as membrane extraction [6], continuous steady-
state transport of analytes through the extraction phase is accomplished by
simultaneous re-extraction of analytes. The membrane extraction can be
exhaustive by designing appropriate membrane modules and optimising the
sample and striping flow conditions [7] or it can be optimised for throughput and
sensitivity in a non-quantitative open bed approach [8].
In addition to classification of methods based on more fundamental
principles as discussed above, it is also instructive to divide techniques according
to particular characteristics. For example, there has recently been a trend
towards solvent-free techniques (see Fig. 9.3) [9]. This is an important direction,
not only because it addresses health and pollution prevention issues, but also
because in most cases such approaches are easier to implement for on-site
monitoring in field conditions. This direction has been creating a lot of interest
and research opportunities recently and it is expected to continue to be a very
active area in the near future, The most promising solventless techniques are
256
ISolvent-free Sample Preparation Methodsl
headspace,
sorbent and approaches. Supercritical Fluid Extraction
membrane
traon-siteimExtraction Etractions.
among
exStatic Staticon process. In all tech-
Cartridge|H Direct
Dynamic | H Dynamic Dick Headepaco
In-tube
9.2 FUNDAMENTALS
As the above discussion and Fig. 9.2 indicate, there is a fundamental similarity
among extraction techniques used in the sample preparation process. In all tech-
niques the extraction phase is in contact with the sample matrix and analytes
are transported between the phases. For exhaustive techniques the phase ratio
is higher and geometries are more restrictive compared with non-exhaustive
approaches required to ensure the quantitative transfer of analytes. The ther-
modynamics of the process is defined by the extraction phase/sample matrix dis-
tribution constant. It would be instructive to consider in more detail the kinetics
of processlveoccurring at the extraction phase/sample matrix interface since this
defines the time of the analytical procedure. In many cases the analytes are
re-extracted from the extraction phase, but this step is not discussed here since
this process is analogues and much simpler in principle compared to removing
analytes from a more complex sample matrix. The main objective of this chapter
is to outline the common fundamental principles among various extraction tech-
niques to facilitate a better understanding of selection criteria for appropriate
techniques, device geometries and operational conditions.
9.2.1 Thermodynamics
257
Fig. 9.4. Partitioning between aqueous sample matrix and organic extraction phase.
=
Kes aa = CJCs (9.1)
defines the equilibrium conditions and ultimate enrichment factors achievable
in the technique, where ae and as are the activities of analytes in the extraction
phase and matrix correspondingly, which can be approximated by the appropri-
ate concentrations. Figure 9.4 shows a schematic example of the system for
liquid-liquid extraction. For solid extraction phase adsorption, equilibria can be
explained by the following equation:
Ke = SCs (9.2)
where Se is the solid extraction phase surface concentration of adsorbed
analytes. The above relationship is similar to Eq. (9.1) with the exception that
extraction phase concentration is replaced with surface concentration. The Se
term in the numerator indicates that the sorbent surface area available for
adsorption must also be considered. This complicates the calibration at equilib-
rium conditions because of displacement effects and the non-linear adsorption
isotherm [11]. The above equations can be used to calculate the amount of
analyte in the extraction phase at equilibrium conditions [4]. For example, for
equilibrium liquid microextraction techniques and large samples, including
direct extraction from the investigated system, the appropriate expression is
very simple:
n = KesVC (9.3)
where K,, is the extraction phase/sample matrix distribution constant, Ve is the
volume of the extraction phase and C. is the concentration of the sample. In het-
erogeneous samples (headspace, immiscible liquids and solids) the components
of the sample partition in the multiphase system are less available for extraction.
This effect depends on analyte affinity and volume of the competing phases and
can be calculated if appropriate volumes and distribution constants are known
using equations discussed in Ref. [4]. The distribution constants depend on vari-
ous parameters including temperature, pressure and sample matrix conditions
such as pH, salt and organic component concentration. All these parameters
need to be optimised for maximum transfer of analytes to the extraction phase
during the method development process. In practice, however, kinetic factors
defined by the dissociation constants, diffusion coefficient and agitation condi-
tion frequently determine the amount of extracted analytes from complex sam-
258
pies since the overall rates are slow and therefore extraction amounts for limited
time experiments do not reach equilibrium values.
9.2.2 Kinetics
9.2.2.1 Liquid-liquidextraction
Let us take the simple case of static extraction of water with organic solvent as
illustrated in Fig. 9.4 to consider the effect of different parameters on extraction
kinetics. The appropriate equation showing concentration profiles in each of the
phases can be obtained by solving Fick's Second Law differential equation for
appropriate boundary conditions:
C, = D a2C(x' t) (9.4)
at ax2
If no convection is present in the system, the distribution constant is defined by
Eq. (9.1) and the two phases are placed in contact with each other at t = 0 then
the solution can be found using Laplace Transform approach to be for aqueous
sample phase (x < 0):
z +erf(zVt-/)
C (x,t)=CO Kes (9.5a)
1+
Kes
259
Boundary
Aqueous organic
>t D\
_ \ _
Fig. 9.5. Concentration profiles at the interface between infinite volume sample and extraction
phases for analyte characterized by identical diffusion coefficients in aqueous and organic phase
of 10-5 cm2/s. The profiles correspond to 1 s (A), 10 s (B), 100 s (C) and 1000 s (D) after merging
both phases.
260
---
i
Flow
Convection
A(EP,B)
O Particle Core
O Organic Material
Fig. 9.6. Processes involved in extraction of heterogeneous samples containing porous solid
particles. The terms in the figure are discussed in the text.
restrictor is attached to maintain the pressure at the end of the vessel. During
the process, the extraction phase continuously removes analytes from the
matrix, which are then transferred to the collection vessel after the expansion of
the fluid. This leaching process is very similar to chromatographic elution with
packed columns, particularly to the frontal method. The main difference is that
in sample preparation analytes are dispersed in the matrix at the beginning of
the experiment, while in chromatographic frontal analysis a long plug is intro-
duced into the column at the initial stage of the separation process. The principal
objective of the extraction is to remove analytes from the vessel in the least
amount of time, requiring elution conditions under which the analytes are
unretained. In chromatography, on the other hand, the ultimate goal is to sepa-
rate components of the sample, which requires retention of analytes in the
column. Another major difference is that the packing matrix is usually well char-
acterized in chromatography, but in sample preparation it is often unknown.
One way to develop a mathematical model for this extraction approach is to
establish the mass balance equation for the system after careful consideration of
the individual mass transfer steps occurring during the extraction process (see
Fig. 9.3) and specific boundary conditions [14]. Extensive investigations on
similar topics have already been conducted by engineers who have studied the
mass transfer in porous media [12] and chromatographers [13]. In these studies
the relationship between various matrix parameters and flow conditions on the
elution profile were described mathematically, and verified experimentally. In
chromatography this relationship is usually described as contributions from
each of the mass transfer steps to the height equivalent to a theoretical plate
(HETP). The overall performance of the system can be defined as the sum of the
relevant individual components judiciously selected to reflect the most
significant individual steps present in the elution process. For the purpose of
this discussion, this approach is adopted to develop a model for extraction
kinetics in flow-through techniques.
261
The effect of slow desorption kinetics of analytes from the matrix on the
elution profile, can be described as the contribution to the HETP [8], hRK:
where k is the partition ratio, kd is the dissociation rate constant of the analyte-
matrix complex of reversible process, kh is the ratio of the intraparticulate void
volume to the interstitial void space and is expressed as:
ko i1-c e) (9.7)
h c =2 kI d2 ue (9.9)
3 ( +1)2 D,
where is the tortuosity factor for the porous particle and D. is the diffusion
coefficient of the analyte in the material filling the pores, which in most practical
cases will be an extraction phase and therefore Dp = De, where De is the diffusion
coefficient of the analyte in the extraction phase. This contribution can be quite
important considering the relatively large particle size (about 1 mm) of
262
environmental matrices, and it becomes particularly important when the pores
are filled with dense organic material, such as humic matter rather than the
extraction phase.
In the flowing bulk of the fluid an analyte experiences resistance to mass
transfer associated with eddy diffusion (random paths of the analytes through
the vessel filled with the particles) which is given by hED:
where yM is the obstruction factor which characterise the structure of the matrix.
This component is expected to be small. The analyte concentration profile
generated during the experiment as a function of time C(x,t) can be represented
using the equation which describes the dispersion of a plug of finite width [9]:
where L is the length of the vessel, COis the initial concentration of analyte in the
extraction vessel and a is the mean square root dispersion of the band expressed
as:
= Ht< (9.14)
where m(t) is the extracted mass of analyte and m0 is the total amount of analyte
263
in the vessel at the beginning of the experiment. We will refer to this function as
the "time elution profile", emphasizing the similarity of the extraction process
in this simple case to chromatographic elution.
The resulting function describes a process where elution and mass transfer
between the phases occur simultaneously. In this discussion we will refer to this
function as the "extraction time profile" to emphasize the point that in a
majority of extraction cases these two processes are expected to be combined.
F(t) describes the kinetics of the process, which defines the release rate of
analyte from the sample matrix and can include for example: the matrix-analyte
complex dissociation rate constant, the diffusion coefficient, the time constant
that describes swelling of the matrix that will facilitate removal of analyte, or a
combination of the above. Detailed discussion, graphical representations and
applications of this model to describe and/or investigate processes in supercriti-
cal fluid extraction have been described in detail elsewhere [17,18].
The above conclusion can be stated in a more general way. Convolution
among functions describing individual processes occurring during the extrac-
tion, describes the overall extraction process and represents a unified way to
describe the kinetics of these complex processes. The exact mathematical
solution to the convolution integral is frequently difficult to obtain, but graphi-
cal representation of the solution can be calculated using Fourier Transform or
264
numerical approaches. Frequently, it is possible to incorporate mathematical
functions that describe a combination of the unit processes. In the example of
the flow-through system discussed above, the elution function describes the
effect of porosity and analyte affinity towards the extraction matrix on the
extraction rate. It should be emphasised that the convolution approach consid-
ers all processes equivalently. In practice, however, a small number-and
frequently just one unit process-controls the overall rate of extraction so the
equations can be simplified by considering this fact.
Determination of the limiting step is not possible exclusively by qualitative
agreement with the mathematical model since the effect on recovery of most of
the unit processes has an exponential decay nature. To properly recognise them,
quantitative agreement and/or the effect of extraction parameters needs to be
examined. Identification of the limiting process provides a valuable insight into
the most effective approach to optimise the extraction.
265
9.3.1 Flow-through techniques
For homogeneous samples and flowing fluid extraction phase, the description of
the extraction process is much simpler and can be based directly on the chro-
matographic theory for liquid stationary phases. Let us consider another case of
the flow-through system where the extraction phase is dispersed as a thin layer
inside the extraction bed and the sample flows through the cartridge. The bed
can be constructed of a piece of fused silica capillary, internally coated with a
thin film of extracting phase [25] (a piece of open tubular capillary GC column;
in-tube SPME) [26], or the bed may be packed with extracting phase dispersed
on an inert supporting material (SPE cartridge). In these geometric arrange-
ments, the concentration profile along the axis x, of the tubing containing the
extracting phase as a function of time t, can be described by adopting the expres-
sion for dispersion of a concentration front:
x-
C(x,t) = 0.5C erf (9.17)
where u, is linear velocity of the sample through the tube, k is the partition ratio
defined as:
k =K VI (9.18)
= Ht (9.19)
1+k
266
sample, and inversely related to the partition ratio. For the in-tube SPME and
short capillaries with a small dispersion, the minimum extraction time for
in-tube SPME at equilibrium conditions can be assumed to be similar to the time
required for the centre of the band to reach the end of the capillary:
L 1+Ks
te -= \ v (9.20)
U
where L is the length of the capillary holding the extraction phase. For packed
bed extractors typical for SPE techniques, analogous equations can be devel-
oped. In that case the calculated time corresponds to the maximum extraction
time before breakthrough occurs. As expected, the extraction time is propor-
tional to the length of the capillary and inversely proportional to the linear flow
rate of the sample. Extraction time also increases with an increase in the extrac-
tion phase/sample distribution constant and with the volume of the extracting
phase but decreases with an increase of the void volume of the capillary.
9.3.2 Batch techniques
Coupling equations for systems involving convection caused by flow through a
tube as discussed above, are frequently not available for other means of agitation
and other geometric configurations. In these cases the most successful approach
is to consider the boundary layer formed at the interface between the sample
matrix and the extraction phase. Independent of the agitation level, fluid
contacting the fibre's surface is always stationary, and as the distance from the
fibre surface increases, the fluid movement gradually increases until it
corresponds to bulk flow in the sample. To model mass transport, the gradation
in fluid motion and convection of molecules in the space surrounding the fibre
surface can be simplified as a zone of a defined thickness in which no convection
occurs, and perfect agitation occurs in the bulk of the fluid everywhere else. This
static layer zone is called Prandtlboundary layer (see Fig. 9.8) [27].
,= nha
-rait-n
/ Dounuay Sample
i Layer
,
0,
a
0
A X
Pn-iti-n
Left: Fig. 9.7. Normalized concentration profiles for in-tube SPME calculated using the equation
discussed in the text.
Right: Fig. 9.8. Boundary layer model.
267
9.3.3 Boundary layer model
268
t =°
Absorption Adsorption
a b
Fig. 9.9. Extraction using absorptive (a) and adsorptive (b) extraction phases immediately after
exposure of the phase to the sample (t = 0) and after completion of the extraction (t = t,).
The only way to overcome this fundamental limitation of porous coatings is, as
Fig. 9.9 suggests, to use an extraction time much less than the equilibrium time,
so that the total amount of analytes accumulated onto the porous coating is
269
- Fiji m-
II"I i - pores
i bulk air
Ii movement
i
i
i
iI solid coating
surface (A)
I'- .-~ boundary layer
I
/, concentration profile
Fig. 9.10. Schematic of the diffusion based calibration model. The terms are defined in the text.
substantially below the saturation value. At saturation, all surface available for
adsorption is occupied. When performing such experiments, not only is it critical
to precisely control extraction times, but also convection conditions since they
determine the thickness of the diffusion layer. One way of eliminating the need
for compensation of differences in convection is to normalize (use consistent)
agitation conditions. For example, the use of a stirring means at a well defined
rotation rate in the laboratory, or fans for field air monitoring, will ensure
consistent convection [28,29]. The short time exposure measurement described
above has an advantage associated with the fact that the rate of extraction is
defined by diffusivity of analytes through the boundary layer of the sample
matrix, and corresponding diffusion coefficients, rather than by distribution
constants. This situation is illustrated in Fig. 9.10 for cylindrical geometry of the
extraction phase dispersed on the supporting rod.
The analyte concentration in the bulk of the matrix can be considered con-
stant when a short sampling time is used, and there is a constant supply of an
analyte via convection. These assumptions are true for most cases of sampling,
where the volume of sample is much greater than the volume of the interface,
and the extraction process does not affect the bulk sample concentration. In
addition, the solid coating can be treated as a perfect sink. The adsorption bind-
ing is frequently instantaneous and essentially irreversible. The analyte concen-
tration on the coating surface is far from saturation and can be assumed to be
negligible for short sampling times and relatively low analyte concentrations in a
typical sample. The analyte concentration profile can be assumed to be linear
from C, to C. In addition, the initial analyte concentration on the coating sur-
face (Ce) can be assumed to be equal to zero when extraction begins. Diffusion of
analytes inside the pores of a solid coating controls mass transfer from outer to
inner surface of the coating.
270
The function describing the mass of extracted analyte with sampling time
can be derived [30], which results in the following equation:
where n is the mass of extracted analyte over sampling time (t) in ng; Ds is the
gas-phase molecular diffusion coefficient; A is the surface area of the sorbent; 6
is the thickness of the boundary layer surrounding the extraction phase; B1 is a
geometric factor and C. is the analyte concentration in the bulk of the sample. It
can be assumed that the analyte concentration is constant for very short
sampling times and therefore Eq. (9.20) can be further reduced to:
271
sample temperature (Tg) increases, the kinematic viscosity also increases. Since
the kinematic viscosity term is present in the numerator of Re and in the
denominator of Sc, the overall effect on 6 is small. A reduction of the boundary
layer and an increase of the mass transfer rate for an analyte can be achieved in
at least two ways, i.e., by increasing the sample velocity and by increasing the
sample temperature. However, the temperature increase will reduce the solid
sorbent efficiency. As a result, the sorbent coating may not be able to adsorb all
molecules reaching the surface and therefore stop behaving as a zero sink for all
analytes.
Equations (9.21) and (9.23) indicate that the use of the headspace above the sam-
ple as an intermediate phase might be an interesting approach to accelerate
extraction for analytes characterised by high Henry constants. When a thin
extraction phase is used, the initial extraction rate and hence the extraction time
is controlled by the diffusion of analytes through the boundary layer present in
the sample matrix. The presence of a gaseous headspace facilitates rapid trans-
port into the extraction phase because of the high diffusion coefficients. To
increase the transport from the sample matrix into the headspace, the system
can be designed to produce a large sample/headspace interface. This can be
accomplished by using large diameter vials with good agitation, purge or even
spray systems. At room temperature only volatile analytes are transported
through the headspace. For low volatility compounds, heating of the sample is a
good approach, if loss in magnitude of the distribution constant can be accepted.
The ultimate approach is to heat the sample and cool the extraction phase at the
same time. Heating of the sample does not only increase the Henry constant, but
also induces convection of the headspace due to density gradients associated
with temperature gradients present in the system, resulting in higher mass
transport rates. The cooling of the sorbent increases its capacity. Collection of
analytes can be performed in the same vial [32] or can be separated in space simi-
larly as in the purge and trap technique. In the heating-cooling experiments,
both kinetic and thermodynamic factors are addressed simultaneously. This
technique is discussed in more detail in Chapter 13 on Solid Phase Micro-
extraction. Headspace approaches are also interesting since adverse affects asso-
ciated with the presence of solids, oily or high molecular weight interferences,
which can cause fouling of the extraction phase are eliminated.
272
z -
0 Z
Fig. 9.11. SPME/TWA approaches based on the in-needle fibre and in-needle coating.
developed based on this principle. For example, when the extracting phase in the
SPME device is not exposed directly to the sample, but is contained in a protec-
tive tubing (needle) without any flow of the sample through it (see Fig. 9.11), the
diffusive transfer of analytes occurs through the static sample (gas phase or
other matrix) trapped in the needle. The system consists of an externally coated
fibre with extraction phase withdrawn into the needle (Fig. 9.11). This geomet-
ric arrangement represent a very powerful method, capable of generating a
response proportional to the integral of the analyte concentration over time and
space (when the needle is moved through space) [33]. In this case, the only mech-
anism of analyte transport to the extracting phase is diffusion through the
matrix contained in the needle. During this process, a linear concentration pro-
file (shown in Fig. 9.11) is established in the tubing between the small needle
opening, characterized by surface area A and the distance Z between the needle
opening and the position of the extracting phase. The amount of analyte
extracted, dn, during time interval, dt, can be calculated by considering Fick's
first law of diffusion [4]
n AC
f (t)d (27)t)dx
273
opening, Z. It should be emphasized that Eqs. (9.24) and (9.25) are valid only in a
situation where the amount of analyte extracted onto the sorbent is a small
fraction (below RSD of the measurement, typically 5%) of the equilibrium
amount in respect to the lowest concentration in the sample. To extend integra-
tion times, the coating can be placed further into the needle (larger Z), the
opening of the needle can be reduced by placing an additional orifice (smaller A),
or a higher capacity sorbent can be used. The first two solutions will result in low
measurement sensitivity. An increase of sorbent capacity presents a more
attractive opportunity. It can be achieved by either increasing the volume of the
coating, or its affinity towards the analyte. An increase of the coating volume
will require an increase of the device size. Therefore the optimum approach to
increased integration time is to use sorbents characterized by large coating/gas
distribution constants. If the matrix filling the needle is other than sample ma-
trix then an appropriate diffusion coefficient and distribution constant should be
used in Eq. (9.27) as discussed below in the case of membrane extraction.
The capacity of the extraction phase for the difficult to extract analytes such as
polar or ionic species is frequently enhanced by introducing a derivatisation
step. The objective of this approach is frequently not only to convert the native
analytes into less polar derivatives that are extracted more efficiently, but also to
label them for better detection and/or chromatography. The most interesting
implementation of this approach is simultaneous extraction/derivatisation. In
this technique the derivatisation reagent is present in the extraction phase
during the extraction. The major advantage of this approach is that two steps are
integrated into one. There are two limiting cases describing the combination
between extraction and derivatisation. The first occurs when mass transfer to
the fibre is slow compared to the reaction rate. In this case Eq. (9.22), as
discussed above, describes the accumulation rate of analytes, assuming that the
derivative is trapped in the extraction phase.
In the second limiting case the situation is reversed in that the reaction rate
is slow compared to the transport of analytes to the extraction phase. In other
words, at any time during the extraction, the extraction phase is at equilibrium
with the analyte remaining in a well-agitated sample, resulting in uniform
reaction rate throughout the coating. This is a typical case for thinly dispersed
extraction phase, since the equilibration time for well-agitated conditions is very
short compared to a typical reaction rate constant. The accumulation rate of the
product in the extraction phase n/t can then be defined by:
n = VkrKe X C(t)dt (9.28)
where Cs is the initial concentration of analyte in the sample and kr is chemical
reaction rate constant. In the other words, when the sample is of large volume,
such as direct sampling in the field, the reaction and accumulation of analyte in
274
the extraction phase proceeds with the same rate as long as reagent is present in
an excess amount. It is worth noting that the rate is also proportional to the
extraction phase/sample matrix distribution constant. If the concentration
varies during the accumulation, the collected amount corresponds to the
integral over concentration and time, similarly as discussed above in the case of
time weighted average sampling. For limited sample volume, however, the
concentration of analyte in the sample phase decreases with time as it is
partitioned into the coating and converted to trapped product, resulting in a
gradual decrease of the rate. The time required to exhaustively extract analytes
from a limited volume can be estimated using experimental parameters [4].
Strip
Fig. 9.12. Membrane extraction at good sample agitation and stripping conditions. The terms are
defined in the text.
275
proportional to both the diffusion coefficient (D.) and the distribution constant
(Kes) and inversely proportional to b. De determines the rate of analyte migration
through the membrane and Kes the magnitude of the concentration gradient
generated in the membrane [341. This information can be used to calibrate the
extraction process a priori if these parameters are obtained from tables or
experimental data [35].
The concentration of unknown can be calculated by converting Eq. (9.29):
bn
C, - (9.30)
B2 ADeKe,,t (9
276
structures of analytes, extraction phase and matrix, which adds to the attrac-
tiveness of this approach. For equilibrium microextraction techniques the
extraction phase/sample matrix distribution constant is used to quantify the
concentration of analytes in the sample matrix (see Eq. (9.3)). For extraction
approaches controlled by the mass transfer in the boundary layer, calibration
can be based on the diffusion coefficient in the sample matrix at constant extrac-
tion time under well-defined agitation conditions. When the derivatisation reac-
tion kinetics controls the extraction rate, the rate constant can provide a means
of calibration. In some cases, as in membrane extraction, a combination of con-
stants defines the extraction rate and can be used for calibration as well. The
major argument against using this approach is that physicochemical constants
are affected by many experimental parameters, such as temperature and matrix
conditions. However the impact of temperature change can be compensated for.
This can be accomplished by monitoring the temperature and using correction
factors and therefore direct calibration for simple matrices is possible [36]. For
more complex matrices, internal standard or standard addition calibration,
applied routinely in exhaustive techniques to monitor recoveries, can be used to
compensate for matrix variations [37]. In future research, correlations between
distribution constants and simple measurements such as turbidity and pH may
be found to account for small matrix variations and therefore eliminate the need
for internal calibration for samples of known origin.
The advantages of non-exhaustive extraction are its fundamental simplicity
and fewer geometric restrictions, which facilitate a number of interesting on-site
implementations by integrating the sampling and sample preparation steps.
Additionally, more information can be obtained about the investigated system.
For example, it is possible to speciate and determine distribution of analyte in
multiphase systems since the extraction process does not disturb the equilibria
naturally present in the system. Therefore different forms of an analyte are
extracted and quantified according to their corresponding distribution constants
and/or diffusion coefficients.
A better understanding of the fundamentals of the extraction process
facilitates exploration of new opportunities, which makes sample preparation a
more vital and scientifically exciting part of the analytical process. The focus in
this discussion has been on chemical extraction, purposely neglecting extraction
approaches involving mechanical principles, which would introduce an addi-
tional level of complexity, although these would be interesting to explore in
future developments of the technology.
REFERENCES
1 E.M. Thurman and M.S. Mills, Solid Phase Extraction. Wiley, New York, 1998.
2 J. Dean, Extraction Methods for EnvironmentalAnalysis. Wiley, New York, 1998.
3 A. Handley, Extraction Methods in Organic Analysis. Sheffield Academic Press,
Sheffield, 1999.
4 J. Pawliszyn, Solid Phase Microextraction.Wiley-VCH, New York, 1997.
277
5 B.V. Ioffe and A.G. Vitenberg, Headspace Analysis and Related Methods in Gas
Chromatography.Wiley, New York, 1984.
6 J. Pawliszyn, Chapter 14 in: J. Pawliszyn (ed.), Sampling and Sample Preparation
for Fieldand Laboratory. Fundamentalsand New Directionsin Sample Preparation.
Comprehensive Analytical Chemistry XXXVII. Elsevier, Amsterdam, 2002, pp.
479-502.
7 K. Pratt and J. Pawliszyn, Anal. Chem., 64 (1992) 2101.
8 M. Yang, M. Adams and J. Pawliszyn, Anal. Chem., 68 (1996) 2782.
9 J. Pawliszyn, Trends Anal. Chem., 14 (1995) 113.
10 M. Adams, E. Otu, M. Kozliner, J. Szubra and J. Pawliszyn, Anal. Chem., 67 (1995)
212.
11 T. Gorecki, X. Yu and J. Pawliszyn, Analyst, 124 (1999) 643.
12 F.A.L. Dullien, PorousMedia. Academic Press, San Diego, 1992.
13 C. Horvath and H.J. Lin, J. Chromatogr., 149 (1978) 43.
14 J. Crank, Mathematics of Diffusion. Clarendon Press, Oxford, 1989.
15 J.C. Giddings, Anal. Chem., 35 (1963) 1999.
16 J.A. Cadzow and H.F. van Landingham, Signals, Systems, and Transforms. Prentice
Hall, Englewood Cliffs, NJ, 1985.
17 J. Pawliszyn, J. Chromatogr. Sci., 31 (1993) 31.
18 J. Langenfeld, S. Hawthorne, D. Miller and J. Pawliszyn, Anal. Chem., 67 (1995)
1727.
19 J. Langenfeld, S. Hawthorne, D. Miller and J. Pawliszyn, Anal. Chem., 65 (1993) 338.
20 B.E. Richter, B.A. Jones, J.L. Ezzell, N.L. Porter, N. Avdalovic and C. Pohl, Anal.
Chem., 68 (1996) 1033.
21 J.R.J. Pare, J.M.R. Belanger, K. Li and S.S. Stafford. J. Microcolumn Sep., 7 (1995)
37.
22 N. Alexandrgu and J. Pawliszyn, Anal. Chem., 61 (1989) 2770.
23 Z. Miao, Z. Zhang and J. Pawliszyn, J. Microcolumn Sep., 6 (1994) 459.
24 K.D. Bartle, T. Boddington, A.A. Clifford and N. Cotton, Anal. Chem., 63 (1991)
2371.
25 R. Eisert and J. Pawliszyn, Anal. Chem., 69, (1997) 3140.
26 R. Eisert and J. Pawliszyn, Crit. Rev. Anal. Chem., 27 (1997) 103.
27 A.D. Young, Boundary Layers. BSP Professional Books, Oxford, 1989.
28 F. Augusto, J. Koziel and J. Pawliszyn, Anal. Chem., 73 (2001) 481.
29 K. Sukola, J. Koziel, F. Augusto and J. Pawliszyn, Anal. Chem., 73 (2001) 13.
30 H.S.Carslaw and J.C. Jaeger, Conductionof Heat in Solids. Clarendon Press, Oxford,
1986.
31 J. Koziel, M. Jia and J. Pawliszyn, Anal. Chem., 72 (2000) 5178.
32 Z. Zhang and J. Pawliszyn, Anal. Chem., 67 (1995) 34.
33 M. Chai and J. Pawliszyn, Environ. Sci. Technol., 29 (1995) 693.
34 Y. Luo, M. Adams and J. Pawliszyn, Anal. Chem., 70 (1998) 19.
35 Y. Luo and J. Pawliszyn, Anal. Chem., 72 (2000) 1064.
36 P. Martos and J. Pawliszyn, Anal. Chem., 69 (1997) 206.
37 C. Grote and K. Levsen, in: J. Pawliszyn (ed.), Applications of Solid Phase Micro-
extraction. RSC, Cambridge, 1999.
278
Chapter10
10.1 OVERVIEW
10.1.1 Introduction
The term "headspace GC" is commonly used to describe static headspace but is
sometimes used to refer to dynamic headspace (purge and trap).
In the static headspace technique (Fig. 10.1), the sample is placed in a vial,
which is sealed and usually heated. The volatile analytes begin to partition
between the sample and the gas phase. Eventually an equilibration state is
reached. At this point, an aliquot of the gas phase is transferred to the GC.
With dynamic headspace (Fig. 10.2), an inert gas is bubbled through the
sample and the volatiles are transferred to an absorbent trap. The trap is heated
and the volatiles are released or desorbed and transferred to the GC. Finally the
Left: Fig. 10.1. Static headspace sampling. The sample is collected in a container with no space
over the liquid so that volatiles are not lost (A). An aliquot of the sample is placed in a headspace
vial that is sealed with a septum and a crimp-top vial (B). The volatiles in the liquid phase enter
the gas phase until a state of equilibrium is reached (C). An aliquot of the gas phase is removed
with a gas-tight syringe (D) and injected into the GC.
Right: Fig. 10.2. Dynamic headspace sampling or purge and trap. The top diagram is a schematic
of the purge mode. The purge gas (helium or nitrogen) passes through the sample, extracting the
volatiles, which pass onto an adsorbent trap. The lower diagram shows the desorb mode where
the gas flow is reversed and the trap is heated. The desorb gas sweeps the analytes out of the trap
to the GC column for separation. Finally, the trap is heated to a higher temperature, to clean out
residual analytes and water.
trap is heated to a higher temperature than was used during the desorb step, to
remove residual analytes and moisture, and the system is ready for another
sampling.
Purge and trap is generally more sensitive than static headspace, since the
former technique should theoretically remove all of the analytes from the
matrix, while the latter is limited to removal of a single aliquot of the gas phase.
However, a purge and trap system requires more maintenance and is subject to
problems such as foaming of the sample. The two techniques are compared in
Table 10.1. This chapter will focus on static headspace sampling.
280
TABLE 10.1
Comparison of static and dynamic headspace
Static headspace Dynamic headspace (purge and
trap)
Concentration Parts per billion to low percent levels Parts per trillion to parts per
range million
Hardware Ranges from a simple gas-tight syringe to Complex dedicated instrument
required a sophisticated dedicated instrument.
Automation Can be manual with a gas-tight syringe Purge and trap autosamplers can
but several automated systems can handle handle a maximum of -70
over 100 samples samples
Quantitation Highly matrix-dependent since there is Less matrix-dependent since most
only partial recovery of the analyte of the analyte are removed
Ease of Little maintenance required Labor intensive; traps must be
maintenance cleaned; foaming a problem
Common 1. Blood alcohol 1. Monomers in polymers
applications 2. Residual solvents in pharmaceuticals 2. Flavors in foods
3. Monomers in polymers 3. Volatiles in drinking and
4. Flavors in foods wastewater and soil (United
5. Volatiles in drinking and wastewater States)
(Europe), soil (United States)
significant amount of headspace over the sample; then the vial is sealed and
heated. The analytes (volatile compounds) begin to move into the gas phase
above the sample until a state of equilibrium is reached. At this point, the ratio of
the analyte concentrations in the gas phase and in the liquid or solid phase is a
constant. The constant is defined as the partition coefficient.
= V0 (10.2)
VL
The equilibration process is illustrated in Fig. 10.3 where: CL = concentration in
the liquid phase after equilibration; C, = concentration in the gas phase after
equilibration; Co = original concentration of analyte in the sample; VL = volume
281
-- - A 1 I . I: I I_=I M r =l.C1flnnh
rig. 10.4. Hypotnetlcal curves calculated trom .
Eq. (10.3) for three different analytes with parti- :
tion coefficients of 4, 40 and 400. For all of the 8
compounds, the original concentration of the =-
analyte in the liquid sample was 100 ppb. The
x-axis represents volumes of liquid sample in a '
10-mil vial ranging from 0.5 to 9.5 ml. The y-axis .-
is the concentration of the analyte in the gas o
phase for each volume of liquid sample. Note that
volume of the sample has virtually no effect on E
t1_ L, -1 ---- ,^-
- ---- + - IL, Ri
tile: ll~flubpiCe CVI$UfhtSLldlllUll Wtll tle pitt ta~Lut: E 4 K = 40
coefficient is 400, but sample volume has a sig- 0 ..
nificant effect when the partition coefficient is 4. 0 - -
0.5 2.5 4.5 6.5 8.5
Volume of liquid sample (mL) in a 10-mL vial
282
* No salt
:::Sodium chloride
1 Sodium sulfate
Fig. 10.5. Effect of saturation with salt
on increasing the headspace sensitivity Co
of various compounds in water. Note s
that "salting out" has a significant ,,
effect only on dioxane, the polar
compound where the response was
enhanrrl q-fnld (nff-sarle) when
saturating with potassium carbonate. MeCI2 Benzene TCE Chloroform Toluene Dioxane
TABLE 10.2
Air-water partition coefficients [12] of selected compounds versus temperature
Compound Partition coefficient K
40°C 60°C 80°C
Dioxane 1618 642 288
Ethanol 1355 511 216
Ethyl acetate 62.4 29.3 17.5
Benzene not available 2.27 1.66
Toluene 2.82 1.77 1.27
o-Xylene 2.44 1.31 0.99
Dichloromethane 5.65 3.31 2.07
1,1,1-Trichloromethane 1.65 1.47 1.18
Tetrachloroethylene 1.48 1.27 0.87
Cyclohexane 0.08 0.05 0.02
283
10.3 OPTIMIZATION OF HEADSPACE RESPONSE
10.3.2 Mixing
Most commercially available headspace systems have the ability to mix the
sample during the equilibration period. Mixing tends to reduce the time required
to achieve equilibration [3]. This brings up the question of sampling the head-
space before reaching equilibrium. The main advantage of sampling the head-
space after the analytes have reached full equilibrium between the matrix and
the gas phase, is to achieve maximum sensitivity and precision. In many
laboratories, speed of analysis is more important than maximizing sensitivity
and precision. An example of an application where the GC run time is much
shorter than the time necessary to reach equilibrium is the determination of a
residual monomer in a polymer. In the case of volatiles in a solid, equilibrium
may take hours or even days to achieve. The only practical solution may be to
sample the headspace before equilibrium is attained, and to heat all of the
samples for the same length of time prior to analysis, to assure comparability.
When headspace GC is used in thermodynamic studies such as the
determination of activity coefficients [4-6], it is important to sample after
equilibrium has been reached.
284
200 pL J
In a typical headspace analysis, the vial is sampled only once and then discarded.
After one sampling, some of the analyte is removed from the headspace and after
a new equilibrium occurs, there is less analyte in the gas phase. Therefore,
unlike the situation with liquid injection, only the first injection represents the
original sample.
There is a technique that utilizes multiple sampling from a single vial.
Multiple headspace extraction (MHE) was proposed in 1970 by Suzuki et al. [7]
then modified by Kolb and Ettre [8]. With MHE, a liquid or solid sample is sealed
in a headspace vial and sampled repeatedly at equal time intervals. It is assumed
that the concentration of volatiles under these conditions will decay exponen-
tially (Fig. 10.7). If an infinite number of extractions is carried out, the volatiles
will be completely removed from the vial. The total area count of the peaks
corresponding to the analyte should be equal to the sum of the areas from each
individual extraction. MHE is normally used with solid samples but it can also be
applied to liquids. The technique is useful only if the partition coefficient
between the matrix and the analyte is small, so that a substantial quantity is
removed at each extraction.
285
Fig. 10.7. FID chromatograms of vinyl chloride injected four times from the same vial in a
multiple headspace extraction analysis. The vial contained a polyvinylchloride polymer.
W
0
To
0
0.
o
a
U.
0 4 8 12 16
Gasoline (ppm)
Fig. 10.8. An illustration of the matrix effect on recovery of gasoline from sand, soil and water
by headspace GC.
286
10.4.1.1 Using a blank that exactly matches the sample matrix
This method could be used for the analysis of organic volatiles in drinking water.
It is not difficult to prepare a solution of water that does not contain measurable
volatile organics. The water is spiked with volatile organics of known concentra-
tion and a calibration curve is prepared.
Quantity in original m
sample 0
X = 0.5 mg/mL
Fig. 10.9. Standard additions quantitation in the headspace analysis of ethyl acetate in an
industrial mixture. The detector responses were determined for the unspiked sample and for the
sample spiked with increasing quantities of ethyl acetate. The resulting curve was back-
extrapolated to the x-axis to determine the quantity of ethyl acetate in the original sample.
287
-- _ sample
o
a
0o ar
-
z
1 2 3 4
Run number
Fig. 10.10. Calibration curves for determining vinyl chloride in polyvinylchloride with multiple
headspace extraction.
E, Ai
A A k
(10.5)
1-e
YCA is the total area count, A, is the area from the first sampling, and k is the
slope of the plot obtained by plotting the natural logarithm of the area counts
versus the number of the sampling. Calibration curves for MHE samplings are
shown in Fig. 10.10.
After validating the method by demonstrating a linear response, a simplified
form of Eq. (10.5) can be used, which requires only two samplings:
YlAi is the total area count, A1 is the area from the first sampling, and A2 is the
area from the second sampling. While results would be expected to be more
accurate with more than two samplings, Eq. (10.6) is practical for routine
analysis. The area counts in the calibration sample would also be determined
using Eq. (10.6). Finally, the mass of the analyte would be calculated with Eq.
(10.7).
Wsp is the weight of the analyte in the sample, Am,,p is the total area of the
analyte in the sample as calculated in Eq. (10.6), WStd is the weight of the analyte
in the standards vial, and AId is the total area of the analyte in the standard vial
as calculated in Eq. (10.6).
288
10.4.1.5 Use the full evaporative technique
The full evaporative technique (FET) [10] is an approach for eliminating the
matrix effect in headspace GC, but it is not really an equilibrium headspace
method. With FET, a very small liquid sample (a few microliters) is injected into
an empty headspace vial and the vial is heated until the entire sample is in the
vapor state. Then a sample of the gas phase in the vial is injected into the GC.
Since there is no equilibrium between two phases, there is no matrix effect. The
sensitivity that one can obtain with this method depends on the partition coeffi-
cient of the analyte in the original sample. For example with dioxane in water at
40°C (K = 1618), there is almost no change in sensitivity with FET if one com-
pares the sensitivity from a 10-g sample in a 20-ml headspace vial to the sensi-
tivity when a 10-ml sample is in the same 20-ml vial, However with o-xylene in
water at 40°C (K = 2.44), the sensitivity with FET is lower by a factor of over
100.
This section gives a brief description of the various types of apparatus that are
commercially available.
289
10.5.1 Manual headspace
The simplest method of headspace analysis is to sample the vials with a manual
gas-tight syringe. The vials may be unheated or heated in a temperature bath.
Manual headspace is practical if the lab does not have many samples and the
samples contain volatile compounds with low partition coefficients. This type of
sample does not require precise temperature control and there is no danger of
condensation in an unheated syringe. Manual headspace is an inexpensive way
for a laboratory to evaluate the headspace technique for a particular application.
290
Transfer line to GC
.;'Sp Heated zone
Gasin Sarnpe v A
Restrictor
Gas
Gas
w inb Pressurize
alve ^ Vent valve
to vent
Fig. 10.11. Schematic of a valve-loop automated headspace system. The valve, sample loop,
tubing that leads to the needle that penetrates the sample vial, and the transfer line that leads to
the GC, are heated. Note that there are two sources of carrier gas. The sequence of events is
described in the text. The "inject step" is shown here.
291
SAMUNG
pj
r
Fig. 10.12. Schematic of the Perkin-Elmer pressure-balanced headspace system. The sequence of
events is described in the text.
TABLE 10.3
Summary of advantage and potential disadvantages with dedicated automated headspace
systems
Automated Advantages Disadvantages
headspace
technique
Syringe- Inert sample path. GC injector Gas-tight syringes have Teflon-tipped
based accessible for trouble-shooting in plungers which can leak (once they are used
headspace and for liquid injection. at a given temperature, the syringe should
Simple to vary injector volume not be used at a lower temperature).
Occasional needle bending.
Valve and GC injector usually accessible for Activity in the sample path has been a
loop trouble-shooting in headspace and problem. Ability to change sample loop (and
for liquid injection. Usually easy to sample volume) is limited in some systems.
move from one GC to another.
Pressure- Inert sample path. Simple system GC injector usually not accessible.
balanced with minimum of moving parts. Trouble-shooting difficult. Works best with a
column that requires a high back pressure.
292
been reported [13]. There is a significant matrix effect when comparing the
response of ethanol in blood to the response in water or urine. To minimize
differences from sample to sample and from samples to standards, Penton [2]
diluted the samples and standards four-fold with a solution that was saturated
with sodium chloride and also contained n-propanol internal standard.
In addition to ethanol, there are headspace methods for numerous other
solvents in biological fluids. Sharp [14] developed a method for screening for
more than 40 organic compounds in biological matrices, using a two-column
system with FID detection and MS confirmation. Matrix effects were eliminated
with a 15-fold dilution. Frequent findings in samples were ethanol, toluene and
ethyl ether. Russo and Campanella [15] used headspace to determine benzene at
a level of 10 ng/g in small tissue samples (50-70 mg).
TABLE 10.4
Limits* on residual solvents in pharmaceuticals (United States Pharmacopeia Method 467)
Solvent Limit (g/g)
Dichloromethane 600
Benzene 2
Trichloroethylene 80
Chloroform 60
1,4-Dioxane 380
*As of 2000, benzene removed later.
293
DMI was a better dissolution medium for hydrophobic drugs than DMF or DMA.
Later, the same authors [19] developed a micro method for this analysis. The
sample size was reduced from 100-200 mg to 5-30 mg. The equilibration time
was reduced from 45-60 to 5-10 min.
294
One would not normally think of headspace as a technique for determining
pesticides, but Royer et al. [29] proposed an automated headspace method for
determining dithiocarbamates in plant matrixes. The method was suggested as a
replacement for the spectrophotometric method, European Norm EN 12396-1
(1996), and the manual headspace method, EN 12396-2 (1999). The new method
offers several advantages including greater sample throughput and the need for
fewer reagents. The authors mixed 2 g of plant matrix with 6 ml of a stannous
chloride/HC1 solution and 4 ml of a diethanolamine solution in a headspace vial;
the vial was then heated to 90°C. The dithiocarbamates released carbon
disulfide, which was determined in the headspace. Recoveries from potato, onion
and currants varied from 85 to 103%. The detection limit was 0.05 mg/kg.
Headspace GC is more than an analytical tool; it has been used for determining
thermodynamic constants including partition coefficients [30-32], activity
coefficients [4-6] and Henry's constant [33,34]. An example follows: Kolb et al.
[31] determined the partition coefficients of 12 compounds in water at tempera-
tures ranging from 40 to 80°C with the vapor phase calibration technique. For
calibration, a known quantity of analyte was injected into an empty vial and
allowed to completely vaporize. An aliquot of the vapor phase was injected into
the GC and the area count of the resulting peak was measured. Then a known
quantity of the analyte was injected into a vial containing the matrix, and the
vial was kept at a given temperature until equilibration was reached between the
matrix and the headspace. After equilibration occurred, a portion of the head-
space, equal to the volume injected in the calibration step, was injected into the
GC. The area count of the resulting peak was compared to the area count that
was obtained in the calibration step and the partition coefficient was calculated.
Some of the data from this work is shown in Table 10.2. There was excellent
agreement with partition coefficients derived from other sources.
10.8 CONCLUSION
In this era of rapid technological changes, headspace GC has survived for several
decades as a useful analytical tool for determining volatiles in liquids and solids.
The potential user can determine if the technique is feasible for a given
analytical problem by examining a vast body of literature, and by a few simple
experiments using a GC, some vials and a gas-tight syringe.
Acknowledgements
Figures 10.1-10.11 are printed with the permission of Varian Inc. Figure 10.12 is
printed with the permission of Perkin-Elmer Instruments.
295
REFERENCES
1 B.V. Ioffe and A.G. Vitenberg, Headspace Analysis and Related Methods in Gas
Chromatography.Wiley-Interscience, New York, 1984.
2 Z. Penton, Clin. Chem., 31 (1985) 439.
3 Z. Penton, J. High Resolut. Chromatogr., 15 (1992) 834.
4 N. Asprion, H. Hasse and G. Maurer, J. Chem. Eng. Data, 43 (1998) 74.
5 C.B. Castells, D.I. Eikens and P.W. Carr, J. Chem. Eng. Data, 45 (2000) 369.
6 C.B. Castells, D.I. Eikens and P.W. Carr, J. Chem. Eng. Data, 45 (2000) 376.
7 M. Suzuki, S. Tsuge and T. Takeuchi, Anal. Chem., 42 (1970) 1705.
8 B. Kolb and L. Ettre, Chromatographia,32 (1991) 505.
9 Test Methods for Evaluating Solid Waste Physical/Chemical Methods, SW-846, Vol.
IB, Chap. 4, Sec 4.2.1, U.S. Environmental Protection Agency, Cincinnati, OH,
December 1996.
10 M. Markelov and J. Guzowski, Anal. Chim. Acta, 276 (1993) 235.
11 Z. Penton, J. High Res. Chromatogr., 17 (1994) 647.
12 B. Kolb and L. Ettre, Static Head-space Gas Chromatography: Theory and Practice.
Wiley-VCH, New York, 1997.
13 F. Tagliaro, G. Lubli, S. Ghielmi, D. Franchi and M. Marigo, J. Chromatogr. B, 580
(1992) 161.
14 M.E. Sharp, J. Anal. Toxicol., 25 (2001) 631.
15 M.V. Russo and L. Campanella, Anal. Lett., 34 (2001) 883.
16 USP 24/NF 19, United States Pharmacopeial Convention, Rockville, MD, 1999, 1877.
17 R.B. George and P.D. Wright, Anal. Chem., 69 (1997) 2221.
18 L. Hong and H.R. Altorfer, Pharm. Acta Helv., 72 (1997) 95.
19 L. Hong and H. Altorfer, Chromatographia,53 (2001) 76.
20 L. Alonso and M.J. Fraga, J. Chromatogr.Sci., 39 (2001) 297.
21 E. Fernandez-Garcia, J. Agric. Food Chem., 44 (1996) 1833.
22 F. Mathieu, C. Malosse, A. Cain and B. Frerot, J. High Resolut. Chromatogr., 19
(1996) 298.
23 I. Medina, M.T. Satue-Gracia and E.N. Frankel, J. Am. Oil Chem. Soc., 76 (1999)
231.
24 L. Du, Q. Xu and A. Lin, Fenxi Ceshi Xuebao, 18 (1999) 59.
25 G. Forsgren, H. Frisell and B. Ericsson, Nord. Pulp Pap. Res. J., 14 (1999) 5.
26 A. Kaipainen, S. Ylisuutari, Q. Lucas and L. Moy, Int. Sugar J., 99 (1997) 403.
27 J. Jalbert, R. Gilbert and P. T6treault, Anal. Chem., 73 (2001) 3382.
28 D.M. Levermore, M. Josowicz, W.S. Rees and J. Janata, Anal. Chem., 73 (2001) 1361.
29 A. Royer, M. Menand, A. Grimault and P.Y. Communal, J. Agric. Food Chem., 49
(2001) 2152.
30 T.M, Cummins, G.A. Robbins, B.J. Henebry, R.C. Goad, E.J. Gilbert, M.E. Miller and
J.D. Stuart, Environ. Sci. Technol., 35 (2001) 1202.
31 B. Kolb, C. Welter and C. Bichler, Chromatographia,34 (1992) 235.
32 A. Chaintreau, A. Grade and R. Munoz-Box, Anal. Chem., 67 (1995) 3300.
33 X. S. Chai and J.Y. Zhu, Anal. Chem., 70 (1998) 3481.
34 J. Peng and A. Wan, Environ. Sci. Technol., 31 (1997) 2998.
296
Chapter 11
Liquid-liquid extraction
Frederick F. Cantwell and Manon Losier
There are at least two bulk-liquid phases: The aqueous phase, which initially
contains the dissolved sample, and the organic extractant phase. When the
mixture is agitated, either may become the dispersed phase and the other the
continuous phase, depending on a variety of conditions. Agitation produces not
only a dispersion of one phase as drops, but also convection or stirring in both
phases. Up to a point, more vigorous agitation produces both smaller drops of the
dispersed phase, with a concomitant increase in interfacial area between the two
phases, and more rapid convective mixing within both bulk phases.
The movement of a chemical species from one bulk phase to another is the
result of a thermodynamic driving force in the form of a difference in chemical
potential for a neutral species, or in electrochemical potential for an ionic
species. The movement continues until phase distribution equilibrium is
reached, where the ratio of the concentration of a species i in the organic phase
(C,,) to that in the aqueous phase (Ci,,) is given by the equilibrium distribution
coefficient or partition coefficient (i):
Co
Ki =- (11.1)
The mechanism by which species i moves from one phase to the other is
convective-diffusive mass transfer, which is generally faster for more rapid
convection.
The equilibrium distribution isotherm for i is a plot of Ci,o vs. Ci,. In the
range of concentrations, starting at zero, over which the isotherm is linear, l is
constant. The maximum concentration in this range is sometimes called the
linear sample capacity.
Surface-active species can become adsorbed at the liquid-liquid (LL)
interface in addition to being dissolved in the bulk liquid phase. Depending upon
how tightly packed the adsorbed species can become and on whether multi-
layers can be formed, the adsorbed film can have the properties of a two-
dimensional gas, or liquid, or solid. However, adsorption of sample species at the
LL interface is not very common, and most LLE systems involve only
partitioning between bulk liquid phases.
A sample species may be involved in secondary chemical equilibria with
other dissolved components that are either present in the sample or added as
reagents, in either or both the aqueous or organic bulk phases. In such cases the
sample component (I) may be present as more than one species in these phases so
that the ratio of the sum of its species concentrations in the organic phase to the
sum of its species concentrations in the aqueous phase, or the distribution ratio
D,, is a useful quantity to define.
298
C '
D = - C(i-containing)'o
L (11.2)
C, C(i-containing),a
For systems with secondary chemical equilibria present, the isotherm is a plot of
C,. vs. C, , for which D1 is constant in its linear region. D1 can be expressed in
terms of the c's for the various species. A very common and simple example is
that of a basic sample species, B, which is present as B in the organic phase and
as both B and its conjugate acid BH+ in the aqueous phase, for which:
D CB O KBKEH (11.3)
CB, +CBHa [H30] + Ka,BH
where K,,BH is the acid ionization constant of the species BH + in the aqueous
phase. If Ka,H were equal to zero, D = cB. Quantitatively, the fraction of I
extracted from a volume Va of aqueous phase into a volume V of organic phase,
f,,, at equilibrium is given by the expression:
ID-.Vo
moles I in o V, ..
V at equilibrium. (11.4)
totalmolesI i+Dl.V°
vs
The extent of preconcentration of I can be expressed by the preconcentration
enrichment factor, EF. At equilibrium [7]:
C D
EF I = C, 1- D ,at equilibrium. (11.5)
Clainitia 1+ D· °
Va
Equation (11.5) shows that the maximum value of EF, which is achieved when
D is very large, is equal to the inverse phase ratio, VJVo.
The extent of sample clean-up with respect to a particular component J in
the sample, which interferes with the quantification of analyte component I, is
related to the selectivity of the extraction. This can be expressed as the extent of
separation of I from J in the extract, [8]:
From Eqs. (11.4) and (11.6) it is evident that a large value of D, and a small value
of Dj favors separation of I from J.
Operationally, if it is desired, the forward extraction of a sample component
from aqueous to organic phase can be followed by a back extraction from the
organic phase into a different aqueous phase. Continuing the above example of
base B, the forward extraction of this component into the organic phase is
favored by high pH (i.e. low [H,0+]) and the back extraction is favored by low pH
(i.e. high [H30+]).
299
11.1.3 Advantages of LLE
300
alternative simple and accurate methods of performing LLE, some of which are
noted in Section 11.1.5 and are described in later sections.
Since the problems associated with the separatory funnel mode of operation
arise largely from the agitative drop dispersion/coalescence processes, it is no
surprise that recent LLE techniques, which appear to be growing in popularity
in the literature on chemical analysis, do not employ these processes. Such is the
case for the four techniques identified below.
Single liquid drop techniques with two phases employ a microlitre or smaller
size drop of organic solvent phase suspended in a large volume of aqueous
sample phase. When extraction is terminated, the drop can be injected directly
into an instrument such as a chromatograph or a graphite furnace atomic
absorption spectrophotometer, or it can be interrogated in situ by absorption or
fluorescence spectroscopy.
Unsupported liquid membrane techniques with three phases involve an
aqueous sample phase separated from an aqueous receiver phase by a layer of
organic solvent. The conditions in the aqueous receiver phase are made different
from those in the aqueous sample phase so that the analyte component is
forward-extracted out of the sample phase into the organic phase and back-
extracted out of the organic phase into the receiver phase, all in one simulta-
neous operation.
Supported liquid membrane techniques employ either two or three phases,
with simultaneous forward- and back-extraction in the latter configuration. The
aqueous sample phase is separated from the bulk organic or from an aqueous
receiver phase by a porous, polymer membrane, in the form of either a flat sheet
or a hollow fiber, that has been impregnated with the organic solvent phase. The
sample phase is continuously pumped, the receiver phase may be stagnant or
pumped, and the organic phase in the membrane pores is stagnant and not
replaced between samples. Since this technique of LLE is the subject of a chapter
in this book by Professor Jbnsson, it will not be further described here, except to
note that, in contrast to this supported membrane design, the unsupported
organic membrane phase, discussed above, is not stagnant but rather is con-
vected by momentum transfer from the convected aqueous sample phase.
Segmented flow techniques in small bore tubes performed in the flow
injection analysis mode, were initially developed in the late 1970s and early
1980s as a logical alternative to the Technicon AutoAnalyzer® continuous
extraction system. Continuous, cocurrent flow of alternating aqueous and
organic solvent segments occurs, either with or without air bubble segmenta-
tion. Mechanical phase separators sort the organic from the aqueous segments
at the outlet end of the extraction tube, or on-tube detection is used.
The first three techniques mentioned above, taken as a group, are sometimes
referred to as liquidphasemicroextraction (LPME) techniques.
301
11.1.6 Kinetics and mechanism
302
(this usually occurs quickly if both liquid phases experience convection), then all
of the steps are described by equations that are linear in concentration, and the
overall rate equation is relatively simple. If the CHEM reaction in the aqueous
phase is:
ia -->pa (11.7)
and the extraction is:
Pa 4P 0 (11.8)
then both the original analyte-containing species, i, and the analyte-containing
product of the CHEM reaction, p, constitute the total analyte component I. The
overall extraction rate law is [7,16,19,23,24]:
The first three terms within the parentheses relate to the MT steps (Eq. 11.8)
and include the reciprocal of the mass transfer coefficients (cm/s) for analyte
component movement toward (kMT,a = D, 0/a), across (kINT) and away from (kMT,,
= DJoo), the interface, respectively. The fourth term in the parentheses inc-
ludes the first-order CHEM rate constant kHEM (s-') for Eq. (11.7). In Eq.
(11.10), A is the area of the LL interface (cm 2 ), 6band 60 are the thicknesses (cm)
of the Nernst diffusion layers in the "a" and "o" phases, D1, and D, o0 are the dif-
fusion coefficients (cm2 /s) of the analyte component in the "a" and "o" phases,
and kINT is the interfacial mass transfer coefficient (cm/s) of the analyte compo-
nent.
In Eq. (11.10), the so-called Whitman two-film model has been employed to
describe the convective-diffusive MT in the "a" and "o" phases [10,16,25,27]. In
this view, all of the stirred (bulk) liquid phase is completely and continuously
mixed except for a thin Nernst diffusion layer of thickness 6 which is located
immediately adjacent to the interface (Fig. 11.1). This layer is considered to be
totally stagnant so that analyte crosses it exclusively by diffusion. In this view,
faster convection produces a smaller value of 6.
From Eq. (11.5) it can be seen that the equilibrium concentration of I in the
organic phase which should be used in Eq. (11.9) is:
303
Fig. 11.1. Schematic view of the Whitman two-film model for Bulk Aq. 8. Bulk Org.
convective-diffusive mass transfer at a LL interface. The Phase a o Phase
dark vertical line is the interface. Completely stagnant I
Nernst diffusion layers exist on either side. Their thickness 6 -I - - - N-
is smaller for greater convection. Bulk solutions are com- cton -
pletely convectively homogenized at all times. Solid arrowsonvection Diffusion onvecon
represent pure convective MT and dashed arrows represent Films
pure diffusive MT.
For LLE which has not yet reached equilibrium, the fraction extracted at time t
is:
EF, = D
DI1(1-
I - e-o......)
Vo- ) before equilibrium (11.14)
(11.14)
1+D .°
Va
and the extent of separation of I from J at time t is:
Equations (11.13) and (11.14) differ from Eqs. (11.4) and (11.5) by the presence
of the rate term (1 - e - k.. t.)
It is often the case in LLE that the rate of crossing the liquid-liquid interface
and the rates of all chemical reactions are much faster than the rates of diffusion
across the Nernst diffusion films. When this is so, k,,, and kH,,,, are both so large
that the terms containing them can be ignored and Eq. (11.10) simplifies to:
k.-alail A( D D d LDI
V+1] (11.16)
Furthermore, it may happen that only one of the liquid phase mass transfers
becomes the rate-determining step. This is true for example if D. a > 8>
6, in
which case Eq. (11.16) may be approximated as:
k =A . D DI
.[D V. +1i (11.17)
304
If, in addition, it is also true that DI · VJV a >> 1, then Eq. (11.16) may be
approximated as:
koveral A D (11.18)
Conversely, when the opposite is the case, i.e. 6o> > D- 6a, Eq. (11.16) becomes:
Since two liquid phases are in direct contact with one another, in the absence of
any porous solid support at the interface, then when one of the phases is
mechanically stirred (i.e. directly convected), the other will also experience
convective mixing. This is because the first phase transfers momentum to the
second phase as a result of frictional drag at the LL interface [16,19,25,28]. This
is true regardless of whether the second phase is layered over (or under) the first
or whether it has the form of a small drop suspended in or moving through the
first phase.
The significance of this indirectly induced convection in unsupported liquid
phases is two-fold. First, the rate of the resulting convective-diffusive MT in
each phase is faster than if MT were occurring by diffusion alone and second, it
can be treated quantitatively in terms of the Whitman two-film model, as
embodied in Eq. (11.10).
When discussing MT rates it is necessary to mention the possible effect of
adsorption of a species from solution [29]. Surface-active substances coming
from a sample or from reagents used in an analysis can adsorb at the LL
interface and thereby may reduce the interfacial MT coefficient (h,,,N) and/or
increase the diffusion film thickness (6). The latter would reduce the liquid
phase MT coefficient (D6/8), in any liquid phase which is convected via moment-
um transfer from an adjacent liquid phase [13,19,24-26,30-36]. If, in an extreme
case, convection is completely stopped in a liquid phase then diffusion becomes
the only liquid phase mass transfer process in that stagnant liquid phase and the
MT coefficient can no longer by approximated by D1 /6 but must be represented
by a different expression. The form of the expression for purely diffusive MT
depends on the geometry of the phase (e.g. planar, spherical). The subject is
treated in Section 11.3.2.
305
11.1.8 Quantitative analysis by LLE
Determining the amount of an analyte component I in a sample involves
subjecting both the aqueous sample solution and an aqueous standard solution
to the same LLE process, followed by instrumentally obtaining signals on both
the sample and standard extracts that are proportional to the concentration of I.
Instrumental techniques commonly employed include molecular absorption
spectroscopy, fluorescence, atomic spectroscopy, chromatography, etc. If only a
forward extraction is involved, then the instrument signals would be proportion-
al to C,, (sample) and C,o,(standard). The value C,o O (sample) can be obtained by
multiplying the known value of C,o (standard) by the ratio of the sample and
standard signals.
An essential precondition for a quantitative determination is that the frac-
tion of I extracted (f,) either must be quantitative (fi, = 1) or it must be the
same for both sample and standard. Equation (11.4) shows that when the extrac-
tion is allowed to proceed to equilibrium, assuming that the same ratio V/Va is
used, then conditions such as pH, ligand concentration, etc., used with both sam-
ple and standard, must be such that either D, is very large or it has the same
value. On the other hand, Eq. (11.13) shows that when the extraction is termi-
nated before equilibrium is reached the requirements are more stringent: not
only must D, be the same, but also both koer,,,, and t must be the same for sample
and standard. Another way of stating this is that unless extraction is unambigu-
ously quantitative, then for equilibrium extraction only those "matrix effects"
and operating conditions that influence Dr need to be made the same in standard
and sample; while for non-equilibrium extraction "matrix effects" and operating
conditions that influence both Dr and kove 1 1 must be made the same. For many
types of samples there are no additional matrix effects on kverall than there are on
D, so that the only operating conditions that require more careful control are
the interfacial area (A), the details of stirring (which effects 6a and 6) and tem-
perature. Usually a specially designed extraction vessel and a constant speed
stirring device are used for non-equilibrium extraction, and room temperature is
often adequate.
When surface-active substances are present in a sample the risk exists of a
smaller value of kover.ll for the sample than for the standard, as discussed in
Section 11.1.7. This particular type of matrix effect on non-equilibrium extrac-
tion can be addressed by "matrix-matching" or by employing the standard
addition technique in which the standard substance is added to the sample.
Since the presence of a surface-active component generally has a negligible effect
on DI, it does not exert a matrix effect on equilibrium extraction.
11.2.1 Introduction
The automation of analytical methods received a major boost in the mid-1950s
306
with the introduction of the Technicon AutoAnalyzer®, an air-bubble seg-
mented, continuous flow system [37-39]. In this system, in which aqueous
samples were injected periodically into a flowing aqueous stream, the presence
of air bubbles served both to reduce the longitudinal zone broadening of the
sample zone and to facilitate mixing of reagents with the sample. The subse-
quent implementation of LLE in the Technicon system in the mid-1960s
involved cocurrent flow of aqueous and organic solvent segments through coiled
glass tubes, sometimes packed with glass beads, either with or without simulta-
neous air-bubble segmentation [39-42].
Two decades after the development of the Technicon Autoanalyzer®, the
concept of continuous flow analysis in the "flow injection analysis" (FIA) mode
of operation was introduced [39,43,44]. In contrast to the AutoAnalyzer, FIA
involves the use of smaller diameter flow tubing and does not employ air-bubble
segmentation. The implementation of LLE in the FIA mode took place in the late
1970s [45-48]. Shown in Fig. 11.2 is a simple FIA instrument such as might be
used for segmented flow LLE [49]. The pump(s), sample injector, segmentor,
extraction tube, phase separator, detector and readout device are shown. Pumps
were initially of either the peristaltic type [45,46] or constant pressure type [50],
but now are as likely to be of the constant displacement syringe or reciprocation
piston type [51,52]. Sample injectors are valves, not unlike those used in HPLC
[39].
Segmentors appear in a wide variety of designs, the simplest being a
T-shaped tube [50,53]. Many of these designs have been reviewed [51,54]. The
function of the segmentor is to create a uniform sequence of alternating aqueous
and organic segments which then flow through the extraction tube. The pres-
ence of alternating segments facilitates mass transfer of the extractable analyte
component from the aqueous into the organic phase (see Section 11.2.2) and,
coincidentally, favors both reduced sample zone broadening and enhanced
reagent mixing. The extraction tube is typically smaller than 2 mm internal
diameter. Its material of construction, length and tightness of coiling all affect
the rate of solvent extraction and the sample zone broadening, as discussed in
Section 11.2.2. Phase separators, just like segmentors, appear in a wide variety
Fig. 11.2. Simple segmented flow, flow injection analysis (FIA) instrument for LLE showing
cocurrent alternating segments of organic (black) and aqueous (white) phases in the Teflon
extraction tube. I is the injector, S the segmentor and PS the phase separator. From Ref. [49].
307
I l{
2.0
W
a 10
C
C
0
.Q I
of designs, with their principles of operation based on gravity (i.e. density differ-
ence), wetting (i.e. contact angle) and interfacial tension [13,29,51]. After sepa-
ration of the organic extractant phase from the aqueous phase, the former flows
to a detector. Each sample injection produces a peak, the height or area of which
can be used for quantification [50,55] based on similarly extracted standards.
Figure 11.3 [50,56] presents the absorbance signals obtained from replicate
injections of a caffeine solution into a segmented flow LLE instrument which
employs chloroform as the extractant.
The types of instruments used for detection of the analyte component in the
extract at the end of a segmented flow LLE extraction include molecular absorp-
tion and fluorescence spectrometers, atomic absorption and emission spectrom-
eters, mass spectrometers and electrochemical instruments. In addition, it is
common for the extract to be introduced into a high efficiency separation instru-
ment such as a gas or liquid chromatograph, or a capillary electrophoresis col-
umn, or related instrument [51]. The way in which the segmented flow extractor
is "interfaced" with the detection instrument can be "on-line", with the detector
located in the flow stream after the extraction column and phase separator;
"off-line", by collecting the extract and transferring it to a separate detection
instrument; or "on-tube", with the detector located at some point along the
extraction tube. One of the currently most challenging instruments with which
to interface a continuous flow extractor is a capillary electrophoresis instru-
ment. Common ways of doing so have been reviewed [57]. A requirement of any
on-line or on-tube detector is that it must have a relatively small hold-up volume
so as to minimize additional zone broadening and the resulting reduction of
detector signal [51].
As is true for LLE in general, a great variety of secondary chemical equilibria
are routinely employed in segmented flow LLE to control the values of the
distribution ratios of various sample components for purposes such as: (i)
increasing the fraction of analyte component extracted (f,o) when using a small
phase ratio (VoVa), for greater preconcentration; (ii) increasing the extent of
separation () of the analyte component over other sample components, for
greater selectivity in sample clean up; and (iii) converting the analyte compo-
308
Fig. 11.4. Somewhat more complicated segmented flow, FIA instrument for LLE based on
chloroform extraction from an alkaline aqueous phase. The injector is V4 . The segmentor is T,.
The phase separator M, which passes some of the chloroform phase containing extracted
pheniramine, is located between two sections of the Teflon extraction coil. The phase separator
Ma which passes some of the aqueous phase containing phenylephrine is located after the second
section of the Teflon extraction coil. See Fig. 11.5 for detector output. From Ref. [58], with
permission.
nent into a more detectable species. Reagents commonly employed for these
purposes include buffers, ligands, ion-pair reagents and crown ethers [51].
Because of this, a segmented flow LLE instrument is often more complicated
than that which is shown in Fig. 11.2. Pumps and tubes for aqueous reagents are
commonly present.
By way of illustration, such a system is pictured in Fig. 11.4 [49,58]. In this
example the reagent is 0.2 M NaOH which raises the pH of the aqueous phase
prior to segmentation with chloroform. Two phase separators are used in series,
the first of which isolates some of the organic phase and sends it to the detector
and the second of which isolates some of the aqueous phase and sends it to the
detector. Shown in Fig. 11.5 [49,58] are the peaks obtained from one injection of
a commercial nasal spray containing the antihistamine pheniramine maleate
Pheniramine 1
1,o
8.05
309
and the decongestant phenylephrine hydrochloride. At high pH pheneramine is
present as the neutral free-base form which extracts into chloroform, while
phenylephrine is present as its anionic conjugate base form which stays in the
aqueous phase. Also included in the instrument are a cation exchange resin trap
and an anion exchange resin trap which remove potential interferents from the
aqueous phase prior to the segmentor.
Analyte components that are routinely determined with the aid of secondary
chemical equilibria include metal ions, inorganic anions, ionic and nonionic
surfactants, pharmaceuticals and, in general, any analyte components for which
secondary chemical equilibria can be employed in conventional "separatory
funnel" LLE.
When the segmented flow of two immiscible phases has proceeded through a
sufficient length of extraction tube, phase distribution equilibrium is achieved
and the concepts that were introduced in Section 11.1.2 such as distribution
coefficient of a species (), secondary chemical equilibria and distribution ratio
of a component (D,), describe the system.
The mechanistic details of mass transfer kinetics in segmented flow LLE
have been studied [49,51,59-62] and a nodding acquaintance with them is
essential for understanding any of the several variations of this technique. The
first important characteristic of segmented flow is the fact that one of the liquid
phase solvents (i.e. the one with the smaller contact angle with the wall of the
extraction tube) forms a wetting film on the wall. As examples, a Teflon tube is
wetted by the organic phase and a glass tube is wetted by the aqueous phase.
This is illustrated in Fig. 11.6. The thickness of the wetting film is greater for
larger tube radius, higher viscosity of the wetting phase, higher flow velocity and
lower interfacial tension [49,51,60,62,63]. For typical segmented flow LLE
conditions the wetting film thickness has an order of magnitude value of 10 Am.
As a result of the presence of the wetting film, the wetting phase is a "continuous
phase" while the non-wetting phase is a "dispersed" phase whose segments are
completely surrounded by the continuous phase.
The second important characteristic of segmented flow, which is also illus-
trated in Fig. 11.6, is that convection is present in both the wetting and the
non-wetting phase segments. For the wetting phase this is the result of frictional
drag on the wall, while for the non-wetting phase it results from frictional drag
on the wetting film.
In segmented flow LLE, the injected aqueous sample solution occupies
several, to many, aqueous segments. The transfer of analyte component from
the aqueous to the organic phase segments occurs by the following steps:
liquid-phase MT through the aqueous phase to the liquid-liquid interface;
crossing of the interface; and liquid-phase MT away from the interface in the
310
FLOW -
I :,"
Fig. 11.6. Longitudinal section of a Teflon extraction tube showing alternating segments of
organic (o) and aqueous (a) phases. The convective toroidal circulation patterns are shown in
each segment. The wetting film of organic phase can be seen between the aqueous segments and
the wall. Its relative thickness is exaggerated. From Ref. [491.
311
arises primarily from the wetting film, zone broadening is minimal when the
analyte component is present nearly exclusively in the segments of the non-
wetting phase. In some, rather rare, cases analyte zone broadening occurs
because the analyte component adsorbs on the wall of the extraction tube [67].
The discussions of segmented, two-phase flow LLE in Sections 11.2.1 and 11.2.2
have described mainly instruments having the traditional design which employs
continuous, segmented flow of both phases and includes a phase separator prior
to the detector. However, over time, designs have evolved toward the employ-
ment of more sophisticated electronic/computer control of instrument compo-
nents such as pumps, valves and detectors, and toward more elegant exploitation
of those fundamental properties of two-phase flow which were identified in
Section 11.2.2. Two trends, in particular, are noted. One trend is the greater use
of "on-tube" spectroscopic detectors employing fluorescence and absorbance
measurements, which can continuously measure analyte concentrations in the
segments as they flow past a point on the extraction tube [62,68,69]. This
eliminates the need for a phase separator. The second trend is toward new ways
of exploiting the presence of a wetting film on the walls of the extraction tube.
The terms "monosegmented flow extraction" [68,70,71] and "sequential injec-
tion extraction" [72] have been used for several related techniques. In these
techniques a single, short segment of an aqueous sample solution is injected into
a capillary extraction tube. Introduced just behind it is a short segment of the
organic extractant. Surrounding these two segments may be an aqueous carrier
fluid or air bubbles. The extraction tube often goes to an on-tube detector where
the concentration of the analyte component is measured photometrically in the
organic segments as they pass through it. Extraction tubes may be made of glass
or fused silica, in which case the wetting film is aqueous. For these, mass
transfer from the aqueous into the organic phase occurs largely via the wetting
film as described in Section 11.2.2. Alternatively, extraction tubes may be made
of Teflon or polyethylene, in which case the wetting film is organic. Mass
transfer from the aqueous sample segment into the organic segment is slow in
this case because there is no organic wetting film present on the tube wall
adjacent to the aqueous segment because it is ahead of the organic segment.
Something must be done to speed up the rate of transfer. Two quite different
approaches to circumventing this slow MT problem may be mentioned for their
clever exploitation of the fundamental principles.
In one approach [68], a homogeneous solution containing an aqueous
sample, a nominally immiscible organic solvent and the co-solvent ethanol, all in
one phase, is present as a monosegment between two air bubbles in a polyethyl-
ene extraction tube and is pumped toward an on-tube detector. Before it reaches
the detector, a concentrated aqueous NaNO3 solution is added to this single
segment causing it to separate, via the phenomenon of salting-out, into two
312
adjacent segments: one being a water-rich phase and the other being an organic-
rich phase. Since the analyte component becomes part of the organic-rich
segment during the process of two-phase formation in the tube, there is no need
for MT between the two segments after they have formed.
Another approach to speeding up the rate of extraction in a sequential
injection system with a hydrophobic extraction tube [72] involves drawing a
short organic solvent segment into a Teflon extraction tube followed by a
segment of aqueous sample solution whose composition is such that D for the
analyte component is large. Long air-bubble segments are present before and
after the two adjacent liquid segments, but not between them. The leading
organic phase segment lays down a wetting film on the Teflon tubing wall and
analyte component is extracted into this wetting film as the following aqueous
phase flows past it. In fact, the volume of the organic segment used is small
enough that, after flowing some distance, all of it becomes coated on the wall and
the organic phase is no longer present as a segment. Now the direction of flow is
reversed so that the aqueous phase flows back past the organic wetting film and
out of the tube to waste. Next, a different aqueous solution, for which D is very
small, is drawn into the extraction tube and through the organic wetting film on
the wall where it back-extracts the analyte component out of the film. Reversing
the flow direction now causes this aqueous back-extractant phase to flow back
past the film and out of the tube, where it is collected and subjected to HPLC
analysis.
For many "monosegmented" and "sequential injection" extractions, phase
transfer equilibrium is not achieved [69], so that both the concepts of segmented
flow that were discussed in Section 11.2.2 and the kinetic and mechanistic
concepts discussed in Section 11.1.6, are essential to a proper understanding of
these techniques.
313
that perform back-extractions [74], and is as high as 50 only in systems that
exploit the wetting film in special ways to enhance back-extraction [52].
In the 1996 survey mentioned above [151 it was noted that laboratory
workers do not often use segmented flow LLE in sample preparation. This is
probably due to the considerable instrumental complexity required, compared to
the relatively simple solid phase extraction systems that are commercially
available in profusion.
11.3.1 Introduction
314
5-H GC
|Syringe Z
As is true for all LLE techniques, when the extraction process is allowed to occur
for a sufficiently long time that equilibrium is achieved, then the equations
presented in Section 11.1.2 apply. The equations which describe the time-course
(rate) of extraction and therefore indicate how long it takes to reach equilibrium,
depend on which, if either, of the two phases experience convection, regardless of
whether convection in a given phase is caused directly or by momentum transfer
from the other phase, as described in Section 11.1.7. For Single Drop-LPME,
three MT situations are commonly encountered:
Case I: Directly convected aqueous phase, and indirectly convected organic drop
Microdrops of >0.1 l experience convection via momentum transfer from the
aqueous phase [19,25,30,31]. The convective circulation pattern for such a
system is shown in Fig. 11.7. The overall extraction rate equation is given by Eq.
(11.12) and kveral is given by Eq. (11.10) or by one of its approximate forms, Eqs.
(11.16-11.20).
315
Case II: Directly convected aqueous phase, in the absence of convection in the
organic drop
Drops smaller than about 0.1 L1 (i.e. nanodrops and picodrops) do not experience
convection via momentum transfer from the aqueous phase. The exact cut-off
volume below which a drop becomes stagnant depends on such factors as the
vigorousness with which the aqueous phase is stirred (i.e. its Reynolds number),
the interfacial tension between the organic and aqueous phases and their
density difference, and the presence of minute traces of surface-active impurities
which exist even in highly purified solvents [19,25,30,31].
However, even microliter-size drops can have their convection reduced or
even completely suppressed by the presence of an adsorbed interfacial film.
When the drop is not convected, MT through the aqueous phase toward the
interface is convective-diffusive and is represented by the term D. ba/Di. a in Eq.
(11.10). On the other hand, MT through the organic drop away from the
interface is purely diffusive and can be approximated by an equation for
diffusion into a sphere [82]. This complex equation, which must be evaluated
numerically, does not lead to a first order rate equation and is not correctly
represented by the term SJD,,o in Eq. (11.10).
If k,,, and h,,,HEM are large and if the rate determining MT step occurs in the
aqueous phase, then the first order Eqs. (11.9) and (11.12) describe the kinetics,
and koveral is given by Eq. (11.17) or (11.18); but if the rate determining MT step
occurs in the non-convected organic drop, then Eqs. (11.9) and (11.12) do not
describe the kinetics and Eqs. (11.19) and (11.20) do not apply. To make an
order-of-magnitude estimate of the conditions under which the control of Case II
MT kinetics will reside in the drop phase rather than in the aqueous phase, the
fact that drop phase MT is not first order can be ignored and the inverse mass
transfer coefficient term o/D,,o in Eq. (11.10) can be replaced by tadius/rdrop, where
rdrop is the drop radius and t,.adus is the time required for an average molecule of I
to diffuse the distance rd,,,. Rearrangement of the Einstein diffusion equation
gives [25,83,84]
Values of tadi/rdrop are presented in Table 11.1 for various drop volumes V,
assuming a typical value of 3x10 ° cm2 /s for D, 0o [17]. Also shown in Table 11.1
are values of Di. 6aD,, for various values of the distribution ratio D1, assuming
typical values of 20x 10-4 cm for 6a [7,81] and 1 x 10 -5 cm 2/s for Dia [17].
If tradius/rdrop > > D b8/dD,, then the extraction rate is controlled by diffusive
316
TABLE 11.1
Drop size, distribution ratio and the rate-determining step for Case II mass transfer kinetics
D, D-,6a (s/cm)
1 200
10 2,000
100 20,000
1000 200,000
MT in the drop, and the rate equation will be the one for spherical drop diffusion,
rather than Eq. (11.12). From Table 11.1 it can be seen that drop diffusion con-
trols the extraction rate for microdrops (Vo = 1 to 100 Al) when D is relatively
small.
Iftradius/rdrOP< < D aI/Di,, then convective-diffusive MT in the aqueous phase
controls extraction rate. Equation (11.12) is the rate equation and kovera is given
by either Eq. (11.17) or (11.18) depending on whether DrVVV is greater than 1.
From Table 11.1, the aqueous phase MT is seen to be rate determining for small
microdrops and large nanodrops (V, = 0.1 to 10 gl) when D1 is large, and it is
always rate-determining for small nanodrops and picodrops (V = 0.01 l).
If traiuJr, 0rp and D· 6Ja/D, are not very different from one another then MT in
both phases affect the extraction rate and the "mixed" rate equation will be even
more complicated than both Eq. (11.12) and the spherical drop diffusion
equation.
In the above discussion it has been assumed that 1/,,NT is negligibly small,
even though an adsorbed film may be present at the LL interface. This is a
reasonable approximation since the suppression of drop convection is usually
found to be a more important effect of adsorbed surface-active substances than is
their diffusional barrier effect [19,85].
317
diffusive MT in the aqueous phase. Because the phase ratio VoVa is small (i.e.
10- 5 to 10 - 6) and because the extraction process is terminated short of equilib-
rium, the fraction extracted, f, 0, will generally be very small. This means that the
concentration of analyte in the aqueous phase will not be significantly decreased
during the extraction (i.e. C,a,equi C,ainitial). The differential rate law for revers-
ible semi-infinite spherical diffusion, which applies well before equilibrium is
achieved, has the form [84,86-89]:
For Picodrop-LPME (rdrop < 6 Am), (Di,t)" / 2 < < rp, even at times as short
as 1 second, so that the term in square brackets becomes [1/rdrop], and Eq. (11.23)
can be integrated to give the following first-order rate equation:
The rate constant is larger for smaller drops; however the fraction of the total
analyte component which is extracted (f,o) in a given time is greater for large
drops.
For Nanodrop-LPME and especially for Microdrop-LPME, drop is smaller
so that Eq. (11.23) cannot be simplified and must be integrated as is. The
integrated rate equation obtained from Eq. (11.23) (not shown) gives a generally
convex shape rate curve, but because of the presence of t /2, it is not a first-order
equation.
318
where Corg,a,sat is the solubility of the organic solvent in the aqueous phase (g/ml)
and Porg is its density, at the temperature of interest.
Often in LPME equilibrium is not reached, so that a knowledge of the
dissolution rate equation is useful. The rate of dissolution of a spherical drop of
organic solvent in water at constant temperature depends upon the solubility of
the solvent in water, the vigorousness with which the aqueous phase is con-
vected, the radius of the drop and presence of an interfacially adsorbed film
[17,19].
For an organic microdrop in a convected aqueous phase, a very thin layer (i.e.
< < 6 a) of aqueous phase immediately adjacent to the interface is and remains
saturated with organic solvent and serves as a "source-layer" away from which
dissolved organic solvent experiences convective-diffusive MT in the aqueous
phase [19]. This involves diffusion across a Nernst film of thickness a . The
resulting first-order dissolution rate law is:
1 dnorg Dorg. C
Ad -r
= (Dorga,sat- org,a) (11.27)
A dt 6 (,
Equation (11.27) can be transformed to take into account the fact that A
decreases with drop dissolution, and the resulting expression can be integrated
to give a rather complicated equation. However, for the purpose of assessing the
suitability of an organic solvent for use as an extraction drop, a "worst case" can
be estimated from the initial rate of dissolution. Substituting A = constant and
Corg,a = 0 into Eq. (11.27); dividing both sides by the initial grams of organic
solvent norg,initial, rearranging and integrating, gives an expression for the fraction
of the drop dissolved in time t:
the fraction of a 1-l drop which will have dissolved after 30 min (i.e. 1800 s) is
4x 104, or 0.04%. At equilibrium, after a longtime, Eq. (11.25) gives the fraction
dissolved as 6.6 x 104, or 0.066%, which is still negligible.
For an organic picodrop in a non-convected aqueous phase the dissolution
rate is controlled by pure spherical diffusion into the aqueous phase from the
saturated source-layer surrounding the drop. Assuming that the experiment is
terminated well before drop dissolution-equilibrium is achieved, which is usually
the case, the integrated fractional rate equation is for semi-infinite spherical
diffusion, which has the form [84,86]:
Anorg, a
4
t drop Dorg,a Corg,a,sat t(11.29)
og gitioongitil(11.29)
norg, initial norg, initial
319
If the drop radius is assumed to stay constant, this equation pertains to the
initial dissolution rate. The fact that the rate decreases with time because rdrop
decreases, is ignored.
As an example, consider a 40 pl (i.e. 40 x 10 - 9 ml) drop of n-octane in contact
with any volume of non-stirred aqueous phase that is > >40 pl. Here rdrop = 21
Am and the other parameters are the same as those given in the example above.
After 30 min the fraction of the drop that will have dissolved is 0.004, or 0.4%,
which is negligible. The fraction dissolved when dissolution equilibrium has
been reached depends on Va, as shown in Eq. (11.25). For instance, at equilib-
rium, 24% of a 40 pl drop will be dissolved in a 10 Al aqueous phase.
A point which has been ignored in the above discussions of drop dissolution is
the fact that Crga,sat is larger for smaller drops, as given by the Kelvin equation
[91]. However, the effect is negligible for drops with rdrop > 0.1 Avm and can
justifiably be ignored, even in picodrops.
Loss of an organic solvent from a small drop can occur not only by dissolution
in the aqueous phase but also by vaporization into the headspace of the vessel in
which the LLE is being performed. The number of moles of organic solvent vapor
in the headspace at equilibrium, n'orgvapor, can be estimated from the vapor
pressure of pure organic solvent, Porg, assuming ideal gas behavior:
Porg Vheadspace
norg,vapor = (11.30)
R.T
Of course, before dissolution and vaporization equilibria are reached n'org,vapor
will be smaller than the value given by Eq. (11.30).
Experimentally, drop dissolution and vaporization can be completely pre-
vented by pre-equilibratingthe aqueous phase and headspace with vapor of the
organic solvent, from an independent source, prior to introducing the organic
drop into the system. In this way, even organic solvents having considerable
solubility in water and considerable vapor pressure can be used for Single
Drop-LPME.
A variety of experimental designs has been employed to contact the single drop
of organic phase with the aqueous sample solution. Also, a variety of analytical
measurement techniques has been used to quantify the amount of analyte
extracted into the drop. These include molecular spectroscopic and electrochem-
ical techniques that are used directly on the drop, in situ, and chromatographic,
electrophoretic, atomic spectroscopic and mass spectrometric techniques that
require injection of the organic extract into the instrument in either an off-line
or on-line manner. In this section, the discussion will include modes of contact-
ing the phases, modes of quantifying the analyte and experimental results.
Microdrop-LPME, Nanodrop-LPME and Picodrop-LPME will be discussed in
Sections 11.3.4.1, 11.3.4.2 and 11.3.4.3, respectively.
320
It will be recalled that, no matter what the experimental design, LLE is being
performed for sample preparation, usually to accomplish sample clean-up and/or
analyte-component preconcentration, and that the quantitative measurement
step is being made directly on the organic drop, without the need to remove the
extracted analyte from it. A recent review gives an overview of the formation and
properties of aqueous and organic microdrops and of drop-based analytical mea-
surements [92].
11.3.4.1 Microdrop-LPME
One of the simplest ways of contacting an organic microdrop with a stirred
aqueous sample solution is to suspend the drop in the sample solution from the
tip of a microsyringe needle, as is shown in Fig. 11.7. Experimental characteriza-
tion of the equilibrium and kinetic features of this system was performed on both
an 8-jl drop [7] and a 1 -ul drop [81], of n-octane in 1 ml of a stirred aqueous
sample solution. The presence of indirectly-induced convection in the organic
drop was made visible by suspending charcoal powder in the drop. The kinetics
are therefore expected to be Case I mass transfer (Section 11.3.2). Shown as the
points in Fig. 11.8 are plots of measured concentration of the analyte component
4-methylacetophenone (4-MAP) in the 1 Al drop of n-octane versus extraction
time, for experiments performed at four different stirring rates. The concentra-
tion of 4-MAP was measured by injecting the 1 l n-octane drop directly into a
gas chromatograph. The solid lines in Fig. 11.8 are from nonlinear least squares
fits of Eq. (11.12) to the data points. From the measured limiting concentration,
C4M,o,equ, which all four curves are approaching at long times, the value D4 MA,
= 41 ml/ml is obtained by rearranging Eq. (11.4). The value of k,,,over,, increases
with stirring rate as shown in Table 11.2. Since D4 P is large, kovera, is expressed
by Eq. (11.17), from which the value of the mass transfer coefficient D4M a/6a,
shown in Table 11.2 is calculated. The diffusion coefficient D4 MAP was measured
to be 7.68 x 10- 6 cm 2 /s in a different experiment and is used to obtain the Nernst
diffusion film thicknesses that are shown in the last column of Table 11.2. As
expected, a decreases at faster stirring rates.
A Microdrop-LPME device, based on the design shown in Fig. 11.7, with Case
I MT, employed a 2 -/tl drop of organic solvent in a stirred aqueous sample
solution of between 1 and 9 ml volume in order to preconcentrate and measure
E 4-
321
TABLE 11.2
Equilibrium and kinetic parameters for the extraction of 2.06 x 104 mo/i of 4-MAP from water
into 1.00 l1of n-octane
6
Stirring rate C4 -MAPoe uil koral D-MAPaa/
4 * a
-1
(rpm) (mol/cm ) X 106 (S ) X 103 (cm/s) x 103 (cm) x 104
900 7.97 ± 0.08 4.68 ± 0.11 4.53 ± 0.15 16.9 + 0.7
1200 7.92 + 0.07 6.42 0.17 6.18 ± 0.22 12.4 ± 0.5
1500 7.87 ± 0.03 7.98 ± 0.09 7.64 ± 0.12 10.1 ± 0.3
1800 7.98 ± 0.03 9.05 ± 0.11 8.78 ± 0.14 8.8 + 0.2
2
*D4_MAP, a = 7.68 x 104 cm /s from a separate experiment.
the concentrations of pesticides in river water [93,94] and cocaine and its
metabolites in urine [95]. In these studies a variety of different organic solvents
were used for the microdrop including hexane, chloroform, toluene and mixtures
of these. Extracted analyte was quantified in the drop by gas chromatography
employing an internal standard. Typically, stirring rates were rather slow and 5
min extraction times were used so that the fraction of analyte extracted was
small (f,0 < 1%). Also, no special precautions were taken to keep the stirring rate
constant. These conditions led to rather high relative standard deviations (RSD
= 10%) which, however, were quite adequate to screen for the drugs and
pesticides.
The microsyringe set-up shown in Fig. 11.7 has also been used with a
different mode of contacting the organic microdrop with the aqueous sample
solution [96,97]. In this mode, which the authors call "Dynamic-LPME", the
aqueous phase in the sample vial is not stirred and the organic phase does not
have the shape of a drop during the mass transfer process but, rather, is a
segment inside the microsyringe needle. Convection is created in both the
aqueous phase and the organic phase by alternately moving the microsyringe
plunger in and out 15 or 20 times in succession. To begin with, at the time the
microsyringe needle is immersed in the 4-ml of aqueous sample solution the 1-gl
of organic solvent is present just inside the needle tip. As the plunger is then
retracted the organic solvent is pulled farther back in the needle and is followed
by a specified, few 1lof aqueous sample solution (Va,,amp plug) After a few seconds
pause (the dwell time) the plunger is advanced to expel all of the aqueous
solution, leaving the 1 Aldof organic solvent just inside the tip again. Both the
organic and aqueous liquids in the needle experience convection due to frictional
drag during the times that the plunger is moving out and in. The majority of the
extraction of analyte component occurs from the aqueous sample plug that is in
the needle, into a wetting film of the organic solvent that coats the inside of the
needle, which is made of stainless steel.
Because extraction is occurring into a thin wetting film of organic phase
rather than a drop, the kinetic conditions for Case I in Section 11.3.2 do not
apply. Furthermore, the extraction occurs as a series of steps, one for each cycle
322
of the plunger, rather than continuously. If it can be approximated that the mass
transfer between the thin wetting film and the 3 Al aqueous sample plug is fast so
that they reach distribution equilibrium, then after the first plunger cycle the
concentration in the organic phase is:
2. DI C, a,initial 'o, film Va,samp plug (11.31)
(rneedle +2 D, o, flm ) Vo
where 6 ,fim is the wetting film thickness, r,eedle is the inner radius of the needle
and V0 is the total 1 l volume of the organic solvent [96]. If the following
additional approximations are made: (i) Ci,a CI,,initiM, because Va >> V0 and
because the extraction is always stopped far short of extraction equilibrium; (ii)
the organic wetting film is completely renewed for each plunger cycle; and (iii) n,
the number of plunger cycles, is not very large so that nV,,,lm < < Vo; then the
relationship between the fraction of analyte component extracted and n is:
6
Cl,. Vo 2 n D,- o, Vasampplg
film (11.32)
Cl,a,initi Va needlee +2' D, '5o,fi ) V
Equation (11.32) predicts that the fraction extracted should increase approxi-
mately linearly with the number of plunger cycles and with the volume of the
aqueous sample plug that is drawn into the needle. This has been experimentally
demonstrated to be true [96]. The wetting film thickness and the closeness of
approach to equilibration between the wetting film and the aqueous sample plug
depend on how fast the plunger is moved [97]. It should be evident from the
details of the mode of contacting the phases that "Dynamic-LPME" more closely
resembles the "sequential injection extraction" mode of segmented flow LLE
(Section 11.2.3) than it does Microdrop-LPME.
Analytically, after the extraction operation the 1 l of organic solvent is
directly injected into a gas chromatograph with either electron capture or mass
spectrometric detection. Since the detector signal is proportional to C,o. V, it is
evident from Eq. (11.32) that if n, V,, pl,ug and the details of plunger movement
are kept constant then a linear calibration curve will be obtained for detector
signal vs C,a,initiaVa in the similarly-extracted aqueous standard which can be
used to determine I in the sample. In the studies cited the solvents toluene,
chloroform, butyl acetate and isooctane were all employed as the organic phase
and a variety of chlorobenzenes were determined at 2 Ag/l concentration in
industrial wastewater with RSDs < 10% in most cases.
Another mode of contacting the phases in Microdrop-LPME, rather than
stirring the aqueous sample solution as in Fig. 11.7, is to cause it to flow past the
microdrop which is positioned in a flow cell of small dimensions. The flow
induces MT in the drop via momentum transfer, producing Case I kinetics. The
flow also continuously brings fresh aqueous sample solution into contact with
the drop, making C,,equil = CI,a,initia in Eq. (11.12).
323
Chloroform
Aqueous
_I,
CHl
P-
Left: Fig. 11.9. Diagram of a flow cell containing a hydrophobic PEEK capillary tube with a
microdrop of organic solvent standing on top of it. Aqueous sample solution flows out the tip end
of the PEEK tube and past the drop. It fills the flow cell and leaves via the top of the flow cell. The
syringe is used to withdraw some of the organic drop for analysis. Adapted from Ref. [98].
Right: Fig. 11.10. Diagram of a flow cell containing an organic microdrop (CHC1) suspended
from the tip of a capillary into a flowing aqueous sampling solution. In the design, the flowing
aqueous phase oscillates in size from filling the cell to the size shown by the dashed curved line.
Detection is performed photometrically in the chloroform drop, in situ. Adapted from Ref. [99].
In one flow design shown in Fig. 11.9, the organic microdrop (1-5 l) is
positioned at the tip of a hydrophobic capillary tube (polyetheretherketone,
PEEK, 0.254 mm i.d.) and the aqueous sample solution is pumped through the
PEEK tube and past the microdrop [98]. The concentration of analyte compo-
nent in the drop is determined by withdrawing 0.5 Al of the drop into the syringe
and injecting it into a gas chromatograph fitted with an electron capture
detector. A series of eleven nitro- and chloro-aromatics were used as analyte
components. The rate of extraction is reported to increase with increasing flow
rate of the aqueous solution, consistent with a decrease in thickness of the
Nernst diffusion films in Eq. (11.16). The fact that the analyte components all
have large D, values suggests that MT rate control may reside in the aqueous
phase, rather than in both phases, but this has not been experimentally verified.
In fact, there is no need to know this in order to perform extractions for
analytical purposes. For a 10-min extraction time with a 1- 1I drop of toluene and
a flow rate of 0.1 ml/min, preconcentration enrichment factors (EF) ranged
from 260 to 1600 depending on D, for the compound, and the fractions extracted
(f,,) were greater than 0.7. RSDs were < 10%. Seawater samples were analyzed.
A second flow design involves an approximately 1-1 drop of chloroform
suspended in an approximately 50-L1 glass flow cell through which the aqueous
sample is pumped (Fig. 11.10) [99]. The analyte dodecylsulfate anion (DS-) was
extracted as a colored ion pair with methylene blue (MB + ) cationic reagent which
was initially added to the aqueous sample solution. After the extraction, water
was pumped instead of sample in order to backwash the organic drop and then
the concentration of MBDS in the chloroform drop was determined by measur-
ing the absorbance of the drop in situ by means of an LED/optical fiber-based
detector (not shown in Fig. 11.10). At the end of the experiment the organic drop
324
Left: Fig. 11.11. Diagram of a drop-in-a-drop which is
used for Microdrop-LPME in an acoustic levitation field.
was pumped out and a fresh one was formed by pumping in chloroform. The
entire instrument is computer controlled. For a 60-s extraction time the extrac-
ted concentration increased with an increase in sample flow rate which is
consistent with Case I or Case II kinetics. It is likely that in this determination,
the MT kinetics are Case II with rate control in the aqueous phase, because DS-
is surface-active, MB-DS is probably surface active and D, is large.
Yet another mode of contacting the phases in Microdrop-LPME employs
acoustic levitation of a drop-in-a-drop (Fig. 11.11). The levitation force is pro-
vided by the acoustic radiation pressure of an ultrasound source [100,101]. In
the experiment described, 0.74 r1 of an aqueous sample solution and about 4 pAl of
toluene were levitated at a pressure node in an ultrasonic field. The drop-in-
drop, with the water drop inside the toluene drop, forms spontaneously. After 2
min, 0.2 Al of the toluene phase is withdrawn from the outer drop with a
microsyringe and 0.1 l of it is injected into a gas chromatograph. The analyte
component was n-hexanol. It was necessary to include an internal standard
(n-octanol) because evaporation of toluene during the 2 min levitation is signifi-
cant. Since the ultrasonic field induces convection in the drops, this is probably
Case I kinetics. RSD was -7% and was attributed to the precision of the GC step
rather than to that of the extraction step.
In some Microdrop-LPME studies the microsyringe set-up shown in Fig. 11.7
has been used with no convection in either phase [96,102]. This gives Case III
kinetics, in which the integral of Eq. (11.23) is the rate equation. In these studies
a 1 or 2 -ul drop of toluene or hexane is suspended in the non-stirred aqueous
sample solution (e.g. 4 ml) for a specified period of time after which the micro-
drop is injected into a gas chromatograph. For a given extraction time the
fraction of I extracted is greater for a large drop volume, but it is significantly
less than when the aqueous phase is stirred. Linear calibration curves are
obtained when the extraction conditions are kept constant. For a 15-min extrac-
tion, typical preconcentration enrichment factors (EFI) of about 10 are obtained.
11.3.4.2 Nanodrop-LPME
LLE from a volume of 1-3 Al of a convected aqueous sample solution into a 1-30
nl drop of chloroform at the tip of a fused silica capillary was used to pre-
concentrate and clean up analyte components in preparation for microspot
matrix-assisted laser desorption-ionization mass spectrometry (MALDI-MS)
[103]. The extraction device is shown in Fig. (11.12). Drop sizes up to 30 nl are
325
Pipette Tip
Capillary Aqueous
Filled with a Sample ( 1-5 pL)
Water-lmmiscible
Organic Solvent
Organic,
Solvent
Drop L
(1-30 nL) A
Rotation by
Small Electric
Motor
Microscopic
Observation [40x]
Right: Fig. 11.12. Diagram of a Nanodrop-LPME device for extracting analyte components from
a rotationally-stirred aqueous phase into a nanodrop formed at the tip of a fused silica capillary.
The microscope is used to help position and monitor the drop. From Ref. [103], with permission.
very stable even at a 3000-rpm rotation speed of the plastic pipet tip containing
the sample solution. Extraction times of 15 min are typical. After the extraction
the organic drop was withdrawn into the capillary, the capillary was removed
from the pipet tip and all the chloroform in the capillary (e.g. 200 nl) was pushed
out and deposited onto the MALDI matrix layer where the chloroform rapidly
evaporates leaving the analyte component as a microspot. Hydrophobic peptides
were separated from large concentrations of interfering hydrophilic peptides
and from interfering metal ions. Also, it is claimed that, since the organic drop
does not touch the wall of the container holding the aqueous sample solution, the
extraction of contaminants that might have been present on the wall is
minimized.
Because the organic drop is in the nanoliter range it will not experience
momentum-induced convection and Case II kinetics prevails, with MT in the
aqueous phase being rate determining. MALDI is not a quantitative technique,
so fraction extracted (fi, ) and preconcentration enrichment factor (EFI ) were not
quantified, but both preconcentration and extent of separation from inter-
ferents () were inferred to be very high.
11.3.4.3 Picodrop-LPME
LLE into picoliter-size drops has been performed with both convected and
non-convected aqueous phases. Shown in Fig. 11.13 is an apparatus for "micro-
injection-microspectroscopy" in which a drop containing between 1 and 80 pl of
tributylphosphate organic phase is formed at the tip of a microcapillary in a glass
microtube (400 m deep x 1 mm side x 10 mm long [104]) The aqueous sample
solution is pumped through the glass microtube at flow rates from 5 to 20 l/min.
The sample solution contains the extractable metal-ligand complex Al-2,2'-di-
hydroxyazobenzene. The concentration of complex in the picodrop is determined
by in situ measurement of its absorbance at 480 nm using the associated optical
microscope system. The kinetics here are Case II, with film-diffusion rate
326
Objective
Lens
Drop - -Pinhole
| Detector |
Fig. 11.13. Diagram of a flow cell containing an organic picodrop past which aqueous sample
solution is continuously pumped. Detection of analyte component in the picodrop is performed
photometrically on the drop in situ using microspectroscopy. Adapted from Ref. [104].
control in the convected aqueous phase. Because fresh aqueous sample solution
is continuously brought to the drop, Cl,, equi = C a,initial.
Since extraction rate studies of C,,o vs. t were performed for only short times,
Ci,1 << Clequi in the first-order rate equation (11.9), and the integrated rate
equation for initial rate is:
As predicted from these equations, experimental plots of C, o vs. t were linear and
their slopes increased with increasing flow rate of aqueous sample, which gives a
smaller 6, and with decreasing drop radius. The preconcentration enrichment
factor of the aluminum complex, for which D, = 162, was EFt - 103 in a 17-pl drop
after a 10-min extraction.
Picoliter-LPME with laser trapping of the drop in an unconvected aqueous
phase has been the subject of several reports [84,87,88,105,106]. More recently,
a microcapillary injector has been used to position a 31-pl drop of nitrobenzene
on top of a small gold disk electrode in a cell containing 2 ml of aqueous sample
solution [89]. Laser trapping was not needed. The concentration of the electro-
active analyte compound 1-ferrocenyl-2-butanol in the nitrobenzene drop was
327
1' . . . I I
determined in situ by a coulometric technique. The fact that the mass transfer
kinetics for this system are Case III was demonstrated experimentally by the
good fit of Eq. (11.24) to the C, vs. t data (Fig. 11.14) and by the linearity of plots
of observed rate constant vs. raT20p [84], as predicted by Eq. (11.25).
328
Since the drop to be analyzed after Single Drop-LPME is composed of a
water-immiscible organic solvent it is readily compatible with direct injection in
gas chromatography, normal phase liquid chromatography, mass spectrometry
and atomic spectrometry. However, it may not be compatible with direct injec-
tion in capillary electrophoresis or reversed-phase liquid chromatography where
the carrier fluid is mainly aqueous. One way of dealing with this potential
problem is to evaporate the drop, perhaps on a small wire loop [107,108] and
redissolve it in a suitable aqueous solvent. Another way is via back-extraction of
the analyte component into an aqueous solution. This latter technique is the
subject of Section 11.4.
It has generally been considered impractical to perform headspaceanalysis
of volatile analyte components above an aqueous sample solution using
Microdrop-LPME in which the microdrop would be suspended in the air above
the aqueous phase. However, very recently it has been demonstrated to be quite
feasible [109]. A 1-1l drop of 1-octanol suspended from the tip of a microliter
syringe in the headspace was shown to be an excellent preconcentration medium
for the BTEX components (benzene, toluene, ethylbenzene and o-xylene) in a
stirred aqueous sample. Quantification of analyte concentration in the micro-
drop was achieved by split-mode injection into a gas chromatograph using a
quadrupole mass spectrometer for detection, with the aid of n-decane as internal
standard. Detailed extraction rate studies of C,, vs. time showed that the
extraction rate is limited by mass transfer in both the aqueous sample solution
and the organic drop. In headspace analysis the drop is not in contact with the
stirred aqueous sample solution so that the drop cannot be convected by momen-
tum transfer. When desired, the drop was convected by gently vibrating the
needle from which it was suspended. This study opens the realm of headspace
analysis for volatile organic compounds (VOCs) by Microdrop-LPME.
In this chapter the emphasis is on the theoretical aspects of LLE, an
understanding of which is useful, and sometimes essential, in designing, opti-
mizing and troubleshooting an LLE technique. However, once a procedure has
been established it is not at all necessary to deal with equations such as Eq.
(11.10) and it is not necessary to know the values of parameters such as D, D,, 6
and A, in order to perform routine LLE operations, such as Microdrop-LPME,
for sample preparation in quantitative analysis. In routine use standards are
always run through the same extraction steps as the sample.
11.4.1 Introduction
329
I I
II ... nr LI
NaH 2PO4 (pH2.1)
.........a2 ) -
-4 Teflon Ring
Fig. 11.15. Unsupported Membrane-LPME in which the Y N
n-octane membrane separates the stirred aqueous sample al N 80 or 40 pL
solution from a layer of aqueous receiving phase. The 14Q n-Octane
wetting properties of the liquid membrane keep it
touching the Teflon ring. The shape of the membrane . \n .= -
results from frictional drag of the toroidally circulating
aqueous sample that can be seen in Fig. 11.16b. From Ref.
[9], with permission. V Aqueous Sample
(pH13)
330
a k
Fig. 11.16. Diagrammatic representation of: (b) The convective-flow circulation patterns in the
device shown in Fig. 11.15; and (a) the resulting Nernst diffusion films of thickness 6. The
horizontal components of circulation are omitted for clarity. From Ref. [9], with permission.
brane and a2. This condition is typically achieved by a pH difference when the
analyte components are acids or bases, as was discussed in Section 11.1.2.
Va 1 V,
in which
and
D 2 = CIo3equil (11.38)
The
fraction
euil of extracted into a2,
331
D' V.
f EF at equilibrium (11.39)
al D 2 + DI,, D 2 V + Dll Va2
From Eq. (11.36) it can be seen that a high preconcentrationis favored by very
small D, 2 and Vo < V02 < V,,al, so that EFrtends toward a high limiting value equal
to Vai/Va,. However, if quantitative extraction of I from al into a2 is the goal, Eq.
(11.39) shows that a very small D,2, is desired along with V < V, and V < V,,, 2 so
that fl, 2 tends toward the value 1. Thus, completeness of extraction is favored by
larger V,, while preconcentration is favored by smaller V,,2. This consideration
has led to two variations in design for Unsupported Membrane-LPME which will
be discussed in connection with Figs. 11.15 and 11.18.
The kinetics of extraction can be treated in terms of the Whitman two-film
model at each LL interface, assuming that both hkINT and kCHEM are much larger
than the mass-transfer coefficients for convective-diffusive MT in the three
liquid phases. Shown in Fig. 11.16A are the locations of the four Nernst diffusion
films. The two steps in Eq. (11.35) can be treated as two reversible, consecutive
first-order processes [9]. The four rate constants in Eq. (11.35) can be expressed
as functions of the four mass transfer coefficients, D 0l/60 l, D,o/6, ,, D,6,,o2 and
D,,/6a2. A set of three differential equations can be written and explicitly solved
to obtain theoretical expressions for the concentrations of I in each of the three
phases as a function of time (see Appendix in Ref. [9]).
Shown as points in Fig. 11.17 are experimentally measured rate curves of
C,2, vs. time for the extraction of two drug compounds, mephentermine and
2-phenylethylamine. The solid lines through the points are nonlinear least
squares fit lines of the theoretical equation to the data. Also shown on the figure
are the theoretically-predicted curves for the concentrations in the organic
membrane phase. The extractions represented in Fig. 11.17 involved 1 ml of
stirred (2050 RPM) aqueous sample (al) at pH 13, 80 ul of n-octane liquid
membrane (o) and 200 ul of pH 2.1 aqueous receiving phase (a2). Quantification
332
was performed by injecting 10 Culof the a2 extract into an HPLC. In al, both
drugs exist nearly exclusively as their neutral free-base species, B, while in a2
they exist nearly exclusively as their cationic conjugate acid species, BH + . At pH
13 D, 1 = 45 for mephentermine and D,1 = 1.1 for 2-phenylethylamine; while at
pH 2.1 D1,2 = 3.2x10 7 for mephentermine and D, 2 = 1.8x10 -8 for 2-phenyl-
ethylamine. Note that the extraction is slower for the 2-phenylethylamine, for
which D, is smaller; but that at equilibrium, which is shown by the horizontal
dotted line, even this compound is quantitatively extracted from 1 ml of al into
0.2 ml of a2. This of course is due to the very small value of D1 ,2 at pH = 2.1,
which provides a large thermodynamic driving force for the overall process.
A much simpler rate equation for C,, vs. t than the one given in the Appen-
dix of Ref. [9] and used in Fig. 11.17 can be obtained by employing the steady-
state approximation for C,, in the organic membrane phase [10], at which moles
of I leaving al and entering a2 are approximately the same. This is justified by
the small fraction of I that is in the membrane at any time. The rate equation is:
overall - k3 (11.41)
kg +k3
and tag is the "lag time" constant that is related to the time required to reach
diffusional steady-state across the Nernst film [82,112]. (Note that there are two
different steady-state concepts here.) Equation (11.40) can be rearranged to give
the preconcentration enrichment factor as a function of time:
It has been shown that a larger value of kov, 0rl is favored by faster convection
(smaller 6's) and larger Va2 (larger A2).
When high preconcentration is desired, as opposed to quantitative
extraction, then V= should be minimized because the maximum EF I is Va1/V2.
The apparatus shown in Fig. 11.18 employs a microdrop of receiving phase
suspended in the organic liquid membrane. Shown in Fig. 11.19 are extraction
rate curves for mephentermine and 2-phenylethylamine plotted as EF vs. t [10].
Here V., is 1.6 ml (pH 13), V is 30 gl and V, is 0.5 Al (pH 2.5). The solid lines are
for nonlinear least squares fitting of Eq. (11.42) to the data points. The lag time
at which the line intersects the time axis is 1.2 min for mephentermine and 0.7
min for 2-phenylethylamine. The fact that, after their lag times, these two
first-order rate curves deviate only slightly from linearity is an indication that
after 25 min they are still far from extraction equilibrium. This is because the
small V, has resulted in small rate constants (ko-el = 0.024 min- and 0.0056
min' for mephentermine and 2-phenylethylamine, respectively). Nevertheless,
333
Fig. 11.18. Apparatus for Unsupported Membrane-LPME with a microliter drop of receiving
phase suspended in the membrane from the tip of a microsyringe. The shape of the liquid
membrane results from frictional drag by the toroidally circulating aqueous sample phase. From
Ref. [10], with permission.
151
even after only 10 min, 2-phenylethylamine for which D,1 = 1.1 has already been
preconcentrated over 100 fold. By contrast if only forward extraction from
aqueous pH 13 into n-octane were employed, this compound would have a
maximum possible preconcentration of only 1.1-fold, at equilibrium (Eq. 11.5).
334
preconcentrated analyte component is present in a microdrop which can be
injected in toto avoids the need to evaporate the extract prior to injection. The
technique also represents a second clean-up step for separating potentially
interfering sample components from the analyte component. The liquid
membrane itself is replaced for every sample and standard, eliminating concerns
of carry-over and memory. The membrane is very easy to prepare.
On the negative side, back extraction for sample clean up and precon-
centration is practical only when it is possible to make D, 2 < < D,,l by means of
differences in composition of the aqueous sample and receiving phases such as
different pH for acids and bases, the presence of ligands for metal ions and the
employment of other types of secondary chemical equilibria.
The possibility of increasing preconcentration (EF,) dramatically by
replacing microliter drops of Va2 with nanoliter or picoliter drops is very real.
Such drop volumes are compatible with the injection volumes required by
capillary separation techniques. One can easily imagine an electrophoresis
capillary replacing the microsyringe that protrudes into the liquid membrane
phase in Fig. 11.18; with a nanoliter drop of its running buffer replacing the
microliter drop on the syringe. After extraction the nanodrop could be brought
back into the capillary, the capillary removed from the vial and CZE commenced.
11.5.1 Introduction
The fact that an analyte component (I) may be present in an aqueous sample
solution as more than one kinetically labile (i.e. rapidly interchanging) species
was discussed briefly in Section 11.1.2. It led to using the component distribu-
tion ratio D, in place of the species distribution coefficient Ki. Other aspects of the
present discussion were foreshadowed in Section 11.1.8. Here we consider this
situation from another perspective, asking the question: "When I perform a
sample preparation step that involves distribution of the analyte between two or
more phases, have I recovered an amount of analyte that is proportional to the
concentration of the analyte component or to a species of that component?"
These considerations apply equally to LLE, SPE, SPME, purge-and-trap
methods, etc.
Consider for simplicity the example that was started in Section 11.1.2: that
of a basic analyte compound which is present in the sample solution as two
species B and BH+. The fraction of each is determined by the pH. Let us say that
the pH happens to be equal to PK,,BH. Now imagine that we also prepare a
standard of this compound by dissolving the pure free base, B, in water. We
perform LLE on both the sample and standard solutions, completely oblivious of
the fact that in the sample only 50% of the analyte component is present as the
species B, while in the standard nearly 100% is present as B. The extraction
equation is:
335
Ba B (11.43)
for which the species distribution coefficient is iB. Since the species BH+ does not
distribute between the two phases, ,,H+ = 0. We now perform GC on the extracts
from both sample and standard and calculate the concentration of the compound
in the sample as the product of the ratio of sample GC signal to standard GC
signal times the concentration in our original aqueous standard solution. Will we
be determining the component concentration (B + BH+) in the sample or will we
merely have determined the concentration of the species B?
Combining the equilibria
BH + + H20 - =>B a + H0 +
,T' KB (11.44)
Bo,
If the extraction conditions are exhaustive, such as a very large KB, so that D in
Eq. (11.3) is large then f, in Eq. (11.4) could well be close to 1 meaning that the
extraction of B has shifted the acid ionization equilibrium in Eq. (11.44) far to
the right and we have quantitatively removed from the aqueous sample solution
virtually all of both BH + B. That is, we have determined component concen-
tration. We say that the acid-base equilibrium has been completelyperturbedby
the extraction.
On the other hand, if the extraction conditions that we are using yield a very
small B or if we are operating very far from extraction equilibrium then the
extraction of B will be slight so that [B]a,,,eqi = [B]a,iitia, and [B] o, equil = 1KB [B]a,initi.a
and we have determined species B concentration. Here we say that the acid-base
equilibrium has been left completely unperturbed. Obviously, intermediate
extraction conditions yield intermediate degrees of perturbation.
Often the conditions that are used for solvent extraction remove only some
fraction of the analyte component and do not quantitatively (exhaustively)
extract it; while a different fraction is extracted from the standard (see Section
11.1.8). The obvious and most notable condition leading to non-perturbation of
the sample occurs in those techniques in which the aqueous sample solution is
caused continuously to flow past a small drop of organic phase, maintaining the
aqueous concentration at C,ainitia. If the analyte component is present as more
than one species in the sample but not the standard then we will not have
determined the total concentration of the analyte component.
Looked at from a different perspective, an awareness of perturbation and its
causes tells us how we can use LLE (or SPE, SPME, etc.) to measure the
concentration of a particular species in a sample solution. This is called a
speciationmeasurement. One example is the determination of the concentration
of free (i.e. unbound) progesterone species (P) in the presence of the protein
bovine serum albumin (BSA) which binds P, using Microdrop-LPME [113].
336
The equilibria are:
Pa + BSAa i P. BSAa
I? (11.45)
Po
Cu(Lix63)o
REFERENCES
1 L.C. Craig and D. Craig, in: A. Weissberger (Ed.), Techniques of Organic Chemistry,
Vol. III, 2nd Edn. Interscience, New York, 1956.
2 H. Irving and R.J.P. Williams, in: I.M. Kolthoff and P.J. Elving (Eds.), Treatise on
Analytical Chemistry, Vol. 3, Part I. Interscience, New York, 1961, Ch. 31.
3 R.F. Majors, LC-GC, 14 (1996) 754.
4 J. Stary, The Solvent Extraction of Metal Chelates. MacMillan, New York, 1964.
5 A. Ringbom, Complexation in Analytical Chemistry. Wiley, New York, 1973.
6 G. Schill, in: J. Marinsky and Y. Marcus (Eds.), Ion Exchange and Solvent Extraction,
Vol. 6. Marcel Dekker, New York, 1974, Ch. 1.
7 M.A. Jeannot and F.F. Cantwell, Anal. Chem., 68 (1996) 2236.
8 P.R. Rony, Separat.Sci., 3 (1968) 239.
9 M. Ma and F.F. Cantwell, Anal. Chem., 70 (1998) 3912.
10 M. Ma and F.F. Cantwell, Anal. Chem., 71 (1999) 388.
11 D.J. Shaw, Introduction to Colloid and Surface Chemistry. Butterworths, London,
1980, Ch. 4 and 10.
12 T.E. Belk, Chem. Eng. Progr., 61 (1965) 72.
13 J.C. Lee and T.D. Hodgson, Chem. Eng. Sci., 23 (1968) 1375.
14 G.S. Laddha and T.E. Degaleesan, in: T.C. Lo, H.M. Baird and C. Hanson (Eds.),
Handbook of Solvent Extraction. Wiley, New York, 1983.
15 R.E. Majors, LC-GC, 14 (1996) 936.
16 P. Danesi and R. Chiarizia, Crit.Revs. Anal. Chem., 10 (1980) 1.
337
17 E.L. Cussler, Diffusion: Mass Transfer in Fluid Systems. Cambridge University
Press, London, 1984.
18 I. Nahringbauer and B. Larsson, Anal. Chim. Acta, 151 (1983) 153.
19 J.T. Davies and E.K. Rideal, Interfacial Phenomena. Academic Press, New York,
1961, Ch. 7.
20 M. Harada, T. Imamura, K. Fujiyoshi and W. Eguchi, J. Chem. Eng. Jpn., 8 (1975)
233.
21 W.J. Albery, J.F. Burke, E.B. Leffler and J. Hadgraft, J. Chem. Soc. FaradayTrans.,
72 (1976) 1618.
22 E.L. Cussler, Diffusion: Mass Transfer in Fluid Systems. Cambridge University
Press, London, 1984, Ch. 9, 11 and 13.
23 0. Levenspiel, Chemical Reaction Engineering. Wiley, New York, 1972, Ch. 11 and
13.
24 G.J. Hanna and R.D. Noble, Chem. Rev., 85 (1985) 583.
25 G.S. Laddha and T.E. Degaleesan, TransportPhenomenain Liquid Extraction. Tata
McGraw-Hill, New Delhi, 1976, Ch. 3, 6, 7 and 15.
26 W. Nitsch and A. Hoffmann, Chem. Ing. Tech., 53 (1981) 367.
27 R. Fleming, R.H. Guy and J. Hadgraft, J. Pharm. Sci., 72 (1983) 142.
28 V.G. Levich, Physicochemical Hydrodynamics, Prentice-Hall, Englewood Cliffs, NJ,
1962, Ch. 8.
29 G. Persaud and F.F. Cantwell, Anal. Chem., 59 (1987) 2.
30 M. Linton and K.L. Sutherland, Proc. 2nd Int. Conf. Surface Activity, 1 (1957) 494.
31 M.J. Slater, in: J.C. Godfrey and M.J. Slater (Eds.), Liquid-Liquid Extraction
Equipment. Wiley, New York, 1994, Ch. 4.
32 R.G. Borwanker and D.T. Wassan, Chem. Eng. Sci., 43 (1988) 1323.
33 D.C. England and J.C. Berg, Am. Inst. Chem. Eng., 17 (1971) 313.
34 G.C. Quintana, in: R.P. Chhabra and D. De Kee (Eds.), Transport Processes in
Bubbles, Drops and Particles. Hemisphere Publishing, New York, 1992, Ch. 4.
35 W. Nitsch and K. Roth, Colloid Polym. Sci., 256 (1978) 1182.
36 W. Nitsch and G. Weber, Chem. Ing. Tech., 48 (1976) 715.
37 L.T. Skeggs, Am. J. Clin. Pathol., 28 (1957) 311.
38 W.B. Furman, Continuous Flow Analysis: Theory and Practice.Marcel Dekker, New
York, 1976.
39 C.J. Patton and A. Wade, in: G.W. Ewing (Ed.), Analytical Instrumentation
Handbook. Dekker, New York, 1990, Ch. 27.
40 V. Wallace, Anal. Biochem., 20 (1967) 411.
41 D.A. Burns, Anal. Chem., 53 (1981) 1403A.
42 L. Valentine, Advances in Automated Analysis, Technicon International Congress,
Vol. II. Mediad, Inc., White Plains, New York, 1970, 87.
43 K.K. Stewart, G.R. Beecher and R.E. Hare, Fed. Proc., 33 (1974) 1439.
44 J. Ruzicka and E.H. Hansen, Anal. Chim. Acta, 78 (1975) 145.
45 B. Karlberg and S. Thelander, Anal. Chim. Acta, 98 (1978) 1.
46 H. Bergamin, J.X. Madeiros, B.F. Reis and E.A.G. Zagatto, Anal. Chim. Acta, 101
(1978) 9.
47 B. Karlberg, P-A. Johansson and S. Thalander, Anal. Chim. Acta, 104 (1979) 21.
48 J.F.M. Kinkel and E. Tomlinson, Int. J. Pharmaceutics, 6 (1980) 261.
49 C.A. Lucy, Ph.D. Thesis, University of Alberta, 1988.
50 L. Fossey and F.F. Cantwell, Anal. Chem., 54 (1982) 1693.
51 V. Kuban, Crit. Rev. Anal. Chem., 22 (1991) 477.
52 C.A. Lucy and S. Varkey, Anal. Chem., 67 (1995) 3036.
53 F.F. Cantwell and J.A. Sweileh, Anal. Chem., 57 (1985) 329.
54 V. Kuban and F. Ingman, Crit. Revs. Anal. Chem., 22 (1991) 491.
338
55 J.A. Sweileh and F.F. Cantwell, Can. J. Chem. 63 (1985) 2559.
56 L. Fossey, Ph.D. Thesis, University of Alberta, 1985.
57 M. Valcarcel, A. Rios and L. Arce, Crit. Revs. Anal. Chem., 28 (1998) 63.
58 C.A. Lucy and F.F. Cantwell, Anal. Chem., 58 (1986) 2727.
59 L. Nord, K. Backstrom, L.G. Danielsson, F. Ingman and B. Karlberg, Anal. Chim.
Acta, 194 (1987) 221.
60 L. Nord, Ph.D. Thesis, Royal Institute of Technology, Stockholm, 1984.
61 C.A. Lucy and F.F. Cantwell, Anal. Chem., 61 (1989) 101.
62 C.A. Lucy and F.F. Cantwell, Anal. Chem., 61 (1989) 107.
63 H.T. Evensen, D.R. Meldrum and D.L. Cunningham, Rev. Sci. Instr., 69 (1998) 519.
64 E. Castaneda, M.Sc. Thesis, University of Alberta, 1986.
65 C.A. Lucy and K.K.C. Young, Anal. Chem., 66 (1994) 2220.
66 C.A. Lucy and B.P. Housermann, Anal. Chim. Acta, 307 (1995) 173.
67 C.C. Lindgren and P.K. Dasgupta, Talanta, 39 (1992) 101.
68 I. Facchin, J.W. Martins, P.G.P. Zamora and C. Pasquini, Anal. Chim. Acta, 285
(1994) 287.
69 K. Carlsson and B. Karlberg, Anal. Chim. Acta, 415 (2000) 1.
70 I. Facchin and C. Pasquini, Anal. Chim. Acta, 308 (1995) 231.
71 L. Nord and B. Karlberg, Anal. Chim. Acta, 164 (1984) 233.
72 K.L. Peterson, B.K. Logan, G.D. Christian and J. Ruzicka, Anal. Chim. Acta, 337
(1997) 99.
73 L. Nord and B. Karlberg, Anal. Chim. Acta, 145 (1983) 151.
74 T.M. Rossi, D.C. Shelby and I.M. Warner, Anal. Chim., 54 (1982) 2056.
75 W. Nitsch, Dechema Monogr., 55 (1965) 143.
76 A. Brodin and A. Agren, Acta Pharm. Suecica, 8 (1971) 609.
77 T. Tanimura, J.J. Pisano, Y.Ito and R.L. Bowman, Science, 169 (1970) 54.
78 Y. Ito, in: N.B. Mandava and Y. Ito (Eds.), Counter-CurrentChromatography:Theory
and Practice. Dekker, New York, 1988, Ch. 3.
79 K. Hostettmann and A. Marston, in N.B. Mandava and Y. Ito (Eds.), Counter-Current
Chromatography:Theory and Practice, Dekker, New York, 1988, Ch. 5.
80 W. Conway, in: N.B. Mandava and Y. Ito (Eds.), Counter-CurrentChromatography:
Theory and Practice. Dekker, New York, 1988, Ch. 3.
81 M.A. Jeannot and F.F. Cantwell, Anal. Chem., 69 (1993) 235.
82 J. Crank, The Mathematicsof Diffusion. Clarendon Press, Oxford, 1975, Ch. 4 and 5.
83 W.J. Moore, Physical Chemistry. Prentice-Hall, New York, 1962, Ch. 9.
84 K. Nakatani, T. Uchida, N. Kitamura and H. Masuhara, J. Electroanal.Chem., 375
(1994) 383.
85 P.C. Blokker, Proc. 2nd Int. Conf. Surface Activity, 1 (1957) 503.
86 A.J. Bard and L.R. Faulkner, Electrochemical Methods: Fundamentals and
Applications. Wiley, New York, 1980, Ch. 5.
87 K. Nakatani, T. Uchida, H. Misswa, N. Kitamura and H. Mashuhara, J. Phys. Chem.,
97 (1993) 5197.
88 K. Nakatani, M. Wakabayshi, K. Chikama and N. Kitamura, J. Phys. Chem., 100
(1996) 6749.
89 K. Nakatani, M. Sudo and N. Kitamura, Anal. Chem., 72 (2000) 339.
90 M.A. Jeannot, Ph.D. Thesis, University of Alberta, 1997.
91 P.C. Hiementz, Principles of Colloid and Surface Chemistry. Dekker New York,
1986, Ch. 6.
92 K.E. Miller and R.E. Synovec, Talanta, 51 (2000) 921.
93 L.S. de Jager and A.R.J. Andrews, Chromatographia,50 (1999) 733.
94 L.S. de Jager and A.R.J. Andrews, Analyst (London), 125 (2000) 1943.
95 L.S. de Jager and A.R.J. Andrews, J. Chromatogr.A, 911 (2001) 97.
339
96 Y. He and H.K. Lee, Anal. Chem., 69 (1997) 4634.
97 Y. Wang, Y.C. Kwok, Y. He and H.K. Lee, Anal. Chem., 70 (1998) 4610.
98 W. Liu and H.K. Lee, Anal. Chem., 72 (2000) 4462.
99 H. Liu and P.K. Dasgupta, Anal. Chem., 68 (1996) 1817.
100 E. Welter and B. Neidhart, Fres. J. Anal. Chem., 357 (1997) 345.
101 R. Eberhardt and B. Neidhart, Fres. J. Anal. Chem., 365 (1999) 475.
102 T. Ligor and B. Buszewski, ChromatographiaSupplement, 51 (2000) S279.
103 B.O. Keller and L. Li, Anal. Chem., 73 (2001) 2929.
104 0. Kogi, H.B. Kim and N. Kitamura, Anal. Chim. Acta, 418 (2000) 129.
105 K. Nakatani, T. Uchida, H. Misawa, N. Kitamura and H. Masuhara, J. Electroanal.
Chem., 367 (1994) 109.
106 K. Nakatani, M. Sudo and N. Kitamura, J. Phys. Chem. B, 102 (1998) 2908.
107 P.K. Dasgupta and K. Surowiec, Anal. Chem., 68 (1996) 4291.
108 P.K. Dasgupta, K. Surowiec, Anal. Chem., 68 (1996) 1164.
109 A.L. Theis, A.J. Waldack, S.M. Hansen and M.A. Jeannot, Anal. Chem. ASAP Article
AC015569c, Web Release Date October 30, 2001.
110 S. Watanabe, T. Tani, S. Watanabe and M. Seno, Biochim. Biophys. Acta, 1073 (1991)
275.
111 M. Ma, D. He, Q. Wang and Q. Xie, Talanta, 55 (2001) 1109.
112 R.H. Guy and R. Fleming, J. Colloid Interface Sci., 83 (1981) 130.
113 M.A. Jeannot and F.F. Cantwell, Anal. Chem., 69 (1997) 2935.
114 Y. Wang, Ph.D. Thesis, University of Alberta, 2000.
340
Chapter 12
12.1 INTRODUCTION
Solid-phase extraction (SPE) is a method used for the isolation and concentra-
tion of selected analytes from a gas, fluid or liquid (usually) flowing sample
stream by their transfer to and retention (sorption) on a solid phase. The solid
phase is then isolated from the sample and the analytes recovered by elution
using a liquid or fluid, or by thermal desorption into the gas phase. Occasionally
the sampling step is performed by suspension of a suitable SPE device in a liquid,
fluid or gas phase long enough to reach equilibrium with the sample, or some
other reproducible end-point, followed by mechanical separation of the solid
phase from the sample. The analytes are then recovered by elution or thermal
desorption, as before. The principal goals of SPE are trace enrichment (concen-
tration), matrix simplification (sample clean-up) and medium exchange
(transfer from the sample matrix to a different solvent or to the gas phase). SPE
was initially developed as a complement or replacement for liquid-liquid extrac-
tion (LLE). It is now the most common sampling technique in many areas of
chemistry, including environmental, pharmaceutical, clinical, food and indus-
trial chemistry. Over time various sampling formats and sorbents have been
developed to facilitate the convenient processing of different sample types and to
extend the scope of the method. A high level of automation is also possible using
robotics or on-line interfaces to separation and spectroscopic instruments. Full
system integration for unattended extraction, separation and detection is
possible although flexible, inexpensive manual sample processing remains the
more common practice in most laboratories.
In this chapter I will describe the principles and practice of SPE for the
analysis of liquid samples using cartridge and disc devices. SPE using cartridges
for atmospheric sampling and solid-phase microextraction, in-tube solid-phase
microextraction and stir bar extraction are discussed elsewhere in this book,
while specific applications of SPE can be found in the applications section and in
Refs. [1-6].
Fortuitous application of the SPE principle predates written history, but its gen-
eral use in laboratory and field sciences is more recent. Inorganic oxide columns
were used from the 1930s for the clean-up of samples in organic solvents, but
this was rarely considered an extraction process. Probably the first significant
analytical application of SPE was the use of metal pipes filled with activated car-
bon in the early 1950s for the isolation of organic contaminants from various
water sources [7]. These studies were performed to isolate sufficient material for
structural identification with the relatively insensitive instruments available at
TABLE 12.1
Milestones in the development of solid-phase extraction
Early 1950s Pipes filled with activated charcoal used to isolate microcontaminants from
surface waters.
Late 1960s Macroreticular porous polymers introduced as a replacement for active charcoal.
Standardized protocols developed using small columns. Applications to biological
fluids and air sampling developed.
Mid 1970s Silica-based chemically bonded phases used in solid-phase extraction and
eventually dominate the practice of solid-phase extraction. Disposable cartridges
introduced in 1978 and the open syringe barrel format in 1979.
Early 1980s Short precolumns introduced for automated solid-phase extraction and liquid
chromatography. This did not become a routine technique until the early 1990s.
Mid 1980s Start of the development of sorbents with improved molecular recognition
properties. Mixed mode (reversed phase/ion exchange) sorbents introduced for
improved isolation of drugs from biological fluids (1985). Restricted access
materials introduced for the selective isolation of drugs in biological fluids
without prior removal of proteins (1985). Immunosorbents developed to take
advantage of the specificity of antibodies for the isolation of target analytes from
complex mixtures (Biomedical applications developed in the early 1980s followed
by specific procedures for food contaminants and then environmental
applications by the mid-1990s). Silica or other supports with tethered mixed
oxygen-nitrogen donor cryptands introduced for the selective isolation of metals
by a molecular recognition mechanism (1988).
1989 Matrix dispersion introduced as a sample processing technique.
1989 Particle-loaded membranes for disc extraction introduced. Soon followed by
particle-embedded glass fiber discs and laminar discs. Starts the development of
disc technology as a complement to cartridge sampling devices.
1992 Solid-phase microextraction is brought to commercial fruition.
1994 Molecularly imprinted polymers introduced as synthetic analogs of
immunosorbants.
Late 1990s Multiwell plates for high throughput sample processing introduced. Stir bar
extraction (1999) and in-tube solid-phase microextraction (early work in the
1980s but no general acceptance until late 1990s as an interface for microcolumn
liquid and gas chromatography) continue the development of new formats for
convenient sampling using a solid-phase sorption mechanism.
342
that time and for toxicity assessment. The large volume of water generally sam-
pled (>1000 liters over several days) meant that LLE procedures were inappro-
priate or difficult to implement. Since then a number of different sorbents and
formats for SPE have been introduced, as summarized in Table 12.1.
The heterogeneous nature of activated carbons available at that time ham-
pered any extensive development of SPE. These materials had poorly defined
properties characterized by variable and weak adsorption of some analytes, irre-
versible adsorption and low recovery of others, and occasional chemical modifi-
cation of some analytes [8]. The introduction of poly(styrene-divinylbenzene)
and poly(methacrylate) macroreticular porous polymer beads, prepared by sus-
pension polymerization, in the late 1960s was responsible for rekindling interest
in the SPE of organic contaminants from water and for extending the scope of
SPE to air sampling and the analysis of biological fluids [6,7,9-13]. Polymer
sorbents played a key role in the development of purge-and-trap techniques for
the routine determination of volatile organic compounds in water [14,15]. These
chemically inert sorbents had reasonable mechanical strength (at least com-
pared to gels), with a large surface area, high sample capacity, low water reten-
tion and provided high sample recoveries by either solvent elution or thermal
desorption. Compared with carbon analyte recovery was generally better and
irreversible adsorption and catalytic activity greatly diminished. Standardized
sampling procedures using small columns, early versions of modern cartridge
devices, were introduced but the problem of extract contamination by sorbent
impurities remained a persistent problem.
SPE for liquid samples became a widely used laboratory technique in the
early 1980s with the introduction of disposable cartridges containing silica-
based chemically bonded sorbents of a suitable size for sample processing by
gentle suction. Typical cartridge devices consist of short columns (generally an
open syringe barrel) containing a sorbent with a nominal particle size of 50-60
Am, packed between porous plastic or metal frits (Fig. 12.1) [1-4,6,16-19]. A
large number of non-selective and selective sorbents (including restricted access
media, mixed-mode sorbents, immunosorbents, molecularly imprinted poly-
mers, chemically bonded cryptands, etc.) are available today providing for the
diverse application base of modern SPE. Low volume cartridge or precolumn
devices are the basis of on-line hyphenated systems (SPE-LC, SPE-GC) for
Sample
(polyl
Fritted
(20pmI
Sorben
Fritted
Luer tip
Fig. 12.1. Schematic diagram showing the typical construction of a solid-phase extraction
cartridge and a vacuum manifold for parallel sample processing.
343
Reservoir
Clamn
" ' r
disk.
Collection
( tube
Fig. 12.2. Typical cartridge and vacuum filtration formats for solid-phase extraction using discs.
integrated sample processing and separation that matured into robust practical
systems in the mid-1990s [6,19-23].
SPE discs were developed initially to provide higher sample processing rates
for large sample volumes and to minimize plugging by suspended particles and
matrix components [24-29]. Since small-diameter discs are easily prepared,
discs are also favored for handling small-volume samples as well. At least three
different disc formats are offered today. Particle-loaded membranes contain
sorbent particles of 8-12-utm diameter immobilized in a web of short poly(tetra-
fluoroethylene) (PTFE) fibrils [24-27]. These are formed into 0.5-mm thick discs
of various sizes containing about 90% by weight of sorbent. The discs are flexible
and superficially resemble filter paper discs. They are used with some support-
ing structure such as a porous glass or plastic support (Fig. 12.2). Particle-loaded
membranes are also available in a conventional cartridge format. In this case,
the sorbent bed contains particles of a larger size, about 50 Arm, in a thicker disc
of about 1.0 mm, sealed into the base of an open syringe barrel. These discs have
an integral poly(propylene) prefilter attached to their top surface. Parti-
cle-embedded glass fiber discs contain particles of about 10-30 Am in diameter
woven into a glass fiber-supporting matrix [28,29]. Small diameter discs are
self-supporting but larger discs require a supporting structure similar to the par-
ticle-loaded membranes. Particle-embedded glass fiber discs are also available
with an integral prefilter attached to their top surface. Laminar discs contain
10-,tm sorbent particles in a consolidated 0.5- or 1.0-mm thick bed (usually)
retained by two glass fiber filters held in place by screens and a retaining ring in
a preassembled cartridge [301]. Disc technology has contributed directly to the
automation of SPE through development of the multiwell extraction plate (Fig.
12.3), which is used for isolation and sample clean-up in high-throughput screen-
ing techniques, such as combinatorial chemistry and drug development [31-36].
344
Fig. 12.3. Multiwell plate for automated SPE. (From Ref. [31]; copyright Elsevier).
345
study where one technique emerges as a preferred choice from consideration of
the positive and negative features of each method. Increasingly as external
factors dictate a reduction in the use of popular solvents this has tended to favor
SPE. Also, as laboratories look at approaches to increase sample throughput and
reduce labor costs, this has also tended to favor SPE methods. For field sampling
SPE provides a simpler approach than solvent extraction methods.
Cartridge and disc devices use the same sorbent technology and are
distinguished only by differences in format. Cartridges are easily prepared in the
laboratory while discs, so far, have only been produced in a manufacturing
setting. Discs, therefore, are only available in a limited range of sorbent chemis-
tries, largely dictated by market forces. Also, it is easier to scale-up cartridge
devices to handle large sample loads and to perform sample clean-up than it is
for discs. Discs are significantly more expensive than cartridges and for
relatively simple, routine applications economic consideration often dictate the
use of cartridges. Discs, on the other hand, offer specific format advantages that
favor their use for some applications.
For large sample volumes containing suspended particles, discs are likely to
function better than cartridges. Discs provide shorter sample processing times
on account of their larger cross-sectional area and decreased pressure drop,
allowing higher sample flow rates. This is important in environmental surveil-
lance programs, such as the analysis of surface waters for trace contaminants,
where large sample volumes are commonly employed to achieve adequate detec-
tion limits.
Discs have a large surface area per unit bed mass. This facilitates their use
for passive sampling by immersing the disc in the sample as opposed to the
conventional approach of passing the sample through the disc in a manner
similar to filtration [37-391. Passive sampling is convenient for both field and
laboratory applications, but has been little explored so far. One reason is the
slow equilibrium of the extraction process even for stirred solutions. The sorp-
tion of organic contaminants from water by an octadecylsiloxane-bonded silica
disc using passive sampling under conditions that avoid sample depletion was
used as an indirect measure of bioconcentration by biota [40,41]. The disc
concentrates the more hydrophobic compounds from water in preference to
hydrophilic compounds and is theorized to be a more realistic approach to
assessing potential environmental consequences of organic contaminants in
complex environmental systems. This is a relatively new approach to toxicity
assessment and requires further work to establish a satisfactory relationship
between the sorption properties of the disc and those of target aquatic species.
Because of the low packing density of typical cartridge devices, longer
sorbent beds than are needed for extraction are used to compensate for reduced
retention resulting from channeling. Increased bed mass results in increased
346
non-specific matrix adsorption and dirtier extracts. The use of smaller particles
and the greater mechanical stability of discs minimize channeling, and the opti-
mized use of bed mass results in a cleaner chemical background due to reduced
matrix adsorption. For small sample sizes it is easier to miniaturize discs than
cartridges, and several disc devices (e.g. microdiscs, pipette tips, etc.) that con-
tain only a few milligrams of sorbent are available. Immobilized analytes on
microdiscs facilitate integrated sample processing techniques such as in-vial
desorption [42-44] and on-disc derivatization [42,44,45]. Solid-phase deriva-
tization techniques for conventional sorbent cartridges are also described [46].
In-vial desorption reduces the amount of organic solvent used for recovery and
eliminates tedious sample preparation steps compared to conventional sample
processing methods. The in-vial desorption technique is an equilibrium-based
approach to sample recovery. After extraction the dried disc is placed in an
autosampler vial and covered with a few ml of solvent. Recovery depends on the
selection of the solvent, temperature and equilibration time. An equilibration
time of 4 h, or overnight, at room temperature is sufficient in many cases to pro-
vide acceptable recovery (> 90%) and method precision as well as being compatible
with automated sample analysis in gas and liquid chromatography. Addition of a
derivatizing reagent to the solvent used for desorption allows derivatization and
analyte recovery to be performed simultaneously in the same vial without removal
of the disc. The vial can be sealed and heated to enhance the rate of reaction with
the derivatizing reagent. Common reactions are trialkylsilylation and alkylation
for gas chromatographic analysis. Ion-exchange discs catalyze the rate of alkyla-
tion of acid herbicides, surfactants and pesticides using a solution of an alkyl
iodide as the derivatizing reagent. An example is shown in Fig. 12.4 [42].
The disc format supports other, if minor, applications at present, such as in
situ detection using radioactivity counting [47], phosphorescence [48] and
MALDI mass spectrometry [49,50]. The particle-embedded glass fiber discs can
be inserted directly into a hole in a glass fiber sheet and developed in a conven-
tional manner for thin-layer chromatography combining recovery and separa-
tion into a single step [51,52]. This approach is the basis of the Toxi-Lab®
system for toxicological screening of biological fluids.
The most important inorganic oxide adsorbents for SPE are silica gel, alumina,
Florisil and diatomaceous earth. Silica gels used for SPE have surface areas of
about 300-800 m2/g, pore sizes from 4-10 nm, and an apparent pH of 5.5-7.5
(measured indirectly as the pH of a 5% (m/m) sorbent suspension in water).
Alumina used for SPE has a surface area of about 150 m2/g and a pore size of 6
nm. Depending on processing conditions it is available as neutral (pH 7.5 ± 0.5),
weakly acidic (pH 6.0 ± 0.5), acidic (pH 4.5 ± 0.5) and basic (pH 9.5 ± 0.5) forms.
347
C
C
0
o
Q.
2
a
W
O
LUI
Time (min)
Fig. 12.4. Use of in-vial desorption and derivatization for the determination of chlorinated acid
herbicides in an extract from a spiked (< 10 aug/l each) river water sample by gas chroma-
tography. A 500-ml sample was processed through a 25-mm strong anion-exchange disc, the disc
dried by suction, sealed in a autosampler vial with 1 ml of acetonitrile and 0.2 ml of methyl iodide
and heated at 80°C for 1 h, and excess methyl iodide removed by evaporation. (From Ref. [42];
copyright Elsevier).
348
TABLE 12.2
General application of inorganic oxide adsorbents
Uses:
· Isolation of low and medium polarity analytes from non-aqueous solutions
· Isolation of cations (alumina and silica) and anions (alumina) from buffered aqueous
solutions
* Matrix simplification by fractionation into groups containing a similar number and type of
functional group
Examples:
· Isolation of organochlorine pesticides and polychlorinated biphenyls from transformer oil,
animal fats and oils, etc., using Florisil.
* Isolation of lipids by chromatography over silica gel using chloroform to elute simple lipids,
acetone to elute glycolipids and methanol to elute phospholipids [54].
* Group fractionation of polycyclic aromatic compounds (hydrocarbons, N-containing, and
OH-containing) in synthetic fuels over alumina using a step solvent gradient.
* Isolation of paraquat and diquat from high moisture crops in a pH 9 aqueous extract using
silica gel [55,56]
* Mycotoxins in feeds using silica gel
* Pesticides in foods, feeds, and soil extracts [57]; alkaloids, pigments and flavor compounds
from plants; sugars and caffeine in cola beverages, inorganic anions, and organic acids in
aqueous solution using alumina; steroids and vitamins from creams and oil-based
suspensions
Low-specificity sorbents are widely used for the isolation of contaminants from
aqueous solution. Because of the prevalence of water as a general solvent in
environmental, biological and industrial chemistry these methods represent the
most common applications of SPE. Typical sorbents include silica-based,
chemically bonded materials, porous polymers [18,58-60] and graphitic carbon
[8,61]. Silica-based, chemically bonded sorbents can be prepared with a wide
range of bonding density, pore sizes and functional group types (Table 12.3).
They are generally synthesized by reaction of monofunctional or trifunctional
silanes with silica gel followed by endcapping in some cases. Trifunctional
reagents result in sorbents with a polymeric bonded layer of higher carbon
loading and greater pH stability. Silica gels of high surface area, 500-600 m2 /g,
are generally used to prepare sorbents for the isolation of small molecules.
Chemically bonded sorbents with high surface areas, long alkyl chains and high
phase loading maximize retention of small molecules from aqueous solution
while wide pore materials with a low phase loading and short alkyl chains are
used to isolate macromolecules (Table 12.4) [1,4,5,62-69]. Chemically bonded
sorbents with polar functional groups are used mainly to isolate analytes from
organic solvents based on their selective interactions with analyte polar
functional groups. They are less retentive than alkylsiloxane-bonded phases for
aqueous samples. When the sampling process is not dominated by breakthrough
349
TABLE 12.3
Structures of silica-based chemically bonded sorbents
Type Functional group Structure
C18 Octadecyl Si-C18H27
C8 Octyl -SiCH,7
C2 Ethyl Si-C2H5
CH Cyclohexyl S
PH Phenyl -i-
350
TABLE 12.4
Samples in aqueous solution
Samples in Aqueous Solution
* Isolation of neutral and ionic analytes from aqueous solution. Weak acids and bases by ion
suppression. Strong acids and bases using ion-pair extraction (alternative to ion exchange)
* Retention increases with solute size and is reduced by polar interactions (particularly
hydrogen-bonding) and ionization
* Polar bonded phases provide only weak retention and are not particularly useful unless
elution of the analyte is a problem from non-polar sorbents
Examples
* Isolation of agricultural and industrial chemicals from surface waters using C18, Carbon or
PS-DVB
* Isolation of polycyclic aromatic compounds from water by C18 and PS-DVB [62,63]
* Isolation of phenols from water by C18, Carbon and PS-DVB [64]
* Isolation of drugs from biofluids using C18, C8, PS-DVB, or CN [65]
* Isolation of macromolecules from biofluids and fermentation broth using C4
* Isolation of pesticides from biofluids using C18 and C8 [66,67]
· Isolation of antibiotics from biomatrices using C18 and C8 [68]
* Isolation of pigments and coloring materials from beverages and food extracts using C18
* Isolation of carbohydrates and nucleosides from biofluids using AMINO
* Isolation of proteins, peptides and surfactants using DIOL
Samples in Organic Solvents
* Retention depends on the type and number of functional groups. Solute size is less
important.
CN Strong dipole-type interactions and weak hydrogen-bond acidity
AMINO Strong hydrogen-bond base, weak hydrogen-bond acid and weak dipole-type
interactions.
DIOL Strong hydrogen-bond acid, weak hydrogen-bond base and significant
dipole-type interactions
Examples
· Isolation of polar pesticides from fats and oils
· Isolation of polycyclic aromatic compounds from fuel oils
Active ingredients from ointments and suppositories
are either similar in chemistry to porous polymers developed for HPLC and have
moderate surface areas (< 600 m2 /g), or are biporous and highly crosslinked with
surface areas from 700-1200 m2/g. The larger surface areas of the highly
crosslinked polymers results in higher retention. The highly strained porous
structure is swollen to different extents by aqueous and organic solvents and is
partially responsible for the favorable retention and elution properties of these
materials [58]. Porous polymer particle-loaded membranes are selectively
solvated by organic solvents, which are added to the sample as a processing aid
(1% v/v). The solvated polymers exhibit significant selectivity differences result-
ing in significant compound-specific changes in breakthrough volumes (Table
12.5) [72]. The retention properties of poly(styrene-divinylbenzene) polymers
can also be increased by light surface modification with polar functional groups,
such as acetyl or sulfonate [59,73]. The polar substituents reduce the interfacial
351
TABLE 12.5
Breakthrough volumes (cm 3) of some organic compounds on porous polymer particle-loaded
membrane discs with 1% (v/v) organic solvent in water
Compound Processing solvent
Acetonitrile Tetrahydrofuran 2-Propanol Methanol
Anisole 575 300 1750 1300
Acetanilide 95 25 100 75
2-Phenylethanol 150 25 100 200
Heptan-l-ol 2700 300 2700 1000
Propyl propanoate 1400 200 1250 750
tension between the polymer surface and aqueous sample increasing contact
between the analyte and polymeric sorbent. For sulfonated polymers increased
retention of neutral polar compounds was strongly dependent on the concentra-
tion of sulfonate groups with an optimum concentration of 0.6 mmol sulfonate/g
providing the largest increase in retention. The same goals are achieved by a
macroporous poly(divinylbenzene-co-N-vinylpyrrolidone) polymer (Oasis TI
HLB), which has a surface area of about 800 m2/g [74]. The sulfonated polymers
and vinylpyrrolidone copolymer have the added advantage of being fully water
wettable. Consequently, a sorbent conditioning step prior to sample application
is no longer required and samples can be fully recovered even if the sampling
device runs dry during sample processing.
The modern forms of carbon used for SPE are graphitized carbon blacks and
porous graphitic carbon [8,61,75-78]. Graphitized carbon blacks are (largely)
non-porous with moderate surface areas of 100-210 m2/g. Their surfaces are con-
taminated by oxygen complexes (hydroquinone, quinones, chromene) and benz-
pyrylium salts. Analyte retention occurs by a combination of adsorption and
anion exchange. Porous graphitic carbon is a more refined material developed
for HPLC and produced in a larger particle size for SPE. The retention mecha-
nism is different to chemically bonded phases with high retention obtained for
planar molecules containing several polar functional groups and for systems rich
in polarizable electrons [8,77,78]. Recovery of adsorbed analytes from carbon
sorbents using conventional solvents, such as methanol and acetonitrile, is often
more difficult than for chemically bonded sorbents. Stronger (less polar) binary
solvent mixtures are often more efficient and desorption by backflushing is rec-
ommended for well-retained compounds.
Low-specificity sorbents are commonly used for multiresidue analysis of
pesticides, herbicides, etc., and their decomposition products in environmental
and food samples [30,79-82]. These applications are relatively recent but
represent a natural extension of SPE methods for use in surveillance programs.
No single sorbent is considered ideal for these applications at present, sustaining
interest in the development of new sorbents with optimized retention properties
352
to meet regulatory requirements. Ion-pair extraction is a useful approach for
isolating ionized compounds by low-specificity sorbents but is not widely used at
present [83].
TABLE 12.6
General applications of ion-exchange sorbents
· In general strong ion exchanges are used to isolate weak acid/bases of opposite charge and
weak ion exchanges strong acid/bases
· Retention selectivity can be adjusted by manipulating the sample pH and ionic strength
· Choice of competing ion, its concentration and eluent pH controls selectivity for matrix
simplification and elution
· Isolation of macromolecules in an active form may require special non-denaturing sorbents
based on cellulose, agarose or dextran
Examples
· Isolation of carboxylic, sulfonic and phosphoric acids, phenols, amines and inorganic ions
from water
· Isolation of amino acids, organic acids, nucleosides and nucleotides from biofluids
· Isolation of organic acids and bases from coal-derived and synthetic fuels
Isolation of organic acids, phenols and amines from wine, fruit juices and food extracts
* Isolation of acid herbicides from water [84]
353
performed by disc extraction using a sulfonic acid-functionalized resin in the
hydrogen, silver or barium forms. The hydrogen form is used to remove hydrox-
ide and carbonate ions from samples by neutralization. The silver form is useful
for the removal of excess halide ions from samples by formation of insoluble sil-
ver halide salts that precipitate in the disk matrix. The barium form is used to
remove sulfate from samples through the formation of insoluble barium sulfate.
354
12.3.3.2 Restricted access materials
Restricted access sorbents were initially developed for the isolation of low-
molecular-mass drugs from biological fluids with minimum sample pretreat-
ment [94-100]. Recent applications include the isolation of acidic herbicides
from surface waters burdened by humic and fulvic acids [101]. Restricted access
materials work by preventing access of macromolecules (proteins) to those
regions of the sorbent where analyte retention occurs (Fig. 12.5). Restricted
access to the retentive part of the sorbent is provided by either a physical
diffusion barrier, such as a pore diameter in internal surface reversed-phase
sorbents, or by a chemical diffusion barrier, such as a polymer network at the
outer surface of the particle in semi-permeable surface materials. In addition,
the outer surface of the particles must be non-adsorptive and matrix-compatible.
Restricted access sorbents are commonly used for automated on-line sample
processing in liquid chromatography. In this case a short precolumn (0.5 to 3.0
cm) packed with the restricted access material is interfaced to a separation
column by a multiport switching valve (see Section 12.6.1). The biological fluid is
injected directly onto the precolumn, which retains the analytes of interest.
Interfering macromolecules (proteins) pass through the precolumn unretained
and are flushed to waste. The analytes retained on the precolumn are eluted
on-line to the separation column and detected. Simultaneously the precolumn is
reconditioned (or exchanged) before processing the next sample. An important
consideration for automated sample processing is the ability of the restricted
...
;.k kr---;rkir
""" - -~
workof
rophilic
r phase
Hydrophobic
inner phase
Fig. 12.5. Schematic representation of the separation mechanism of restricted access materials.
(From Ref. [101]; copyright Elsevier).
355
access material to repeatedly extract the analyte without change in properties or
accumulation of sample matrix components.
12.3.3.5 Immunosorbents
Immunosorbents (or immunoaffinity sorbents) have been used for a long time
for sample pretreatment in medicine, biology and food science but more general
applications, such as to environmental samples, are relatively recent [108-113].
In part this is due to the difficulty of making antibodies selective to small
356
molecules as well as a lack of familiarity among analytical chemists with the
procedures used to make specific antibodies. The interest in immunosorbent
extraction results from its high selectivity, which allows extraction, concentra-
tion and clean-up from complex matrices in a single step and from large sample
volumes when required. The high degree of molecular selectivity is due to the
specificity of the antibody-antigen (analyte) spatial fitting and interactions.
Cross-reactivity, considered a disadvantage in the development of single
compound schemes, is advantageous for multiresidue methods. Class-specific
immunosorbents for mycotoxins, phenylurea herbicides, and polycyclic aromatic
hydrocarbons, for example, are commercially available. Manufactured immuno-
sorbents have been available for a relatively short time and the range of products
remains narrow. Most applications continue to emanate from a few research
groups. It is clear, however, that immunosorbents have a good potential for use
with difficult sample matrices. Products based on immunochemistry have been
well received in other areas of analytical chemistry [114].
Immunosorbents are prepared by covalently bonding a suitable antibody to
an appropriate sorbent. These sorbents are used in disposable cartridges or as
short precolumns in coupled-column liquid chromatography, which has received
the most attention in research laboratories. For the latter application it is
important that the sorbent is pressure stable and the precolumn reusable for
many samples. Coupled-column SPE-GC is not as advanced [115] as SPE-LC but
has good potential for future growth.
The breakthrough volume of an immunosorbent cartridge or precolumn
depends on the total number of accessible binding sites and on the analyte-anti-
body binding constant. It is desirable to minimize non-specific analyte and
matrix interactions with the support. As a consequence there is no universal
antibody support suitable for all applications. The analyte capacity of immuno-
sorbents is generally low, but since these materials are usually used for trace
analysis and their selectivity for the target analytes is high, this is not normally a
problem. The strong analyte-antibody binding facilitates rinsing the sorbent
with solvents to eliminate matrix interference. Finding conditions that release
the analytes in a small elution volume can be problematic, particularly if the
immunosorbent is to be reused. Aqueous organic solvent mixtures are commonly
used for elution but may denature the antibodies; alternatives include the use of
competitive binding (displacing) agents, change in pH or temperature variation.
It is clear that the selection of sample processing conditions with immuno-
sorbents requires some adaptation of conventional SPE procedures, but when
fully optimized they are no more difficult to use than conventional sorbents.
357
group of similar analytes). The imprint is obtained by the polymerization of
functional and cross-linking monomers in the presence of a template molecule
(the analyte). The template-monomer system is chosen such that in solution the
imprint molecule complexes one or several of the functional monomers, which
then become spatially fixed in a solid polymer by the polymerization reaction.
The resultant imprints possess a steric (size and shape) and chemical (spatial
arrangement of complementary functional groups) memory for the template
molecule. Removal of the template from the polymer matrix creates vacant
recognition sites that enable the polymer to selectively rebind the imprint
molecule from a mixture of closely related compounds. In some cases the binding
affinity and selectivity for template-polymer binding approach those
demonstrated by antibody-antigen complexes.
For the time being, it is difficult to predetermine the experimental condi-
tions for successful imprinting of target analytes. In the case of non-covalent
binding, the most common synthetic method (other approaches include revers-
ible covalent interactions and metal ion mediated interactions), the key to a suc-
cessful imprint polymer is the formation of a stable complex between monomer
and template molecules in the pre-polymerization mixture. Maximal efficiency
of imprint formation occurs when the solvent (porogen) has sufficient solvating
power to adequately solubilize the reaction mixture without interfering with the
formation of template-monomer interactions. In current practice apolar organic
solvents for synthesis of the imprint polymer are generally used, which are usu-
ally different to solvents used for sample application. Shrinking or swelling of
the polymers in different solvents during sample processing and recovery can
have a profound effect on the affinity of the polymer for the template molecule
I117,122-124]. To date, block polymerization using UV radiation or thermal ini-
tiation to create a macroporous polymer, followed by grinding and particle siz-
ing, has been the technique most often used for the preparation of polymer
particles for cartridge SPE. The synthesis of the imprint occurs in the presence
of a large amount of the template molecule, which can be difficult to quantita-
tively extract from the polymer reducing its binding capacity, but more seri-
ously, residual template molecules may leak into sample extracts disqualifying
analyte quantification, particularly at trace concentrations. One solution to this
problem is to use a template molecule similar to the analyte for the imprint pro-
cess together with an analytical method that is able to separate the template and
analyte for the determination step. Only a few practical applications of molecu-
larly imprinted polymers for SPE have been demonstrated so far [117,119], most
of which are for the isolation of drugs from biological fluids, but there is consid-
erable promise for this technology. Methods for controlling non-specific adsorp-
tion and polymer morphology and stability are required, but seam well within
the bounds of polymer technology used to prepare sorbents for HPLC. Molecu-
larly imprinted polymers should be easier and cheaper to produce in chemical
laboratories than antibodies while, at least in theory, they should be capable of
similar specificity.
358
12.4 THEORY OF SOLID-PHASE EXTRACTION
Most of the parameters that describe the sequence of processing steps in SPE are
amenable to measurement by liquid chromatography or estimation using theo-
retical principles derived from the theory of liquid chromatography
[4,6,125-127]. Analyte concentrations are generally low and the volume of sam-
ple that can be processed, and therefore the amount of analyte isolated, is deter-
mined by the breakthrough volume of the sampling device. The volume and type
of rinse solvent for matrix simplification is determined by the type and amount
of sorbent required for the isolation step. Here, the need is to identify the volume
of the strongest solvent that eliminates matrix components from the sorbent
without analyte loss. Optimum conditions can be established by selecting eluting
conditions that preserve a certain minimum value for the retention factor of the
least retained analyte. To accomplish a significant concentration of the analytes
with minimal further sample manipulation it is desirable to recover the analytes
in a small solvent volume. Generally the minimum elution volume that can be
safely employed is about 2-3 times the hold-up volume for the sampling device.
This corresponds to a retention factor less than 2 for the most retained analyte.
In addition, the sorption mechanisms for sample application occur according to
the characteristics of frontal analysis and the rinse and desorption step by elu-
tion. Typical SPE devices contain short sorbent beds and cannot be expected to
provide a high plate number. The possibility that retention may be influenced by
kinetic properties has also to be taken into consideration.
Co
z 0.841 -
359
5% or 10% is arbitrary defined as the breakthrough volume (VB). The chosen
level reflects the difficulty of experimentally determining small changes in the
concentration of the analytes at the outlet of the sampling device. For the
purpose of modeling the breakthrough process a value of 1% is generally chosen
in keeping with the desire to define a maximum sample volume that can be
processed with minimal (acceptable) analyte loss. A second point on the break-
through curve, V, corresponds to the sample volume at which the retention
capacity of the sorbent is saturated and the concentration of analyte exiting the
sampling device is the same (actually 100% used to define the breakthrough
volume) as that entering the sampling device. It corresponds to the volume of
sample that will result in the isolation of the maximum amount of analyte but
with a lower overall recovery, because a fraction of the sample is lost during the
sorption process. The point of inflection for the breakthrough curve corresponds
to the chromatographic retention volume, VR, provided that the plate number for
the sampling device is not too small [128].
In general, there are two common causes of premature breakthrough in fron-
tal chromatography. The retention capacity of the sorbent bed is overloaded due
to a high concentration of either analyte or sorbed matrix components. Or, the
sorbent bed may fail to adequately retain the analytes due to the provision of an
insufficient plate number for retention volumes to be independent of the plate
number for the sampling device.
360
curve from which the breakthrough volume is estimated from the best-fit line
through the experimental data. A similar approach has been used to determine
the breakthrough volume of precolumn traps in on-line SPE-LC [133-135].
where VM is the interparticle volume of the sorbent bed (hold-up volume), k the
retention factor, and N the plate number for the sorbent bed calculated by Eq.
(12.3)
2
N = VR(VR - C)/v (12.3)
In principle it should be possible to calculate VB from Eqs. (12.1) to (12.3) by
determining VM and N for the sampling device and measuring VR (or k) for the
analytes of interest. N and VM are easily determined for precolumn cartridges in
coupled SPE-LC since the on-line arrangement allows direct recording of the
breakthrough curves. There is no convenient way to make measurements of
these parameters for off-line sampling conditions, and therefore estimates must
be made from results obtained for the SPE device by HPLC [127,128,142].
Equations (12.1) to (12.3) are derived assuming that the conditions of linear
chromatography apply and the plate number for the sorbent trap is reasonably
361
TABLE 12.7
Coefficients for the Lovkvist and Jonsson model, Eq. (12.4), for sorbent beds with a low plate
number
Breakthrough level (%) Coefficients
a. al a,
large. For sorbent beds with low plate numbers the above equations can result in
a poor estimate of breakthrough volumes.
Lovkist and Jonsson [128] proposed a model described by Eq. (12.4) to
characterize the sampling properties of sorbent beds with a small plate number
that has been adopted by other groups [127-131,142-144] to calculate break-
through volumes under SPE conditions. The coefficients ao, a, a2 are character-
istic of the breakthrough level and are summarized in Table 12.7 [128]
VB = (ao + a1 /N + a2 /N2)-1(1 + k)VM (12.4)
The calculation of the breakthrough volume requires the determination of N and
VM for the sampling device and either a measurement or estimation of the reten-
tion factor. The influence of these parameters with respect to the performance of
solid-phase extraction devices can now be discussed. The main requirements are
the optimization of size (VM and N), kinetic properties (N, particle size, flow rate)
and retention (N, k and VM).
For off-line or on-line sampling the sorbent bed cannot be too large because it is
desirable to recover the analytes in a small solvent volume. In addition, in
off-line sampling the pressure drop available to transport the sample through
the sampling device at a practical velocity is limited. The pressure drop per unit
length of sorbent bed is given by Eq. (12.5) [127]
where AP is the pressure drop across a sorbent bed of length L, u the linear
velocity of the sample solution through the sorbent bed, Tjthe viscosity of the
sample solution, the flow resistance parameter for the sorbent bed (typically
103 for cartridges and particle-loaded membranes [126]) and d the average
362
10
0 25 50 75 100 125
Particle Size
Fig. 12.7. Plot of (AP/L) against the average particle size (dr) for different sample processing
rates (u). A = 0.11 mm/s, B = 0.22 mm/s, C = 0.43 mm/s and D = 1.08 mm/s. The horizontal line
represents the cut off for a pressure drop of 0.9 atm for a 1 cm bed length. (From Ref. [127];
copyright Elsevier).
particle diameter for the sorbent. For off-line sampling with cartridge devices
the available pressure drop is limited to about 0.9 atm. In practice this means
that sorbent beds are restricted to particle diameters greater than about 40 Am
with larger particles providing a wider range of sample flow rates (Fig. 12.7)
[127]. For coupled-column systems high-pressure pumps are used for sample
application and pressure is no longer the factor that dictates the sorbent particle
diameter employed. These columns are also generally short resulting in
conditions favorable for fast sample processing. For off-line sampling the easiest
solution to increase sample-processing rates is to increase the diameter of the
sampling device at a constant bed height, since the flow rate is proportional to
the diameter squared. This is conveniently achieved using particle-loaded mem-
branes and particle-embedded glass fiber discs.
Sorbent cartridges with short beds packed with coarse particles, or for that
matter, very short beds packed with fine particles, cannot be expected to provide
large plate numbers. A plot of the plate height against the interparticle velocity
for some typical SPE sorbents is shown in Fig. 12.8 [126]. There is no observed
minimum in the plots over the interparticle velocity range of 0.5-5.0 mm/s
(equivalent to a flow rate of about 3-30 ml/min through a 1-cm diameter car-
tridge). The main contribution to the plate height arises from flow anisotropy in
the packed bed and resistance to mass transfer. The sorbents in Fig. 12.8 provide
from 5-40 plates per centimeter of bed length with considerable variation for the
different sorbent types. A cartridge with a 1-cm diameter operated at a sample
flow rate of 1 ml/min cannot be expected to provide more than about 20 plates
per centimeter of bed length. At higher sample flow rates only lower plate num-
bers are expected. For cartridges with a lower packing density than shown in
Fig. 12.8, a realistic scenario based on measurements of average cartridge pack-
ing density [126] even smaller plate numbers are to be expected.
363
11n _ 1n
2 A
:ZF
;~100 I
aH 9
8S~~~~~W
a
9~~~m
S~~~~~~~~
1.5-
(D
Li
0
U80 7n
LiI
H W
H
I
1- l54f
H f 41
Alt -
4
0 0.5 i 1.5
INTERPARTICLE VELOCITY
0 0 50 .
100
1 2 3 4 5 o
INTERPARTICLE VELOCITY uJbr rLk. 1MH
Left: Fig. 12.8. Plot of the plate height (mm) against mobile phase interparticle velocity (mm/s)
for silica-based cartridge sorbents. The test compound was anthracene and the mobile phase
methanol-water (80:20 v/v). 1 = Octadecylsiloxane-bonded sorbent (light loading); 2 = spacer
bonded propanediol sorbent; 3 = cyanopropylsiloxane-bonded sorbent; 4 = butylsiloxane-
bonded sorbent. (From Ref. [126]; copyright Elsevier).
Fig. 12.9. Variation of the efficiency of a particle-loaded membrane as a function of the mobile
phase velocity. (A) Plot of the observed plate height (am) as a function of the interparticle mobile
phase velocity, and (B) the transformation of(A) to indicate the plate number (N) as a function of
the sample flow-rate through a 0.5 mm disk with a 38 mm-diameter active sampling area (From
Ref. [126]; copyright Elsevier).
364
0.7
.u
O0.
C
0.2
0.2
0
Sample Volume
Fig. 12.10. Plot of the breakthrough curves for a sampling device with a hold-up volume of 0.42
ml, retention factor 100 and (A) 5, (B) 20, and (C) 100 plates per cm of bed length. (From Ref.
[127]; copyright Elsevier).
with relatively large values of N provide sharper desorption front profiles and
require a smaller elution volume to quantitatively recover the analyte from the
cartridge.
12.4.3 Retention
365
12.4.3.1 Experimental determinationof retention factors
There are two general approaches for determining retention factors on sorbents
used for SPE. The equilibrium method uses a fixed volume of sample solution
with a known concentration of analyte that is continuously circulated by a
mechanical pump through the sampling device and returned to the sample
solution reservoir until a steady state is reached [36,140,141]. On-line monitor-
ing of the analyte solution exiting the sampling device is a convenient method to
establish the time required to reach a steady state. The amount of analyte
retained by the sorbent is then determined by elution with a small volume of
strong solvent using any suitable method for quantification. Provided that the
sorption isotherm is linear the retention factor can be calculated from the
volume and concentration of the analyte in the sample solution, the hold-up
volume for the sorbent bed, and the amount of analyte taken up by the sorbent
under steady-state conditions.
The alternative method of determining the retention factor is by direct
measurement of retention in a typical chromatographic experiment using col-
umns of different lengths packed with the SPE sorbent [139,142,146-150] or
forced-flow planar chromatography [28,145,151] for discs. Compounds with
retention factors up to about ten thousand can be determined in this way.
Compounds with a retention factor greater than ten thousand are more than
adequately retained for most likely sampling conditions and an accurate deter-
mination of their retention factors is rarely required.
366
TABLE 12.8
Comparison of experimental and extrapolated values for log kw for an octadecylsiloxane-bonded
silica sorbent
Compound Log kw
Linear Eq. (12.8)' Quadratic Eq. (12.9)2 Experimental
Methanol-Water
2-Phenylethanol 2.00 2.36 2.45
4-Chlorophenol 2.52 2.50 2.73
4-Nitrobenzyl alcohol 1.58 2.15 2.40
Acetanilide 1.60 2.19 2.52
Acetophenone 2.26 2.81 3.01
Benzaldehyde 1.87 2.43 2.56
Hexan-2-one 1.91 2.49 2.62
Acetonitrile-Water
2-Phenylethanol 1.27 2.05 2.45
4-Chlorophenol 1.81 2.35 2.73
4-Nitrobenzyl alcohol 1.21 1.89 2.40
Acetanilide 1.03 1.92 2.52
Acetophenone 1.71 2.33 3.01
Benzaldehyde 1.66 1.94 2.56
Hexan-2-one 1.74 2.31 2.62
1From 40-70% (v/v) methanol or acetonitrile.
2
For the full data range from 1-100 %(v/v) methanol or acetonitrile.
367
the solute's effective hydrogen-bond acidity and hydrogen-bond basicity, S ah
and A S, respectively. The solute's characteristic volume and excess molar
refraction are additive properties that can be calculated for any analyte whose
structure is known [154,155]. The solute dipolarity/polarizability and hydro-
gen-bond acidity and basicity parameters can be obtained experimentally from
gas-liquid chromatographic data or water-solvent distribution constants [156].
For some solutes such as anilines, pyridines, and sulfoxides, which exhibit vari-
able hydrogen-bond basicity in aqueous and non-aqueous solutions, the CH
descriptor is replaced by I O for the extraction of samples from aqueous solu-
tion. The P descriptor is retained for extraction of analytes from organic sol-
vents. Solute descriptors are available for about 4000 compounds with others
available through parameter estimates for compounds of known structure [157].
The system constants in Eq. (12.10) are defined by their complementary
interactions with the solute descriptors. The r constant determines the differ-
ence in capacity of the solvated sorbent and sample solution to interact with
solute n- or -electrons; the s constant to the difference in capacity of the
solvated sorbent and sample solution to take part in dipole-dipole and dipole-
induced dipole interactions; the a constant is a measure of the difference in
hydrogen-bond basicity of the solvated sorbent and the sample solution; the b
constant is a measure of the difference in hydrogen-bond acidity of the solvated
sorbent and sample solution; and the m constant is a measure of the relative ease
of cavity formation for the solute in the solvated sorbent and sample solution
together with contributions from dispersion interactions that fail to cancel when
the solute is transferred between phases. For any sampling system, the system
constants can be obtained by multiple linear regression analysis of experimental
retention properties acquired for a group of varied solutes with known
descriptors. The salvation parameter model is strictly applicable to neutral
compounds and ionizable compounds in their neutral form. Ionization tends to
reduce retention compared to the neutral form of the solute in SPE from
aqueous solution.
A plot of the system constants as a function of solvent composition provides a
system map such as the one shown in Fig. 12.11 [127]. System maps are the most
6
4 m
o r C
= '
b
Fig. 12.11. System map for a heavily loaded octadecyl- -4
siloxane-bonded silica sorbent with methanol-water as the 0 25 50 75 100
sample solvent. (From Ref. [127]; copyright Elsevier). Volume %of Methanol
368
TABLE 12.9
System maps for selecting SPE conditions for aqueous samples using the solvation parameter
model
Sorbent* Type Solvent Ref.
IST C-18 (HL) Octadecylsiloxane-bonded silica (high Methanol 142
carbon loading)
JTB C-18 (LL) Octadecylsiloxane-bonded silica (light Methanol 132,136
carbon loading) Acetonitrile
Tetrahydrofuran
Empore C-18 Octadecylsiloxane-bonded silica Methanol 151
particle-loaded membrane
SPEC-C18AR Octadecylsiloxane-bonded silica Methanol 28
particle-embedded membrane
JTB C-4 Butylsiloxane-bonded silica Methanol 150
2-Propanol
Acetonitrile
Tetrahydrofuran
JTB CN Cyanopropylsiloxane-bonded silica Methanol 147,149
2-Propanol
Acetonitrile
Tetrahydrofuran
JTB DIOL Silica-based spacer bonded propanediol Methanol 148
2-Propanol
Acetonitrile
Tetrahydrofuran
PLRP-S Poly(styrene-divinylbenzene) Methanol 146
2-Propanol
Acetonitrile
PGC Porous graphitic carbon Methanol 77
*IST = International Sorbent Technology; JTB = J.T. Baker; PLRP-S = Polymer Laboratories;
PGC = Hypercarb; Empore = 3M; and SPEC = Ansys.
369
TABLE 12.10
System constants for the estimation of solute retention in solid-phase extraction for aqueous
samples (water or 1% v/v methanol in water)
Sorbent* System constants
m r s a b c
IST C-18 (HL) 4.39 0 0 -0.79 -1.90 -0.27
JTB C-18 (LL) 3.92 0 -0.11 -0.54 -1.53 -0.90
IST C-8 4.91 0 -0.84 -1.43 -2.27 -0.54
IST CHX 3.94 0 -0.32 -0.83 -1.98 -0.31
IST CH 3.22 0.35 0 -0.92 -1.61 -0.37
JTB C-4 3.36 0 0 -0.46 -1.53 -1.38
JTB CN 2.06 0.53 0 -0.51 -1.45 -0.88
JTB DIOL 1.57 0.61 0 -0.45 -0.80 -1.05
PLRP-S 5.22 0.84 -0.49 -1.39 -4.01 -0.18
Oasis HLB 3.48 1.16 0 0 -2.94 -0.32
PGC 5.14 0.82 0.36 0 -2.76 -2.32
*IST = International Sorbent Technology; JTB = J. T. Baker; PLRP-S = Polymer Laboratories
(styrene-divinylbenzene porous polymer); Oasis HLB = poly(divinylbenzene-co-N-vinylpyrro-
lidone); PGC = Hypersil porous graphitic carbon (Hypercarb); HL = high loading; LL = light
loading; CHX = cyclohexanesiloxane-bonded; CH = phenylsiloxane-bonded; CN = cyanopropyl-
siloxane-bonded; and DIOL = spacer bonded propanediol.
370
TABLE 12.11
System constants for retention in liquid-solid chromatography with organic mobile phases
Stationary Mobile phase System constants
phase*
m r s a b
DIOL Hexane -1.05 0 1.63 2.10 3.86
AMINO -0.85 0 1.40 1.65 3.81
CYANO -0.37 0 1.88 2.47 0.99
Silica Hexane-methanol -0.83 0 1.06 2.23 1.56
DIOL (99:1) -0.85 0 1.07 2.37 1.47
AMINO -0.72 0 0.94 2.94 1.20
CYANO -0.61 0 0.95 1.86 1.15
Silica Hexane-methyl 0 -0.86 1.67 1.84 3.00
CYANO t-butyl ether (95:5) -1.20 0 1.43 1.10 2.80
Silica Hexane-methyl -0.46 -0.21 1.10 1.16 3.02
CYANO t-butyl ether (80:20) -1.08 0 1.01 0.53 2.26
*AMINO = 3-aminopropylsiloxane-bonded silica; CYANO = 3-cyanopropylsiloxane-bonded
silica; and DIOL = spacer bonded propanediol siloxane-bonded silica.
371
phase conditions is the capacity of the solvated sorbent for polar interactions.
Increasing solute size generally reduces retention, in contrast to reversed-phase
extraction where the m system constant is invariably positive and the most
important term controlling solute extraction. Electron lone pair interactions are
unimportant for many applications. The main contribution to retention by the
3-aminopropylsiloxane-bonded sorbent is the hydrogen-bond basicity of the
solvated sorbent (a constant). For the 3-cyanopropylsiloxane-bonded sorbent it
is the capacity of the solvated sorbent for dipole-type interactions (s constant).
For the spacer bonded propanediol siloxane-bonded sorbent it is the capacity for
hydrogen-bond interactions, particularly as a hydrogen-bond acid (b constant),
combined with a significant capacity for dipole-type interactions (s constant)
that are important. Silica is a strong hydrogen-bond acid and significantly
hydrogen-bond base and dipolar. The range of sorbent selectivity variation is
much greater than observed for extraction of aqueous solutions and is very
strongly dependent on the composition of the sample solvent. This is seen in the
large changes in the system constants observed for hexane and hexane contain-
ing 1% (v/v) methanol. For this reason it is unwise to attempt to reduce retention
properties to a sorbent property alone. The limited data available indicates that
it will be difficult to obtain reasonable breakthrough volumes for analytes that
lack polar functional groups and that breakthrough volumes will change signifi-
cantly with small changes in the composition of the sample solvent. Although the
solvation parameter model can model retention on chemically bonded sorbents
using organic solvents without apparent difficulty it probably does not correctly
account for solute size effects and site-specific interactions for inorganic oxide
sorbents [158].
372
Fig. 12.12. Method selection guide for the isolation of organic compounds from solution. SAX =
strong anion exchanger; SCX = strong cation exchanger; WCX = weak cation exchanger; RP =
reversed-phase sampling conditions; NP = normal-phase sampling conditions; IE = ion-
exchange sampling conditions. (From Ref. [127]; copyright Elsevier).
adequate sample volume for analysis. The sample volume that can be processed
by a particular SPE device depends primarily on the breakthrough volume of the
analyte, the concentration of the analyte matrix, sample flow rate, and the
sorbent mass. SPE cartridges are available in a range of sizes containing from
about 35 mg to 10 g of sorbent with the 100 mg and 500 mg sorbent cartridges
being the most widely used for extraction and the larger cartridge sizes for sam-
ple clean-up. As a rough guide the sample capacity of a SPE cartridge is about
1-5% of the sorbent mass. The parameters of interest for selecting a disc for a
particular application are summarized in Table 12.13. The large diameter discs
are used for processing large sample volumes in an acceptable time. Small discs
are used for processing small sample volumes and to recover analytes in a small
solvent volume to eliminate the need for solvent evaporation prior to analysis.
Small discs are frequently used in clinical, forensic and pharmaceutical analysis.
Sample processing involves four distinct steps. Initially, the sorbent is condi-
tioned with solvent to remove impurities and to facilitate sorption of the
analytes. The conditioning step is critically important for processing aqueous
samples using particle-loaded membranes. The high surface tension of water
combined with the microporosity of the discs results in slow and uneven flow
through the discs and low analyte recovery if the discs are not first conditioned
with an organic solvent. The conditioning solvent is then replaced with the same
solvent as the sample solvent and the sample passed through the sampling
device at a controlled flow rate. For large sample volumes a small amount of
organic solvent (1% v/v) is added to aqueous samples to maintain a constant sam-
pling velocity. This usually has little influence on the breakthrough volumes
373
TABLE 12.12
Sorbent selection for disc applications
Sorbent General applications
Octadecylsiloxane- and Largely used for extracting aqueous solutions. Applications
octylsiloxane-bonded dominated by octadecylsiloxane-bonded sorbents. Many
silica applications to the extraction of non-polar and moderately polar
pesticides (all classes), herbicides, phthalate and adipate esters,
polycyclic aromatic compounds, food additives, pharmaceutical
compounds, hydrocarbons and grease, etc. A large number of
approved regulatory methods for water analysis.
Poly(styrene-divinyl- Used for compounds poorly extracted by octadecylsiloxane-bonded
benzene) silica sorbents because of high water solubility (polar pesticides,
herbicides, phenols and pharmaceutical compounds). Higher
recoveries of neutral polar compounds claimed for polar group
functionalized sorbents because of better surface contact with
aqueous samples.
Activated carbon Applications similar to poly(styrene-divinylbenzene) but less
frequently used. Methods proposed for triazine herbicides,
some polar pesticides and N-nitrosodialkylamines.
Mixed mode Usually co-bonded alkylsiloxane and benzenesulfonic acid groups
(or sorbent mixtures) used predominantly in toxicological and
pharmaceutical analysis for the simultaneous isolation of drugs
and their metabolites. A number of established methods for drugs
of abuse (marijuana and cocaine metabolites, amphetamines,
phencyclidine and opiates) in biological fluids.
Sulphonic acid and Isolation of acidic and basic compounds as well as some metals.
quaternary amine ion Many methods for acidic herbicides and pesticides in water and
exchangers basic drugs in biological fluids. Hydrogen-ion and metal-loaded
cation exchangers available for clean-up of samples analyzed by ion
chromatography and capillary electrophoresis.
Crown ethers Commonly used for the isolation of precious metals (Pd, Pt, Rh)
and radionuclides (Cs, Sr, Pb) from high concentrations of other
metals to determine environmental burden and for dating
geochemistry samples. Other applications include the removal of
base metal impurities (Bi, Sb, Fe, Pb, Bi, Cu, Hg) from refinery
streams, plating baths, etc.
TABLE 12.13
Guide to the physical properties of particle-loaded extraction discs
Property Disc diameter (mm)
4 7 10 25 47 90
2
Surface area (cm ) 0.13 0.38 0.80 4.9 17 64
Bed mass (mg) (silica-based sorbents) 4 10 25 140 500 1850
Flow rate (ml/min) 0.5 1.5 3 20 60 250
Elution volume (ml) 0.15 0.25 0.5 3 10 35
Typical sample volume (ml) <1 <5 <25 <250 <1000 <5000
374
except for some porous polymer sorbents, which selectively absorb the organic
solvent changing sorbent selectivity and the system phase ratio [72]. Porous
polymer sorbents, such as poly(divinylbenzene-co-N-vinylpyrrolidone), that are
solvated by water alone, allow samples to be processed without a conditioning
step. A sample processing solvent is not usually required for small sample vol-
umes.
Optionally, after the sample is processed, the sorbent can be rinsed with a
weak solvent to displace undesired matrix components from the sorbent without
displacing the analytes. Finally the analytes of interest are eluted from the
sorbent in a small volume of strong solvent for subsequent determination. The
drying step between processing aqueous samples and eluting the retained
analytes with a water-miscible organic solvent is also important. Water retained
by the sorbent (and the support structure when discs are used) contaminates the
elution solvent and may cause difficulty if the solvent is to be reduced in volume
or the analytes analyzed by gas chromatography. Water may also reduce the
efficiency of the elution step, particularly for solvents only partially miscible
with water. Drying, by suction or vacuum desiccation, is usually sufficient to
remove water trapped in the pores, but excessive drying can result in low analyte
recovery from vaporization or inefficient elution. General advice for establishing
optimum sample processing conditions for cartridge and disc sampling devices is
summarized in Table 12.14. A generic outline for processing samples using disc
technology is outlined in Table 12.15. This can be re-scaled for different disc and
sample sizes using the data in Table 12.13.
375
TABLE 12.14
Sample processing parameters and their influence on recovery by solid-phase extraction
Conditioning solvent (typically 3-5 bed volumes)
· Ensures reproducible retention and flow. Critical step for particle-loaded membranes
· Helps to minimize contamination of extracts by sorbent impurities
· Replace by sample solvent before processing sample
Flow rates (typical range 0.2-1.5 mm/s)
* More critical for cartridges than discs due to their variable and heterogeneous packing
density (channeling)
* More critical when the sample volume exceeds the breakthrough volume as typical
sampling devices provide too few theoretical plates for flow independent retention
Sample properties
· Dilute viscous samples with a weak low viscosity solvent to reduce sample processing time
* Remove excessive particle matter by filtration or centrifugation to maintain a constant
sample-processing rate.
* Add small volume of organic solvent (1-3% v/v) to large volume water samples to ensure
sorbent remains solvated and to maintain a constant (fast) sample-processing rate.
Important for particle-loaded membranes
* Adjust pH to reduce ionization of weak acids and bases for reversed-phase sampling
* Maintain approximately constant ionic strength for samples and standards when using
reversed-phase sampling conditions. Ionic strength is a critical parameter for ion-exchange
extraction
* Deproteination of biofluids may be required for acceptable recovery of low molecular mass
analytes for reversed-phase sampling
* Precipitation of inorganic acids (sulphate, phosphate, etc.) by barium hydroxide is
sometimes required for acceptable recovery of organic acids from biofluids using
ion-exchange extraction
Drying time (typically 1-5 min, but sometimes considerably longer)
· Sufficient to remove all sample solvent trapped in the sorbent pores
* Excessive drying may result in low recovery of analytes from evaporation or retention in
poorly solvated regions of the sorbent
Rinse solvent (optional)
* Small volume of intermediate strength solvent to elute matrix components. Analytes
remain immobilized on the sorbent
* Biological fluids, plant extracts and soil extracts often require a rinse step but surface
waters may not
Eluting solvent (ideally 2-3 bed volumes but often larger)
· Should be a strong solvent able to displace all analyte from the sorbent in a small volume
· Normally should be volatile and miscible with the sample solvent
376
TABLE 12.15
Generic guide for sample processing using solid-phase extraction discs. In this example a 47-mm
disc and a 1-1 water sample are used for illustration
Decontamination With disc installed in filtration apparatus (or cartridge) rinse with 10 ml
of solvent (acetonitrile) by allowing the solvent to soak into the
membrane for a few minutes and then remove by suction.
Condition As described above using 10 ml of methanol. Before the last drop of
methanol has been sucked through the disc add 10 ml of deionized water.
The disc should not be allowed to suck dry until after the sample has been
processed.
Sample The sample containing 1% (v/v) methanol is passed through the disc with
a suitable reservoir in place. Filter aid or a prefilter may be required for
samples with a heavy burden of particulate matter. The sample is
processed at a flow rate between about 50 and 100 ml/min using a vacuum
of about 10-20 mm Hg.
Drying Bulk water is removed from the disc and sampling apparatus by sucking
air through the disc under full vacuum for about 3 minutes. Volatile
compounds may be lost.
Recovery The analytes are recovered by passing two 5 ml volumes of acetonitrile
through the disc. The first volume of acetonitrile is allowed to soak into
the disc for a few minutes and then gently sucked through the disc.
Without letting the disc run dry the second volume of solvent is added to
the disc and sucked through the disc. The disc is then allowed to suck dry.
SPE lends itself to automation to an extent that has been difficult to achieve
using LLE procedures [1,4,166]. A major impetus for the development of auto-
mated laboratory procedures has come from increasing demand for high
throughput methods in clinical, pharmaceutical and environmental laborato-
ries. Automation is viewed as an essential feature of achieving this goal by reduc-
ing time and cost of labor while improving method accuracy and precision,
reducing assay tedium and enhancing safety by limiting exposure to chemicals
and pathogens. In the most favorable case automated procedures provide for
out-of-hours operation and require a minimum of operator intervention. This is
not always achieved, however.
Automated methods can be considered in three categories. The most com-
mon are semi-automated instruments where operator intervention is required
at some stages during the extraction. These instruments are designed to reduce
the tedium of repetitive manual operations. They may increase sample through-
put and improve accuracy and precision, but are really method assistants.
377
Workstations, on the other hand, are dedicated instruments that usually carry
out all stages of the SPE method without intervention, and are usually easily
coupled to separation instruments. Most early versions, however, operated in a
serial sampling mode or semi-batch mode and offered limited improvement in
sample throughput. The best of these systems can extract between 25-50
samples per hour, but the majority are nearly an order of magnitude slower than
this. For coupled-column systems this was not a problem when the separation
was slow, but with increasing use of fast separation methods and mass spectro-
metric detection, which is less demanding of extraction selectivity, this was no
longer adequate. Really high sample throughput was achieved by parallel sample
processing using workstations designed around the 96-well plate format. The
best of these systems can extract up to about 400 samples per hour, but are
usually somewhat limited in the sample volume that can be handled, and are
mainly used in clinical and pharmaceutical laboratories, where large sample
volumes are rarely required for quantification of target analytes. A higher level
of automation can be achieved using robotic systems that are capable of many
activities besides SPE (e.g. weighing, centrifugation, filtration, capping, evapo-
ration, bar code reading, etc). These can often process the sample from an earlier
stage in the analysis through to injection of extracts into the separation system.
Automation is probably not a high priority for all laboratories. In the absence
of large sample numbers requiring a short turn around time fully automated
systems are difficult to justify. The best of the modern systems use convenient
graphical interfaces that are easy to use but earlier and some existing systems
require specialized knowledge of computer programming and electronics for
operation and maintenance. Carryover from the sharing of common modules by
the samples can limit performance and reduce accuracy through contamination
and limit throughput by the need for repetitive programmed rinse procedures.
Systematic errors arise when the instrument fails to take account of sample
properties, such as incomplete withdrawal of sample volumes due to changes in
viscosity, precipitation or clotting. Sample stability can be a problem for some
samples and conditions, for example, when samples are delayed for various times
prior to being processed. Differences in liquid management, for example, the use
of positive pressure versus vacuum, often means that manually developed
procedures do not function well on automated systems without extensive re-
optimization. On the other hand, automated workstations provide a more
controlled environment for method development with the possibility of running
larger sets of trial experiments in a shorter time to identify rugged extraction
conditions.
378
feasible but largely restricted to use in research laboratories for the present
[169-171]. The majority of applications are for aqueous samples reflecting the
main use of the technique in clinical, pharmaceutical and environmental labora-
tories for the analysis of biological matrices and surface waters. Typical sample
volumes for SPE-LC are 1-10 ml while larger volumes, 10-1000 ml, are required
for environmental applications. For SPE-GC sample sizes are often an order of
magnitude smaller because of the greater sensitivity provided by typical gas
chromatographic detectors. The design of precolumn sampling cartridges is
similar for SPE-LC and SPE-GC as are the sorbents used. Low-specificity
sorbents are used for general applications, mainly octadecylsiloxane-bonded
silica gel and porous polymers [167,169,172], and class-specific sorbents, partic-
ularly restricted access materials and immunosorbents, for selective extraction
of target analytes [96-101,173-177]. Restricted access materials have gained
acceptance for the direct analysis of drugs in biological fluids and the analysis of
polar pesticides in the presence of humic acid in surface waters. Breakthrough
volume considerations are generally unimportant for sampling biological fluids,
since only small sample volumes are required for adequate detection of target
analytes but for environmental applications this is a critical parameter and
dominates sorbent selection. High surface area porous polymers and to a lesser
extent graphitized carbon sorbents, are generally used for environmental appli-
cations [6,43,178-181], but no single sorbent is suitable for all applications.
The choice of precolumn dimensions and sorbent properties is a compromise
between the need for sufficient retention to accommodate the desired sample
volume and efficient analyte desorption for transfer to the separation column
without unacceptable loss in its separation power. Typical precolumns are 10-30
mm long with internal diameters of 1-3 mm when used with standard 4-4.6 mm
internal diameter separation columns in liquid chromatography or 2-20 mm
long and 1-4.6 mm internal diameter for coupling to gas chromatography. These
columns contain about 10-50 mg of sorbent with an average particle diameter in
the range of 10-40 gm. Less frequently precolumns are prepared from several
small particle-loaded membrane discs assembled to form a continuous bed. For
automated operation the precolumn must be either reusable or automatically
exchangeable.
The on-line coupling of precolumn and separation column in liquid chroma-
tography is achieved using a single 6-port automated switching valve with flow
paths indicated in Fig. 12.13, or by using two automated valves for additional
versatility in automating a wider range of sample processing steps. The control
system has to provide for regeneration and conditioning of the precolumn,
loading of the sample, rinsing the precolumn to minimize matrix constituents
and transfer of the analytes to the analytical column for separation. Usually the
sample processing steps occur concurrently with the separation of the previous
sample. All events are programmable and optimized by time and flow control.
For strongly retained analytes, backflushing the precolumn is often the best
approach for solvent desorption. The solvent used for desorption is usually the
379
(I) (I)
same or a weaker solvent than the mobile phase used for the start of the
separation. The analytes are also distributed over a significant length of the
precolumn and their transfer usually occurs in a solvent volume that is too large
and perhaps too strong to avoid additional band broadening compared with
conventional injection techniques. The precolumn dimensions (short column of
smaller internal diameter than the separation column), sorbent particle size
(small particle diameter) and the retention capacity of the sorbent for the
analytes (weak retention capacity) are optimized to assist in minimizing band
broadening due to the transfer process. On the other hand for successful
sampling the sorbent should have strong retention properties, the precolumn
should be as large as practical and packed with coarse particles to assist in
sample application when large volumes are used. For environmental samples,
sample volumes usually exceed 100 ml and sample application rates of 5-15
ml/min are common. The flow resistance of the precolumn is an important
consideration for these sampling conditions. The criteria for extraction and
desorption are generally opposite and practical operating conditions are always a
compromise.
In most cases preconcentration of analytes occurs with simultaneous matrix
accumulation. Matrix simplification can be achieved through the sampling pro-
cess, the use of a rinse solvent prior to analyte transfer, and selective transfer of
front-, heart- or end-cuts of the desorbed sample extract. Dual precolumn sys-
tems allow solvent exchange and separate trapping of analytes with a wide range
of breakthrough volumes. The first column of a dual precolumn system can be
used to retain strongly sorbed matrix components that would contaminate the
second precolumn or interfere in the separation of the transferred extract.
380
Fig. 12.14. Typical valve switching arrangement for automated SPE-GC. (From Ref. [168];
copyright Elsevier).
381
a solvent film is already present in the retention gap when extract transfer
starts, can be used to minimize this loss.
Loop-type interfaces with partial concurrent solvent evaporation in a long
retention gap are more robust and easier to automate than the on-column
interface but are restricted to the transfer of analytes of low volatility [20,
185-187]. The interface is equipped with a 6-port and a 14-port switching valve.
The 6-port valve is used to divert the carrier gas either to the gas chromatograph
or to the 14-port valve, on which the SPE cartridge and two loops for storing
organic solvent are mounted. The first loop contains the presolvent and is
usually of smaller volume (e.g. 50 Al). The second loop contains the solvent used
to desorb the analytes from the SPE cartridge and to transfer them to the gas
chromatograph. After the sample processing steps are completed and the
cartridge dried, both valves are switched simultaneously and the solvent vapor
exit opened. The carrier gas pushes the contents of both loops, the second loop
via the SPE cartridge, into the retention gap. Both valves are switched to the
load position when a preselected pressure drop corresponding to near comple-
tion of the evaporation process is recorded and the solvent vapor exit closed after
a selected delay time. The transfer is carried out at a temperature just above the
solvent boiling point.
12.7 CONCLUSIONS
REFERENCES
1 E.M. Thurman and M.S. Mills, Solid-Phase Extraction. Principles and Practice.
Wiley, New York, 1998.
2 J.R. Dean, ExtractionMethods for EnvironmentalAnalysis. Wiley, Chichester, 1998.
382
3 J.S. Fritz, Analytical Solid-Phase Extraction. Wiley, New York, 1999.
4 N.J.K. Simpson (Ed.), Solid-Phase Extraction: Principles, Strategies and
Applications. Marcel Dekker, New York, 2000.
5 I.D. Wilson, E.R. Adlard, M. Cooke and C.F. Poole, Encyclopedia of Separation
Science. Academic Press, London, Vol. 1-10, 2000.
6 M.-C. Hennion, J. Chromatogr.A, 856 (1999) 3.
7 I. Liska, J. Chromatogr. A, 885 (2000) 3.
8 E. Matisova and S. Skrabakova, J. Chromatogr. A, 707 (1995) 145.
9 G.A. Junk, J.J. Richard, M.D. Grieser, D. Witiak, J.L. Witiak, M.D. Angello, R. Wick,
H.J. Svec, J.S. Fritz and G.V. Calder, J. Chromatogr., 99 (1974) 745.
10 M. Dressier, J. Chromatogr., 165 (1979) 167.
11 A. Zlatkis, H.A. Lichtenstein and A. Tishbee, Chromatographia,6 (1973) 67.
12 M. Harper, J. Chromatogr. A, 885 (2000) 129.
13 J.S. Fritz and M. Macka, J. Chromatogr.A, 902 (2000) 137.
14 J.W. Eichelberger, T.A. Bellar, J.P. Donnelly and W.L. Budde, J. Chromatogr. Sci.,
28 (1990) 460.
15 B. Kolb, J. Chromatogr. A, 842 (1999) 163.
16 M.-C. Hennion and V.C. Pinchon, Environ. Sci. Technol., 28 (1994) 576A.
17 L.A. Berrueta, B. Gallo and F. Vicente, Chromatographia,40 (1995) 474.
18 M. Gilar, E.S.P. Bouvier and B.J. Compton, J. Chromatogr. A, 909 (2001) 111.
19 N. Masque, R.M. Marce and F. Borrull, Trends Anal. Chem., 17 (1998) 384.
20 J.J. Vreuls, A.J.H. Louter and U.A.Th. Brinkman, J. Chromatogr.A, 856 (1999) 279.
21 E.R. Brouwer, S. Kofman, U.A.Th. Brinkman, J. Chromatogr.A, 703 (1995) 167.
22 A.J.H. Louter, J.J. Vreuls and U.A.Th. Brinkman, J. Chromatogr.A, 842 (1999) 391.
23 E.C. Goosens, D. de Jong, G.J. de Jong and U.A.Th. Brinkman, Chromatographia,47
(1998) 313.
24 D.R. Hagen, C.G. Markell, G. Schmitt and D.D. Blevins, Anal. Chim. Acta, 236 (1990)
157.
25 H. Lingeman and S.J.F. Hoekstra-Oussoren, J. Chromatogr.B, 689 (1997) 221.
26 E.M. Thurman and K. Snavely, Trends Anal. Chem., 19 (2000) 18.
27 J.S. Fritz and J. Masso, J. Chromatogr. A, 909 (2001) 79.
28 M.L. Meyer, C.F. Poole and M.P. Henry, J. Chromatogr.A, 695 (1995) 267.
29 I. Urbe and J. Ruana, J. Chromatogr. A, 778 (1997) 337.
30 V. Pinchon, M. Charpak and M.-C. Hennion, J. Chromatogr.A, 795 (1998) 83.
31 R.S. Plumb, R.D.M. Gray and C.M. Jones, J. Chromatogr., B 694 (1997) 123.
32 J. Hempenius, J. Wieling, J.P.G. Brakenhoff, F.A. Maris and J.H.G. Jonkman, J.
Chromatogr. B, 714 (1998) 361.
33 U.J. Nilsson, J. Chromatogr.A, 885 (2000) 305.
34 M.S. Lee and E.H. Kerns, Mass Spectrom. Rev., 18 (1999) 187.
35 M. Jemal, Biomed. Chromatogr., 14 (2000) 422.
36 M.C. Rouan, C. Buffet, L. Masson, F. Marfil, H. Humbert and G. Maurer, J.
Chromatogr., B 754 (2001) 45.
37 C.E. Green and M.H. Abraham, J. Chromatogr.A, 885 (2000) 41.
38 W.C. Koskinen and B.L. Barber, J. Environ. Qual., 26 (1997) 558.
39 J.D. Mattice, S.K. Park and T.L. Lavy, Bull. Environ. Contam. Toxicol., 60 (1998)
202.
40 E.M.J. Verbruggen, W.M.G.M. Van Loon, M. Tonkes, P. Van Duijn, W. Seinen and
J.L.M. Hermens, Environ. Sci. Technol., 33 (1999) 801.
41 A.P. Freidig, E.A. Garicano, F.J.M. Busser and J.L.M. Hermens, Environ. Toxicol.
Chem., 17 (1998) 998.
42 J.A. Field and K. Monohan, J. Chromatogr. A, 741 (1996) 85.
43 L.E. Sojo and J. Djauhari, J. Chromatogr.A, 840 (1999) 21.
383
44 L.K. Ng, P. Lafontaine and J. Harnois, J. Chromatogr.A, 873 (2000) 29.
45 J.W. King and L.J. King, J. Anal. Toxicol., 20 (1996) 262.
46 J.M. Rosenfeld, J. Chromatogr.A, 843 (1999) 19.
47 N.A. Marley, J.S. Gaffney, K.A. Orlandini, P.J. Drayton and M.M. Cunningham,
Radiochim. Acta, 85 (1999) 71.
48 E.D. Hagestuen, A.F. Arruda and A.D. Campiglia, Talanta, 52 (2000) 727.
49 R.L. Vermillion-Salsbury and D.M. Hercules, Int. J. Envuiron. Anal. Chem., 73 (1999)
297.
50 A.W.T. Bristow, C.S. Creaser, S. Nelieu and J. Einhorn, Analyst, 121 (1996) 1425.
51 D.M. Steinberg, L.J. Sokoll, K.C. Bowles, J.H. Nichols, R. Roberts, S.K. Schultheis
and C.M. O'Donnell, Clin. Chem., 43 (1997) 2099.
52 P.D. Whitter and P.L. Cary, J. Anal. Toxicol., 10 (1986) 68.
53 S.A. Barker, J. Chromatogr.A, 885 (2000) 115.
54 V. Ruiz-Gutierrez and M.C. Perez-Camino, J. Chromatogr.A, 885 (2000) 321.
55 T.M.P. Chichila and D.M. Gilvydis, J. AOAC Int., 76 (1993) 1323.
56 Y. Pico, G. Font, J.C. Molto and J. Manes, J. Chromatogr.A, 885 (2000) 251.
57 G.R. van der Hoff and P. van Zoonen, J. Chromatogr.A, 843 (1999) 301.
58 C.W. Huck and G.K. Bonn, J. Chromatogr.A, 885 (2000) 51.
59 M.E. Leon-Gonzalez and L.V. Perez-Arribas, J. Chromatogr.A, 902 (2000) 3.
60 N.A. Penner, P.N. Nesterenko, M.M. Ilyin, M.P. Tsyurupa and V.A. Davankov,
Chromatographia,50 (1999) 611.
61 M.-C. Hennion, J. Chromatogr. A, 885 (2000) 73.
62 R.M. Marce and F. Borrull, J. Chromatogr.A, 885 (2000) 273.
63 G. Michor, J. Carron, S. Bruce and D.A. Cancilla, J. Chromatogr. A, 732 (1996) 85.
64 I. Rodriguez, M.P. Llompart and R. Cela, J. Chromatogr.A, 885 (2000) 291.
65 M.A. Moeller, S. Steinmeyer and T. Kraener, J. Chromatogr., B 713 (1998) 91.
66 A. Pauwels, D.A. Wells, A. Covaci and P.J.C. Schepens, J. Chromatogr., B 723 (1999)
117.
67 T. Kumazawa and 0. Suzuki, J. Chromatogr., B 747 (2000) 241.
68 R.W. Fedeniuk and P.J. Shand, J. Chromatogr.A, 812 (1998) 3.
69 J.M. Soriano, B. Jimenez, G. Font and J.C. Molto, Crit. Revs. Anal. Chem., 31 (2001)
19.
70 K. Albert, R. Brindle, P. Martin and I.D. Wilson, J. Chromatogr. A, 665 (1994) 253.
71 P. Martin, A. Taberner, A. Fairbrother and I.D. Wilson, J. Pharm. Biomed. Anal. 11
(1993) 671.
72 S.K. Poole and C.F. Poole, Analyst, 120 (1995) 1733.
73 L. Schmidt and J.S. Fritz, J. Chromatogr., 640 (1993) 145.
74 Y.-F. Cheng, D.J. Phillips and U. Neue, Chromatographia,44 (1997) 187.
75 M.-C. Hennion, V. Coquart, S. Guenu and C. Sella, J. Chromatogr.A, 712 (1995) 287.
76 S. Guenu and M.-C. Hennion, J. Chromatogr. A, 725 (1996) 57.
77 C. Lepont, A.D. Gunatilleka and C.F. Poole, Analyst, 126 (2001) 1318.
78 S.K. Poole and C.F. Poole, Anal. Commun., 34 (1997) 247.
79 V. Pinchon, J. Chromatogr. A, 885 (2000) 195.
80 H. Sabik, R. Jeannot and B. Rondeau, J. Chromatogr.A, 885 (2000) 217.
81 V. Pignon, R. Jeannot and E. Sauvard, Int. J. Environ. Chem., 75 (1999) 345.
82 S.R. Rubern, W.M. Draper and S.K. Perera, J. Agri. Food Chem., 48 (2000) 4109.
83 M.C. Carson, J. Chromatogr. A, 885 (2000) 343.
84 M.J.M. Wells and L.Z. Yu, J. Chromatogr. A, 885 (2000) 237.
85 R. Saan-Nordhaus and J.M. Anderson, J. Chromatogr. A, 706 (1995) 563.
86 P.R. Haddad, F, Doble and M. Macka, J. Chromatogr. A, 856 (1999) 145.
87 X.-H. Chen, J.-P. Franke, J. Wijsbeek and R.A. de Zeeuw, J. Chromatogr., 617 (1993)
147.
384
88 P. Degel, Clin. Biochem., 29 (1996) 529.
89 O.H. Drummer, J. Chromatogr.B, 733 (1999) 27.
90 S. Paterson, R. Cordero, S. McCulloch and P. Houldsworth, Ann. Clin. Biochem., 37
(2000) 690.
91 E.H.R. Vanderwal, E.R. Brouwer, H. Lingeman and U.A.Th. Brinkman,
Chromatographia,39 (1994) 239.
92 I. Ferrer, D. Barcelo and E.M. Thurman, Anal. Chem., 71 (1999) 1009.
93 B.A. Tomkins and W.H. Griest, Anal. Chem., 68 (1996) 2533.
94 K.-S. Boos and C.-H. Grimm, Trends Anal. Chem., 18 (1999) 175.
95 K.K. Unger, Chromatographia,31 (1991) 507.
96 J. Haginaka, Trends Anal. Chem., 10 (1991) 17.
97 K.-S. Boos, A. Rudolphi, S. Vielhauer, A. Walfort, D. Lubda and F. Eisenbeiss,
Fresenius'J.Anal. Chem., 352 (1995) 684.
98 Z. Yu and D. Westerlund, Chromatographia,44 (1997) 589.
99 C.-H. Grimm, K.-S. Boos, Ch. Apel, K. K. Unger, P. Onnerfjord, L. Heintz, L.-E.
Edholm and G. Marko-Vargra, Chromatographia,52 (2000) 703.
100 R. Koeber, C. Fleischer, F. Lanza, K.S. Boos, B. Sellergren and D. Barcelo, Anal.
Chem., 73 (2001) 2437.
101 E. Hogendoorn and P. van Zoonen, J. Chromatogr.A, 892 (2000) 435.
102 R.M. Izatt, J.S. Bradshaw and R.I. Bruening, Pure Appl. Chem., 68 (1996) 1237.
103 R.M. Izatt, J. Inclusion Phenom. Mol. Recogn. Chem., 29 (1997) 197.
104 N.E. Izatt, R.L. Bruening, K.E. Krakowiak and S.R. Izatt, Ind. Engin. Chem. Res., 39
(2000) 3405
105 D.M. Beals, W.G. Britt, J.P. Bibler and D.A. Brooks, J. Radioanal. Nucl. Chem., 236
(1998) 187.
106 P. Martin, B. Leadbetter and I.D. Wilson, J. Pharm.Biomed. Anal., 11 (1993) 307.
107 H. Tsuchiya, J. Chromatogr., B 720 (1998) 225.
108 J.M. Van Emon, C.L. Gerlath and K. Bowman, J. Chromatogr.B, 715 (1998) 211.
109 V. Pichon, M. Bouzige, C. Miege and M.-C. Hennion, Trends Anal. Chem., 18 (1999)
219.
110 N. Delaunay, V. Pichon and M.-C. Hennion, J. Chromatogr., B 745 (2000) 15.
111 D. Stevenson, J. Chromatogr., B 745 (2000) 39.
112 P.B. Carrasco, R. Escola, M.P. Marco and J.M. Bayona, J. Chromatogr. A, 909 (2001)
61.
113 N. Delaunay-Bertoncini, V. Pichon and M.-C. Hennion, Chromatographia,53 (2001)
S224.
114 M.-C. Hennion and D. Barcelo, Anal. Chim. Acta, 362 (1998) 3.
115 J. Dalluge, T. Hankemeier, R.J.J. Vreuls and U.A.Th. Brinkman, J. Chromatogr.A,
830 (1999) 377.
116 D. Stevenson, Trends Anal. Chem., 18 (1999) 154.
117 B. Sellergren, Trends Anal. Chem., 18 (1999) 164.
118 L. I. Anderson, J. Chromatogr.B, 739 (2000) 163.
119 L. I. Anderson, J. Chromatogr. B, 745 (2000) 3.
120 F. Lanza, B. Sellergren, Chromatographia,53 (2001) 599.
121 A. Martin-Esteban, Fresenius J. Anal. Chem., 370 (2001) 795.
122 J.G. Karlsson, L.I. Andersson and I.A. Nicholls, Anal. Chim. Acta, 435 (2001) 57.
123 A. Martin-Esteban, F. Turiel, D. Stevenson, Chromatographia, 53 (2001) S434.
124 P. Martin, I.D. Wilson and G.R. Jones, Chromatographia,52 (2000) S19.
125 M.-C. Hennion, C. Cau-Dit-Coumes and V. Pichon, J. Chromatogr.A, 823 (1998) 147.
126 C.F. Poole, S.K. Poole, D.S. Seibert and C.M. Chapman, J. Chromatogr.B, 689 (1997)
245.
127 C.F. Poole, A.D. Gunatilleka and R. Sethuraman, J. Chromatogr.A, 885 (2000) 17.
385
128 P. Lovkvist and J.A. Jonsson, Anal. Chem., 59 (1987) 818.
129 I. Liska, J. Chromatogr. A, 655 (1993) 163.
130 M.L. Mayer and C.F. Poole, Anal. Chim. Acta, 294 (1994) 113.
131 M.L. Larrivee and C.F. Poole, Anal. Chem., 66 (1994) 139.
132 D.S. Seibert and C.F. Poole, J. High Resolut. Chromatogr., 21 (1998) 481.
133 N. Masque, M. Galio, R.M. Marce and F. Borrull, J. Chromatogr.A, 771 (1997) 55.
134 V. Pichon and M.-C. Hennion, J. Chromatogr., 665 (1994) 269.
135 V. Pichon, L. Chen, S. Guenu and M.-C. Hennion, J. Chromatogr.A, 711 (1995) 257.
136 C.F. Poole, S.K. Poole and A.D. Gunatilleka, Adv. Chromatogr., 40 (2000) 159.
137 HG.J. Mol, J. Staniewski, H.-G. Jansson, C.A. Cramers, R.T. Ghijsen and U.A.Th.
Brinkman, J. Chromatogr., 630 (1993) 201.
138 H.G.J. Mol, H.-G. Janssen and C.A. Cramers, J. High Resolut. Chromatogr., 16
(1993) 413.
139 M.-C. Hennion and V. Coquart, J. Chromatogr., 642 (1993) 211.
140 A. Gelencser, G. Kiss, Z. Krivacsy, Z. Varga-Puchony and J. Hlavay, J. Chromatogr.
A, 693 (1995) 217.
141 A. Gelencser, G. Kiss, Z. Krivacsy, Z. Varga-Puchony and J. Hlavay, J. Chromatogr.
A, 693 (1995) 227.
142 K.G. Miller and C.F. Poole, J. High Resolut. Chromatogr., 17 (1994) 125.
143 E. Baltussen, F. David, P. Sandra, H.-G. Janssen and C.A. Cramers, J. Chromatogr.
A, 805 (1998) 237.
144 E. Baltussen, H. Snijders, H.-G. Janssen, P. Sandra and C.A. Cramers, J.
Chromatogr. A, 802 (1998) 285.
145 W.P.N. Fernando, M.L. Larrivee and C.F. Poole, Anal. Chem., 65 (1993) 588.
146 D. Bolliet and C.F. Poole, Chromatographia,46 (1997) 381.
147 D.S. Seibert and C.F. Poole, J. High Resolut. Chromatogr., 18 (1995) 226.
148 D.S. Seibert, C.F. Poole and M.H. Abraham, Analyst, 121 (1996) 511.
149 D.S. Seibert and CF. Poole, Chromatographia,41 (1995) 51.
150 D.S. Seibert and C.F. Poole, Anal. Commun., 35 (1998) 147.
151 M.L. Mayer, S.K. Poole and C.F. Poole, J. Chromatogr. A, 697 (1995) 89.
152 T. Baczek, M. Markuszewski, R. Kaliszan, M.A. van Straten and H.A. Claessens, J.
High Resolut. Chromatogr., 23 (2000) 667.
153 C.F. Poole and S.K. Poole, J. Chromatogr.A (2002), in press.
154 M.H. Abraham and J.C. McGowan, Chromatographia,23 (1987) 243.
155 M.H, Abraham, G.S. Whiting, R.M. Doherty and W.J. Shuely, J. Chem. Soc. Perkin
Trans., 2 (1990) 1451.
156 M.H. Abraham, C.F. Poole and S.K. Poole, J. Chromatogr. A, 842 (1999) 79.
157 J.A. Platts, D. Butina, M.H. Abraham and A. Hersey, J. Chem. Inf. Comput. Sci., 39
(1999) 835.
158 W. Kiridena and C.F. Poole, Analyst, 123 (1998) 1265.
159 J.H. Park, M.H. Yoon, Y.K. Ryu, B.E. Kim, J.W. Ryu and M.D. Jang, J. Chromatogr.
A, 796 (1998) 249.
160 F.Z. Oumada, M. Roses, E. Bosch and M.H. Abraham, Anal. Chim. Acta, 382 (1999)
301.
161 J. Li and D.A. Whitman, Anal. Chim. Acta, 368 (1998) 141.
162 J.C. Pearce, A. Churchill and A.C. Terry, Chemom. Intell. Lab. Sys., 17 (1992) 213.
163 P. Hubert, P. Chiap, M. Moors, B. Bourguignon, D.L. Massart and J. Crommen, J.
Chromatogr. A, 665 (1994) 87.
164 D.T. Rossi and N. Zhang, J. Chromatogr.A, 885 (2000) 97.
165 I. Liska, J. Chromatogr., 655 (1993) 163.
166 G. Font, J. Manes, J.C. Malto and Y. Pico, J. Chromatogr., 642 (1993) 135.
167 L. Bovanova and E. Brandstetenova, J. Chromatogr. A, 880 (2000) 149.
386
168 R. Sasano, T. Hamada, M. Kurano and M. Furuno, J. Chromatogr. A, 896 (2000) 41.
169 J.R. Veraart, H. Lingeman and U.A.Th. Brinkman, J. Chromatogr.A, 856 (1999)
483.
170 D. Martinez, M.J. Cugat, F. Borrull and M. Calull, J. Chromatogr.A, 902 (2000) 65.
171 M. Valcarcel, L. Arce and A. Rios, J. Chromatogr.A, 924 (2001) 3.
172 P. Simon, M. Lafontaine, P. Delsaut, Y. Morele and T. Nicot, J. Chromatogr.B, 748
(2000) 337.
173 S. Salado and L.E. Vera-Avila, J. Chromatogr.B, 690 (1997) 195.
174 E. Schoenzetter, V. Pichon, D. Thiebaut, A. Fernandez-Alba and M.-C. Hennion, J.
Microcol. Sep., 12 (2000) 316.
175 E.A. Hogendoorn, E. Dijkman, B. Baumann, C. Hidalgo, J.V. Sancho and F.
Hernandez, Anal. Chem., 71 (1999) 1111.
176 Z. Yu and D. Westerlund, Chromatographia,47 (1998) 299.
177 R.A.M. vander Hoeven, A.J.P. Hofte, M. Frenay, H. Irth, U.R. Tjaden, J. vander
Greef, A Rudolfi, K.-S. Boos, G.M. Varga and L.E. Edholm, J. Chromatogr.A, 762
(1997) 193.
178 R. Wissiack, E. Rosenberg and M. Grasserbauer, J. Chromatogr. A, 896 (2000) 159.
179 J. Slobodnik, H. Lingeman and U.A.Th. Brinkman, Chromatographia,50 (1999) 141.
180 N. Masque, R.M. Marce and F. Borrull, J. Chromatogr.A, 793 (1998) 257.
181 A.C. Hogenboom, M.P. Hofman, D.A. Jolly, W.M.A. Niessen and U.A.Th. Brinkman,
J. Chromatogr.A, 885 (2000) 377.
182 A.J.H. Louter, U.A.Th Brinkman and R.T. Ghijsen, J. Microcol. Sep., 5 (1993) 303.
183 J. Hankemeier, S.P.J. van Leeuwen, R.J.J. Vreuls and U.A.Th. Brinkman, J.
Chromatogr.A, 811 (1998) 117.
184 T. Hankemeier, A.J.H. Louter, J. Dalluge, R.J.J. Vreuls and U.A.Th. Brinkman, J.
High Resolut. Chromatogr., 21 (1998) 450.
185 T.H.M. Noij and M.M.E. van der Kooi, J. High Resolut. Chromatogr., 18 (1995) 535.
186 A.J.H. Louter, S. Ramalho, D. Jani, J.J. Vreuls and U.A.Th. Brinkman, J. Microcol.
Sep., 8 (1996) 469.
187 K.K. Verma, A.J.H. Louter, A. Jain, E. Pocurull, J.J. Vreuls and U.A.Th. Brinkman,
Chromatographia,44 (1997) 372.
387
Chapter 13
13.1 INTRODUCTION
Solid phase microextraction was developed to address the need for rapid sample
preparation both in the laboratory and on-site. With this the technique a small
amount of extracting phase dispersed on a solid support is exposed to the sample
for a well defined period of time. In one approach a partitioning equilibrium
between the sample matrix and extraction phase is reached. In this case
convection conditions do not affect the amount extracted. In a second approach
utilizing short time pre-equilibrium extraction, if convection/agitation are
constant, then the amount of analyte extracted is related to time. Quantification
can then be performed based on timed accumulation of analytes in the coating.
Figure 13.1 illustrates several implementations of SPME which have been
considered. They mainly include open bed extraction concepts such as coated
fibres, vessels, agitation mechanism disks, but in-tube approaches are also
considered. Some better address issues associated with agitation while others
address ease of implementing sample introduction to the analytical instrument.
It should be noted that solid phase microextraction was originally named after
the first experiment using an SPME device which involved extraction on solid
fused silica fibres, and later as such, as a reference to the appearance of the
extracting phase, relative to a liquid or gaseous donor phase, even though it is
recognized that the extraction phase is not always technically a solid.
Solid phase microextraction is often mistakenly considered as another form
of solid phase extraction or micro-SPE. However, there are significant differ-
ences between the methods. Solid phase extraction is essentially a three-step
process (see Chapter 12 "Principles and Practice of Solid-Phase Extraction" ). A
sample is initially passed through the sorbent bed, and analytes present in the
sample are exhaustively extracted from the sample matrix to the solid sorbent.
In a second step, unwanted analytes are selectively desorbed from the solid
sorbent by washing with a solution capable of desorbing unwanted analytes, but
leaving desired analytes retained on the sorbent. In the final step the wash solu-
tion is changed for one able to desorb analytes of interest. The resulting eluent
may then be concentrated by evaporation, to the desired volume. Solid phase
microextraction however takes advantage of equilibrium extraction and selec-
ComprehensiveAnalytical Chemistry XXXVII
J. Pawliszyn (Ed.)
© 2002 Elsevier Science B.V. All rights reserved 389
WIN Extraction Phase
Tube
I
tive sorption from the matrix onto the coating. In the first step, the coating is
exposed to the sample and analytes with a high affinity for the sorbent are selec-
tively extracted. In the second step, everything extracted by the fibre is desorbed
into the analytical instrument. No intermediate clean-up step is normally imple-
mented. Micro-SPE is more related to SPE as it is a total extraction method, but
utilizes a reduced sample and sorbent volume. A comparison with SPME is
therefore inappropriate.
A degree of selectivity is required for any sample preparation method. It is
impractical to introduce all compounds present in a sample to an analytical
instrument. The method developed must eliminate compounds incompatible
with the instrument including matrix components. It is also desirable to remove
as many of the unwanted compounds as possible, to make the resulting data
interpretation as clean and simple as possible. Thus, with selective extraction,
sample preparation is simplified and typically results in significant savings in
time and precision.
Selectivity is therefore quite important when choosing an SPME coating.
High capacity, even for a range of analytes, is more important for solid phase
extraction, where prevention of break-through is a significant concern. Because
break-through is not an issue to be addressed in an equilibrium extraction
method such as SPME, more emphasis may be placed on sorbent selectivity.
SPME differs from SPE in another significant way: because of the large
volume of sorbent required relative to SPME, SPE sorbent has the potential to
retain non-adsorbed components in the void volume. It is difficult to design a
wash regime that removes unwanted compounds completely, without impacting
retention of the analyte(s) of interest. In this way there is the potential that
unwanted compounds may remain, either adsorbed, or present as non-adsorbed
analytes in the bulk of the sorbent. Because of the geometry of the SPME device,
and the modes of extraction used, unwanted analytes are not normally present
in the sorbent at the time of desorption.
SPME devices have an open-bed structure, relative to SPE devices where the
extraction medium is packed into a cartridge-like device. In SPME the surface of
390
the extraction phase is itself accessible for analysis. This is less true with in-tube
SPME, although limited surface characterization has been performed with the
capillaries as well. Therefore, with SPME it is possible to perform convenient
spectroscopic analysis of surface adsorbed components and not only extracted
chemical species, but also composition of collected aerosols or particulates. This
can have an important advantage in speciation and characterization of natural
system.
In this chapter the focus is placed on two primary implementations of the
technique to date: fibre and in-tube SPME. These have been explored theoretic-
ally and experimentally to a large extent. The first is based on externally coated
fibres mounted in a syringe-like needle for protection, and the second on
internally coated tubes or capillaries. General conclusions obtained in the
studies involving these systems can be extended to other geometries.
391
syringe needle syringe barrel plunger cap
Fig. 13.2. The custom-made SPME device based on Hamilton 7000 series syringe.
attach enbhub
a-attachment hub
" \\
sealing septum _
cooling
Coils
- septum piercing T
needle
microtubing -coating
coated fused silica fiber
Fig. 13.3. Simple versions of the SPME device using (a) coated fibre and (b) internally coated
tubing.
392
Plunger
Barrel
-Plunger Retaining Screw
Z-slot
Hub-Viewing Window
Adjustable Needle
GuidelDepth Gauge
TensioningSprin
g , - SeptumPiercingNeedle
Sealing Septum
4 - FiberAttachment
Tubing
F u s ed
Coated SilicaFiber
Fig. 13.4. Design of the first commercial SPME device made by Supelco.
393
by increasing the surface to volume ratio of the extraction phase. The main dis-
advantage of these approaches, however, is loss of convenience associated with
syringe configuration, in particular in introduction to the analytical instrument.
This step necessitates the use of high volume desorption devices and creates dif-
ficulties in automation of the extraction process as well as handling volatile com-
pounds which are lost during transfer of the extraction phase from sample to the
injection system. Since high sensitivities are obtained for hydrophobic high
molecular weight compounds with fibre SPME the advantages of using larger
volume phases are limited, especially for small sample volumes [11].
Although to date SPME devices have been used principally in laboratory
applications, more current research has been directed toward remote
monitoring, particularly for clinical, field environmental, and industrial hygiene
applications. In their operating principles, such devices are analogous to the
devices described above, but modifications are made for greater convenience in
given applications. For example, as shown in Fig. 13.5a, adding a tube with a
small opening to cover the needle of the SPME syringe results in a useful device
for breath analysis in a non-invasive clinical application [12]. This design can be
further improved by adding a one-way valve mounted on the aperture, but the
concept of operation remains the same.
An important feature of a field device is the ability to preserve extracted
analytes in the coating. The simplest practical way to accomplish this goal is to
seal the end of the needle with a piece of septum. Additionally, cooling extends
the storage time. Polymeric septum material, however, may cause losses of
analytes from the fibre. Therefore, a more appropriate approach is to use metal
to metal (or solid polymer) seals. Figure 13.5b illustrates an example of a device
construction based on a "leaf"' closure. It is anticipated that future devices
designed for field applications will be more rugged than the current laboratory
versions and will look more like "sticks" or "pens" than syringes.
Recently, different configurations of field samplers were evaluated for the
analysis of volatile compounds [13]. A percentage of compound retained in the
coating was evaluated for the four devices, at different storage times and
exposed fiber
aperture
inert tubing
(a)
Fig. 13.5. SPME device modified for (a) breath analysis and (b) field application.
394
temperatures, and for different fibre coatings. The PDMS fibre demonstrated
the lowest ability to retain these compounds. Carboxen-PDMS had the highest
and PDMS-DVB was intermediate. The devices employed various sealing
methods, to preserve the samples on the fibre, and to protect from external
contamination. The Supelco field sampler seals by retracting the outer needle
behind a silicone septum. Alternatively, a conventional manual sampler may be
sealed by pressing the end of the needle into a piece of septum. Two prototype
devices constructed were sealed by either a leaf system (Fig. 13.5b), which opens
automatically when the outer needle is exposed, or by capping the outer needle
after sampling, with a teflon cap.
Another promising technique uses a valve syringe, where the fibre is
withdrawn into the barrel of the syringe along with a sample of the air being
analyzed, by means of retracting the syringe plunger. The valve is then sealed
until analysis. Initial data describing the new approaches clearly demonstrate
advantages of the new designs to seal the fibre in the needle, compared to the
commercial devices, by eliminating the losses and contamination to the septa
material of volatile components.
395
(a)
Left: Fig. 13.6. Two different implementations of the SPME technique: (a) polymer coated on
outer surface of fibre; (b) polymer coated on internal surface of capillary tube.
Right: Fig. 13.7. Microextraction with SPME. Vt, volume of fibre coating; Kr,, fibre/sample
partition coefficient; V., volume of sample; CO, initial concentration of analyte in the sample.
n K= ,Vf Co (13.1)
KrfVf +V
where n is the number of moles extracted by the coating, Kf, is a fibre coating/
sample matrix distribution constant, Vf is the fibre coating volume, Vs is the
sample volume, and CO is the initial concentration of a given analyte in the
sample.
Strictly speaking, this discussion is limited to partitioning equilibrium
involving liquid polymeric phases such as poly(dimethylsiloxane). The method of
analysis for solid sorbent coatings is analogous for low analyte concentration,
since the total surface area available for adsorption is proportional to the coating
volume if we assume constant porosity of the sorbent. For high analyte concen-
trations, saturation of the surface can occur, resulting in nonlinear isotherms as
discussed later. Similarly high concentration of a competitive interference com-
pound can displace the target analyte from the surface of the sorbent.
Equation (13.1), which assumes that the sample matrix can be represented
as a single homogeneous phase and that no headspace is present in the system,
can be modified to account for the existence of other components in the matrix
by considering the volumes of the individual phases and the appropriate
distribution constants. The extraction can be interrupted and the fibre analyzed
prior to equilibrium. To obtain reproducible data, however, constant convection
conditions and careful timing of the extraction are necessary.
Simplicity and convenience of operation make SPME a superior alternative
to more established techniques in a number of applications. In some cases, the
technique facilitates unique investigations. Equation (13.1) indicates that, after
equilibrium has been reached, there is a direct proportional relationship
between sample concentration and the amount of analyte extracted. This is the
basis for analyte quantification. The most visible advantages of SPME exist at
the extremes of sample volumes. Because the set-up is small and convenient,
coated fibres can be used to extract analytes from very small samples. For
example, SPME devices are used to probe for substances emitted by a single
flower bulb during its life span; the use of sub-micrometer diameter fibres
396
permits the investigation of single cells. Since SPME does not extract target
analytes exhaustively, its presence in a living system should not result in
significant disturbance. In addition, the technique facilitates speciation in
natural systems, since the presence of a minute fibre, which removes small
amounts of analyte, is not likely to disturb chemical equilibria in the system. It
should be noted however, that the fraction of analyte extracted increases as the
ratio of coating to sample volume increases. Complete extraction can be achieved
for small sample volumes when distribution constants are reasonably high. This
observation can be used to advantage if exhaustive extraction is required. It is
very difficult to work with small sample volumes using conventional sample
preparation techniques. Also, SPME allows rapid extraction and transfer to an
analytical instrument. These features result in an additional advantage when
investigating intermediates in the system. For example, SPME was used to
study biodegradation pathways of industrial contaminants [15]. The other
advantage is that this technique can be used for studies of the distribution of
analytes in a complex multiphase system [16] and speciate different forms of
analytes in a sample [17].
In addition, when sample volume is very large, Eq. (13.1) can be simplified to:
n = KfVfCo (13.2)
which points to the usefulness of the technique for field applications. In this
equation, the amount of extracted analyte is independent of the volume of the
sample. In practice, there is no need to collect a defined sample prior to analysis
as the fibre can be exposed directly to the ambient air, water, production stream,
etc. The amount of extracted analyte will correspond directly to its concentra-
tion in the matrix, without being dependent on the sample volume. When the
sampling step is eliminated, the whole analytical process can be accelerated, and
errors associated with analyte losses through decomposition or adsorption on
the sampling container walls will be prevented. This advantage of SPME could
be enhanced practically, by developing portable field devices on a commercial
scale.
Three basic types of extraction can be performed using SPME: direct extraction,
headspace configuration, and a membrane protection approach. Figure 13.8
illustrates the differences among these modes. In the direct extraction mode
(Fig. 13.8a), the coated fibre is inserted into the sample and the analytes are
transported directly from the sample matrix to the extracting phase. To
facilitate rapid extraction, some level of agitation is required to transport
analytes from the bulk of the solution to the vicinity of the fibre. For gaseous
samples, natural convection of air is sufficient to facilitate rapid equilibration.
For aqueous matrices, more efficient agitation techniques, such as fast sample
flow, rapid fibre or vial movement, stirring or sonication are required [6,18].
397
These conditions are necessary to reduce the effect caused by the "depletion
zone" produced close to the fibre as a result of fluid shielding and slow diffusion
coefficients of analytes in liquid matrices.
In the headspace mode, the analytes need to be transported through the bar-
rier of air before they can reach the coating. This modification serves primarily
to protect the fibre coating from damage by high molecular mass and other
non-volatile interferences present in the sample matrix, such as humic materials
or proteins. This headspace mode also allows modification of the matrix, such as
a change of the pH, without damaging the fibre. Amounts of analyte extracted
into the coating from the same vial at equilibrium using direct and headspace
sampling are identical as long as sample and gaseous headspace volumes are the
same. This is caused by the fact that the equilibrium concentration is independ-
ent of fibre location in the sample/headspace system. If the above condition is not
satisfied, a significant sensitivity difference between the direct and headspace
approaches exists only for very volatile analytes.
The choice of sampling mode has a very significant impact on extraction
kinetics, however. When the fibre coating is in the headspace, the analytes are
removed from the headspace first, followed by indirect extraction from the
matrix, as shown in Fig. 13.8b. Overall mass transfer to the fibre is typically
limited by mass transfer rates from the sample to the headspace. Therefore,
volatile analytes are extracted faster than semivolatiles since they are at a
higher concentration in the headspace which contributes to faster mass
transport rates through the headspace. Temperature has a significant effect on
the kinetics of the process by determining the vapour pressure of analytes. In
fact, the equilibration times for volatiles are shorter in the headspace SPME
mode than for direct extraction under similar agitation conditions. This outcome
is produced by two factors: that a substantial portion of the analyte is in the
headspace prior to extraction, and that diffusion coefficients in the gaseous
phase are typically four orders of magnitude larger than in liquid media. Since
concentrations of semivolatiles in the gaseous phase at room temperature are
typically small, however, overall mass transfer rates are substantially lower and
result in longer extraction times. They can be improved by using very efficient
agitation or by increasing the extraction temperature [19].
Fig. 13.8. Modes of SPME operation: (a) direct extraction; (b) headspace SPME; (c) membrane-
protected SPME.
398
a
' 15
o tO
I 10
xc
A
5
0 . 0
0 10 20 30 40 50 60 70 80
Fig. 13.9. Time extraction profile obtained for headspace solid phase microextraction of several
PAHs from aqueous samples at (a) 75% and (b) 100% stirring rates. A, naphthalene; B,
acenaphthene; C, phenanthrene; D, chrysene.
Figure 13.9 illustrates the effect of agitation on the extraction time profile
obtained for polynuclear aromatic hydrocarbons (PAHs). As the rotational speed
of the magnetic stirrer increases, the equilibration time of naphthalene and
acenaphthene decreases from 8 min to 3 min and from 25 min to 10 min,
respectively. For less volatile analytes, phenanthrene and chrysene, the equili-
bration is not reached during the experimental period in either case, but the
amount of analyte extracted after 70 min extraction at low agitation (Fig. 13.9a)
is about the same as after 45 min of more efficient stirring (Fig. 13.9b). The use
of even more efficient agitation techniques, such as direct sonication, further
cuts the extraction time [20].
The effect of elevated sampling temperature is a lowering of fibre sample
partition coefficients. This decreases the amount extracted at equilibrium but it
may be acceptable if target limits of detection still can be reached. This effect is
demonstrated dramatically with the analysis of amphetamines. As shown in Fig.
13.10, room temperature extraction produces a very long equilibrium extraction
time, but ultimately the highest amount extracted. Conversely, the highest
temperature tested (73°C) produces a very short equilibration time (ca. 5 min)
but a significantly lower equilibrium amount extracted. An increase in the
Henry constant and diffusion coefficient improves kinetics at higher
temperature combined with a decrease in fibre capacity determined by the
distribution constant contributes to a substantial decrease in extraction time.
To enable the simultaneous analysis of very volatile substances (gases) and
less volatile analytes, the headspace SPME technique can be combined with
static headspace sampling by means of the gastight/SPME device illustrated in
Fig. 13.11 [21]. A fused silica fibre, coated with poly(dimethylsiloxane) (PDMS),
is connected with 30 gauge stainless steel (SS) tubing. The empty end of this SS
tubing/fibre assembly is then mounted to the plunger of a Hamilton 500 lI
gas-tight syringe. When the fibre is withdrawn into the syringe needle, a certain
volume of gas is also withdrawn into the gas-tight syringe through the needle
opening. For sampling, the fibre is first withdrawn into the syringe needle,
399
plunger-
-fWs
syringe
barrel
f
!
Teflon
tip-
SS
tubing
needle -
fiberand
0 20 40 60 80 100 coating -
Time (min)
Left: Fig. 13.10. Effect of extraction temperature on equilibrium time and amount extracted, for
headspace methamphetamine analysis. Extraction conditions: PDMS 100 Am fibre, 0.5 M KOH,
saturated NaCi, 2 ml sample in a 4 ml vial, extraction temperature as shown, extraction time 15
min., analysis by GC-FID, desorption time 15 min. (): 22C; (): 40°C; (): 60°C; (): 73°C.
Right: Fig. 13.11. Gastight SPME device.
which is used to punch through the sample vial septum. The fibre is then
exposed into the headspace by lowering the plunger for a predetermined period
of time to establish analyte equilibrium among the coating, the headspace, and
the sample matrix. Movement of the plunger up and down can be used to
increase mass transport from the headspace to the coating and shorten the
equilibration times. Then the plunger is raised to a pre-marked position, which
allows a predetermined volume of headspace gas to be withdrawn into the
gastight/SPME device. The raised plunger also withdraws the fibre into the
needle for protection. Upon injection, the volume of headspace is injected at the
same time as analytes are desorbed.
Figure 13.12 illustrates the distribution of 28 typical volatile analytes,
between a PDMS coating of about 1 L1lvolume and the 110 /1 headspace con-
tained in the sampling device from Fig. 13.11. The presence of the gaseous
headspace in the gastight SPME device improves sensitivity only for very vola-
tile analytes with boiling temperatures below room temperature. The resulting
limits of detections (LODs) are mid to low parts per trillion (ppt) for the full
range of volatiles [21].
Figure 13.8c shows the principle of indirect SPME extraction through a
membrane. The initial purpose of the membrane barrier was to protect the fibre
against damage, similar to the use of headspace SPME when very dirty samples
are analyzed. Membrane protection is advantageous for determination of
analytes having volatilities too low for the headspace approach. In addition, a
membrane made from appropriate material can add a certain degree of
selectivity to the extraction process. The kinetics of membrane extraction are
substantially slower than for direct extraction though, because the analytes
400
Combination
1.2
U) 0.8
)- 0.6
)
aZ 0.4
0.2
0
0 5 10 15 20 25 30
Elution Order
Fig. 13.12. Normalized peak area of 28 VOCs, which were sampled from the headspace of a
salt-saturated water sample at 100 ppb using three extraction methods; the peak area of VOCs
sampled by the gastight SPME was normalized to 1 for all analytes. The x axis gives the elution
order numbers of VOCs.
must diffuse through the membrane before they can reach the coating. The use
of thin membranes and increased extraction temperatures will result in faster
extraction times [22]. The thicker membranes can be used to slow down the
mass transfer through the membrane resulting in the time weighted average
(TWA) measurement for aqueous and gaseous samples discussed latter.
Figure 13.13 illustrates that there are two fundamental approaches to in-tube
SPME: active or dynamic where the analytes are passed through the tube, and
passive or static where the analytes are transferred into the sorbent using
diffusion. In either approach, the coating may be supported on a fused silica rod,
or coated on the inside of a tube or capillary. Below we briefly discuss the
theoretical aspects of the extraction processes which use these geometric
arrangements.
401
a b
- ~~~~~~~~~--·---
~_ --f-
II
P I
I
:6 "j i
5
mc:
9
Fig. 13.13. Comparison of (a) passive versus (b) dynamic modes of in-tube extraction.
of the capillary and inversely proportional to the linear flow rate of the fluid.
Extraction time also increases with an increase in the coating/sample distribu-
tion constant and with the thickness of the extracting phase but decreases with
an increase in the void volume of the capillary. An increase in the coating/sample
distribution constant produces an increase in the absolute amount extracted. It
has been observed that increases in amounts extracted can be achieved in many
cases by pre-conditioning the capillary with methanol or some other appropriate
solvent, prior to extraction. Enhancement has even been observed when a plug
of methanol is aspirated into the capillary before the sample is drawn in, and
follows the sample plug in the capillary during the extraction aspirate/dispense
steps. This is analogous to the solvent pre-conditioning used in SPE to enhance
extraction.
In practice, in-tube SPME is implemented by replacing a section of the
tubing in a commercially available autosampler, and then programming the
autosampler to pass sample in and out of the extraction capillary until equilib-
rium or a suitable extraction level has been reached. Several options for
implementing in-tube SPME are summarized in Fig. 13.14.
It should be emphasized that the above discussion is valid only for direct
extraction when the sample matrix passes through the capillary. This approach
is limited to particulate-free gas and clean water samples. The headspace SPME
approach can broaden the application of in-tube SPME. In that case, careful con-
sideration to the mass transfer between sample and headspace should be given
in order to describe the process properly. Also, if the flow rate is very rapid pro-
ducing turbulent behaviour and the coating/sample distribution constant is not
very high, then perfect agitation conditions are met and Eq. (13.3) can be used to
402
HPLC Pump HPLC Pump
P
c) Flush Loop In-tibe SPME To column
te = t95% b (13.3)
2Dr
In this equation, bt refers to the thickness of the sorbent material, and Df refers
to analyte diffusion coefficient in the sorbent.
The removal of analytes from a tube is an elution problem analogous to
frontal chromatography and has been discussed in detail [24]. In general, if the
desorption temperature of a GC is high and thin coatings are used, then all the
analytes are in the gas phase as soon as the coating is placed in the injector, and
the desorption time corresponds to the elution of two void volumes of the
capillary. For liquid desorption (into a liquid chromatography system, for exam-
ple) (Fig. 13.15), the desorption volume can be even smaller since the analytes
can be focused at the front of the desorption solvent [25].
403
Dispenser
LOAD position
ample vial
Fig. 13.15. SPME with modified needle to allow in-needle extraction and automated flow-
through analysis of small samples and automation of a coated tubing solid phase microextractor
using Spark Holland Autosampler.
dn = AD dt = AD Ctdt (13.4)
dz Z
where AC(t)/Z is a value of the gradient established in the needle between needle
opening and the position of the extracting phase, Z; AC(t) = C(t)-Cz, where C(t) is
a time dependent concentration of analyte in the sample in the vicinity of the
needle opening, and C. is the concentration of the analyte in the gas phase in the
404
A
a i) I
-C=-
s M RX, concentration
gradient inside tube
C=O0
L Oto30mm
Fig. 13.16. Use of SPME for in-tube time weighted average sampling: (a) schematic, (b)
adaptation of commercial SPME manual extraction holder.
n =D - C(t)dx (13.5)
405
The exploitation of restricted access to the absorbing medium allows the
implementation of SPME for time-weighted average (TWA) sampling. Where
diffusion to the sorbent surface is limited, the sorbent can act as a sort of "zero
sink" such that extraction is very far from equilibrium under normal sampling
conditions. In practice then, any analytes reaching the sorbent surface are
absorbed, essentially exhaustively. The rate of diffusion however, is still depend-
ent on the sample concentration, so the total amount absorbed by the coating is
proportional to the average of analyte concentrations over time, hence time-
weighted average sampling is achieved. This has been implemented to date with
the conventional fibre assembly, by retracting the fibre a known distance inside
the needle (Fig. 13.16b). The small size of the needle orifice limits diffusion to
the sorbent surface, and the ultimate diffusion rate is also a function of the
distance between the fibre tip and the end of the needle. Depending on the
volatility and concentration of the analyte of interest, the fibre may be
positioned either closer to or further from the end of the needle, to achieve the
desired degree of non-equilibrium extraction and sensitivity. It would also be
possible to implement this type of sampling with the sorbent coated on the
interior wall of a capillary. To date however, the retractable needle implementa-
tion has gained the most attention, due to its ease of use and adjustability for the
analyte and sample at hand.
Time-weighted averagesamplers
The theory of TWA sampling using an SPME device, was evaluated both in the
laboratory using a standard gas mixture of airborne hydrocarbons, and by
comparison to standard methods for field sampling [28,29]. One hundred ,um
PDMS fibre was used in the device shown in Fig. 13.16b to determine mass
uptake rates or sampling rates (SR). These values were determined for Z values
of 0.3 cm, 1 cm, and 3 cm. The intercept value for the linear uptake rate equation
represents the amount adsorbed on the stainless steel tubing. The higher the
molecular mass of the compound, the higher is the intercept value. However, for
a specific analyte, the SR value is proportional to the concentration and
diffusion coefficient of the analyte in the gas phase and inversely proportional to
the pathlength Z. Therefore if these SR values are used to determine an
unknown concentration without accounting for analyte adsorbed to the needle, a
positive error will be introduced, which is the intercept value. This error
becomes more significant as the volatility of the compound decreases.
Experimental values of uptake rates for the most volatile compounds like
pentane and hexane and PDMS as a coating were less than expected from the
theory. These compounds have high vapour pressure and diffusion coefficients,
which allows a fraction of them to leave the fibre coating before desorption in the
GC. In addition, due to their low partition coefficients for PDMS, these com-
pounds reach equilibrium in a short time which makes it difficult to load less
than 5% of the equilibrium amount. These problems were solved by using an
adsorptive coating like PDMS-DVB.
406
The problem of adsorption of the least volatile compounds on the stainless
steel tubing is aggravated by retracting the fibre further into its housing needle
(3 cm). The area of the adsorptive surface increases as we retract the fibre
further into the needle. However, retracting the fibre further slows the
equilibrium process and minimizes the loss of the most volatile compounds. It
may be possible to use the PDMS-DVB fibre to minimize the loss of volatile
compounds from the fibre coating.
TABLE 13.1
Summary of data for analysis of styrene, obtained from field study using (A) TWA charcoal
tubes, (B) TWA by SPME with 100 gam PDMS fibre (retracted 0.3 cm)
Sample type SPME 100 Am PDMS Charcoal tubes Passive badge PID
5 min sample (g/l) 130 97 90 50-250
30 min TWA (g/1l) 56 54 72 N/A
407
ful for designing TWA devices for aqueous sample monitoring using hydrophilic
membranes since this approach reduces effect of needle surface adsorbtion.
Equations (13.1) and (12.2) indicate that the efficiency of the extraction process
is dependent on the distribution constant Kfs. This is a characteristic parameter
that describes properties of a coating and its selectivity toward the analyte
versus other matrix components. Specific coatings can be developed for a range
of applications. Coating volume determines method sensitivity as well (see Eq.
(13.1)), but thicker coatings result in longer extraction times (Eq. (13.3)).
Therefore, it is important to use the appropriate coating for a given application.
This is clearly demonstrated in Fig. 13.17, which compares the performance of
two different coatings for analysis of polar and nonpolar compounds from an
aqueous matrix. The distribution constant and the sensitivity of the method
drop over two orders of magnitude for o-xylene, and increase by an order of
magnitude for 2,4-dichlorophenol when the film is changed from nonpolar
PDMS (Fig. 13.17a) to polar poly(acrylate) polymer (PA) (Fig. 13.17b) [31].
Coating selection and design can be based on chromatographic experience.
To date, several experimental coatings have been prepared and investigated
for a range of applications. In addition to liquid polymeric coatings such as
PDMS for general applications, other more specialized materials have been
developed. For example, ion exchange coatings were used to remove metal ions
and proteins from aqueous solutions [32,33], liquid crystalline films to extract
planar molecules and carbowax for polar analytes [2], metal rods to electro-
deposit analytes [19,34], pencil "leads" to extract pesticides [35], and Nafion
coatings to extract polar compounds from nonpolar matrices [36].
Polypyrole coatings have been recently developed to extract polar or even
ionic analytes and possibly to explore the conductive polymer properties of the
polymer [37]. This could involve applying a charge to the polymer during
extraction in order to selectively extract analytes of interest, and then reversing
the charge to facilitate desorption. One of the main difficulties limiting the wide
oxylene
a E b
E)
:
/ K=794 2,4-dichlorophenol
0C .
0 o-xylene K\\
K =4
a
-FU
-~ i , \ --
`
A~~~~~~~~1
~ ~15 - -
-7-~~~~~~~~~~~~~~-
a6
-~~~~~~~~~~~~~~~~~~~
" -- --
10 10 15
Retention time (min) Retention time (min)
Fig. 13.17. Total ion current GC-MS chromatogram of benzene, toluene, ethylbenzene, and
o,m,p-xylenes (BTEX) and phenol in water extracted with: (a) a poly(dimethylsiloxane) coating,
and (b) a poly(acrylate) coating.
408
application of SPME/LC is the absence of a suitable stationary phase that not
only has high extraction ability for the polar analytes but is also stable in
solutions of various matrices. Polypyrole coatings have been investigated to fill
this gap.
Conducting polymers are versatile materials in which molecular/analyte
recognition can be achieved in different ways, including: (1) the incorporation of
counter ions that introduce selective interactions, (2) utilising the inherent and
unusual multifunctionality (hydrophobic, base-acid and TI-it interactions, polar
functional groups, ion-exchange, hydrogen bonding, electroactivity, and etc.) of
the polymers, (3) the introduction of functional groups to the monomers, (4) the
co-deposition of metals or other monomers within the polymer, and (5) the
application of appropriate electrochemical potentials. Based on these properties,
conducting polymers have found wide application in many fields including
separation science [38], chemical sensors [39] and electrochemical analysis [40].
So far, the widely used conducting polymers are based on polypyrrole, poly-
thiophene and polyaniline. Of these three classes of materials, polypyrrole (PPY)
and its derivatives have seen the most use and have been intensively studied in
recent years, due to additional advantages: (1) they can be easily polymerized
from organic or aqueous media at neutral pH by electrochemical or chemical
methods, (2) they are relatively stable in air and solution, and (3) pyrrole
monomer and some of its derivatives are available commercially.
Polypyrrole has been coated on the inner surface of a fused silica GC
capillary by the chemical polymerisation method. This PPY coated capillary has
been used successfully for automated in-tube SPME and determination of
B-blockers in urine and serum samples when coupled with liquid chromatogra-
phy/electrospray ionisation mass spectrometry (LC/ESI-MS) [41]. Compared
with the previous study using Omegawax 250 capillary for extraction of
P-blockers [42], PPY coated capillary has shown better extraction ability for
most of the compounds studied and therefore lower detection limits have been
achieved. This method can be extended to other compounds such as amine
containing drugs and anionic species due to the acid-base interaction and ion
exchange properties of polypyrrole coating.
A very pronounced difference in selectivity toward target analytes and
interferences can be achieved by using surfaces common to affinity chromatog-
raphy. Using the method of polymer imprinting [43], antibody mimics can be
generated with specificities to an analyte of choice. Briefly, the desired affinity
can be introduced by adding an amount of the compound of interest to the
polymerization reaction. This "pattern" chemical may be removed after poly-
merization, leaving vacant sites of a specific size and shape, suitable for binding
the same chemical again from an unknown sample. In this technique, we have
observed that non-specific binding should be controlled for, but that enhance-
ments in sensitivity are seen, particularly at low analyte concentrations.
Antibodies shows high affinity and extraordinary specific recognition ability
towards their complementary antigen in biological systems. Immunoassays,
409
Absorption Adsorption Adsorption
(large pores) (small pores)
Fig. 13.18. Schematic representation of absorptive vs. adsorptive extraction, and adsorption in
small vs. large pores.
410
Originalsampleconsistingof myoglobin Proteinsadsorbed onto and
ontothen
(M),cytochrore C (C), and ysozyme (L), releasedfrom a po yacrylc acid
coatedfiber L
C
L
L C
Fig. 13.19. Protein extraction using poly(acrylic) acid coated fibre. Demonstration of myoglobin
and cytochrome C displacement by lysozyme with time.
stronger affinity for the coating, replaces the other two compounds during
extraction. This effect is associated with the fact that there is only limited
surface area available for adsorption. If this area is substantially occupied then
the displacement effects occur [46,47] and the equilibrium amount extracted can
vary with concentrations of both the target and other analytes. In extraction of
analytes with liquid coatings on the other hand, partitioning between the sample
matrix and extraction phase occurs. In this case, equilibrium extraction
amounts vary only if the coating property is modified by the extracted compo-
nents, which only occurs when the amount extracted is a substantial portion (a
few percent) of the extraction phase. This is very rarely observed since SPME is
typically used to determine trace contamination samples.
One way to overcome this fundamental limitation of the porous coatings is,
as the above figure suggests, to use an extraction time much less than the equi-
librium time, so that the total amount of analytes accumulated onto the fibre is
substantially below the saturation value. When performing such experiments,
not only is it critical to precisely control extraction times, but also convection
conditions must be monitored to ensure that they are constant or can be com-
pensated for. One way of eliminating the need for compensation of convection is
to normalize (use the same) agitation conditions. For example, the use of stirring
means at well defined rotation rates in the laboratory or fans for field air moni-
toring will ensure consistent convection.
The short time exposure SPME measurement described has an advantage
associated with the fact that the rate of extraction is defined by diffusivity of
analytes through the boundary layer of the sample matrix, and their
corresponding diffusion coefficients, rather than distribution constants. This is
demonstrated in Fig. 13.20.
The amouNt of analytes accumulated on the fibre as a function of time for
fibre geometry can be estimated as:
n= 2DLCt (13.6)
ln((b +6)/ b)
411
benzene y= 546.7x -1188, 10 = 0.9964
2=
,.l, ,, v = 514.7x
+ 1. R 09986
3 E+04-
Diffusion Coefficients
E
benzene: 0.088
2.E+04
-
toluene: 0.084
0E+00
0 10 20 30 40 50 60 70
Sampling time (s)
Fig. 13.20. Correlation of uptake rate with diffusion coefficient, for short sampling time
(non-equilibrium) extraction of VOCs by PDMS-DVB fibre.
where n is the mass of extracted analyte over sampling time (t) in ng; D, is the
analyte molecular diffusion coefficient in sample matrix (cm2/s); b is the outside
radius of the fibre coating (cm); L is the length of the coated rod (cm); 6 is the
thickness of the boundary layer surrounding the fibre coating (cm); and C is
analyte concentration in sample matrix (ng/ml).
The differences in diffusion coefficients between compounds are small
compared to the differences in distribution constants. This makes it easier to
calibrate the system. Because of the large differences in distribution constants
between analytes, the resulting chromatograms are characterized by small peak
areas for compounds with small distribution constants, and large areas for those
with large constants. With uptake dependent on diffusion coefficients, all
compounds in a chromatogram with similar molecular masses will have similar
peak areas, given similar detector responses. Also, it is relatively simple to
calculate the diffusion coefficients for a given analyte and therefore correct for
the small differences in it. It must be understood that this system is only suitable
for trace analysis. When sample concentrations become too high, saturation of
the active sites occurs, and uptake rates are no longer linear. Shorter exposure
times, where smaller amounts are extracted, can solve this problem. Also, at
these higher concentrations, samples are easily extracted and analyzed with
PDMS fibre, using conventional SPME extraction methods. Results of extrac-
tion by the diffusion type approach are shown in Fig. 13.21. Accumulation of
volatile components on the solid coating in 10 s is much larger compared to the
10 min equilibrium extraction on PDMS. This approach to extraction is not
limited to devices using the fibre geometry, but is generally applicable.
At this point it might be instructive to draw parallels between developments
and applications of SPME and electrochemical methods. The coulometric tech-
nique corresponds to the total extraction method. Although the most precise,
this technique is not often used because of time required to complete it. SPME is
capable of producing exhaustive extraction when the volume of the extraction
phase is large enough combined together with high distribution constants. In
fibre geometry, the larger volume translates into thicker coatings which results
412
A -PDMS - 10min, B - PDMS/DVB -10 s
0.8
A
1 2 netn
. .. I ----
...
2. hexane
3 heptane
4 octane
5 nonane
6. decade
7 undecane
8. dodecane
7 8
-11n
Fig. 13.21. Comparison of VOC mass loading I I
between PDMS and PDMS/DVB. 1 2 3 4 5 6 (min)
in long extraction times. The alternative approach is to disperse the whole vol-
ume of the extraction phase onto a larger surface area, resulting in a thinner
coating and faster equilibration times. For example, solid support material may
include particulate matter, a stirring mechanism or the vessels walls (see Fig.
13.1). In this case, however, there would be more handling required to conve-
niently introduce the extraction phase into the sample introduction system (GC
or HPLC). It might necessitate the use of organic solvent to desorb the analytes
from the extraction phase. Equilibrium potentiometric techniques are more fre-
quently used (pH electrode), particularly in cases where the sample is a simple
mixture and/or selectivity of the membrane is sufficient to quantify target
analyte in complex matrices. The equilibrium SPME method has some advan-
tages in this regard, since the technique is typically coupled with separation
and/or mass spectrometry detection methods, which allows identification and
quantification of many components simultaneously. The advantage of electro-
chemical methods is response time because of low capacities of electrodes.
Some electrochemical methods, like amperometry, are based on mass trans-
port through the boundary layers as in pre-equilibrium SPME. Analogously in
SPME, calibration based on diffusion coefficients can be accomplished as long as
the agitation conditions are constant and the extraction times are short and
coating has high affinity towards the analytes. Figure 13.21 illustrates the
results obtained for the 10 s extraction times using solid coating. In some
implementations of the technology the rate of mass transfer to the extraction
phase can be purposely restricted by placing it in the needle and therefore
achieving the time-weighted average measurements of concentration in a
specific time period (Fig. 13.16).
The compounds adsorbed onto the Carboxen coating are strongly bound.
This fact could be used to prepare solvent-free standards. Table 13.2 shows that
413
TABLE 13.2
Effect of fibre exposure time on the amount of BTEX remaining on the 75-gAm thick carboxen
coated fibre
Compounds Amount of BTEX (%) remaining on the fibre";
2 min 3 min 4 min 5 min 6 min
extraction extraction extraction extraction extraction
2 min 3 min 4 min 5 min 6 min
exposure exposure exposure exposure exposure
Benzene 98.6 99.0 97.7 96.1 89.9
Toluene 99.5 99.2 100.3 101.2 93.6
Ethylbenzene 100.7 99.1 100.9 97.7 94.5
p-Xylene 99.8 100.3 103.1 95.7 95.2
o-Xylene 100.6 99.9 99.2 95.5 93.7
1,3-Dichlorobenzene 99.3 100.8 100.1 98.6 97.2
1,1,2-Trichloroethane 98.8 99.7 98.1 96.7 93.5
Tetrachlloro-ethylene 100.1 101.3 98.5 97.2 94.1
*'Compared with the corresponding data obtained by 2-min, 3-min, 4-min, 5-min and 6-min
extraction without exposure to the air, respectively.
Fibrepreparation
There are several different methods of depositing coatings onto fibres. The
dipping technique typically consists of placing a fibre in a concentrated organic
solvent solution of the material to be deposited for a short time. After removal of
the fibre from the solution, the solvent is evaporated by drying and the deposited
material can be crosslinked [2]. An extension of this method is electrodeposition,
which can be used to deposit selective coatings on the surface of metallic rods
[31]. The limitation of this approach is coating thickness variance from fibre to
414
fibre. Therefore, the preparation of films for commercial devices is carried out
simultaneously during the drawing of the fused silica rod. While this process
requires dedicated and very expensive equipment, reproducibility of the coating
thickness is excellent. This process is identical to the preparation of optical
fibres [49], and so the equipment needed is commercially available. Similar
results can be obtained using a piece of hollow fibre membrane (small i.d.
tubing), made from the desired extraction material. Preparation consists of
swelling the membrane by means of an appropriate volatile solvent, placing the
enlarged membrane onto the tip of the fibre, and evaporating the solvent. The
thickness of the coating is determined by the membrane thickness. Therefore
the volume of the coating can be large, reaching to 3 gl for a 300 m thick PDMS
hollow fibre membrane. A porous hollow fibre membrane can also be used for
adsorption of target analytes, or its pores can be filled with organic solvent to
allow for solvent microextraction [50].
x 3.E-03
I :A benzene
I1 a toluene
'
E 2.E-03 + p-xylene
o ethylbenzene i
N
E 1.E-03
0 P f
I I
* r U
P
0.E+00
Fig. 13.22. Effect of wind speed on 0 30 60 90
non-equilibrium extraction of
Air velocity (cm/s)
VOCs (exposed fibre).
415
Fig. 13.23. Constant agitation device for field air sampling by SPME.
increases from 0 to 5 cm/s. No further change was observed as the wind speed
further was increased from 5 to 20 cm/s. This indicates that the thickness of the
boundary layer between the fibre and air is diminished as the wind speed
increases, which results in an increase of the mass transfer rate and the mass
loading on the fibre. After the thickness of the boundary layer has been
decreased to certain degree, uptake becomes limited by diffusion within the
pores of the polymer, and no further increase in extraction is seen with
increasing wind speed. In practice, when using SPME to sample in the field, it
would be helpful to use a fan to move air samples across the fibre during
sampling, to eliminate possible imprecision in extraction due to variations in
wind speed. Figure 13.23 shows an example of an agitation device for field air
sampling, consisting of a modified hair-dryer fan with a mounting for the SPME
device. Appropriate devices for water sampling can be designed as well.
Control of salt concentration and sample pH can be used to enhance extrac-
tion; the principle is similar to that for solvent extraction procedures. An appro-
priate salt or buffer is added directly to the sample without the need for a specific
instrument. The other parameter that is very important to optimize is extrac-
tion temperature. At elevated temperatures native analytes can effectively dis-
sociate from the matrix and move into the headspace for rapid extraction by the
fibre coatings. However, the coating/sample distribution coefficient also
decreases with an increase of temperature, resulting in a diminution in the equi-
librium amount of analyte extracted, as described previously. To prevent loss of
sensitivity, the coating can be cooled simultaneously with sample heating. This
idea was implemented in the design of the device shown in Fig. 13.24. In this
device, a fused silica tubing is sealed and coated at one end (outer surface of the
capillary). Liquid carbon dioxide is delivered via the inner capillary to the coated
end of the outer capillary resulting in a coating temperature lower than that of
the sample. This "cold finger" effect results in accumulation of the volatilized
416
inner capillary
C02in X C0o2 out
-outer capillary
- syringe barrel
ferrule
.<~needle
headspaceoating
sample vial
Fig. 13.24. Design of internally cooled SPME device. heater
analytes at the tip of the fibre. There is additional enhancement in the sample
matrix-fibre coating distribution constant associated with the temperature gap
present in the system which can be given by the following equation:
417
TABLE 13.3
Quantitative extraction of BTEX from different matrices
Analytes %Extracted
Water (80°C) Soil (15% water) Sand (110°C) Clay (5% water)
(120°C) (170°C)
_
Benzene 50 48 84 64
Toluene 65 101 105 84
Ethylbenzene 85 96 106 91
m,p-Xylene 90 99 102 95
o-Xylene 100 104 100 100
_ , -
ExtractIon cell r
I i
Fig. 13.25. Hot water dynamic extraction system combined with SPME for analysis of solid
samples.
Fig. 13.26. Static hot water-SPME system.
418
extracted analytes are desorbed into the analytical instrument and typically
quantified, based on an isotopically labelled standard spike (introduced to the
sample prior to extraction).
13.3.5 Derivatization
The main challenge in organic analysis is polar compounds: they are difficult to
extract from environmental and biological matrices and difficult to separate on
the chromatographic column. Derivatization approaches are frequently used to
address this challenge. Figure 13.27 summarizes various derivatization tech-
niques that can be implemented in combination with SPME [55]. Some of the
techniques, such as direct derivatization in the sample matrix, are analogous to
well-established approaches used in solvent extraction. In the direct technique,
the derivatizing agent is first added to the vial containing the sample. The
derivatives are then extracted by SPME and introduced into the analytical
instrument (Fig. 13.28). For example, this approach has been applied to extract
and separate phenols from aqueous samples by first converting the target
analytes to their acetate derivatives [56].
Because of the availability of polar coatings, extraction efficiency for polar
underivatized compounds is frequently sufficient to reach the sensitivity
required. Occasionally, however, there are problems associated with the separa-
Derivatization in
SPME Fiber Coating
Simultaneous Derivatization
derivatization following
and extraction extraction
419
(in coating post-extraction derivatization)
4
3
a
2
c
,- 0
S 8 palmitic acid ataric acid
F 6 b methyl ester methyl ester
.2)
2 0 n . a~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i
0
10 11 12 13 14 15 16
Time (min.)
Fig. 13.29. Separation of derivatized (b) and underivatized (a) long chain carboxylic acids.
420
a 0
IAa r- A l
W DDDM1DPt w
DDmjD
it a
Fg__J
Doping the SPME Placing the doped fiber into Fiber desorption,
fiber with the gaseous phase or headspace above separation, and
derivatizing reagent aqueous phase in reaction vial for quantitation
in-fiber derivatization/SPME
pie. Volatilities of the pyrenylmethyl esters formed during the reaction also are
low, resulting in the accumulation of the product onto the fibre until analyte or
reagent is exhausted or decomposed. At a high injector temperature the deriva-
tized analytes are removed from the coating and the fibre can be reused.
A similar approach is used for the analysis of formaldehyde from gaseous
samples [59]. This approach is not an equilibrium extraction, but it is based on
the fast reaction. Overall extraction rate is therefore controlled by mass trans-
port, analogous to the non-equilibrium extraction of benzene, toluene and
p-xylene using porous coatings, described earlier (Fig. 13.20). The derivatizing
agent, o-(2,3,4,5,6-pentafluorobenzyl)hydroxylamine is first doped onto the fibre
by room temperature headspace extraction from an aqueous solution. Formalde-
hyde is subsequently extracted from an unknown sample, and converted to the
oxime derivative in the fibre. In this case the kinetics of the derivatizing reaction
are fast, and uptake is controlled by mass transport. If reaction kinetics were
slow, uptake may be controlled by both the distribution constant and the reac-
tion kinetics [51]. With fast derivatization reaction kinetics, the process is simi-
lar to the mass transport controlled extraction described above for VOCs
analysis from air, where uptake rates for short sampling times are proportional
to diffusion coefficients. The significant difference between the two approaches
is that the formaldehyde analysis approach has a much higher capacity.
This simple, but powerful procedure as described above for this analysis, is
limited to low volatility reagents. The approach can be made more general by
chemically attaching the reagent directly to the coating. The chemically bound
product can then be released from the coating either by high temperature in the
injector, light illumination or change of applied potential etc. The feasibility of
this approach was recently demonstrated by synthetizing standards bonded to
silica gel, which were released during heating. This approach allowed solvent-
free calibration of the instrument [60]. The other interesting option is to use an
SPME device similar to the one shown in Fig. 13.24, but with the derivatization
reagent, instead of carbon dioxide, delivered to the coating.
In addition to using a chemical reagent, electrons can be supplied to produce
redox processes in the coating and convert analytes to more favourable deriva-
tives. In this application, the rod as well as the polymeric film must have good
421
C
4
Fig. 13.31. SPME-electrochemistry minicell. 1, Reference electrode;
2, 100-ml plastic syringe; 3, Teflon capillary; 4, 8-ml Teflon vial; 5,
SPME fibre working electrode (WE); 6, platinum wire counter
electrode (CE).
Because of its solvent-free nature and the small size of the fibre or capillary,
SPME can be interfaced conveniently to analytical instruments of various types.
Only extracted analytes are introduced into the instrument, since the extracting
phase is non-volatile. Thus there is no need for complex injectors designed to
deal with large amounts of solvent vapour, and these components can be
simplified for use with SPME. The sensitivity of determinations using the SPME
technique is very high, facilitating trace analysis. Although in most cases the
entire complement of analytes is not extracted from the sample, all extracted
material is transferred to the analytical instrument, resulting in good perform-
ance. Also, the solvent-free process results in narrow bands reaching the
instrument, giving taller, narrower peaks and better quantification.
SPME-GC interface
The analytical instrument used most frequently with SPME has been the gas
chromatograph. Standard GC injectors, such as split/splitless can be applied to
SPME as long as a narrow insert with an inside diameter close to the outside
diameter of the needle is used. The narrow inserts are required to increase the
linear flow around the fibre, resulting in efficient removal of desorbed analytes.
The split should be turned off during SPME injection. Under these conditions,
the desorption of analytes from the fibre is very rapid, not only because the
coatings are thin but because the high injector temperatures produce a dramatic
decrease in the coating/gas distribution constant and an increase in the diffusion
coefficients. The speed of desorption in many cases is limited by the time
422
carrier -
supply
upply
0 6 12 18
Time (s)
Left: Fig. 13.32. Schematic diagram of the flash SPME injector. 1, Injector body; 2, washer; 3,
septum; 4, nut; 5, needle guide; 6, 0.53 mm i.d. fused silica capillary; 7, nut; 8, ferrule; 9, heater;
10, butt connector; 11, relay; 12, capacitor; 13, switch.
Right: Fig. 13.33. Rapid analysis of BTEX in water by SPME-flash injector-GC. 1, Benzene; 2,
toluene; 3, ethylbenzene; 4, m,p-xylenes; 5, o-xylenes.
423
Leads connected to capacitive
discharge supply
oblong openingmm
in the plunger x,.32 mm I.D
fused silica 0.057 mm O.D.
capillary / heating wire
silicone rubber
seal ,
nut capillary
needle .
GC Oven
5
To lon-Trap MS
Fig. 13.35. Direct capacitive discharge desorption system. 1, SPME syringe; 2, electric
connection I; 3, injector body; 4, steel wire; 5, gold coating; 6, electric connection II; 7, transfer
line; 8, capacitor; 9, relay; 10, butt connector.
424
Fig. 13.36. High sensitivity of toluene detection by
directly hyphenating SPME device to an ion trap mass o
spectrometer. Time (min)
SPME-HPLCinterface
Research has also been focused on designing interfaces for liquid phase separa-
tion techniques to address the need for analysis of nonvolatile and thermally
labile analytes. The interface to high performance liquid chromatography
(HPLC) can be a straightforward analogue of the traditional loop injection sys-
tem. A typical SPME-HPLC interface consists of a custom-made desorption
chamber and a six-port injection valve (Fig. 13.37) [64]. The upper part of the
PEEK tubing (c), fitted into a tee-union, is enlarged to fit the needle of the
syringe. The internal tubing of the SPME device, which holds the fibre, can be
sealed by the PEEK tubing and the tee-union tightly enough to withstand sol-
vent pressures as high as 4500 psi. The desorption chamber is placed in the posi-
tion at which the injection loop normally resides on the injection valve. When the
injection valve is in the "load" position, it allows the fibre to be introduced into
the desorption chamber under ambient pressure. It also allows for the introduc-
tion of a desorption solvent if different from the mobile phase. The valve is then
switched to "inject" to transfer the desorbed analytes to the column. A heater
425
Desorpt
chamber
¢tPcim
cbU-
Frame
Enlarged
Interface
SPME
Device
UV-VIS detector
Fig. 13.37. SPME-HPLC interface: (a) stainless steel (SS) 1/16 in tee; (b) 1/16 in SS tubing; (c)
1/16 in PEEK tubing (0.02 in i.d.); (d) two-piece, finger-tight PEEK union; (e) PEEK tubing
(0.005 in i.d.) with a one-piece PEEK union.
can be installed in the device to facilitate the desorption process. The interface
performs well, and its desorption volume is similar to the volume of the typical
injection loop. The use of small-volume desorption chambers results in very effi-
cient supercritical fluid chromatography separation (see Fig. 13.38) with very
narrow bore capillary columns [65].
In-tube SPME-HPLC
There is a significant need to automate SPME-HPLC analyses, and this is
addressed conveniently through the development of in-tube SPME using
internally coated tubing [66]. In this method, a section of coated fused silica GC
capillary is placed between the needle of an HPLC autosampler and the injection
valve. The capillary sections selected have coatings similar to common commer-
cially available SPME fibres. Figure 13.15 illustrates the system based on a
modified Spark Holland micro LC autosampler [67]. In this system the analytes
are extracted first into the coating by passing sample through the tubing,
followed by desorption of the compounds using a small volume of solvent or
mobile phase. These approaches, which are suitable for analysis of very small
samples, also offer convenient interfacing to micro HPLC instrumentation.
The theory of extraction and desorption by this method was investigated and
validated with the automated analysis of six phenylurea pesticides. For extrac-
tion, sample is aspirated and dispensed from the capillary a number of times,
preferably until an equilibrium is nearly achieved. It was expected that the
426
3
427
gradient composition, and MS fragmentor voltage. Both standard and capillary
LC formats were employed. For capillary LC detection limits were reduced by a
factor of 10. The effect of sample alcohol content was determined, and the
method was applied to the analysis of spiked water (12% ethanol) and wine
samples.
Automated in-tube solid-phase microextraction (SPME) coupled with liquid
chromatography/electrosprayionisation mass spectrometry (LC-ESI-MS) using
the HP1100 system, was evaluated for the analysis of 5-blockers [42], ranitidine
[69] and amphetamines [70] in biological and pharmaceutical samples. Extrac-
tions required 10-15 min per sample. For quantification, LC-MS analyses were
initially performed by liquid injection onto the LC column with analysis in
selected ion mode. Several in-tube SPME parameters were optimized. These
included capillary column stationary phase selection, sample pH, extraction flow
rate and number and volume of aspirate/dispense steps. The compounds
extracted by the capillary column were easily desorbed by either mobile phase
flow or a small quantity of methanol. Carryover was not observed.
A serum sample from a patient administrated propranolol was analyzed
using this method and both propranolol and its metabolites were detected. This
was a significant finding as metabolites are typically more challenging to
analyze, due to their higher polarity. For ranitidine analysis both tablet and
urine samples were analyzed with a detection limit at S/N = 3 of 1.4 ng/ml. and
within-day and between-day variations in ranitidine analysis were 2.5 and 6.2%
(n = 5), respectively. For the analysis of amphetamine, methamphetamine, and
their 3,4-methylenedioxy derivatives in urine samples detection limits (S/N = 3)
of 0.38-0.82 ng/ml were observed.
Mutagenic heterocyclic amines were also analyzed by this technique [71].
The amines extracted by the capillary column were desorbed by aspiration of 30
/b1 of methanol prior to injection, and carryover of heterocyclic amine was not
observed. The detection limits (S/N = 3) were 0.2-3.1 ng ml- l and IQ, MeIQx and
PhIP concentrations in grilled beefsteak were determined.
Other interfaces
Smaller injection volumes for applications of SPME to micro-HPLC and capil-
lary electrophoresis can be accomplished by modifying microinjector designs;
the sliding injector developed for capillary isotachophoresis is one example [72].
The other approach is to design an appropriate sample introduction system
based on guides, to introduce the fibre with the extracted components directly
into the capillary (see Fig. 13.39). Using this method very efficient separation
was achieved (Fig. 13.40) [73].
SPME can be directly combined to optical detection, based on reflectometric
interference spectrometry [74-76]. A light beam passing through an optically
transparent fibre coated with transparent sorbing material interacts with ab-
sorbed substances through internal reflection. Therefore, if any of the extracted
analytes strongly absorb the transmitted light, there is a loss in intensity that
428
a Capillary electropherograph 3
ler 7
0.001 AU 4
By ~~~~~~~1
l
To detector - f B
Separation capillary 0 5 10 15 20
(75 nmId. x 360pm o.)/ S fiberassembly Time (min)
Buffer reservoir Pt electrode
Left: Fig. 13.39. Schematic of the SPME-CE system (a) and interface (b).
Right: Fig. 13.40. Electropherogram of ten phenolic compounds obtained by SPME-CE using (A)
a 40 gLm o.d. PA-coated silica fibre and (B) a 40 gAm o.d. bare silica fibre. Peak identities: 1:
2,4-dimethylphenol, 2: phenol, 3: 4-chloro-3-methylphenol, 4: pentachlorophenol, 5:
2,4,6-trichlorophenol, 6: 2-methyl-4,6-dinitrophenol, 7: 2,4-dichlorophenol, 8: 2-chlorophenol, 9:
4-nitrophenol, 10: 2-nitrophenol.
can be detected with a simple optical sensor. These devices demonstrate poor
sensitivity primarily because it is difficult to find light wavelengths that are spe-
cifically adsorbed by the analytes and not by the coating or interferences. In an
alternative design, the light can be passed directly through the absorbing poly-
mer which is then cooled to facilitate high sensitivity of determination [77]. Flu-
orescence can be used to detect analytes in the coating. The selectivity of the
extraction process and spectroscopy can be combined with selectivity of the elec-
trochemical process resulting in a spectroelectrochemical sensor [78].
Fibre assemblies: The first commercial version of the laboratory SPME device
was introduced by Supelco in 1993 (see Fig. 13.4). The device is similar in
operation principle to the custom-made device shown in Figs. 13.2 and 13.3a.
Some additional improvements include the adjustment for depth of the fibre
with respect to the end of the needle, which allows control of the exposure depth
in the injector and extraction vessel. The device incorporates such useful
features as colour marking of the fibre assemblies to distinguish among various
coating types. In addition to standard PDMS coatings of various thickness and
polyacrylate (PA), Supelco developed new mixed phases based on solid/liquid
sorption, such as Carbowax-DVB and PDMS-DVB. Supelco also introduced an
HPLC interface (Fig. 13.41) that integrates the original concept with the injec-
tion valve (Fig. 13.37).
429
SPMEfiberholder
Septumpiercingneedle
le I
Needleguide
Doubletapered
ferrule ression union
Comp
SPMEfiber_
Solvent
desorption
chamber From To
-r interface interface
r-' Solvent - Solvent
from from
To syringe From synnge Vdte
I~Slvn interface
VWaste
for programming SPME analyses. Samples are loaded onto trays accommodat-
ing five vial sizes, and samples are heated and agitated during extraction, using a
separate sample preparation chamber. To facilitate agitation, the extraction
chamber is rotated at a programmable rotation speed during extraction. Addi-
tional vials/stations are present to accommodate wash solutions, derivatizing
agents, temperature control, derivatization and fibre conditioning, to facilitate
operation of SPME at optimal conditions. While the built-in software can be
used to perform basic SPME analyses, extra programming flexibility is provided
by the Cycle Composern software, available as an accessory. We expect this new
instrument with its greatly enhanced flexibility, will significantly expand the
430
43 1 6 7 10
range of SPME methods amenable to automation. New coatings and devices are
expected to follow, as interest in SPME grows along with the unprecedented
numbers of new applications appearing in the literature.
Field portable SPME-FastGC: Solid phase microextraction coupled to high
speed gas-chromatography is a good combination to perform rapid and cost-
effective investigations in the field, even of complex organic samples. As dis-
cussed above, SPME is particularly suited for fast GC, as it is solvent-free, and
the thin coatings can provide very fast desorption of analytes at high tempera-
tures. Some instrumental modifications were performed recently in order to
achieve successful fast separations [62].
A portable system was optimized for SPME-fast GC field investigations and
was commercialized by SRI Instruments (model 8610C, SRI Instruments,
Torrance, CA). The instrument was tested in combination with a flame ioniza-
tion detector (FID), a photoionization detector (PID), and a dry electrolytic
conductivity detector (DELCD). A dedicated injector, presented in Fig. 13.42,
was mounted on the portable system in order to use SPME for high-speed
separation. The injector guarantees very fast thermal desorption of the analytes
from the SPME fibre [80].
The injector for high-speed GC should produce as narrow an injection band
as possible. Internal volumes of regular injector ports are too large (e.g.,
split/splitless injector), since they have been designed to accommodate large
volumes of gaseous samples or vapours produced by solvent injection. Thermal
focusing for separation improvement is not convenient for fast separations, since
temperature programming is impractical for high-speed GC. Hence, an injector
port with a small internal volume was required for this application. Also, very
fast thermal desorption from the SPME fibre was required to produce a narrow
injection band and achieve effective separation. In the dedicated injector for
SPME-fast GC, the injector port was maintained cold during needle introduction
and was rapidly heated only when the fibre was exposed to the carrier gas
stream. The desorption area of the injector was heated by capacitive discharge
that allowed heating rates as fast as 4000°C/s, and very narrow injection bands
were observed, as required by fast GC.
Fibre conditioners: SRI has also has a dedicated fibre injector available for
their systems, and CTC Analytics has introduced a fibre cleaning station for
their SPME autosampler. New SPME fibres require initial conditioning at
manufacturer-recommended temperatures ranging from 210-320°C, for time
periods ranging from 0.5-4 h. Conditioning is also recommended at the begin-
ning of the work day, and between runs in the case where analyte carry-over is a
possibility. In addition, many SPME-fast GC applications, such as field
431
iA) C'··li--·· , nh"· .......
LCcolimnll
S slt,
co.,iolr
aid daa anolysis
ad a1tllyl
OtsisfO
432
Small custom modifications of commercial devices can lead to new possibili-
ties. For example, the replacement of part of the loop in an Agilent 1100 instru-
ment results in automation of the sample preparation and introduction into an
LC instrument (Fig. 13.43). This in-tube SPME system was applied to the analy-
sis of drugs and pesticides as well as their degradation products in real samples
such as waste water, blood plasma and cell culture extracts with minimum sam-
ple processing prior to placement into the autosampler. These ideas are explored
in Chapter 20-29.
The addition of input and output connections to the GC autosampler vial
allows the system to be used to continuously monitor flowing streams, as shown
in Fig. 13.44. The flow-through design facilitates agitation of the sample [83].
Alternatively, when a connection is added directly to the needle of the auto-
sampler syringe, as in Fig. 13.13b, the system can analyze samples present in the
vial without the need to expose the fibre [4]. This modified approach, which
would facilitate simplified automated sampling, can be designed to rely on air
pressure to push the sample through the needle. Then the fibre containing the
extracted analytes in its coating is introduced to the instrument for desorption.
The in-needle concept can be expanded to the trap in-needle system [84].
Example of such design is illustrated in Fig. 13.45a. In this embodiment the side
hole domed tip needle was used for convenience of packing the needles. This
design is able to perform conveniently sampling, sample preparation and sample
introduction to analytical instrument. The device can be used as active sampler,
by drawing the gas or liquid sample through the needle using a gas tight syringe
attached to the needle. This can be performed in exhaustive mode, by careful
packing of the trap and drawing limited amount of the sample to prevent break-
through or it can be more convenient equilibrium or diffusion based as discussed
above. This device can also perform well as passive sampler when the needle is
exposed to the investigated system directly allowing components of the sample
a
Trap Gas tight syringe
Z a
;-Yr~~~~~
T~~1
autosampler carousel
Left: Fig. 13.44. On line monitoring of flowing streams by Varian autosampler using the
modified vial design.
Right: Fig. 13.45. Alternative embodiments of SPME: (a) In-needle trap and (b) membrane
SPME.
433
~~
a.~~ _·
434
726.4
100 .
5 pmol/uL
so
1347.7
" j4
200 400 600 800 1000
. . .
1200 1400 1600 1800 2000 2200 2400
m/z, amu
2600
.I- . .
2800 3000 3200
Fig. 13.47. Mass spectrum of enkephalin (m/e = 726.4) and substance P (m/e = 1347.7) formed by
SPME/MALDI.
435
challenges to develop successful methods are analogous. The nature of target
analytes and complexity of sample matrix determine the level of difficulties in
accomplishing a successful extraction. The simplicity, speed and convenience of
the extraction devices primarily impacts the costs of practical implementation
and automation of the developed methods.
The potential savings in analysis time, reduced solvent use and apparent
simplicity of SPME techniques will continue to attract interest among analytical
chemists searching for improved analysis methods. So long as analysts have a
sound understanding of the theory and principles behind this technique, good
accuracy and precision will follow. This part of the chapter describes method
development and optimization.
13.5.1.1 Thermodynamics
Solid phase microextraction is a multiphase equilibration process. Frequently,
the extraction system is complex, as in a sample consisting of an aqueous phase
with suspended solid particles having various adsorption interactions with
analytes, plus a gaseous headspace. In some cases specific factors have to be
considered, such as analyte losses by biodegradation or adsorption on the walls
of the sampling vessel. In the discussion below we will only consider three
phases: the fibre coating, the gas phase or headspace, and a homogeneous matrix
such as pure water or air. During extraction, analytes migrate between all three
phases until equilibrium is reached.
The mass of an analyte extracted by the liquid polymeric coating is related to
the overall equilibrium of the analyte in the three-phase system. Since the total
mass of an analyte should remain constant during the extraction, we have:
n= KfhKhVCOV (13.9)
K, Khs Vf +Kh, V + Ve
436
stant, Khs, by the gas/sample distribution constant, Kgs, if the effect of moisture
in the gaseous headspace can be neglected. Thus, Eq. (13.2) can be rewritten as:
n Kf VfCOVs (13.11)
Kf V + Ks Vh +Vs
The equation states, as expected from the equilibrium conditions, that the
amount of analyte extracted is independent of the location of the fibre in the
system. It may be placed in the headspace or directly in the sample as long as the
volumes of the fibre coating, headspace, and sample are kept constant. There are
three terms in the denominator of Eq. (13.11) which give measures of the analyte
capacity of each of the three phases: fibre (KfsVf), headspace (KhSVh), and the
sample itself (Vs). Therefore, it is always important to optimize volumes of all the
phases present in the system. If we assume that the vial containing the sample is
completely filled (no headspace), the term KhSVh in the denominator, which is
related to the capacity (CVh) of the headspace, can be eliminated, resulting in
Eq. (13.1).
Equation (13.1) describes the mass absorbed by the polymeric coating after
equilibrium has been reached in the system. In most of determinations, Kfs is
relatively small compared to the phase ratio of sample matrix to coating volume
(Vf < < V). In that situation the capacity of the sample is much larger compared
to capacity of the fibre resulting in a very simple relationship defined by Eq.
(13.2), which states that the amount extracted at equilibrium is proportional
only to the distribution constant, volume of the fibre coating and the concentra-
tion.
Equation (13.2) emphasizes the field sampling capability of the SPME
technique. It is not necessary to perform a separate sampling procedure where a
well defined volume of the matrix is obtained since the amount of analyte
extracted is independent of V, as long as KSVf < Vs. The SPME device can be
placed directly in contact with the investigated system to allow quantitation.
Strictly speaking, the above discussion is limited to partitioning equilibrium
involving liquid polymeric phases such as poly(dimethylsiloxane). The method of
analysis for solid sorbent coatings is analogous for low analyte concentration,
since the total surface area available for adsorption is proportional to the coating
volume if we assume constant porosity of the sorbent. For high analyte concen-
tration the saturation of the surface can occur resulting in nonlinear isotherms.
Similarly high concentration of competitive interference compound can displace
the target analyte from the surface of the sorbent. The simplest way to consider
these high concentration effects is to replace the volume of the fibre coating, Vf in
the above equations as a measure of the total fibre surface area by a fraction of
the original coating volume corresponding to a free surface area available for
adsorption.
The way to overcome these effects is to use diffusion based sampling by
exposing fibre to the sample for a short period of time. The amount of analytes
437
collected on the coating is given by Eq. (13.7), which also does not contain
volume of sample. The extraction mass is dependent only on the diffusion
coefficient of given analyte in the sample matrix and agitation conditions, which
needs to be constant during sampling. This approach is therefore also suitable
for field sampling.
Predictionof distributionconstants
In many cases, the distribution constants present in equations above which
determine the sensitivity of SPME extraction can be estimated from physico-
chemical data and chromatographic parameters. This approach eliminates need
for calibration. For example, distribution constants between a fibre coating and
gaseous matrix (e.g., air) can be estimated using isothermal GC retention times
on a column with stationary phase identical to the fibre coating material. This is
possible because the partitioning process in gas chromatography is analogous to
the partitioning process in solid phase microextraction, and there is a well-
defined relationship between the distribution constant and the retention time.
The nature of the gaseous phase does not affect the distribution constant, unless
the components of the gas, such as moisture, swell the polymer, thus changing
its properties. A most useful method for determining coating-to-gas distribution
constants uses the linear temperature programmed retention index (LTPRI)
system, which indexes compounds' retention times relative to the retention
times of n-alkanes. This system is applicable to retention times for tempera-
ture-programmed gas-liquid chromatography. The logarithm of the coating-to-
air distribution constants of n-alkanes can be expressed as a linear function of
their LTPRI values [87]. The LTPRI values for many compounds are available in
the literature, hence this method allows estimation of Kfg values without experi-
mentation. If the LTPRI value for a compound is not available from published
sources, it can be determined from a GC run. Note that the GC column used to
determine LTPRI should be coated with the same material as the fibre coating.
More discussion on this topic is conducted in Section 13.6 on applications.
Estimation of the coating/water distribution constant can be performed
using Eq. (13.10) [88]. The appropriate coating/gas distribution constant can be
found by applying techniques discussed above, and the gas/water distribution
constant (Henry's constant) can be obtained from physicochemical tables or can
be estimated by the structural unit contribution method [89].
Some correlations can be used to anticipate trends in SPME coating/water
distribution constants for analytes. For example, a number of investigators have
reported the correlation between octanol/water distribution constant Kow and
Kf,. This is expected, since Ko, is a very general measure of the affinity of
compounds to the organic phase. It should be remembered, however, that the
trends are valid only for compounds within homologous series, such as aliphatic
hydrocarbons, aromatic hydrocarbons or phenols; they should not be used to
make comparisons between different classes of compounds, because of different
analyte activity coefficients in the polymer.
438
Effect of extractionparameters
Thermodynamics theory predicts the effects of modifying certain extraction
conditions on partitioning and indicates parameters to control for reprodu-
cibility. The theory can be used to optimize the extraction conditions with a
minimum number of experiments and to correct for variations in extraction
conditions, without the need to repeat calibration tests under the new condi-
tions. For example, SPME analysis of outdoor air may be done at ambient
temperatures that can vary significantly. A relationship that predicts the effect
of temperature on the amount of analyte extracted allows calibration without
the need for extensive experimentation. Extraction conditions that affect K,
include temperature, salting, pH, and organic solvent content in water. A brief
discussion about the use of extraction parameters in SPME method optimization
can be found the next section.
13.5.1.2 Kinetics
The kinetic theory is very useful in optimization of the extraction conditions by
identifying "bottlenecks" of solid phase microextraction and indicates strategies
to increase extraction speed. In the discussion below we will limit our consider-
ation to direct extraction (Fig. 13.48).
Perfect agitation:Let us first consider the case where the liquid or gaseous
sample is perfectly agitated. In other words, the sample phase moves very rap-
idly with respect to the fibre, so that all the analytes present in the sample have
access to the fibre coating. In this case, the equilibration time, defined as the
time required to extract 95% of the equilibrium amount of an analyte from the
sample, corresponds to:
439
Fig. 13.49. Mass absorbed versus time from 0 1 2 3 4 5
agitated solution of infinite volume. D t / SKf(b-a)
Using this equation, one can estimate the shortest equilibration time possible
for the practical system by substituting appropriate data for the diffusion
coefficient of an analyte in the coating (Df) and the fibre coating thickness (b - a).
For example, the equilibration time for the extraction of benzene from a
perfectly stirred aqueous solution with a 100 Am PDMS film is expected to be
about 20 s. Equilibration times close to those predicted for perfectly agitated
samples have been obtained experimentally for extraction of analytes from air
samples (because of high diffusion coefficients in gas) or when very high sonica-
tion power was used to facilitate mass transfer in aqueous samples. However, in
practice, there is always a layer of unstirred water around the fibre. A higher
stirring rate will result in a thinner water layer around the fibre.
Practical agitation: Independent of the agitation level, fluid contacting a
fibre's surface is always stationary, and as the distance from the fibre surface
increases, the fluid movement gradually increases until it corresponds to bulk
flow in the sample. To model mass transport, the gradation in fluid motion and
convection of molecules in the space surrounding the fibre surface can be
simplified by a zone of a defined thickness in which no convection occurs, and
perfect agitation in the bulk of the fluid everywhere else. This static layer zone is
called the Prandtl boundary layer and is discussed in detail in Chapter 9. Its
thickness is determined by the agitation conditions and the viscosity of the fluid.
Figure 13.49 shows theoretical equilibration time for this case. The equili-
bration time can be estimated for practical cases from the equation below:
8K
t e = t95 = 3 (b -a) (13.13)
D,
440
13.5.2 Method development
Developing a new SPME method in most cases requires the following steps: (1)
Selection of fibre coating. (2) Selection of the derivatization reagent, if required.
(3) Selection of the extraction mode. (4) Selection of the agitation method. (5)
Selection of separation and/or detection technique. (6) Optimization of desorp-
tion conditions. (7) Optimization of sample volume. (8) Determination of the
extraction time profile in a pure matrix. (9) Determination of extraction time.
(10) Calculation of the distribution constant. (11) Optimization of extraction
conditions (pH, salt, temperature). (12) Determination of the linear range for
pure matrix at optimum extraction conditions. (13) Selection of the calibration
method. 14) Optimization of the extraction conditions for heterogeneous sam-
ples. (15) Verification of the equilibration time, sensitivity and linear dynamic
range for complex samples. (16) Determination of method precision. (17) Deter-
mination of method detection limit. (18) Validation of the method. (19) Automa-
tion of the method
In most cases, not all the steps need to be performed. Knowledge gained from
previous experiments, as well as from literature, can often be applied to the
problem at hand. Most SPME methods developed to date are used in combina-
tion with gas chromatographic separation and an appropriate detection method.
Hyphenation with HPLC and other techniques should be also considered. A dis-
cussion of the particular steps in method development follows, with the main
focus on aqueous samples.
441
t 100 prn
jM
30 pm PDMS
lM 7pm J
MEN Poly(acrylate)
* PDMS\DVB
,*Carbowax\DVB (also TR)
= Carboxen
442
size of a molecule. The template resin has uniform pore diameters, therefore it
exhibits uniform extraction efficiency.
When a group of analytes of different characteristics (e.g., pesticides) is to be
determined by SPME, primary consideration should be given to the analytes
that are the most difficult to extract. A coating should be chosen that enables the
determination of all the analytes with enough sensitivity. When none of the
commercially available coatings meets the requirements, custom coatings can
sometimes be prepared. For example, Fig. 13.47 shows the separation of alcohols
extracted from unleaded gasoline using a fibre coated with commercially avail-
TM
able Nafion perfluorinated resin [92]. SPME device also facilitates micro-
solvent extraction. For example, porous coatings can be used to support organic
solvent by placing the fibre into the solvent prior to extraction [93]. Solvent
located in the pores extracts analytes from the matrix, when the fibre is placed in
contact with a sample.
443
3 5
25°ZDy
CH 2 OCOC 2H5
.t~j X I ------r
I UU I 0U0 150UU
- 0u0 40 120 200 280 360
Retention Time (s) Mass / Charge
1. Acetic; 2. Propionic; 3. Isobutyric; 4. Butyric; 5. Pivalic 6. Isovaleric; 7. Valeric; 8. Hexanoic Acids
Fig. 13.51. (a) Reconstructed GC/MS chromatogram illustrating the presence of short-chain
fatty acids in a real sewage sample (SPME extraction with PA-coated fibre). Peak assignment: 1,
acidic; 2, propionic; 3, isobutyric; 4, butyric; 5, pivalic; 6, isovaleric; 7, valeric; and 8, hexanoic
acids. The peaks shown correspond to pyrenylmethyl esters of these acids. (b) An example of the
mass fragmentogram and the structure of propionic acid/PDAM ester.
volatile. The doped fibres not only facilitate spot sampling, but can also measure
long-term exposure to a given contaminant, since analytes are continuously
accumulated onto the fibre until all the reagent is consumed.
444
TABLE 13.4
Sampling mode selection criteria
Sampling mode Analyte properties Matrices
Direct medium to low volatility Gaseous
Headspace high to medium volatility liquid (including complex); solid
Membrane protection low volatility complex samples
TABLE 13.5
Agitation methods in SPME
Method Advantages Disadvantages
Static (no agitation) simple, performs well for gaseous limited to volatile analytes and
samples headspace SPME
Magnetic stirring common equipment, good requires stirring bar in the vial
performance
Intrusive stirring very good performance sealing the sample vial difficult
Vortex/moving vial good performance, no need for a stress on needle and fibre
stirring bar in the vial
Fibre movement good performance, no need for a stress on needle and fibre, limited
stirring bar in the vial to small sample volumes
Flow through good agitation at rapid flows potential for cross contamination,
requires constant flow
Sonication very short extraction times noise, heating of the sample
445
-S
Care must be taken when using magnetic stirring to ensure that the rota-
tional speed of the stirring bar is constant and that the base plate does not
change its temperature during stirring. This usually implies the use of high
quality digital stirrers. Alternately, with cheaper stirrers, the base plate should
be thermally insulated from the vial containing the sample to eliminate varia-
tions in sample temperature during extraction. Magnetic stirring is efficient
when fast rotational speeds are applied. Figure 13.53 illustrates the dependence
of the equilibration time for benzene on the rotational speed. The equilibration
time progressively decreases with stirring speed increase, and at 2500 rpm is
only about 200 s. Intrusive stirring can improve the agitation further, but it
requires a direct connection between the stirrer and the motor, which is difficult
to seal.
Needle vibration technique uses an external motor and a cam to generate a
vibrating motion of the fibre and the vial. This technique has been implemented
by Varian in the SPME autosampler. In the Vortex technique the vial is moved
rapidly in a circular motion. Both techniques generally provide good agitation,
resulting in equilibration times similar to those obtained for magnetic stirring.
However, for the needle vibration technique, good performance is limited to
small vials and direct extraction mode. Both techniques can be conveniently
applied to process a large number of samples since the sample vials do not
require any manipulations such as introduction of stirring bars.
Flow through techniques are very useful in continuous monitoring applica-
tions and can also be automated. However, some additional flow metering
devices may be required to ensure reproducible mass transfer conditions. The
most efficient agitation method evaluated to date for SPME applications is the
direct probe sonication, which can provide very short extraction times, fre-
quently approaching the theoretical limits calculated for perfectly agitated sam-
ples. However, this technique has substantial drawbacks associated with large
amounts of energy introduced into the system, which cause the sample to heat
up and in some cases can lead to decomposition of the analyte. Most of these dis-
advantages are eliminated when sonication is applied in the flow-through mode.
446
13.5.2.5 Selection of separationand/or detection technique
So far, most SPME applications have been developed for gas chromatography,
but other separation techniques, including HPLC, capillary electrophoresis (CE)
and supercritical fluid chromatography (SFC), can also be used in conjunction
with this technique. The complexity of the extraction mixture determines the
proper quantitation device. Regular chromatographic and CE detectors can
normally be used for all but the most complex samples, for which mass spectrom-
etry should be applied. As selective coatings become available, the direct
coupling to MS/MS and ICP/MS becomes practical.
447
TABLE 13.6
GC injectors and their compatibility with SPME
Split/splitless PTV SPI
Can be used for SPME in splitless Can be used for SPME; low Low internal volume; best
mode; low volume insert required. volume insert required. suitable for SPME.
TABLE 13.7
Effect of temperature on gas/coating distribution constant, Kf5
Compound Kfg
25°C 250°C
Benzene 300 0.4
Xylene 3,000 1.4
Undecane 26,000 13
onto the front of the column when narrow bore inserts are used might still be too
long to produce narrow bands for very volatile analytes. It is necessary, there-
fore, to refocus the band after the desorption is finished. A thick film column or
cryofocusing can be used for this purpose. It is always advisable to consider the
application of a short length of deactivated narrow bore fused silica tubing as a
retention gap. This will eliminate the broadening effect created by the tempera-
ture gradient existing in the column close to the injector. Except for the most vol-
atile analytes (gases), the combination of a thick film column and the retention
gap is usually sufficient to refocus the injection band, and cryofocusing does not
have to be applied.
Parameters which control the desorption process in the HPLC interface are
analogous to those in GC applications. In addition to temperature and flow rate,
the composition of the mobile phase also affects the process. In many cases, it is
possible to use the mobile phase without any modifications, as in PAH analysis
[98]. In some instances, addition of an appropriate solvent to the interface will
assist desorption. The linear flow rate of the mobile phase should be maximized
by choosing a small i.d. tubing for the desorption chamber. Temperature also
plays an important role in accelerating the desorption process as illustrated in
Table 13.8. At ambient desorption temperatures, a significant carryover of
Triton X-100 was observed, which was corrected when the temperature was
raised to 135°C [99].
448
TABLE 13.8
Effect of desorption chamber temperature on carryover of Triton X-100 using Carbowax/DVB
fibre
No. of units in the ethoxylate chain (n =) %Carryover
21°C 135°C
4 15 0
5 20 2
6 30 2
7 25 3
8 26 1
9 26 2
10 27 2
11 27 3
12 26 2
13 28 3
14 26 0
extraction. This results in optimum sensitivity and better precision since the
variation in sample volume does not affect the amount of analyte extracted. In
the case when the sample volume is low (Fig. 13.54B) then substantial depletion
of the sample occurs during extraction, resulting in loss of sensitivity and
precision. For example, in Fig. 13.54B, only half the amount of analytes is
extracted compared to Fig. 13.54A. Therefore the optimum sample volume of the
sample should be selected based on the estimated distribution constant Kfs. The
distribution constant can be estimated using literature values for the target
analyte or a related compound with the coating selected. For the PDMS coating,
it can also be directly estimated from the chromatographic parameters. In all
other cases, Kfs has to be determined experimentally by equilibrating the sample
with the fibre and determining the amount of analyte extracted by the coating.
Care must be taken to avoid analyte losses via adsorption, evaporation,
microbial degradation, etc., when very long extraction times are required to
reach the equilibrium.
449
The sensitivity of the SPME method is proportional to the number of moles
of the analyte n extracted from the sample and, for direct extraction method, is
given by Eq. (13.1).
As the sample volume Vs increases, so does the amount of analyte extracted
until the volume of the sample becomes significantly larger than the product of
the distribution constant and volume of the coating (fibre capacity K 6 < < Vs). At
this point, the sensitivity of the method does not increase with further increase
in volume. In practice, the limiting volume can be calculated from the following
dependence:
v = 100KV
E
(13.14)
where E is the magnitude of the relative measurement error (%). For example,
for a 5% error, it can be estimated that Vs = 20KfsVf. This means that for Kf,
values of up to about 200 and a 100 m coating, a 2 ml vial is sufficient to give
maximum sensitivity, while a 40 ml vial can be used for Kfs values smaller than
4000, etc. Using sample volumes larger than the limiting value not only maxim-
izes the sensitivity, but also results in better precision since variations of the
sample volume do not affect the results. Figure 13.55 illustrates this relationship
a 20
XO
K I0
Q000
40
0 10 20 30 40
Simple volme [tL
20
0 coal&K
ting0
80
m K=I 5000
o ltuame
Salpie [r ml
0 10 20 30 40 50
Fig. 13.55. Dependence of n/n o on V. for (a) 100 Am fibre (V, = 0.65 l), (b) 30 xm fibre (Vg 0.14
/l)
,u1) for direct SPME from small volume samples (two-phase system).
450
4 mL vial, 100 pm coating 40 mL vial, 100 pm coating
101
1060 ' _ --- 160 _ ...-
20 20
0 2 C 3
0 1 2 3 4 0 10 20 30 40
Vhs Vhs
15 mL vial, 100 pm coating
100 ) . _
---
--·e-e
80
·-,
60
-x
`·- --- chloroform
40
- 1,1,1-trichloroethane
20
- carbon tetrachloride
fl
0 2 4 6 8 10 12 14
Vhs
Fig. 13.56. The effect of headspace volume on the amount of analyte extracted in a system of a
constant volume of (a) 4, (b) 15, and (c) 40 ml by a 100 um PDMS fibre for chloroform,
1,1,1-trichloroethane and carbon tetrachloride.
451
of the analyte in the sample remains relatively high. For the two other curves
and low headspace volumes, n/no is higher for carbon tetrachloride than for
1,1,1-trichloroethane, which is consistent with the values of their distribution
constants Kf,. However, as carbon tetrachloride has the largest Kh,, n/no drops
faster for this compound than for 1,1,1-trichloroethane, as a result of which the
respective two lines cross each other. The biggest relative increase in n/n, when
moving from small to large vials (-44%) is observed for 1,1,1-trichloroethane,
characterized by the highest Kf, value. For compounds with lower Kf values the
relative increase in the amount extracted is much lower (-16% for chloroform
and - 12% carbon tetrachloride). It is therefore the combination of Kf, and Kh, for
a given compound that determines the magnitude of the effect of sample volume
on the amount of analyte extracted by the fibre in three-phase system involving
headspace [100].
Headspace volume can have a significant effect on equilibration times
(extraction kinetics). If the amount of the analyte extracted by the fibre at equi-
librium is negligibly small compared to the amount present in the headspace
equilibrated with the sample prior to the extraction, the analyte is extracted
almost exclusively from the gaseous phase, and the process is much faster com-
pared to the case when significant amounts of the analyte have to transported
from the liquid sample to the fibre via the headspace. Assuming that this situa-
tion occurs when 95% of the analyte is extracted exclusively from the headspace,
the criterion that must be fulfilled can be described by Eq. (13.15). The assump-
tion is reasonable, as a 5% difference usually falls within the limits of experimen-
tal error for trace SPME/GC analysis:
Kffi Vf (13.15)
KhVh
The above criterion means that the capacity of the headspace (KhVh,,) needs to be
at least 20 times larger than the capacity of the fibre (KfV) to achieve rapid
extraction. For a given sample volume V,, this can be achieved by using a large
enough headspace volume, or by increasing Khs. The latter can be accomplished
by increasing the temperature or by salting the analyte out of the liquid phase.
When the criterion (13.15) is fulfilled, equilibration can take as little as a few
minutes, and is almost independent of the agitation conditions (provided the
analyte is equilibrated between the liquid phase and its headspace before the
extraction begins). It should be emphasized, however, that the increase in
headspace capacity causes a significant loss in sensitivity. In any case, it should
be remembered that changing the vial size during the optimization process can
not only affect the sensitivity, but also the equilibration time if the headspace/
sample volume ratio is different in the new vial.
452
A p,p'- DDT 13 Dichlorvos
GC Response(area counts) GC IResponse (areacounts)
100,000 2,5C
80,000 2,0c0 ii
60000 15Eo O %
rel.error
40000 ; : 7% rel. error 1,0
20,000 5 10 5/i m ro
20% rel error
o 0
0 50 100 150 200 250 300 350 i
0 50 100 150 200 250 300 350
Timeminutes) Time(minutes)
PDMS fiber, 100 pm
Fig. 13.57. Selection of the extraction time based on extraction time profiles.
kinetics of analyte transfer from the matrix to the fibre under given mass trans-
fer conditions. The profile is determined by extracting samples of identical com-
position for progressively longer periods of time and plotting the resultant area
counts vs. time. The mass of the analyte extracted can be quantified by a proper
calibration procedure, e.g. syringe injection of a standard solution of the analyte
in an appropriate solvent. In gas chromatography, this type of calibration is reli-
able only when non-vaporizing injectors are used (PTV, SPI, heated on-column).
453
TABLE 13.9
Ranges of K values obtained when the amount of the analyte extracted by the fibre n determined
experimentally falls within 5% (relative) of the true value, for two different fibre coating
thickness and two sample volumes
2 ml sample 35 ml sample
K Range of K K Range of K
100 um
100 105-95 100 105-95
1,000 1067-935 1,000 1051-949
10,000 12537-8172 10,000 10598-9413
100,000 -36190 100,000 115748-86928
1,000,000 -55072 1,000,000 4700000-492593
10,000,000 -58104 10,000,000 -923611
100,000,000 -584 26 100,000,000 o-1012177
7 gm
100 105-95 100 105-95
1,000 1051-949 1,000 1050-950
10,000 10574-9434 10,000 10504-9496
100,000 112903-88785 100,000 105422-94622
1,000,000 3500000-558824 1,000,000 1093750-913462
10,000,000 o-1187500 10,000,000 17500000-6785714
100,000,000 -1338028 100,000,000 -19000000
curve rises rapidly. Figure 13.57b shows an example when a 1 min deviation
from 10 min sampling time results in 50% relative error of determination.
Autosamplers can measure the time very precisely and the precision of
analyte determination can be very good, even when equilibrium is not reached in
the system. However, this requires the mass transfer conditions and the temper-
ature to remain constant during all experiments.
454
Kh nV (13.16)
Vf(CoV -n)
It is important to consider the sample volume when calculating the value of Kf.
Table 13.9 presents the ranges of K values that are obtained when the amount of
the analyte extracted by the fibre n determined experimentally falls within + 5%
(relative) of the true value, for two sample volumes (2 and 35 ml) and two fibre
coating thicknesses (100 and 7 m). A +5% error can be assumed typical for
trace analysis by SPME/GC. To correctly estimate the distribution constants
over 1000 with 100 gm thick PDMS fibre, volumes of the sample larger than 2 ml
need to be used. For very large Kfs values (>1,000,000), typical for high
molecular weight hydrocarbons in water or air, no reliable estimates ofKf, can be
obtained even with 35 ml sample when the 100 gm fibre is used, therefore only
the 7 m fibre should be used for this purpose.
If the distribution constant is low and the total volume of the sample is high,
a good estimate of Kfs can be obtained from the following relationship:
n
Kf = n (13.17)
C0Vf
However, extreme caution is advised when using this equation, since it assumes
that after the extraction, the analyte concentration in the sample does not
change significantly. If this equation is used by mistake when the Kfs value is
large or the sample volume is too small, the result of the calculation does not cor-
respond to Kf,, but rather to the sample/coating phase volume ratio (V /Vf).
To determine the coating/matrix distribution constant when headspace is
present in the vial, knowledge of the matrix/headspace distribution constant is
required:
Ks = KHRT (13.19)
RT
For analytes with an unknown Henry constant, the following equation can be
used:
Kfs = nKf (13.20)
Kf VfVsCo -nKfh V -nVh
Equation (13.20) is obtained by substituting Kh, = KfS /K into Eq. (13.18) and
rearranging it. The fibre/gas distribution constant can be found by extracting
target compounds from air mixtures by using a simple bulb experiment [101]. In
addition, Kfg can be obtained directly from a chromatographic run when the sta-
tionary phase is made of the same material as the fibre coating. For compounds
455
I I
a)
* pH 1035
* p 11
C0 2 H 13|
Amphetamine Methamphetamine
which have low vapour pressures, the amount of analyte present in the head-
space is very small, especially if the volume of the headspace is kept to a mini-
mum. Therefore, in this situation, the headspace volume can be neglected
altogether and Eq. (13.1), which corresponds to direct extraction, can be applied.
456
5000000
4000000
0)
U 3t)0000
IE 15o salt
O iti0 5 ~1
al 2000000
1000000
0
Amphetamine Methamphetanile
13.5.2.12 Determinationof the linear dynamic range of the method for a pure
matrix at optimum extraction conditions
Modification of the extraction conditions affects both the sensitivity and the
equilibration time; therefore, it is advisable to check the previously determined
extraction time before proceeding to the determination of the linear dynamic
range. This step is required if substantial changes of the sensitivity occur during
the optimization process.
SPME coatings include polymeric liquids such as PDMS, which by definition
have a very broad linear range. For solid sorbents, such as Carbowax/DVB or
PDMS/DVB, the linear range is narrower because of a limited number of
sorption sites on the surface, but it still can span over several orders of magni-
tude for typical analytes in pure matrices. In some rare cases when the analyte
has extremely high affinity towards the surface (e.g., basic proteins adsorption
on poly(acrylic acid) [103] or bioaffinity coatings), saturation can occur at low
analyte concentrations. In such cases, the linear range can be expanded by
shortening the extraction time. In the majority of practical applications
457
developed to date, it is the dynamic range of the detector that limits the linear
response of SPME methods.
458
TABLE 13.10
Quantitation of PAHs from urban air particulate (NIST 1649) by static high temperature water
extraction
Compound Cert conc. Lg/g Estimated concentration as %of certified concentration
(%RSD) (%RSD)
100 mg/270°C 50 mg/270°C 50 mg/300°C
Fluoranthene 7.1 (7) 137 (6) 149 (8) 128 (2)
Pyrene 7.2 (7) 113 (9) 121 (7) 105 (7)
Benzo(a)pyrene 2.9 (17) 49 (14) 72 (10) 70 (4)
459
static hot water extraction at 250°C, followed by a cool-down period to increase
the amount of analytes absorbed by the PDMS coating.
460
ture; (4) sample volume; (5) headspace volume; (6) vial shape; (7) condition of
the fibre coating (cracks, adsorption of high M.W. species); (8) geometry of the
fibre (thickness and length of the coating); (9) sample matrix components (salt,
organic material, humidity, etc.); (10) time between extraction and analysis; (11)
analyte losses (adsorption on the walls, permeation through Teflon, absorption
by septa); (12) geometry of the injector; (13) fibre positioning during injection;
(14) condition of the injector (pieces of septa); (15) stability of the detector
response; (16) moisture in the needle. The majority of these factors and their
effect on method precision have already been discussed. A brief explanation of
the remaining parameters follows.
The reproducibility of the sample and headspace volumes can impact the
precision of SPME methods. The volume of the sample has an impact on the
amount extracted when only a few ml sample is used (see the discussion above).
Therefore, small samples such as 1 ml need to be measured very carefully. For
large samples, small variations in the sample volume do not have a substantial
effect on the precision of the data. Obviously, the volume of all the samples and
standards must be the same to obtain valid results.
Headspace volume can be an important factor determining precision of the
results in three-phase systems. It is relatively easy to measure sample volume
accurately. However, vials are not manufactured to have exactly the same total
volume. Wall thicknesses and bottom shapes may differ from vial to vial. Also,
the shape of the septum in a closed vial can vary from concave to convex. All
these factors will affect the total volume of the system, and therefore the
headspace volume, which is the difference between the total volume and the
sample volume. Such differences (especially related to septum shape) are usually
more pronounced in small vials, therefore worse precision can usually be expect-
ed when they are used. In addition to keeping the sample and standard volumes
constant throughout all the experiments, vials of the same volume should be
used for all samples and standards, so that the headspace volume is constant
from vial to vial. This becomes a bigger challenge for small vials, when a small
absolute variation in headspace volume could translate to a large relative
difference.
The condition of the fibre is very important: the presence of high molecular
weight species, such as proteins and humic matter, adsorbed on the surface of
the fibre, causes a change in sorption characteristics. In some cases, the sub-
stances adsorbed can be removed by washing the fibre in appropriate solvents.
When the coating is badly cracked, a small amount of the matrix itself kept in the
cracks by capillary forces might be transferred into the analytical instrument
and adversely affect its operation.
The volume of the coating determines the amount of analytes extracted.
Fibres which begin to lose the coating phase should be replaced immediately.
The length of the fibre should be monitored to ensure equivalency when
changing the fibre. If the length of the fibre (thus the volume of the coating) is
different, an appropriate correction factor should be incorporated.
461
Time between preparation of standards or sample collection and extraction
should be minimized. After placing aqueous samples or standards in a vial,
analytes begin to interact with the walls of the container. Very hydrophobic sub-
stances are frequently strongly adsorbed on the surfaces and are not available
for extraction by the coating. This results in lower than expected extracted
amounts. One way to prevent this problem is to use vials with more inert walls,
e.g., silanized glass or Teflon. For volatile analytes, it is crucial that new Tef-
lon-lined septa are always used. Once the Teflon lining is pierced, the analytes
can come into direct contact with the septum material (typically polydimethyl-
siloxane), which usually results in losses through absorption. Finally, it should
be remembered that many volatile compounds can permeate through Teflon at a
relatively high rate. This phenomenon can also contribute to analyte losses
when the samples or standards are stored for prolonged periods of time.
The time between extraction and analysis should also be minimized. When
this time is longer that a few seconds, the needle of the fibre assembly should be
sealed with a septum and preferentially cooled down to reduce losses and
interferences. After the fibre is withdrawn from the sample, it begins to equili-
brate with ambient air. This process is slow when the fibre is withdrawn into the
needle. Therefore, the few seconds that are normally required to transfer the
fibre to the injection port of a gas chromatograph do not result in any substantial
losses of the analytes. However, when the fibre is to be stored for some time
before the analysis, sealing of the needle is necessary. In addition, this protection
prevents contamination of the fibre with interferences or even target analytes
which might be present in ambient air. Cooling of the SPME device might be
necessary to extend the stability of the volatile analytes in the coating [11].
13.5.2.17 Determinationof method detection limit
There are several methods described in the literature to determine the method
detection limit. The most widely accepted definition is based on estimating the
detection limit using low concentration spikes and calculating the standard
deviation of the determination. The detection limit is then defined as three times
the standard deviation obtained for the measurement value not higher than 10
times the method detection limit.
TABLE 13.11
Regression line parameters for SPME vs. purge and trap for clean water analysis
Compound Regression parameters-purge and trap on x-axis
Slope y-Intercept Correlation
Benzene 1.07 + 0.087 -5.9 + 6.6 0.9951
Toluene 1.06 + 0.060 -4.6 + 4.6 0.9976
Ethylbenzene 1.06 + 0.046 -3.8 - 3.7 0.9986
m/p-Xylene 1.05 + 0.073 -8.4 + 11 0.9964
o-Xylene 1.07 + 0.060 -4.0 + 4.6 0.9977
462
TABLE 13.12
Statistical characteristics of the results obtained in the round robin test on pesticide analysis by
SPME (sr = repeatability standard deviation, sL = inter-laboratory standard deviation, s, =
reproducibility standard deviation, r = repeatability, R = reproducibility, G.A. = gross average,
C.I. = confidence interval of the gross average, T.V. = confidence interval of the "true" value. All
values are expressed in g/l)
Compound sr sL sR r R G.A. C.I. T.V.
Dichlorvos 2.06 5.04 5.44 5.83 15.4 227.3 27±5.8 25±1.35
EPTC 0.56 1.56 1.66 1.57 4.7 9.9 10+1.6 10+0.54
Ethoprofos 0.82 4.79 4.86 2.32 13.74 15.5 16±2.3 17+0.92
Trifluralin 0.27 0.57 0.63 0.76 1.79 1.6 1.6±0.76 2±0.11
Simazine 2.34 3.45 4.17 6.61 11.79 23.6 24+6.6 25±1.35
Propazine 1.21 2.04 2.37 3.42 6.71 9.5 10-3.4 10±0.54
Diazinon 0.63 2.13 2.22 1.79 6.29 8.2 8+1.8 10-0.54
M. chlorpyriphos 0.12 0.32 0.34 0.35 0.97 1.6 1.6±0.35 2-0.11
Heptachlor 2.03 2.89 3.53 5.75 10 8.9 9+5.8 10+0.54
Aldrin 0.54 0.73 0.91 1.53 2.58 2.0 2+1.5 2+0.11
Metolachlor 0.73 2.83 2.92 2.07 8.28 15.7 16±2.1 17±0.92
Endrin 0.87 3 3.13 2.47 8.85 8.8 9±2.5 10±0.54
463
values. This might be due in part to losses of analytes through adsorption (as
described above) in cases where the aqueous pesticide solutions were not pre-
pared directly before the analysis, as required by the test protocol.
This section discusses the more fundamental concepts related to the application
of SPME for analysis of real matrices and physicochemical applications. The
applications of the technique are covered in more detail in the application chap-
ters, as well as in the bookApplications of Solid PhaseMicroextraction[109]. An
updated list of SPME publications is posted in a database at the University of
Waterloo website (www.spme.uwaterloo.ca). A basic understanding is emerging
regarding approaches which can be taken to the basic matrix types: gaseous, liq-
uid and solid. Progress in this direction is summarized in the first three parts of
this section. General approaches to facilitate successful extraction and quantifi-
cation in these systems are emphasized. The objective of this section is not to
demonstrate a comprehensive review, but rather to discuss fundamental princi-
ples facilitating successful quantification. Physicochemical applications empha-
size the capability of SPME to characterise distribution of analytes in the
natural system; the section on investigation of living objects briefly summarizes
this exciting new direction.
All SPME methods developed to date for gas analysis are directed towards air
analysis. However, the approaches described are also suitable for the analysis of
other gas mixtures.
Model studies on extraction from air have been performed by two methods:
static and (the more realistic) dynamic. In the static method, the target com-
pounds are introduced to a glass bulb equipped with a septum. After the analytes
vaporize completely, SPME fibre is introduced to the bulb through a septum. In
practical ambient air measurements, the system is not static, since convection is
always present. It is more appropriate, therefore, to use dynamic gas chambers
464
logK = a-)+b
4.50
1.50
Fig. 13.60. Representative plots for pentane, benzene, p-xylene, mesitylene and n-undecane
demonstrating the linear relationship between logKf and 1/T.
for the modelling studies [110]. Equilibration times for extraction of trace
contaminants from moving air are short, as predicted from the theory. Increased
air flow decreases the equilibration times for less volatile analytes.
PDMS coating is generally not affected by major air components. A small
decrease in the response for organics was observed only for samples of relative
humidity close to 100%. The primary experimental parameter which controls
the response is temperature. However, the effect of temperature on the distribu-
tion constant can be easily predicted, since logKfg is linearly related to 1/T and
the heat of vaporization of the pure solute. In other words, Kfg can be calculated
for given extraction conditions using the temperature measured and the heat of
vaporization for the target compound. Figure 13.60 illustrates the above rela-
tionship graphically for a range of compounds varying in volatility. The linear
relationship is clearly illustrated. In addition, as has been described in the
section on theory, the heat of vaporization and activity coefficient are related to
the retention time of a compound on the GC column using the same coating
material as the SPME fibre. Thus, the appropriate distribution constants can be
determined directly from the retention indices of the target analytes. For PDMS
coating, the following relationship between the gas/coating distribution con-
stant and the linear temperature programmed retention index of a compound
(LTPRI) has been found (see Fig. 13.61):
logoKfg = 0.00415- LTPRI- 0.188 (13.16)
Thus, the LTPRI system permits interpolation of the Kfg values from the plot of
logKfg versus retention index. The differences between the values of distribution
constants calculated from the retention indices and determined directly were
465
nn
5.00
*E 4.00
0 I
m_ 3.00
2.00
1.00
0.00
400 500 600 700 800 900 1000 1100 1200 1300 1400 1500
466
n~ l
100%
(a)
TIC
i I I I I
0.46% Naphthalene
128 (b)
(a) TIC
wI I I 1 * (b) m/z = 128
3.14% Phenanthrene (c) m/z = 178
(C) (d) m/z = 202
178 Anthracene
Fluoranthene
(d)
202
Fig. 13.62. GC/MS traces corresponding to polycyclic aromatic hydrocarbons present in diesel
exhaust sampled by SPME/PDMS. (a) Total ion chromatogram; selected ion traces at: (b) m/z =
128, (c) 178, (d) 202.
The majority of methods for liquid matrices reported in the literature were
developed for aqueous samples. The fibre coating/water distribution constant
can be calculated from the following equation:
467
1
iogK ,= 0.0042*LTPRI - 0.188 + logK,,aw I
5.50
*
4.00 -
350
3 00
3; -
Cycopentane and
250 ,' BenzeneSubstituted Aromatics
*Cylohexe
2.50 Derivatives ' IogK,
ogK, = 0 0043LTPRI -0
00043*LTPR1 88
osS
~- ~IgK,
0.0040
ULTPRI 0.69
2.00 C C C =K
.i50 K C cf h K1
.oo
1.00 ,i,
10 200 400 600 800 1000 1200 1400
LTPRI
Fig. 13.63. LogK, (PDMS) as a function of LTPRI for isoparaffins, substituted benzenes and
cycloalkanes at 25C.
where Kfg can be determined from chromatographic data as discussed above, and
Kaw is the air/water distribution constant for a given analyte which can be found
in the tables of Henry's constant values. For example, the equation for a PDMS
coating and aqueous matrix can be calculated from the equation below:
This equation is obtained after substituting the expression for Krg given by Eq.
(13.17) into Eq. (13.16). The Kfw values found using this equation agree very well
with experimental data considering that typically the errors in Henry's constant
determination are larger than 10%. In addition, Henry's constants are similar
for closely related compounds, as illustrated in Fig. 13.63. As could be predicted
from Eq. (13.18), the slopes of the lines are similar, but their intercepts vary
because of the differences between the average values of Henry's constant for
the different classes of compounds. This relationship enables quantitation of
unknown analytes in water provided that the class to which a compound belongs
is known.
Several reports indicated the existence of a linear relationship between the
coating/water distribution constant and octanol-water distribution constant,
Kow,, [114]. Considering the discussion above, such relationships are expected to
exist only for individual classes of compounds, such as aromatic hydrocarbons.
Because the activity coefficients of various classes of analytes in octanol are
expected to be different than in PDMS or other fibre coatings, the relationship
between the two extracting phases will vary with the change of chemical proper-
468
ties of the analytes. It is possible, however, to predict the general trends in Kf,
within a group of related compounds by using the corresponding values of Kow.
The above discussion pertains to pure water as the sample matrix. The
presence of other components in water might affect the distribution constants of
the analytes. Also, liquid chromatographic experience gives some clues about the
trends in the distribution constant change with modification of the matrix. For
example, addition of salt would generally result in an increase in the distribution
constant for neutral organics, but the change is expected to be noticeable only if
the concentration of the salt exceeds 1%. Also, the presence of water miscible
polar organic solvents may change the properties of the matrix by reducing its
polarity. In addition, swelling of the polymer with the solvent might occur for
polar coatings, resulting in a change of the coating volume and possibly also of
Kf,. However, the change is not expected to be substantial when the concentra-
tion of the solvent is below 1% [115]. When samples contain higher amounts of
salt and/or dissolved organics, but they are well defined so that pure matrix can
be prepared, external calibration using standards in the matrix may still be
appropriate. Otherwise, standard addition (preferentially of isotopically labelled
analytes) should be used to compensate for variations in the matrix composition.
A major analytical challenge is always associated with the analysis of sam-
ples containing solids, such as sludge. Several approaches can be implemented
with SPME. Sometimes modification of the extraction conditions, such as tem-
perature, pH, salt and other additives, facilitates the displacement of analytes
into the aqueous phase or the headspace, resulting in distribution constants sim-
ilar to those for pure water. In certain cases, direct extraction is not possible
because of a very dirty matrix or pH conditions which damage the fibre (above
pH 12 for the PDMS coating). In such situations, the headspace mode is more
suitable for many applications. Even semivolatiles can be analyzed by this
method as long as the extraction temperature is sufficiently high, and good agi-
tation conditions are provided.
469
the fibre decreases. The inherent loss of sensitivity associated with the decrease
of the distribution constant at high temperatures can be compensated for by
cooling the fibre, as discussed earlier. However, this approach has not yet been
commercialized. Alternatively, matrix modifier can be added to the solid sample.
Water has been found to be very effective in displacing the analytes from the
solid surfaces for many types of samples, especially at elevated temperatures.
Another successful approach to SPME analysis of solids involves the use of
water or a polar organic solvent, such as methanol, for the extraction of the
analytes from the solid matrix prior to SPME analysis. When a polar solvent is
used, the extract is spiked into pure water. Low temperature water extraction
followed by SPME was found to be a very useful approach for polar compounds,
such as herbicides [116]. Application of methanol with water spiking, on the
other hand, has been found to be useful for the analysis of volatile hydrocarbons
[117]. An interesting modification to the above procedure involves volatizing the
extract followed by fibre extraction from the gaseous phase [118]. Quantitation
of analytes in the gas phase is easier, as discussed above.
For less volatile analytes, the methanol approach can give good results.
High-pressure hot water extraction is also a suitable solvent-free alternative
(see Chapter 18). Both static and dynamic hot water extraction have their advan-
tages: the static method is very simple and inexpensive as it does not use high
pressure pumps; the dynamic approach, on the other hand, provides an extract
which is much cleaner than the original matrix. However, it might be possible to
extract many semivolatile target analytes from very dirty matrices with the high
temperature static system, when using the headspace mode of SPME.
Solid phase microextraction can be applied not only for extraction purposes, but
also to perform measurements which better characterize the extraction system.
The measurements can include studying the properties of the fibre coating and
investigation of multiphase equilibria in the matrix. These investigations are
convenient because SPME is and equilibrium technique and the fibres are small
and therefore to not disturb chemical equilibria in the system as it is illustrated
on Fig. 13.64. In addition, since detailed mathematical treatment of the solid
470
illlall II CVaLU
---I-A RUMS
TiYlllr riht
IIVlt·
l
phase microextraction process is available, SPME can be applied to study the
parameters involved in the extraction, including diffusion coefficients in both
the coating and the extracted phase, as well as various distribution constants.
For understanding solute-solvent interactions at the molecular level and the
thermodynamic processes involved in forming the solution, the study of infi-
nite-dilution activity coefficients of probe solutes in a polymer phase is a funda-
mental approach. SPME can be a useful tool for determining these coefficients.
The stationary phases of interest can be coated on fibres made of suitable mate-
rials (fused silica, stainless steel, etc.). The SPME device with a selected fibre
coating can be used to extract a group of probe compounds, which are then sepa-
rated on a standard commercially available column and quantified by a GC/MS.
The SPME measurements of infinite-dilution activity coefficients of probe
solutes in a selected stationary phase consist of two steps. The first step is to
measure the infinite-dilution partition coefficients of the probe solutes between
the stationary phase and the gas phase. Then, from these partition coefficients,
the corresponding activity coefficients are determined [119].
Investigation of analyte distribution in natural systems is critical for many
disciplines of science including environmental chemistry and biochemistry.
Information gained during these studies is also very important to analytical
chemists, since it facilitates the development of optimum sampling protocols and
can lead to more rational optimization of extraction parameters. One approach
to study multiphase equilibria involves separation of the individual phases after
equilibrium has been reached and analysis of the individual phases to determine
the concentrations of target compounds. For example, SPME was used to study
the concentration of organic substances from octanol saturated water [12].
However, the separation of phases is not always necessary. SPME has been
applied to measure the distribution of chemicals between the components of a
system, for example, consisting of water and macromolecules such as dissolved
humic organic matter (HOM) [121]. There are two methods for performing this
experiment [122]. Typically, the calibration curve is first constructed by employ-
ing the amount of the analytes partitioned on the fibre vs. the free analyte
concentration for pure water. In method I designed for large volume systems
(see Fig. 13.65), the extraction is then performed in the binding macromolecule
SPME
fibre
vial
sample solution
dissolved protein
471
l lI .Protein Soiition
Polethylee Inseralt
solution (e.g., protein, humic material) with a known total amount of the analyte
(e.g., drug, pesticide). In method II designed for small samples (see Fig. 13.66), a
known amount of the analyte is first loaded onto the fibre and then this fibre is
put into the protein solution for delivery of the analyte. The amount of analyte
remaining on the fibre is analyzed after the system reaches equilibrium. The
free analyte concentration is then calculated from the calibration curve obtained
during the first step in both of the methods. Since only a very small amount (150
Al for each extraction) of the macromolecule solution is required, method II is
very, useful for situations where the macromolecule is very limited, for example,
in the case of some proteins. Method I does not require determination of absolute
masses of analyte since the only information needed is the area count change
(ratio) of the fibre injections before and after the macromolecule was introduced
into the system, to calculate the distribution in this two phase system. These
methods have been applied to drug binding studies [123]. The procedures
described above can be extended to natural systems containing native analytes
by using isotopically labelled analogues [124]. This topic is discussed in more
detail in Chapter 8.
472
TABLE 13.13
Spacial distribution of pesticides in tomato plant stem
Height from base (cm) Mass extracted (pg)
Atriazine Simazine Prometryn
113 2 2 0
99 26 52 9
92 0 0 8
68 0 26 0
43 0 0 0
26 0 7 0
10 9 35 0
6 43 101 148
473
8
E .5
D
6
,0 . 4
@
c
o
0 2 4 6 8 0 2 4 6 8
Migration Time / minutes Migration Time / minutes
REFERENCES
474
8 M.A. Nickerson, Sample screening and preparation within a collection vessel. U.S.
patent number 5,827,944.
9 E. Baltussen, P. Sandra, F. David, H-G. Janssen and C. Cramers, Anal. Chem., 71
(1999) 5213-5216.
10 X. Liu, I. Bruheim, J. Wu and J. Pawliszyn, Anal. Chem., in preparation.
11 P. Popp and M. Moeder, in: ExTech'2001. Barcelona, Spain, September, 2001.
12 C. Grote and J. Pawliszyn, Anal. Chem., 69 (1997) 402.
13 L. Miller, in: J. Pawliszyn (Ed.), Applications of Solid Phase Microextraction. RSC
Chromatography Monographs, The Royal Society of Chemistry, Cambridge, UK,
1999.
14 D. Louch, S. Motlagh and J. Pawliszyn, Anal. Chem., 64 (1992) 1187.
15 J. Al-Hawari, in: 78th Canadian Society for Chemistry Conference and Exhibition.
Memorial University of Newfoundland, St. Johns, NF, 1996.
16 J. Poerschmann, F-D Kopinke and J. Pawliszyn, Environ. Sci. Technol., 31 (1997)
3629.
17 Z. Mester and J. Pawliszyn, Rapid Comun. Mass Spectrom., 13 (1999) 1999.
18 S. Motlagh and J. Pawliszyn, Anal. Chim. Acta, 284 (1993) 265.
19 Z. Zhang and J. Pawliszyn, Anal. Chem., 65 (1993) 1843.
20 S. Motlagh and J. Pawliszyn, Anal. Chim. Acta, 284 (1993) 265.
21 Z. Zhang and J. Pawliszyn, J. High Resolut. Chromatogr., 19 (1996) 155.
22 Z. Zhang, J. Poerschmann and J. Pawliszyn, Anal. Con., 33 (1996) 219.
23 J. Crank, Mathematics of Diffusion. Clarendon Press, Oxford, 1989, p. 14.
24 J. Pawliszyn, J. Chromatogr.Sci., 31 (1993) 31.
25 R. Eisert and J. Pawliszyn, Anal. Chem., 69 (1997) 3140.
26 M. Chai and J. Pawliszyn, Environ. Sci. Technol., 29 (1995) 693.
27 J. Crank, Mathematics of Diffusion. Clarendon Press, Oxford, 1989, p. 2.
28 A. Khaled and J. Pawliszyn, J. Chromatogr., 892 (2000) 455.
29 P. Martos and J. Pawliszyn, Anal. Chem., 71(8) (1999) 1513.
30 B. Shurmer, Investigation of the Bioavailability of Organic Contaminats by Solid
Phase Microextraction. University of Waterloo, Waterloo, Canada, 1999.
31 Z. Zhang, M.J. Yang and J. Pawliszyn, Anal. Chem., 66 (1994) 844A.
32 E. Otu and J. Pawliszyn, J. Mikrochim. Acta, 112 (1993) 41.
33 J.L. Liao, C.M. Zeng, S. Hjerten and J. Pawliszyn, J. Microcol. Sep., 8 (1996) 1.
34 F. Guo, T. Gorecki, D. Irish and J. Pawliszyn, Anal. Con., 33 (1996) 361.
35 H.B. Wan, H. Chi and M.K. Wong, C.Y. Mok, Anal. Chim. Acta, 298 (1994) 219.
36 T. Gorecki, P. Martos and J. Pawliszyn, Anal. Chem., 70 (1998) 19.
37 J. Wu, X. Yu, H. Lord and J. Pawliszyn, Analyst, 125 (2200) 391.
38 H. Ge, K. Gilmore, S. Ashraf, C. Too and G. Wallace, J. Liq. Chromatogr., 16 (1993)
1023.
39 G.G. Wallace, M. Smyth and H. Zhao, Trends Anal. Chem., 18 (1999) 245.
40 A. Michalska, A. Ivaska and A. Lewenstam, Anal. Chem., 69 (1997) 4060.
41 J. Wu, H. Kataoka, H. Lord and J. Pawliszyn, J. Microcol. Sep., 12 (2000) 255.
42 H. Kataoka, S. Narimatsu, H. Lord and J. Pawliszyn, Anal. Chem., 71 (1999) 4237.
43 G. Vlatakis, L.I. Andersson, R. Miller and K. Mosbach, Nature, 361 (1993) 645.
44 D.S. Hage, Anal. Chem., 65 (1993) 420R.
45 H. Yuan, W. Mullett and J. Pawliszyn, Analyst, 126 (2001) 1456.
46 J-L. Liao, C-M. Zeng, S. Hjerten and J. Pawliszyn, J. Microcolumn Sep., 8 (1996) 1-4.
47 T. Gorecki, X. Yu and J. Pawliszyn, Analyst, 124 (1999) 643-649.
48 G. Xiong, Y. Chen and J. Pawliszyn, Anal. Chem., in preparation.
49 P. Cheo, Fiber Optics. Prentice-Hall, Englewood Cliffs, NJ, 1985, p. 88.
50 J. Chongrong and J. Pawliszyn, J. Microcol. Sep., 10 (1998) 167-173.
51 J. Pawliszyn, Solid Phase Microextraction:Theory and Practice.Wiley, NY, 1997.
475
52 Z. Zhang and J. Pawliszyn, Anal. Chem., 67 (1995) 34.
53 S. Hawthorne, Y. Yang and D. Miller, Anal. Chem., 66 (1994) 2912.
54 H. Diamon and J. Pawliszyn, Anal. Corn., 33 (1996) 421.
55 L. Pan and J. Pawliszyn, Anal. Chem., 69 (1997)196.
56 K. Buchholz and J. Pawliszyn, Anal. Chem., 66 (1994) 160.
57 L. Pan and J. Pawliszyn, Anal. Chem., 69 (1997) 196-205.
58 L. Pan and J. Pawliszyn, Anal. Chem., 67 (1995) 4396.
59 P. Martos and J. Pawliszyn. Anal. Chem., 70 (1998) 2311.
60 P. Konieczka, L. Wolska, Ei Luboch, J. Namiesnik, A. Przyjazny and J. Biernat, J.
Chromatogr.A, 742 (1996) 175.
61 E.D. Conte and D.W. Miller, J. High Resolut. Chromatogr., 19 (1996) 294.
62 T. G6recki and J. Pawliszyn, J. High Resolut. Chromatogr., 18 (1995) 161.
63 T. Gorecki and J. Pawliszyn, Anal. Chem., 67 (1995) 3265.
64 J. Chen and J. Pawliszyn, Anal. Chem., 67 (1995) 2530.
65 Y. Hirata and J. Pawliszyn, J. Microcolumn Sep., 6 (1994) 443.
66 R. Eisert and J. Pawliszyn, Anal. Chem., 69 (1997) 3140.
67 R. Eisert and J. Pawliszyn, Crit. Rev. Anal. Chem., 27 (1997) 103.
68 Y. Gou, C. Tragas, H. Lord and J. Pawliszyn, J. Microcol. Sep., 12(3) (2000) 125.
69 H. Kataoka, H. Lord and J. Pawliszyn, J. Chromatogr. B, 731 (1999) 353.
70 H. Kataoka, H. Lord and J. Pawliszyn, J. Anal. Toxicol., 24 (2000) 257.
71 H. Kataoka and J. Pawliszyn, Chromatographia,50 (9/10) (1999) 532.
72 T. McDonnell and J. Pawliszyn, Anal. Chem., 63 (1991) 1884.
73 C-W. Whang and J. Pawliszyn, Anal. Commun., 35 (1998) 353.
74 B.L. Wittkamp and D.C. Tilotta, Anal. Chem., 67 (1995) 600.
75 G.L. Klunder and R.E. Russo, Anal. Chem., 49 (1995) 379.
76 H.M. Yan, G. Kraus and G. Gauglitz, Anal. Chim. Acta, 312 (1995) 1.
77 J. Pawliszyn, Device and Process for Increasing Analyte Concentration in a Sorbent,
U.S. Patent 5,496,741.
78 William R. Heineman, University of Cincinnati, USA, personal communication.
79 C. Arthur, L. Killam, K. Buchholz, J. Berg and J. Pawliszyn, Anal. Chem., 64 (1992)
1960.
80 T. Gorecki and J. Pawliszyn, Field Anal. Chem. Technol., 1(5) (1997) 277-284.
81 J.A. Koziel, M. Jia, A. Khaled, J. Noah, and J. Pawliszyn, Anal. Chim. Acta, 400
(1999) 153.
82 J. Koziel, B. Shurmer and J. Pawliszyn J. High Resol. Chromatogr., 23 (2000) 343.
83 R. Eisert and K. Levsen, J. Chromatogr.A, 737 (1996) 59.
84 J. Koziel, M. Odziemkowski and J. Pawliszyn, Anal. Chem., 73 (2001) 47.
85 X. Liu, J. Wu and J. Pawliszyn, Analyst, in preparation.
86 H. Tong, N. Sze, B. Thomson, S. Nacson and J. Pawliszyn, Analyst, in preparation.
87 P. Martos, A. Saraullo and J. Pawliszyn, Anal. Chem., 69 (1997) 402.
88 P. Martos, A. Saraullo and J. Pawliszyn, Anal. Chem., 69 (1997) 1992.
89 R. Schwarzenbach, P. Gschwend and D. Imboden, Environmental Organic
Chemistry. Wiley, New York, 1993, pp. 109-123.
90 V. Mani (Supelco), private communication.
91 K. Buchholz and J. Pawliszyn, Environ. Sci. Technol., 27 (1993) 2844.
92 T. G6recki, P. Martos and J. Pawliszyn, Anal. Chem., 70 (1998) 19.
93 C. Jia, Y. Luo and J. Pawliszyn, J. Microcolumn. Sep., 10 (1998)167.
94 L. Pan and J. Pawliszyn, Anal. Chem., 69 (1997) 196.
95 Y. Cai and J. Bayona, J. Chromatogr., 696 (1997) 113.
96 J. Poerschmann, F-D Kopinke and J. Pawliszyn, Environ. Sci. Technol., 31 (1997)
3629.
97 L. Pan, M. Adams and J. Pawliszyn, Anal. Chem., 67(1995) 4396.
476
98 J. Chen and J. Pawliszyn, Anal. Chem., 67(1995) 2530.
99 H. Daimon and J. Pawliszyn, Anal. Comm., 33 (1996) 365.
100 T. G6recki and J. Pawliszyn, Analyst, 122 (1997) 1079.
101 M. Chai and J. Pawliszyn, Environ. Sci. Technol., 29 (1995) 693.
102 L. Wang, Determination of Amphetamine and Methamphetamine in Urine and
Blood Plasma by Headspace Solid Phase Microextraction, M.Sc. Thesis, University of
Waterloo, Waterloo, Canada, 1996.
103 J-L. Liao, C-M Zeng, S. Hjerten and J. Pawliszyn, J. Microcol. Sep., 8 (1996) 1.
104 K. Buchholz and J. Pawliszyn, Anal. Chem., 66 (1994) 160.
105 H. Daimon and J. Pawliszyn, Anal. Corn., 33 (1996) 421.
106 Z. Zhang, J. Poerschmann and J. Pawliszyn, Anal. Corn., 33 (1996) 129.
107 B. MacGillivray, P. Fowlie, C. Sagara and J. Pawliszyn, J. Chromatogr. Sci., 32
(1994) 317.
108 T. G6recki and J. Pawliszyn, Analyst, 121(1996)1381.
109 J. Pawliszyn (Ed.), Applications of Solid Phase Microextraction. Royal Society of
Chemistry, Cambridge, UK, 1999.
110 P. Martos and J. Pawliszyn, Anal. Chem., 69(1997) 206.
111 M. Odziemkowski, J. Koziel, D. Irish and J. Pawliszyn, Anal. Chem., 73 (2001) 3131.
112 P. Martos and J. PawliszynAnal. Chem., 70(1998) 2311.
113 L. Muller, T. Gorecki and J. Pawliszyn, Fresenius J. Anal. Chem., 364 (1999) 610.
114 J. Dean, W. Tomlinson, V. Makovskaya, R. Cumming, M. Hetheridge and M. Comber,
Anal. Chem., 68 (1996)130.
115. C. Arthur, L. Killam, K. Buchholz and J. Pawliszyn, Anal. Chem., 64 (1992) 1960.
116 A. Boyd-Boland and J. Pawliszyn, J. Chromatogr., 704 (1995)163.
117 B. MacGillivray, Analysis of Substituted Benzenes in Environmental Samples by
Headspace Solid Phase Microextraction, M.Sc. Thesis, University of Waterloo,
Waterloo, 1996.
118 A. Saraullo, Determination of Petroleum Hydrocarbons in the Environment by
SPME, M.Sc. Thesis, University of Waterloo, 1996.
119 Z. Zhang and J. Pawliszyn, J. Phys. Chem., 100 (1996) 17648.
120 P. Popp, A. Paschke, U. Schroter and G. Oppermann, ChemiaAnalityczna (Warsaw),
40 (1995) 897.
121 J. Poerschmann, Z. Zhang, F. Kopinke and J. Pawliszyn, Anal. Chem., 69 (1997) 597.
122 H. Yuan and J. Pawliszyn, Anal. Chem., 73 (2001) 4410.
123 H. Yuan, Bioapplications of Solid Phase Microextraction, Ph.D. Thesis, University of
Waterloo, Waterloo, 2000.
124 J. Pawliszyn, Solid PhaseMicroextraction: Theory and Practice.Wiley, NY, 1997, pp.
185-189.
125 B. Smith, C. Zinni, E. Caramdo, J. Pawliszyn, M. Tyler andY. Hayasaka, Chem. Ecol.,
17 (2000) 215.
126 H. Lord, M. Moeder and J. Pawliszyn, Anal. Chem., in preparation.
127 A. So, N. Sze, A. Namera and J. Pawliszyn, J. Chromatogr.(2002) in press.
477
Chapter 14
Membrane extraction
Janusz Pawliszyn
The topic of membrane separation is very broad since it covers both basic mate-
rial science as well as numerous engineering implementations of the technology.
However, analytical chemists familiar with the theory of separation methods
know fundamental physicochemical principles of membrane separations. Ana-
lytical scale separations, such as solvent extraction or size-exclusion chromatog-
raphy, are typically performed as batch processes. Membrane implementation of
the same fundamental principles facilitates continuous separation. The optimi-
sation of mass transfer in membrane separations often differs compared to cor-
responding batch processes because of different geometric arrangements
implemented to facilitate continuous operation. It is not surprising that most
developments and commercial applications have been conducted by engineers
focused on large-scale continuous separation of products in process streams.
Analytical chemists frequently adopted new membrane materials to develop
their applications. Numerous books and reviews have been published on the
topic of membrane separations, principally by engineers [1-3] but more recently
by analytical chemists as well [4-6]. The objective of this chapter is briefly to
introduce different membrane separation technologies, and then give a funda-
mental and practical discussion of membrane extraction since they are in princi-
ple more selective and therefore more interesting for analytical applications. In
this chapter, the focus is a discussion on polymeric extraction membranes and in
particular their applications in gas chromatography.
A membrane is a selective barrier between two phases. Analytes are trans-
ferred from the sample (donor phase) through the membrane into the acceptor
(or stripping) phase. An essential characteristic of membrane-based separation
processes is the ease with which they can be adopted for continuous operation.
This feature makes membranes very attractive sample preparation tools since
their application allows conversion of the analytical separation and detection
instruments into sensor-like devices suitable for monitoring operations. It is
expected that chemical monitors manufactured using micro-machining technol-
ogies and incorporating membranes will be used directly on-site, similarly to pH
and other selective membrane electrodes presently used.
ComprehensiveAnalyticalChemistry XXXVII
J. Pawliszyn (Ed.)
( 2002 Elsevier Science B.V. All rights reserved
A SapeEffluent containing concentrated
Sample C0 * . 0 species above cut-off
B
Sample [Aj1 Purified effluent
Fig. 14.1. Schematic representation of separations in (A) size-exclusion and (B) extraction
membranes.
There are two basic types of membrane separations: one involves the mech-
anical separation of sample components based on size exclusion (Fig. 14.1A). A
wide variety of membrane materials can be used, each with its own advantages
and disadvantages [7]. In many cases, a membrane is a porous network of a syn-
thetic polymer, such as polypropylene, polysulfone or a cellulose derivative. Sep-
aration with these membranes is based on size-exclusion principles: sufficiently
small molecules can permeate through the pores, whereas larger ones cannot. A
membrane is usually characterised by its molecular weight cut-off (MWCO)
value. This value is defined as the molecular mass of the smallest compound that
is retained by the membrane to a larger extent than 90%. A usual way of classify-
ing size-exclusion membrane separation processes is by means of the driving
force applied. The filtration process is achieved by applying a pressure differ-
ence; dialysis is driven by a concentration difference and electrodialysis by an
electric potential difference coupled with a concentration gradient.
The other type of membrane separation (see Fig. 14.1B) involves a chemical
interaction between the membrane material and the analytes resulting in parti-
tioning of target components between the membrane material and the sample
matrix. These extraction-type membranes can be homogeneous liquids or solids
with microporous structures of different selective characteristics (e.g., affinity
membranes [8]). As the pore size of the porous membrane decreases, the selec-
tivity of transport becomes defined by the sorption characteristics of the mem-
brane material rather than their porosity. In fact, liquid membranes can be
viewed as having molecular size channels. These nonporous membranes can be
composed of a solvent or polymeric liquid film (for example silicone), into which
a molecule must actually dissolve in order to be able to pass through. In this case,
the efficiency of membrane transport for a particular compound is largely
dependent on its partition coefficient between the different parts of a membrane
separation system. Only compounds that are easily extracted from the donor
phase into the membrane and, in addition, easily extracted from the membrane
into the acceptor phase will be transported. The separation of different com-
pounds is based on the same principle as liquid extraction followed by a back-
extraction. Therefore, molecules with different physicochemical properties can
480
I Membrane Separation Methods i
Po ymeric Membranes
be separated even if they are of equal size. In selecting a material, the objective
for a size-exclusion membrane is to minimise chemical interaction between the
membrane and the sample matrix as well as the analytes, while in extraction
type membranes, the basis of separation is defined by the way different sample
components chemically interact with the material. The classification of mem-
brane separation techniques discussed in this chapter is shown in Fig. 14.2.
There are two basic ways membrane extraction can be implemented practi-
cally. One involves the use of flat sheet membranes. This configuration requires
the extraction module to support the membrane and to define the access of the
donor phase (typically a sample) and the acceptor phase to the membrane. Fig-
ure 14.3a gives an example of such a module used in Membrane Extraction with
Sorbent Interface (MESI), which is discussed in detail later. It has been designed
to expose one side of the membrane to the sample and the other side to the gas
that strips analytes that permeate through the membrane. The other implemen-
tation is the use of a hollow-fibre membrane which is, in principle, a piece of
microtubing made of suitable material with the desired permeation characteris-
tics. This structure is naturally self-supporting. Figure 14.3b illustrates the
implementation of this geometry as an MESI sampler. It is a very simple struc-
ture consisting of a hollow-fibre membrane and two connecting tubings, where
the stripping gas flows through the centre of the membrane and the whole out-
side surface area of the membrane is exposed to the sample. In addition to the
simplicity of making hollow-fibre modules, they are characterised by a high sur-
face area to volume ratio, resulting in faster mass transfer compared to a flat
sheet format for the same thickness. The major drawback is the need to use
thicker membranes to facilitate a self-supporting structure. They can only with-
stand a small pressure difference across the membranes. Most common mem-
brane materials such as silicone are available commercially in both flat sheet and
hollow-fibre formats.
It is possible to consider three basic types of mass transfer through mem-
branes: passive, facilitated and active [9]. In the first, most commonly used in
analytical applications, the membrane acts as a barrier through which the com-
ponents are transported under the influence of a gradient in their chemical
potential. In facilitated transport, together with a gradient in the chemical
481
(a) (b)
Bolt Stripping fluid in , ' Stripping fluid out
Supporting _ - ~
Connection
Wire rmesh tubing
G Z Z , J sheet
Fiat
membrane
Membrane
Teflon
__ __ =~washer
I I L Bottom
rig-~~~ ~plate
Threaded hole
Carrier gas in Carrier gas out
Fig. 14.3. Membrane module designs for (a) flat sheet membrane and (b) hollow-fibre membrane.
14.2.1 Filtration
Jv = -HpA(dP/dx) (14.1)
which defines volume flux Jv associated with pressure gradient dP/dx across the
membrane, where A is surface area of the membrane and Hp is hydrodynamic
permeability defined by the sample viscosity and the resistance of the mem-
brane, which depends on pore size. The influence of pore size on volume flow is
very substantial; for example, the flux obtained for pure water increases about
25 times by an increase in the MWCO value from 10 to 54 kDa, at a constant
membrane area and thickness [10].
482
Filtration offers some advantages over other size-exclusion approaches such
as a higher speed of separation and analyte recovery. However, it has one major
disadvantage associated with concentration polarisation, which arises due to sol-
ute concentration next to the membrane. The resistance to mass transfer is not
only caused by the membrane itself, but also by a layer which is formed on the
membrane surface by the accumulation of compounds which cannot pass
through the membrane pores. The contribution of this so-called concentration
polarisation layer is dependent on the applied pressure: at low pressures the con-
centration polarisation resistance is negligible (which indicates that no build-up
of retained compounds occurs), whereas at high pressures it accounts for a major
part of the total resistance. For an efficient and constant flux through the mem-
brane it is essential that the concentration polarisation layer be removed from
the membrane as much as possible. This can be achieved by applying a stirring
action during filtration on the sample side of the membrane. More frequently,
however, cross-flow filtration is used. In this case, the sample is pumped through
a flow channel or a hollow fibre and the sample flow itself removes accumulated
compounds from the membrane. Turbulent rather than laminar sample flow
along the membrane is more efficient in removing the polarisation effect. Lami-
nar flows result in zero velocity at the membrane surface, whereas turbulent
flows have nonzero velocities, which facilitate the removal of the accumulated
layer of sample components. In practice, increasing the sample flow-rate and
increasing the diameter of the channel facilitate the generation of turbulent
flow.
Practical filtration is achieved by placing a sample on one side of the
membrane and applying a pressure difference to drive all molecules of appropri-
ate size including the solvent, through the membrane pores to the other side. In
traditional off-line applications with disposable membranes, the driving force is
often applied by a vacuum or a centrifugal force. In on-line filtration the sample
is pumped on the feed side (the name for the donor side in the filtration
processes) of the membrane and a pressure is created by restricting the outlet
tubing of the filtration unit. In analytical applications, the filtrate flow thus
obtained is typically directed to an injection loop, from where it is introduced
into an analytical system. Higher filtration speeds can be obtained by increasing
the membrane surface area, which is most conveniently achieved by using
hollow-fibre membranes instead of planar ones.
14.2.2 Dialysis
In dialysis a sample is placed on the donor side of a porous membrane and analytes
diffuse through the membrane pores to the acceptor side as the result of a concen-
tration gradient dC/dx. The mass flux, Jm in this case is defined by Fick's law:
483
sample viscosity, the temperature and the molecular dimensions of the analyte
in comparison to the membrane pore size. Equation (14.2) indicates that flux is
dependent upon the surface area and the thickness of the membrane as well as
the diffusion coefficient. In order to maximise the flux and thus obtain a high
analyte recovery, these parameters as well as sample viscosity (because it affects
the diffusion coefficient) should be optimised.
To increase the membrane area, hollow-fibre membranes can be used
instead of planar ones. Typically, both ends of a bundle of hollow fibres are glued
into a fitting and the bundle is immersed into the sample, while the acceptor
phase is present inside the fibres. Since the area available for diffusion is much
larger for hollow-fibre than for planar membranes, the number of analyte mole-
cules recovered per unit time is much higher, which leads to distinctly improved
detection limits. Another important parameter with a distinct influence on the
flux is the membrane pore size or, since pores of varying size exist within a mem-
brane, rather the pore-size distribution. For each application, a proper MWCO
value should be selected so that the interfering material is sufficiently retained,
but at the same time rapid transport of analyte molecules is ensured. In many
cases on-line dialysis is applied to samples of biological origin and the major aim
is to retain proteins and other large biopolymers in order to protect the separa-
tion system designed to separate small molecules. For these applications, mem-
branes with MWCO values of 10-15 kDa are almost exclusively used, since they
have been proven to efficiently remove interfering bio-macromolecules. These
membranes are very rugged and can be used many times before a degradation in
analytical performance. If smaller compounds are to be removed, such as humic
substances from environmental samples, smaller pores are more appropriate
It is very important to use a properly constructed dialysis block and to select
a suitable membrane. It has been shown that analyte recovery can be improved
not only by decreasing the membrane thickness, but also by decreasing the
depth of the donor channel 11]. This indicates that mass transfer resistance not
only takes place in the membrane but also in the boundary layer on the donor
phase side in typical flow conditions. It is crucial to keep the concentration
gradient across the membrane as high as possible. If both the donor and acceptor
phases are kept stagnant, as it is in equilibrium dialysis, then after a while the
analyte concentration gradients extend into these phases resulting in reduced
flux and longer separation times. Maximum recovery of 50% in equilibrium
dialysis can be obtained (for equal donor and acceptor volumes). To improve
recovery, the concentration gradient must be kept higher and this can be
achieved by removing the analytes from the acceptor channel using an acceptor
phase that moves either continuously or in pulses. This continuous dialysis
technique, in principle, allows the quantitative transfer of an analyte from the
sample to the analytical system. However, the moving acceptor phase gives rise
to dilution, which must be overcome by reconcentration of the analytes for trace
analysis. The efficiency of dialysis is decreased when analytes bind to the
membrane material. Both electrostatic and hydrophobic interactions are
484
possible. Low surfactant concentrations, present in the sample below the critical
micelle concentration, decreases the effect.
The application of a dialysis membrane is very attractive in combination
with micro separation technologies. This approach has been used for on-line
extraction of living systems [12]. The coupling of dialysis membranes with CE
was proposed recently [13]. The sample is pumped through the outside of a
hollow-fibre dialysis membrane that connects two sections of a capillary and,
after diffusion of the analytes to the interior, an appropriate voltage is applied to
introduce them into the electrophoresis system. The dialysis membrane allows
not only the coupling of the micro-separation instrument to the investigated
system [14], but can also constitute an interface between two orthogonal
separation techniques resulting in two-dimensional separation [15].
14.2.3 Electrodialysis
485
supported liquid can be used as the membrane material and although both
options have the same general characteristics, SLM offers more flexibility
because selectivity of the extraction can be adjusted by selecting an optimal
solvent. In both cases, separation is accomplished in two consecutive extraction
processes between the membrane and the donor phase and between the acceptor
phase and the membrane.
The main advantage of liquid homogeneous membranes, as compared with
polymeric membranes, is the greater transport velocity through them, which is
due to the greater diffusivity of species in a liquid medium. Additionally, it is
easier to incorporate carriers in these membranes with a view to selectively
increasing the permeability of certain species, giving rise to facilitated or coupled
transport processes. However, the membrane lifetime is usually longer for
polymeric membranes. Liquid membranes tend to have solvent leakage, which
becomes more pronounced as the polarity of the solvent increases.
Frequently, the liquid membrane is used to separate two aqueous phases and
the pH of the sample in the donor channels are adjusted to such a value that the
analytes of interest are uncharged and easily extracted into the membrane liquid
or the silicone polymer film. The acceptor phase has the proper pH to effect
ionisation of the analytes immediately after passing the membrane, which
implies that they cannot be back-extracted into the membrane and are trapped
in the acceptor channel. The fundamentals and applications of these and other
similar approaches are presented in Chapter 15.
14.3.1 Fundamentals
486
Membrane Stripping Membrane Sample
z C-4foS---..
----
Be
La
'E
a
o
1
-_ - ~
I ---- I
away from the membrane surface by the stripping phase. There are two basic
approaches to membrane extraction. The first is based on quantitative recovery
of analytes and the other involves steady state mass transfer. The exhaustive
approach results in easier quantification since all analytes are removed from a
well defined volume of the sample, while steady state conditions result in higher
sensitivities since the mass transport through the membrane is maximised. In
the next two sections, the impact of different parameters on the extraction per-
formance for each approach is discussed theoretically. For the purpose of this
discussion, we will assume that hollow-fibre geometry is used and the stripping
phase is a gas, but the general conclusion from this discussion can be extended to
other geometries and stripping phases.
Figure 14.5 shows an example of a module which can be used for quantitative
removal of analytes from a sample. It is constructed of a hollow fibre positioned
in the centre of a glass tube and extending beyond the tube. Glue is used to fix
the membrane in a position and to connect the two tubings. Typically, a sample
is introduced through either end of the hollow-fibre membrane. The stripping
fluid is delivered to the module using the tubings and flows around the fibre.
When the flows are appropriate, the exhaustive transfer of analytes from sample
to gas phase occurs. The value of suitable flows can be found by solving the
appropriate differential equation which describes the mass transfer in the
system illustrated schematically in Fig. 14.6. The aqueous (or gas) sample passes
through the centre of the fibre and the stripping gas phase flows in a counter-
current direction. The appropriate mathematical solution to the equation is
, .......
~. inlet
!Gas
Sample inlet
Glass tube
Fig. 14.5. Design of the membrane extraction module for exhaustive extraction.
487
Gas flow
Membrane wall b
Fig. 14.6. Schematic of the hollow-fibre membrane
Gas flow
extraction system. Gas flow
Fig. 14.7. Dimensionless plot showing analyte concentration distribution inside the hollow-fibre
membrane.
488
100
90
7 80
t 70
.) 60
60
= so
. 40 - Analycal aoltion
,30 U trichloroethene
V 20 0 tetrachloroethene
Fig. 14.8. Comparison of theoretical W A 1,1,l1trichloroethan
modelling and experimental 10
extraction efficiency data for flow ° I 40 80 120 160
through module system. Volumetric flow [pUminl
cross-sectional area. The only way to improve mass transfer is to generate a tur-
bulent component in the sample flow through the hollow fibre, for example, by
coiling the membrane. However, it is very difficult to generate substantial tur-
bulence in such a small diameter tube. The above discussion clearly demon-
strates that, although it is feasible to reach exhaustive extraction, the sensitivity
obtained in the process is low since the amount of sample extracted under the
exhaustive conditions is very low. It is possible to increase the time of extraction
and/or use multi-fibre modules to obtain higher sensitivities, but both
approaches lead to the greater use of the stripping phase and difficulties in refo-
cusing analytes. This exhaustive approach is more effective when extracting gas
samples, since the diffusion coefficient in gas is higher compared to liquid, how-
ever, good metering pumps will always be required since the extraction recovery
is dependent on the flow conditions.
489
a: model prediction
b: expermental result
0.2
Fig. 14.9. Comparison of theoretical and experi- ction time profi (benzene)
mental transients and steady state extraction rate X
for direct exposed membrane configuration 0 50 100 150 200 250 300
system. ime (sec)
Z ACsDKes,,t (14.3)
ADe/e5 +2Rln R+b
Q R
where R and b are the inner radius and the thickness of the hollow-fibre
membrane, respectively; A is the membrane inner surface area; Cs is the analyte
concentration in the sample; De is the diffusion coefficient in the membrane; fis
the ratio of the average concentration of the stripping gas and the analyte
concentration in the stripping gas at the exit of the membrane (constant at
steady-state conditions); Ke, is the membrane/air partition coefficient; Q is the
stripping gas flow rate; and t is the trapping time.
This equation includes analyte and membrane physico-chemical properties
reflected in the distribution constant between the membrane and gas phase and
the diffusion coefficient of analyte molecules in the membrane. Equation (14.3)
illustrates that steady-state transport can be calculated for given geometric
configurations and experimental conditions as long as K and D are known. These
values can be stored in the computer or calculated based on a theoretical formula
or even determined experimentally as illustrated for the MESI system discussed
latter. Equation (14.3) also indicates that several experimental parameters
affect the extraction efficiency. A longer membrane probe leads to a higher
extraction rate if the stripping gas flow rate is high enough. Membrane thickness
is another factor affecting extraction rate and the membrane response time for a
given analyte. A thinner walled membrane would be beneficial for this applica-
tion. Temperature is an important factor in the extraction, affecting both the
490
distribution constant and the diffusion coefficient of an analyte. The effects are
opposite for these two parameters. A relatively low extraction temperature
results in a higher extraction rate.
In summary, the mass transfer in the open configuration illustrated in Fig.
14.3b is much faster than in the closed module design from Fig. 14.5 since
efficient agitation of the sample can be easily generated. The calibration of the
exhaustive extraction method (Fig. 14.5) is simpler since it requires only infor-
mation about the volume of sample passing through the module; however, the
use of Kes and De as a basis for calibration of the open configuration (Fig. 14.3b)
eliminates the need for external calibration of this system, making it as attrac-
tive. The temperature should be monitored during the experiment and results
corrected for its change.
The calibration approach described above for well agitated clean homoge-
neous samples can also be extended to quantification of volatile analytes in more
complex aqueous and even solid matrixes by the headspace membrane arrange-
ment, as long as the gas/matrix distribution constant (Henry constant for aque-
ous and solid samples) is known. This approach to calibration is analogous to one
recently described for the SPME technique (see Eqs. 13.7 and 13.18) [20]. Practi-
cally, the headspace membrane extraction arrangement can be implemented by
using the "cap" sampling system (see Fig. 14.10) [21]. To increase mass transfer
through the headspace created by the cap, helium can be injected into the cap.
Also, a micro-fan can be installed in the cap to enhance the mass transfer of
analytes through the headspace. The extraction cap can be placed at the water
surface or at a required depth to perform on-site or on-line monitoring. In this
arrangement headspace extraction is used and the membrane does not come in
contact with the sample matrix directly. This reduces the possibility of mem-
brane contamination and mechanical damage, which prevents a decrease of
membrane performance with time. This ensures that the extraction probe can be
placed in the sampling environment for long-term monitoring without the need
for replacement. This technique can also be extended to solid sample analysis by
covering for example soil, by the "cap". Figure 14.11 shows the range of linear
hydrocarbons that are able to permeate through a 0.15-mm thick PDMS hol-
low-fibre membrane. As expected, the relative permeation rate initially
increases with the increase in the molecular weight of the compound due to
increase in the sample matrix-PDMS distribution constant. The diffusion coeffi-
cient is at the same time decreasing, but this effect is much smaller. However,
above C-12 there is a dramatic decrease in the amount of extracted analytes
since the stripping of semivolatile compounds is not very efficient at room tem-
perature. The headspace approach to membrane extraction can be extended to
less-volatile analytes by heating the sample, the membrane, and the transfer
lines. By increasing the membrane temperature from room temperature to 80°C
hydrocarbons up to C-16 can be transmitted through the same membrane. An
alternative approach is to periodically heat the membranes since this process
cleans the system and produces concentration injection pulses of the semiv-
491
I-
mbrane
ule .>
8
6 7 8 9 10 11 12 13 14 15 16
arhn nllmhr
492
14.4 MEMBRANE EXTRACTION WITH SORBENT INTERFACE (MESI)
The MESI approach to membrane extraction was developed to address the need
for enhanced sensitivity in membrane extraction and simple continuous opera-
tion. To efficiently transfer analytes through the membrane, the stripping gas
flows should be high to prevent the accumulation of analytes in the membrane.
This process results in the dilution of analytes. As Fig. 9.14 (in Chapter 9) indi-
cates, the concentration of analytes in the stripping gas is low at the optimum
permeation rates. The use of high sensitivity detectors, such as mass spectrome-
ters, allows effective on-line determination of the diluted sample components.
However, to achieve high sensitivities, refocusing of analytes prior to detection
would be beneficial. In addition, some analytical instruments, for example gas
chromatographs, require pulse sample introduction, which can be generated by
a sorbent interface present during desorption.
Figure 14.12 illustrates that the MESI process consists of two steps: select-
ive permeation and trapping of extracted analytes by the sorbent, and then
desorption of accumulated components from the trap into an analytical instru-
ment. In the trapping mode, the sorbent is kept at room temperature, or if a
cooler is used the temperature can be considerably lowered to increase the trap
capacity. The power supply is off, and the carrier gas is free of analytes after
passing over the sorbent. Therefore, only pure gas enters the instrument.
During the desorption step, the sorbent is rapidly heated and the accumulated
analytes are released to enter the instrument as a high concentration narrow
pulse, which facilitates high sensitivity determinations. In addition, use of a
sorbent on line with membrane extraction eliminates the need for valves to
produce pulse injection into the analytical instrument.
The sensitivity of the MESI system in the trapping mode is directly related to
the trapping time-the longer the trapping time, the more analytes are accumu-
lated at the sorbent interface. Trapping time is limited by the breakthrough time
11
carrier
493
Sorbentrap
Fig. 14.14. Design of solid sorbent trap with direct heating. 1, Sorbent material; 2, quartz wool; 3,
gold-plated ferrule; 4, Valco connector; 5, stainless steel tubing connecting the membrane
module to the sorbent trap; 6, stainless steel tubing connecting the sorbent trap to the GC
column; 7, electrical connection to power supply.
494
an electrical pulse is applied to the trap, the gas inside expands very quickly and
the analytes can be pushed back partially towards the carrier gas lines instead of
going directly into the analytical instrument, which is placed at the opposite end
of the sorbent trap. The injection band would be broadened in this case, and the
separation would be worsened. To eliminate this effect, the stainless steel tubing
of very low i.d. is used to connect to the carrier gas lines at the entrance to the
trap producing flow restriction and, therefore, preventing back flush. With this
configuration, when a heating pulse is applied, the carrier gas, following the less
restricted path, flows mainly towards the instrument, moving the analytes in
this direction.
14.4.1 MESI-GC
MUemrane -Colurmn
t
probe
Sorbent
interface
495
Fig. 14.16. Chromatogram obtained
from the MESI system with prog-
ressively longer trapping times (30 s
increments). The bottom trace
corresponds to a trapping time of 1
min, and the top trace corresponds
to 3 min. Peaks: 1.benzene: 2 tolu- t
-U.U3m Pek 1 e . 1.0 L.U .3 . 3.0
ene; 3, m, p-xylenes 4, o-xylene. Minutes
self-supported they cannot be connected to the gas lines without the use of
special holders (see Fig. 14.3A).
Once the analytes cross the membrane, the carrier gas stream carries them
away to the sorbent interface, where concentration occurs. The sorbent interface
includes a sorbent trap, a heating means and a power supply as discussed above.
Figure 14.16 presents a set of five chromatograms obtained with the previ-
ously described MESI system, composed of the membrane module illustrated in
Fig. 14.3a and a trap as in Fig. 14.14, using progressively longer trapping times
(30, 60, 90, 120 and 180 s) for a dynamically generated standard gas mixture.
The bottom trace corresponds to a trapping time of 1 min, and the top trace to 3
min. The first peak in each group corresponds to traces of water and air, which to
some extend also permeate through the membrane and are trapped by the
sorbent. The remaining four peaks in each group correspond to benzene,
toluene, m,p-xylene and o-xylene (from left to right). It is clear from Fig. 14.15
that within the time frame examined, the response of the system was propor-
tional to the trapping time. For example, the benzene peak height for 1 min
trapping time was -0.5 V, while for 3 min trapping time it was 1.5 V. This proves
that no breakthrough of analytes occurred in the trap. Each individual separa-
tion was completed in -25 s, and the peak shapes were satisfactory. The width of
the benzene peak at half height was -250 ms, which was very good taking into
account that the desorption of the analytes was carried out without reversing
the direction of the carrier gas flow through the sorbent trap. This proves that
the MESI system in the configuration examined can be successfully used for
semi-continuous monitoring of trace levels of volatile analytes.
The most important aspect of MESI is its ability to selectively concentrate
volatile analytes on-line, before fast separation. This can result in dramatically
improved sensitivity, as illustrated in Fig. 14.17. The lower trace in this figure is
a MESI chromatogram obtained for a standard BTEX mixture and micro GC
equipped with a low sensitivity universal thermal conductivity detector (TCD
[31]. The trapping time was 1 min. The scale for this chromatogram is presented
496
A r
4.1
4.
5.0
3.
.5
2.5
2.0
1.5 0.0
Minutes
Fig. 14.17. Comparison of results obtained using the MESI system with those obtained using
regular gas-phase injection through a built-in injector. The lower chromatogram (lefty axis) has
been obtained using the MESI system; the upper trace (righty axis) is a chromatogram for direct
gas injection. BTEX concentration -4 ppm (v/v) of each compound. Peaks: 1, benzene; 2,
toluene; 3, ethylbenzene; 4, m,p-xylenes; 5, o-xylene.
on the lefty axis. The upper trace is a chromatogram obtained for the same mix-
ture with a regular micro-sampling loop injection, using the second GC module
in the same instrument. This chromatogram is presented in a 100 times larger
scale (see right y axis). It is clear from Fig. 14.17 that even with a precon-
centration time of only 1 min, the sensitivity of MESI was higher by more than
two orders of magnitude compared to direct injection of a gas sample. The use of
a selective membrane results in the elimination of a sloping baseline noticeable
in the upper trace associated with the presence of large amounts of oxygen and
moisture in the loop injection system. It has been demonstrated that use of
MESI coupled to a micro-GC allows trace analysis with a low sensitivity TCD
detector. For example, chloroform in drinking water present at low ppb level has
been detected by this system (see Ref. [31]).
Calibration is an important issue in MESI applications since it uses a
non-exhaustive extraction method. External calibration shows good precision
and wide linear range [32] in MESI similar as has been demonstrated before for
MIMS [33] systems. However, for field application, because of the demands of
fast, simple and accurate monitoring, traditional calibration methods such as
internal and external calibration are not appropriate choices, and in some cases
they are not applicable. Alternatively, an on-line calibration method can be
based on the mathematical model that was introduced in a previous section of
this chapter. Equation (14.3) can be rearranged to express the relationship
between the concentration in air, Cs, and other parameters:
C
c, Zt ( Qf 2Rln((R +b)IR)
ADK
(14.4)
(14.4)
497
In Eq. (14.4), R, b, A and Q are constant. The trapping time (t) can be
experimentally fixed. De and Kes are constant if the extraction temperature does
not change and assuming low sample concentration. Kes and De values are
available in the literature or can be measured on site. Parameter f is a constant
when extraction conditions such as stripping gas flow rate and extraction
temperature, are constant. Therefore, Eq. (14.4) can be simplified to:
Cs = BZ (14.5)
i( f 2Rln((R+b)/R
t Q ADKe
5
498
97
I-
25
___ _
32151.0
, 25720.8
= 19290.6
1~~~~~~~~
i
.6 12860.4
L
6430.2 I
Fig. 14.18. (A) Plot of the temperature profile during the experiment. (B) Monitoring chromato-
gram of BTEX extraction during the temperature change. 1, Benzene; 2, toluene; 3, ethyl-
benzene; 4, o-xylene.
involves pulse heating of the membrane and observing the transients generated
by the system.
Figure 14.19 shows a chromatogram obtained for benzene by membrane
pulse heating. The two highest peaks, 1 and 2, correspond to two heating pulses,
hence thermal desorption from the membrane. The two peaks are the same
height, which means that a constant amount was desorbed. The amount of
analyte desorbed from the membrane material into the stripping gas is repre-
sented by the difference between the intensity of peak 1 (or peak 2) and the
average intensity of the peak obtained at steady state. Curve a is a smooth plot of
each peak height from non-steady state to steady state. The non-steady state
transient obtained after a heating pulse, as shown in curve b, is identical to the
transient associated with curve a corresponding to the exposure of the
membrane to the sample. This transient can be used to calculate the diffusion
499
5:50 PM - 'JL 3:50 P '-God1 methanol
2 acetone
3:30 PM __J L L ' 3 propanol
4. butan-2-ol
5. 1I 1-trichloroethane
6 benzene
1:30 PM ul 7 toluene
11:50 AM
9:30 AM -
Fig. 14.20. Laboratory air monitoring with 0.5 1 1.5 2 2.5 3 3.5 4 4.5
MESI-GC. Time (min)
MESI-GC has been used for analysing volatile organic compounds in various
samples. Environmental monitoring has been demonstrated for tap water and
various types of air. Figure 14.20 illustrates the results of air quality monitoring
in a research laboratory. It is possible to notice an increase in air contamination
as a result of increased research activity during the day. In particular, the
presence of toluene is detected at 1:30 p.m. associated with the use of the Soxhlet
extraction system.
Monitoring of human breath by MESI could become a very powerful tech-
nique for clinical chemistry and toxicology since the last part of the breath (end
tidal volume) represents the headspace of blood [35]. Figure 14.21 illustrates the
acetone measured in breath over a 4.5 h period. During this time, the subject ate
various amounts of food. Acetone is used to monitor diets and diabetic patients.
---
OUu,
a 400
0
to
300
a 200
100
Fig. 14.21. Breath acetone
profile recorded over a 4.5 h
period with MESI-GC-Ion 0 20 40 90 180 210 240 270
Mobility Spectrometer. Time (min)
500
0
0)
C
C.
a
To
q
C
C)
a
Fig. 14.22. Breath ethanol
profile with MESI-GC after
application on the skin of 1 ml 5
of vodka. Time (min)
700
650
C-~----*
600
C 550
I~~~`-----~"-
2 500
t 450
o 400
Fig. 14.23. Monitoring of 35o
breath for alcohol during
application on the skin of 1 ml 3 20 40 60 80 100 120 140
of vodka every 5 min. Time (min)
The correlation between the level of the acetone in the breath and the amount of
food ingested is showed in Fig. 14.20. After lunch, the concentration of acetone
in the breath drops, and it begins increasing again after 4 h followed by another
drop after a snack. It can be seen that the concentration of acetone in breath is
directly related to the timing of food consumption.
Exposures monitoring with MESI system are illustrated in Figs. 14.22 and
14.23. In this experiment alcohol was used as a chemical marker to visualise the
capabilities of the MESI-GC approach for this application. 1 ml of vodka was
applied on the skin. The collection of the breath started 3 min after the applica-
tion. One breath sample was analysed each time, using a trapping time of 4 min.
It can be seen from the figure that the ethanol travelled rapidly through the body
to reach the breath. The profile is very sharp, with the concentration dropping
rapidly after the first sampling. Different vodka application time intervals were
used, and the concentration of ethanol was measured in the breath. It has been
found that by applying 1 ml of vodka every 5 min, the concentration of ethanol
maintains constant over a long period of time (see Fig. 14.23). Samples were
recorded every 15 min, using 4-min trapping times and one breath per sample.
The above examples illustrate the flexibility of the membrane-sorbent sys-
tem in combination with GC. This instrumental approach results in a single
device being able to perform sampling, sample preparation, separation and
quantification and therefore it enables GC to act as a sensor for environmental
501
and clinical monitoring. Such applications are expected to be used widely in the
future, when small size separation micro-machining devices become available.
REFERENCES
502
Chapter 15
15.1 INTRODUCTION
504
industrial separations, for example extraction of metal ions [15,16] and organic
acids [17,18] from wastewater. Also, extraction of large polyelectrolytes as
lignosulfonate [19] and proteins [20] was described.
There are a number of other ways for arranging a three-phase liquid
membrane extraction; e.g., as a bulk liquid between semipermeable membranes
or in completely unsupported forms. These types are commonly called bulk
liquid membranes (BLM). This principle leads to slow mass transfer and long
extraction times. Membrane extraction can also be arranged in the form of an
emulsion-an emulsion liquid membrane (ELM) [16,21]. The ELM is useful for
removal of compounds from the feed solution, but the quantitative recovery of
the extract is difficult. The techniques mentioned are therefore not used in
analytical chemistry, but extensively in industrial processes and in fundamental
membrane extraction studies.
A two-phase membrane extraction system with a porous membrane
(support) that separates one aqueous and one organic phase is called micro-
porous membrane liquid liquid extraction (MMLLE) [22]. In principle there are
two approaches: with a hydrophobic or a hydrophilic porous membrane. Using a
hydrophobic membrane, the organic liquid fills the pores and a direct contact
between the phases is obtained close to the surface of the membrane where the
interfacial mass transfer takes place. This is a liquid membrane system where
the liquid membrane is the organic liquid in the pores of the porous membrane
support, in analogy with SLM. The membrane could also be hydrophilic, which
would lead to an aqueous phase in the membrane pores, but this does not yet
seem to have been tried for analytical purposes.
In MMLLE, virtually the same extraction chemistry can be obtained as with
LLE and it can be seen as a way of performing LLE in a continuous and
instrumental way. There are other, more instrumentally complicated ways of
performing continuous LLE, as described in the literature [23]. These typically
involve mixing the organic and aqueous phases as segments in a flow system,
and then sequentially separating these phases again, the latter operation being
quite critical.
505
other hand, the membrane is virtually insoluble in most common solvents, so
any combination of aqueous and organic liquids can be used as the donor and
acceptor phases. Application of polymeric membranes has been described both
with an aqueous, trapping acceptor and with an organic solvent in the acceptor
channel [24-26]. The latter version is sometimes termed membrane-assisted
LLE [27], and is somewhat similar to MMLLE, with the additional feature that
the dissolution of the analytes into the membrane polymer will influence the
mass transfer, leading to slower extraction but a more stable system.
There are a large number of chemical principles that have been used for SLM
extraction of various classes of compounds. A number of such principles are
summarised in Table 15.1 and discussed in some detail below.
506
TABLE 15.1
Schematic overview of different chemical principles for extraction and trapping used for SLM
extraction. TOPO = trioctylphosphine oxide; DTPA = diethylenetriaminepentaacetic acid;
DEHPA = diethylhexylphosphoric acid.
Analytes Donor Membrane Acceptor Transported Trapping Ref.
species
Simple permeation
Acids acidic org (+TOPO) basic neutral anions 6,28
Bases basic org acidic neutral cations 6
Carrier-mediated transport
Metal ions 8-hydroxy- org DTPA complexes charged 29
quinoline complexes
Metal ions, acidic, DEHPA acidic, complexes counter- 30
Amino acids pH = 3 pH = 0 transport
of H+
Amino acids, basic trioctylmethyl- acidic, ion pairs cations 31-34
amino ammonium chloride
phosphonates
Amino acids basic Pd-complex acid complexes cations 35,36
Sugars, diol- neutral boronic acid acid covalently protonation 37,38
containing carrier bound of carrier
compounds complexes
Anionic acidic + org basic ion pairs deprotonation 39
surfactants amine of amine
carrier
Immunological trapping
Triazine neutral org atrazine permeation immunological 40
herbicides antibodies as antigen-
antibody
complex
507
MeL+- C (HD) 2 - FMe +
+
C
2 H' e MeD 2 - - 2H
Fig. 15.2. Membrane extraction schemes for of metal ions (Me). HQ = 8-Hydroxyquinoline,
DTPA 5- = diethylenetriaminepentaacetic acid anion, R4N+ = methyltrioctylammonium cation,
HD = diethylhexyl phosphoric acid. For a,b,c, see the text.
508
Donor phase Membrane Acceptor phase
-
A- R4NCI - HA
CI
C[ew RNA - IT H'
a
H+ H:AD(HD) 3 - EH
b
pH- 3 pH<
S-
RNHS- -RNH2
RNH;
OH
Fig. 15.3. Extraction schemes for membrane extraction of amino acids (HA) and anionic
surfactants (S-). RNH 2 = hexylamine. Other abbreviations as for Fig. 15.2.
509
100
90
k
80
70
60
E(%) 50
40
30
20
Fig. 15.4. Influence of TOPO 10
content in di-n-hexyl ether on 0
extraction efficiency E. (After
Shen et al. [28], with permission
from Elsevier Science B.V.) TOPO (%)in di-n-hexyl ether
with simple permeation. The addition of an aliphatic amine (e.g., hexyl amine) in
acidic donor solution permitted the formation of an ion pair that could readily be
extracted. As shown in Fig. 15.3c, a higher pH in the acceptor caused the amine
to be non-charged, leading to the destruction of the ion pair, so the surfactants
were efficiently trapped in an indirect way.
There is an example of covalent binding between the analyte and a mem-
brane additive, namely the extraction of sugar and other compounds with vicinal
hydroxyl groups by borate reagents. This principle has been studied by Smith
[37,38], but it seems not yet to have been applied to analytical problems.
A final example of a membrane-phase additive is TOPO (tri-octyl phosphine
oxide). An example of this is shown in Fig. 15.4, where extraction of carboxylic
acids of different polarities is strongly influenced by the contents of TOPO in the
membrane [28]. The most polar acid, lactic acid, was not extracted at all without
TOPO, but the extraction was significantly improved by the additive. Butanoic
acid, being the most hydrophobic acid in the study, was well extracted without
TOPO and essentially unaffected by its concentration. TOPO is known to form
hydrogen bonds, thereby improving the extraction of relatively polar
compounds.
510
.AC = YD CD- aA C (15.1)
where CD and CA are the concentrations in the donor and acceptor phase, respect-
ively; a, and aA are the fractions of the analytes that are in extractable
(uncharged) form in the actual phase. Typically, extraction conditions are such
that aD is close to 1 and aA is a very small value (corresponding to, e.g., an
alkaline donor and an acidic acceptor intended for extraction of bases, c.f. Fig.
15.1). CA is zero from the beginning of the extraction and increases during the
operation, usually to values well over CD. The maximum concentration enrich-
ment factor possible is reached when AC has eventually reached zero:
Ee(m) = (CA/CD)max = QD/(A (15.2)
The rate at which the conditions in Eq. (15.2) are approached depends on many
parameters, as detailed earlier [47] and briefly discussed in Section 15.3.4. There
are in principle two different cases: Membrane-controlledextraction and donor-
controlled extraction. If the extraction is membrane-controlled, the rate-limiting
step is the diffusion of the analyte compound through the membrane, which
usually leads to a slow extraction. On the other hand, with a donor-controlled
process the mass transfer rate is typically considerably larger. Here, it is limited
by diffusion in the donor phase, so the rate of mass transfer depends mainly on
the diffusion coefficient in the donor phase, D, and on the donor flow condi-
tions. If the partition coefficient K between the donor and membrane phases is
less than about 1, the extraction is typically membrane-controlled, while the
mass transfer is mainly donor-controlled when K > 10. Other than this, the
value of the partition coefficient has no large influence on the efficiency of
extraction although the rate of mass transfer increases slowly with K. As K is not
involved in Eq. (15.2), it does not influence the maximum possible enrichment
factor, but only the rate of mass transfer. This is in contrast to the conditions for
classical LLE.
Further, it seems that too large partition coefficients are not favourable [49],
probably as the transfer of analyte out of the membrane into the acceptor phase
may become less efficient. This influence of K is further discussed in Section
15.3.3.
The extraction efficiency E is usually expressed as the fraction of analyte
amount input to the system that is recovered in the acceptor:
E = nA/n, (15.3)
where n1 and nA are the number of moles input during the extraction time and
those collected in the acceptor, respectively. This parameter is analytically
important and it is not identical to the concept of recovery. An alternative to Eq.
(15.3) is the formula:
E' = (n,-n )/n, (15.4)
where nw is the number of moles leaving the donor channel. Equation (15.3) thus
tells how much of the input material is recovered in the acceptor, while Eq.
511
(15.4) measures how much is removed from the donor phase. We can define a
recovery R as:
R = E/E' (15.5)
If E = E', the recovery is 100% and no analyte is lost in the process. If E < E'
some analyte is either adsorbed in the apparatus or left in the membrane. This is
the memory effect, which is discussed in a number of cases [28,39,50], and
constitutes a limitation of SLM extraction. In practice, the problem can be
overcome by careful design of the experimental conditions. This is also related to
the adverse effects of too high partition coefficients, as mentioned above.
From Eqs. (15.2-15.3) it can be seen that for efficient SLM extraction, a neutral,
extractable species should be formed in the donor phase (or at the donor-
membrane interface) so aD is practically unity. Then, this species should be
moved though the membrane and into the acceptor phase, where it shall become
transformed to a nonextractable species, i.e. a, should be close to zero. This
transformation is termed trapping and it can be arranged in several chemical
ways.
512
constitutes an efficient trapping. The a,-values that are obtained are easily
calculated. For a monoprotonic basic analyte, the fraction aA of non-ionised
molecules in the acceptor is given by:
aA = Ka/([H +] + Ka) (15.6)
Ka is the dissociation constant of the corresponding ammonium ion. It was found
[47] that a sufficiently low value of cA in practice is 0.0005. With conditions
giving such a value, the extraction efficiency will be constant over a reasonable
period of time. This, together with Eq. (15.6), leads to the rule of thumb that the
acceptor pH should be 3.3 units below the pKa to obtain sufficient trapping of
amines.
Obviously, acidic compounds are trapped in an analogous way by a high
acceptor pH, and analogous equations can be set up for other acidic compounds.
This leads to the rule that the acceptor pH should be 3.3 units above the pKa of
acidic analytes, for the trapping to be satisfactory.
With acceptor conditions fulfilling these rules and with a donor pH leading to
a, = 1, Eq. (15.2) leads to a maximum enrichment factor of 2000 times.
15.3.2.2 Indirecttrapping
There are a number of SLM-systems where the transport is driven by principles
other than the simple direct trapping by forming a non-extractable charged spe-
cies as a result of a pH gradient or the addition of a reagent for the purpose
described above. In the application of DEHPA for extraction of various cationic
compounds (metal ions, amino acids, amines) (Figs 15.2c and 15.3b, the trans-
port is driven by a pH gradient, by providing an excess of protons in the acceptor
phase compared to the donor phase. The DEHPA anion is then protonated, so
the metal ion is released and the return of the carrier from the acceptor to the
donor is facilitated. This is an example of a counter-transport mechanism, and in
this case, the transported analyte is in the same form (in itself non-extractable)
on both sides of the membrane. A similar situation is found when anions are
extracted by means of Aliquat-336 (Figs. 13.2b and 13.3a). Here a gradient of
counter ions (chloride in the figures, but other ions could be more efficient) from
the acceptor to the donor phases, also in a counter-transport mechanism. In
those cases a direct trapping also helps to drive the mass transfer (by the employ-
ment of DTPA in Fig. 15.2b and protonation in Fig. 15.3a).
In the case of ion-pair transport as exemplified above for anionic surfactants
(Fig. 15.3c) [39], the ion-pair former (aliphatic) amine is neutralised so the
ion-pair is "killed" in the acceptor. In this case, the amine can in principle be
extracted back into the membrane and become "trapped" in the donor, but if this
happens in the actual system has not been investigated.
513
0.9
>, 0.8
the antigen can no longer re-dissolve in the organic membrane, so the analyte is
trapped in the acceptor. With a surplus of antibody, the concentration of analyte
in the acceptor phase will easily exceed the initial sample concentration. In a
preliminary work [40], the immunocomplex was quantified on-line, using a
fluorescein flow immunoassay in a sequential injection analysis (SIA) setup. The
outlined scheme was successfully applied for the extraction of 4-nitrophenol
(4-NP) from spiked water solutions as well as from a spiked wastewater sample,
indicating that the immunoextraction can be suitable when dealing with
difficult matrices.
514
according to Eq. (15.5). It is clear that the SLM technique is best applicable to
compounds with moderate hydrophobicity, i.e., log Kow less than 3-4. On the
other hand, it is possible to extract compounds with very low values for Ko. Most
hydrophilic compounds hitherto extracted with SLM seem to be some biogenic
amines that were successfully extracted by Romero et al. [45], using DEHPA as a
carrier (see Section 15.3.1.2). For these compounds like spermine, spermidine,
putrescine, etc. log K,,ow is in the range of -0.5 to -1.0. Also lactic acid (c.f. Fig.
15.4) is very hydrophilic with log Kow -- 0.6.
The extraction efficiency depends on the donor flow rate and other physical
parameters as the following equation (assuming complete trapping (A = 0) in a
stagnant acceptor) [47,53]:
E=l-·expch
E=l-expLh D
3D
3D K kl+
lnK1+ Kl+aDKkM
3
23DD
.DD+
6D 1
j3.
(15.7)
E =E EVA = E FD t/VA
JVA (15.8)
Vs and VA are the volumes of the extracted sample and the acceptor channel,
respectively, and t is the time of enrichment. When the donor flow rate is
increased, E decreases as shown above, but this is compensated for by the
increasing amount of analyte being input to the system, so the enrichment factor
typically increases with donor flow rate for a given time, as seen in Fig. 15.6. On
515
120
100
0.8
80
0.6
uJ 60
0.4
40
0.2
20
Fig. 15.6. Extraction efficiency, E and
anvi-1rnon+ Cantor Vx it ---r- a 0
e-lmlllmllL au-culrVne ttarllu-.ry unitl) a 0 0.05 0.10 0.15
functions of the reduced flow parameter
(see text). )
the other hand, a high flow rate will consume a large volume of sample, which is
usually no problem for extraction of environmental samples (e.g., river water) or
similar. If the available sample volume is limited (e.g., biological samples),
extraction could be better performed at low flow rates to maximise the
extraction efficiency.
With SLM extraction, very high enrichment factors can be potentially obtained,
considerably higher than with LLE. Generally, Eq. (15.2) describes the maxi-
mum possible enrichment factor. With a knowledge of the relevant pKa values,
Eqs. (15.2) and (15.6) easily give estimations of the theoretically obtainable
enrichment factors for acidic and basic compounds. For a number of weakly
basic compounds, such calculations have been performed and compared with
experimental data [48]. In Fig. 15.6 some results are seen. For aniline withpKa =
4.6, a value of aA = 2.3 10 4 is calculated at the experimental conditions (acceptor
pH = 1), leading to a maximum enrichment factor of about 4300 times (assuming
a, = 1). This value is not reached even after long extraction times (the experi-
ment was ended after 25 h of extraction and 6 1 of sample, giving a final enrich-
ment factor of about 2000 times and still increasing). The weaker base
3,5-dichloroaniline, pKa = 2.5, gives acA 3.10- 2 so the maximum enrichment fac-
tor is only 33, and a similar value is quickly reached after a short time of extrac-
tion. This situation can be improved by increasing the acid concentration in the
donor, as shown in the cited work. High acid concentrations in the acceptor
might lead to differences in the ionic strength between the donor and acceptor
phases. In that case, activity effects will have to be considered in connection with
Eq. (15.2), as described in the cited work.
For ultra trace analysis, very high enrichment factors can in principle be
obtained by realising very low values of aA by an efficient trapping reaction. It is
also convenient in analytical sample preparation that the extraction is linear, i.e.
that the extraction efficiency is independent of extraction time or volume, so the
516
amount of analyte collected in the acceptor is directly proportional to the
concentration in the extracted sample. This condition will be true if the enrich-
ment factor obtained is considerably less than the theoretical maximum enrich-
ment factor, which is the same as aAcA << ac,. This is the basis for the
recommendation [47] that the acceptor pH shall be 3.3 units below the pKa for
basic analytes and correspondingly for other types of analytes.
15.3.6 Selectivity
The term selectivity can mean two slightly different things in the present
context of sample preparation of small molecules in biological and environmen-
tal matrices. The first consideration is the discrimination between different
small molecules and the second is the discrimination between small and large
molecules, usually with the objective of removing the large molecules and
recovering the small ones. From the discussions above, it is obvious that there
are a large number of possibilities to tune the membrane transport processes so
a certain class of compounds is efficiently extracted while other classes are not
extracted to appreciable extents. It is also clear that for several reasons macro-
molecules are not usually extracted at all by SLM. Many biological macro-
molecules (proteins, peptides etc.) are charged at most conditions and thus not
extractable. Even if it is possible to design extraction systems that can extract
such compounds, they are totally excluded from systems designed to extract
small molecules. A similar situation exists with humic compounds in environ-
mental samples (e.g., water). Completely uncharged macromolecules exist, but
even so, these compounds have, by virtue of their size, transport properties that
make their transport through an SLM exceedingly slow. There are many
examples of analytical applications to biological and environmental systems,
which show nearly complete selectivity of both the above types. For example,
triazine herbicides were extracted from spiked river water both with SLM and
with SPE [54]. The results in Fig. 15.7 show that in the SLM extract virtually
only the expected compounds are found, while a typical "humic hump" from
humic acids together with disturbing signals from unknown compounds are seen
in the SPE extract.
250
200
150
517
4 50
a
400
7 350.
'
.6
300
2501 LI
e
- 200
150
°-I l
AV lj
5 _ a. Blood plasma
V
U ll I l l
0 2 4 6 8 o
200 b
80
1
160
140
l 20.
100 I Il
30
30
40
00 ____I__I
J
0 2 4 6 8 10 0 5 10 150 5 10 15
Time (min) Minutes Minutes
Right: Fig. 15.9. (a) Chromatograms of Amperozide (I), its metabolite (II) and homologue (III)
with the subsequent blank after enrichment from blood plasma. (b) Corresponding
chromatograms after enrichment from an aqueous buffer solution. Concentrations 4g/ml of I
and II, 8 ug/ml of III. (Reprinted from Lindeghrd et al. [55]. Copyright 1994 American Chemical
Society.)
Another example is the SLM extraction of a basic drug from blood plasma, as
shown in Fig. 15.8 [55]. Here, the chromatogram from spiked blood plasma is vir-
tually identical to that from a water solution of the actual drugs. A difference is
that the extraction efficiency from the blood plasma is lower than from water.
This is traced to the influence of drug-protein binding, and the phenomenon can
be used to study such binding constants, possibly including the corresponding
kinetics.
15.4.1 General
518
hydrophobic membranes are used: typically PTFE or polypropene membranes.
The acceptor phase is an organic solvent, also filling the pores of the hydrophobic
membrane. This forms an aqueous-organic two-phase system, with the organic
phase partly in the membrane pores and partly in the acceptor channel. MMLLE
extends the concept of membrane extraction, as it is best applicable to hydropho-
bic, preferably uncharged compounds. i.e. those that cannot easily be extracted
with SLM. The extract is typically organic, not aqueous as with SLM. Thus
MMLLE is more easily interfaced to gas chromatography and normal-phase
HPLC than SLM, which is better compatible with reversed-phase HPLC.
With SLM, the concept of trapping in a stagnant acceptor is central, but with
MMLLE and an organic acceptor, no trapping reactions are used. The maximum
extraction efficiency is directly given by the partition coefficient as in LLE, the
only driving force for the mass transfer being the attainment of distribution
equilibrium between the aqueous and organic phases. If the partition coefficient
is very high, it is possible to work with a stagnant acceptor and still get a
considerable enrichment into a very small extract volume. The extraction effi-
ciency will be higher if the hydrophobicity of the analyte is large; i.e. if the
partition coefficient is large. With smaller partition coefficients, the mass
transfer can be improved if the acceptor phase is continuously or intermittently
flowing, removing the extracted molecules successively from the acceptor and
maintaining the diffusion through the membrane.
There is a recent example of MMLLE with the phases reversed, i.e. extrac-
tion of phenolic compounds from an organic phase into an aqueous phase [56]. In
this case the phenols were trapped in a highly basic aqueous acceptor, which was
kept stagnant during the extraction.
519
E = 1/(1/K + FA/FD) (15.10)
With a stagnant acceptor, the maximum value of the enrichment factors is:
Ee(max) = K (15.11)
This equation for MMLLE should be compared to the corresponding equation
for SLM, Eq. (15.2), which does not contain the value of the partition efficiency
and the maximum value of the enrichment factor is determined by the trapping
conditions.
MMLLE is conceptually similar to continuous dialysis. Assuming equilib-
rium between the phases and with the value of K equal to unity, Eqs. (15.10) and
(15.11) are also valid for that technique. In these equations it is assumed that the
acceptor flow is parallel to the donor flow. By reversing the acceptor flow
(counter-current flow) the concentration gradient between the two phases will
be larger leading to more efficient mass transfer. The matter was discussed by
Valcdrcel [23, p. 66]. This situation is mathematically far more complex than
when parallel flow is assumed. For the corresponding problem in dialysis,
numerical solutions to the appropriate partial differential equation system have
been derived [57].
Some typical membrane extraction units or contactors are shown in Fig. 15.10.
They are made of two blocks of inert material with a machined groove in each.
After clamping the blocks together with a membrane between, flow-through
channels (donor and acceptor) are formed on each side of the membrane.
Channel volumes are typically in the range 10-1000 Al. Another type of mem-
brane unit is based on a hollow fibre membrane. The inside (lumen) of the fibre
contains the acceptor and the annular volume between the outside of the fibre
and the inside of a surrounding tube or cylindrical hole contains the donor. Such
(a) (b)
A
....... Membrane ~ _
Acceptor
Fig. 15.10. Membrane units for liq- channel | \
uid membrane extraction. (a) Flat
membrane module with 1 ml chan-
nel volume (A = blocks of inert
(c) Donor
material, B = membrane). (b) Flat inlet ~
membrane module with 10 Al chan- Acceptor
outlet
nel volume. (c) Hollow- fibre module Acceptor inlet -- IAcceor o
with 1.3 l acceptor channel t - Porous hollow
(lumen) volume. Donor outlet j fibre <2pl
520
(a) Peristaltic
Acid pump Mixing
coil
Waste
Sample Membrane
unit
(b)
Vast
Dot
pH 1
Mix
Wash
(C
Fig. 15.11. Apparatus for liquid membrane extraction. (a) Manual off-line instrument based on a
peristaltic pump. (b) Instrument with on-line connection to HPLC for environmental studies. (c)
Experimental set-up for SLM-HPLC determination of biomolecules in blood plasma or urine.
units can be made with channel volumes as small as 1 bl. [58]. These types of
membrane units are in principle applicable to all versions of membrane
extraction for analytical sample preparation or sampling. Membrane units can
be purchased from Global FIA (Gig Harbor, WA, USA) or other suppliers of FIA
equipment. They can also be prepared in-house or obtained from the Lund
University Workshop via the author of this chapter.
Simple and cheap membrane extraction flow systems for reasonably large
sample volumes can be built up around a peristaltic pump. An example of such a
system is seen in Fig. 15.11a. With this set-up, the acceptor phase is manually
removed by the use of a syringe after each extraction. Such systems have been
used both for laboratory work [50] and for sampling in natural waters [52].
521
15.5.2 On-line connection to chromatography
With large membrane units (channel volumes around 1 ml) direct transfer of the
acceptor phase to a HPLC can be arranged by means of a precolumn [59,60] in
order to inject as much as possible of the extracted analyte (see Fig. 15.11b).
Such systems can be completely or partially automated with pneumatically or
electrically actuated valves controlled by timers or by computer systems.
Smaller membrane channels can also be used. Then a heart-cut which contains a
major part of the extract can be accommodated in the injection loop for direct
injection into the HPLC column without a precolumn [25,61,62].
For smaller samples (<1 ml) the liquid delivery precision of peristaltic
pumps is not adequate. For such applications, membrane extraction equipment
based on syringe pumps connected to so called robotic liquid handlers have been
designed [55,63]. A typical example is shown in Fig. 15.11c. Here, a robotic nee-
dle connected to a syringe pump submits reagent to adjust the pH of samples in
vials, picks up an aliquot and passes it through the donor channel of a small
membrane unit (channel volume around 10 l). The entire extract collected in
the acceptor channel is transferred to an injection loop injector connected to the
HPLC system in the usual manner. Thus, the entire extract from, say, 1 ml sam-
ple ends up in one chromatographic injection. It is possible to arrange this appa-
ratus so that while one sample is chromatographed, the next sample is extracted,
so the cycle time of the system is determined by the chromatogram time.
For gas chromatography, the most suitable membrane extraction technique
is MMLLE. The organic acceptor is better compatible with GC than with LC, as
are the analytes that are best extracted in such a system, i.e., relatively hydro-
phobic compounds. One approach is that the organic acceptor can be injected
directly into a GC by means of a large-volume injector [22]. A new development
is the ESy instrument (Personal Chemistry AB, Uppsala) [64] where a MMLLE
extraction in microscale (extract volume ca 1 Il) is automatically performed and
the organic extract is directly injected into the GC by means of an injection
needle, directly connected to the extraction cell (see Fig. 15.12).
522
15.5.3 Off-line equipment
.DPE-Membrane
Analyte
Fig. 15.13. Off-line membrane extraction devices according to (a) Hauser and Popp [27], (b)
Mullins [65], and (c) Gr6nhaug et al. [71]. (With permission from the authors and Wiley VCH.)
523
15.6 APPLICATIONS
TABLE 15.2
Applications of membrane extraction to biological samples. Abbreviations: LPME2 = Liquid
phase microextraction, two-phase system, and LPME3 = Liquid phase microextraction, three-
phase system. Other abbreviations are found in the text or are generally used.
Analytes Matrices Membrane Analytical Ref.
technique technique
Aliphatic amines Urine SLM GC 72
Aliphatic acids Manure SLM GC 73
Aliphatic amines Blood plasma SLM GC 74
Amperozide Blood plasma SLM HPLC 55
Bambuterol Blood plasma SLM CE 58,75,
76
Bambuterol Blood plasma SLM LC-CE 77
Diprivan (Propofol) Urine SLM HPLC 78
Phenols Blood plasma SLM HPLC-biosensor 79
Lead Urine SLM AAS 42
Lead Urine SLM PSA 80
Local anaesthetics Blood plasma SLM GC 22
Local anaesthetics Blood plasma MMLLE GC 81
Amphetamine Blood plasma, urine LPME3 (LLLME) CE 66
Amhetamines, benzodia- Blood plasma, urine LPME2, LPME3 HPLC, GC, CE 68
zepines, Naproxen,
Citalopram
Benzodiazepines Blood plasma, urine LPME2 GC 69
Ibuprofen, naproxen, Urine LPME3 CE 67
ketoprofen
Metabolites of Ropivacaine Urine SLM HPLC 63
Citalopram and metabolites Blood plasma LPME3 CE 70
Amphetamines Blood, urine LPME3 FIA-MS/MS 71
Organo-phosphate esters Blood plasma MMLLE GC-MS 82
Sulfonylurea drugs Blood plasma PME (off-line) LC-MS/MS 65
Amphetamine Urine SLM HPLC 83
Benzimidazole anthelmintics Urine, tissue, milk SLM HPLC, LC-MS 84
524
15.6.2 Environmental analysis
TABLE 15.3
Applications of membrane extraction in environmental analysis. For abbreviations of membrane
techniques, see the text
Analytes Matrices Membrane Analytical Ref
technique technique
Phenoxy acids Natural water SLM HPLC 85
Chlorinated organics Water PME GC 86
Phenoxy acids Natural water SLM HPLC 87
Sulfonylurea herbicides Natural water SLM HPLC 88
Phenoxy acids Natural water SLM HPLC 52
Aliphatic amines Air (Impinger samples) SLM (gas) GC 11
Aliphatic amines Air (Impinger samples) SLM (gas) GC 89
Uncharged organics Water PME GC 26
Carboxylic acids Air (Impinger samples) SLM IC 90
Sulfonylurea herbicides Natural water SLM HPLC 59
Cu Natural water SLM AAS 91
Cu, Cd, Co, Ni, Zn Natural water SLM AAS 29
Phenolics Natural water SLM HPLC 60
Phenolics Nutrient solutions SLM HPLC 50
Carboxylic acids Soil liquid SLM IC 61
Carboxylic acids Soil liquid SLM IC 92
Triazine herbicides Natural water SLM HPLC 93
Triazine herbicides Natural water MMLLE FIA 94
Anionic surfactants Natural water SLM HPLC 39
Triazine herbicides Natural water SLM HPLC 95
Anilines Natural water SLM HPLC 96
Cu, Pb, Cd Natural water SLM AAS 97
Cu, Pb, Cd Natural water SLM AAS 41
Triazine herbicides Natural water SLM HPLC 98
Zn, Ni, Co, Mn Natural water SLM HPLC 99
Carboxylic acids Air SLM IC 100
continued
525
TABLE 16.3 (continuation)
TABLE 15.4
Various applications of membrane extraction in chemical analysis of foods and industrial
materials
Analytes Matrices Membrane Analytical Ref.
techniques techniques
Phenols Kerosene, naphtha PME FIA 108
Triazine herbicides Cooking oil MMLLE FIA, HPLC 109
Phenols Crude oil PME HPLC 110
Phenols Gasoline, Kerosene PME HPLC 111
Phenols Crude oil, fuels PME HPLC 112
Vitamin E Butter PME HPLC 62
Nicotine Snuff SLM (solid sample) UV 113
Caffeine Coffee, Tea SLM (solid sample) UV 114
Pesticides Eggs PME HPLC 115
Vanillin Various food SLM (solid sample) Amperometric 116
Anionic surfactants Detergents MMLLE FIA 117
Phenols Pyrolysis oil MMLLE LC-LC 56
Biogenic amines Wine SLM HPLC 45
526
Acknowledgements
The work of the Lund membrane group was supported by grants from the
Swedish Natural Science Research Council (NFR), the Swedish Environmental
Protection Agency (SNV), the Swedish Council for Forestry and Agricultural
Research (SJFR), the Swedish Institute, the Swedish International Develop-
ment Co-operation Agency (SIDA), the Crafoord Foundation and the European
Community (DG XII). The companies Pharmacia AB, Astra Draco AB and Astra
Pain Control AB have contributed with funds and interesting applications. The
work would not have been possible without a number of colleagues, especially
Prof. Lennart Mathiasson, graduate and undergraduate students as well as
guest researchers who have made important contributions.
REFERENCES
527
24 R.G. Melcher, Anal. Chim. Acta, 214 (1988) 299.
25 R.G. Melcher and S.A. Bouyoucos, Process Contr. Qual., 1 (1990) 63.
26 P.L. Morabito and R.G. Melcher, Process Contr. Qual., 3 (1992) 35.
27 B. Hauser and P. Popp, J. Sep. Sci., 24 (2001) 551.
28 Y. Shen, L. Gr6nberg and J.A. Jonsson, Anal. Chim. Acta, 292 (1994) 31.
29 M. Papantoni, N.-K. Djane, K. Ndung'u, J.A. J6nsson and L. Mathiasson, Analyst,
120 (1995) 1471.
30 P. Wieczorek, J.A. J6nsson and L. Mathiasson, Anal. Chim. Acta, 346 (1997) 191.
31 P. Dzygiel, P. Wieczorek, L. Mathiasson and J.A. J5nsson, Anal. Lett., 31 (1998) 1261.
32 P. Dzygiel and P. Wieczorek, J. Chromatogr.A, 889 (2000) 93.
33 P. Dzygiel and P. Wieczorek, J. Sep. Sci., 24 (2001) 561.
34 M. Rak, P. Dzygiel and P. Wieczorek, Anal. Chim. Acta, 433 (2001) 227.
35 J.A. Calzado, C. Palet, J.A. Jonsson and M. Valiente, Anal. Chim. Acta, 417 (2000)
159.
36 J.A. Calzado, C. Palet and M. Valiente, J. Sep. Sci., 24 (2001) 533.
37 B.D. Smith, Supramol. Chem., 7 (1996) 55.
38 M. Di Luccio, B.D. Smith, T. Kida, C.P. Borges and T.L.M. Alves, J. Membr. Sci., 174
(2000) 217.
39 T. Miliotis, M. Knutsson, J.A. J6nsson and L. Mathiasson, Int. J. Environ. Anal.
Chem., 64 (1996) 35.
40 E. Thordarson, J. Emneus and J.A. J6nsson, Anal. Chem., 72 (2000) 5280.
41 N.-K. Djane, K. Ndung'u, F. Malcus, G. Johansson and L. Mathiasson, Fresenius J.
Anal. Chem., 358 (1997) 822.
42 N.-K. Djane, I.A. Bergdahl, K. Ndung'u, A. Schiitz, G. Johansson and L. Mathiasson,
Analyst, 122 (1997) 1073.
43 N.-K. Djane, K. Ndung'u, C. Johnson, H. Sartz, T. Tornstrom and L. Mathiasson,
Talanta, 48 (1999) 1121.
44 P. Dzygiel, P. Wieczorek, J.A. Jnsson, M. Milewska and P. Kafarski, Tetrahedron,
55 (1999) 9923.
45 R. Romero, J.A. J6nsson, D. Gdzquez, M.G. Bagur and M. Sdnchez-Vifias, J. Sep. Sci.
(2002), in press.
46 J.A. Calzado, C. Palet and M. Valiente, Anal. Chim. Acta, 403 (2000) 101.
47 J.A. J6nsson, P. L6vkvist, G. Audunsson and G. Nilve, Anal. Chim. Acta, 227 (1993)
9.
48 L. Chimuka, N. Megersa, J. Norberg, L. Mathiasson and J.A. Jbnsson, Anal. Chem.,
70 (1998) 3906.
49 L. Chimuka, L. Mathiasson and J.A. Jonsson, Anal. Chim. Acta, 416 (2000) 77.
50 M. Knutsson, J. Lundh, L. Mathiasson, J.A. Jnsson and P. Sundin, Anal. Lett., 29
(1996) 1619.
51 L. Chimuka, M.M. Nindi, M.E.M. ElNour, H. Frank and C. Velasco, J. High Resol.
Chromatogr., 22 (1999) 417.
52 M. Knutsson, G. Nilv6, L. Mathiasson and J.A. J6nsson, J. Agric. Food Chem., 40
(1992) 2413.
53 P. Wieczorek, J.A. J6nsson and L. Mathiasson, Anal. Chim. Acta, 337 (1997) 183.
54 N. Megersa, T. Solomon and J.A. Jonsson, J. Chromatogr.A, 830 (1999) 203.
55 B. Lindegdrd, H. Bj6rk, J.A. J6nsson, L. Mathiasson and A.-M. Olsson, Anal. Chem.,
66 (1994) 4490.
56 T. Hy6tylainen, T. Andersson, M. Jussila, S.K. Wiedmer, M. Rautiainen and M.-L.
Riekkola, J. Sep. Sci., 24 (2001) 544.
57 B. Bernhardsson, E. Martins and G. Johansson, Anal. Chim. Acta, 167 (1985).
58 S. Palmarsdottir, E. Thordarson, L.-E. Edholm, J.A. Jnsson and L. Mathiasson,
Anal. Chem., 69 (1997) 1732.
528
59 G. Nilv6, M. Knutsson and J.A. Jonsson, J. Chromatogr.A, 668 (1994) 75.
60 M. Knutsson, L. Mathiasson and J.A. Jonsson, Chromatographia,42 (1996) 165.
61 Y. Shen, V. Obuseng, L. Gr6nberg and J.A. Jonsson, J. Chromatogr. A, 725 (1996)
189.
62 M.M. Delgado Zamarreno, A. Snchez Prez, M. Bustamante Rangel and J.
Hernandez M6ndez, Anal. Chim. Acta, 386 (1999) 99.
63 J.A. Jnsson, M. Andersson, C. Melander, J. Norberg, E. Thordarson and L.
Mathiasson, J. Chromatogr. A, 870 (2000) 151.
64 J. Norberg and E. Thordarson, Analyst, 125 (2000) 673.
65 F. Mullins, J. Sep. Sci., 24 (2001) 593.
66 S. Pedersen-Bjergaard and K.E. Rasmussen, Anal. Chem., 71 (1999) 2650.
67 S. Pedersen-Bjergaard and K.E. Rasmussen, Electrophoresis,21 (2000) 579.
68 K.E. Rasmussen, S. Pedersen-Bjergaard, M. Krogh, H.G. Ugland and T. Gr0nhaug,
J. Chromatogr.A, 873 (2000) 3.
69 H.G. Ugland, M. Krogh and K.E. Rasmussen, J. Chromatogr.B, 749 (2000) 85.
70 T. Gr0nhaug Halvorsen, S. Pedersen-Bjergaard and K.E. Rasmussen, J.
Chromatogr. A, 909 (2001) 87.
71 T. Gr0nhaug Halvorsen, S. Pedersen-Bjergaard, J.L.E. Reubsaet and K.E.
Rasmussen, J. Sep. Sci., 24 (2001) 615.
72 G.A. Audunsson, Anal. Chem., 60 (1988) 1340.
73 L. Mathiasson, M. Knutsson, G. Bremle and L. Martensson, Swedish J. Agri. Res., 21
(1991) 147.
74 B. Lindeghrd, J.A. Jonsson and L. Mathiasson, J. Chromatogr., 573 (1992) 191.
75 S. Palmarsdottir, B. Lindegard, P. Deininger, L.-E. Edholm, L. Mathiasson and J.A.
Jinsson, J. CapillaryElectroph., 2 (1995) 185.
76 S. Palmarsdottir, L. Mathiasson, J.A. Jonsson and L.-E. Edholm, J. Chromatogr.B,
688 (1997) 127.
77 S. Palmarsdottir, L. Mathiasson, J.A. Jnsson and L.-E. Edholm, J. Capillary
Electroph., 3 (1996) 255.
78 J. Trocewicz, Z. Suprynowicz and J. Markowicz, J. Chromatogr.B., 685 (1996) 129.
79 J. Norberg, J. Emneus, J.A. J6nsson, L. Mathiasson, E. Burestedt, M. Knutsson and
G. Marko-Varga, J. Chromatogr. B., 701 (1997) 39.
80 N.-K. Djane, S. Armalis, K. Ndung'u, G. Johansson and L. Mathiasson, Analyst, 123
(1998) 393.
81 Y. Shen, L. Mathiasson and J.A. Jonsson, J. Microcol. Sep., 10 (1998) 107.
82 O.B. Jonsson, E. Dyremark and U.L. Nilsson, J. Chromatogr.B, 755 (2001) 157.
83 J. Trocewicz, J. Sep. Sci., 24 (2001) 587.
84 T. Msagati and M.N. Nindi, J. Sep. Sci., 24 (2001) 606.
85 G. Nilv6, G. Audunsson and J.A. J6nsson, J. Chromatogr.,471 (1989) 151.
86 R.G. Melcher and P.L. Morabito, Anal. Chem., 62 (1990) 2183.
87 L. Mathiasson, G. Nilve and B. U16n, Int. J. Environ. Anal. Chem., 45 (1991) 117.
88 G. Nilve and R. Stebbins, Chromatographia,32 (1991) 269.
89 L. Gronberg, P. Lvkvist and J.A. Jonsson, Chemosphere, 24 (1992) 1533.
90 L. Gronberg, Y. Shen and J.A. J6nsson, J. Chromatogr., 655 (1993) 207.
91 N. Parthasarathy and J. Buffle, Anal. Chim. Acta, 284 (1994) 649.
92 Y. Shen, L. Str6m, J.A. J6nsson and G. Tyler, Soil Biol. Biochem., 28 (1996) 1163.
93 J. Trocewicz, J. Chromatogr.A, 725 (1996) 121.
94 R. Carabias Martinez, E. Rodriguez Gonzalo, M.P. Santiago Toribio and J.
Hernandez M6ndez, Anal. Chim. Acta, 321 (1996) 147.
95 L. Chimuka, M.M. Nindi and J.A. Jonsson, Int. J. Environ. Anal. Chem., 68 (1997)
429.
96 J. Norberg, A. Zander and J.A. Jonsson, Chromatographia,46 (1997) 483.
529
97 N. Parthasarathy, M. Pelletier and J. Buffle, Anal. Chim. Acta, 350 (1997) 183.
98 N. Megersa and J.A. J6nsson, Analyst, 123 (1998) 225.
99 K. Ndung'u, N.-K. Djane and L. Mathiasson, J. Chromatogr.A, 826 (1998) 103.
100 L. Martensson, M. Magnusson, Y. Shen and J.A. Jonsson, Agri. Ecosys. Environ., 75
(1999) 101.
101 N. Megersa, T. Solomon, B.S. Chandravanshi and J.A. Jnsson, Bull. Chem. Soc.
Ethiopia, 14 (2000) 9.
102 K. Ndung'u and L. Mathiasson, Anal. Chim. Acta, 404 (2000) 319.
103 J. Norberg, E. Thordarson, L. Mathiasson and J.A. J6nsson, J. Chromatogr.A, 869
(2000) 523.
104 M. Sandahl, L. Mathiasson and J.A. J6nsson, J. Chromatogr.A, 893 (2000) 123.
105 M. Sandahl, E. DUlfsson and L. Mathiasson, Anal. Chim. Acta, 424 (2000) 1.
106 X. Guo and S. Mitra, J. Chromatogr.A, 904 (2000) 189.
107 N. Megersa, L. Chimuka, T. Solomon and J.A. Jonsson, J. Sep. Sci., 24 (2001) 567.
108 E. Rodrigues Gonzalo, J.L. Perez Pav6n, J. Ruzicka, G.D. Christian and D.C. Olson,
Anal. Chim. Acta, 259 (1992) 37.
109 R. Carabias Martinez, E. Rodriguez Gonzalo, E. Herndndez Fernandez and J.
Herndndez M6ndez, Anal. Chim. Acta, 304 (1995) 323.
110 T. Garcia Sanchez, J.L. PBrez Pav6n and B. Moreno Cordero, J. Chromatogr.A, 766
(1997) 61.
111 E. Fernandez Laespada, J.L. Perez Pav6n and B. Moreno Cordero, J. Chromatogr.A,
823 (1998) 537.
112 E. Ferndndez Laespada, J.L. P6rez Pav6n and B. Moreno Cordero, J. Chromatogr.A,
852 (1999) 395.
113 E. Luque-Perez, A. Rios, M. Valcdrcel, L.-G. Danielsson and F. Ingman, Anal. Chim.
Acta, 387 (1999) 155.
114 E. Luque-Prez, A. Rios, M. Valcarcel, L.-G. Danielsson and F. Ingman, Lab. Autom.
Inform. Managem., 34 (1999) 131.
115 R. Carabias Martinez, E. Rodriguez Gonzalo, P.H. Paniagua Marcos and J.
Herndndez MBndez, J. Chromatogr.A, 869 (2000) 427.
116 M. Luque, E. Luque-Prez, A. Rios and M. Valcdrcel, Anal. Chim. Acta, 410 (2000)
127.
117 L. Jing-fu and J. Gui-bin, Microchem. J., 68 (2001) 29.
530
Chapter 16
16.1 INTRODUCTION
Volatile organic compounds can be sampled and analysed directly from soils,
liquids and gases by membrane inlet mass spectrometry (MIMS). This unusual
flexibility in matrixes that can be analysed is achieved through an inert polymer
membrane, which is used as the only interface between the sample and a mass
spectrometer. The membrane is simply brought into contact with the sample
where, after the volatile organic compounds diffuse through it and evaporate
into the mass spectrometer, they are ionized and analysed. In this fashion most
volatile organic compounds can be analyzed directly and in most cases a chang-
ing chemical environment in the sample can be monitored on-line.
Membrane inlet mass spectrometry is a relatively old technique; it was
introduced by Hoch and Kok in 1963 [1] as a method for the continuous
monitoring of gases dissolved in water. Soon after it was tested as an interface
between a gas chromatograph and a mass spectrometer [2]. However, a special
characteristic of membrane inlet systems-a very high selectivity in the trans-
port through the membrane-has limited this application. Normally the MIMS
membranes are very hydrophobic and hydrophobic compounds dissolve very
well in the membrane. They have low detection limits (ng/l in water and -1
Ag/m 3 in air), whereas hydrophilic compounds have detection limits in the high
jg/il range in water and mg/m 3 in air.
For many years the main applications of the MIMS technique were in vivo
monitoring of blood gases [3] and monitoring of dissolved gases and small
organic compounds (MW < 100 Dalton) in fermentors [4]. This changed in the
late eighties when the direct insertion membrane probe was developed [5]. The
direct insertion membrane probe turned the strategy of membrane inlet design
upside down. Instead of having the membrane mounted at the end of a probe
which was inserted into the sample, the membrane probe was inserted into the
mass spectrometer and the sample had to be transported to the membrane inlet.
With this design volatile organic compounds are delivered directly to the ion
source of the mass spectrometer and do not have to be transported through an
evacuated tube, where chromatographic effects at the surfaces can take place
[6]. The direct insertion membrane probe initiated a very productive decade
Comprehensive Analytical Chemistry XXXVII
J. Pawliszyn (Ed.)
© 2002 Elsevier Science B.V. All rights reserved
when the MIMS technique evolved from a specialist technique into a mature
method for the detection of volatile organic compounds in environmental
samples [7] and in biological matrixes [8].
During the nineties, the application of the MIMS technique was expanded
beyond the analysis of gases and small organic compounds in aqueous matrixes.
Methods that made it possible to analyse complex organic solutions like gasoline
were developed [9] and an interest in the analysis of gaseous samples appeared
[10]. For example, a fast and sensitive method for the detection of sulfur
compounds in air was recently presented [11]. Solid analysis is a new application
area for the MIMS technique and optimized methods for the analysis of residual
volatiles in soil samples are emerging [12,13]. A review of MIMS applications can
be found in Srinivasan et al. [14] and Johnson et al. [15].
At the moment the detection of volatile organic compounds (boiling point <
200°C) have become a routine matter. However, effort is put into the develop-
ment of methods that will allow the direct detection of less and non-volatile
organic compounds [16-19]. Among these methods the desorption chemical
ionization method [20] is currently the most versatile allowing the detection of
large and complex biomolecules like steroid hormones [21], atrazine, PCB and
phthalates [22].
As well as the main areas of MIMS development and application described
above, many specialised MIMS methods have been presented in the past, and
this chapter gives an overview of the many possibilities of the technique. Empha-
sis is on the flexibility of the MIMS technique, when it comes to the analysis of
samples having completely different matrixes. First, it describes methods for the
characterization of gaseous samples, next characterization of aqueous and org-
anic liquids, followed by analysis of sediments, and finally, solid samples.
532
I, =AxDxKxx (16.1)
1
where A is the surface area of the membrane, D is the diffusion constant, K is the
distribution ratio of the analyte between the concentration in the membrane and
in the sample, C is the analyte concentration in the sample, and I the thickness of
the membrane. In other words, the flow through the membrane is proportional
to the surface area of the membrane, the product between the diffusion constant
and the distribution ratio (a parameter called the permeability) and the concen-
tration gradient across the membrane.
The response time (t1 -90 %) defined as the time it takes the signal to increase
from 10% to 90% following a step change in concentration is given by
12
t0-90 = 0.237 x - (16.2)
D
It increases with the square of the membrane thickness and decreases
proportionally to the diffusion constant. In contrast to the steady-state flow the
response time does not depend upon the distribution ratio.
The distribution ratio is properly the most important parameter, when it
comes to the detection limits of a MIMS system. As mentioned above, most
MIMS systems use a very hydrophobic silicone membrane and this membrane
discriminates strongly against compounds containing polar groups. This is best
demonstrated by a comparison of detection limits obtained from structurally
similar compounds. For example, the detection limit for toluene is normally in
the ng/l range, whereas that of the anisol (methoxybenzene) is 10 times higher,
benzaldehyde is 10-100 times higher, benzyl alcohol more than 100 times higher
and benzoic acid as the most polar compound in the series is more than a
thousand times higher. The extreme variation in solubility of relatively similar
compounds in the membrane makes MIMS a special technique in that it gives a
biased picture of the hydrophobic components in a sample. Often, constituents
in a sample that hitherto has been unnoticed using other analytical methods,
suddenly give a signal that dominates the analytical output and suppresses the
signal from well-known constituents. The identification of 3-Cl-4-methoxy-
benzaldehyde in growth media of the white-rot fungus Bjerkandera adusta is a
good example of this [23]. Unfortunately this feature occasionally makes it
impossible to determine the concentration of polar constituents present at high
concentration.
The hydrophobic nature of some compounds can cause analytical difficulties
when present at high concentrations. They might act as plasticizers and cause
softening (swelling) of the membrane. The signal becomes nonlinear with con-
centration. 3-Cl-4-methoxy-benzaldehyde is an example of a compound that
softens silicone membranes. It gives linear signals from the detection limit (< 10
gg/l) and up to about 10 mg/Il, where after the signal increases too fast up to about
50 mg/l before it becomes linear again (Fig. 16.1). This behaviour can be
explained by the simple assumption, that the sample molecules interact with the
533
1 _
0. -
/
0.6-
0.4-
polymer chains and cause a local softening of the membrane, where diffusion is
faster and perhaps the distribution ratio is also higher. At low concentration the
distance between individual sample molecules in the membrane is so big that
one molecule will not experience the soft areas created by other molecules. At a
certain point the soft areas begin to overlap and a molecule diffusing through the
membrane will experience both soft and dense places in the membrane. The
higher the sample concentration the more soft areas are experienced and the
result is a signal that increases faster than linearity. At a certain concentration
the membrane is fully softened (swelled), and the signal becomes linear again,
but the permeability has increased.
The effect of membrane swelling can be difficult to observe due to saturation
effects [24] in the mass spectrometric system. When it is observed, it does not
just cause nonlinearities for the hydrophobic constituent itself, it will also influ-
ence the detection of other compounds in the sample. A related effect can be
observed in connection with the analysis of liquid samples that have a high salt
concentration (-40 gl). Such a matrix can influence the distribution ratio
between the sample and the membrane with nonlinear effects as the result.
The diffusion constant influences both the steady-state flow through the
membrane and the response time. In general, the larger the compound the
slower it diffuses through the membrane, since more free space between the
polymer chains is needed before the molecule can pass. As a result large com-
pounds have long response times, whereas small compounds have short response
times. However, the small diffusion constant for large compounds does not nec-
essarily mean that the steady-state flow through the membrane becomes small.
As already mentioned, the steady-state flow through the membrane depends on
the product between the diffusion constant and the distribution ratio. Often the
increase in distribution ratios by far exceeds the decrease in diffusion constant
and the result is a higher flow through the membrane for the large compound.
This is exemplified by octanol which has a steady-state flow through the mem-
brane which exceeds that of ethanol by approximately a factor of 50 [6].
The evaporationof volatile organic compounds from the membrane surface
into the mass spectrometer is normally not a problem. The vapour pressure is
high and the flow through the membrane so low that heat conduction can
compensate for the energy used in the evaporation process. It is after the
534
evaporation from the membrane has taken place that problems may arise.
Interactions between the vaporized analyte and surfaces in the vacuum [6,25]
may cause prolonged response times or even a complete suppression of the
signal. Most membrane inlets are therefore designed to have as short a distance
as possible from the membrane to the ion source of the mass spectrometer.
Compounds of low volatility (boiling point > 250°C) cannot be detected by
standard MIMS. They stick to the membrane and have to be released by some
sort of stimulated evaporation. Thus far, three methods have been tested: in
trap-and-release MIMS [16,17] a sudden thermal heating of the membrane up to
300°C is used; in laser desorption MIMS [19] the vacuum side of the membrane is
irradiated by laser; and in desorption chemical ionization MIMS [20,21] a
plasma-assisted evaporation is used.
The membrane temperature is very important for the performance of a
membrane inlet. Both the diffusion constant, the distribution ratio and the
permeability (equal to the product of the diffusion constant and the distribution
ratio) depend on the temperature, according to Arrhenius equations [26]:
K = K xexpAHRT )) (16.4)
where Do, Ko and P0 represent the diffusion constant, the distribution ratio and
the permeability at some initial temperature, To. Ed is the energy of activation
for diffusion, AH. is the difference in heats of solution between the membrane
and the sample matrix and Ep the activation energy for permeation.
In general the activation energy for diffusion, Ed, is larger than zero, which
means that the diffusion constant increases with temperature, whereas for most
organic compounds the difference in heats of solution between the membrane
and the matrix, AHs, is less than zero, which means that the analyte concentra-
tion inside the membrane decreases with temperature. In all circumstances, the
response time, which is inversely proportional to the diffusion constant, becomes
shorter the higher the temperature. The steady-state flow through the mem-
brane can both increase or decrease, when the temperature is increased. This is a
question of the sign of the sum of Ed and AH. For gas analysis AHs is often larger
than Ed, whereas for water analysis AH is smaller than Ed. The result is that for
gas analysis the steady-state flow through the membrane decreases with an
535
increase in temperature, whereas it increases for water analysis. Gas analysis is
therefore carried out at a temperature that is chosen as a compromise between a
desire for maximum signals (low temperature) and short response time (high
temperature). Water analysis is carried out at the practically maximal tempera-
ture, which is normally around 70°C. At higher temperatures bubble formation
in front of the membrane can cause serious instability in the signals [27].
Membrane inlets exist in many different designs, each created to fit a specific
application. However, most of the designs belong to one of six categories (Fig.
16.2): (a) membrane probes, where the membrane is mounted at the end of a
probe to be inserted directly into the sample matrix; (b) the direct insertion
membrane probe, where the membrane is mounted inside the mass spectrome-
ter and the sample (gas or liquid) has to be transported to it; (c) measuring cells,
where a sample (liquid or solid) is transferred to a small measuring cell that has
a membrane built into the wall; (d) the helium purge inlet, where permeate
molecules at the inside of the membrane is purged by helium to the mass
spectrometric ion source; (e) two stage inlet systems, where two different
systems (one being the membrane) for separation and/or preconcentration of the
sample are used in series; and (f) membrane inlets using a stimulated desorption
for the analysis of semi- and/or non-volatile organic compounds.
a) _____ _ d)
He
Membrane
b) e)
bII He
Member
_:Membrane Memban
Membrane I Jet separator
Membrane 1 Sample in
Fig. 16.2. Different types of membrane inlet design. (a) Membrane probe mounted directly in a
reactor. (b) Flow-through membrane probe. (c) Sample cell mounted directly at the mass spectro-
metric ion source. (d) Helium purge membrane inlet. (e) A two-stage membrane inlet system using
a helium purge membrane inlet followed by ajet separator that reduces the relative amount of the
helium in the carrier gas. (f) A membrane inlet using stimulated desorption to release less volatile
organic compounds from the membrane surface. In the system presented in (f) (T&R-MIMS) the
membrane is continuously irradiated by light and electrons from the filament.
536
The membraneprobe is probably the simplest membrane inlet to use. It con-
sists of a capillary (steel) perforated, and covered by a polymer membrane in one
end, while the other end is connected via an evacuated tube to the mass spec-
trometer. The probe can be inserted directly into almost any sample. For exam-
ple they have been inserted into arteria of humans [3], into plants [28], into
sediments [29] and into chemical and biological reactors of all kinds. The draw-
back of the membrane probe is that it can only be used to analyse gases and
highly volatile organic compounds (b.p. < 100°C). Compounds of a lower volatil-
ity interact with the surfaces in the evacuated tube which connects the inlet to
the mass spectrometer and very long response times are often the result.
In order to circumvent the problem with condensation of sample molecules
in the tube that connects a membrane probe to the mass spectrometer, the direct
insertion membrane probe was invented [5]. In this inlet the membrane is
mounted at the end of a probe which is inserted directly into the mass spectro-
metric ion source. The liquid or gas sample is then flushed through the probe.
The advantage of the inlet is the elimination of the connection tube between the
membrane and the ionizing region and volatile organic compounds with a boiling
point up to around 200°C can be measured. Further, the inlet has the advantages
that the sample can be chemically modified on-line prior to the detection and
calibration using an external standard is straight forward. The inlet has found
extensive use in connection with detection of volatile contaminants in water [7]
and with monitoring of growing microbial cultures [30].
The measuringcell is a small vessel (1-5 ml) mounted as close as possible to
the ionizing region of the mass spectrometer. It is interfaced directly to the mass
spectrometer via a sheet membrane that constitutes a part of one of the walls.
The sample, liquid or solid, is simply transferred to the vessel and after a few
minutes a MIMS spectrum of the volatile constituents can be recorded. When
the measuring cell is attached directly to the ion source [31] it can be used to
measure the same compounds as the direct insertion membrane probe.
An alternative to the close coupling of the membrane inlet to the ionizing
region of the mass spectrometer is found with the helium-purge membrane inlet
[32]. In this inlet a liquid or gas sample is flushed across one side of the
membrane, whereas the other side is continuously purged by a helium stream
that carries the permeated molecules to the ionizing region of the mass spectro-
meter. The purging of molecules from the membrane to the ion source reduces
the problem with condensation effects in connection tubes and the inlet has
almost the same applications as the direct insertion membrane probe.
In order to improve either selectivity or sensitivity, many devices have been
developed which combine the membrane inlet with another analytical chemical
technique for separation or concentrating. The two-stage inlets can be split into
systems using: two steps of membrane separation, a pre-separation or pre-
concentrating device before the membrane inlet; and systems using a post-
concentrating device after the membrane inlet. Most of the two-step devices
have only been introduced or briefly discussed in the literature and will not be
537
described here. One combination that is frequently used is the combination of a
helium-purge membrane inlet and a jet separator [33]. In this combination the
jet separator is used for selective removal of helium from the purge stream
before it reaches the mass spectrometer. Examples have recently been published
where either a sorbent (Tenax) [34] or a cold trap [35] has been used after the
membrane inlet to concentrate the membrane permeate prior to the detection.
With these systems a considerable improvement in detection limits is to be
expected. The sorbent trap can also be used prior to the membrane inlet in
connection with gas analysis. Here, a pre-separation of related compounds can
even be achieved in addition to the preconcentration if the adsorbed molecules
are released from the trap using a controlled temperature gradient [36].
The five inlet types presented are designed for the measurement of gases
and/or volatile organic compounds and are only of limited use for the measure-
ment of less volatile (b.p. >200°C) organic compounds. In order to measure such
compounds with a membrane inlet, it is necessary to stimulate their evaporation
from the membrane surface. Currently, three techniques are in use: the trap-
and-release technique [16,17], laser desorption [19], and desorption chemical
ionization [20]. In the trap-and-release technique the sample (gas or liquid) is
passed through the membrane inlet and the less volatile organic compounds dif-
fuse into the membrane, but do not evaporate from it. After a sampling period
the membrane is rapidly heated and the less volatile organic compounds evapo-
rate from the membrane into the mass spectrometer. The desorption chemical
ionization method is an improved version of the trap-and-release method, where
the membrane is positioned in the centre of a chemical ionization plasma. The
plasma assists in the liberation of nonvolatile compounds from the membrane
during the heating process by ionizing the analyte molecules directly at/on the
membrane surface. With this method it is possible to measure compounds like
polyaromatic hydrocarbons and estrogenic compounds like phthalates at ng/l
concentrations in water, whereas many steroid hormones and pesticides can be
detected at low pg/l concentrations [22]. In laser desorption MIMS the evapora-
tion of less volatile organic compounds from the membrane is stimulated by
irradiation of the membrane surface with laser light. This technique has success-
fully been applied to measure polyaromatic hydrocarbons in water at low ng/l
concentrations.
16.4.1 Characterization
538
TABLE 16.1
Detection limits, response times and linear ranges measured by MIMS for a selected set of
environmentally significant compounds from air samples (adapted from Refs. [37,38])
Compound Response time Detection limit Linear range
(s)a
3
(Lg/m )b (Ag/m3 )
Benzene 1 1 1-10 000
Toluene 1 0.5 1-10 000
Xylenes 2 2 2-10 000
Chlorobenzene 2 0.5 2-10 000
Chloroform 1 5 5-30 000
Carbon tetrachloride 1 3 10-20 000
1,2-Dichloroethene 1 1 5-10 000
1,1-Dichloroethane 1 1 1-5 000
1,1,1-Trichloroethane 2 4 10-30 000
1,1,2,2-Tetrachloroethane 3 1 1-5 000
Tetrachloroethene 1 0.5 1-10 000
Dimethyl sulfide 1 1 1-40 000
Dimethyl disulfide 2 2 2-40 000
Ethanethiol 2 5 5-50 000
aMembrane (polydimethylsiloxane) thickness 25 tm. Response time is defined as the time it
takes for the signal to rise from 10 to 90%. bS/N = 3.
ones published, but they are levels that should be relatively easy to obtain with a
flow-through membrane inlet and a new mass spectrometer. Many of the VOCs
can be measured at low /tg/m 3 levels directly from air. As an example of the
results, single ion monitoring of tetrachloroethene close to the detection limit is
shown in Fig. 16.3. The response times that can be obtained using a MIMS
system (seconds) are very good when measurements are done from air using a
thin membrane (25 /tm). Table 16.1 also shows that wide linear dynamical
ranges are obtained, typically 3-5 orders of magnitude. The reliability of the
quantitative results has also been confirmed by comparing the total VOC-
contents measured by MIMS to those obtained with an on-line FID-analyzer in
analysis of paintshop exhaust air [39], the agreement of the results was very
good.
16
14
12
10
539
In addition to the typical VOCs listed in Table 16. 1, it has been demonstrated
that MIMS can be used to measure organometallic (molybdenum hexacarbonyl
and ferrocene) [40], semivolatile [18] and polar [41] organic compounds directly
from air. However, because the polydimethylsiloxane membrane discriminates
against polar compounds it was proposed that direct sampling glow discharge
mass spectrometry is better suited for polar compounds [41]. Semivolatiles, such
as malathion, nitrobenzene, 2-chlorophenol, and dimethyl methylphosphonate
could be measured at ppb levels or lower [18].
A very innovative application of MIMS in air monitoring was a study by Colo-
rado et al. [42]. They built an automatic MIMS apparatus utilizing a membrane
and cryogenic concentration unit for monitoring isoprene from greenhouse air
below ppb levels. The necessary selectivity for the measurement was obtained
using chemical ionization with vinyl methyl ether (VME) reagent gas and tan-
dem mass spectrometry. VME reacts with isoprene to form an adduct ion at m/z
94 corresponding to [isoprene+VME-MeOH] +.Under collision induced dissocia-
tion the ion m/z 94 fragments by methyl radical loss to ion m/z 79, on-line moni-
toring this single reaction gave the necessary selectivity and sensitivity for the
on-line greenhouse air monitoring experiments performed.
Figure 16.4 shows an example of the results obtained in the study where a
method called temperature-programmed desorption membrane inlet mass
spectrometry (TPD-MIMS) was introduced [36]. In this technique VOCs are con-
centrated, typically from air, into an adsorbent material, and subsequently de-
sorbed using a rapid temperature program and detected with a MIMS apparatus.
The method allows separation of mixtures within a few minutes and detection at
sub-nanogram levels.
Monitoring of fermentation products/off-gases (e.g., N2, O,, CO, and ethanol)
from outlet gas of fermentors is a very important application of MIMS [30,
0.5
0.4
U 0.3
0.2
0.1
0
0 1 2 3
Time, min
Fig. 16.4. Separation of a four component mixture using the TPD-MIMS technique. Compounds
in the mixture are: 1 dichloromethane (6 ng, m/z 84), 2 trans-1,2-dichloroethene (7.8 ng, m/z 61),
3 chloroform (5.5 ng, m/z 83) and 4 tetrachloroethene (3.6 ng, m/z 166). The volume of the air
sample was 25 ml and the adsorbent material used was HayeSepD. (Reproduced with permission
of John Wiley & Sons Limited from R.A. Ketola, G. Gron, F.R. Lauritsen, Temperature-
programmed desorption for membrane inlet mass spectrometry, Rapid Commun. Mass
Spectrom., Vol 12, p. 773-778, Copyright 1998 by John Wiley & Sons Limited.)
540
43-47]. The use of adsorbents has also been adapted with MIMS methods to
increase selectivity and detection limits [36,48]. Another interesting demonstra-
tion of the wide applicability of MIMS is on-line monitoring of industrial solvents
(e.g., styrene, 1,1,1-trichloroethane and toluene) in breath [49].
To our knowledge, MIMS has not hitherto been used to follow gas phase
reactions-a potentially very good application of MIMS. The paper which comes
closest to this is a study in which MIMS was used to introduce monochloramine,
a reactive initiator of chlorination reactions, into the ion source of a mass
spectrometer [50].
16.5.1 Water
16.5.1.1 Characterization
Measurement of volatile organic compounds directly from aqueous phase is the
standard application of MIMS [8,14,15,26,30,32,37,44-47,51-60] and recent
work has shown that many semivolatile organic compounds can also be mea-
sured reliably [20-22,61-65]. Table 16.2 summarizes the typical performance
characteristics of MIMS in direct measurement of VOCs and SVOCs from water.
The detection limits given for VOCs are not necessarily the lowest published val-
ues, but are levels that are relatively easy to obtain with most MIMS systems
using a flow-through membrane inlet and a new mass spectrometer. Many of the
VOCs can be measured at ng/l (pptr, parts-per-trillion) levels directly from aque-
ous phase and in an exceptional case detection of certain compounds at high pg/l
(ppq, parts-per-quadrillion) have been demonstrated, as shown in Fig. 16.5.
Detection limits of SVOCs are much higher with standard MIMS methods, but
with the newly developed trap-and-release methods (T&R-MIMS) SVOCs can be
measured at ,ugfl levels (ppb, parts-per-billion) [21,22,61]. Figure 16.6 shows
that steroid hormones can be reliably measured using desorption chemical ion-
ization membrane inlet mass spectrometry (DCI-MIMS) [21], a recently intro-
duced technique combining trap-and-release MIMS and chemical ionization.
Response times in the MIMS method are typically in the range of 1-3 min when
VOCs are measured from water using a 100 m thick silicone membrane. In the
trap-and-release method a pre-set concentration time is used before the analytes
are desorbed from the membrane. The complete analysis time (5-30 min) is
therefore more important than membrane response times.
Table 16.2 also shows that with standard MIMS a wide linear dynamical
range is obtained for VOC analysis, typically 3-5 orders of magnitude, whereas
the trap-and-release method only has a linear dynamical range of 2-3 orders of
magnitude. The reliability of the quantitative results obtained by MIMS has
541
TABLE 16.2
Detection limits, response times and linear ranges measured by MIMS for a selected set of
environmentally significant compounds from aqueous samples (adapted from Refs. [21,22,
37,611)
Compound Response Detection limit Linear range
time (min)' (ig/l)d (jig/l)
VOCsa
Benzene 2.0 0.1 0.1-10 00
Toluene 1.5 0.1 0.3-2 000
Xylenes 2.0 0.1 0.1-5 000
Chlorobenzene 1.5 0.2 ND'
Chloroform 1.5 0.1 0.1-1 000
Carbon tetrachloride 2.0 0.1 0.1-1 000
1,2-Dichloroethene 1.0 0.1 ND'
1,1-Dichloroethane 1.0 0.4 NDe
1,1,1-Trichloroethane 1.5 0.1 0.4-1 0OO
e
Dimethyl sulfide ND 0.5 ND'
Dimethyl disulfide NDe 0.5 NDe
Ethanethiol ND' 0.1 ND'
542
?n
+
0
E
a
Fig. 16.5. The response of trans-dichloroethylene (m/z 96 and 98) in the analysis of water in
which samplings were made at two different concentration levels, 1 pptr and 500 ppq. (Reprinted
with permission from M. Soni, S. Bauer, J.W. Amy, P. Wong, R.G. Cooks, Direct determination of
organic compounds in water at parts-per-quadrillion levels by membrane introduction mass
spectrometry, Analytical Chemistry, Vol 67, pp. 1409-1412, Copyright 1995 American Chemical
A _ are~'"
1.0 n
{D)
O 0.4 .0
< 1.0 -
a 0.2 0.5 -
blank
0.0 0.0
270 280 290 300 310 320 0 100 200 300 400
m/z Time (cycles)
Fig. 16.6. (a) DCI-MIMS mass spectrum of a birth control pill and (b) multiple ion monitoring
data for extracts of three birth control pills containing the synthetic sex hormones ethynyl-
estradiol (30 gLg,m/z 279 water loss from protonated molecule) and levonorgestreol (75 ag, m/z
313 protonated molecule) dissolved in 20 ml of distilled water and analyzed as is. (Reproduced by
permission of The Royal Society of Chemistry from F.R. Lauritsen, J. Rose, Determination of
steroid hormones by membrane inlet mass spectrometry and desorption chemical ionization,
Analyst, Vol 125, p.1577-1581, Copyright 2000 by The Royal Society of Chemistry.)
543
snn as_
180
160
140
120
0 100
80
60
40
20
Fig. 16.7. On-line MIMS measurement of the waste water stream of an oil refinery. (Reproduced
with permission of VSP International Science Publishers from T. Kotiaho, R. Kostiainen, R.A.
Ketola, T. Mansikka, I. Mattila, V. Komppa, T. Honkanen, K. Wickstrim, J. Waldvogel, O. Pilvi6,
Development of a fully automatic membrane inlet mass spectrometric system for on-line
industrial waste water monitoring, Proc. Control and Qual., Vol 11, pp. 71-78, Copyright 1998 by
VSP International Science Publishers.)
rates were also used to aid the identification of the metabolites of Trichomonos
vaginalis [68]. Recently, it has also been shown that mathematical methods can
be used for identification of the individual compounds based on the electron
ionization multicomponent mass spectra measured by MIMS [70].
One of the most interesting recent applications of MIMS is on-site and
on-line monitoring of environmentally significant compounds from water. A cus-
tom-made MIMS apparatus to be used in a mobile laboratory has been built for
on-site measurements [52], a specialized MIMS system has been designed for
on-line measurement of VOCs of river water in a boat [51] and for on-line
wastewater monitoring of chemical plants (Fig. 16.7) [55], a comprehensive kit
for analytical field work utilizing a GC/MS-instrument equipped with a
membrane inlet has been constructed [71] and a submersible membrane inlet
GC/MS-instrument has been built for marine pollution measurements [72].
Other interesting instrumental developments have been, e.g., development of a
cryotrap MIMS for measurement of VOCs at low pptr levels [35] and demonstra-
tion that polar organic compounds can be measured at ppb levels using a
microporous membrane combined with glow discharge ionization [73]. Some
nonconventional MIMS applications published are measurement of active prin-
ciples in tisanes of vegetable drugs [74], studies of aroma fraction of wines [75],
and classification of cola beverages [76]. The latter study clearly demonstrated
that MIMS (Fig. 16.8) is an excellent alternative or complementary method to
the electronic nose sensors used for classification of foodstuffs.
544
2 . 4
1.5 ; c- Classes
x
0 2
- I x
x x x 3
0 *0 +4
05
0 0 7
7 8
.. *~~~~~~~~ A 9
;> < 10
D >11
** * 12
-1 It* *13*
Zip a 13
-1.5 0
-6 -4 -2 0 2 4 6 8 10 12
1. principal component
Fig. 16.8. Separation of thirteen cola beverages according to their first two principal components.
(Reproduced with permission of John Wiley & Sons Limited from R.A. Ketola, J. Heikkonen, S.
Piepponen, F.R. Lauritsen, T. Kotiaho, Classification of Cola beverages on the basis of mass
spectra measured by membrane inlet mass spectrometry, Rapid Commun. Mass Spectrom. Vol
12, pp. 1011-1017, Copyright 1998 by John Wiley & Sons Limited.)
545
7
, EI
-6 0
2.0- *POAA Penicillin-V ,
-5 (
a. 1.5-
-4 o
to
0.0- 0_
50 100 150 200
Time (hours)
Fig. 16.9. On-line monitoring ofphenoxyacetic acid during a 200-h penicillin-V fermentation (A).
Penicillin-V was measured off-line using the hydroxylamine method (D). (Reprinted by permission
of Wiley-Liss, Inc., a subsidiary of John Wiley &Sons, Inc., from K.F. Hansen, F.R. Lauritsen, H.
Degn, An on-line sampling system for fermentation monitoring using membrane inlet mass
spectrometry (MIMS): Application to phenoxyacetic acid monitoring in penicillin fermentation,
Biotechnol. Bioeng. Vol 44, pp. 347-353, Copyright 1994 by John Wiley &Sons, Inc.)
16.5.2.1 Characterization
A generic name, reversed-phase membrane inlet mass spectrometry, has been
adapted for MIMS methods in which water or organic compounds are measured
from organic matrix [9]. The name was borrowed from terminology used in
liquid chromatography, since typically MIMS is used to measure VOCs from
aqueous samples. The reversed phase MIMS papers published are dealing with
two different application areas, namely measurement of water activity in org-
anic liquids of low dielectric constant [93-95] and measurement contaminants of
organic solvents [9]. However, measurement of styrene and tetrachloroethylene
directly from olive oil [96] and measurement of Kathon CG (an antimicrobial
agent) from cosmetic emulsions [97] can also be considered reversed-phase
MIMS studies.
The method for determination of water activity developed by Degn et al.
utilizes a hydrophilic polyethylene terephthalate membrane instead of the nor-
546
35 + 37
NH 2 CI / N C1+' 51
90"] (11-L,
30.6-
NH237CI+ 53
- 27.4- NH35CI2+ 85
35 37
NH CI CI+. 87
+ .
N35CI3 119
in
00.0
4.2- 35 37 +.
.. ~ N CI2 CI 121
35C2
+' 70
639.
35C137CI+. 72
Fig. 16.10. On-line monitoring of the reactions of monochloramine with HCI. Additions of HCl
were made at the times indicated by the arrows. Data were collected by scanning the full El mass
spectra (m/z 33-300). (Reprinted from T. Kotiaho, A.K. Lister, M.J. Hayward, R.G. Cooks,
On-line monitoring of chloramine reactions by membrane introduction mass spectrometry,
Talanta, Vol 38, pp. 195-200, Copyright (1991), with permission from Elsevier Science.)
547
5 i A N
A
LUo d 1.0 - b
0
o 0.8 o 0.8 -
0
0 6
= . c 0.6-
e 0.4-
.01
0 0.2-4
0.2 c 0.2 I
10
U0
I 1..
4
I 50 6b
I
f0
T
8U 0 100
V. V
'^
IU
.
^^
4U
I
.
^^ '^..
4
^I.
50
_I
^^
o
-^
o
^^
0
^^ '^^I
9 100
mlz m/z
Fig. 16.11. Collision induced dissociation spectra of m/z 89 obtained from (a) gasoline and (b) 100
ppm of tert-butyl methyl ether in hexane. (Reprinted with permission from F.R. Lauritsen, T.
Kotiaho, T.K. Choudhury, R.G. Cooks, Direct detection and identification of volatile organic
compounds dissolved in organic solvents by reverse phase membrane introduction tandem mass
spectrometry, Analytical Chemistry Vol 64, pp. 1205-1211, Copyright 1992 American Chemical
Society.)
548
organic compounds in blood [104] and for measurement of partial gas pressures
in bone tissue [102]. Most of these studies have been done using animals
[3,101-104], but interestingly subcutaneous tissue gas tensions have also been
measured for humans [105]. A recent application of MIMS is the demonstration
that NO can be measured from very small human plasma sample volumes
(10-300 Al) at gM levels [106]. In addition to the animal tissue gas tension
measurements, MIMS was shown to give valuable information about the gas
concentrations in living plants [28].
16.5.3.2 Urine
A MIMS method for determining the toluene metabolite p-cresol from urine has
been developed [107]. Urine was not directly analyzed, even though that is also
possible; instead the metabolite was measured from 1-propanol solutions after a
three-step sample preparation (hydrolysis, extraction, dissolution). Average
p-cresol concentration of unexposed workers was measured to be 11 Ag/ml.
16.5.3.3 Saliva and stomach fluid
Membrane inlet mass spectrometry has played a very import role in the studies
of changes in gases dissolved in the rumen fluid in situ and in vitro [108]. MIMS
has been used for measurement of dissolved gases (e.g. CH4, CO2, H2 and O,2)
directly in the rumen liquor of ten fistulated sheep on variety of diets. The in situ
MIMS experiments also allowed one to design laboratory experiments that
mimic the natural conditions much better than before.
MIMS has also been used to measure the time persistence of monochlor-
amine in human saliva and stomach fluid [109]. This information is very
important since monochloramine is widely used for disinfection of drinking
water and can be toxic. The study was done by monitoring on-line the decay of
monochloramine signal when an aliquot of a saliva or stomach fluid sample was
added to the monochloramine solution continuously circulated through a direct
insertion membrane probe. The conclusion of the study was that any toxicity
associated with the ingestion of monochloramine is most probably due to the for-
mation of disinfection byproducts rather than absorption of intact inorganic
chloramine(s).
Measurement of dissolved gases (e.g., CH4, N2 , N2 0O,02 and CO,) in peat and
sediment samples with MIMS was shown to give valuable information about the
microbial physiology and processes (e.g., denitrification and methanogenesis) in
sediments and peat cores [29,110-113]. The advantage of the MIMS technique
compared to other methods is the possibility for simultaneous measurement of
multiple species [111]. Gas concentrations can be measured at 1 .M levels and
with spatial resolution of the order of 1 mm [112]. As an example vertical profiles
of dissolved oxygen, methane and carbon dioxide in a sub-core of a peat monolith
is shown in Fig. 16.12.
549
CH4 (pm)
I
Fig. 16.12. Dissolved oxygen (O), methane (A), and carbon dioxide (0) in a a
sub-core of a peat monolith. Water table corresponds to the surface of the
peat. (Reprinted from D. Lloyd, K. Thomas, D. Price, B. O'Neil, K. Oliver,
T.N. Williams, A membrane-inlet mass spectrometer miniprobe for the 10
direct simultaneous measurement of multiple gas species with spatial
- -,l of
-inn T T,/rrnl__l MtA_ I, A T-1 nn 1 A 1 lnanvrht
-tU -
UIU -
u.tJ -
lle ---
IJ~IUVULI. -l- ll-U V Vl ,]l.'-m3 V,3, 0 CO,(Jm) so
(1996), with permission from Elsevier Science.) CO2 (pm)
16.7.1 Soil
Recent studies have shown that the use of a static [13] or a dynamic [12,114,115]
headspace method combined with MIMS is very suitable for measurement of
VOCs from soil samples, slurry samples or other types of solid samples, widening
the applicability of MIMS considerably. It has also been shown that MIMS can be
used to measure the VOC content of subsurface soil samples using a penetrom-
eter membrane interface probe [116].
The performance characteristics of the dynamic headspace membrane inlet
method, i.e. the purge-and-membrane mass spectrometry (PAM-MS) method
have been studied thoroughly [12,114,115]. One of the most important results of
these studies is the observation that for most types of real soil samples quanti-
tation can be done using only one standard soil sample. This is due to the fact
that in principle in the PAM-MS analysis the entire amount of a particular VOC
is measured, since it is completely purged out of the sample, i.e. quantitation is
matrix independent. The PAM-MS method is a sensitive method with detection
limits of chlorinated and aromatic solvents in the range of a few /tg/kg, the
linearity is 3-4 orders of magnitude and relative standard deviations in determi-
nations of VOCs from soil samples is around 5%. The potential of the PAM-MS
method for fast screening of soil samples was proven by analysis of many real
environmental samples and comparing the results with data obtained by head-
space gas chromatographic methods in two different laboratories (Fig. 16.13)
[115]. Good correlation between the various methods was obtained.
Another very interesting MIMS application is the simultaneous in vivo mon-
itoring of gases in plants and soil around the plants [28]. The experimental
set-up of a multi-capillary membrane inlet system and an example of the results
obtained are shown in Fig. 16.14. The capillary membrane inlets are mounted
either into the soil or into the plant cavities. The gas concentrations in the soil
550
1000
0n -
800
o 700
pi 600
E 500
a 400
=) 3000
200
Fig. 16.13. Correlation of the 100
quantitative results obtain- 0
ed by HSGC and PAM-MS 200 400 600 800 1000
method. PAM, mg/kg
a) b) 4 Soil:brownforestsoil Plant:cornI
N,
3
1 0CO,
0-
lobo 2obo
S Soil: brownforestsoil Plant:tomato
CO,
.2
.0 0.8; u, N
-0
S
1.4 0o,
oo
0.
00 2doo 3doo00
O6- Soil: slightlyacidicsandysoil Plant:corn
W
a CO,
2.
8:
a 41 lo A, Nz ____ .k___
0-
10UU 2000 3000
Time (h)
Fig. 16.14. (a) The experimental set-up for in vivo monitoring of gases in plants and soils; (b)
measured gas concentrations in three different types of soils. The plant grown was corn or
tomato. (Reproduced with permission of John Wiley &Sons Limited from S. Bohdtka, Process
monitoring in fermentors and living plants by membrane inlet mass spectrometry, Rapid
Commun. Mass Spectrom. Vol 11, p. 656-661, Copyright 1997 by John Wiley &Sons Limited.)
have been observed to be characteristics of life in the soil and of the quality of the
soil.
16.7.2 Adsorbents
Adsorbents together with MIMS are used, as with other techniques, for sample
concentration before the concentrated sample is directed into the mass spectro-
meter [34,36,48]. This set-up can also be used to determine breakthrough curves
for various trapping materials and a thorough study of the stability of common
adsorbents such as Chromosorb, Carboxen, Tenax and Carbosieve has been
conducted [117].
551
16.7.3 Characterization of polymers
During the various development steps of MIMS, many different polymers have
been tested for their potential use as membranes in MIMS [118-121]. Polymers
tested include materials such as polyethylene [118,119], regenerated cellulose
[118], polypropylene [118], silicone rubber [118], polyethylene vinyl acetate
copolymer [118], poly(dimethylsiloxane) [119,120], latex [119], polyurethane
[119], nitrile [119], poly(vinyl) chloride [119], polyimide [119], Teflon [119],
poly(dimethylsiloxane) with zeolite [120], polyvinyl alcohol [120], and poly-
etherimide (polyester)/silicone composite membrane [121]. As an outcome of
these studies transport parameters such as the distribution ratio, the diffusion
constant and the permeability constant for many analytes diffusing through
various polymers have been established. In practice this possibility of determin-
ing transport parameters through polymers has been used to characterise the
transport of flavour compounds through polymer packing materials [122].
From a MIMS point of view, the study of many polymer materials has not
resulted in major progress. The poly(dimethylsiloxane) is still the best all-round
membrane for gas and VOC analysis, although Teflon offers a very good alterna-
tive for gas analysis. However, recent studies with low vapour pressure liquids as
membranes [123] and zeolite membranes for isomer separation [124], have
demonstrated interesting scientific advancement in this field.
In its present form, standard MIMS has reached a mature stage where it is hard
to imagine fundamental changes in the future. It is currently used routinely for
the detection of VOCs in liquid matrixes and every year new applications within
the analysis of gas and solid samples occur. There are still many unexplored
possibilities for the practical use of MIMS and here the simple sample handling,
the possibility for fast analysis and on-line monitoring are tempting advantages
for many potential users. From an industrial and environmental point of view
552
the technique might become very interesting along with the development of
compact and/or miniaturized MIMS systems for on-site monitoring purposes.
For example, at the 2001 American Society for Mass Spectrometry Conference
an underwater MIMS system that can operate autonomously or under user
control via a wireless rf-link was presented [130]. Such on-site MIMS systems
[131] are unique compared to other on-site sensors, because they allow analysis
of a relatively broad spectrum of compounds and the interference from matrix
components is relatively limited.
Recently, new MIMS methodologies have been developed that expand the
analytical capabilities of the MIMS technique with respect to the analysis of
larger and polar compounds. Steroid hormones, estrogenic compounds, pesti-
cides and polyaromatic hydrocarbons are currently detectable at relatively low
levels using stimulated desorption from the membrane surface. When fully
developed, these new systems might set new standards for on-site equipment.
Acknowledgements
TK thanks Raimo Ketola and Marja Ojala for the help in preparation of the
figures. The Danish Center for Water Quality Sensors (VAKS) and the National
Technology Agency (Tekes) and acknowledged for financial support.
REFERENCES
553
18 M.E. Cisper and P.H. Hemberger, Rapid Commun. Mass Spectrom., 11 (1997)
1449-1453.
19 M.H. Soni, J.H. Callahan and S.W. McElvany, Anal. Chem., 70 (1998) 3103.
20 R.A. Ketola and F.R. Lauritsen, Rapid Commun. Mass Spectrom., 13 (1999) 749-751.
21 F.R. Lauritsen and J. Rose, Analyst, 125 (2000) 1577-1581.
22 T. Aggerholm and F.R. Lauritsen, Rapid Commun. Mass Spectrom., 15 (2001)
1826-1831.
23 F.R. Lauritsen, T. Kotiaho and D. Lloyd, Biol. Mass Spectrom., 22 (1993) 585-589.
24 S. Chauvatcharin, K.B. Konstantinov, K. Fujiyama, T. Seki and T. Yoshide, Microb.
Util. Renewable Resour., 9 (1996) 493-506.
25 K.F. Hansen, S. Gylling and F.R. Lauritsen, Int. J. Mass Spectrom. Ion Process., 152
(1996) 143-155.
26 M.A. LaPack, J.C. Tou and C.G. Enke, Anal. Chem., 62 (1990) 1265-1271.
27 M. Bier, T. Kotiaho and R.G. Cooks, Anal. Chim. Acta, 231 (1990) 175-190.
28 S. Bohdtka, Rapid Commun. Mass Spectrom., 11 (1997) 656-661.
29 J. Benstead and D. Lloyd, FEMS Microbiol.Ecol., 13 (1994) 233-240.
30 F.R. Lauritsen, in: J.F.M. Van Impe et al. (Eds.), Advanced Instrumentation, Data
Interpretation,andControl of BiotechnologicalProcesses.Kluwer, 1998, pp. 67-103.
31 F.R. Lauritsen and A. Lunding, Enzyme Microbiol. Technol., 22 (1998) 459-465.
32 L.E. Slivon, M.R. Bauer, J.S. Ho and W.L. Budde, Anal. Chem., 63 (1991) 1335-1340.
33 L.E. Dejarme, S.J. Bauer, R.G. Cooks, F.R. Lauritsen, T. Kotiaho and T. Graf, Rapid
Commun. Mass Spectrom., 7 (1993) 935-942.
34 A.A. Rivlin, Rapid Commun. Mass Spectrom., 9 (1995) 397-399.
35 M.A. Mendes, R.S. Pimpirn, T. Kotiaho and M.N. Eberlin, Anal. Chem., 68 (1996)
3502-3506.
36 R.A. Ketola, C. Gr0n and F.R. Lauritsen, Rapid Commun. Mass Spectrom., 12 (1998)
773-778.
37 T. Kotiaho, R.A. Ketola, M. Ojala, T. Mansikka and R. Kostiainen, Am. Env. Lab.,
3/97 (1997) 19-21.
38 R.A. Ketola, Method Development in Membrane Inlet Mass Spectrometry, Air
Analysis and Desorption Techniques. VTT Publications, 364, 1998, p. 90.
39 R.A. Ketola, M. Ojala, H. Sorsa, T. Kotiaho and R.K. Kostiainen, Anal. Chim. Acta,
349 (1997) 359-365.
40 M.E. Cisper and P.H. Hemberger, Rapid Commun. Mass Spectrom., 11 (1997)
1454-1456.
41 S.M. Gordon, P.J. Callahan, D.V. Kenny and J.D. Pleil, Rapid Commun. Mass
Spectrom., 10 (1996) 1038-1046.
42 A. Colorado, D.J. Barket Jr., J.M. Hurst and P.B. Shepson, Anal. Chem., 70 (1998)
5129-5135.
43 E. Pungor Jr., C.R. Perley, C.L. Cooney and J.C. Weaver, Biotechnol. Lett., 2 (1980)
409-414.
44 H. Degn, J. Microbiol.Meth., 15 (1992) 185-197.
45 E. Heinzle, J. Biotechnol., 25 (1992) 81-114.
46 E. Heinzle and M. Reuss (Eds), Mass Spectrometry in Biotechnological Process
Analysis and Control. Plenum Press, New York, 1987, 241 pp.
47 T. Kotiaho, F.R. Lauritsen, T.K. Choudhury, R.G. Cooks and G.T. Tsao, Anal. Chem.,
63 (1991) 875A-883A.
48 G. Matz, A. Walte, W. Miinchmeyer and H.-E. Rikeit, Soc. Automot. Eng. [Spec.
Publ.] SP-1161 (1996) 175-179.
49 H.K. Wilson and T.W. Ottley, Biomed. Mass Spectrom., 8 (1981) 606-610.
50 T. Kotiaho, B.S. Shay, R.G. Cooks and M.N. Eberlin, J. Am. Chem. Soc., 115 (1993)
1004-1014.
554
51 B.J. Harland and P.J. Nicholson, Sci. Total Environ., 135 (1993) 37-54.
52 V.T. Virkki, R.A. Ketola, M. Ojala, T. Kotiaho, V. Komppa, A. Grove and S. Facchetti,
Anal. Chem., 67 (1995) 1421-1425.
53 M. Soni, S. Bauer, J.W. Amy, P. Wong and R.G. Cooks, Anal. Chem., 67 (1995)
1409-1412.
54 R.A. Ketola, V.T. Virkki, M. Ojala, V. Komppa and T. Kotiaho, Talanta, 44 (1997)
373-382.
55 T. Kotiaho, R. Kostiainen, R.A. Ketola, T. Mansikka, I. Mattila, V. Komppa, T.
Honkanen, K. Wickstr6m, J. Waldvogel and O. Pilvio, Process Contr. Qual., 11 (1998)
71-78.
56 M. Ojala, R.A. Ketola and T. Kotiaho, LC-GC, 16 (1998) 1026-1032.
57 R.M. Alberici, R. Sparrapan, W.F. Jardim and M.N. Eberlin, Environ. Sci. Technol.,
35 (2001) 2084-2088.
58 T. Kotiaho, R. Kostiainen, R.A. Ketola, M. Ojala, I. Mattila and T. Mansikka, in: E.J.
Karjalainen, A.E. Hesso, J.E. Jalonen and U.P. Karjalainen (Eds.), Advances in Mass
Spectrometry. Elsevier, Vol. 14, 1998, pp. 501-527.
59 F.R. Lauritsen and T. Kotiaho, Rev. Anal. Chem., 15 (1996) 237-264.
60 P.S.H. Wong, R.G. Cooks, M.E. Cisper and P.H. Hemberger, Environ. Sci. Technol.,
29 (1995) 215A-218A.
61 F.R. Lauritsen and R.A. Ketola, Anal. Chem., 69 (1997) 4917-4922.
62 G. Matz, M. Loogk and F. Lennemann, J. Chromatogr. A, 819 (1998) 51-60.
63 G. Matz, G. Kibelka, J. Dahl and F. Lennemann, J. Chromatogr. A, 830 (1999)
365-376.
64 F.R. Lauritsen, M.A. Mendes and T. Aggerholm, Analyst, 125 (2000) 211-215.
65 M.A. Mendes and M.N. Eberlin, Analyst, 125 (2000) 21-24.
66 M.J. Hayward, T. Kotiaho, A.K. Lister, R.G. Cooks, G.D. Austin, R. Narayan and
G.T. Tsao, Anal. Chem., 62 (1990) 1798-1804.
67 H.C. Beck, F.R. Lauritsen, J.S. Patrick and R.G. Cooks, Biotechnol. Bioeng., 51
(1996) 23-32.
68 F.R. Lauritsen, L.T. Nielsen, H. Degn, D. Lloyd and S. Bohdtka, Biol. Mass
Spectrom., 20 (1991) 253-258.
69 D. Lloyd, F.R. Lauritsen and H. Degn, J. Gen. Microbiol., 137 (1991) 1743-1747.
70 R.A. Ketola, M. Ojala, V. Komppa, T. Kotiaho, J. Juujarvi and J. Heikkonen, Rapid
Commun. Mass Spectrom., 13 (1999) 654-662.
71 G. Matz, W. Schr6der, A. Harder, A. Schillings and P. Rechenbach, Field Anal. Chem.
Technol., 1 (1997) 181-194.
72 G. Matz and G. Kibelka, in Proceedingsof the PITTCON'98,New Orleans, Lousiana,
March 1-5, 1998, 1864 pp.
73 F.R. Lauritsen, T.K. Choudhury, L.E. Dejarme and R.G. Cooks, Anal. Chim. Acta,
266 (1992) 1-12.
74 F.F. Vincieri, N. Mulinacci, G. Mazzi, D. Favretto, M. D'Alpaos and P. Traldi, Eur.
Mass Spectrom., 1 (1995) 503-505.
75 D. Favretto, G. Grandis, G. Allegri and P. Traldi, Rapid Commun. Mass Spectrom.,
12 (1998) 1595-1600.
76 R.A. Ketola, J. Heikkonen, S. Piepponen, F.R. Lauritsen and T. Kotiaho, Rapid
Commun. Mass Spectrom., 12 (1998) 1011-1017.
77 K.F. Hansen, F.R. Lauritsen and H. Degn, Biotechnol. Bioeng., 44 (1994) 347-353.
78 F.R. Lauritsen and S. Gylling, Anal. Chem., 67 (1995) 1418-1420.
79 G. Matz and F. Lennemann, J. Chromatogr. A, 750 (1996) 141-149.
80 J.-P. Arcangeli, E. Arvin, M. Mejlhede and F.R. Lauritsen, Water Res., 30 (1996)
1885-1893.
81 R.C. Johnson, N. Srinivasan, R.G. Cooks and D. Schell, Rapid Commun. Mass
555
Spectrom., 11 (1997) 363-367.
82 P.J. Savickas, M.A. LaPack and J.C.Tou, Anal. Chem., 61 (1989) 2332-2336.
83 T. Kotiaho, A.K. Lister, M.J. Hayward and R.G. Cooks, Talanta, 38 (1991) 195-200.
84 T. Kotiaho, M.J. Hayward and R.G. Cooks, Anal. Chem., 63 (1991) 1794-1801.
85 M. Gazda, L.E. Dejarme, T.K. Choudhury, R.G. Cooks and D.W. Margerum, Environ.
Sci. Technol., 27 (1993) 557-561.
86 E.J. Pedersen, E.T.Urbansky, B.J. Marinas and D.W. Margerum, Environ. Sci.
Technol., 33 (1999) 4239-4249.
87 C. Shang, W.-L. Gong and E.R. Blatchley III, Environ. Sci. Technol., 34 (2000)
1721-1728.
88 H. Degn and F.R. Lauritsen, J. Phys. Chem., 93 (1989) 2781-2783.
89 R.C, Johnson, K. Koch and R.G. Cooks, Ind. Eng. Chem. Res., 38 (1999) 343-351.
90 R.F.P. Nogueira, R.M. Alberici, M.A. Mendes, W.F. Jardim and M.N. Eberlin, Ind.
Eng. Chem. Res., 38 (1999) 1754-1758.
91 P.SH. Wong, N. Srinivasan, N. Kasthurikrishnan, R.G. Cooks, J.A. Pincock and J.S.
Grossert, J. Org. Chem., 61 (1996) 6627-6632.
92 R. Augusti, A.O. Dias, L.L. Rocha and R.M. Lago, J. Phys. Chem. A, 102 (1998)
10723-10727.
93 S. Bohatka and H. Degn, Rapid Commun. Mass Spectrom., 5 (1991) 433-436.
94 H. Degn, S. Bohdtka and D. Lloyd, Biotechnol. Techniques, 6 (1992) 161-164.
95 H. Degn and N.-U. Frigaard, Anal. Chim. Acta, 274 (1993) 355-359.
96 T. Kotiaho, S. Gylling, A. Lunding and F.R. Lauritsen, J. Agric. Food Chem., 43
(1995) 928-930.
97 D. Favretto, P. Traldi, C.A. Benassi and A. Bettero, Biol. Mass Spectrom., 20 (1991)
669-676.
98 P.R. Jones and S.K. Yang, Anal. Chem., 47 (1975) 1000-1003.
99 S. Ouyang, Y.H. Chen and Y. Xu, Anal. Chim. Acta, 337 (1997) 165-172.
100 S. Ouyang, Y. Xu and Y.H. Chen, Anal. Chem., 70 (1998) 931-935.
101 J.S. Lundsgaard, B. Jensen and J. Gronlund, J. Appl. Physiol.: Respirat. Environ.
Exercise Physiol., 48(2) (1980) 376-381.
102 T. Kiaer, B. Dahl and G. Lausten, J. Orthop. Res., 10 (1992) 807-812.
103 M. Roberts, E.T. Colton III, G. Owens, D.D. Thomas and G.M. Watkins, Med. Biol.
Eng., 13 (1975) 535-538.
104 J.S. Brodbelt, R.G. Cooks, J.C. Ton, G.J. Kallos and M.D. Dryzga, Anal. Chem., 59
(1987) 454-458.
105 W.E. Donavan and B. Myers, in: H.I. Bicher and D.F. Bruley (Eds), Oxygen Transport
to Tissue, Instrumentation,Methods and Physiology. Plenum Press, New York, 1973,
pp. 67-72.
106 E.V. Trushina, N.J. Clarke, L.M. Benson, A.J. Tomlinson, C.T. McMurray and S.
Naylor, Rapid Commun. Mass Spectrom., 12 (1998) 985-987.
107 A. Sturaro, L. Doretti, G. Parvoli and P. Traldi, Anal. Chim. Acta, 224 (1989)
119-122.
108 D. Lloyd, J.E. Ellis, K. Hillman and A.G. Williams, J. Appl. Bacteriol. Symp. Suppl.,
73 (1992) 155S-163S.
109 T. Kotiaho, J.M. Wood, P.L. Wick Jr., L.E. Dejarme, A. Ranasinghe, R.G. Cooks and
H.P. Ringhand, Environ. Sci. Technol., 26 (1992) 302-306.
110 K.L. Thomas and D. Lloyd, FEMS Microbiol. Ecol., 16 (1995) 103-114.
111 K.L. Thomas, D. Price and D. Lloyd, J. Microbiol. Meth., 24 (1995) 191-198.
112 D. Lloyd, K. Thomas, D. Price, B. O'Neil, K. Oliver and T.N. Williams, J. Microbiol.
Meth., 25 (1996) 145-151.
113 G. Cowie and D. Lloyd, J. Microbiol. Meth., 35 (1999) 1-12.
114 M. Ojala, I. Mattila, T. Srme, R.A. Ketola and T. Kotiaho, Analyst, 124 (1999)
556
1421-1424.
115 M. Ojala, I. Mattila, V. Tarkiainen, T. Srme, R.A. Ketola, A. Miittanen, R.
Kostiainen and T. Kotiaho, Anal. Chem., 73 (2001) 3624-3631.
116 J. Costanza and W.M. Davis, Field Anal. Chem. Technol., 4 (2000) 246-254.
117 R.J.B. Peters and H.A. Bakkeren, Analyst, 119 (1994) 71-74.
118 L.B. Westover, J.C. Ton and J.H. Mark, Anal. Chem., 46 (1974) 568-571.
119 A.J. Maden and M.J. Hayward, Anal. Chem., 68 (1996) 1805-1811.
120 N. Kasthurikrishnan, R.G. Cooks and S. Bauer, Rapid Comm. Mass Spectrom., 10
(1996) 751-756.
121 R. Alberici, R. Sparrapan, M.E. Eberlin, D. Windmoller and R. Augusti, Anal.
Commun., 36 (1999) 221-223.
122 J.C. Tou, D.C. Rulf and P.T. DeLassus, Anal. Chem., 62 (1990) 592-597.
123 R.C. Jonhson, K. Koch, N. Kashuriksrishnan, W. Plass, J.S. Patrick and R.G. Cooks,
J. Mass Spectrom., 32 (1997) 1299-1304.
124 K.H. Bennett, K.D. Cook, J.L. Falconer and R.D. Noble, Anal. Chem., 71 (1999)
1016-1020.
125 R.J. Vreeken and R. Houriet, Anal. Chim. Acta, 313 (1995) 237-241.
126 A.D. Hendricker, F. Basile and K.J. Voorhees, J. Anal. Appl. Pyrol., 46 (1998) 65-82.
127 A.D. Hendricker, C. Abbas-Hawks, F. Basile, K.J. Voorhees and T.L. Hadfield, Int. J.
Mass Spectrom., 190/191 (1999) 331-342.
128. N. Kasthurikrishnan, R.G. Cooks and M.J. Thompson, J. Am. Soc. Mass Spectrom., 9
(1998) 234-241.
129 M. Ojala, M.Poutanen, I. Mattila, R.A. Ketola, T. Kotiaho and R. Kostiainen, Rapid
Commun. Mass Spectrom., 14 (2000) 994-998.
130 D.P. Fries, R.T. Short, G. Kibelka and M.L. Kerr, 49th ASMS Conference on Mass
Spectrometry andAllied Topics, Chicago, 2001.
131 G. Matz, W. Schr6der and T. Kotiaho, in: R.M. Meyers (Ed.), Encyclopedia of
Analytical Chemistry: Applications, Theory, and Instrumentation, Vol. 5. Wiley,
2000, pp. 3783-3804.
557
Chapter 17
17.1 INTRODUCTION
Addition of solvent
[Heat pressuree
EXTRACTION TECHNIQUE |
MAE/SFE/PFE, which might not only be vague in terms of solvent choice but
might have situations where the operation of the instrument requires some
intervention. These method variabilities are not the fault of the regulators who
produce the methods, but are necessary because of the complex interactions of
the analytes with it.
Analyte-matrix interactions obviously need to be overcome in order to
release the analyte and make it available for subsequent analysis. The nature of
these interactions is very dependent upon the type of molecule to be removed.
Organic molecules occur in different forms ranging from the non-polar (e.g.,
PAHs) through to more polar molecules (e.g., phenols). While the former may be
only weakly bound (Van der Waals forces) to the matrix, polar molecules may be
strongly bound and have the possibility of electrostatic attraction, depending
upon the pH of the matrix. In this way it is difficult to generalise on the exact
nature of extraction technique and solvent choice. This is where the skill and
experience of the analyst becomes important. Often the choice of solvent to be
used to remove an analyte from a matrix is based on past experience. That
assumes that the analyte(s) to be extracted are known. For unknown samples,
several extractions may be required with at least two different sample types to
assess the effectiveness of a particular solvent system.
Extraction of organic pollutants from solid environmental samples has
frequently been done using organic solvents with or without the addition of heat
(Fig. 17.1). This process is typified by the technique of Soxhlet extraction. A flow
diagram detailing a procedure for the extraction of analytes from soils using
Soxhlet extraction is shown in Fig. 17.2. Soxhlet extraction normally requires
large volumes (up to 150 ml) of often chlorinated solvent to be refluxed through
the solid sample for an extended period of time (6-24 h). While it is not
uncommon for the analyst to assemble several experimental set-ups at the same
time it can be labour-intensive at the start and end of the process. As the process
is time-consuming, alternative extraction strategies have been developed, based
on instrumentation that allows the process to be controlled more effectively.
This chapter mainly focuses on these newer instrumental approaches to
compare and contrast both their application base and robustness for extraction.
560
Soxhlet apparatus assembled. Sample e.g. soil, accurately weighed
Solvent placed in round-bottomed I and mixed with similar weight of
flask on isomantle anhydrous sodium sulphate.
Concentration of analyte by
solvent evaporation (e.g.
rotary evaporator, solvent
blow-down) or SPE.
I~~~~~ \\\
=Analysis
cP
gas
561
TABLE 17.1
Critical temperature and pressures for selected solvents
Solvent Critical temp. (C) Critical pressure
Atmospheres psi
Carbon dioxide 31.1 74.8 1070.4
Nitrous oxide 36.6 73.4 1050.1
Water 374.4 224.1 3208.2
Xenon 16.7 59.2 847.0
17.2.2 Instrumentation
All SFE systems contain six basic components, namely a supply of high purity
carbon dioxide (CO2), a supply of high purity organic modifier, the pumps, the
oven for the extraction cell, the pressure outlet or restrictor, and the collection
vessel. A schematic diagram of a typical SFE arrangement is shown in Fig. 17.4.
The CO2 is supplied in a cylinder fitted with a dip tube that allows liquefied CO 2
only to be obtained. This is in contrast to the common cylinder type which
562
comprises liquid CO2 at the bottom and vapour at the top (when positioned
vertically). Modified CO2 can be produced via the use of two pumps or via a single
cylinder containing CO2 and an organic modifier; the latter is the preferred
approach.
Two pumps are required to pump the pressurized CO2 and organic modifier
through the SFE system. After pumping out from the cylinder, CO2 and organic
modifier are mixed with a T-piece before introducing into the extraction cell.
The combination of the two pumps allows a degree of control in terms of modifier
composition (1-20% v/v) and flexibility of solvent choice, e.g., methanol, toluene.
The ideal pump should deliver a constant flow rate (ml min -') at a suitable pres-
sure (3500-10000 psi). Two kinds of pump-the reciprocating pump (piston
pump) and syringe pump-are usually used in SFE systems, but the former is
the most common due to its lower cost. However, when the reciprocating pump
is employed, external cooling of the pump head is required to prevent cavitation,
i.e., gas entrapment. No such modification of the pump head is required for the
modifier.
An oven is used to generate the critical temperature of the solvent. Located
within the oven is the sample cell or vessel. The temperature range of the oven is
ideally located between T. and 250°C. The sample vessel, which is typically made
of stainless steel, must be capable of safely withstanding high pressures (up to
10000 psi). It is essential to remember that some equilibration time is required
to achieve the pre-set temperature for the newly prepared and inserted sample-
containing vessel prior to starting the extraction process. Almost all commercial
extraction vessels are of the flow-through design. This allows fresh, unadulter-
ated supercritical fluid to pass over the sample.
The first commercial instruments relied on an inadequate method of pres-
sure regulation based on the use of narrow-bore silica or stainless steel tubing
which acted as the restrictor. These were largely superseded in the 1990s by
more robust methods of pressure regulation based on the use of 'moving ori-
fices', typified by the use of such terms as 'nozzles' or 'back-pressure regulators'.
For the extraction of real sample it is more appropriate to use a variable
restrictor.
Supercritical fluid extraction relies on the diversity of properties exhibited
by the supercritical fluid to (selectively) extract analytes from solid, semi-solid or
liquid matrices. The important properties offered by a supercritical fluid for
extraction are: (a) good solvating power; (b) high diffusivity and low viscosity,
and (c) minimal surface tension.
SFE can be operated in two modes: off-line and on-line. While a significant
part of the early work was done on utilising the compatibility of SFE to
chromatographic separation, the lack of robust instruments and the inflexibility
of such systems has probably precluded their continued usage. The situation is
reversed for off-line SFE. In this situation, the flexibility of off-line operation
allows the analyst to focus on the sample preparation only. This allows SFE to be
optimised to maximise analyte recovery, to process larger samples without fear
563
of overloading the chromatographic column, a certain amount of freedom of
choice pertaining to the analysis (e.g. GC, HPLC, IR etc.), and finally, that the
analytical measurement instrument is available to analyse other samples.
It would appear however that the initial growth in the popularity of SFE,
based on its environmentally friendly solvent (CO,), has largely been ignored in
recent years, at least in the scientific literature. However, SFE has found a niche
market. Three U.S. EPA methods exist for the SFE of PAHs (method 3561), total
petroleum hydrocarbons, TPHs (method 3560) and PCBs and organochlorine
pesticides (method 3562) from soils.
Many reviews on SFE have been published with an environmental bias [1-7]. It
is not the intention of this chapter to review the extensive application literature
that exists for SFE. Accordingly, a short table of applications has been provided
(Table 17.2) to highlight suitable analytes and conditions. Based on this and
other information a set of recommendations for the use of SFE are proposed.
TABLE 17.2
A summary of supercritical fluid extraction (SFE) using for extracting PAHs, PCBs and
pesticides from soils
continued
564
Analyte(s) Sample types/ Operating conditions Recovery (%) Ref.
matrices
PCB conge- Samnle: Fluid: pure CO. or CO. with quantitative 11
ners industrial soil methanol modifier (2-5%).
(CRM 481) Different conditions employed in
21 different laboratories.
565
SFE system operational and Sample e.g. soil, accurately weighed
connected to electric and gas and mixed with similar weight of
supply. A two-pump system is drying agent and copper (for soils
required for extraction of polar with high sulphur content).
analytes.
i-1
Fig. 17.5. Flow diagram for supercritical fluid extraction.
566
resulted in temperatures far greater than the atmospheric boiling points of the
liquids. This elevated boiling phenomenon is due to the special way in which
microwaves induce heat.
On a standard hot-plate or heating mantle heating system, heat is trans-
ferred from the hot-plate, through the walls of the vessel and then into the liquid
contained within it. This is a relatively slow process, which results in a tempera-
ture gradient whereby the liquid is hottest at the walls of the vessel and coolest
in the centre.
In contrast, microwave heating works by exciting the molecules in the liquid
causing heating by either rotation of dipoles or charges within the molecules of
the liquid. Microwaves have a specific wavelength of 12.2 cm or a frequency of
2.45 GHz set so as not to interfere with radar and telecommunication networks.
The presence of an electric field generated by the microwaves would cause polar
molecules to align themselves with the field. However, since microwave fields
oscillate, the molecules align and relax with the field of electromagnetic radia-
tion. It is the energy dissipated during such movements, which leads to the
heating effect (Fig. 17.6).
The physical parameter, which describes the ability of a material to heat
when placed in the microwave field, is termed the dielectric loss coefficient. The
larger this value, the more polar the solvent and the greater the ability of the sol-
vent to transform electromagnetic energy into heat via mechanical motion. In
order for microwave heating to occur, one of the following three conditions
should be used [18]: (1) a single solvent or mixture of solvents that have high
dielectric loss coefficients; (2) a mixture of solvents with high and low dielectric
loss; (3) a microwave susceptible sample that has a high dielectric loss in a low
dielectric loss solvent.
The increase in temperature, which can be reached by microwave heating,
accelerates the rate at which extraction takes place by increasing the solubility
of the analytes in the extraction solvent. This means that the sample throughput
can be increased. In addition, since far smaller volumes of solvent are needed for
MAE as compared to Soxhlet methods, this reduces costs. Parr and co-workers,
who first showed how microwaves could be used in extractions from plant and
animal tissues, described the fundamental physical phenomena involved [19].
The disadvantage of this technology is the fact that the solvent needs to be physi-
567
cally removed from the sample matrix upon completion of the extraction prior to
analysis.
17.3.2 Instrumentation
The first use of MAE for the extraction of analytes with organic solvents
appeared in 1986 [17]. In MAE, both the organic solvent and the sample are
subjected to radiation from a magnetron and since microwaves are able to cause
rapid heating of samples, safety is an important issue.
Although domestic microwave ovens can be used for extraction, safety
considerations have led to a number of manufacturers producing microwave
ovens specially designed for laboratory use. The advantages of these dedicated
instruments over domestic ovens especially in terms of safety are great,
although they are considerably more expensive.
Initially, MAE was always performed using closed-vessels and involved
heating the solvents under pressure beyond their boiling point and consequently
these experiments carried a risk of explosion. For example, reactions that cause
decomposition using acids often produce gaseous by-products, which can result
in a rapid increase in pressure in a closed vessel. If a sample of a polymer
containing a large number of the same type of bonds is heated, when the bonds
break there will be a rapid release of carbon dioxide. It is because of this fact that
sample size is limited to less than 2 g in closed vessels.
This problem led researchers to develop microwave systems that could be
open to the atmosphere which, although limiting the temperature to which the
solvent could be heated, would handle larger sample sizes since gaseous release
would no longer be a problem. Such systems are referred to as "open-vessel
systems" and can be used for sample sizes greater than 2 g.
Laboratory microwave ovens are available in both closed-vessel style and
open-vessel style. The former is typified by the MES-1000 Microwave Solvent
Extraction System supplied by CEM Corp., USA (recently superseded by the
MARS5 system), and the latter by the Soxwave system from Prolabo Ltd.,
France (now, CEM Corp.). In addition, some laboratory microwave ovens such as
QLAB 6000 developed by Questron Technologies Corp., Canada, can be set for
closed or open vessel operation or flow-through. Figures 17.7 and 17.8 are sche-
matic diagrams of pressurised (closed) and atmospheric (open) MAE systems. As
an example of the current technology, the MARS5 system from CEM Corp.
allows up to 14 extraction vessels (XP-1500 plusTM ) to be irradiated simulta-
neously. Safety features include a function for monitoring both pressure and
temperature and the instrument is equipped with a solvent alarm to call atten-
tion to an unexpected release of flammable and toxic organic solvent. Pressure
(up to 800 psi) is continuously measured (measurements taken at the rate of 200
s-]), while the temperature (up to 300°C) is monitored for all vessels every 7 s.
The sample and solvent are placed into the inert vessel liner, which has a
volume of 100 ml and is made from a TFM fluoropolymer. The vessels are
568
Fig. 17.7. Pressurised-vessel
microwave system.
equipped with an AutoVent Plus "' system, which allows venting of excess
pressure within each extraction vessel. This works by lifting of the vessel cap
slightly to release any excess pressure and then immediately resealing to prevent
loss of sample. If any solvent leaks from the extraction vessel(s), a solvent
monitoring system will automatically shut off the magnetron but leaves the
exhaust fan to continue working.
The use of microwaves for environmental sample preparation has increased
at a very fast rate over the past 15 years. Many traditional Soxhlet extraction
procedures were adapted to be carried out much faster and with much less
solvent in a microwave oven. The biggest hindrance to this process was the
reduced in sample size, which could be accommodated safely in a pressurised
microwave system.
However, with the advent of atmospheric microwave systems, much larger
sample sizes can be used and this, coupled with the lower limits of detection
offered by modern chromatographic equipment, has meant that most traditional
extractions can be performed in a microwave oven. To illustrate this, there are
now a number of EPA methods, based on microwave extraction, which would
previously have employed a different extraction technique such as Soxhlet.
Although microwaves have found a place in many routine environmental
analysis laboratories, there is a problem with the through-put capabilities of
these instruments. In comparison with Soxhlet they are certainly no worse, but
there is a growing need to further increase the number of sample preparations
which can be performed per day. The systems described above are batch pro-
569
cesses, i.e. each microwave vessel holds a single discrete sample. A maximum of
14 samples can be irradiated simultaneously and all vessels must be allowed to
cool before opening to prevent loss of volatile materials. If it were possible to
perform these extractions in a continuous system, the increase in through-put
would be vast.
CEM and Milestone (now part of CEM) have manufactured continuous flow
systems but the problem with these systems lies in the way in which the
magnetron produces the microwaves. The magnetrons on these systems work on
a duty cycle. This means that the time the microwaves are produced is varied
rather than the power of the microwaves. For example, to obtain 50% power, the
magnetron would be on full-power for 10 s and then off for 10 s so on average,
50% power. This is a problem for continuous flow samples, because if the sample
is passing through the microwave cavity during the power-off section of the duty
cycle, there will be no irradiation of that sample.
The problem caused by duty cycles in closed-vessel systems has been elimi-
nated by the production of the MARS5 instrument, which works on a variable
power not variable time magnetron. If this technology could be introduced into
continuous-flow systems, the impact on MAE would be great. Routine analyses
could be performed much quicker and would require much lower volumes of
solvents than used with traditional Soxhlet extraction methods causing consid-
erable cost reductions.
Many reviews on MAE have been published with an environmental bias [18,
20-21]. It is not the intention of this chapter to review the extensive application
literature that exists for MAE. Accordingly, a short table of applications has
been provided (Table 17.3) to highlight suitable analytes and conditions. Based
on this and other information a set of recommendations for the use of MAE are
proposed.
570
TABLE 17.3
A summary of microwave-assisted extraction (MAE) using for extracting PAHs, PCBs and
pesticides from soils
Analyte(s) Sample types / Operating conditions Recovery (%) Ref.
matrices
PAHs, OCPs, etc. Samples: standard Closed-vessel system; Sample 75 (average) 22
reference sediments size: 5 g. Extractants: ace-
(HS-3, HS-4, HS-5, tone-hexane (1:1, v/v), 30 ml;
and SRM 1941) and A temperature of 115°C and
soils (SRS 103-100, an extraction time of 5 min
and ERA lot no. were selected.
321)
PAHs, PCBs, pes- Samples: marine Domestic oven; Sample size: 97-102 25
ticides, etc. sediment. 2-10 g. Solvents: 10-ml toluene
+ 1-ml H2O; Extraction at 660
w for 6 min.
PAHs Sample: sewage Closed-vessel system; 97.7-117.3 26
sludge. Extractants: acetone-hexane
(1:3), 30 ml;
Temperature: 130°C.
PCBs Sample: soil, sewage Idem, but temperature 80°C. 123.4-149 26
sludge.
(continued)
571
TABLE 17.3 continuation
572
MAE system: atmospheric or Sample e.g. soil, accurately weighed,
pressurised. Check that system is and placed in extraction vessel along
operational and connected to with solvent (hexane:acetone, 1:1, v/v,
electric supply. acetone or DCM only). 30-45 ml of
solvent per 2-5 g of sample.
I-
MAE conditions: pressurised system. MAE conditions: atmospheric system.
Pressure, <200 psi; Temperature, Temperature, bp of solvent; extraction
110-145 C; extraction time, 5-20 time, 5-20 mins; microwave power,
mins; microwave power, 100% at 900 100% at 300 W. Cooling time of
W. Cooling time of extraction vessels, extraction vessel, negligible.
20-30 mins, if unaided.
Analysis
- When extracting non-polar substances such as PAHs, use a 1:1 v/v mixture of
acetone and hexane to ensure high temperatures and good solubility of
analytes in solvent.
A flow diagram detailing a procedure for the extraction of analytes from soils is
shown in Fig. 17.9.
Pressurised fluid extraction (PFE) has become the generic name for the extrac-
tion technique originally launched by Dionex Inc., in Spring 1995, called acceler-
ated solvent extraction (ASE ). The technique relies on the use of temperature
and pressure to extract organic compounds from different environmental,
industrial and food matrices.
573
Solubility and mass transfereffects
Three factors are considered important:
- Higher temperature increases the capacity of solvents to solubilise analytes.
- Faster diffusion rates occur as a result of increased temperature.
- Improved mass transfer and hence increased extraction rates occur when
fresh solvent is introduced, i.e., the concentration gradient is greater
between the fresh solvent and the surface of the sample matrix.
Disruptionof surface equilibria
As both temperature and pressure are important both are discussed separately.
Temperature effects
- Increased temperatures can disrupt the strong solute-matrix interactions
caused by van der Waal's forces, hydrogen bonding, and dipole attractions of
the solute molecules and active sites on the matrix.
- A decrease in viscosity and surface tension of solvents occurs at higher
temperature thus allowing improved penetration of the matrix, and hence
improved extraction.
Pressure effects
- The utilisation of elevated pressures allows solvents to remain liquified
above their boiling points.
- Extraction from within the matrix is possible, as the pressure allows the
solvent to penetrate the sample matrix.
17.4.2 Instrumentation
action
574
extractor can operate with up to 24 sample-containing extraction vessels and up
to 26 collection vials plus an additional four vial positions for rinse/waste
collection. All the sample vessels have the same internal diameter of 19 mm, but
six volume sizes are available: 0.5, 1, 5, 11, 22 and 33 ml. With two removable end
caps, the sample vessel is easy to clean and fill with sample. The step-by-step
sample-filling procedure is as follows: screw one of the sample vessel's end caps
on to finger-tightness; then, introduce a filter paper into the vessel followed by
the sample itself; lastly, screw the other end cap on to finger-tightness and then
place the vessel in the carousel. Before performing extraction, an auto-seal
actuator places the selected extraction vessel into the oven.
For doing extraction, the sample cell vessel is positioned vertically in the
oven, and filled with the selected solvent (or solvent mixture) by the solvent sup-
ply system. Then, the cell is heated to a pre-set temperature (up to 200°C) and
pressure (up to 20 MPa) and held for a few minutes (typically 5 min). After com-
pletion, the static valves are released. Then, a few ml of clean solvent passes
through the extraction cell to exclude the existing solvent(s) and extracted
analytes. Finally, a gas-purging process follows in which N2 gas evacuates both
the sample vessel and the stainless steel transfer line. All extracted analytes pass
through the stainless steel tubing that punctures a septum (solvent resistant)
located on the top of the collection vial (40 or 60 ml capacity). The vessel will be
automatically returned to the carousel after extraction. Moreover, the inherent
safety features adopted by the PFE system include an IR sensor to monitor the
arrival and level of solvent in the collection vial, as well as an automatic shut-off
procedure that initiates in the case of system failure.
575
TABLE 17.4
A summary of pressurised fluid extraction (PFE) using for extracting PAHs, PCBs and pesticides
from soils
Analyte(s) Sample types/ Operating conditions Recovery Ref.
matrices (%)
PAHs Sample: native Sample size: 7-9 g. Solvents: acetone- DCM quantitative 35
contaminated (1:1, v/v); Temperature: 100 C; Pressure:
soil. 2000 psi; Time: 5-min heat-up plus 5-min
static extraction.
PAHs Sample: real soil. Sample size: 20 g. Solvents: hexane- quantitative 36
acetone-toluene (10:5:1); Temperature:
100°C; Pressure: 13.8 MPa; Time: 5-min
heat-up plus 10-min static extraction.
PAHs, Sample 1: urban Sample size: 1-5 g. quantitative 37
OCPs, dust (SRM 1649a) Solvents: DCM; Temperature: 100°C;
PCB Sample 2: sediment Pressure: 2000 psi; Time: 5-min heat-up
congeners (SRM 1944). Sample plus 5-min static extraction.
3: marine sediment
(SRM 1941a)
PAHs Sample: contam- Solvents: acetone-hexane (1:1, v/v) quantitative 38
inated soils or acetone-DCM (1:1, v/v); Temperature:
100°C; Pressure: 10 MPa/14 MPa; Time:
5-min heat-up plus 2x5-min static extrac-
tion. Solvents: toluene; Temperature:
175/200C; Pressure: 14 MPa; Time: 8- min
heat-up plus 2x5-min static extraction.
OCPs Sample: contam-i idem quantitative 38
nated soils
PCDDs, Sample: fly ash Solvents: toluene; Temperature: quantitative 38
PCDFs 175/200°C; Pressure: 14 MPa; Time: 8-min
heat-up plus 2x10-min static extraction.
PCB Sample 1: marine Sample size: 0.5-1.5 g. Solvents: For sample 1: 39
congeners sediment (SRM 1944) acetone-DCM (1:1, v/v); Temperature: 83-112; For
Sample 2: sewage 100°C; Pressure: 2000 psi; Time: 5-min sample 2: 91-
sludge (BCR 392) heat-up plus 5-min static extraction. 109; For
Sample 3: harbor Others idem, but 2x5 min static sample 3: 93-
sediment (CRM 536) extraction. 123; For sam-
ple 3: 99-131.
Herbicides Sample: agricultural Sample size: 20 g. Solvents: methanol; 47-99 40
soil and spiked soil. Temperature: 125°C; Pressure: 10 MPa;
Time: combined heat-up and static
extraction, 10 min.
576
analytes. A mathematical approach for the selection of the optimum solvent
has been developed [42,43] (for further details see later).
- Optimise static/flush cycles. The success of Soxhlet extraction is based on its
ability to recycle fresh, clean solvent through the sample at the rate of
approximately 4 cycles per hour. While this means that a Soxhlet extraction
may take up to 24 hours to achieve quantitative recovery of analytes, the
process of successive flushing with fresh solvent is an important concept. It is
recommended in PFE that the cycling of solvent through the sample is
mimicked. In order to achieve this PFE can perform up to three static-flush
cycles in any single extraction, allowing the PFE to mimic the action of
Soxhlet extraction. It has been reported [44] that the number of static flush
cycles required to quantitatively extract chlorinated pesticides from soil is
two. It was concluded [44] that two successive 5 min cycles was more
effective at extraction than longer (10 and 15 min) static extraction times.
- Optimization of PFE conditions: the main PFE operating variables are:
temperature, pressure, time. In order to evaluate these three variables it is
necessary to select operational (safe working) limits prior to commencement.
Typically, the following limits can be used: pressure, 1000 and 2400 psi;
temperature, 40 and 200°C; and time, 2 and 16 min.
- The preferred approach in our own laboratory is to use a multivariate
approach using chemometrics [45].
A flow diagram detailing a procedure for the extraction of analytes from soils is
shown in Fig. 17.11.
17.5 MISCELLANEOUS
One approach is to use solid phase microextraction (SPME) for the extraction of
analytes from soil slurries. This is achieved by stirring a known quantity of soil
with a solvent (e.g., water) and then exposing the SPME fibre directly to the
resultant slurry. An initial attempt to demonstrate the applicability of the
approach was done by Cisper et al. [46]. The group were able to determine
pyrene and anthracene at the 10 ppm level by exposure of an SPME fibre to a
soil/methanol slurry spiked sample for 3 min. Analysis was via laser desorption
ion trap mass spectrometry. It was suggested that the addition of a suitable
internal standard to the slurry sample would allow quantitation. A polar
poly(acrylate) fibre was utilised by Boyd-Boland and Pawliszyn [47 ] for the
qualitative extraction of nitrogen-containing herbicides from an aqueous slurry.
The soil sample (garden lawn) had been previously treated with benfluralin; a
GC-MS chromatogram indicated the presence of the herbicide. The same group
[48] were able to obtain quantitative data for the assay of metachlor, a pesticide,
on soil. The soil sample was prepared by stirring 0.5 g in 4 ml of water and then
exposing the SPME fibre directly to the resultant slurry for 50 min. It was found
577
PFE system operational and Sample e.g. soil, accurately weighed
connected to electric and gas and mixed with drying agent and
supply. copper (for soils with high sulphur
content).
Analysis
that the level of metachlor in soil was 1.84 mg kg- and 1.85 mg kg by SPME and
Soxhlet extraction, respectively. This approach demonstrated the applicability
of SPME for the analysis of analytes from solid matrices. However, some concern
was expressed by the authors in terms of the limited applicability of the
approach as metachlor is a relatively water-soluble compound (solubility = 530
mg 1) and this is not likely to be the case for other compounds.
An alternative to the slurry method is to extract the solid sample with hot water
and then isolate the analytes from the water using SPME prior to chromato-
graphic separation and detection. This approach has been applied to a range of
semivolatile compounds of environmental interest including polycyclic aromatic
hydrocarbons (PAHs) [49]. Sub-critical water (250°C) is used to extract soil sam-
ples for up to 60 min. Then, the sample cell is cooled (by placing under tap water)
and a known volume (1.8 ml) of the water removed and the solubilised organic
analytes are isolated by SPME prior to GC-MS analysis. A 100 Lm film thickness
poly(dimethylsiloxane)-coated fibre was used. For the determination of PAHs,
internal standards of the perdeuterated forms of the PAHs under investigation
were added directly to the water in the extraction cell prior to the heating step.
578
Quantitation was based on a comparison of the GC/MS peak areas found for the
sample PAHs versus the peak areas (and spike quantities) of the same deuter-
ated PAH internal standards (e.g. phenanthrene concentrations were based on
the response of the phenanthrene-d,, spikes). The effect of recoveries, relative to
24 h Soxhlet extractions, for 11 PAHs from railroad bed soil were assessed in
terms of water extraction temperature (200 and 250°C), extraction time (15 and
60 min) and operator. The highest recoveries were obtained at 250°C where
recoveries for the PAHs ranged from 45 to 197% (15 min extraction) and from 61
to 278% (60 min extraction). Similar recoveries, using a 250°C extraction tem-
perature, were obtained from an urban dust sample (SRM 1649) for 8 PAHs (or
grouped isomers): 33-149% for a 15 min extraction time and 58-141% for a 60
min extraction time. In addition to PAHs a range of volatile and polar organics
were extracted from an industrial soil using hot water (at 250°C) and a 15 min
SPME. For the determination of the aromatic amines and other contaminants
an internal standard, 2,4-diethylamine, which was added directly to the extrac-
tion cell after the heating step but prior to removal of the water. Quantitation
was based on the use of external standard solutions of the target organics pre-
pared in water. The results were compared with an 18 h sonication method and a
24 h Soxhlet extraction (both using a 1:1 mixture of dichloromethane:acetone).
The authors concluded that the three extraction procedures produced reason-
ably similar concentrations for the less volatile organics, including aniline and
the chlorinated anilines. However, a significantly higher concentration (at least
>33 times) was obtained for N-methylaniline by SPME. It was postulated that
analyte degradation had lead to this spuriously high concentration being
recorded for N-methylanaline. The authors stressed the importance of being vig-
ilant for possible analyte degradation at high temperature.
An extension to this work was provided by Daimon and Pawliszyn [50] who
compared two different approaches for the combined high temperature water
extraction: SPME of PAHs in solid matrices. A 30 Am poly(dimethylsiloxane)
coated fibre was used for SPME. The two approaches evaluated were: dynamic
and static extraction. In the dynamic extraction mode, analytes leached from the
solid matrix with hot water are collected in a vial and simultaneously isolated
from the water using SPME. In the static extraction mode a high pressure cell is
used in to which is inserted the SPME fibre. In this approach analytes released
from the solid matrix are partitioned into the fibre coating as the cell is cooling.
After SPME sorption the fibre is removed from either the collection water or the
high pressure cell and inserted in to the injector of the gas chromatograph for
separation and detection without the need for any further clean-up or pre-
concentration. Both approaches were evaluated for the analysis of PAHs from
urban air particulate (NIST SRM 1649). In the dynamic approach, extraction
was done for 15 min at 250°C and 50 atm while for the static extraction the
sample was heated to either 270 or 300°C for 120 min in the presence of water
and SPME was done for 120 min while the cell was cooling. Recoveries for the
three PAHs investigated were 134.0, 87.5 and 72.0% for fluoranthene, pyrene
579
and benzo[a]pyrene, respectively, by the dynamic extraction approach. Recov-
eries by the static approach ranged from 137, 149 and 128% for fluoranthene at
270°C (100 mg sample), 270°C (50 mg sample) and 300°C (50 mg sample),
respectively; 113, 121 and 105% for pyrene at 270°C (100 mg sample), 270°C (50
mg sample) and 300°C (50 mg sample), respectively; and 49, 72 and 70% for
benzo[a]pyrene at 270°C (100 mg sample), 270°C (50 mg sample) and 300°C (50
mg sample), respectively. It was suggested that the advantage of the dynamic
approach was the simple external calibration (samples and spikes, 40 ml volume,
were isolated by the SPME fibre for 70 min with rapid stirring) and fibre
protection, since the fibre coating is exposed to relatively pure water under low
temperature conditions. On the other hand, static high temperature water
extraction with simultaneous SPME uses only simple apparatus and procedure,
and extends the approach for analytes tightly bound to soils. Quantitation for
2
this approach was achieved by spiking isotopically labelled standards i.e. 10 [ H]
2 2
fluoranthene, 10 [ H] pyrene and 12 [ HI benzo[a]pyrene, as internal standards
in to the sample prior to extraction.
580
PAHs compounds or more volatile compounds. This approach was then applied,
in addition to aqueous samples, to a sewage sludge and a soil sample known to be
contaminated with PAHs. For the sewage sludge sample, 0.2 Al of 10 gg ml- l
standard solution of each deuterated PAH (naphthalene-d8 and phenanth-
rene-d,,), in acetone, were used as internal standards. Using a sampling time of
30 min and maximum stirring of the sample it was possible to quantify the levels
of the two PAHs (2.0 and 0.3 ng ml- l for naphthalene and anthracene, respect-
ively) in the sewage sample using single-ion mode GC-MS. Using the same
approach the levels of four PAHs were determined in a soil sample. In this
situation, phenanthrene-d,, was used as an internal standard for anthracene
and phenanthrene while chrysene-d,2 was used as the internal standard for
benz[a]anthracene and chrysene. The deuterated PAHs (0.1 ml of a 10 g ml- l
standard solution) were spiked into a 1.0 g sample of the soil. The concentrations
determined were as follows: anthracene (17.0 pg gl), phenanthrene (4.0,g gl),
benz[a]anthracene (0.23 ,g gl) and chrysene (0.23 g g).
James and Stack [53] exposed a 100 Am poly(dimethylsiloxane) fibre to the
headspace above a soil sample (1.0 g) heated up to a temperature of 60°C (22, 30,
40, 50 and 60°C). Both spiked soils (control site) and soil samples from a landfill
site were used in the study. After 180 min the fibre was inserted into the injector
of a gas chromatograph and analysed for a selected range of volatile organic
compounds. Calibrations, using soil samples spiked with selected solvents
(1,1,1-trichloroethane, benzene, tetrachloromethane, trichloroethene, methyl-
benzene, tetrachloroethene, ethylbenzene, p-xylene and styrene) in the range
0.5-75 gg g' were linear (r2 > 0.9889). This approach (50°C for 30 min) was
applied to eight sites within a municipal landfill, at three depths, and resulted in
the detection of xylene isomers in 18 out of the 24 samples. The maximum
concentration of xylene determined was 2.8 pg g-l. In addition, a number of other
non-target organic contaminants were identified e.g. 4-methylphenol (found in
four samples) and alkylated benzene compounds and terpenes. 4-Methylphenol
is used as a wood preservative and so it is not uncommon to find it in municipal
landfill sites. The authors concluded that the application of headspace SPME
provided a rapid protocol for the quantification of residual solvents in landfill
soil samples.
581
TABLE 17.5
TPHs: total petroleum hydrocarbons; PAHs: polycyclic aromatic hydrocarbons; OCPs: organo-
chlorine pesticides
582
place). MAE is capable of extracting multiple samples simultaneously in a short
time; however, additional cleanup is required to remove the sample matrix from
the analyte-containing solvent, after cooling of the sample vessels. PFE allows
multiple samples to be extracted sequentially in an automated system. The
major disadvantage of PFE is the highest capital cost (of any of the techniques
described) and the need for sample treatment for high sulphur-containing soils,
or risk the blockage of tubing between the hot sample cell and the cooler collec-
tion vial.
Five main factors are considered important when making a decision as to
which technique to use for extracting environmental samples: capital cost,
operating costs, requirements for method development, environmental impact
(risk) and level of automation. It is possible that after taking these factors into
account that the 'easy option' is to retain the traditional, tried and tested
techniques currently available in your laboratory (Soxhlet extraction). However,
it is likely that a cost-benefit analysis of the instrumental techniques might well
enable the user to improve sample throughput, reduce solvent consumption (and
hence, disposal costs) and re-deploy staff in other activities.
So, while the authors have a tendency to favour the newer instrumental
extraction techniques because of their speed of extraction and smaller consump-
tion of organic solvent, they also know that current approaches are by no means
the end commodity. The current research focus towards miniaturisation should
lead to next generation systems that incorporate complete automated extrac-
tion-analysis systems.
583
TABLE 17.6
Calculation of individual group contributions for methanol
Group Group contribution to Group contribution to Group contribution to
dispersion
3
(Fd) polarity (F ) hydrogen bonding (Uh)
J1 2 cm /2 mo m mol' J mol-'
CH 3 420 0 0
OH 210 500 20000
Total 630 500 20000
2 6 2 2 6 2
5t = h + 5p + d (17.2)
d (Z'Fd)/V (17.3)
* (17.4)
6p = (zZZFp)/V
2 1 /2
6p = ( Fp ) /V (17.5)
2
6h = ((zZ Uh)/V)' (17.6)
":For molecules with more than one polar group present, Eq. (17.5) must be used
instead of Eq. (17.4), to take into account the interactions between the polar
groups [58].
An example calculation of the individual components of the solubility
parameter for a solvent (methanol) is shown in Table 17.6. The values for the
group contributions were taken from Ref. [58].
Fractional parameters of the Hildebrand solubility parameter can be calcu-
lated using Eqs. (17.3)-(17.6) and plotted on a triangular graph in order to give a
visual representation of the extent of contribution from the three components
(polarity, dispersion and hydrogen bonding). A visual representation between
the three components is via a triangular graph using fractional parameters (Fig.
17.12). The triangle can be used to predict the best solvent for dissolving a com-
pound. The optimum solvent should be in the same position as the target
analyte. This thinking can be applied to extraction techniques. To quantitatively
extract an analyte from a matrix, the solvent should have similar properties as
the compound, i.e. non-polar solvents are better at extracting non-polar
analytes. Using this approach, an informed selection of the most suitable solvent
for a particular analyte can then be made.
584
100
olvent value
onalyte values
Methanol
Acetone
fichloromethane
,cetonitrile
so-hexane
Foluene
,cetonitrile:dichloromethane (:1, v/v)
0o V / t 100
Fig. 17.12. Prediction of the optimum solvent for extraction: modelling the process.
REFERENCES
585
21 S.J. Chalk, H.M. Kingston, P.J. Walter, K. McQuillin and J. Brown, in: H.M.
Kingston and S.J. Haswell (Eds.), Microwave-Enhanced Chemistry: Fundamentals,
Sample Preparationand Applications. ACS, 1997, Chapter 15, p. 667.
22 V. Lopez-Avila, R. Young and W.F. Beckert, Anal. Chem., 66 (1994) 1097.
23 V. Lopez-Avila, R. Young, J. Benedico, P. Ho, R. Kim and W.F. Beckert, Anal. Chem.,
67 (1995) 2096.
24 I.J. Barnabas, J.R. Dean, I.A. Fowlis and S.P. Owen, Analyst, 120 (1995) 1897.
25 A. Pastor, E. Vazquez, R. Ciscar and M. de la Guardia, Anal. Chim. Acta, 344 (1997)
241.
26 B. Enders and G. Schwedt, GITFachz. Lab. (German), 40 (1996) 172.
27 G.-H. Xiong, X.-Q. He and Z.-X. Zhang, Anal. Chim. Acta, 413 (2000) 49.
28 K. Ganzler, A. Salgo and K. Valko, J. Chromatogr., 371 (1986) 299.
29 F.I. Onuska and K.A. Terry, Chromatogr., 36 (1993) 191.
30 S.T. Stout, A.R. dacunha and D.G. Allardice, Anal. Chem., 68 (1996) 653.
31 T.R. Steinheimer, J. Agric. Food Chem., 41 (1993) 588.
32 G.-H. Xiong, J.-M. Liang, S.-C. Zou and Z.-X. Zhang, Anal. Chim. Acta, 371 (1998) 97.
33 G.-H. Xiong, B.-Y. Tang, X.-Q. He, M.-Q. Zhao, Z.-P. Zhang and Z.-X. Zhang, Talanta,
48 (1999) 333.
34 B.E. Richter, B.A. Jones, J.L. Ezell and N.L. Porter, Anal. Chem., 68, 1033 (1996).
35 N. Saim, J.R. Dean, Md. P. Abdullah and Z. Zakaria, J. Chromatogr., 791 (1997) 361.
36 J.D. Berset, M. Ejem, R. Holzer and P. Lischer, Anal. Chim. Acta, 383 (1999) 263.
37 M.M. Schantz, J.J. Nichols and S.A. Wise, Anal. Chem., 69 (1997) 4210.
38 P. Popp, P. Keil, M. M6der, A. Paschke and U. Thuss, J. Chromatogr. A, 774 (1997)
203.
39 E. Bjbrklund, S. Bowadt, T. Nilsson and L. Mathiasson, J. Chromatogr.A, 836 (1999)
285.
40 E. Conte, R. Milani, G. Morali and F. Abballe, J. Chromatogr.A, 765 (1997) 121.
41 J.R. Dean, Anal. Commun., 33, 191 (1996).
42 L.J. Fitzpatrick and J.R. Dean, 21st InternationalSymposium on Capillary Chroma-
tography and Electrophoresis, Park City, Utah, USA, 20-24 June 1999.
43 L.J. Fitzpatrick and J.R. Dean, Anal. Chem., 74 (2002) 74.
44 P. Popp, P. Keil, M. Moder, A. Paschke and W. Thuss, J. Chromatogr., 774A, 203
(1997).
45 N. Saim, J.R. Dean, Md. P. Abdullah and Z. Zakaria, Anal. Chem., 70, 420 (1998).
46 M.E. Cisper, W.L. Earl, N.S. Nogar and P.H. Hemberger, Anal. Chem., 66 (1994)
1897-1901.
47 A.A. Bowland and J.Pawliszyn, J. Chromatogr., 704 (1995) 163-172.
48 A.A. Boyd-Bowland, S. Magdic and J.B. Pawliszyn, Analyst, 121 (1996) 929-938.
49 K.J. Hageman, L. Mazeas, C.B. Grabanski, D.J. Miller and S.B. Hawthorne, Anal.
Chem., 68 (1996) 3892-3898.
50 H. Daimon and J. Pawliszyn Anal. Comm., 33 (1996) 421-424.
51 Z. Zhang and J. Pawliszyn, J. High Res. Chromatogr., 16 (1993) 689-692.
52 Z. Zhang and J. Pawliszyn, Anal. Chem., 65 (1993) 1843-1852.
53 K.J. James and M.A. Stack, J. High Resolut. Chromatogr., 19 (1996) 515-519.
54 J.R. Dean, Extraction Methods for Environmental Analysis. John Wiley and Sons,
Chichester, 1998.
55 R.F. Fedors, Polymer Eng. Sci., 14, 147 (1974).
56 C.M. Hansen, J. PaintTechnol., 39, 105 (1967).
57 A.F.M. Barton, The Handbook of Solubility Parametersand other Cohesion Para-
meters. CRC Press, Florida, 1983.
58 D.W. van Krevelen and P.J. Hoftzyer, Propertiesof Polymers; Their Estimation and
Correlation with Chemical Structure. Elsevier, Amsterdam, 1976.
586
Chapter 18
18.1 INTRODUCTION
The desire to replace organic solvents used for extracting organic compounds
from solids and semi-solids has resulted in a large volume of literature describ-
ing the use of supercritical carbon dioxide. One other fluid, which certainly fits
"green" solvent criteria, is water. However, water at ambient conditions is much
too polar to be generally useful for extracting non-polar and moderately polar
organics from solid and semi-solid matrices, ranging from soils and sludges to
biological tissues. The polarity of water can be reduced dramatically (and the sol-
ubility of organics increased) by heating and pressurizing water to its supercriti-
cal state. However, the very high pressure (> 220 bar) and temperature (> 374°C)
required complicates its use for analytical extractions. In addition, supercritical
water is quite reactive (stainless steel is generally not sufficiently resistant to
corrosion). Thus, the major uses of supercritical water for organic compounds
have focused on their destruction, rather than extraction and recovery of
organics for analytical or process purposes. Similarly, the characteristics and
uses of superheated steam are well known, but generally are not considered
broadly applicable to organic extractions.
Interestingly, the characteristics of one area of water's phase diagram has
largely been ignored by the scientific community. As shown by the shaded area
in Fig. 18.1, the subject of this chapter is the use of water as an extraction fluid,
where the water is heated anywhere up to its critical temperature, and where
enough pressure is applied that the water remains liquid rather than turning to
steam. This technique has variously been named "subcritical water", "hot
water", or "superheated water" extraction, but regardless of the terminology,
the approach is defined by hot water kept under sufficient pressure to maintain
the liquid state.
When we performed our initial work with "subcritical" water [1], we had
approximately 400 references related to supercritical water extraction in our
files, but could only locate a few citations which were even vaguely related to hot
water extractions. None of these exploited the properties of hot water for analyt-
ical extractions. The lack of scientific literature on the properties of water in this
ComprehensiveAnalytical Chemistry XXXVII
J. Pawliszyn (Ed.)
© 2002 Elsevier Science B.V. All rights reserved 587
I
I
200
150-
Ice
. 100-
Liquid
Melting point curve --
50-
10000
0 . I ,
Fig. 18.1. Phase diagram of water. Note the shaded area representing hot subcriticali" water.
Generated using NIST/ASME Steam Formulation Database 10, version 2.2, A.H. Harvey, A.P.
Peskin, S.A. Klein [43].
region is particularly surprising since those who make cappuccino coffee, or who
use a home pressure cooker have long been exploiting these beneficial proper-
ties. While analytical chemists have largely ignored hot water, the venerable
Professor Franck had published one data set in 1983 on the solubility of
anthracene under "subcritical" water conditions which was very enlightening
[2]. Franck's data showed that the solubility of anthracene increased by a factor
of 1000 over its ambient solubility by simply heating the water to 150°C (at 60
bar). This surprising data set (although ignored for many years) clearly demon-
strates the potential uses of hot water for organic extractions.
In subsequent years since our first work with hot water extraction [1], the
amount of published work in this field has grown steadily. We hope that this
chapter will give the reader a realistic view of the potential of water as an
extraction fluid (both positive and negative aspects), as well as provide a survey
of work in the field. Since the majority of readers will be more familiar with
supercritical carbon dioxide, some comparisons of the two fluids will be made
throughout the text based on the qualitative comparisons shown in Table 18.1.
18.2 BACKGROUND
Three major characteristics of water change when it is heated and kept in the
liquid state, i.e., the polarity (as measured by the dielectric constant, ), the
surface tension, and viscosity all drop dramatically as shown in Figs. 18.2, 18.3,
and 18.4 [3]. In fact, when water is heated from ambient to 250°C, the changes in
588
TABLE 18.1
Properties of supercritical carbon dioxide versus hot "subcritical" water
CO2 HO
Solubility control 10-100x 50-100000Ox
Major parameter T&P T
"Easiest analyte" non-polar polar
"Hardest analyte" polar non-polar
Reactivity low low to moderate
Concentrating analyte extract usually simple simple to difficult
Selectivity for solute polarity fair good
Selectivity against matrix (e.g. soils) good poor
Polarity range (e) 1-2 10-80
these three parameters are nearly identical to changing the fluid to pure
methanol or acetonitrile (Figs. 18.2 to 18.4). While the drop in viscosity and
surface tension would be expected to enhance extraction rates based on mass
transfer considerations, the major advantage gained is the drop in water's
polarity from a very polar fluid at room temperature ( = 80) to a fairly low
polarity fluid at higher temperatures (e.g., E = ca. 20 at 300°C).
The practical effect of being able to control water's polarity with tempera-
ture is the ability to control the solubility of various organic compounds. The
surprisingly large increase in anthracene's solubility has subsequently been con-
589
%Methanol or Acetonitrile in Water at 20 C %Methanol or Acetonitrile in Water at 20 °C
on An n A
... . ..0 20 40 60 80 100
u AU 4U OU OU IUU
E
0
C
0
.2 a.
C
U)
a
0
C
V)
0
2A
I
a,
Left: Fig. 18.3. Effect of temperature and concentration of organic solvent in water on the
surface tension [3].
Right: Fig. 18.4. Effect of temperature and concentration of organic solvent in water on the
viscosity [3].
firmed by measuring the solubility of several solids (mostly PAHs and pesticides)
[4-7]. In fact, the solubility of such compounds tends to increase by an order of
magnitude for every 50°C increase in water's temperature, as shown for
benzo[a]pyrene and the pesticides atrazine and chlorothalonil in Fig. 18.5. Thus,
solubility increases of four to five orders of magnitude are typical for heating
water to 250°C and continue to increase at hotter temperatures. In fact, the solu-
bility of the PAH benzo[e]pyrene increases by an amazing factor of 25 million
fold when the water is heated to 350°C [7]. For comparison purposes, the solubil-
ity changes which occur for similar compounds over the useful temperature and
pressure range for supercritical carbon dioxide are typically in the one- to
two-orders of magnitude range, rather than the four- to six-orders of magnitude
increases found for hot water [8].
Although the solubility of liquid solutes such as alkyl benzenes and flavor
and fragrance compounds do not increase quite as dramatically as solid solutes,
the increase in solubility with temperature is still quite dramatic [9-11] with sol-
ubility changes of ca. 50 to 100-fold for nonpolars (like monoterpenes) occurring
in 200°C water. As might be expected, solubility increases are much more limited
if the solute already has significant solubility at room temperature [10]. For
example, the solubility of nerol only increases from ca. 2 x 104 mole fraction at
590
80
Benzo[a]pyrene
62
60 _-- -
J
40
-5
co
20
6
0.004 0.04
23 100 150 200
0D
Temperature, C
2000
Atrazine 1780
1600
CI
E 1200
CH3 N N
2- 800 _
O H3C NH N NH \CH 3
500
400 210
70
50 75 100 125
Temperature, °C
N C I
ci C
CI
Temperature, °C
Fig. 18.5. Solubilities of different organic pollutants at different water temperatures. Adapted
from [4-6].
591
as long as the pressure is sufficient to maintain the liquid state. While increasing
pressure most often increases the solubility of organics in carbon dioxide,
increasing pressure actually increases water's dielectric constant, and therefore,
decreases organic solubilities. For example, the solubility of the PAH, pyrene, in
100°C water drops from 9 x 10 - 7 mole fraction at 40 bar to 1 x 10- 7 mole fraction at
400 bar [5].
Much of the early work in hot water extraction was performed using equipment
analogous to that used for supercritical CO2 extraction, as shown in Fig. 18.6. In
essence, a pump is used to provide the water to the sample extraction cell via a
pre-heating coil. The pump can be run in a constant flow mode by utilizing a
small HPLC back-pressure regulator [5], or by using a flow control valve at the
outlet of the unit. Care must be taken to control the internal pressure so that the
phase of the water (liquid versus steam) in the extraction unit is controlled.
Analyte collection methods depend on the solubility of extracted compounds
in ambient water, and the final method used to analyze the extracted com-
pounds. For solutes fairly soluble in water, simply cooling and collecting the
water can suffice. For non-polar solutes, cooling the water prior to collection can
cause loss of the extracted compounds in the transfer line between the oven and
the collection device. (Remember, if the solubility of a compound increases sev-
eral orders of magnitude in the heated water, it decreases when the water is
cooled for collection.) For such compounds, we suggest introducing a small flow
of organic collection solvent in the heated zone as shown in Fig. 18.6. This allows
the collection solvent to mix with the hot water, and extract the organic solutes
from the water as cooling occurs. The result is a very nicely controlled flow of
cooled (and phase separated) water and organic solvent at the collection device.
Typical solvents good for non-polar organics (e.g., hexane, toluene, methylene
chloride) which are not miscible in water perform well. However, at higher tem-
peratures (ca. >250°C), chlorinated solvents can dehalogenate forming hydro-
chloric acid, and must not be used. Since nonpolar organics are generally
I~~~~~c~ollectio Miniature Back
Solvent >4 * Pressure Regulator
Pump Inlet Safety
Ives _ffRelief
Water - Valve
Collection
Vial
Preheating- Extraction
Coil Cell
592
analyzed by GC, their final collection in a solvent which is compatible with GC is
convenient.
An alternative to collecting analytes in the extract water, or in an organic
solvent is to pass the extract water through a solid-phase trap, such as C18
bonded silica [12-16]. This approach eliminates the use of organic solvents and
has been used to couple hot water extraction with HPLC. However, extreme care
must be taken to ensure that any non-polar organics are not deposited in the
transfer line as the water is cooled. Alternate approaches to solid phase trapping
have utilized membrane discs [17-19], and solid-phase microextraction (SPME)
[20-23].
Although the concept of off-line hot water extractors is similar to supercriti-
cal CO 2 extractors, the required characteristics of the various hardware compo-
nents are somewhat different, especially the pump and the extraction cell
heater. Overall, water is much easier to pump at a consistent flow rate than CO2,
since water is a non-compressible fluid in the pump. While many commercial
CO2 pumps are effective for pumping water, simpler and cheaper HPLC pumps
also suffice. Also, while the pressure required for supercritical CO2 extractions is
often 400 bar (or higher), the pressure required for subcritical water extractions
is only a few bar, as shown in Fig. 18.1. Thus, at the highest temperatures
typically used to extract non-polar organics with hot water (e.g., 300"C), the
pump need only be capable of working at ca. 100 bar (just above the pressure of
86 bar which is required to maintain the liquid state of water at 300°C).
Naturally, the cell heater must also be capable of 300°C, which is normally
accomplished by placing the sample cell (and a water pre-heating coil) inside a
GC oven.
Much lower temperatures and pressures will suffice if the operator only
desires to extract moderately polar compounds (e.g., most pesticides, oxygenated
flavor and fragrance compounds). For many such compounds, a temperature of
150°C or lower will suffice, which only requires a pump capable of delivering
water at ca. 5 bar. This keeps the water liquid, although for the sake of
simplicity, extractions are often performed at 50 bar. In fact, some commercial
SFE and PSE (pressurized liquid extraction) units can perform hot water
extractions at these conditions.
An approximate idea of the temperature required to extract a particular
class of organic compounds is shown in Fig. 18.7. Again, subcritical water is
essentially opposite to supercritical CO,, i.e., the most extreme conditions (high-
est temperature) are required for hot water extraction of non-polar solutes such
as PAHs and PCBs. In fact, the "easiest" compounds for extraction by supercriti-
cal carbon dioxide, alkanes (which are very non-polar) are virtually impossible to
extract with hot water (except by decompressing the water to steam, as
described later in the text). Similarly, the "easiest" organics to extract with hot
water, polar organics, are generally the most difficult to extract with supercriti-
cal CO 2 (and require the addition of an organic modifier to be extracted at all in
supercritical CO,).
593
Hard Nonpolar 280 °C
PCBs
i J PAHs I
organochlorine pesticides
monoterpenes
triazine and nitro-pesticides
explosives (HMX,RDX,TNT)
flavor oxygenates
Fig. 18.7. Water temperature
ran r-irrl to extrrt m- phenols, amines
poundswith differentpolarities. Easy Polar 100 °C
Similar to SFE, the time required for a particular extraction can be deter-
mined by collecting and analyzing fractions. As shown in Fig. 18.8, the time
required to extract chloransulam-methyl from soil can be dramatically reduced
by increasing the extraction temperature to 150°C, and reasonably good recover-
ies are achieved in ca. 100 min [24].
Special mention should be made of the plugging problems associated with
hot water extraction since this problem was so important during the develop-
ment of SFE. Unfortunately, plugging problems also exist with hot water extrac-
tion. However, at least in our laboratory (which mostly extracts soils and waste
sludges, as well as aromatic herbs), plugging at the outlet restrictor (e.g., the
back pressure regulator shown in Fig. 18.6), is not a significant problem.
Plugging of the outlet line (where the water is cooled) can be major if very high
concentrations of non-polars (e.g., a soil highly contaminated with PAHs) are
extracted at high temperatures (e.g., 250°C), since the solutes deposit when the
water temperature is dropped. However, adding the organic collection solvent to
the heated zone (as shown in Fig. 18.6), has completely solved this problem in
Ann
91
81
7(
S 6'
8 5
' 4
3
2
1
0
Time (min)
Fig. 18.8. The effect of time and temperature on the extraction efficiency of cloransulam-methyl
from soil. Used by permission from Krieger et al. [24].
594
our lab. Plugging of the sample matrix itself remains a significant problem. For
example, many soils appear to swell by 20% or more during extraction. Thus, a
void must be left in the sample cell upon filling to allow for this expansion. Fre-
quently, the sample swells enough at the higher water temperatures that mixing
it with a dispersant (or reducing the sample size) is required.
For SFE, "static" extractions are often performed to allow a modifier or chemical
reagent to be in contact with a sample for a certain time, after which a "dyna-
mic" extraction was performed in a conventional manner to collect the analyte.
While such experiments are certainly reasonable with hot water extractions, the
steam/water equilibrium can also be exploited to allow the construction of
extremely simple and inexpensive hot water extraction units [17]. The unit
consists simply of a threaded stainless steel pipe with threaded end caps sealed
with Teflon tape. At the time of this writing, the total cost per extractor was ca.
$10 (USD), and if tightened properly, the extractors typically last for more than
100 uses.
To perform an extraction with these simple static units, a sample is placed in
the cell, water is added, and the cell is capped. A headspace of air (or an inert gas
like nitrogen) is left above the water. When the water is heated (typically in an
old GC oven), the internal pressure is controlled by the steam/liquid equilibrium
pressure, i.e., the bottom of the cell contains the sample and the liquid phase of
water, while the top contains steam. After 15-60 min of heating, the cells are
cooled under tap water, and opened to collect the water extract.
It is critically importantnot to overfill these cells. To avoid excessive pressure
buildup, sufficient headspace must remain so that the pressure is controlled by
the steam/water equilibrium. (A convenient reference for the steam/liquid water
equilibrium pressures at various temperatures can be found in Refs. [25] and
[43], along with several other useful tables of water characteristics.)
The initial use of these cells was to provide a simple way to extract organic
pollutants from contaminated soils, and then determine their concentrations in
the extract water using SPME [20]. (In essence, we were attempting to turn a
soil sample into a water sample suitable for SPME analysis.) For more polar
analytes (e.g., chloroanilines), the solubility in ambient water was sufficient
enough that the target organics remained in the water phase after the static hot
water extraction of the soil. Thus, the water could be removed and the concen-
trations of the pollutants determined using normal SPME calibration
procedures.
However, non-polar solutes like PAHs showed substantial partitioning back
to the soil samples when the cell was cooled. While this initially made determin-
ing the concentration of the PAHs on the soil impossible, the addition of
perdeuterated PAHs as internal standards to the soil/water mixture prior to
extraction allowed good quantitative results from this simple static cell extrac-
595
Fig. 18.9. Comparison ofPAH concentrations obtained by 1 h 250°C hot water/SPME extraction
and by 48 h Soxhlet extraction (certified values) for urban dust (SRM 1649) [201.
tion followed by SPME and GC/MS analysis, as shown in Fig. 18.9 by the analysis
of urban dust [20]. However, the internal standard must behave the same as the
target analyte both in partitioning between the soil and water (upon cooling the
extraction cell) and in the SPME sorption step from the extract water. For
compounds where isotopically labeled internal standards are available, this
approach yields good quantitative data. However, if isotopically labeled internal
standards are not available, quantitation is less reliable. For example, semi-
quantitative determinations of soil PCB concentrations were possible using two
congeners (PCB 103 and 169), which are not present in environmental samples,
as internal standards [23]. SPME has more recently been used to determine
chlorophenols and methyl mercury in subcritical water extracts [21,22].
The static extraction cells have also been used with solid-phase trapping by
including an "Empore" sorbent disc in the cell during the static extraction step
[17,18]. In this technique, the cell is slowly rotated during the heating step so
that the water, soil, and sorbent disc are well-mixed. The mixing continues dur-
ing cooling, and the organics collect on the disc as their solubility in water is
reduced. The disc is then removed and extracted in a few ml of organic solvent to
recover the extracted organics.
Surprisingly, this very simple (and inexpensive) procedure is capable of
yielding quantitative extraction and collection of organic pollutants from soil.
Collection efficiencies of non-polars is quite good, with > 90% of the mass of each
PAH from a coal tar contaminated soil being collected on the sorbent disc, rather
than repartitioning to the soil during the cooling step [18]. This "low-tech"
approach yields quite good quantitative agreement with the concentrations of
PAHs determined in a coal tar contaminated soil, as compared to the concentra-
tions determined using 18 h of Soxhlet extraction (Fig. 18.10) [18].
596
500
400
E 300
E 200
w
100
Fig. 18.10. Comparison of PAH concentrations obtained by 1 h 250°C hot water/SPE (Empore
disc) extraction and by 18 h Soxhlet extraction for manufactured plant soil [18].
597
120 SDS
100
-
-80 1
0 i / water
o 40
0
0 10 20 30 40 50
Dynamic extraction time Imin)
Left: Fig. 18.11. Hot water extraction of benzo[ajpyrene from soil with and without the addition
of sodium dodecyl sulfate (SDS). Adapted from Ref. [271.
Right: Fig. 18.12. Extraction of metals from industrial oils with 150°C hot water modified with
4% (v/v) HNO3 + 0.1 M KCI. Adapted from [29].
Supercritical CO2 has often been praised as being a "selective" extraction fluid,
since pressure and temperature can be used to change its solvent strength. As
auuu
2500
2000
E
-8 1500oo
1000
500
598
1UU
80 374.9 °C
6.
0 Liquid water
'E
0
25 C
251C S water,
8
S0
Q 40
Z -methano
o-nitrololuene
5
20
-pnethyamiP _
CCI4 f PAHs _
00 200 300 400 500
Temperature,°C
Fig. 18.14. Dielectric constants of various solvents and solutes at room temperature compared to
the dielectric constant of water at different temperatures (220 bar).
shown in Table 18.1, the polarity of CO 2 (as measured by its dielectric constant)
can be controlled from ca. £ = 1 to 2 (or roughly in the range of alkane solvents).
In contrast, the polarity of hot water can be controlled from ca. 13 (at 350°C) to
80 (at ambient temperature). Thus, while hot water cannot achieve the low
polarity of CO,, its polarity can easily be controlled from E = 80 to that of
solvents like acetone ( = 18) and ethanol ( = 24), simply by raising the
temperature (Fig. 18.14). In our lab, this ability to control water's polarity
enables us to selectively extract different classes of compounds better than with
supercritical CO,.
An example of water's selectivity is shown in Fig. 18.15 by a comparison of
extracts from urban air particulate matter [31]. While both techniques yield
quantitative recoveries of PAHs compared to Soxhlet extraction, supercritical
CO 2 extracts large quantities of n- and branched alkanes, which completely
obscure the PAHs in the GC chromatogram (Fig. 18.15). In contrast, the
chromatogram from the hot water extract shows primarily PAHs, with essen-
tially no interfering alkanes. In fact, hot water is not capable of extracting
alkanes larger than ca. C14, unless the water is depressurized to steam [32,33].
While we have found the selectivity of hot water for extracting different
compound classes to be much better than that of supercritical CO2, it must be
noted that hot water is much less selective than supercritical CO 2 in regards to
extracting matrix components, especially for soils and sediments [31]. Hot water
extracts of soils and sediments are often dark brown, especially at higher
extraction temperatures, while supercritical CO2 extracts of the same soils will
typically be clear or yellow. A recent comparison of various extraction methods
for extracting PAHs from soil showed that water at 250°C and supercritical CO2
both extracted ca. 8 mg/g soil of matrix material (based on the residue weight
after evaporation of the collection solvent). However, when 300°C water
599
0
41 0
2 .0 0
I
IS
0 0B
At Subcritical Waer
K
0 00
U
0
0.1 s~~~~~~~~~~ 0
21
AL
0 e
N 0h
0 0
- 0
9;
0
3. i5 Retention Time (min) 35
la
40
V)
3.
Ai
Fig. 18.15. Comparison of GC/MS chromatograms of urban air particulate extracts obtained with
subcritical water, SFE with pure CO2 , and Soxhlet extraction [31].
600
100
80
0 60
-40
20
Fig. 18.16. Selective extrac-
tion of oxygenates versus
-
extractions were used, the extracted residue increased to 13 mg/g soil, which is
similar to the 15 mg/g extracted by pressurized liquid extraction. However, each
technique removed less matrix material than Soxhlet extraction, which removed
ca. 100 mg/g soil. SFE with pure CO2 also removed the least organic matrix
material from the soil of all of the techniques, and its extracts showed less matrix
material in the early parts of the GC/MS chromatograms [31].
The relative ability to selectively and quantitatively extract flavor and
fragrance compounds has also been demonstrated [34]. For example, hot water
extraction of savory yields much faster recovery of desirable (and more valuable)
oxygenates such as thymol, carvacrol, borneol, and linalool than of the less
desirable (and less valuable) monoterpenes (Fig. 18.16). With 100°C water, ca.
80% of the oxygenated compounds can be recovered in 20 min, while the extract
contains only a few percent of the mono- and sesquiterpenes. In contrast,
supercritical CO 2 shows little or no selectivity between oxygenates and terpenes.
Supercritical CO2 extracts also contain substantial concentrations of the
undesirable plant wax alkanes, while hot water extracts contained no detectable
concentrations of the alkanes (Fig. 18.17) [34].
A very elegant combination of supercritical CO 2 extraction with hot water
extraction utilizes the strengths of each fluid [19]. The non-polar herbicide
"Dacthal" (the dimethyl ester of 2,3,5,6-tetrachloro-1,4-benzenedicarboxylate)
was first extracted from the soil using SFE with pure CO2 . Next, the same soil
sample was extracted with hot water to extract the mono- and diacid metabo-
lites. This combination of SFE and hot water's strengths can be performed on a
single instrument, and deserves further attention.
Although the focus of this chapter is on hot water extractions, its use for
reverse-phase liquid chromatographic separations has received some recent
attention [35-42,44]. The concept for reversed-phase LC using water is easily
601
a) subcritical water extract
a
/:
0
o
o
CD
i0I 2I I -
10 20 30
Retention time, min
a)
c
0
0~
5a
P
C:
Fig. 18.17. Comparison gCD
of GC chromatograms
of savory extracts CD
obtained with selective
subcritical water and 1- I I I I I I
SFE with pure CO 2 0 10 ... . 20 30
[34]. Retenton time, min
seen in Fig. 18.2. Note that heating the water to ca. 200°C achieves the same drop
in solvent polarity as solvent programming from 0 to 100% methanol or aceto-
nitrile. Also, water's surface tension and viscosity also drop with temperature
(Figs. 18.3 and 18.4), both of which enhance LC separations. Several examples
utilizing this idea have been reported for polar (e.g., underivatized amino acids)
to moderately-polar (e.g., phenols and aliphatic alcohols).
In addition to the desire to reduce the use of organic solvents, one of the more
intriguing characteristics of water is that it has no response in a flame ionization
detector (FID), which has led to successful use of a GC/FID detector for aliphatic
and aromatic alcohols and underivatized amino acids [38]. However, subsequent
investigators have had less success with this approach, likely because the effect
of various manufacturer's FID designs (and related flame gas flow rates) on the
ability of a particular FID detector to accommodate high water flow rates has not
been well-understood [44].
A major concern with the use of hot water is the potential for degrading the
organics one wishes to extract. While this certainly occurs with some com-
pounds, the fact that the samples are not exposed to oxygen (water purged with
602
0
U 100 C 40 min [ 150C 20 min ' 175°C 12 min
0
'o
V
55
Fig. 18.18. Degradation of some savory oxygenates at water temperatures higher than 100°C [34].
603
the mixture of many different esters, but are normally analyzed after extraction
and derivatization as a single compound. For acid herbicides, this is often
performed by extracting the herbicides from soil using a Soxhlet apparatus,
hydrolyzing to convert the various esters to the acid form of each herbicide, and
then derivatizing (e.g., with diazomethane) the acid form to a single ester form
for analysis. Thus, for example, even though several different esters of 2,4-di-
chlorophenoxyacetic acid (2,4-D) may be used on a field, they are all converted to
their methyl esters for analysis.
Hydrolysis of acid herbicides on soil with water at 100-150°C has been
performed using the simple static cell described above, with a small piece of an
"Empore" anion exchange disc being placed in the cell during the heating step
[17]. Thus, the herbicides are extracted from the soil, hydrolyzed to their acid
form, then collected on the anion exchange disc as the static cell is cooled. The
acid forms of the herbicides are then derivatized with 1 ml of BSTFA (N,O-
bis(trimethylsilyl)-trifluoroacetamide), and analyzed by GC/MS. Recoveries for
a variety of acid herbicides were > 80% with a 30-min extraction, and the entire
procedure only required 3 ml of water and 1 ml of BSTFA reagent.
Other reactions occur which may or may not be useful, depending on the
investigator's goal. For example, attempts to develop hot water extraction meth-
ods for polychlorinated dioxins (PCDDs) from incinerator fly ash were complete
failures, and even spiked PCDDs could not be recovered with hot water. Subse-
quent studies showed that hot water may be an excellent method for destroying
PCDDs, but is not likely to be a useful analytical extraction method [49].
Similarly, even though PCBs can be quantitatively extracted from soils and sedi-
ments at 250-300°C [23,50,51], higher temperatures can result in dechlorin-
ation of the PCBs, particularly with added reactants such as iron [52]. Some
organics react at even lower temperatures. Attempts to use subcritical water
extraction to recover high explosives (TNT, RDX, and HMX) from contaminated
soils showed substantial degradation even at temperatures as low as 100-125°C.
This observation led to the development of a method to decontaminate soil by
destroying the explosives with subcritical water. Soils contaminated with as
much as 12 wt.% TNT have been cleaned to low ppm levels with just one hour of
exposure to hot water in a static system [53].
604
TABLE 18.2
Applications of hot water extraction
Matrix Analytes Comments Ref.
Soil and sediments alkanes SPE' 33
selective extraction 31
alkyl benzenes SPE online HPLC 15
SPME 2 20
aromatic amines SPME 20
chlorinated hydrocarbons SPME 20
chlorophenols modifiers, SPME 21
fungicide (tricyclazole) 54
herbicides SPE 13
consecutive SF CO 2 19
and subcritical water
24
chlorinated acid herbicides SPE, hydrolysis 17
metals (methylmercury) SPME 22
pesticides modified water 55
C18, modified water 12
56
PAHs SPME 20
SPE 33, 18
selective extraction 31
SPE-online LC-GC 14
modifier, SPE 28, 27
SPE online HPLC 15
PCBs 51, 57
SPE 50
SPME 23
PCDFs 58
soil organic carbon, phenanthrene matrix interactions 59
Air particulate/ fly alkanes selective extraction 31
ash PAHs SPME 20
selective extraction 31
SPE 18
Waste/sludges alkanes, alkyl benzenes, phenols, PAHs selective extraction 32
surfactant-nonylphenol polyethoxy modified (30% ethanol) 26
carboxylate
metal (Se) derivatization 60
organic waste (epoxy resin) hydrolysis 61
Food fungicides 62
Industrial oil metals acidification 29
Coal metals (Se, As, Hg) modifier, derivatization 30
Chrysanthemum insecticide hydrolysis 46
Clove buds eugenol, eugenyl acetate, caryophyllene - 63
Eucalyptus terpenes, oxygenates - 64
Fennel monoterpenes, oxygenates - 65
continued
605
Matrix Analytes Comments Ref.
Ground tea waste lignocellulosic materials - 66
Kava lactones - 67
Laurel terpenes, oxygenates - 68
Marjoram terpenes, pinenes, alcohols - 69
Oleander glycosides - 70
Peppermint oxygenates, caryophyllene - 71
selective extraction 34
Rosemary terpenes, oxygenates 72
Savory terpenes, oxygenates selective extraction 34
Coconut oil fatty acids from triglycerides hydrolysis 47
Linseed oil fatty acids from triglycerides hydrolysis 47
Soybean oil fatty acids from triglycerides hydrolysis 47, 48
Veronica longifolia glycosides, terpenes 73
2
1SPE = solid phase extraction; SPME = solid phase microextraction.
short time. In our laboratory, hot water extractions have been easier to perform
than SFE (especially in the early days when all SFE instrumentation had to be
fabricated in the lab), and the development of quantitative extraction methods
has gone more quickly. For samples such as soil, hot water extracts do tend to be
"dirtier" than SFE (at least those generated with pure CO,) in terms of contain-
ing more soil organic matter. However, hot water extracts of herbs can be
"cleaner" than SFE extracts, since plant waxes are not recovered with hot water.
The ability of hot water extraction to selectively extract different chemical
classes of analyte compounds far exceeds any selectivity we have been able to
achieve with SFE. In our laboratory, we have attempted to selectively extract
different polarities of compounds from samples ranging from waste sludges to
herbs using both SFE and hot water. In every case, the hot water extractions
yielded better selectivity among the various compound classes.
Quantitative hot water extraction methods for organics ranging from quite
low polarities (e.g., PAHs, PCBs, monoterpenes) to very high polarities (e.g.,
pesticide metabolites, amino acids) have been reported by a variety of laborato-
ries. In contrast, SFE methods for many of the more polar analytes were never
successfully developed, especially without the need for high concentrations of
organic modifiers. Given the characteristics of the two fluids, perhaps the
combination of SFE for non-polar organics, and hot water extraction for moder-
ate- and high-polarity compounds will be the optimal approach in the future.
REFERENCES
1 S.B. Hawthorne, Y. Yang and D.J. Miller, Anal. Chem., 66 (1994) 2912.
2 G.L. Rossling and E.U. Franck, Ber. Bunsen-Ges. Phys. Chem., 87 (1983) 882.
3 Y. Yang, M. Belghazi, A. Lagadec, D.J. Miller and S.B. Hawthorne, J. Chromatogr.A,
810 (1998) 149.
4 D.J. Miller and S.B. Hawthorne, Anal. Chem., 70 (1998) 1618.
5 D.J. Miller and S.B. Hawthorne, A.M. Gizir, A.A. Clifford, J. Chem. Eng. Data, 43
606
(1998) 1043.
6 M.S.S. Curren and J.W. King, Anal. Chem., 73 (2001) 740.
7 N.D. Sanders, Ind. Eng. Chem. Fundam,. 25 (1986) 169.
8 D.J. Miller, S.B. Hawthorne, A.A. Clifford and S. Zhu, J. Chem. Eng. Data, 41 (1996)
779.
9 Y. Yang, D.J. Miller and S.B. Hawthorne, J. Chem. Eng. Data, 42 (1997) 908.
10 D.M. Miller and S.B. Hawthorne, J. Chem. Eng. Data, 45 (2000) 315.
11 D.M. Miller and S.B. Hawthorne, J. Chem. Eng. Data, 45 (2000) 78.
12 C. Crescenzi, A. Di Corcia, M. Nazzari and R. Samperi, Anal. Chem., 72 (2000) 3050.
13 C. Crescenzi, G. D'Ascenzo, A. Di Corcia, M. Nazzari, S. Marchese and R. Samperi,
Anal. Chem., 71 (1999) 2157.
14 T. Hyotyl'ainen, T. Andersson, K. Hartonen, K. Kuosmanen and M.-L. Riekkola,
Anal. Chem., 72 (2000) 3070.
15 Y. Yang and B. Li, Anal. Chem., 71 (1999) 1491.
16 B. Li, Y. Yang, Y. Gan, C.D. Eaton, P. He and A.D. Jones, J. Chromatogr. A, 873
(2000) 175.
17 X. Lou, D.J. Miller and S.B. Hawthorne, Anal. Chem., 72 (2000) 481.
18 S.B. Hawthorne, S. Trembley, C.L. Monoit, C.B. Grabanski and D.J. Miller, J.
Chromatogr.A, 886 (2000) 237.
19 J.A. Field, K. Monohan and R. Reed, Anal. Chem., 70 (1998) 1956.
20 K.J. Hageman, L. Mazeas, C.B. Grabanski, D.J. Miller and S.B. Hawthorne, Anal.
Chem., 68 (1996) 3892.
21 L. Wennrich, P. Popp and M. Moder, Anal. Chem., 72 (2000) 546.
22 A. Beichert, S. Padberg and B.W. Wenclawiak, Appl. Organomet. Chem., 14 (2000)
493.
23 S.B. Hawthorne, C.B. Grabanski, K.J. Hageman and D.J. Miller, J. Chromatogr.A,
814 (1998) 151.
24 M.S. Krieger, J.L. Wynn and R.N. Yoder, J. Chromatogr., 897 (2000) 405.
25 L. Haar, J.S. Gallagher and G.S. Kell, National Bureau of Standards/National
Research Council Steam Tables. Hemisphere Publishing, Bristol, PA, 1984.
26 J.A. Field and R.L. Reed, Environ. Sci. Technol., 33 (1999) 2782.
27 V. Fernandez-Perez and M.D.L. de Castro, J. Chromatogr., 902 (2000) 357.
28 B. Vallejo-Pecharroman, L.E.G. Ayuso and M.D.L. de Castro, Chromatographia,53
(2001) 5.
29 V. Fernbndez-Prez, M.M. Jimbnez-Carmona and M.D.L. de Castro, Anal. Chim.
Acta, 433 (2001) 47.
30 V. Ferndndez-PBrez, M.M. Jim6nez-Carmona and M.D.L. de Castro, J. Anal. Atom.
Spectrom., 14 (1999) 1761.
31 S.B. Hawthorne, C.B. Grabanski, E. Martin and D.J. Miller, J. Chromatogr.A, 892
(2000) 421.
32 Y. Yang, S.B. Hawthorne and D.J. Miller, Environ. Sci. Technol., 31 (1997) 430.
33 K. Hartonen, G. Meissner, T. KesalB and M.-L. Riekkola, J. Microcolumn Sep., 12
(2000) 412.
34 A. Kubatovd, A.J.M. Lagadec, D.J. Miller and S.B. Hawthorne, Flavour Fragr.J., 16
(2001) 64.
35 R.M. Smith, O. Chienthavorn, S. Saha, I.D. Wilson, B. Wright and S.D. Taylor, J.
Chromatogr.A, 886 (2000) 289.
36 R.M. Smith and R.B. Burgess, Anal. Commun., 33 (1996) 327.
37 R.M. Smith and R.B. Burgess, J. Chromatogr.A, 785 (1997) 49.
38 D.J. Miller and S.B. Hawthorne, Anal. Chem., 69 (1997) 623.
39 Y. Yang, A.D. Jones and C.D. Eaton, Anal. Chem., 71 (1999) 3808.
40 T.M. Pawlowski and C.F. Poole, Anal. Commun., 36 (1999) 71.
607
41 S.M. Fields, C.Q. Ye, D.D. Zhang, B.R. Branch, X.J. Zhang and N. Okafo, J.
Chromatogr.A, 913 (2001) 197.
42 T.E. Young, S.T. Ecker, R.E. Synovec, N.T. Hawley, J.P. Lomber and C.M. Wai,
Talanta, 45 (1998) 1189.
43 Generated using NIST/ASME Steam Formulation Database 10, version 2.2, A.H.
Harvey, A.P. Peskin and S.A. Klein.
44 B.A. Ingelse, H.-G. Janssen and C.A. Cramers, J. High Resol. Chromatogr., 21 (1998)
613.
45 W.L. Marshall and E.U. Franck, J. Phys. Chem. Ref. Data, 10 (1981) 295.
46 M. Krappe, S.B. Hawthorne and B.W. Wenclawiak, Fresenius J. Anal. Chem., 364
(1999) 625.
47 R.L. Holliday, J.W. King and G.R. List, Ind. Eng. Chem. Res., 16 (1997) 932.
48 J.W. King, R.L. Holliday and G.R. List, Green Chemistry (1999) 261.
49 I. Windal, S.B. Hawthorne and E. De Pauw, Organohalogen. Compounds, 40 (1999)
40 591.
50 K. Hartonen, K. Inkala, M. Kangas and M.-L. Riekkola, J. Chromatogr.A, 785 (1997)
219.
51 Y.Yang, S. Bowadt, S.B. Hawthorne and D.J. Miller, Anal. Chem., 67 (1995) 4571.
52 H.K. Yak, B.W. Wenclawiak, I.F. Cheng, J.G. Doyle and C.M. Wai, Environ. Sci.
Technol., 33 (1999) 1307.
53 S.B. Hawthorne, A.J.M. Lagadec, D. Kalderis, A.V. Lilke and D.J. Miller, Environ.
Sci. Technol., 34 (2000) 3224.
54 M.S. Krieger, W.L. Cook, L.M. Kennard, J. Agric. Food Chem., 48 (2000) 2178.
55 A. di Corcia, A.B. Caracciolo, C. Crescenzi, G. Giuliano, S. Murtas and R. Samperi,
Environ. Sci. Technol., 33 (1999) 3271.
56 M.M. Jim6nez-Carmona, J.J. Manclus, A. Montoya and M.D. Luque de Castro, J.
Chromatogr., 785 (1997) 329.
57 S. Pross and B.W. Wenclawiak, Fres. J. Anal. Chem., 367 (2000) 89.
58 B. van Bavel, K. Hartonen, C. Rappe and M.L. Riekkola, Analyst, 124 (1999) 1351.
59 M.D. Johnson, W.L. Huang, Z. Dang and W.J. Weber, Environ. Sci. Technol., 33
(1999) 1657.
60 C.M.R. Varade and M.D.L. de Castro, J. Anal. Atom. Spectrom., 13 (1998) 787.
61 C. Fromonteil, Ph. Bardelle and F. Cansell, Ind. Eng. Chem. Res., 39 (2000) 922.
62 T.M. Pawlowski and C.F. Poole, J. Agric. Food Chem., 46 (1998) 3124.
63 A.A. Clifford, A. Basile and S.H.R. Saidi, Fres. J. Anal. Chem., 364 (1999) 635.
64 M.M. Jimenez-Carmona and M.D.L. de Castro, Chromatographia,50 (1999) 578.
65 L. Gdmiz-Gracia and M.D.L. de Castro, Talanta, 51 (2000) 1179.
66 A. Demirbas, A. Caglar, A. Ayas and S. Karslioglu, Fuel Sci. Technol. Int., 14 (1996)
395.
67 A. Kubatova, D.J. Miller and S.B. Hawthorne, J. Chromatogr., 923 (2001) 187.
68 V. Ferndndez-P6rez, M.M. Jim6nez-Carmona and M.D.L. de Castro, Analyst, 125
(2000) 481.
69 M.M. Jim6nez-Carmona, J.L. Ubera and M.D.L. de Castro, J. Chromatogr., 855
(1999) 625.
70 X. M. Wang, J.B. Plomley, R.A. Newman and A. Cisneros, Anal. Chem., 72 (2000)
3547.
71 A. Ammann, D.C. Hinz, R.S. Addleman, C.M. Wai and B.W. Wenclawiak, Fres. J.
Anal. Chem., 364 (1999) 650.
72 A. Basile, M.M. Jim6nez-Carmona and A.A. Clifford, J. Agric. Food Chem., 46 (1998)
5205.
73 J. Suomi, H. Siren, K. Hartonen and M.-L. Riekkola, J. Chromatogr., 868 (2000) 73.
608
Chapter 19
19.1 INTRODUCTION
19.1.1 Overview
610
A RdXH B R.XH
II o I I0
R CI RR CO
"o (CH,) 3S
0~CR
Io
+
RXH,
70 1
R.XSi(CH 3 )3
OH
(Et)3N R O
R R /C\ =N\OR
RdNH ECF
NHR
611
A 0 B H H
C=N-N-R d
,C\R
Hi H2NNRd
I H'
4H
7 RI OHNNR
H
R. HNNRJ
AC Rd
H H
H.
R, NHRd
R, NHRd
C=N "' Rd
C=N HH~ ~R
/¥ R2
19.2 FLUORESCENCE
19.2.1 Amines
612
o 0
II II
c H + H2NR k'=± /
OPA II / O NR
HO H
O
O
11 JZZ/" II
SC-H
CC'H
.1,
11
NR S
R,
H
/H
H SR2
I-
7
C
cI NR
SR2
A -alkylthiol-2-alkyl substituted isoindole
613
with OPA/Thiols [10,11] demonstrated the utility and practicality of this
approach.
New ideas on, and new applications of, derivatizations with OPA continue to
be reported, particularly the use of linked multiple detectors. Vasanits et al. [12]
described an extensive and detailed study of the determination of 31 amino acids
present in apples via derivatization with OPA using either MPA or NAC. A
major difference in the use of either of these two thiols is that formation of
OPA/MPA derivatives required a higher temperature than those obtained from
OPA/NAC derivatization. Optimization of chromatographic conditions and the
linking of fluorescent and photodiode array detectors in series maximized the
information derived from these analyses. Sensitivity remained unaffected by the
type of column used, allowing wide choice in selection of columns for developing
complex separations without undue concern about sacrificing sensitivity for
resolution.
Van Eijk et al. [14] simultaneously determined both plasma concentrations
of amino acids and their isotopic enrichment in a study on relative rates of syn-
thesis and degradation of proteins. The amino acids were derivatized with
OPA/MPA. Fluorescence detection provided concentrations and the effluent
from the fluorimeter was fed directly to an interfaced mass spectrometer for
determination of the isotopic composition. Differentially lower sensitivities for
mass spectrometric detection of the more polar analytes were attributed to
excess hydration interfering with the ionization process and not a matter inher-
ent in the formation or stability of the derivative. Sensitivity to the more polar
analytes was improved to some extent by increased temperature of the transfer
line rather than modification of the derivatizing reagent or reaction conditions.
Sampling by microdialysis is one of the more advanced techniques for in uivo
studies in physiology [15]. Yang et al. [16] combined this technique with analyti-
cal derivatization to determine extracellular glutamate and other amino acids in
cerebrospinal fluid (CSF). That biofluid was sampled via microdialysis and
mixed on-line with OPA to form fluorescent derivatives preparatory to
HPLC/FLU. The authors pointed out the importance of developing the relatively
standard HPLC/FLU as well as the more advanced CZE with laser induced
fluorescence (LIF) as the former was more widely available. Glutamate is an
excitatory amino acid involved in a variety of neurodegenerative disorders such
as Parlinson's and Alzheimer's diseases. Many investigators will wish to under-
take studies that involve the in vivo determination of glutamate and other amino
acids and these efforts would benefit from a wider range of instrumental options
such as those described by Yang and co-workers.
Wu and co-workers [17] as well as Ruberu et al. [18] described innovative
techniques using post-column derivatization by OPA/2ME and detection by
fluorescence. The latter group determined phenylurea pesticides. After separa-
tion these analytes by HPLC were decomposed to monoalkylamines by post-
column photolysis. The monoalkylamines were then derivatized with OPA/2ME.
Wu and co-workers [17] used a double function of quaternary ammonium
614
surfactants to determine N-methylcarbamate (NMC) pesticides. The surfactant
served as a hydrophobic pseudophase in micellular electrokinetic capillary chro-
matography (MECC) as well as (particularly cetyltrimethylammonium bromide,
CTAB) the catalyst for the thermal decomposition of NMCs to methylamine.
Again the alkylamine was derivatized with OPA/2ME for high sensitivity
detection by fluorescence. Separation, thermal decomposition and analytical
derivatization of the methylamine were performed in the capillary without any
interfacing. Despite multiple uses of the single capillary, the authors reported
column efficiencies of 50,000 and limits of detection below 0.5 ppm. The
processes and chemistry involved are complex and the methods required careful
optimization, yet they are simple in actual implementation and require minimal
sample preparation and interfacing.
Herraez-Hernandez and Campins-Falco [19] extended the OPA/THIOL
derivatization to include secondary amines. In developing the method for deter-
mination of ephedrine, a secondary amine, NaOC1 or chloramine-T provided a
Cl- that catalyzed the formation of an imine with subsequent hydrolysis to the
primary amine which was derivatized with OPA/NAC. The authors detailed the
optimization of this technique but did not clearly identify the structure of the
compounds.
Naphthalene-2,3-dicarboxaldehyde (NDA) also found considerable applica-
tion to the determination of amines. Again, a nucleophile is required to produce
a stable and highly fluorescent derivative. In this case, however, cyanide (CN)
rather than a thiol proved superior [20-27]. The structure of these 1-cyano-2-
substituted-benz[flisoindoles has been characterized by matrix assisted laser
desorption ionization (MALDI) and time-of-flight (TOF) mass spectrometry
[20]. Shah et al. [21] tested the stability of the cyano derivatives at 10°C over a 15
h period. Under these conditions there was no evidence of decomposition. In
contrast, OPA derivatives were stable for 5 h at 0°C.
Complex analytical problems are readily addressed through the use of
NDA/CN. Rapid derivatization of peptides [22], including those containing
highly basic lysine residues [23], expanded the biological applications. The
derivatives were readily separated by MECC and derivatives of these analytes
[22] were detected by LIF. Applications and analytical techniques were not
limited to biogenic molecules. Chiang et al. [24,25] separated the enantiomers of
bacolfen (4-amino-3-p-chlorophenylbutyric acid-an anti-inflammatory drug)
by derivatizing with NDA/CN and using CZE with a cyclodextrin in the buffer as
a chiral selector.
Advanced automated techniques also exploited sensitivity available for
derivatization with this reagent. A microchip device [26] combined enzymatic
reactions, electrophoretic separation of the reactants from the products and
post-column derivatization of proteins and peptides with NDA. Gottschlich et al.
[26] constructed this device and used it to measure peptides and amino acids
derived from insulin. They digested an oxidized insulin B-chain with trypsin
under stopped flow conditions in a heated channel of the device. Subsequent
615
separation was completed in 1 min and the separated peptides and proteins etc.
were derivatized with NDA in a post-column reactor and detected by laser-
induced fluorescence. It was also possible to reduce the disulphide bonds of
insulin on this device (reduction of disulphide bonds is discussed below). The
work demonstrated that microchip devices can perform all reactions, separa-
tions, derivatizations and detections involved in studies of proteins, peptides and
amino acids.
Parmentier et al. [27] described elegant chemistry and instrumentation used
for simultaneous determination of reactive oxygen species (ROS) and gluta-
thione (GSH). The bio-protective peptides, GSH and y-glutamylcysteine were
derivatized at the amine NDA/CN to form the corresponding fluorescent
products although the structure of these were not identified (see below re
derivatization of GSH with OPA). The toxic ROS were reduced by dihydro-
rhodamine-123 (DHR-123) which was itself oxidized to the fluorescent dye
rhodamine-123 (Rh-123). All fluorescent compounds were then separated by
CZE and detected LIF. The technique was sophisticated in all its components
but it provided a rapid and simultaneous measure of the cellular concentrations
of the toxic ROS and their endogenous scavenger.
19.2.1.2 5-Furoylquinoline-3-carboxaldehyde
Despite a structure that is quite dissimilar to either OPA or NDA, 5-furoyl-
quinoline-3-carboxaldehyde (FQ) also reacts with the amino acids in the
presence of CN to produce fluorophoric isoindoles [28]. As in the case of both
OPA and NDA, this compound has the desirable property of being a fluorogenic
rather than a fluorophoric reagent [29]. Sensitivity of analysis engendered by
the derivatives is exceptionally high-although this can vary with analyte
structure. In addition, since FQ is neutral its mobility is approximately equal to
the electro-osmotic flow and further reduces likelihood of interference.
The reaction has been very thoroughly studied. Activation energies and
reaction rates constant for derivatization of several amino acids with FQ were
recently determined for a number of amino acids [30]. While the activation
energy was relatively low (11-34 kJ), so was the rate constant. For most amino
acids the low rate constant indicated that complete reaction required about an
hour at 60°C. There was also high variation in sensitivity amongst the different
amino acids. This was attributed to differences in spectral properties (molar
absorbtivity, emission spectrum, fluorescence quantum yield) rather than less
efficient yield. Dovichi's group has provided considerable data on mechanisms
for derivatization of amino acids with FQ but there is yet more to learn regarding
this very important reaction. For instance, it is not clear why the derivatization
of proteins with FQ at 65°C [31] and was complete within minutes rather than
the hour required for amino acids [32]. Although, in some respects, detailed
knowledge of the reaction may be incomplete, the sensitivities achieved were
sufficiently high to detect amino acids and biogenic amines in 4 ul of brain
616
dialysate [28]. The slower reaction rates for these analytes necessitated analysis
of the samples in the off-line mode.
19.2.1.4 9-Fluorenylmethylchloroformate
Although 9-fluorenylmethylchloroformate is a long established and useful
reagent for determination of amines [6], it remains the subject of on-going study.
The simple question of its reactivity, surprisingly, is still being discussed.
Aymard et al. [37] determined the secondary amines ephedrine (Eph) and
norephedrine (NEph) from human plasma using derivatization with FMOC and
subsequent determination by HPLC with fluorometric detection. They argued
that formation of the FMOC derivatives provided specificity by two mechanisms.
The first was detection at an emission wavelength that has no corresponding
617
signal from endogenous compounds. The second was selective reaction with
primary and secondary amines.
Sparidans et al. [38] demonstrated that FMOC has a more extensive
reactivity towards amines than that proposed by Aymard and co-workers [37].
The former group developed a method for determination of olpadronate, a drug
used to treat osteoporosis. This analyte is a bis-phosphonate that contains a
tertiary amine. Sparidans and co-workers found that FMOC derivatized the
tertiary amine to give the secondary amine-FMOC [38]. This was unexpected
and the authors reported difficulty in obtaining confirmation of the structure by
mass spectrometry. Nevertheless, there was corroborating and recent data in
the literature that tertiary dimethylamines do react with FMOC and that there
is congruent demethylation [39].
As with many other techniques, sample preparation prior to or after derivati-
zation was a factor in optimizing yield of FMOC-derivatives. Sample handling
after derivatization of Eph and Neph requires considerable care [38]. FMOC
derivatives contain an ester functionality and are labile to hydrolysis. Conse-
quently, injection solutions were made in anhydrous acetonitrile, a solvent that
is quite hygroscopic. The requirement for anhydrous acetonitrile precluded
dissolution in the mobile phase and resulted in peak distortion when 150 Al were
injected, thus limiting sensitivity. These observations demonstrated the struc-
ture and stability of a derivative can affect various aspects of the analytical
technique.
Derivatization of olpandronate was sensitive to sample handling at almost
all stages of sample preparation [38]. Extraction of bisphosphonates typically
involves co-precipitation with Ca and subsequent removal of the metal ion
exchange chromatography. In this case cation exchange, but not anion exchange,
resin was compatible with derivatization using FMOC. Moreover, oplandronate
bound irreversibly to the resins. Addition of an analogous bisphosphonate to the
sample inhibited this binding. This may have had direct consequences to the
derivatization. At higher concentration of analyte (or other amines such as the
added carrier) the pH of the reaction and, consequently, the reaction rate
decreased as the reaction progressed and buffer was required to maintain the
pH. Even dissolution of reagent required optimization as did selection of the
buffer for the reaction. Acetone, but not other solvents, dissolved sufficient
reagent to allow a fast reaction.
In their study on determination of amino acids and peptides with FMOC,
Shangguan and co-workers [40] addressed problems resulting from excess
fluorophoric reagent, from additional reagent derived peaks and from sample
preparation. These workers performed the derivatization on a solid support of
alkaline silica gel as the sorbent (herein termed solid phase analytical deriva-
tization or SPAD) and compared the results with those from derivatization in
solution. Solid phase analytical derivatizations substantially reduced formation
of byproducts of FMOC. In addition, excess FMOC, a neutral molecule, was
selectively eluted from the phase with EtOAc. Since FMOC reacts with the
618
amine functionality, the derivatized amino acids and acids retained a free
carboxyl group. As a result during elution of excess reagent and side reaction
products with organic solvent, the FMOC derivatives adhered to the column and
were eluted with more polar solvents that contained methanol and water.
19.2.1.5 Isothiocyanates
Derivatization of amines and amino acids with fluoresceneisothiocyanate isomer
1 (FITC) produces highly fluorescent derivatives. Molina and Silva [41] deter-
mined the phosphorus-containing amino acid herbicides (glufosinate and
glyophosate) as well as aminomethyl phosphonic acid (the major metabolite of
glufosinate) as their FITC derivatives. They identified the optimal pH, FITC
concentrations, reaction temperature and reaction time to provide optimal yield.
The derivatives were separated by MECC and detected by LIF. Excess reagent
was an initial problem. With the first running buffer, derivatized analytes eluted
between 5 and 6 min but the excess FITC eluted as an intense broad peak at 3-5
min. Inclusion of Triton X-100, a non-ionic surfactant, in the running buffer
shifted the retention time of the FITC to less than 3 min but did not affect the
retention times of the analytes. With effective separation of FITC from the
reagent, the linear dynamic range was 2-3000 gg/ml and this was sufficient to
obviate the need for pre-concentration.
Asymmetric dimethyl-L-arginine (ADMA) is an endogenous inhibitor of
nitric oxide synthase (NOS). High concentrations of ADMA would indicate a
reduced formation of nitric oxide, which is a potent vasodilator. Reduced capac-
ity for vasodilation is thrombogenic since blood vessels cannot expand in
response to an occlusion. In principal, elevated ADMA concentrations may be
markers of vascular disease. This is borne out in practice where individuals with
renal failure, with peripheral arterial occlusive disease, or with clinically asymp-
tomatic hypercholesterolemia all exhibit elevated plasma concentrations of
ADMA. Determination of this analogue of arginine is therefore an important
analytical problem. Asymmetric dimethyl-L-arginine and symmetric dimethyl-
L-arginine (SDMA) as well as other methylated arginines were derivatized with
FITC preparatory to CZE with detection by LIF. Causse et al. [42] demonstrated
an effective separation of these methylated analogues of arginine that was com-
plete within 14 min. These basic amino acids also separated from neutral and
acidic amino acids that have more negative charges at the basic pH of the run-
ning buffer. The authors, however, report a slow decomposition of the FITC
derivatives of both ADMA and SDMA to products that elute prior to the parent
compounds but can be accounted for by proper sample storage.
Toyo'oka et al. [43] exploited the reactivity of isothiocyanates and combined
this moiety with one containing a chiral centre and another containing a
fluorophore. The technique was used for the resolution of racemic mixtures of
amines and amino acids with subsequent detection by HPLC/FLU). The chiral
tag R(-)-4-(3-isothiocyanatopyrrolidin-1-yl)-7-(N,N-dimethylaminosulfonyl)-2,
1,3-benzoxadiazole [R(-)-DBD-PyNCS] reacted readily with amines although the
619
yield for the more hindered amines such as those in the P blockers was reduced
by steric hindrance. Despite these difficulties, the method provided useful
information of chiral amines and amino acids. Separation of the enantiomers
was accomplished by simple isocratic elution which decreased analysis time. In
addition, high excitation (x = 460) and emission (em = 550) could be expected
to add to the specificity of the detection.
A reagent consisting of a near infrared dye linked to an isothiocyanate
functionality (NN382, LI-COR Inc) gave ultrasensitive determination for
peptides [44]. The isothiocyanate functionality provided facile derivatization of
peptides. Absorption and emission wavelengths of this dye are 790 and 800 nm,
respectively. At the high wavelengths in the near infrared (670-1000 nm) there
are few endogenous interferences. Moreover, scatter (both Raman and Rayleigh)
is substantially reduced at higher wavelengths. The study by Baars et al. [44]
found that concentrations of reagents, buffers and temperatures were important
determinants of the derivatization yield. Reaction did not occur at room temper-
ature but required 35-45°C. The traces from derivatization at 35°C, however,
were cleaner than those at higher temperature. A surprising finding was that
the ionic strength as defined by [NaCl] was also critical in maximizing yield and
had to be maintained at 2 M.
19.2.1.6 4-Fluoro-7-nitro-2,1,3-benzoxadiazole
Hu and Li 45] discussed 4-fluoro-7-nitro-2,1,3-benzoxadiazole (NBD-F) as a
reagent for derivatization of amino acids. They noted that the reaction was
complete within 1 min at 60 ° and this was considerably faster than the rate noted
for reagents such as FITC. In addition, since these derivatives can be detected
using a less expensive Argon laser there are cost efficiencies. Finally, there were
only two extraneous peaks and these are attributed to hydrolysis products of
NBD-F (i.e., NBD-OH and NBD-NH2).
Workers investigating D and L leucine required and exploited the high
sensitivity obtained with NBD. Determination of naturally occurring D-leucine
is an essential component of elucidating its possible physiological role. This
analyte must be separated from and detected in the presence of L-leucine which
is present in much higher concentrations. Derivatization with NBD-F provided
the requisite sensitivity and Inoue et al. [46] also reported the added benefit of a
very rapid reaction. Linked ODS-(C2 chiral) columns provided the separations.
Sensitivity of 1 fmol (S/N = 5) for leucine was reported and this was sufficient to
detect native D enantiomer in the brain. Kato et al. [47] tested enantiomeric
separations on a monolithic column in which the silica gel had been modified
with chiral phases. Amino acids derivatization with NBD-F were detectable at
very high sensitivity facilitating the use of capillary electro-chromatography.
Reductive amination (discussed below) of the aldehyde on a reducing sugar
with methylamine converts the aldehyde into and N-methylglyeamine [48]. The
secondary amine reacts with NDB-F to produce N-methylglycamines tagged
with NBD. Reaction between the secondary amine and NDB-F was more rapid
620
than for primary amine. This reactivity was attributed to a higher electron
density at the nitrogen. The NBD-tagged N-methylglycamines were separated
by CZE and detected by LIF. Although preparation of the NBD tagged
N-methylglycamine is a two-step reaction, Honda and co-workers [48] stressed
that it is a "one-pot" synthesis that is quantitative and complete within 50 min.
Moreover the derivatization is the last step and occurs in borate buffer which is
suited for the subsequent CZE. Detection by an argon laser-induced fluorescence
yielded an on-capillary sensitivity in the attomole range. This was sufficient to
detect N-glycans in a microgram quantity of a glycoprotein.
621
involved derivatization of the anilines with fluorescamine, the separation of the
derivatives formed by MECC, and fluorescence detection. Detection limits were
comparable (in the order of 1 zg/l). The analytical derivatization/fluorimetric
detection technique, however, gave higher separation efficiency and shorter
analysis times (approximately 4 min) and was reported as being more reliable as
well as having greater ease of handling.
Several groups [52-54] reported on derivatization with 4-(4,5-diphenyl-
1H-imidazol-2-yl)-benzoyl chloride (DIBC) preparatory to HPLC analysis with
fluorescence detection. Al-Dirbashi et al. [52] determined concentration of meth-
amphetamine and amphetamine in human urine. The highly fluorescent deriva-
tives could be detected when the analytes were analyzed from as little as 10 Al of
urine. Such a small volume did not require any work-up other than evaporation.
Moreover, the small sample size facilitated use of miniaturized equipment
suggesting even more extensive applications where small samples sizes are
important. Other reports describe application of this reagent to determination of
phenols (discussed below) [53,54].
Wang et al. reported [55] synthesized N-hydroxysuccinimidyl fluorescein-
O-acetate (SIFA), which was developed for determination of catecholamines by
HPLC with fluorescence detection. SIFA proved to be an amine selective reagent
that reacted with NE, E and DA under mild conditions. The detection limits of
the pure compounds were 3.2, 12, and 56 fmol, respectively, and a determination
of these catecholamines in urine was demonstrated. The issues with this tech-
nique were the high dilution factor and both reagent and matrix interferences.
Only 20 / 1 of 25 ml derivatized solution was injected onto the HPLC column.
Hydrolyzed reagent, derivatized amino acids and aliphatic amines could be
separated from the derivatized catecholamines but under such conditions, the
retention time for derivatized catecholamines was longer than for other meth-
ods. There appears to be considerable scope for enhancing the sensitivity and
speed of analysis for catecholamine determination using this useful reagent.
19.2.2 Thiols
622
analytical methods, as reviewed by Ubbink [58]. Analytical derivatizations are
fundamental in almost all methods for measuring the biogenic thiols.
There are several problems present in the determination of thiols. There
must, of course, be the usual derivatization to incorporate a group to enhance
detection. It is also often necessary to block oxidation of thiols (RSH) to the
disulphides (RS-SR). A third and frequent requirement is the reduction of the
RS-SR dimer to the RSH monomer. Although measurement of the disulphides
such as oxidized glutathione (GS-SG) may be important, it is the reduced thiols
that are biologically active. Reduction of the disulphide bond and maintaining
the reduced state during sample preparation is arguably, according to MacCoss
[59], the "greatest problem associated with measuring homocysteine" and the
same can be claimed for the measurement of all biogenic or exogenous thiols
found in plasma.
DeGraff et al. [60] reported a prototypical example of the chemistry involved
in the analyses of thiols. These authors studied diagnosis and management of
cystinosis, a rare metabolic disorder of cystine transport. At issue was the
determination of intracellular cystine concentrations. Inter and intra-acellular
cysteine, however, can be oxidized during sample handling and preparation.
Blocking the thiol of cysteine with excess N-ethylmaleimide (NEM) [60,61]
prevented oxidation of cysteine to cystine during sonication of cells. It also
blocked the disulfide exchange reactions of cystine with available sulfhydryl
moieties such as glutathione and proteins. Once all the possibilities of thiol
oxidation and/or exchange were blocked, cystine was reduced with sodium
borohydride. The resulting cysteine was derivatized with monobromobimane
(discussed below) to produce a highly fluorescent product suitable for HPLC
separation with detection by fluorescence.
623
19.2.2.2 Blocking/protectingthe thiol group
Just as there are varied approaches to the problem of disulphide reduction, there
are also a variety of blocking/protecting reagents for the reduced thiols.
Hammermeister [68] took a very classical approach and blocked reduced thiols
from cell extracts by acetylation to prevent oxidation to the RS-SR dimers.
Haj-Yehia et al. [69] took a similar approach and derivatized the thiols of
dihydrolipoc acid (DHLA) with ethylchloroformate. A more standard reagent is
N-ethylmaleimide which has been successfully used by many investigators
[60,61,70-72] offers substantial specificity for reaction at the mercapto group.
Another class of thiol protecting/blocking reagents are based on iodo groups
[9,73]. Iodoacetic blocked the thiol of homocysteine and cysteine leaving the
amine to react with OPA [9]. As a result, the method could be used to determine
the reduced forms of the two thiol amino acids as well as five other, non-sulphur
containing, amino acids such as glutamic acid. Boyd et al. [73] utilized iodo-
acetamide, not to derivatize a thiol amino acid for but to scavenge excess
t-butylthiol subsequent to an OPA derivatization. It is likely that iodoacetamide
would react equally well with the reduced sulphur containing acids. MaCoss et
al. [59] drew on the literature on protein sequencing and selected 4-vinyl-
pyridine (4-VP) as a thiol protecting/blocking reagent and this produced a very
stable product.
624
hydrolysis product of mBrBi and these are potential interferences. If these
caveats are recognized and accounted for, then excellent sensitivities can be
achieved by derivatization with mBrBi.
In contrast, and as would be expected, monochlorobimane is less reactive and
derivatization with this reagent requires enzymatic catalysis [74] with GSH
transferase. While this increases specificity of the derivatization it limits the
application to analytes that are enzymatic substrates. For instance, reaction
between GSH and monochlorobimane can only occur when catalyzed by GSH
transferase but the enzyme will not catalyze derivatization of homocysteine.
This affords considerable specificity and permits determination of intracellular
GSH simply by adding monochlorobimane to cultured cells. The reagent crosses
the membrane and intracellular GSH transferase catalyses the derivatization to
produce a highly fluorescent derivative that allows measurement of intracellular
GSH. Kamencic et al. [74] reviewed this technique and extended the scope of
derivatizations with monochlorobimane. They fortified tissue homogenate with
this GSH transferase and found that it allows a direct measure of GSH in
homogenate as well as in intact cells. It is likely that the enzyme catalyzed
derivatization would also be applicable to the determination of GSH in such
matrices as plasma of cerebrospinal fluid that do not contain this enzyme.
4-Aminosulfonyl-7-fluoro-2,1,3-benzoxadiazole (ABD-F): A highly fluores-
cent derivative is obtained from reaction between thiols and the fluorogenic
reagent [62,75]. Salazar and co-workers [62] reacted ABD-F with homocysteine
and determined this analyte from plasma by HPLC/FLU. They also confirmed
the structure of the derivative with electrospray ionization-mass spectrometry.
In this work, concentrations of homocysteine in plasma were determined by the
HPLC/FLU, by fluorescence polarization immunoassay and by capillary gas
chromatography-mass spectrometry. There was good correlation between the
three techniques providing investigators with options for determining valid
concentrations of this important amino acid.
In keeping with the current requirements of analysis, Kang et al. [75] com-
bined on-column derivatization with ABD-F and separation/detection by
CZE/LIF to develop a fully automated technique for determination of homo-
cysteine. The analyte was directly injected into the capillary in a "sandwich"
between two plugs of reagent solution and derivatized at 50°C prior to separa-
tion. A theoretical examination of reaction rates and mobility under the CZE
conditions permitted calculation of injection rates. The result was a reproducible
on-column derivatization technique that improved the cost-effectiveness of the
determination but increased somewhat the detection limit. The limit of detec-
tion for on-column derivatization with LIF detection was 5.0 nM, as compared to
2.5 nM when a pre-column derivatization was used. The method is a very simple,
fast, and practical approach to fully automated determination of homocysteine
and other thiols contained in low-volume and low-concentration samples.
6-Iodoacetamidofluoresceine: Causse et al. [67,76,77] and Ducros et al. [78]
built upon the known reaction of iodoacetic acid and iodoacetamide with thiols
625
[9,68]. They investigated 6-iodoacetamidofluoresceine (IAF) as a thiol specific
reagent and concluded that, compared to other reagents, it had superior
properties of speed of reaction, selectivity, sensitivity, stability of derivatives and
absence of hydrolysis products. The sensitivity and specificity for determination
of homocysteine following derivatization with IAF enabled comparison between
high performance liquid chromatography/conventional fluorescence detection
CZE/LIF and fluorescence polarization immunoassay. Although all three
technique produced equivalent results CZE/LIF was the most cost-effective [76].
The reports by Ducros and co-workers [78] also highlighted the importance
of matrix effects that must always be considered in analytical derivatization.
Derivatization of GSH, homocysteine, etc. with IAF was dependent on the
method of precipitating the protein in the sample of plasma or serum. A detailed
study demonstrated that protein precipitation with sulfosalicylic acid and
acetonitrile provided high yields of derivative whereas the use of other solvents
or reagents was considerably less efficacious.
Maleimides: There are several reagents that combine maleimide functional-
ity with a fluorescent moiety and these are thiol specific. Ercal and co-workers
[79] studied 9- acetoxy-2-(4-(2,5-dihydro-2,5-dioxo-lH-pyrrol-1-yl)phenyl)-3-
oxo-3H-naphtho [2,1-b]pyran maleimide (ThioGloTM3) as a derivatizing agent.
Derivatization with this reagent was rapid at room temperature and the fluores-
cent ThioGlo-SR adducts were detected fluorimetrically (e = 365 nm, m =
445 nm). This reagent also produced less hydrolysis products than found for
derivatization with a more standard maleimide reagent, N-(1-pyrenyl)-male-
imide. Oxidized glutathione (GSSG) was determined by difference. All free thiol
groups were blocked with 2-vinylpyridine, GSSG was reduced enzymatically
with GSH reductase and the resulting GSH was derivatized with ThioGlo.
Reactions at sulphur in determinationof GSSG and GSH: Cereser et al. [80]
and Lenton et al. [81] recently addressed the difficult problem of determining
both GSH and GSSG. Concentrations of these two compounds reflect the oxida-
tive demand within a cell. Their quantitative determination has considerable
import for physiology, pharmacology and toxicology. Cereser and co-workers
took a more classical approach and used pre-column derivatization of GSH with
OPA [80] but in the absence of thiol reagent. Formally, under these conditions
(the exact mechanism is unknown) the reaction of the nitrogen with OPA can be
followed by a intramolecular reaction between the thiol terminus of GSH with
the isoindole [61] (Fig. 19.5). The resulting tricyclic derivative with an unusual
10 member ring is highly fluorescent. The relationship between yield vs pH was
a curve with a relatively narrow maximum at pH = 10 and indicated the need for
626
close attention to buffering [80]. The optimization produced a detection limit for
GSH at 50 fmol. It was not possible to directly determine GSSG under these con-
ditions but the disulphide could be measured indirectly by difference. The first
step was the determination of GSH and this was followed by reduction of GSSG
with dithiothreitol. Subtraction of the GSH concentrations before reduction
from those following reduction produced the value of the disulphide.
Post-column derivatization, however, provided a simultaneous determina-
tion of GSH and GSSG via derivatization with OPA. The reaction conditions
were pH = 11.6 and 60°C [81]. Again thiol was omitted to allow derivatization at
both the NH2 and SH functionalities. Lenton et al. [81] demonstrated that GSH
was not oxidized to GSSG during sample preparation and that their direct
measure of GSSG reflected the true concentration of this analyte. It is not clear
whether GSSG forms a OPA derivative or whether the elevated temperature and
pH in the post column reaction produces some other compound which then
reacts with OPA. The hypothesis of reaction at both the NH2 and SH of the GSH
remains unproven. Elucidation of derivative structure may be an important
problem. That intramolecular reaction product exhibits a substantially
increased fluorescence relative to that obtained for the simpler thiols; for
instance, these investigators reported 1000 times greater sensitivity for GSH
and GSSG than for cysteine.
627
Volpi et al. [83] derivatized unsaturated disaccharides with DANSYL hydrazine.
The disaccharides were prepared by enzymatic digestion of glycoaminoglycans
and then converted to their DANSYL hydrazones. Following derivatization a
series of seven unsaturated disaccharides with various degrees of sulphation
(0-2 sulphate groups) was separated by single column and an isocratic system.
Detection of less substituted and early eluting analytes was achieved but it
appeared that peaks from reagent or other sources could interfere at lower
concentrations.
19.2.3.2 1,2-Diamino-4,5-methylene-dioxybenzene
Inoue's group investigated derivatized sialic acids with 1,2-diamino-4,5-
methylene-dioxybenzene (DMDB) [84,85]. Unlike most derivatizations of sac-
charides which occur on the carbonyl, the sialic acids were determined as
ketoacids (Fig. 19.6). Derivatization with DMDB produced a sensitivity of 13
fmols of a silaic acid dimer and Sato et al. [84] found that this compared to the
sensitivity obtained by electrochemical detection. Both techniques, electrochem-
ical detection of native analyte and derivatization with DMDB, proved useful in
analysis of oligomers and polymers of sialic acids (Sia chains). Lin and
co-workers argued that derivatization methodology, however, provided higher
sensitivity and, more importantly, was effective in the presence of proteins [85].
This could be expected. The reagent, DMDB, will not react with proteins and
these biopolymers will remain invisible to the detector. At alkaline conditions,
HO
~Rs~~i
j H
N
.
XAOJ
/ .
2
C.
CHCH
OH
HO
R2 =R r
628
however, the electrochemical detector would respond quite sensitively to the
sulphydryl and hydroxyl groups on the proteins produce a high background.
Although there are advantages to this derivatization, the acidic reaction
conditions hydrolyzed the Sia chains.
O
II
JI
H R,
Fig.
19.7.
Reductive
examination.
629
A second side reaction can be generated by the -7.5% acetic acid used in
many of these derivatizations as typified by the conditions used by Charlwood et
al. [86,87]. Acidic conditions can hydrolyze sialic acids residues from the poly-
saccharide [47,84,88,89], and this is a significant matter [82]. Sialic acids are
nine carbon carboxylated sugars located at the terminal end of glycoproteins and
glycolipids [89]. This location enables the biological functions of sialic acids for
cell-cell signalling and adhesion which are fundamental to processes of multi-
cellular life. Their association with polysaccharides is an important measure in
elucidating many biological processes and their loss during derivatization limits
the quality and interpretation of the data produced.
Honda et al. [48] argued that a combination of a monoamine and a dimethyl-
amine-borane complex as reactant and reductant, respectively, is a preferred
condition for reductive amination. Unwanted side reactions are minimized or
eliminated by a reaction temperature at 40°C and maintaining pH = 4.5. In a
technique described in Section 2.1.5 reductive amination with methyl amine and
the dimethylamine-borane complex produced a N-methylglycamine, with
minimal loss of sialic acids and with little or no reduction to glucitols.
A widely used reaction mixture for reductive amination [86,87], includes the
saccharide, amine, sodium cyanoborohydride in a solution of dimethylsulph-
oxide containing a 7.5% acetic acid. The reaction temperature varies from
70-80°C and the reaction time from 1-2 h depending on the temperature. These
conditions appear harsh, but for derivatization of saccharides with 2-amino-
acridone (2-AMAC) or 3-acetylamino-6-aminoacridine (AA-Ac) there are consis-
tent reports giving mass spectral evidence of sialic acid residues retained on the
oligosaccharides prepared for analysis by this technique [90-92].
It should also be recognized that there are examples of saccharide analysis
where such issues are not a concern. The specifics of the analytical problem need
to identified. For instance, in a study on inflammatory bowel disease, Araki et al.
[93] required the determination of dextran sulphate from rat faeces to demon-
strate retention of this polysaccharide in the GI tract. They described reductive
amination of dextran sulphate sodium (DSS) with 2-aminopyridine and sodium
cyanoborohydride. Reaction conditions were stringent requiring 20 h at 90°C but
were adequate for this particularanalyticalproblem.The fluorescent derivatives
were subsequently separated by size exclusion chromatography.
630
map. Methyl-4-aminomethyl benzoate (M4-AB) was used to derivatize a dextran
ladder for a specific and useful application which is discussed in more detail
below [87,92].
19.2.3.5 Aminoacridines
Charlwood and co-workers [86,91,92,95-97], Birrell et al. [87], Hanrahan et al.
[90] and Huterrer et al. [98] of Camilleri's group conducted extensive studies on
quantitative determination of mono, oligo and polysaccharides and also
determined the complex branched structure of the oligomers and polymers.
Derivatives of the aldehydic groups were prepared by reductive amination with
either 2-AMAC or more recently AA-Ac and with sodium cyanobohydride. The
authors reported basic developments including preparation and validation of a
new aminoacridine reagent as well as numerous applications.
In the course of these studies [86,87] the authors addressed the problem of
excess reagent and other impurities. They demonstrated that semi-preparative
column chromatography of the reaction mixture on an OASIS column allowed a
selective elution of derivatives and used this method extensively [91,92,95-97.
Introduction of the 2-AMAC group provided the fluorescence necessary for
high sensitivity detection and also increased lipophilicity as well as including the
aminoacridine group that ionized at acid pH. The latter two properties facili-
tated separations by HPLC or CZE [86,87]. When used in combination with the
appropriate enzymatic hydrolyses, analytical derivatization with 2-AMAC and
chromatographic separation allowed assignment of structures for glycans in
complex mixtures. These derivatives were compatible with CZE/ LIF,
HPLC/FLU, HPLC-MS/MS and MALDI-TOF [86-88,90-93,95-97]. Hutterer
and co-workers [98] further extended the utility of the 2-AMAC derivatives.
They reported that oligosaccharides tagged with 2-AMAC were well suited to the
MECC at 100 kV which considerably improved the resolution over that obtained
at 20 kV - an important finding. Increase in resolution adds to the quality of
data available from the electrophoretic analysis.
More recently Charlwood and co-workers [86] prepared AA-Ac and compared
it to 2-AMAC. They reported a two-fold increase in the detection sensitivity. For-
mation of the AA-Ac as the products were apparently more lipophilic, were
better retained and resolved MECC relative to that found for the 2-AMAC deriv-
atives. As a result, approximately twice as many oligosaccharides were detected
with the AA-Ac derivatives. In addition, the AA-Ac derivatives carry more posi-
tive charge and fluoresce under acidic conditions. With these properties it is fea-
sible to use CZE and MECC. Accordingly the analyst can select the order of
elution based on molecular size of the AA-Ac derivatives. Capillary zone electro-
phoresis elutes the derivatives in order of size (for instance AA-Ac glucose is the
first peak because of its high charge density) and the order is reversed for MECC.
A double derivatization with M4-AB and 2AMAC [87,92] provided an elegant
technique for determination of molecular weights of mono, oligo and polysaccha-
rides. A dextran ladder was derivatized with M4-AB whereas the saccharide
631
were derivatized with 2-AMAC. The derivatized dextran ladder was added then
to the solution containing the 2-AMAC derivatized saccharides. The 4MAB
derivatives were monitored by absorption at 305 nm and 2-AMAC derivatives
were monitored by fluorescence. Fluorescence properties of 4MAB did not
interfere with that of the 2-AMAC and the two classes of compounds could be
detected independently. The derivatized dextran ladder provided an internal
calibration of molecular weight for derivatized oligosaccharides and the
independent detection ensured that there was no cross-interference.
632
R
NH1
+ CI-CH 2CHO
iR
cl-c~.
CH
2
Xcr
R=sugay residue
Fig. 19.8. Reaction of an adenine derived drug or nucleoside with chloroacetaldehyde. R = Ribose
or deoxyribose.
633
matography with fluorescence detection provided limit of detection for cGMP at
10 fmol injected on column. Such sensitivity was sufficient to assay guanylate
cyclase (GCase) activity by measuring GTP and cGMP. The two nucleotides are
the substrate and the product of GCase, respectively. Although this method was
less sensitive than generally used radioisotopic methods, it was sufficient to
determine GCase activity in human platelets. Moreover, the HPLC method is
simpler and does not require the use of radio-isotopes.
634
to allow for switching and washing. While this is an effective methodology, the
run time is long-just under 50 min.
Haj-Yehia et al. [69] determined a-lipoic acid (LA) and DHLA in biological
fluids. After extraction and protection of the thiols in DHLA, the carboxyl
groups were derivatized with 2-(4-aminophenyl)-6- methylbenzothiazole using a
diimide under basic conditions to couple the analyte and reagent. In addition to
providing a determination of LA and DHLA the reagent may prove useful in the
analytical derivatization of other carboxylic acids.
635
19.2.8 Photolytic activation
Photochemical reactions produce highly fluorescent compounds and this can be
used to advantage most often with post-column derivatization [116-118].
Sharma [116] studied the formation of adducts between tamoxifen, an anti-
cancer drug, and DNA. This work demonstrated that during biotransformation,
tamoxifen (or its 4-hydroxy biotransformation product), is activated and binds
covalently to DNA. Post-column photochemical derivatization cyclized the
4-hydroxytamoxifen-DNA adduct to a phenanthrene and production of this fluo-
rescent compound increased sensitivity by two orders of magnitude. Adduct for-
mation between an oxidatively activated compound and DNA is often mutagenic
in its own right. It is therefore possible that treatment of cancer with tamoxifen
may also induce cancer at a later stage. A similar photochemical induced
cyclization was used to determine the cycloxyenase II inhibitor 5-chloro-3-(4-
methanesulfonylphenyl)-6'-methyl-[2,3']bipyridinyl. This analyte rapidly
formed fluorescent products when exposed to UV light at 254 nm in a post col-
umn reactor [117]. These compounds were detected by excitation at 260 nm and
emission 375 nm. At 5 ng/ml the response was well above the base-line noise and
the reproducibility was excellent.
On-line post-column irradiation at = 254 nm produced photo-induced
modifications of equilin and equilenin to fluorescent derivatives [118]. Limits of
detection were 10 fmol for equilenin ( = 280 nm; Kem = 410 nm) and 50 fmol
for equilin (x = 280 nm; Xk, = 312 nm). Differences in kex and hXm indicated that
irradiation at 7 = 254 did not simply oxidize the dihydronaphthalene of the A-B
ring in equilin to the naphthalene of the A-B ring in equilenin.
Determination of melatonin in rat and mouse pineal gland was by HPLC
with fluorescence detection [119]. Melatonin was derivatized with hydrogen
peroxide in carbonate and produced a fluorophore detected at the emission
wavelength of 380 nm with the excitation at 245 nm. The structure of the
derivative was not defined but may be N-acetyl-N-formyl-5-methoxy-ky-
nuramine or some analogue of it.
636
pGal(l -- 4)pGIcNAC-O-TMR
a fuc(l -- 3)
z(1-03)FTase t [LeX]
pGal(l - 4)pGIcNAC-O-TMR
[LacNac--TMR]
tr(-2)FTase
pGal(l - 4)PGlcNAC-O-TMR
a fuc(1-02) Il +3)Frase
Gal(1 - 4)PGIcNAC-O-TMR
a fiic(-2) a lc(1 -3)
[LeY]
Fig. 19.9. Biosynthetic pathways for LacNac-O-TMR to for LeX and LeY.
The amine functionality then provided the reactive moiety for derivatization by
5-CTRSE. Detection limits for the tetramethylrhodamine derivatives were on
the order of 60 molecules injected on-capillary. This derivatization was not
applicable to the study of biological problems because of the lengthy deriva-
tization procedure and high concentrations of analyte required to compensate
for hydrolysis or reagent. Nevertheless it was a powerful proof-of-principal.
Following demonstration of this very low detection limit, a biochemical
derivatization was developed [122]. A methoxycarbonyloctyl group was attached
to the oligosaccharide gal (1->4)PGlcNac (N-acetyl-lactosamine abbreviated
LacNac) and this product was reacted with ethylenediamine to form Pgal(1-4)
PGlcNac-O(CH 2 ),CONH(CH 2)2 NH2 . The amine group was the site of reaction
with 5-CTRSE to form the derivative gal(1-*4)jGlcNacO (CH2 )sCONH(CH
2) 2NHTMR (LacNac-O-TMR). Upon incubation, cells took up this reagent which
then became a substrate enzymatic process within the cell. Both degradation
and biosynthetic pathways were observed [33] and are summarized in Fig. 19.9.
In the degradation pathways galactosidase cleaved LacNacTMR to GlcNacTMR
and subsequent reaction with hexosaminidase cleaved the sugar and released
the HO(CH2)8CON(HCH2)2NHTMR moiety. In the presence of GDP-fucose
(GDPF), fucosyl transferases (FTase) incorporated the LacNac-TMR into more
complex oligosaccharides labelled LeX and LeY as shown in Fig. 19.9. Three
types of activities were defined. In the G1 phase of the cell cycle neither LeX nor
LeY was produced as would be expected in the resting phase. In the G2/M phase
some cells remained inactive (i.e., no degradation or biosynthesis) or produced
either LeX or LeY. No cell was found to produce both LeX and LeY. Thus the cell
cycle of individual cells can be probed by chemical cytometry.
637
19.3 ABSORPTION OF ULTRAVIOLET AND VISIBLE LIGHT
Detection in the ultraviolet and visible (UV/VIS) wavelengths does not, as a rule,
provide the high sensitivities typical of many modern methods. Nevertheless the
techniques are well understood, the equipment is rugged and there are many
applications that can benefit from use of such reliable methods. This was
exemplified by continuing reports [123-141] describing methods based on
derivatization followed by detection by absorption in the UV/VIS region.
19.3.1 Amines
19.3.1.1 Isothiocyanate
Miller et al. [123] recognized that study of complex biochemical problems
requires rapid and simple methods as well as rugged, relatively inexpensive
equipment. They derivatized isopeptides with phenylisothiocyanate (PITC) and
applied this technique in studies on cross-linking of proteins during clotting.
Their's was a simplified method for determining the frequency of cross-linking
as it eliminated complex multicolumn systems or the more expensive fluori-
metric detectors used in other methodologies.
New information continues to refine derivatizations with PITC; for instance,
Albin et al. [124] recently reported that standard hydrolysis conditions (6 M HCl
at 110°C for 24 h) were not optimal. They suggested a sequential hydrolysis
technique to optimize yield for derivatization of amino acids by PITC. Zahou and
co-workers [125] used derivatization with PITC but relied on the high extinction
coefficient to detect at 200 nm to provide the high sensitivity. CZE separated the
18 amino acids produced by the hydrolysis of peptides with 10-600 residues.
Sensitivities were in the 50-60 fmol range or three orders of magnitude better
than the ninhydrin technique used in routine analysis of amino acids.
19.3.1.2 Nitroaromatics
Nitroaromatic reagents continue to find extensive application in the determina-
tion of carbonyls, amines, alcohols and acids. The nitro groups impart high
molar extinction coefficients thus increasing the sensitivity. Nitroaromatic
moieties are attached to hydrazines or acyl chlorides to provide reactivity. In the
case of 1-fluoro-2,4-dinitrobenzene (FDNB), the nitro groups also exert a power-
ful electron withdrawing effect and in conjunction with the electron withdraw-
ing effect of F produce an electrophilic centre on the aromatic ring. Nucleophilic
attack and displacement of the fluorine by amines tags the analyte with
2,4-dinitrobenzene (DNB).
The neuro-excitatory non-proteogenic amino acid 3-N-oxalyl-2,3-diamino-
propionic acid (-ODAP) and other such non-proteogenic amino acids were
determined as their DNB derivatives. These amino acids are potentially toxic to
humans and animals. Under certain environmental conditions these compounds
can be present in some food crops such as grasspea (Lathyrussativus). Measure-
638
ment of ODAP is useful in assessing potential toxicity. Extracts of grasspea were
treated with FDNB and ODAP as well as other amino acids were derivatized at
the amine [126]. The derivatives could be detected in the pmol range. The
resulting method is an important addition to the options available to the food
chemist.
Determination of gentamicin, an aminoglycoside antibiotic, presents an
analytical problem that is somewhat more complex than usual. First, the drug
itself is a mixture of gentamicins C(1), C(la), and C(2) with other gentamicins
also possibly being present in the pharmaceutical preparations. Second, pharma-
cokinetic and other pharmacological or toxicological studies require that all
three compounds be measured. Third, the drug has no functionalities that can be
used for detection. Isoherranen and Soback [127] determined the gentamicins in
plasma or urine via a SPAD technique. They extracted the gentamicins from the
bio-fluids by SPE and the sorbed gentamicins were contacted with a solution of
1-fluoro-2, 4-dinitrobenzene directly in the extraction cartridge itself. This
considerably simplified the manipulations required in sample preparation.
Studies on diffusion of gentamicin through agar matrix [128] presented less
severe requirements than pharmacokinetics but the selection of the preferred
method is not completely resolved. In developing this technique, the authors
compared derivatization with FDNB and OPA as well as stability of the resulting
derivatives. The former reagent provided a more facile reaction and more stable
reaction products than those achieved by the OPA reaction. On the other hand
Kaale et al. [129], using a central composite experimental design, optimized an
NDA/CN derivatization and followed detection by absorption at Xm = 330 nm.
The fully optimized capillary zone electrophoresis provided a baseline separa-
tion of gentamicin C1, Cla, C2, C2a, C2b, sisomicin and several minor
components.
Biogenic amines are readily derivatized with 3,5-dinitrobenzoyl chloride
(DNBC) and this reaction was developed to determine such analytes in food
[130]. This reaction has a distinct advantage of rapidity, being complete within 3
min in a solution of 1 M NaOH, acetonitrile and 2-propanol. Kirschbaum et al.
[130] confirmed structures of the derivatives by HPLC-MS with electrospray
ionization. Rapid derivatization of amines with this reagent was also reported by
Herraez-Hernandez et al. [131]. These authors compared sensitivities for
derivatization by DNBC with those obtained from FMOC and recommended the
latter reagent for determination of amines.
19.3.2 Carbonyls
639
derivatives of numerous carbonyls found in car exhaust. They reported the
concentrations and speciation of carbonyls collected in Tuscarora Mountain
tunnel as an example of the diversity of these environmental contaminants [132]
produced by heavy diesel trucks. Despite its extensive use over time, application
of 2,4-DNPH to a variety of problems requires careful study and attention to the
needs of the specific problem and the chemistry of the analytes.
Innovation in this field remains essential-even when the analytical
problem consists of a frequently measured analyte such as formaldehyde and a
long used reagent such as 2,4-DNPH. Determination of formaldehyde, a potent
toxin, is required for quality control in the manufacture of water soluble poly-
mers used in waste water treatment. In this case, however, the standard pre-
column derivatization with 2,4-DNPH produces unreliable measures of
formaldehyde as the polymers themselves react with this reagent. Michels [133]
circumvented the problem by using post-column derivatization. This allowed a
separation of the polymer from the analyte prior to formation of the chromo-
phoric derivative.
Despite the long experience with this derivatization there are still potential
problems with specific analytes. Dusi and Gamba [134] studied determination of
polyether antibiotics (PEs) lasalocid, monensin, salinomycin and narasin that
are used in poultry feeds. Under the acidic conditions of the reaction, these
polyethers are hydrolyzed to carbonyls that react with this reagent. While the
yield of the derivatization of the other antibiotics was acceptable, that for
lasalocid was low. This was attributed to steric hindrance. Lasalocid has one
carbonyl with two highly substituted a carbons. Monensin, in contrast, has a
ketal and hemiketal which, at acidicity of the reaction, are converted to ketones
but the a positions are either methylenes or hydroxy methyl groups.
Reaction of the hydrazone group via intramolecular condensation [135] was
recently described. Such reactions occur when there are additional carbonyls or
conjugated unsaturation in an analyte. Following derivatization of 3-butyl-2-one
with the 2,4-DNPH the expected product butynone DNP-hydrazone was formed.
A subsequent rearrangement (as in Fig. 19.3B) led to 3-methyl-1-(2',4'-dinitro-
phenyl)pyrazol (DNPP). This did not interfere with determination of the
analytes but co-eluted near the 2,4-DNPone of formaldehyde. Similar reaction
products were observed for analytical reactions between other hydrazines and
polyfunctional carbonyls.
The usual analytes selected for derivatization with hydrazines contain a car-
bonyl functionality. Nevertheless, Miwa [136] successfully used 2-nitrophenyl-
hydrazine hydrochloride to form the 2-nitrophenylhydrazides of low molecular
weight mono and di-carboxylic acids. The derivatization was catalyzed by
diimide to condense the reagent with the carboxylic acid and required heating to
60°C. A subsequent treatment with strong alkali, also carried out at 60°C, was
640
required to remove excess diimide and other interferences. Although the
reaction was somewhat involved and lengthy, the resulting derivatives have
excellent chromatographic and detection properties.
Busulfan [139] is an antineoplastic drug that acts by alkylating DNA via the two
sulphate leaving groups. The drug itself is a neutral molecule and so was readily
isolated from plasma by liquid/liquid extraction. In alkaline medium one
molecule of busulfan reacted with 2 molecules of 2,3,5,6-tetrafluorophenol
(TFTP) to form di-TFTP-butane. This derivative was readily analyzed by HPLC
and at 275 nm the detection limit was 50 ng/ml. Results from analysis of
busulfan by this HPLC/UV technique correlated well with those obtained by
mass spectrometry. These data indicate a high specificity with the chromophoric
detection method. This analytical derivatization reagent and technique can
likely be applied to other alkylating drugs used in cancer chemotherapy. In
addition a thiophenol with a higher molar absorptivity would undoubtedly
increase the sensitivity of the technique.
Diethyldithiocarbamate (DDTC) also proved a useful nucleophile for deter-
mining busulfan [140]. A facilitating feature of derivatization with this reagent
641
is that it is compatible with plasma and can be carried out directly in that
biofluid. The subsequent sample preparation, however, was somewhat extens-
ive. It required extraction with ethyl acetate followed by further separation of
the analyte by chromatography on C8 columns. Despite the complexity of the
method, recoveries were high (99%) and the overall method had a sensitivity of
25 ng/ml. More importantly, it only required 200 Al of plasma. This is a particu-
larly important characteristic when studying pediatric patients. Their low body
weight combined with the illness and cancer chemotherapy, make children a
very difficult population to study: they can be very anemic and unable to give
large volumes of blood. This sensitivity was sufficient to determine plasma
concentrations after oral administration and to perform both therapeutic drug
monitoring and studies of the pharmacokinetics of this drug in children.
Another pairing of a thiol and an antineoplastic drug produced a successful
method. Cisplatin has no functional group that can be used to detect low levels of
cisplatin by HPLC. This is a standard feature of analytes that require derivatiza-
tion. The required product formed when DDTC reacted with the platinum in
cisplatin to yield a platinum-DDTC (Pt-DDTC) complex [141]. This derivative
has a high molar-extinction coefficient and can thus be detected at high sensitiv-
ity after chromatographic separation. Raghavan et al. [141] indicated that
DDTC has a use beyond analysis. They suggested that the reagent can be used to
destroy cisplatin by converting it to platinum-DDTC. This can be used to
inactivate pharmacy counter-tops or any other area where the drug may be on
surfaces or in solution.
Boyd et al. [73] combined the hindered t-butylthiol with OPA to derivatize and
determine amino acids in microdialysate. The t-butylthio derivatives were more
stable and more lipophilic, the latter property enhancing preconcentration on a
capillary column. Electrochemical detection of the 2-t-butylthio isoindole deriva-
tives provided the requisite sensitivity. A critical aspect of this work, however,
was removal of the excess thiol by scavenging with iodoacetamide and excess
OPA with Na2SO3 . The latter reaction produced OPA-based N-alkyl-1-isoindole-
sulfonate which eluted in the solvent front. The double scavenging reduced the
background by 90% and substantially increased the sensitivity. High sensitivity
and automated throughput allowed sampling with a 10 s temporal resolution. As
many neural processes occur on time scales of a second or less, the temporal
resolution is a crucial feature of the methodology.
Wang et al. [142] developed an integrated miniaturized analytical device
which also exploited the OPA/2ME reaction as well as the electroactivity of the
final derivative. Micromachining technology and a miniaturized electrochemical
detector provided the instrumentation that allowed derivatization, separation
and detection on a glass micro chip. The authors reported on the influence of a
number of variables on the analytical process. Relationships between yield and
642
reagent concentrations showed a fairly sharp maximum indicating that small
error in these variables can have a large effect on recoveries of the derivatives.
The sample/reagent (S/R) ratio had a more "forgiving" profile in that the
maximum yield was obtained over a wide range of S/R values. As with all current
analytical methods speed was an important consideration. A step of the driving
voltage reduced migration times for late eluting components. The entire process
is easily automated and integrated into high throughput techniques.
Tcherkas and co-workers reported on the use of sulphite as a nucleophilic
reagent for OPA derivatization preparatory to HPLC with electrochemical
detection [143]. This is an interesting innovation and is analogous to the
scavenging reaction described by Boyd et al. [73]. At present, however, it is
limited to the more lipophilic amino acids. The more polar analytes in this group
are eluted in the solvent front when derivatized with OPA/sulphite.
Many of the following reagents affect several functionality groups. Silylation can
occur on carboxyl hydroxyl and amine functionalities. Similarly, alkylation can
form esters, ethers and enol ethers from carboxyls, hydroxyls and carbonyls.
Indeed, there are instances where these different groups are on the same mole-
cule. This section is therefore organized according to derivatization reaction.
19.5.1 Silylation
Silylation was one of the first analytical derivatizations [144]. The enhanced
volatilization resulting from silylation of compounds with amino, hydroxyl and
carboxyl functionalities was fundamentally important in both qualitative and
quantitative gas chromatographic analysis of numerous and structurally diverse
organic compounds. Little [145] reviewed artifact formation in the course of
silylation but this authoritative report also summarized reagents, functional
groups and reactions involved in this derivatization technique. Despite the
decades-long use of this derivatization, it remains fundamental to many current
techniques.
Trimethyliodosilane (TMIS) and other strong silyating methods have seen
considerable use. Thevis and co-workers [146] demonstrated selective trimethyl-
silylations of steroidal glucuronides. Derivatization with perdeutero N, O-bis (tri-
methylsilyl) acetamide (BSA) gave the perdeuterotrimethylsilyl ethers of
hydroxyl groups but left the keto groups free. The former were predominantly of
the glucuronyl moiety, whereas the carbonyls were found only on the steroid.
Subsequent reaction with the highly reactive TMIS enolized and trimethyl-
silylated the keto groups. This sequential derivatization tagged hydroxyl and
ketone groups with different isotopes yet finally produced a pertrimethysilylated
derivative which provided valuable structural information.
643
Other applications of TMIS have also appeared recently. Shinohara et al.
[147] addressed derivatization of a sterically hindered steroidal alcohol. Regula-
tory requirements for the International Olympic Committee (IOC) were the
driving force for the study and so the focus was detection in urine. These
investigators developed a method for determining the two urinary methyl-
testosterone metabolites: 17a-methyl 5a-androstan-3a, 17p-diol and 17a-methyl
5a-androstan-3(,17-diol. Again, both hydroxyl groups in these molecules are
hindered: the 3a hydroxyl is an axial configuration and the 17P hydroxyl is a
tertiary position and is also a to the 18 methyl group. Reaction with TMIS,
trimethylsilyated these highly hindered hydroxyls and provided the necessary
gas chromatographic methodology.
The IOC also generated a study on determination of plasma expanders such
as hydroxyethyl starch (HES). Plasma expanders are usually administered in
cases of hypovolaemic shock. In normal individuals, however, they increase
plasma volumes and dilute concentrations of drugs in blood. The IOC banned the
use of HES from January 2000 and required methods for the determination of
this compound in biological fluids. Thevis and co-workers [148] hydrolyzed HES
which produced ethylated monosaccharides. Reaction with TMIS produced the
pertrimethylsilyl derivatives of these monosaccharides and they were
determined by GC.
These reports demonstrated the utility of TMIS. This reagent, however, is
both highly reactive and unstable. It must be prepared with care and the
reaction is usually performed in situ by a combination of N-methyl-N-tri-
methylsilyltrifluoroacetamide, ammonium iodide and ethanethiol accompanied
by heating to 60°C.
The problem of silylating a hindered hydroxyl groups was also encountered
in the determination of gentrisone (13P ethyl-17a-ethynyl-17-hydroxygona-
4,9,11-trien-3-one), an oral contraceptive possessing androgenic as well as anti-
estrogenic and anti-progestagenic properties. These properties make it a
possible doping agent for the high performance athlete and this again generated
an analytical problem. The hydroxyl group is very hindered and the triene
structure coupled with the 3 keto makes this molecule difficult to derivatize for
confirmation of doping by GC-MS. Kim et al. [149] showed that trimethyl-
silylimidazole saturated with sodium acetate coupled with heat will give both
17p-trimethylsilyl (TMS) and 3 TMS-enol. The latter reaction produces three
tautomeric forms with extended conjugation. These can be separated chromato-
graphically and have quite distinct mass spectra. Although formation of the
three tautomers reduces sensitivity, it can add substantially to the specificity of
analysis by providing distinct patterns of peaks and three unique spectra from
one compound.
Ohie [150] et al. compared the efficacy of t-butyldimethylsilyl (TBDMS) and
TMS derivatives for gas chromatographic analysis amino acids, their metabo-
lites and hydroxy fatty acids. Trimethylsilyl derivatives fragment via the loss of a
TMS group (m/z = 73) accompanied by extensive fragmentation. t-Butyl-
644
dimethylsilyl ethers and esters fragment via the loss of the t-butyl group (m/z =
57). The resulting intense ion at M-57 retains most of the structure of the mole-
cule. Accordingly, the signal is highly specific as it is attributable to an ion that
truly represents the structure of the analyte. Selected ion monitoring at higher
masses provided greater sensitivity for most of the compounds studied and in
some instances detection limits were five-fold lower for the TBDMS compounds.
The coefficient of variation for the TBDMS derivatives was approximately half
that of analyses using the TMS ethers/esters. The superior analytical perform-
ance of TBDMS products, however, translated into only a moderate improve-
ment in identifying genetic diseases. Analyses based on the TBDMS derivatives
gave the correct diagnosis in 52 of 53 cases, whereas those based on MS
derivatives gave a correct diagnosis in 49 of the 53 cases.
One frequent problem with all silylations is that reagents and derivatives are
variously labile to water. Mol et al. [151] address this issue in an extensive study
of analysis of phenol endocrine disruptors. They optimized derivatization of
these organic acids with several silylating reagents including N-methyl-N-
(t-butyldimethylsilyl) trifluoroacetamide (MSTFA). A variety of temperatures,
reaction times and solvents were tested to determine optimal conditions. In the
course of this study they found derivatization with MSTFA but not other
silylating reagents can tolerate up to 0.5% water in the silylating solution.
Compatibility of this reagent with a small amount of water was a particularly
important consideration for one of the most complex analytical problems under
investigation. This is the determination of amino acids in the Martian terrain
using GC-MS aboard the spacecraft [152]. That planetary atmosphere contains
some water but is below 0.5%. Rodier et al. [152] found derivatization with
MSTFA was simple, compatible with some water contamination and readily
automated. The investigators, however, pointed out that the TBDMS adds to the
molecular weight of the reaction products and puts the molecular ion of the
larger amino acid derivatives above the mass range of the instrument on the
spacecraft. The structure of these will have to be inferred from the gas chromato-
graphic properties alone.
Techniques for reducing sample handling and reducing solvent use during
sample preparation derivatization attract interest. Staerk and Kulpmann [153]
combined trimethylsilylation with head-space sampling at elevated tempera-
tures using solid-phase microextraction (SPME). In this technique, drugs (e.g.
amphetamines, cocaine, benzodiazepines, barbiturate) were isolated from aque-
ous biological sample by liquid/liquid extraction using a small volume. The
extract was evaporated to dryness in a head-space vial. The vial was sealed and
depending on the analytes two alternative procedures were used. For determina-
tion of opiates or cocaine the analytes were silylated in the vial using MSFTA. An
SPME syringe was inserted and the temperature increased to volatilize the
silylated derivative which were then sorbed onto the fibre for solventless injec-
tion. Alternatively, the fibre was doped with acetic anhydride/pyridine and
inserted into the vial. Upon heating, the analytes were volatalized, sorbed onto
645
the fibre and derivatized in situ (the on-fibre derivatization technique is dis-
cussed further below). Following derivatization, the fibre was inserted into the
injector port and the derivatives desorbed for gas chromatographic separation.
Miniaturization of classical derivatization techniques facilitate analysis of
small samples. Harrison and co-workers [154] developed and applied such
techniques to the determination of oxidized lipids. The analytes were deposited
in capillary tubes which were placed in an all-glass apparatus containing a
receptacle for reagents: either N,O-bis(trimethylsilyl)trifluoroacetamide
(BSTFA) or methoxylamine hydrochloride. Upon heating, reagent gas (BSTFA
or methoxylamine) was released and the vapours derivatized the analytes.
Excess reagent was then removed under vacuum. Silylated analytes were
prepared for electrospray ionization (ESI) analysis by dissolution in aqueous
ammonium acetate-methanol. This treatment hydrolyzed the TMS esters but
left TMS ethers of alcohols intact. Increases in masses of 72 for trimethyl-
silylation and 29 for oximation were used to identify hydroxyl and carbonyl
groups. Ozonolysis and subsequent derivatization readily identified the number
and position of double bonds. Ozonolysis was performed simply by exposing the
sample in a capillary to ozone. Further oxidation of the ozonides with HO 2 2 and
formic acid, however, was likewise performed by heating the reagents in the
sealed container containing the capillaries with the analytes.
Anari et al. [155] combined t-butylmethylsilylation with pentafluorobenzyl-
ation and NICI to determine valproic acid and its products from enzymatic
oxidation. As is found for many PFB esters, NICI of the mixed TBDMS/PFB
derivatives exhibited an uninformative base peak at m/z = 181 corresponding to
the PFB group. The fragmentation of the TBDMS/PFB derivatives, however,
was not as extensive as that found for the TMS/PFB products. The mixed
TBDMS/PFB derivatives of the hydroxylated products exhibited reasonably
intense peaks corresponding to loss of the tbutyl and TBDMS groups and
provided the necessary structural information. The resulting method produced
sensitivities of 2 ng/ml with excellent precision. This was sufficiently sensitive
for studies with the small amounts of cDNA-expressed human cytochrome that
are available to investigators. In brief, a judicious selection and combination of
derivatization pathways and ionization techniques provided a highly sensitive
and selective method.
Yang et al. [156] proposed pentafluorophenyldimethylsilyl (flophemesyl)
chloride as a potential alternative to PFBBr which is quite noxious. As a
silylation, the reaction does not require catalysis and derivatization of saturated
fatty acids is complete at room temperature within 15 min. The flophemesyl
derivatives have excellent chromatographic properties. Electron impact spectra
show reasonably intense M-15 peaks that provide specificity and high sensitiv-
ity. The base peak for all derivatized analytes, however, was at m/z = 121
resulting from a loss of 2 fluorines from the fluorohydrocarbon tropylium ion at
m/z = 159. Flophemsyl was also used to derivatize and determine unsaturated
fatty acids [157]. Although the chromatographic properties were again excellent,
646
in the case of unsaturated fatty acids fragmentation was controlled by the double
bond position and there were relatively fewer ions of higher mass.
19.5.2 Alkylation
647
reaction temperature, and reaction time. Optimization was based on the central
composite design. The method achieved virtually quantitative recoveries of
analytes from spiked water samples (96-103%) with excellent precision (RSD <
3.5%). Detection limits ranged from 0.05 to 14 gug/l.
Aging of old masters' paintings can be evaluated by studying oxidation of
abietic acids. Van den Berg et al. [163] investigated several derivatizations for
determining these oxidation products in varnish from the old works of art. They
pointed out that methylation with either TMS-diazomethane or diazomethane
(the former is a less toxic methylating agent) itself is insufficient as many
hydroxylated compounds would remain undetected. Trimethylsilylation of hydr-
oxyl groups-including the tertiary hydroxyl group at the tertiary position-aids
in chromatography of the hydroxylated materials but the mass spectra are domi-
nated by fragmentation via loss of a methyl group. Methylation of the carboxylic
acids followed by silylation was viewed with concern as more extended sample
handling could contaminate the samples which are exceedingly limited in both
size and supply. The authors opted for an on-line derivatization with tetra-
methylammonium hydroxide (TMAH). At elevated temperatures TMAH affects
the following reactions: (1) it esterifies the C-19 carboxyl, (2) it forms the methyl
ether of the tertiary hydroxyl at C-15, and (3) it enolizes and methylates the
ketone at C-7 but also produces the side products via a number of pathways. The
permethylated abietic acids and various reaction products can be separated by
GC. The TMAH was preferred to the "softer" methylating reagents such as
trimethyltrifluorotoluylammonium hydroxide (TMTFH) because the latter does
not methylate tertiary hydroxyls. Van den Berg and co-workers [163] argued
that although there are more degradation products with TMAH, these are
known. It is therefore possible to compensate for the degradation in the analysis.
The advantages of minimal sample handling predominate for the analysis of
these very rare and unique samples.
648
ble organic solvents with lipophilic chloroformates simultaneously extract and
derivatize analytes from water. Kim and Kim [165] described extractive deriva-
tization of phenols from natural waters. The analytes were phenol itself, which
is very hydrophilic, and a substantial number of increasingly substituted (and
more lipophilic) phenols in water. The aqueous phase was acidified and
contacted with methylene chloride containing isobutylchloroformate as the
reagent and trimethylamine as a basic catalyst. The phenols were extracted and
derivatives were retained in the organic phase which was evaporated and the
isolate was further purified prior to analysis.
Extractive derivatization provided a method for determination of estrone
(El) and estradiol (E2). Agitation of mixture consisting of the aqueous sample at
pH = 7, isobutyl-chloroformate and hexane, simultaneously derivatized and
extracted the analytes [166]. Under these conditions, only the phenolic group at
C-3 was derivatized. Prior to gas chromatographic analysis the secondary
hydroxyl at C-17 was derivatized with chlorodifluoroacetic anhydride to produce
the dichlorofluoroacetyl product. The resulting derivatives fragmented to give
high intensity fragments resulting from a loss of the isobutyryl group.
Blair et al. [167] derivatized y-hydroxybutyrate (GHB) with hexylchloro-
formate directly in aqueous matrices, including urine. The derivatives were
isolated from the sample by SPME and analyzed with GC-MS. Results obtained
by the SPME-GC-MS method were comparable to that obtained by standard
methods. The SPME is simpler, faster and is a solvent-less injection. In addition,
the overall technique can differentiate GHB from y-butyrolactone (GBL).
Reaction between amphetamines and 2,2,2-trichloroethylchloroformate
produces an electrophoric reagent suitable for separation by GC and detection
by the classical electron capture detector or mass spectrometry. March and
co-workers developed high quality methods based on this reagent [168].
Halogenated acid anhydrides are a classical derivatization for amines and this
derivatization was combined with SPME to yield techniques that meet the more
current requirements of speed, simplicity and automation. Jurado and co-
workers [169] and Lee et al. [170] used SPAD for determination of amphet-
amines. In both instances, the analytes were first sorbed from alkalinized matrix
onto a polydimethylsiloxane fibre. The syringe was then transferred to another
vial and exposed to vapours of trifluoroacetic anhydride [169] or heptafluoro-
butyric anhydride [170]. Finally, the fluorinated derivatives were desorbed by
heating in the injector for 5 min. Reaction time and temperature were critical
both for efficient sorption and derivatization. The range of conditions was quite
varied. Jurado et al. [169] carried out their procedures at temperatures between
60 and 100°C and 20 min extraction time to isolate the analytes and prepare the
derivatives. Lee's group [170] performed the derivatization at 270°C to achieve a
649
reaction with heptafluorobutyric anhydride in 10 s, but perhaps the need for ele-
vated higher also reflected the reduced volatility of the reagent.
Namera et al. [171] isolated basic drugs (fenfluramine, amphetamine and
methamphetamine) from whole blood via head-space SPME and using heat to
volatilize the analytes. The approach differed from the SPME derivatizations
described above [169,170]. In this technique heptafluorobutyric anhydride was
first injected into the injector of the GC-MS and upon insertion of the SPME
fibre the compounds were simultaneously desorbed and derivatized. Sensitiv-
ities were as low as 5 ng/G of blood. This unusual measure for concentration was
used because the samples were of forensic origin and the blood had clotted.
There were two recent reports on pentafluorobenzoylation for determina-
tion of alcohols by GC-MS [172,173]. The work was undertaken because in both
instances other reagents did not produce the necessary chromatographic proper-
ties, informative mass spectra and high sensitivities. Xiao and McCalley [172]
compared fragmentation patterns and sensitivity obtained for pentafluoro-
benzoyl (PFBO) derivatives with those obtained from trifluoroacetyl, hepta-
fluorobutyl, pentadecafluorooctanoyl and perfluorotolyl derivatives. PFBO
derivatization in quantitative yield occurred at each hydroxyl in the three
natural estrogens. The phenol of El was pentafluorobenzoylated as were the
phenol and 17a hydroxyl on E2 and the phenol and the 16,17 vicinal diol on E3.
There was, however, no derivatization of the synthetic and highly hindered 173
hydroxyl of 17a ethinylestradiol. Base peak of mass spectra for PFBO deriva-
tives was the molecular ion with diagnostic fragments arising from the loss of
one, two or three PBO groups.
Weintraub and co-workers [173] reviewed the literature and discussed their
own work on analytical derivatization of platelet activating factor (PAF). They
also found that pentafluorobenzoylation of PAF coupled with GC-MS/negative
ion chemical ionization (NICI) provided the necessary separations and sensitiv-
ity. Initially they reacted PAF with pentafluorobenzoyl chloride (PFBOC1) but
found this problematic because of the reactivity of the Cl group. They recently
prepared and used pentafluorobenzoic anhydride (PFBanh). The highest yields
were obtained with pure molten PFBanh at 110°C. The resulting transacylation
produced mixtures of 1-Oacyl-2-acetyl-3-PFB and 1-O-acyl-2-PFB-3-acetyl
glycerols.
650
As previously noted, formaldehyde is a frequently observed environmental
toxin. Koziel et al. [175] determined formaldehyde in air using SPAD with
0-(2,3,4,5,6-pentafluorobenzyl)hydroxylamine hydrochloride (PFBONH2 ) as
the derivatizing reagent that was deposited on an SPME fibre. A theoretical
treatment of the reaction mechanism demonstrated that sorption-reaction times
should be relatively short and that the loading of reagent on the fibre should be
high. This combination ensures that the amount of reagent used up in the
reaction would be negligible and the reaction rate would be high. Results
obtained by this method were equivalent to those obtained by NIOSH methods,
and, as in the case of most SPME methods, it was found to be simpler in
implementation.
Pentafluorobenzaldehyde (PFBAld) derivatized the biogenic primary alkyl-
amines in wines by GC-MS to form the Schiff base [176]. This is the converse of
an analogous reaction reported by Cancilla et al. [177] who determined
carbonyls in natural waters by reaction with PFBONH2 under acid conditions.
In the former work, however, because the analytes were amines the optimum
yield was achieved at pH = 12.
Alpha ketoisocaproic acid (KIC) and other keto acids present interesting
analytical problems. The need for methods for determining KIC derives from its
function as a dietary replacement for L-leucine and its role in the conversion of
D-leucine into L-leucine. It is relatively difficult to extract and isolate this and
other low molecular weight analytes of intermediary metabolism. Matsukawa et
al. [178] derivatized KIC with N-phenyl-1,2-phenylenediamine to form a stable
heterocyclic derivative of the analyte. This product exhibited a sharp peak on GC
and provided spectra with strong molecular ions as well as informative frag-
ments. The limit of quantitation using 50 tkl of plasma was 50 nM. Since the
reaction was based on the a ketoacid moiety, this derivatization could be used to
determine other a ketoacids. Reaction conditions used in this work are harsh
(100°C) and may have to be modified for other analytes which may not be as
stable.
651
Despite the high quality of information available from mass spectrometry,
determination of D and L-leucine by this instrumental technique remains
dependent on proper selection of derivatizing reagent. Hasegawa and co-
workers [179,180] selected a-methoxy-a-trifluoro-methylphenylacetyl chloride
(MTPA-Cl) as a chiral reagent because the fully substituted chiral centre is
impervious to racemization. The derivatization and chromatographic
procedures were relatively straightforward with good separation between D and
L-leucine with both the + and - MTPA but the presence of L-isoleucine, an
isomer of leucine, provided the analytical challenge. It was necessary not just to
separate the two enantiomers bot to all three amino acids prior to mass
spectrometry this was achieved only with +MTPA.
19.5.6.3 Bromoacetonitrile
Investigators are beginning to use bromoacetonitrile (BrACN) for analytical
derivatizations. Very long chain fatty acids [184] and phenols [185] were
derivatized with BrACN to the cyanomethyl esters and ethers. Separation of
these products by GC and detection with a nitrogen phosphorus selective (NPD)
detector sensitivity and specificity. Detection limits of 100 pg/ml [185] were
achieved and this was sufficient to detect the analyte in tap water as well as
environmental waters.
652
the mass spectrometer itself can provide substantial separation and quanti-
tation. Despite the technological and scientific sophistication of the technique,
there are numerous instances where derivatization can enhance both the
sensitivity of detection and the quality of information on fragmentation and
structure of the analyte.
Pentafluorobenzyl bromide (PFBBr) has been used extensively for gas chroma-
tographic determination of organic acids [137]. Derivatization is very straight
forward and the five fluorine atoms give very high sensitivity via the electron
capture detector and by NICI. Singh et al. [188] reasoned that in conventional
atmospheric pressure chemical ionization the displacements of electrons from
the nitrogen sheath gas in the corona discharge creates a source of gas phase
electrons. Under these conditions pentafluorobenzyl (PFB) derivatives should
be detected at very high sensitivity because these derivatives will efficiently
capture electrons and then fragment. They prepared and determined PFB esters
and ethers of amino prostaglandins, phenolic estrogens and adducts of 2'-deoxy-
guanosine produced by the reaction between 4-oxo-2-nonenal and DNA. The
PFB derivatives fragmented by dissociative electron capture and produced very
characteristic spectra. The technique was termed electron capture atmospheric
pressure chemical ionization (ECAPCI). HPLC-MS/ECAPCI coupled with
653
selected reaction monitoring produced detection limits as low as 140 attomole
for prostaglandins and thromboxanes. The results for the prostaglandins are
particularly interesting. Pentafluorobenzylation is only one step in the prepara-
tion of these eicosanoids prior to determination by GC-MS. A highly sensitive
HPLC-MS/MS technique would eliminate the additional silylations and
oximations that are required for analysis by GC-MS.
19.6.3 Saccharides
HO
2
RMP CH,
HO
HO
OH
Fig. 19.10. Reaction between PMP and a O o o
sugar. H H
654
derivatized and underivatized asialo oligosaccharides, but substantial improve-
ments were found for the derivatized mono and disialo compounds. Compared to
the native N-linked oligosaccharides, derivatization of the monosialo compounds
increases the sensitivity by 100-fold and detection of the disialo analyte is
increased by 100% following derivatization [88]. Moreover, the mass spectra of
the native compounds are uninformative as to the branching. In contrast, the
PMP derivatives fragment to produce ions that contained the reducing end of
the molecule and this facilitates interpretation. Derivatization of neutral and
N-acetylated proved equally beneficial. In this case sensitivity of detection by
mass spectrometry was seen for all analytes. Again, PMP derivatives underwent
directed cleavage towards the reducing end of oligosaccharide.
The sulphonated fluorophore, 2-aminonaphthalene trisulphone (ANTS) is a
useful reagent for the analysis of oligo and polysaccharides. The derivatizing
group imparts a charge onto neutral oligo and polysaccharides. As such it
produces derivatives suitable for electrophoresis as well as providing fluores-
cence for high sensitivity detection [190]. Poly and oligosaccharides derivatized
with ANTS were separated by polyacrylamide gel electrophoresis and by CZE.
This utility of this reagent for HPLC-ESI-MS was compared to that of PMP to
determine whether a single reagent can be used for all these separation/detec-
tion techniques. Due to its greater lipophilicity, PMP provided superior
chromatographic properties in both normal and reversed phase HPLC. In
addition, while ANTS was difficult to separate from the highly water derivatives,
a simple liquid/liquid extraction separated excess lipophilic PMP from the
derivatives which were hydrophilic by virtue of their saccharide moiety.
As noted above, sialic acids are critical recognition molecules in biological
processes. Their importance in physiology is matched by the difficulty in produc-
ing fragments in MALDI-TOF that retain the sialic acids. Leavell and Leary [89]
addressed the important problem of stabilizing sialic acid on oligosaccharides
during mass spectral analysis. They prepared the diethylenetriamine (DETA)
derivative of sialylated oligosaccharides and then complexed these compounds
with transition metals. They demonstrated that whereas the underivatized
oligosaccharide lost the sialic acids the cobalt complexed DETA derivative
produced an intense M-H + ion as well as several other fragments, including the
base peak, that retained the sialyl group.
655
phenyl-1,2,4-triazoline-3,5-dione (PTAD) [192] is one dienophile that reacts
with Vit D3 and its pharmaceutical analogue EWB-1089 [193]. PTAD deriva-
tives of the latter analyte in particular require more detailed study. The 22,
23-24, 24a diene in the side chain of EWB 1089 provides an alternative site for
derivatization. Weiskopf and co-workers [185] characterized the structure of the
diagnostic ion at 314. This information will enhance development of other meth-
ods for Vit D3, its analogues and other diene containing analytes. Higashi et al.
[193] derivatized VitD3 with 4-[4-(6-methoxy-benzoxazoyl)phenyl]-phenyl-
1,2,4-triazoline-3,5-dione (MBOTAD) and used chromatography on OASIS and
silica gel columns to purify the resulting derivative. Although sample prepara-
tion is lengthy the method is highly sensitive with a detection limit of 18 fmol
injected on column but in this work the mass spectrometry of the derivatives was
not studied.
19.6.5 Alcohols
Quirke and Van Berkel [194] continued their work on the preparation of
ferrocene derivatives for the derivatization of alcohols. Derivatives are simply
prepared by heating the alcohols with ferrocenoyl azide in anhydrous toluene.
The neutral derivatives are oxidized to the corresponding ferrocenium products
(F+ ) during the processes involved in electrospray ionization. The combination
of a variety of mass spectrometric techniques gives a ready identification of
ferrocenoyl derivatives of the primary, secondary and tertiary alcohols as well as
providing information on other structural features.
656
linked via the maleimido group and this left the biocytin (or biotin) group free for
subsequent sorption onto a column loaded with covalently bound avidin. Avidin
readily binds biotin and provided a selective separation technique. The NEM-
RHO-MBB protein was eluted from the Sepharose column by displacement with
nonapeptide and was hydrolyzed by proteinase K. The peptide fragments were
then added to an avidin-agarose column and those peptides containing the biotin
moiety were sorbed. Washing with water removed the peptides without MBB
tags. Finally the MBB labelled peptides were eluted by displacement with biotin
and analyzed by MALDI-TOF. In the normal RHO, H tagged the cysteine at the
185 position. The MBB label was attached to cysteines at the 110 and 187
position, demonstrating that in the protein these two cysteines had originally
formed an CysS-SCys bond. In misfolded RHO the cysteine at the 185 and the
187 positions were tagged with MBB whereas the cysteine at the 110 position
carried the NEM tag. It follows that the disulphide bond in misfolded RHO was
between 185 and 187.
A similar approach was used to identify the contact sites between RHO and T
[71,72]. Sensitization and desensitization are controlling features in light activa-
tion of visual signal transduction. A critical feature of this process is the binding
of T and rhododopsin kinase to light activated RHO. Competition between T and
the kinase for binding interaction controls activation and quenching of the
biochemical cascades involved in vision. Elucidation and understanding of these
phenomena requires a more detailed knowledge of the binding sites. Analytical
derivatizations at the thiol group of cysteine and other moieties were key in
identifying these sites.
The first step was binding RHO to Sepharose beads through the RHO-
1D4-antibody. Subsequent reactions were then performed on RHO immobilised
on the solid surface (S-RHO). The first reaction was with one of two pyridyl-
dithio reagents that were designed to ultimately cross-link the RHO and T using
either light activation or chemical activation. The light activated reagent was
N-((2-pyridyldithio)ethyl)-4-azidosalicylamide (PEAS) and N-succinimidyl 3-
(2-pyridyldithio) propionate (SPDP) [71] was the chemically activated [72]
reagent. These reagents were linked to the S-RHO via a disulphide group that
could later be reduced.
Addition of a GDP/T complex to the S-RHO-linking group complex followed
by activation with light at >495 nm bound the GDP/T complex with the
S-RHO-linking group [71]. Irradiation at 310 nm activated the PEAS labelled
S-RHO while adjusting the pH activated the SPDP labelled S-RHO [72]. The
activation cross-linked the S-RHO and T. Washing the cross-linked complex
with GTP removed the GDP required for the binding of T to RHO. Deriva-
tization with NEM blocked the remaining CysSH groups and reduction with
DTT broke the link between the S-RHO and the now labelled T. The freed and
tagged T was then derivatized with MBB, as described above. The purification
and hydrolysis were also the same as used for identification of the disulphide link
in the misfolded RHO.
657
19.7 CHIRAL SEPARATIONS
658
where hydrogen bonding interactions enabled the chiral separations. Under
these conditions, a useful and interesting result is that the reagent elutes after
the derivatized analyte as the latter is more hydrophilic.
659
from the mercury derivatization product by reaction with NaBEt4 . Sodium
phenyltetraborate produces unique derivatives for all species of mercury.
Sodium tetraethylborate is also effective at producing volatile derivatives of
tin and lead. These derivatizations occur in the aqueous sample. Subsequent
isolation and injection into the chromatographic system require the develop-
ment of rapid and facile techniques. Milan and Pawliszyn [210] prepared the
derivatives of butyltin species using NaBEt4 . Alkyl lead and inorganic lead were
the subjects of an investigation by Yu and Pawliszyn [211] who derivatized these
analytes with deuterium-labelled sodium tetraethylborate NaB(C2 Ds)4
(DSTEB). Detection by mass spectrometry readily distinguished between Pb2 + ,
Pb(CH 3 )3+ , Pb(C2 H5 )3 +, and Pb(C2 H5 )4 since any derivative had the deuterated
tag which is not found in nature. This was analogous to the use of NaBPh4 to
produce an unambiguous determination of metal species in the environment. In
both instances, for study of tin or lead, the isolation was affected by SPME with
collection in the head-space. The techniques provided results that matched the
requirements of regulatory agencies and because of the SPME technique are
well suited to field analysis.
Methylarsonic acids were converted to lipophilic species with thioglycol
methylate [212]. Derivatization was carried out directly in acidified urine and
the derivatives were extracted into hexane for subsequent analysis by GC-MS.
Although there is a substantial dilution factor in the extraction, the overall
technique is simple, rapid and provides good sensitivity and reproducibility.
19.9 SUMMARY
19.10 GLOSSARY
660
Biogenic amines: Any amine, other than amino acids or nucleic acids that are
part of normal physiology. This includes the catecholamine, indoleamines,
y-hydroxybutyric acid. The catecholamines include epinephrine (adren-
aline) nor epinephrine (nor-adrenaline) and dopamine. Indoleamines
include serotonin, N-acetylserotonin and melatonin. These compounds are
involved in the functioning of the central and autonomic nervous system
and are responsible for the fight or flight response.
Biogenic polyamines: Diamines such as putrescine, cadaverine, spermidine
and spermine. These are compounds involved in signalling between cells
and are thus important markers of normal tissue development and develop-
ment of tumours.
Bio-protective peptides and amino acids: Peptides or amino acids that
contain a free thiol (-SH) that can react with ROS. The peptides are
glutathione (GSH) and y-glutamylcysteine and the amino acid is cysteine.
The major bio-protective peptide is GSH. In addition to reacting with ROS,
the peptide GSH can also react with alkylating species. Alkylating species
are usually drugs or reactive compounds formed from the biotrans-
formation of drugs.
Chemical cytometry: The qualitative and quantitative determination of
compounds produced by a single cell and use of this data to interpret the
biology of the cell.
Dextran ladder: A group of oligosaccharides derived from dextran that provide
a series of saccharides with increasing molecular weight. Used to calibrate
columns for determination of molecular weight of unknown mono, oligo
and polysaccharides.
Liquid chromatography: Any chromatographic technique in which the
mobile phase is liquid and the stationary phase is a solid. This includes high
performance liquid chromatography, thin layer chromatography and
capillary zone electrophoresis.
Metabolic cytometry: The analysis of metabolic interconversions in a single
cell.
Rhododopsin: One of two proteins that control the activation of the retina in
response to light.
Solid phase extraction (SPE): Isolation of an analyte from aqueous matrix
by sorption onto a solid phase and subsequent elution with a volatile
organic solvent.
Solid phase micro-extraction (SPME): Isolation of an analyte from aqueous
matrix by sorption onto a solid phase consisting of a glass fibre coated with
sorbent. The glass fibre is on a device that permits direct insertion into an
injector port of a gas chromatograph for thermal desorption or a port of an
liquid chromatograph for on-line elution.
Solid phase analytical derivatizations (SPAD): A combination of sorption
of analyte onto a solid phase and derivatization on the phase.
Transducin: One of two proteins that controls the activation of the retina in
661
response to light.
Vasodilator: Any endogenous or exogenous compound that increases the
lumen of the blood vessels. Endogenous vasodilators such as nitric oxide
serve to maintain blood pressure and respond to normal or pathophysi-
ological events that may occlude the blood vessel. Exogenous vasodilators
are drugs that are used to overcome deficits in the functioning of the
endogenous vasodilators.
REFERENCES
662
27 C. Parmentier, M. Wellman, A. Nicolas, G. Siest and P. Leroy, Electrophoresis,20
(1999) 2938-2944.
28 Z. Chen, J. Wu, G.B. Baker, M. Parent and N.J. Dovichi, J. Chromatogr.A, 914 (2001)
293-298.
29 Z. Zhang, S. Krylov, E.A. Arriaga, R. Polakowski and N.J. Dovichi, Anal. Chem., 72
(2000) 318-322.
30 J. Wu, Z. Chen and N.J. Dovichi, J. Chromatogr.B, Biomed. Sci. Appl., 741 (2000)
85-88.
31 S. Hu, Z. Zhang, L.M. Cook, E.J. Carpenter and N.J. Dovichi, J. Chromatogr. A, 894
(2000) 291-296.
32 I.H. Lee, D. Pinto, E.A. Arriaga, Z. Zhang and N.J. Dovichi, Anal. Chem., 70 (1998)
4546-4548.
33 S.N. Krylov, Z. Zhang, N.W. Chan; E. Arriaga, M.M. Palcic and N.J. Dovichi,
Cytometry, 37 (1999) 14-20.
34 S.N. Krylov, D.A. Starke, E.A. Arriaga, Z. Zhang, N.W. Chan, M.M. Palcic and N.J.
Dovichi, Anal. Chem., 72 (2000) 872-877.
35 Z. Zhang, S. Krylov, E.A. Arriaga, R. Polakowski and N.J. Dovichi, Anal. Chem., 72
(2000) 318-322.
36 S. Hu, R. Lee, Z. Zhang, S.N. Krylov and N.J. Dovichi, J. Chromatogr.B, Biomed. Sci.
Appl., 752 (2001) 307-310.
37 G. Aymard, B. Labarthe, D. Warot, I. Berlin and B. Diquet, J. Chromatogr. B,
Biomed. Sci. Appl., 744 (2000) 25-31.
38 R.W. Sparidans, J. den Hartig, S. Cremers, J.H. Beijnen and P. Vermeij, J.
Chromatogr. B, Biomed. Sci. Appl., 738 (2000) 331-341.
39 R. Herraez-Hernandez and P. Campins-Falco, Analyst, 125 (2000) 1071-1076.
40 D.H. Shangguan, Y.X. Zhao, H.W. Han, R. Zhao and G.Q. Liu, Anal. Chem., 73 (2001)
2054-2057.
41 M. Molina and M. Silva, Electrophoresis,22 (2001) 1175-1181.
42 E. Causse, N. Siri, J.F. Arnal, C. Bayle, P. Malatray, P. Valdiguie, R. Salvayre and F.
Couderc, J. Chromatogr.B, Biomed. Sci. Appl., 741 (2000) 77-83.
43 T. Toyo'oka, D. Jin, N. Tomoi, T. Oe and H. Hiranuma, Biomed. Chromatogr., 15
(2001) 56-67, 2001.
44 M.J. Baars and G. Patonay, Anal. Chem., 71 (1999) 667-671.
45 S. Hu and P.C. Li, J. Chromatogr.A, 876 (2000) 183-191.
46 T. Inoue, K. Hamase, A. Morikawa and K. Zaitsu, J. Chromatogr.B, Biomed. Sci.
Appl., 744 (2000) 213-219.
47 M. Kato, M.T. Dulay, B. Bennett, J. Chen and R.N. Zare, Electrophoresis, 21 (2000)
3145-3151.
48 S. Honda, J. Okeda, H. Iwanaga, S. Kawakami, A. Taga, S. Suzuki and K. Imai, Anal.
Biochem., 286 (2000) 99-111.
49 C. Molins-Legua, P. Campins-Falco, A. Sevillano-Cabeza and M. Pedron-Pons,
Analyst, 124 (1999) 477-482.
50 H. Nohta, H. Satozono, K. Koiso, H. Yoshida, J. Ishida and M. Yamaguchi, Anal.
Chem., 72 (2000) 4199-4204.
51 A. Asthana, D. Bose, A. Durgbanshi, S.K. Sanghi and W.T. Kok, J. Chromatogr. A,
895 (2000) 197-203.
52 O.Y. Al-Dirbashi, M. Wada, N. Kuroda, M. Takahashi and K. Nakashima, J. Forensic
Sci., 45 (2000) 708-714.
53 Y. Sun, M. Wada, O. Al Dirbashi, N. Kuroda, H. Nakazawa and K. Nakashima, J.
Chromatogr. B, Bioamed. Sci. Appl., 749 (2000) 49-56.
54 M. Wada, S. Kinoshita, Y. Itayama, N. Kuroda and K. Nakashima, J. Chromatogr.B,
Biomed. Sci. Appl., 721 (1999) 179-186.
663
55 H. Wang, J. Li, X. Liu, T.X. Yang and H.S. Zhang, Anal. Biochem., 281 (2000) 15-20.
56 Z. Yan, J.G. Nikelly, L. Killmer, Jr. and J.B. Tarloff, Drug Metab Dispos., 28 (2000)
880-886.
57 F.Q. Schafer and G.R. Buettner, Free Radic. Biol. Med., 30 (2001) 1191-1212.
58 J.B. Ubbink, Semin. Thromb. Hemost., 26 (2000) 233-241.
59 M.J. MacCoss, N.K. Fukagawa and D.E. Matthews, Anal. Chem., 71 (1999)
4527-4533.
60 A. de Graaf-Hess, F. Trijbels and H. Blom, Clin. Chem., 45 (1999) 2224-2228.
61 D. Tsikas, J. Sandmann, D. Holzberg, P. Pantazis, M. Raida and J.C. Frolich, Anal.
Biochem., 273 (1999) 32-40.
62 J.F. Salazar, H. Schorr, W. Herrmann, B. Herbeth, G. Siest and P. Leroy, J.
Chromatogr. Sci., 37 (1999) 469-476.
63 E.D. Bald, R. Glowacki and J. Drzewoski, J. Chromatogr. A, 913 (2001) 319-329.
64 A.R. Ivanov, I.V. Nazimov, L. Baratova, A.P. Lobazov and G.B. Popovich, J.
Chromatogr. A, 913 (2001) 315-318.
65 A.R. Ivanov, I.V. Nazimov and L. Baratova, J. Chromatogr.A, 895 (2000) 157-166.
66 A.R. Ivanov, I.V. Nazimov and L.A. Baratova, J. Chromatogr.A, 870 (2000) 433-442.
67 E. Causse, N. Siri, H. Bellet, S. Champagne, C. Bayle, P. Valdiguie, R. Salvayre and F.
Couderc, Clin. Chem. 45 (1999) 412-414.
68 D.E. Hammermeister, J. Serrano, P. Schmieder and D.W. Kuehl, Rapid Commun.
Mass Spectrom. 14 (2000) 503-508.
69 A.I. Haj-Yehia, P. Assaf, T. Nassar and J. Katzhendler, J. Chromatogr.A, 870 (2000)
381-388.
70 J. Hwa, J. Klein-Seetharaman and H.G. Khorana, Proc. Natl. Acad. Sci. USA, 98
(2001) 4872-4876.
71 K. Cai, Y. Itoh and H.G. Khorana, Proc. Natl. Acad. Sci. USA, 98 (2001) 4877-4882.
72 Y. Itoh, K. Cai and H.G. Khorana, Proc. Natl. Acad. Sci. USA, 98 (2001) 4883-4887.
73 B.W. Boyd, S.R. Witowski and R.T. Kennedy, Anal. Chem., 72 (2000) 865-871.
74 H. Kamencic, A. Lyon, P.G. Paterson and B.H. Juurlink, Anal. Biochem., 286 (2000)
35-37.
75 S.H. Kang, W. Wei and E.S. Yeung, J. Chromatogr.B, Biomed. Sci. Appl., 744 (2000)
149-156.
76 E. Causse, C. Issac, P. Malatray, C. Bayle, P. Valdiguie, R. Salvayre and F. Couderc,
J. Chromatogr.A, 895 (2000) 173-178.
77 E. Causse, P. Malatray, R. Calaf, P. Charpiot, M. Candito, C. Bayle, P. Valdiguie, R.
Salvayre and F. Couderc, Electrophoresis,21 (2000) 2074-2079.
78 V. Ducros, M. Candito, E. Causse, R. Couderc, K. Demuth, M.E. Diop, J. Drai, A.M.
Gachon, I. Garcia, M.F. Gerhardt, C. Philippe-Bourgeois, M.H. Read and M.P.
Sauvant, Ann. Biol. Clin. (Paris), 59 (2001) 33-39.
79 N. Ercal, P. Yang and N. Aykin, J. Chromatogr.B, Biomed. Sci. Appl., 753 (2001)
287-292.
80 C. Cereser, J. Guichard, J. Drai, E. Bannier, I. Garcia, S. Boget, P. Parvaz and A.
Revol, J. Chromatogr. B, Biomed. Sci. Appl., 752 (2001) 123-132.
81 K.J. Lenton, H. Therriault and J.R. Wagner, Anal. Biochem., 274 (1999) 125-130.
82 G. Durand and N. Seta, Clin. Chem., 46 (2000) 795-805.
83 N. Volpi, Anal. Biochem., 277 (2000) 19-24.
84 C. Sato, S. Inoue, T. Matsuda and K. Kitajima, Anal. Biochem., 266 (1999) 102-109.
85 S.L. Lin, Y. Inoue and S. Inoue, Glycobiology, 9 (1999) 807-814.
86 J. Charlwood, H. Birrell, A. Gribble, V. Burdes, D. Tolson and P. Camilleri, Anal.
Chem., 72 (2000) 1453-1461.
87 H. Birrell, J. Charlwood, I. Lynch, S. North and P. Camilleri, Anal. Chem., 71 (1999)
102-108.
664
88 J.A. Saba, X. Shen, J.C. Jamieson and H. Perreault, Rapid Commun. Mass
Spectrom., 13 (1999) 704-711.
89 M.D. Leavell and J.A. Leary, J. Am. Soc. Mass Spectrom., 12 (2001) 528-536.
90 S. Hanrahan, J. Charlwood, R. Tyldesley, J. Langridge, R. Bordoli, R. Bateman and
P. Camilleri, Rapid Commun. Mass Spectrom., 15 (2001) 1141-115.
91 J. Charlwood, J.M. Skehel and P. Camilleri, Anal. Biochem., 284 (2000) 49-59.
92 J. Charlwood, H. Birrell, D. Tolson, J. Connelly and P. Camilleri, Anal. Biochem., 283
(2000) 250-257.
93 Y. Araki, A. Andoh, Y. Fujiyama, K. Hata, J. Makino, T. Okuno, F. Nakanura and T.
Bamba, J. Chromatogr. B, Biomed. Sci. Appl., 753 (2001) 209-215.
94 K.R. Anumula and P. Du, Anal. Biochem., 275 (1999) 236-242.
95 J. Charlwood, J. Langridge, D. Tolson, H. Birrell and P. Camilleri, Rapid Commun.
Mass Spectrom., 13 (1999) 107-112.
96 J. Charlwood, J. Langridge and P. Camilleri, Rapid Commun. Mass Spectrom., 13
(1999) 1522-1530.
97 J. Charlwood, H. Birrell, A. Organ and P. Camilleri, Rapid Commun. Mass
Spectrom., 13 (1999) 716-723.
98 K.M. Hutterer, H. Birrell, P. Camilleri and J.W. Jorgenson, J. Chromatogr. B,
Biomed. Sci. Appl., 745 (2000) 365-372.
99 S.A. Visser, C.J. Smulders, W.W. Gladdines, H. Irth, P.H. van der Graaf and M.
Danhof, J. Chromatogr.B, Biomed. Sci. Appl., 745 (2000) 357-363.
100 D. Shangguan, H. Han, R. Zhao, Y. Zhao, S. Xiong and G. Liu, J. Chromatogr.A, 910
(2001) 367-372.
101 Z. Huang and D.J. Waxman, Anal. Biochem., 273 (1999) 117-125.
102 A. Paci, A. Rieutord, D. Guillaume, F. Traore, J. Ropenga, H.P. Husson and F. Brion,
J. Chromatogr.B, Biomed. Sci. Appl., 739 (2000) 239-246.
103 R.W. Sparidans, A. Veldkamp, B.M. Hoetelmans and J.H. Beijnen, J. Chromatogr.B,
Biomed. Sci. Appl., 736 (1999) 115-121.
104 H. Zhang, H. Ford, J.S. Roth and J.A. Kelley, J. Pharm. Biomed. Anal., 25 (2001)
285-297.
105 F. Dai, J.A. Kelley, H. Zhang, N. Malinowski, M.F. Kavlick, J. Lietzau, L. Welles, R.
Yarchoan and H. Ford Jr., Anal. Biochem., 288 (2001) 52-61.
106 K. Soda, Y. Ohba and K. Zaitsu, J. Chromatogr.B, Biomed. Sci. Appl., 752 (2001)
55-60.
107 B. Frey, R. Haupt, S. Alms, G. Holzmann, T. Konig, H. Kern, W. Kox, B. Rustow and
M. Schlame, J. Lipid Res., 41 (2000) 1145-1153.
108 M. Comesana Losada, J.M. Leao, A. Gago-Martinez, J.A. Rodriguez Vazquez and
M.A. Quilliam, J. Agric. Food Chem., 47 (1999) 618-621.
109 L. Ten-Hage, N. Delaunay, V. Pichon, A. Coute, S. Puiseux-Dao and J. Turquet,
Toxicon, 38 (2000) 1043-1054.
110 A. Hattori, T. Fukushima, K. Hamamura, M. Kato and K. Imai, Biomed.
Chromatogr., 15 (2001) 95-99.
111 A. Hattori, T. Fukushima and K. Imai, Anal. Biochem., 281 (2000) 209-215.
112 H. Yamada, Y. Kuwahara, Y. Takamatsu and T. Hayase, Biomed. Chromatogr., 14
(2000) 333-337.
113 B. Yagen and S. Burstein. J. Chromatogr.B, Biomed. Sci. Appl., 740 (2000) 93-99.
114 Y. Arai, T. Fukushima, M. Shirao, X. Yang and K. Imai, Biomed. Chromatogr., 14
(2000) 118-124.
115 P. Assaf, J. Katzhendler and A.I. Haj-Yehia, J. Chromatogr.A, 869 (2000) 243-250.
116 M. Sharma, Biochem. Biophys. Res. Commun., 273 (2000) 40-44.
117 C.Z. Matthews, E.J. Woolf, L. Lin, W. Fang, J. Hsieh, S. Ha, R. Simpson and B.K.
Matuszewski, J. Chromatogr. B, Biomed. Sci. Appl., 751 (2001) 237-246.
665
118 R. Gatti, M. Franchina, M.G. Gioia and V. Cavrini, Biomed. Chromatogr., 14 (2000)
82-88.
119 K. Hamase, T. Tomita, A. Kiyomizu and K. Zaitsu, Anal. Biochem., 279 (2000)
106-110.
120 S.N. Krylov, E.A. Arriaga, N.W. Chan, N.J. Dovichi and M.M. Palcic, Anal. Biochem.,
283 (2000) 133-135.
121 J.Y. Zhao, P. Diedrich, Y. Zhang, O. Hindsgaul and N.J. Dovichi, J. Chromatogr.B,
Biomed.Appl. 657 (1994) 307-313.
122 Y. Zhang, X. Le, N.J. Dovichi, C.A. Compston, M.M. Palcic, P. Diedrich and 0.
Hindsgaul, Anal. Biochem., 227 (1995) 368-376.
123 M.L. Miller and G.V. Johnson, J. Chromatogr. B, Biomed. Sci. Appl., 732 (1999)
65-72.
124 D.M. Albin, J.E. Wubben and V.M. Gabert, J. Agric. Food Chem., 48 (2000) 1684-
1691.
125 E. Zahou, H. Jornvall and T. Bergman, Anal. Biochem., 281 (2000) 115-122.
126 F. Wang, X. Chen, Q. Chen, X. Qin and Z. Li, J. Chromatogr.A, 883 (2000) 113-118.
127 N. Isoherranen and S. Soback, Clin. Chem., 46 (2000) 837-842.
128 C. Arcelloni, B. Comuzzi, R. Vaiani and R. Paroni, J. Chromatogr. B, Biomed. Sci.
Appl., 753 (2001) 151-156.
129 E. Kaale, S. Leonard, A. Van Schepdael, E. Roets and J. Hoogmartens, J.
Chromatogr.A, 895 (2000) 67-79.
130 J. Kirschbaum, K. Rebscher and H. Bruckner, J. Chromatogr.A, 881 (2000) 517-530.
131 R. Herraez-Hernandez, P. Campins-Falco and J. Verdu-Andres, Analyst, 126 (2001)
581-586.
132 D. Grosjean, E. Grosjean and A.W. Gertler, Environ. Sci. Technol., 35 (2001) 45-53.
133 J.J. Michels, J. Chromatogr. A, 914 (2001) 123-129.
134 G. Dusi and V. Gamba, J. Chromatogr.A, 835 (1999) 243-246.
135 M. Vogel, W. Potter and U. Karst, J. Chromatogr.A, 886 (2000) 303-307.
136 H. Miwa, J. Chromatogr. A, 881 (2000) 365-385.
137 Z. Glatz and H. Maslanova, J. Chromatogr. A, 895 (2000) 179-187.
138 R. Glowacki, K. Wojcik and E. Bald, J. Chromatogr.A, 914 (2001) 29-35.
139 M.H. Quernin, B. Poonkuzhali, Y. Medard, D. Dennison, A. Srivastava, R. Krishna-
moorthy, M. Chandy and E. Jacqz-Aigrain, J. Chromatogr. B, Biomed. Sci. Appl., 721
(1999) 147-152.
140 N. Bleyzac, P. Barou and G. Aulagner, J. Chromatogr.B, Biomed. Sci. Appl., 742
(2000) 427-432.
141 R. Raghavan, M. Burchett, D. Loffredo and J.A. Mulligan, Drug Dev.Ind.Pharm. 26
(2000) 429-440.
142 J. Wang, M.P. Chatrathi and B. Tian, Anal. Chem., 72 (2000) 5774-5778.
143 Y.V. Tcherkas, L.A. Kartsova and I.N. Krasnova, J. Chromatogr. A, 913 (2001)
303-308.
144 D.R. Knapp, Handbook of Analytical Derivatization Reactions. WileyInterscience,
John Wiley &Sons, Inc. 1979.
145 J.L. Little, J. Chromatogr.A, 844 (1999) 1-22.
146 M. Thevis, G. Opfermann, H. Schmickler and W. Schanzer, J. Mass Spectrom., 36
(2001) 159-168.
147 Y. $hinohara, K. Isurugi and T. Hashimoto, J. Chromatogr.B, Biomed. Sci. Appl.,
741 (2000) 271-278.
148 M. Thevis, G. Opfermann and W. Schanzer, J. Chromatogr. B, Biomed. Sci. Appl.,
744 (2000) 345-350.
149 Y. Kim, Y. Lee, M. Kim, Y.H. Yim and W. Lee, Rapid Commun.Mass Spectrom. 14
(2000) 1293-1300.
666
150 T. Ohie, X.W. Fu, M. Iga, M. Kimura and S. Yamaguchi, J. Chromatogr.B, Biomed.
Sci. Appl., 746 (2000) 63-73.
151 H.G. Mol, S. Sunarto and O.M. Steijger, J. Chromatogr. A, 879 (2000) 97-112.
152 C. Rodier, R. Sternberg, F. Raulin and C. Vidal-Madjar, J. Chromatogr. A, 915 (2001)
199-207.
153 U. Staerk and W.R. Kulpmann, J. Chromatogr. B, Biomed. Sci. Appl., 745 (2000)
399-411.
154 K.A. Harrison, S.S. Davies, G.K. Marathe, T. McIntyre, S. Prescott, K.M. Reddy, J.R.
Falck and R.C. Murphy, J. Mass Spectrom., 35 (2000) 224-236.
155 M.R. Anari, R.W. Burton, S. Gopaul and F.S. Abbott, J. Chromatogr. B, Biomed. Sci.
Appl., 742 (2000) 217-227.
156 Y.J. Yang, M.H. Choi, M.J. Paik, H.R. Yoon and B.C. Chung, J. Chromatogr. B,
Biomed. Sci. Appl., 742 (2000) 37-46.
157 M.H. Choi and B.C. Chung, Anal. Biochem., 277 (2000) 271-273.
158 X.P. Lee, T. Kumazawa, K. Kondo, K. Sato and 0. Suzuki, J. Chromatogr.B, Biomed.
Sci. Appl., 734 (1999) 155-162.
159 M.N. Sarrion, F.J. Santos and M.T. Galceran, J. Chromatogr. A, 859 (1999) 159-171.
160 M.N. Sarrion, F.J. Santos and M.T. Galceran, Anal. Chem., 72 (2000) 4865-4873.
161 M.I. Catalina, J. Dalluge, R.J. Vreuls and U.A. Brinkman, J. Chromatogr.A, 877
(2000) 153-166.
162 C.D. Stalikas and G.A. Pilidis, J. Chromatogr.A, 872 (2000) 215-225.
163 K.J. Berg, J.J. Boon, I. Pastorova and L.F. Spetter, J. Mass Spectrom., 35 (2000)
512-533.
164 J. Pietzsch, Biochem. Biophys. Res. Commun., 270 (2000) 852-857.
165 K.R. Kim and H. Kim, J. Chromatogr.A, 866 (2000) 87-96.
166 M.H. Choi, K.R. Kim and B.C. Chung, Analyst, 125 (2000) 711-714.
167 S. Blair, M. Song, B. Hall and J. Brodbelt, J. ForensicSci., 46 (2001) 688-693.
168 C. March, H.T. Karnes, A. McLean and P.S. Mukherjee, Biomed. Chromatogr., 15
(2001) 100-107.
169 C. Jurado, M.P. Gimenez, T. Soriano, M. Menendez and M. Repetto, J. Anal. Toxicol.,
24 (2000) 11-16.
170 M.R. Lee, Y.S. Song, B.H. Hwang and C.C. Chou, J. Chromatogr. A, 896 (2000)
265-273.
171 A. Namera, M. Yashiki, J. Liu, K. Okajima, K. Hara, T. Imamura and T. Kojima,
Forensic Sci. Int., 109 (2000) 215-223.
172 X. Xiao and D. McCalley, Rapid Commun. Mass Spectrom., 14 (2000) 1991-2001.
173 S.T. Weintraub, R.K. Satsangi, E.A. Sprague, T.J. Prihoda and R.N. Pinckard, J. Am.
Soc. Mass Spectrom., 11 (2000) 176-181.
174 C.D. Stalikas and C.N. Konidari, Anal. Biochem., 290 (2001) 108-115.
175 J.A. Koziel, J. Noah and J. Pawliszyn, Environ. Sci. Technol., 35 (2001) 1481-1486.
176 K.K. Ngim, S.E. Ebeler, M.E. Lew, D.G. Crosby and J.W. Wong, J. Agric. Food
Chem., 48 (2000) 3311-3316.
177 D.A. Cancilla and S.S. Que Hee, J. Chromatogr., 627 (1992) 1-16.
178 T. Matsukawa, H. Hasegawa, Y. Shinohara and T. Hashimoto, J. Chromatogr. B,
Biomed. Sci. Appl., 751 (2001) 213-220.
179 H. Hasegawa, T. Matsukawa, Y. Shinohara and T. Hashimoto, Drug Metab. Dispos.,
28 (2000) 920-924.
180 H. Hasegawa, T. Matsukawa, Y. Shinohara and T. Hashimoto, J. Chromatogr. B,
Biomed. Sci. Appl., 735 (1999) 141-149.
181 P.T. Fenn, C. Ning, S. Segal and I.A. Blair, J. Mass Spectrom., 35 (2000) 218-223.
182 C. Ning, P.T. Fenn, I.A. Blair, G.T. Berry and S. Segal, Mol. Genet. Metab., 70 (2000)
261-271.
667
183 B.E. van Dongen, S. Schouten and J.S. Damste, Rapid Commun. Mass Spectrom., 15
(2001) 496-500.
184 M.J. Paik, K.O. Lee and H.S. Shin, J. Chromatogr.B, Biomed. Sci. Appl., 721 (1999)
3-11.
185 H.S. Shin, C.H. Park, S.J. Park and H. Pyo, J. Chromatogr.A, 912 (2001) 119-125.
186 X. Chen, V.E. Anderson and Y.H. Chen, J. Am. Soc. Mass Spectrom., 10 (1999)
448-452.
187 K. Gartenmann and S. Kochhar, J. Agric. Food Chem., 47 (1999) 5068-5071.
188 G. Singh, A. Gutierrez, K. Xu and I.A. Blair, Anal. Chem., 72 (2000) 3007-3013.
189 S. Suzuki, R. Tanaka, K. Takada, N. Inoue, Y. Yashima, A. Honda and S. Honda, J.
Chromatogr.A, 910 (2001) 319-329.
190 J.A. Saba, X. Shen, J.C. Jamieson and H. Perreault, J. Mass Spectrom., 36 (2001)
563-574.
191 X. Shen and H. Perreault, J. Mass Spectrom., 34 (1999) 502-510.
192 A.S. Weiskopf, P. Vouros, J. Cunniff, E. Binderup, F. Bjorkling, L. Binderup, M.C.
White and G.H. Posner, J. Mass Spectrom., 36 (2001) 71-78.
193 T. Higashi, D. Awada and K. Shimada, Biomed. Chromatogr., 15 (2001) 133-140.
194 J.M. Quirke and G.J. Van Berkel, J. Mass Spectrom., 36 (2001) 179-187.
195 K.H. Kim, D.S. Kim, S.P. Hong and O.S. Keon, Arch. Pharm. Res., 23 (2000) 26-30.
196 X. Li, T.W. Yao and S. Zeng, J. Chromatogr. B, Biomed. Sci. Appl., 742 (2000)
433-439.
197 S. Kochhar, B. Mouratou and P. Christen, Meth. Mol. Biol., 159 (2000) 49-54.
198 G. Torok, A. Peter, A. Gaucher, M. Wakselman, J.P. Mazaleyrat and D.W.
Armstrong, J. Chromatogr. A, 846 (1999) 83-91.
199 A. Peter, E. Olajos, R. Casimir, D. Tourwe, Q.B. Broxterman, B. Kaptein and D.W.
Armstrong, J. Chromatogr. A, 871 (2000) 105-113
200 M. Kagawa, Y. Machida and H. Nishi, J. Chromatogr.A, 857 (1999) 127-135.
201 M. Peter, A. Peter and F. Fulop, J. Chromatogr.A, 871 (2000) 115-126.
202 I.J. Kim, S.J. Park and H.J. Kim, J. Chromatogr.A, 877 (2000) 217-223.
203 C.R. Craig and R.E. Stitzel, Modern Pharmacology. Little-Brown, 1994.
204 H. Li, C.J. Meininger and G. Wu, J. Chromatogr. B, Biomed. Sci. Appl., 746 (2000)
199-207.
205 S. Kage, K. Kudo and N. Ikeda, J. Chromatogr. B, Biomed. Sci. Appl., 438 (2000)
363-368.
206 D. Tsikas, F.M. Gutzki, J. Sandmann, E. Schwedhelm and J.C. Frolich, J. Chroma-
togr. B, Biomed. Sci. Appl., 731 (1999) 285-291.
207 D.W. Nierenberg, R.E. Nordgren, M.B. Chang, B. Morris, R.W. Siegler, M.B. Blayney,
F. Hochberg, T.Y. Toribara, E. Cernichiari and T. Clarkson, N. Eng. J. Med., 338 (23)
(1.998) 1672-1676.
208 C. Dietz, Y. Madrid, C. Camara and P. Quevauviller, Anal. Chem., 72 (2000) 4178-
4184.
209 Y. Cai, S. Monsalud, R. Jaffe and R.D. Jones, J. Chromatogr. A, 876 (2000) 147-155.
210 E. Millan and J. Pawliszyn, J. Chromatogr.A, 873 (2000) 63-71.
211 X. Yu and J. Pawliszyn, Anal. Chem., 72 (2000) 1788-1792.
212 Z. Mester, G. Horvath, G. Vitany, L. Lelik and P. Fodor, Rapid Commun. Mass
Spectrom., 13 (1999) 350-353.
668
Chapter20
20.1 INTRODUCTION
The remarkable biological activity of insect pheromones has been known for
over a century, since pioneering experiments in the late 19th century by J.H.
Fabr6 [1] and others demonstrated that male moths were attracted to
pheromone-producing females over distances of a kilometer or more. However,
identification of pheromones from insects and other organisms was hindered by
the small quantities that are produced, often in the nanogram or less range per
individual. Research on the first identification of an insect pheromone began in
earnest in the mid-1930s, and culminated 25 years later in the identification of a
few milligrams of (10E,12Z)-hexadecadienol as the sex attractant pheromone of
the silk moth Bombyx mori, from an extract of 500,000 virgin female moths [2].
The pheromone had remarkable biological activity, stimulating a behavioral
response in male moths at femtogram levels. Since that extraordinary effort, the
development of optimized techniques for collection, analysis, and bioassay of
pheromones, coupled with the quantum leaps in chromatographic and spectro-
scopic methods, has greatly simplified the process of pheromone identification.
Several thousand insect pheromones of various types are now known [3-7],
along with a number of pheromones from other organisms. Identifications of
new pheromones have also been aided by biosynthetic parsimony, in that related
organisms tend to produce pheromone components with similar or identical
structures. In optimal cases, it has been possible to identify a complete phero-
mone blend from less than ten individuals. However, this is usually the excep-
tion rather than the rule. In general, the identification of pheromones remains
challenging, particularly if the bioactive compound(s) are new to science,
because of the very small quantities that can be collected.
It is also instructive to consider the differences between extraction of
analytes from samples such as water or tissue, and collection of pheromone sam-
ples. Although pheromones have been collected numerous times by whole-body
extraction or by extraction of specific tissues, this may be less useful than
dynamic collection of pheromones from living organisms, for several reasons.
ComprehensiveAnalytical ChemistryXXXVII
J. Pawliszyn (Ed.)
C 2002 Elsevier Science B.V. All rights reserved 669
First, although some pheromones are stored in glands or reservoirs, others are
biosynthesized and released as they are needed. Consequently, far more phero-
mone may be collected by continuous dynamic headspace sampling for periods of
minutes to days or longer than would be collected by one-time destructive
extraction of the same number of organisms. Second, headspace collections pro-
vide much cleaner and simpler extracts containing only the volatiles produced by
the test organisms, and are easier to analyze than whole-body extracts. Third,
because some pheromones are biosynthesized at the moment of release, extrac-
tion of tissues that contain both pheromones and pheromone precursors may
result in a chemical profile that is markedly different from that of the correct
pheromone blend.
Furthermore, the difference between developing effective sampling methods
for compounds whose structures are unknown, versus sampling compounds
whose structures, physical properties, and chemical properties are known, must
be borne in mind. That is, although some generalizations can be made with
regard to the chemical and physical properties of different types of pheromones
or related attractants, most sampling methods are biased. For example, acti-
vated charcoal or synthetic polymers effectively trap hydrophobic, semivolatile
compounds, but they are ineffective or even useless at trapping small polar
molecules such as ammonia, formic acid, acetic acid, and ethanol, all of which are
known insect semiochemicals.
Problems with identification of trace amounts of compounds can be further
compounded if the structures are unstable. Identifications of several insect
pheromones were hindered by the instability of the bioactive compounds to gas
chromatography and other typical analytical chemistry methods. For example,
the sex attractant pheromones of the tobacco hornworm moth Manduca sexta
((10E,12E,14Z)- and (0E,12E,14E)-hexadeca-10,12,14-trienals; [8]), the tuss-
ock moth Orgyia pseudotsugata ((6Z,8E)-heneicosadien-11-one [9]), and the
stink bug Thyantapallidovirens(methyl (2E,4Z,6Z)-decatrienoate) [101 were all
partially or totally destroyed during attempted analysis by gas chromatography.
This chapter will review the methods that have been developed for the
collection and isolation of insect pheromones and related compounds. Where
appropriate, examples from other taxa will be included. In practice, the appara-
tus and techniques for collection of trace quantities of pheromones, with adjust-
ments for scale, are dictated more by the chemical and physical properties of the
pheromones, and the physical circumstances required for their production or
extraction, than by the particular organism. For example, the same general
sampling techniques have been used to sample pheromones from organisms as
small as single-celled gametes to as large as Asian elephants [11,12]!
670
produced by one organism for communication with another organism of the
same species. Within this broad definition, pheromones are further classified
into different categories, depending on their function. Several recent reviews
and databases provide compilations of known insect pheromones [3-7].
Some of the more common types of pheromones are briefly described below,
along with their key characteristics and selected examples.
The volatile sex attractant pheromones that many insects produce to attract
mates are probably the most widely known type of pheromones [7]. These
pheromones are produced by one sex, and responded to by the opposite sex.
Males rather than females are frequently the attractive sex, for various
biological and ecological reasons that are beyond the scope of this chapter to
explore. Sex attractant pheromones also are used by many other animals, and
even by single-celled organisms such as yeasts. The pheromone properties vary
with the producing organism, and with the medium in which the signal is
transmitted. Thus, terrestrial organisms whose signals are airborne produce sex
attractants that are volatile or semivolatile, which restricts the molecular
weights to about 350 daltons or less, and also limits the number of functional
groups. In contrast, aquatic and marine organisms may produce polar or nonvol-
atile sex pheromones whose properties may be limited more by constraints on
solubility than on volatility and molecular size. However, this is not always the
case; for example, sex pheromones of brown algae are small hydrophobic polyun-
saturated hydrocarbons [13-17].
Many insects have sex and species recognition pheromones that act only at close
range or on contact [4,5]. These pheromones are used to confirm that an
individual is of the correct species and sex before mating attempts progress.
These types of pheromones are also used by social insects such as ants, termites,
and honeybees to distinguish nestmates from non-nestmates, and to distinguish
caste status. The pheromones tend to be simple straight- or branched-chain
hydrocarbons of limited volatility, with specificity usually being conferred by a
particular blend of compounds rather than the presence or absence of a unique
compound.
Marking pheromones are used, for example, by parasitic insects to mark hosts in
which they have laid an egg [18]. The marks deter oviposition by conspecifics,
limiting competition for the resource. The marks also may prevent the same
female from superparasitizing a host. Scent marks are also ubiquitous in higher
671
animals, being used to define territory and to display the dominance status of
individuals.
Alarm pheromones alert conspecifics to danger, and are widely used by insects
that live in groups or colonies, such as ants, bees, and aphids [4,5]. The alarm
pheromone may cause receivers to move away from the signal or drop off the
plant (e.g. aphids), or the pheromone may recruit nestmates to defend the nest,
such as occurs with Africanized honeybees. In some cases, these pheromones
may also have a secondary role as defensive compounds. For example, many true
bugs (Hemiptera) produce copious quantities of volatile defensive secretions,
and in addition to deterring predators, the compounds also cause conspecifics to
move away [19]. These pheromones tend to be small volatile compounds,
allowing rapid dispersion of the signal.
Trail pheromones are used by organisms such as ants to mark trails to food
sources [20]. Because the food sources are usually ephemeral rather than
permanent, ant trail pheromones tend to be semivolatile, and must be continu-
ally reinforced in order for the trail to persist. Trail pheromones are also used by
processionary insects such as tent caterpillars and bagworms, to establish trails
between the protective tent and food sources [21].
Aggregation pheromones, as distinct from sex pheromones that attract only one
sex, are produced by one (or both) sex(es), and attract both sexes. They are used
by insects such as bark beetles to assemble the critical mass of individuals
needed to mass attack and overwhelm a host tree's defenses [22]. Aggregation
pheromones are also used by some beetle species that infest grain and related
stored products [23], and by first instar stink bugs, causing the nymphs to
cluster together [24].
There is a wide variety of other types of pheromones associated with all aspects
of reproduction. For example, primer pheromones induce physiological changes
such as egg maturation and ovulation, or the inhibition of reproduction by
worker castes in bees [25]. Aphrodisiac pheromones cause sexual stimulation
and excitation in males, or render females compliant to mating attempts
[26-28]. Releaser pheromones trigger the release of specific behaviors, such as
gregariousness in locusts [29], or the release of sperm or eggs into the water by
672
aquatic organisms [30]. Other pheromones transferred during mating reduce
female receptivity to further mating attempts, block further sex attractant
pheromone production by females, and stimulate oviposition [31].
Predators and parasites frequently exploit the pheromones of their hosts.
This exploitation may be eavesdropping, using the pheromones as chemical
beacons to locate hosts [32,33]. Alternatively, exploiting organisms may actively
mimic the pheromones of their hosts. For example, bolas spiders lure male
moths towards their sticky threads by producing the female sex attractant
pheromone for that moth species [34]. Furthermore, the spiders appear to be
able to change the blend that they produce to match the flight times and periods
of specific prey species.
An example may provide insight into the variety and importance of pheromones
in the biology and ecology of organisms. Consider an ant nest or a honeybee hive,
consisting of a queen, males, and numerous female workers of different castes,
frequently totalling thousands or even millions of individuals. The structure,
organization, and functioning of these immense and complicated societies is
largely controlled by the exchange of chemical signals between individuals, with
an intricate vocabulary of chemical "words" providing the "language" of
communication that enables the colony to function.
This also illustrates another important point: when planning to sample and
identify a pheromone, a detailed knowledge of the biological function of the
pheromone is a vital prerequisite. Many behaviors, such as mating or host
location, actually consist of a sequential series of behavioral steps. In some cases,
one pheromone may trigger the whole behavioral cascade that culminates in
copulation, that is, activation of a resting individual, long range orientation and
directed movement towards a potential mate, arrestment, assessment and rec-
ognition, sexual stimulation, and finally copulation. For example, males of some
beetle and moth species, when stimulated with a single female pheromone, will
fly upwind, land, and, in a sexual frenzy, attempt to mate with the pheromone
source, even with such unnatural sources as the end of a glass rod or a pencil
treated with pheromone [5,6]. With other organisms, different steps in the
mating behaviors are mediated by different chemical signals, and if the full
repertoire of chemicals is not present, copulation will not occur. For example,
males of many species of ticks are attracted by 2,6-dichlorophenol [35], but in the
absence of other chemical signals, copulation will not occur. Thus, it is crucial
both to know the biology of the study organism thoroughly, and to develop a
bioassay that tests for the specific behavior mediated by a particular pheromone.
An example may be illustrative. When first perceiving one or more components
of the female sex pheromone blend, male moths begin wing-fanning, and this
behavior has been exploited in simple bioassays for identification of moth sex
pheromones. However, this is only the first step in the behavioral cascade that
673
ends in a male moth copulating with a female, and wing-fanning can be triggered
by incomplete or incorrect pheromone blends that will not result in a male moth
flying upwind, or even taking flight. More sophisticated bioassays that test the
whole series of behavioral steps, such as flight tests in wind tunnels, are required
in order to identify the complete pheromone blend [36].
Control and knowledge of the physiological state of the test organisms is also
crucial because many pheromones are produced only in specific contexts. For
example, sex pheromones are produced only by sexually mature individuals that
are in a reproductively receptive state. It is pointless to try to collect pheromones
from immature individuals, or from organisms that are otherwise in non-
reproductive states, such as insects that are in seasonally induced reproductive
diapause. Production of sex (and other) pheromones also may occur only during
certain specific periods of the day, with production usually being linked to daily
light-dark cycles. Consequently, when collecting pheromones from live organ-
isms, light, temperature, and other physical conditions must be correct in order
for the desired pheromone to be produced. Furthermore, female insects usually
cease production of sex attractant pheromones after mating, either perma-
nently, or fora refractory period while some eggs are laid. Thus, for collection of
sex pheromones, it is prudent to use insects whose reproductive history is
known, and if possible, to use virgin insects. Analogous considerations apply to
other types of pheromones. In short, insects must be provided with a facsimile of
the conditions under which they would normally perform a specific behavior
associated with a pheromone in order for that pheromone to be produced.
Although the volatile sex pheromones are most well known, it must be
remembered that the physical and chemical properties of a pheromone are
suited to its function, the medium through which it is transmitted, and the
target receptor organs. Thus, sex attractant pheromones for terrestrial organ-
isms, whose function is attraction of mates from a distance, tend to be relatively
small molecules. Conversely, pheromones that act on contact, such as sex and
species recognition pheromones, tend to be larger and of limited volatility.
Consequently, sampling methods must take into account the likely properties of
the particular pheromones being sampled.
674
TABLE 20.1
Advantages and limitations of various methods of sampling and collecting pheromones
Method Advantages Limitations
Whole body Extracts most material present; can Active compounds are buried in a complex
extraction provide large samples for background of extraneous material; extensive
fractionation and bioassay cleanup and fractionation may be required
Extract may not be representative of profile of
compounds actually released by living organ-
ism
Organisms are destroyed; can only sample
them once
Enzyme action during tissue maceration may
alter profile of extractables
Amounts and types of compounds extracted bi-
ased by choice of solvent
Need to remove relatively large amounts of sol-
vent without losing or degrading bioactive com-
pound(s)
Solvent purity is critical
Steam Extracts most material present; can Sensitive compounds destroyed during extrac-
distillation provide large samples for fraction- tion
ation and bioassay Extract may not be representative of profile of
Separates volatiles from nonvolatile compounds actually released by living organ-
materials; extracts can go straight ism
on GC Need to extract sample from aqueous distillate
Organisms are destroyed; can only sample
them once
continued
675
TABLE 20.1 (continuation)
Solid phase Collect continuously or repeatedly Sampling method only; cannot be used for bulk
micro- from same group of living organisms collections for fractionation or bioassays.
extraction for periods of minutes-weeks Each collection provides only a single sample;
(SPME) Sample is representative of the sample cannot be split easily, eg., for analysis
blend of compounds actually released and bioassays
by the test organism Fibers are biased in their retention of different
Different fibers optimized for reten- chemical classes and also retain only a small
tion of different classes of analytes fraction of very volatile compounds
Simple and inexpensive; fibers can Cannot be used with thermally sensitive com-
be reused many times pounds
Solvent-free: volatile components
not masked by solvent
Essentially quantitative desorbtion
of sample
Loaded fibers can be shipped between
labs
Cryogenic Traps all components including Traps water; need to isolate active compounds
trapping CO2 and other permanent gases from water
Equipment may be cumbersome Safety concerns due to plugging of apparatus
with ice; pressure buildup on warming
676
20.4.1 Whole body and pheromone gland extracts
677
solvent include the ease of purification and the selective solvating power for the
compounds of interest, which are normally of limited solubility or insoluble in
water. Thus, solvents such as pentane, hexane, methylene chloride, and diethyl
ether have been used extensively in extraction of pheromone-containing tissues.
It is critically important to use appropriate sized apparatus and glassware for
working with the small quantities of pheromone extracts that are normally
available. The fabrication and use of microscale glassware and other apparatus
is described in Millar and Haynes [38,53]. It is pointless to extract nanogram
quantities of volatile pheromones using tens or hundreds of milliliters of solvent
in the glassware normally used in organic chemistry laboratories because of
problems with contamination, loss of the bioactive material during concentra-
tion, or adsorbtion on the surfaces of glassware. The number of handling steps
also must be minimized to reduce losses and contamination. To minimize losses
during concentration, the bulk of the solvent can be removed with a fractional
distillation column [50]. For smaller quantities, solvent is removed by passive
evaporation from an open vial, or is blown off with a gentle stream of nitrogen.
Often a single cleanup step, or no cleanup step at all, is used before going directly
to high resolution chromatographic methods.
678
als are limited, or when following changes in chemical profiles during develop-
ment, for example, during sexual maturation, or before and after mating. Even
when the number of individuals is not limiting, it is usually more economical to
sample the same individuals several times than to use them once in a single sol-
vent extraction.
Several points should be considered when choosing methods for collection of
volatile pheromones. First, the amounts of material obtained are small, being a
function of their rate of release and volatility, and the flow rate in the aeration
setup. Normally, nanograms to a milligram or so of material are obtained per
aeration cycle. If only small quantities of pheromone are required, for example
for quantification of a known chemical, then straightforward sampling methods
using simple aeration chambers are usually satisfactory. However, if larger
amounts of material are required for fractionations and bioassays, or NMR
analysis, then repeated collections may be required.
The characteristics of the bioactive pheromone blend must also be consid-
ered. If the blend consists of nonpolar to intermediate polarity components with
approximately similar volatilities, sampling or preparative-scale collections can
be done using well established methods (see below). However, if the blend
consists of components with widely differing properties, then the collection of
samples that accurately reflect the blend being released becomes more difficult.
For example, adsorbents such as Porapak Q, Tenax, or activated charcoal are
satisfactory for collecting low to medium polarity chemicals of intermediate size
and volatility, but they are poor to useless for trapping low molecular weight
and/or polar compounds such as ethanol, ammonia, and short-chain amines.
This problem is discussed in more detail below. The fact that the blend of
compounds trapped by most sampling and collection methods is quantitatively
and qualitatively biased has been largely ignored. In cases where pheromone
blends have proven to consist of hydrophobic molecules with molecular weights
in the 150-350 dalton range, this bias is not important. However, in other cases,
such as the attraction of mosquitoes to hosts, in which the attractant blend may
consist of components as diverse as carbon dioxide, lactic acid, and more hydro-
phobic alcohols and phenols, this sampling bias may be critically important. It is
undoubtedly responsible in part for slow progress in other areas of research,
such as the attraction of insects to host plants, in which complex blends of
chemicals with widely differing properties are involved.
Furthermore, volatile semiochemicals may be hidden under solvent peaks
during GC analyses of extracts prepared by elution of trapping media. It may be
possible to circumvent this problem by solventless thermal desorbtion of the
trapped volatiles, but purge-and-trap type thermal desorbtion is also not with-
out problems, including degradation of fragile compounds and production of
artifacts from the trapping media, as discussed below.
In the discussion that follows, where appropriate, I will draw distinctions
between methods that are appropriate only for sampling, and methods that are
appropriate for both sampling and preparative scale collection of pheromones.
679
20.4.2.1 Apparatusfor continuous collections
Apparatus for collection of volatiles in the headspace above live insects or other
organisms should be made from glass, metal, Teflon, polycarbonate, or Plexiglas
[38,50]. Other plastics, rubber, glues and adhesives, wood, and other building
materials should not be used because they bleed and/or retain volatiles. Parts
must be thoroughly cleaned and where possible, oven-dried at -125°C before
use. Connections are made with ball and socket ground-glass joints that allow
some motion. Tubes are connected with brass Swagelock unions with Teflon
ferrules. Connections through solid surfaces such as screw-cap lids are made by
drilling holes to accommodate Swagelock bulk-head unions. For small systems,
Teflon tubing connected to syringe needles with Teflon heat-shrink tubing is
used to connect the apparatus to air or vacuum lines. Rubber or plastic tubing
should only be used downstream from the collector. Apparatus of appropriate
sizes usually can be built from commonly available jars or containers. For large
numbers of aerations on a continuing basis, custom-built apparatus may be
worthwhile.
During aerations, all air entering the aeration system is precleaned by
passage through a charcoal filter, readily constructed using the granulated
charcoal used in aquariums. The airstream is usually humidified to prevent
desiccation of the organisms by bubbling the air through a water-filled gas-
washing bottle or passing it through a tube loosely packed with water-soaked
glass wool, placed in front of the charcoal scrubber. Nitrogen can be used if
anaerobic conditions are required, bearing in mind that volatiles produced
under aerobic and anaerobic conditions may differ [54]. Appropriate air sources
include purified compressed air (e.g., so-called "medical air"), or ambient air
supplied by an oil-free diaphragm pump. Alternatively, air can be pulled through
the system by vacuum; in "push-pull" systems, both pressure and vacuum are
used [47,55]. The aeration system must be neither overpressurized nor held
under partial vacuum, particularly when collecting volatiles from live
organisms. Either case may occur due to flow constrictions from the charcoal air
scrubber on the inlet side, or the volatiles collection trap on the outlet end.
More recently, "open" systems for the collection of volatiles from larger
organisms such as intact plants have found widespread use [47,56]. With this
method, a large volume of clean air is passed over a plant in an open-ended sleeve
or tube, with a small portion of the airstream being sampled.
Aeration chambers may be very small (e.g., multiple simultaneous aerations
in small vials [57]) to quite large. Small chambers are usually glass, whereas
intermediate to large scale aeration chambers are fabricated from glass, poly-
carbonate, metal, or Plexiglass. Apparatus such as wide-mouthed containers
with ground-glass joints (e.g., vacuum desiccators, wide-mouthed Erlenmeyer
flasks, etc.) or Teflon-lined screw-on lids is readily adapted for aerations. Cus-
tom-made cylindrical flow-through chambers which open in the middle or at one
end, with a securely clamped, O-ring sealed middle joint, and ground-glass ball
and socket joints at either end for connections are also excellent. Insects often
680
like to have something to climb or perch on, so clean metal screen or analogous
substrates can be inserted.
Larger chambers can be constructed from steam-cleaned, solvent-washed 55
gall drums [58,59], modifying the pouring hole to accept a charcoal filter. The
chamber is loaded through a large hole cut in the bottom, which is then closed
with a bolted-on metal plate and a Teflon gasket, with an outlet for the volatiles
collector.
More recently, sophisticated setups for the aeration of intact plants have
been developed, incorporating techniques for purification of large volumes of
inlet air, and computer controlled sequential sampling and monitoring of envi-
ronmental conditions [47,55,60]. These systems are commercially available in
custom designs (Analytical Research Systems, Inc., Gainesville FL; website
www.ars-fla.com). Related systems also have been described [61,62].
It is also possible to sample from living organisms in the field. For smaller
scale sampling, polyvinyl fluoride [63] or polyacetate cooking bags [64,65] were
used to bag flower heads for volatiles collection, polythene terephthalate bags
were used to bag fungal fruiting bodies [66], and Tedlar® bags were used in
sampling pine needle volatiles [67]. Larger apparatus can be constructed from
aluminum [58] or polyvinyl fluoride sheets. For example, we sampled volatiles
from trunks of living Eucalyptus trees by "bagging" sections with Tedlar®
fabric (Millar et al., unpublished data). Because these large bags cannot be
completely sealed, a push-pull system is advisable, with clean air pumped in, and
only a portion of the output being sampled so that the system remains under a
slight positive pressure.
Field sampling of odor sources can also be carried out in the open air, for
example, with flower odors [68]. Battery-operated personal air sampling pumps
with flows of several 1/min, such as those used for air sampling in industrial
settings, are ideal for these applications. Field sampling can be important
because odors released by organisms in their natural habitats may differ from
those produced under less natural conditions. However, background volatiles
may be problematic.
During the identification of novel semiochemicals, significant quantities are
needed for bioassay, fractionation, and identification of the active principles,
necessitating collection of relatively large samples from groups of insects. For
these bulk collections, headspace volatiles are usually collected on a bed of
granulated adsorbent (-40-80 mesh, to compromise between high surface area
and resistance to air flow), held in place with glass wool plugs in a glass or metal
tube. Porapak Q (and its refined version, Super Q; Waters, Supelco), Tenax GC
(refined version, Tenax TA; Chrompack, Alltech), and activated charcoal are the
most commonly used adsorbents. Porapak and Tenax are hydrophobic polymeric
resins with a high affinity for lipophyllic organic compounds. Tenax is more
thermally stable than Porapak, but this may not be an issue if traps are not
thermally desorbed. For most applications, either adsorbent can be used inter-
changeably [69]. Porapak and Tenax adsorbents need conditioning before use,
681
typically by heating under a slow flow of inert gas overnight (Porapak, 200°C,
Tenax 270-340°C for 2-24 h; [70-72]), followed by Soxhlet extraction with
pentane, methylene chloride, or ether [50,73,74]. Once conditioned, both adsorb-
ents require storage in sealed containers protected from light. Artefacts from
both types of adsorbents have been tabulated (Porapak [72,75,76]; Tenax [71,77,
78]). To control for artefacts, a system blank (i.e., aeration of an empty setup)
should always be run.
Volatiles are recovered from either adsorbent by elution (5-10 ml/g
adsorbent) or Soxhlet extraction with low-boiling solvents such as pentane or
methylene chloride, or for polar compounds, carbon disulfide. Small extracts are
concentrated by passive evaporation from open vials, or blown down under a
stream of clean N2 , whereas larger volumes of solvent are removed by fractional
distillation. Eluted sample tubes can be reconditioned by further elution with
clean solvent, while Soxhlet-extracted adsorbents are usually ready for reuse
after a second extraction with solvent. Residual solvent vapor is removed from
the cleaned resin under a stream of N2 .
The volume of adsorbent required for a particular application depends on the
polarity and volatility of analytes, flow rate, and volume of air to be sampled.
Because collections usually are made under ambient conditions, temperature is
not an issue. Highly volatile and/or more hydrophyllic compounds are not
retained well and will break through traps rapidly. Breakthrough volumes per
gram of adsorbent for analytes of different classes have been tabulated [79,80]
and are available from adsorbent suppliers (e.g. Scientific Instrument Services).
Practically, the easiest way to check for breakthrough is by putting two traps in
series, checking both for the analytes of interest. For most applications with
semivolatile pheromones, volatiles from tens to hundreds of liters of air can be
trapped on a relatively small bed of adsorbent (a few milligrams to a gram or so,
-5-50 mm in length), remembering that longer beds offer greater resistance to
air flow. As a general rule, the quantity of adsorbent is kept small to minimize
artifacts and the quantity of solvent required to elute the analytes.
Activated coconut charcoal is also an excellent adsorbent, with several
desirable characteristics. It is highly retentive with a high capacity, which
minimizes the quantity of adsorbent needed. It is easily cleaned in bulk by
heating under a clean N2 flow overnight at 250C, after which further solvent
cleanup is rarely necessary. Other researchers have recommended conditioning
with exhaustive solvent stripping with, e.g., hot EtOH, CH 2C12 , and CS 2 [81], or
CH 2 Cl2 and pentane [82]. Charcoal traps can be eluted with pentane, methylene
chloride, or CS2 (-5-20 volumes). Unlike the polymeric resins, charcoal is cheap,
so it can be discarded after use to eliminate the chance of cross-contamination.
Alternatively, small reusable traps have been fabricated by permanently sealing
a few milligrams of charcoal between stainless steel screens in glass tubes
[82-85]. These are commercially available (Brechbfihler AG, Urdorf Switzer-
land; http://www.brechbuehler.ch/).
682
Until the advent of SPME, trapping of insect pheromones on Porapak,
Tenax, or charcoal was the method of choice for both sampling and bulk
collection. SPME is now replacing these methods to some extent in sampling
applications, but SPME is not suitable for bulk collections over extended
periods, which will continue to rely on methods based on these adsorbents.
683
more efficiently than hydrocarbons with similar GC retention indices. Con-
versely, coatings such as CarboxenT"-polydimethylsiloxane have been developed
specifically for sampling of very volatile compounds. Thus, a prudent choice of
fiber coating, and calibration with appropriate standards covering the entire
range of interest, is crucial for quantitative work.
Third, adsorbtion is inversely dependent on temperature. However, when
sampling less volatile substrates, it may be necessary to heat the sample to
increase the headspace concentrations of analytes to measurable levels. In
extreme cases, such as sampling long-chain hydrocarbons from insect cuticles,
samples have been heated to as high as 170°C during sampling (see below).
Finally, adsorbtion appears to depend slightly on concentration, with lower
concentrations of analytes being adsorbed better than higher concentrations
[87].
SPME provides only enough material for GC, GC-MS, possibly GC-FTIR
[92], and HPLC with certain types of compounds and detectors. It is an excep-
tional method for analysis of known compounds, but when attempting to
identify novel compounds in trace amounts, the sensitivity may not be good
enough even for some GC-based methods. SPME's primary value is as a sampl-
ing tool for compounds of known structure, for compounds with structures
similar to known compounds, or for compounds with structures simple enough
to be determined solely by GC-linked methods.
Within these limitations, SPME has a number of advantages. First, it is
rapid and simple, and virtually any odor source can be sampled under laboratory
conditions, or in the field with devices modified for field collections [93]. Loaded
devices can even be sealed up and shipped between laboratories. Second, the
dynamics of release of compounds from an odor source can be followed readily by
sequential sampling of the odor source continuously or at any desired interval
with a pair of devices, for any desired duration, from minutes to months.
However, for sampling in a closed container, compounds that are not produced
continuously may become depleted [87]. Third, one of the key advantages of
SPME is that samples are solvent-free, so that volatile analytes are not hidden
under solvent peaks in GC analyses. The problem of loss of volatile samples
during concentration of extracts is also eliminated. Fourth, SPME, with the
appropriate choice of fiber coating and coating thickness, is amenable to samp-
ling analytes spanning the entire range of volatilities and polarities, including
gases such as ammonia and volatile amines [94,95], low-boiling compounds such
as short-chain alcohols and hydrocarbons, all the way through to very high-
boiling compounds such as hydrocarbons with chain lengths of C30 and higher.
The former group of compounds in particular may be difficult to sample
accurately by other methods.
Because the desorbtion of analytes from the SPME fiber takes an appreciable
period of time in the GC injector, injection techniques must be modified to retain
good chromatographic resolution and symmetrical peaks, particularly as there is
no solvent effect to refocus the analytes at the front of the GC column.
684
Refocusing of the desorbed compounds is accomplished by appropriate choice of
temperature (low initial temperature, or even cryogenic focusing at the column
head) and stationary phase (thicker film phases for volatile compounds) [89,96],
and by use of narrow bore, small-volume injector liners (0.75 mm i.d.; Supelco)
developed especially for use with SPME.
Application of SPME for qualitative analysis of volatiles is straightforward.
However, the determination of relative and absolute amounts of each compound
in blends is not trivial, for several reasons. First, the adsorbtion of compounds on
SPME fibers is proportional to the analytes' molecular weights, and to a lesser
extent, their polarities. Thus, for quantitative studies, a proportionality
constant K that relates the mass adsorbed on the fiber to the concentration in
the headspace must be calculated for each analyte [87]. Second, when sampling
from closed systems, at equilibrium, the amount of a given analyte on the SPME
fiber is proportional to the analyte concentration in the headspace. However,
higher molecular weight compounds take longer to equilibrate on the fiber than
smaller molecules. Consequently, sampling times should be long enough to allow
equilibration of all components of interest. Third, GC detector response varies
with molecular structure and molecular formula, and relative response factors
must be calculated for each compound of interest.
To address these problems with quantitation, Bartelt [87] developed a
regression model for closed systems that found a strong correlation between log
K and the GC Kovats retention index, over a number of different chemical
classes. Using this model, the K value can be accurately predicted for compounds
in 13 chemical classes, using only the Kovats retention index (which is trivial to
determine), the temperature at which SPME sampling was conducted, and the
functional group of the analyte. The K values, in combination with the GC
detector response factors, can then be used to calculate the absolute headspace
concentrations of all compounds in a complex blend from the SPME-GC peak
areas. K values were found to vary slightly with analyte concentrations, but this
effect was minor in comparison to the major changes in K values with increasing
molecular size. For example, the SPME detection limits using a PDMS fiber
were calculated to be about 5.6 ng/ml of air for acetaldehyde, versus 0.00036
ng/ml of air for ethyl decanoate.
This landmark work was extended to dynamic systems in two further
studies. In the first [97], a general model was developed that allowed the
determination of the absolute concentrations of analytes in a uniform airstream.
A key feature of the model is that it does not require the analytes to have reached
equilibrium with the fiber, which is of crucial practical importance for sampling
relatively large molecules that take a long time to equilibrate. A further conse-
quence that emerged from the study was the fact that an SPME fiber inserted
into a 1.5 mm to 4.5 mm outlet port, with a slow air flow rate (2 ml/min) provides
essentially quantitative collection of analytes with GC retention indices > 1500.
The second study examined the effects of airflow and SPME sampling port
diameter in more detail [98]. For a given release rate of an analyte (mass/unit
685
time), trapping efficiency was inversely proportional to air flow rate, with low
flow rates providing the highest trapping efficiencies. Surprisingly, outlet port
diameters of 1.5 and 4.5 mm provided approximately equal trapping efficiencies.
These three studies have provided a foundation for using SPME to make
quantitative measurements of complex mixtures of analytes of widely differing
polarities and volatilities, from both closed and dynamic biological systems. The
importance of these studies in the continuing development of SPME methodol-
ogy cannot be underestimated. To further appreciate the differences between
SPME and trapping on polymers such as Porapak, or purge and trap analysis,
several studies have compared the techniques for collecting volatiles from
various biological systems [99-103]. Each has its advantages and limitations,
and no single method can be used in all situations.
686
rately reconstruct release rates from synthetic blends of the identified
compounds.
SPME was used to determine that the rhinoceros beetles Scapanes australis
and Strategus aloeus produce very volatile pheromones with remarkably simple
structures. The pheromone of the former species consisted of a 2:1 blend of the R
and S enantiomers of 2-butanol in combination with 3-hydroxy-2-butanone and
2,3-butanediol, whereas the latter species produced a blend of 2-butanone,
3-pentanone, and sec-butyl acetate [110]. These pheromone components vary
widely in volatility and polarity, and might have been difficult to analyze by
other methods.
Palm weevils, such as the American palm weevil Rhynchophoruspalmarum,
are attracted by a synergistic combination of pheromones and volatiles produced
from decaying host plant material. SPME with PDMS fibers was used in combi-
nation with collection of volatiles on an adsorbent polymer (Supelpak-2) to
identify more than 100 compounds from the headspace above decaying samples
of sugar cane, coconut palm, oil palm, and the tree Jacaratiadigitata [111]. The
combination of SPME with adsorbent trapping was critically important because
the most volatile compounds, such as acetaldehyde and short-chain alcohols and
esters, were minimally retained on the adsorbent traps; in fact, all compounds
with retention times less than 12 min broke through the adsorbent traps to a
greater or lesser extent.
In all of the studies above, the critical role of SPME cannot be overempha-
sized. Traditional methods of trapping the analytes on adsorbents such as
charcoal, Porapak, or Tenax would not have worked because the highly volatile
semiochemicals would not have been retained, and the concentrations of most of
the volatiles would have been too low for direct analysis of headspace samples.
Furthermore, the more volatile constituents would have been hidden under the
solvent peaks during GC analysis.
SPME is also being used in sampling and identification of insect pheromones
and related attractants of moderate to relatively high molecular weight. In one
of the first reported studies of the application of SPME to insect chemical
ecology, SPME was used to quantify the diurnal cycleof production of the
pheromone components 2-methyl-4-heptanol, 2-methyl-4-octanol, 5-nonanol,
3-hydroxy-4-methyl-5-nonanone, and 4-methyl-5-nonanol from sugarcane weev-
ils [112]. With continuous sampling at 30 or 60 min intervals, SPME analysis
provided an accurate picture of the time course of pheromone production [112].
As an example of the critical importance of determining the correct ratio of
compounds in attractant blends, Zhang et al. [113] reexamined volatiles released
from apples, using SPME coupled with electroantennographic detection. The
analyses resulted in the development of an improved blend of attractants for
apple maggot fly consisting of five short-chain esters of 8-10 carbons.
SPME in combination with coupled GC-electroantennographic detection
also played a critical role in the development of improved attractants for house-
flies [103]. Volatiles collected from fresh pig manure that elicited strong
687
responses from housefly antennae included butanoic and 3-methylbutanoic
acids, dimethyldisulfide, dimethyltrisulfide, dimethyltetrasulfide, phenol,
phenylethanol, indole, and skatole. An optimized blend of skatole, butanoic acid,
and dimethyltrisulfide was approximately as attractive as pig manure. The same
research group compared SPME to more traditional methods of headspace
analysis (collection on Tenax resin, purge and trap analysis) when analyzing
buffalo gourd root powder for attractants for corn rootworms [114]. Results from
the three sampling methods were similar, and a number of volatiles were
identified, of which (E)-3-octen-2-one and (E,E)-3,5-octadien-2-one were most
attractive in field tests.
A number of insect species utilize the fruiting bodies of fungi as breeding or
feeding substrates, and the insects locate their hosts by olfaction. SPME in
combination with collection of volatiles on Porapak was used to analyze the
volatiles from intact and chopped fruiting bodies [66]. Compounds identified
included ethanol, acetic acid, mono-and sesquiterpenes, and intermediate chain
length aldehydes, alcohols, and ketones. As in the studies mentioned above,
SPME was crucial to the detection of ethanol and acetic acid. In related studies,
SPME is finding increased use in botany, for example in sampling volatile
chemicals from intact and damaged plants [89,115,116]. SPME was also used to
sample the patterns of emission of odors from whole plants, buds, and flowers of
the plant Boronia megastigma while manipulating the light regimes under
which the plants were held. The identified volatiles included mono- and ses-
quiterpenes, esters, alcohols, and straight-chain hydrocarbons [117].
In an example of SPME used in the extraction of pheromones from aqueous
media, SPME was used to sample the attractant pheromones produced by the
sessile female gametes of the brown algae Laminaria digitata to attract the
motile male gametes [15]. These compounds, consisting of C11 hydrocarbons and
epoxides, were extracted by immersion of the PDMS fiber in 20 ml of medium
from a culture that had released the female gametes. Previous studies of this
type had used closed-loop-stripping analyses (see below), but the SPME method
was much simpler and faster.
A series of papers has reported the use of SPME in the identification of
pheromones from leafminer moths [118-121]. These moths are very small (-2-3
mm long), and SPME had distinct advantages over dissection and extraction of
pheromone glands. In particular, SPME was used to collect pheromone for
extended periods from undisturbed calling female moths, and samples could be
taken from individual females multiple times. The SPME-collected volatiles
from a single calling female equaled the amount obtained from an extract of fifty
pheromone glands [119]. This finding is important because lepidopteran phero-
mones are frequently produced in such small amounts that even identification of
the major components can be difficult. Thus, any technique that significantly
increases the amount of pheromone collected is of tremendous value.
SPME was recently used to study the dynamics of pheromone production by
male southern green stinkbugs [122]. Sexually mature males produced
688
nerolidol, bisabolene, and two bisabolene epoxides, but the amounts produced by
individuals were highly variable. Repeated sampling of the same individuals
revealed that the ratio of compounds produced by individuals from the same
population was variable, but that the ratio produced by a given individual was
stable over time. This type of study would have been more difficult using
traditional methods of pheromone collection on adsorbents.
SPME has also been used to analyze larger molecules with limited volatility
by heating the relevant tissues in a sealed vial to temperatures as high as 170°C,
and sampling the headspace by SPME. This method was used to sample long-
chain fatty acids in sternal glands from paper wasps [123], and hydrocarbons
from cuticular fragments from the wasps Vespa crabro, V. orientalis, and Polistes
dominulus [124].
Finally, SPME is finding increasing use in the sampling of larger molecules
of limited volatility, using an adaptation termed "contact SPME", in which the
SPME fiber is wiped over an insect's cuticle or over an extruded gland. As with
headspace sampling, the method is nondestructive so that the same organism(s)
can be sampled repeatedly, and specific areas of the cuticle or specific glands can
be sampled without fear of contamination from other structures or tissues, as
frequently happens with dissected glands. In one of the first reports of contact
SPME, Frerot et al. [125] reanalyzed the pheromone of the moth Sesamia
nonagriodes,collecting pheromone by wiping the SPME fiber over the extruded
gland. More pheromone was obtained by SPME than by gland washes, the SPME
extracts were cleaner, and the ratio of pheromone components was slightly
different than that obtained by gland washes. In another example, contact
SPME was used to confirm the sternal gland as the site of production of a trail
pheromone, (Z)-3-dodecen-1-ol, in the termite Macrotermesannandalei,by rub-
bing the SPME fiber on the exposed gland [126]. However, because each termite
produced only about 1 ng of the pheromone, the compound was initially identi-
fied from extraction and fractionation of thousands of termites and excised
glands.
Contact SPME has also been used in the analysis of the large cuticular
hydrocarbons of insects that are used for recognition of species, sex, and nest-
mate status. For example, contact SPME was used to determine the hydrocar-
bon profiles of several Reticulitermes and Coptotermes species [127]. In a related
example, SPME proved crucial in studying cuticular pheromones associated
with dominance status [128,129] and fertility [130] in the ants Dinoponera
quadriceps and Harpegnathossaltatorrespectively. In both cases, the ability to
sample the same individuals repeatedly for up to four months was crucial in
following the increase in pheromone with the development of dominant repro-
ductive status within the colony.
Sledge et al. [131] compared solvent extraction with two methods of SPME
extraction in the analysis of Dufours gland secretions of a stenogastrine wasp,
Parischnogastersp. The first SPME method consisted of headspace extraction of
glands held in a sealed vial at 170°C for 10 min, whereas the second consisted of
689
simply puncturing the gland and adsorbing the contents on the SPME fiber. The
results from all three methods were reported to be similar. The same group also
used contact SPME to study mechanisms used by a social parasite, the wasp
Polistes sucifer, to usurp host nests [132,133]. This wasp has no worker caste,
and must take over host nests in order to produce offspring. Repeated sampling
of the cuticular hydrocarbons of the parasite and its hosts by contact SPME
revealed that the parasite rapidly acquires the hydrocarbon profile of its hosts,
and is thus accepted as a legitimate member of the colony.
From the examples above, it is apparent that the SPME technique has
proven enormously useful in chemical ecology, and it is rapidly becoming the
method of choice for nondestructive and repetitive sampling of insect phero-
mones and related compounds. However, the reader is again cautioned that, as
with most other methods of sampling volatiles, the profile of adsorbed volatiles
does not represent the true proportions of the compounds in the headspace, and
calibration is crucial for accurately determining both the relative and absolute
amounts of compounds present.
690
graphic performance. Furthermore, in the latter method, glass fragments from
the capillaries can adsorb analytes. Nevertheless, the latter method has been
used to analyze dissected glands, fragments of insect cuticle, or other body parts
for pheromones of a number of ant species [138].
Standard purge and trap techniques, in which volatiles are collected or ther-
mally desorbed onto an adsorbent cartridge, followed by purging the cartridge
with dry inert gas to remove water and then thermal desorbtion directly into the
GC (short-path thermal desorbtion), have found only limited use in chemical
ecology, despite turn-key systems being available from several suppliers. Purge
and trap analysis has been used, for example, to analyze insect attractants from
ripe or fermenting fruit [139-142]. Purge and trap methods suffer some of the
same limitations as the aeration methods described above, such as the fact that
the profile of chemicals obtained may be biased by the lack of retention of some
classes of compounds. Furthermore, some of the trapping materials, such as sil-
ica gel, may catalyze reactions of the analytes during the rapid heating of the
desorbtion cycle. Artefacts also may be produced by thermal degradation of the
adsorbents. Several reports have addressed these problems [43,70,143,144].
691
20.4.4.5 Cryogenic trapping
Insect semiochemicals have been collected occasionally by cryogenic trapping,
for example, using liquid nitrogen cooled traps to condense air containing
entrained volatiles [58]. To recover the volatiles, the trap is warmed slowly to
distill off the condensed air, leaving a concentrate. Unlike most other methods,
cryogenic trapping can collect compounds with widely differing volatilities and
polarities, including gases like ethylene and CO 2 [160,161]. However, it has a
number of drawbacks which preclude its general use, including the large volume
of water that is also trapped, from which analytes of interest have to be extracted
or otherwise separated. Safety may also be a concern because formation of ice
plugs can create explosion hazards as the condensed air vaporizes.
Since the identification of the first insect pheromone more than 40 years ago, a
series of specialized techniques have been developed to streamline the identifica-
tion process. Whole body extractions and gland extractions in which the test
organisms were sacrificed have largely given way to dynamic collection methods
in which the volatiles released by living organisms are collected for extended
periods of time. Such dynamic collection methods provide a more accurate
representation of the pheromone blends produced by organisms than tissue
extractions, and allow repeated sampling from the same individuals. For small-
scale collections, analytes may be trapped on short beds of Porapak Q, Tenax, or
activated charcoal. Alternatively, analytes may be sampled by SPME, which has
increased flexibility and ease of use in comparison to trapping on adsorbents.
Furthermore, because SPME requires no solvent for desorbtion, volatile
analytes are not obscured by solvent peaks in GC analyses. For large scale
collection of analytes for bioassay or for spectroscopic analysis, large numbers of
organisms can be placed in appropriate sized aeration chambers supplied with
food and water, with collections continued indefinitely, changing the collectors
periodically. Overall, it must be remembered that each of the methods described
above has limitations with regard to the types of analytes that can be collected,
the efficiency of collection of different classes of analytes, and the quantities that
can be collected. No single method is appropriate for all sampling and collection
scenarios.
REFERENCES
1 J.H. Fabr6, Bilder aus der Insektenwelt. Verlag Kosmos, Stuttgart, 1914.
2 E. Hecker and A. Butenandt, in: H.E. Hummel and T.A. Miller (Eds.), Techniques in
Pheromone Research, Springer-Verlag, New York, 1984.
3 M.S. Mayer and J.R. McLaughlin, Handbook of Insect Pheromones and Sex
Attractants. CRC Press, Boca Raton, FL, 1991.
4 W. Francke and S. Schulz, in: K. Mori (Ed.), Comprehensive Natural Products
Chemistry, Vol. 8. Elsevier, Amsterdam and New York, 1999.
692
5 E.D. Morgan and N. Bhushan Mandava (Eds.), Handbook of Natural Pesticides. Vol.
4. Pheromones, PartsA and B. CRC Press, Boca Raton, FL, 1988.
6 J. Hardie and A.K. Minks, Pheromones of Non-lepidopteran Insects Associated with
AgriculturalPlants. CABI Publishing, Wallingford, UK, 1999.
7 The Pherolist, http://www.nysaes.cornell.edu/pheronet/
8 J.H. Tumlinson, M.M. Brennan, R.E. Doolittle, E.R. Mitchell, A. Brabham, B.E.
Mazomenos, A.H. Baumhover and D.M. Jackson, Arch. Insect Biochem. Physiol., 10
(1989) 255.
9 G. Gries, K.N. Slessor, R. Gries, G. Khaskin, P.D.C. Wimalaratne, T.G. Gray, G.G.
Grant, A.S. Tracey and M.J. Hulme, J. Chem. Ecol., 23 (1997) 19.
10 J.G. Millar, Tetrahedron Lett., 38 (1997) 7971.
11 L.E.L. Rasmussen, T.D. Lee, W.L. Roelofs, A.J. Zhang and G.D. Davies, Nature, 379
(1996) 684.
12 L.E.L. Rasmussen, T.D. Lee, A.J. Zhang, W.L. Roelofs and G.D. Davies, Chem.
Senses, 22 (1997) 417.
13 I. Maier, D. G. Muller, G. Gassmann, W. Boland, F.-J. Marner and L. Jaenicke,
Naturwissenchaften, 71 (1984) 48.
14 I. Maier and D.G. Muller, Biol. Bull., 170 (1986) 145.
15 I. Maier, G. Pohnert, S. Pantke-Bocker and W. Boland, Naturwissenschaften, 83
(1996) 378.
16 D.G. Muller, G. Gassmann and K. Luning, Nature, 279 (1979) 430.
17 L. Muller, T. Gorecki and J. Pawliszyn, Fresenius'J. Anal. Chem., 364 (1999) 610.
18 J. Hurter, E.F. Boiler, E. Stadler, B. Blattmann, H.R. Buser, N.U. Bosshard, L.
Damm, M.W. Kozlowski and R. Schoni, Experientia, 43 (1987) 157.
19 W.S. Leal and T. Kadosawa, Biosci. Biotech. Biochem., 56 (1992) 1004.
20 E.O. Wilson and B. Holldobler, The Ants. Belknap Press of Harvard University Press,
Cambridge, MA, 1990.
21 T.D. Fitzgerald and F.X. Webster, Can. J. Zool., 71 (1993) 1511.
22 M.C. Birch, in: W.J. Bell and R.T. Card6 (Eds.), Chemical Ecology of Insects. Sinauer,
Sunderland, MA, 1984.
23 R. Plarre and D.C. Vandervel, in: J. Hardie and A.K. Minks (Eds.), Pheromones of
Non-lepidopteran Insects Associated with Agricultural Plants. CABI Publishing,
Wallingford, UK, 1999.
24 M. Borges and J.R. Aldrich, Experientia, 48 (1992) 893.
25 J. Pettis, T. Pankiw and E. Plettner, in: J. Hardie and A.K. Minks (Eds.), Pheromones
of Non-lepidopteran Insects Associated with AgriculturalPlants. CABI Publishing,
Wallingford, UK, 1999.
26 L.M. Roth and G.P. Dateo, J. Insect Physiol., 12 (1966) 255.
27 L. Sreng, J. Chem. Ecol., 16 (1990) 2899.
28 Q. Wang and J.G. Millar, Ann. Entomol. Soc. Am., 93 (2000) 972.
29 A. Hassanali and B. Torto, in: J. Hardie and A.K. Minks (Eds.), Pheromones of
Non-lepidopteran Insects Associated with Agricultural Plants. CABI Publishing,
Wallingford, UK, 1999.
30 E. Zeeck, T. Harder and M. Beckman, J. Chem. Ecol., 24 (1998) 13.
31 P.S. Chen, Experientia, 52 (1996) 503.
32 K.F. Haynes and K.V. Yeargan, Ann. Entomol. Soc. Am., 92 (1999) 960.
33 W. Powell, in: J. Hardie and A.K. Minks (Eds.), Pheromones of Non-lepidopteran
Insects Associated with Agricultural Plants. CABI Publishing, Wallingford, UK,
1999.
34 C. Gemeno, K.V. Yeargan and K.F. Haynes, J. Chem. Ecol., 26 (2000) 1235.
35 W.F. Wood, M.G. Leahy, R. Galun, G.D. Prestwich, J. Meinwald, R.E. Parnell and
R.C. Payne, J. Chem. Ecol., 1 (1975) 501.
693
36 C.E. Linn and L.K. Gaston, Environ. Entomol., 10 (1981) 751.
37 H.E. Hummel and T.A. Miller (Eds.), Techniques in PheromoneResearch. Springer
Verlag, New York, 1984.
38 J.G. Millar and K.F. Haynes, Methods in Chemical Ecology, Vols. 1 and 2. Chapman
and Hall, New York, 1998.
39 P.E. Howse, I. Stevens and 0. Jones (Eds.), Insect Pheromones and Their Use in Pest
Management. Chapman and Hall, London, 1998.
40 A.R. McCaffery and I.D. Wilson (Eds.), Chromatography and Isolation of Insect
Hormones and Pheromones. Plenum, New York, 1989.
41 G.R. Jones and N.J. Oldham, J. Chromatog.A, 843 (1999) 199.
42 M.A. Golub and I. Weatherston, in: H.E. Hummel and T.A. Miller (Eds.), Techniques
in Pheromone Research. Springer-Verlag, New York, 1984.
43 R. Teranishi, R.G. Buttery and H. Sugisawa, Bioactive Volatile Compounds from
Plants, ACS Symposium Series 525. American Chemical Society, Washington DC,
1993.
44 G.R. Takeoka, R.A. Flath, M. Guntert and W. Jennings, J. Agric. Food Chem., 36
(1988) 553.
45 L. Tollsten and G. Bergstrom, Phytochemistry, 27 (1988) 4013.
46 E. Lorbeer, M. Mayr, B. Hausmann and K. Kratzl, Monat. Fur. Chem., 115 (1984)
1107.
47 R.R. Heath and A. Manukian, J. Chem. Ecol., 18 (1992) 1209.
48 C.S. Charron, D.J. Cantliffe, R.M. Wheeler, A. Manukian and R.R. Heath, J. Am. Soc.
Hort. Sci., 121 (1996) 488.
49 J.G. McConnell, J.H. Borden, R.M. Silverstein and E. Stokkink, J. Chem. Ecol., 3
(1977) 549.
50 H.D. Pierce, Jr., A.M. Pierce, J.G. Millar, J.H. Borden and A.C. Oehlschlager, Proc.
3rd Int. Conf. Stored ProductEntomology, Kansas State University, Manhattan, KS,
1984.
51 J.H. Tumlinson, D.D. Hardee, J.P. Minyard, A.C. Thompson, R.T. Gast and P.A.
Hedin, J. Ecol. Entomol., 61 (1968) 470.
52 T.D. Paine, J.G. Millar, C.C. Hanlon and J.S. Hwang, J. Chem. Ecol., 15 (1999) 433.
53 A.B. Attygalle and E.D. Morgan, Angew. Chem. Int. Ed., 27 (1988) 460.
54 T.R. Hamilton-Kemp, J.G. Rodriquez, D.D. Archbold, R.A. Andersen, J.H. Loughrin,
C.G. Patterson and S.R. Lowry, J. Chem. Ecol., 15 (1989) 1465.
55 R.R. Heath and A. Manukian, J. Chem. Ecol., 20 (1994) 593.
56 J.H. Loughrin, A. Manukian, R.R. Heath, T. Turlings and J.H. Tumlinson, Proc.
Natl. Acad. Sci. USA, 91 (1995) 11836.
57 J.F. Chang, J.H. Benedict, T.L. Payne, B.J. Camp and S.B. Vinson, J. Chem. Ecol., 15
(1989) 767.
58 L.E. Browne, D.L. Wood, W.D. Bedard, R.M. Silverstein and J.R. West, J. Chem.
Ecol., 5 (1979) 397.
59 J.G. Millar, C.-H. Zhao, G.N. Lanier, D.P. O'Callaghan, M. Griggs, J.R. West and
R.M. Silverstein, J. Chem. Ecol., 12 (1986) 583.
60 R.R. Heath, P.J. Landolt, B. Dueben and B. Lenczewski, Environ. Entomol., 21
(1992) 854.
61 B. Weissbecker, J.J.A. van Loon, M.A. Posthumus, H.J. Bouwmeester and M. Dicke,
J. Chem. Ecol., 26 (2000) 1433.
62 N.K. Agelopoulos, K. Chamberlain and J.A. Pickett, J. Chem. Ecol., 26 (2000) 497.
63 M.H. Pham-Delegue, P. Etievant, E. Guichard and C. Masson, J. Chem. Ecol., 15
(1989) 329.
64 H.E.M. Dobson, J. Bergstrom, G. Bergstrom and I. Groth, Phytochemistry, 26 (1987)
3171.
694
65 H.E.M. Dobson, in: H.F. Linskens and J.F. Jackson (Eds.), Essential Oils and Waxes.
Springer-Verlag, New York, 1991.
66 J. Faldt, J.M. Jonsell, G. Nordlander and A.K. Borg-Karlson, J. Chem. Ecol., 25
(1999) 567.
67 M. Kleinheitz, H. Jactel and P. Menassieu, J. Chem. Ecol., 25 (1999) 2741.
68 B.V. Burger, Z.M. Munro and J.H. Visser, J. High Res. Chromatogr., 11 (1988) 496.
69 J. Schaefer, in: Peter Schreier (Ed.), Flavour '81. Walter de Gruyter, New York,
1981.
70 H. Rothweiler, P.A. Wager and C. Schlatter, Atmos. Environ., 25B (1991) 231.
71 G. MacLeod and J.M. Ames, J. Chromatogr., 355 (1986) 393.
72 M.J. Lewis and A.A. Williams, J. Sci. Food Agric., 31 (1980) 1017.
73 K.J. Byrne, W.E. Gore, G.T. Pearce and R.M. Silverstein, J. Chem. Ecol., 1 (1975) 1.
74 L.B. Bjostad, L.K. Gaston and H.H. Shorey, J. Insect Physiol., 26 (1980) 493.
75 A. Sturaro, G. Parvoli and L. Doretti, Chromatographia,33 (1992) 53.
76 P.H. Krumperman, J. Agric. Food Chem., 20 (1972) 909.
77 E.D. Pelizzari and K.J. Krost, Anal. Chem., 56 (1984) 1813.
78 R.J.B. Peters, J.A.D.V. Renesse, V. Duivenbode, J.H. Duyzer and H.L.M. Verhagen,
Atmos. Environ., 28 (1994) 2413.
79 J.J. Manura, Adsorbent Resins, PartI: Calculationand Use of Breakthrough Volume
Data. The Mass Spec. Source 7: 3-11. Scientific Instrument Services, Ringoes, NJ,
1994.
80 I. Maier and M. Fieber, J. High Res. Chromatogr., 11 (1988) 566.
81 P. Matile and R. Altenburger, Planta, 174 (1988) 242.
82 K. Grob and F. Zurcher, J. Chromatogr.117 (1976) 285.
83 R.R. Heath, J.R. McLaughlin, F. Proshold and P.E.A. Teal, Ann. Entomol. Soc. Am,.
84 (1991) 182.
84 J.H. Tumlinson, R.R. Heath and P.E.A. Teal, in: B.A. Leonhardt and M. Beroza
(Eds.), Insect Pheromone Technology: Chemistry and Applications, ACS Symposium
Series 190. American Chemical Society, Washington DC, 1982.
85 P.A. Hollingdale-Smith, Chem. Ind., (1975) 226.
86 J. Pawliszyn, Solid Phase Microextraction Theory and Practice.Wiley-VCH, New
York, 1997.
87 R.J. Bartelt, Anal. Chem., 69 (1997) 364.
88 A.J. Matich, D.D. Rowan and N.H. Banks, Anal. Chem., 68 (1996) 4114.
89 B. Schafer, P. Hennig and W. Engewald, J. High Res. Chromatogr.18 (1995) 587.
90 X. Yang and T. Peppard, LC-GC, 13 (1995) 882.
91 H. Verhoeven, T. Beuerle and W. Schwab, Chromatographia,46 (1997) 63.
92 J. Auger, S. Rousset, E. Thibout and B. Jaillais, J. Chromatogr.A, 819 (1998) 45.
93 L. Muller, T. Gorecki and J. Pawliszyn, Fresenius'J.Anal. Chem., 364 (1999) 610.
94 D.C. Robacker and R.A. Flath, J. Chem. Ecol., 21 (1995) 1861.
95 D.C. Robacker and R.J. Bartelt, J. Agric. Food Chem., 44 (1996) 3554.
96 J.J. Langenfeld, S.B. Hawthorne and D.J. Miller, J. Chromatogr.A, 740 (1996) 139.
97 R.J. Bartelt and B.W. Zilkowski, Anal. Chem., 71 (1999) 92.
98 R.J. Bartelt and B.W. Zilkowski, Anal. Chem., 72 (2000) 3949.
99 N.K. Agelopoulos and J.A. Pickett, J. Chem. Ecol., 24 (1998) 1161.
100 J. Faldt, M. Eriksson, I. Valterova and A.K. Borg-Karlson. Zeit. Naturforsch., C 55
(2000) 180.
101 J.S. Elmore, D.S. Mottram and M.A. Erbahadir, J. Agric. Food Chem., 45 (1997)
2638.
102 J. Vercammen, P. Sandra, E. Baltussen, T. Sandra and F. David, J. High Res.
Chromatogr. 23 (2000) 547.
103 A.A. Cossb and T.C. Baker, J. Agric. Entomol., 13 (1996) 301.
695
104 D.C. Robacker and R.J. Bartelt, J. Chem. Ecol., 23 (1997) 2897.
105 D.C. Robacker, J.A. Garcia and R.J. Bartelt, J. Chem. Ecol., 26 (2000) 1849.
106 M.J. Avery and G.A. Junk, Anal. Chem., 57 (1985) 790.
107 L. Pan, J.M. Chong and J. Pawliszyn, J. Chromatog.A, 773 (1997) 249.
108 R.J. Bartelt and D.T. Wicklow, J. Agric. Food Chem., 47 (1999) 2447.
109 R.J. Bartelt and B.W. Zilkowski, J. Chem. Ecol., 24 (1998) 535.
110 D. Rochat, P. Ramirez-Lucas, C. Malosse, R. Aldana, T. Kakul and J.P. Morin, J.
Chromatogr.A, 885 (2000) 433.
111 D. Rochat, P.N. LeMeillour, J.R. Esteban-Duran, C. Malosse, B. Perthuis, J.P. Morin
and C. Descoins, J. Chem. Ecol., 26 (2000) 155.
112 C.P. Malosse, P. Ramirez-Lucas, D. Rochat and J.P. Morin, J. High Res.
Chromatogr., 18 (1995) 669.
113 A. Zhang, C. Linn, S. Wright, R. Prokopy, W. Reissig and W.L. Roelofs, J. Chem.
Ecol., 25 (1999) 1221.
114 A.A. Coss6 and T.C. Baker, J. Chem. Ecol., 25 (1999) 51.
115 S.F. Vaughn and R.A. Boydston, J. Chem. Ecol., 23 (1997) 2107.
116 C.S. Charron and C.E. Sams, J. Am. Soc. Hort. Sci., 124 (1999) 462.
117 H.S. MacTavish, N.W. Davies and R.C. Menary,Annals Bot. (London), 86 (2000) 347.
118 A.K. Borg-Karlson and R. Mozuraitis, Zeit. Naturforsch., 51c (1996) 599.
119 R. Mozuraitis, A.-K. Borg-Karlson, A. Eiras, P. Witzgall, A. Kovaleski, E.F. Vilela and
C.L. Unelius, Abstracts of the Int. Soc. Chem. Ecol. 13th Annual Meeting, Prague,
Aug. 18-22, 1996, pp. 193.
120 R. Mozuraitis, A.K. Borg-Karlson, V. Buda and P. Ivinskis, J. Appl. Entomol., 123
(1999) 603.
121 R. Mozuraitis, V. Buda, V. Jonusaite, A.K. Borg-Karlson and R. Noreika, Entomol.
Exp. Appl., 94 (2000) 15.
122 N. Miklas, M. Renou, I. Malosse and C. Malosse, J. Chem. Ecol., 26 (2000) 2473.
123 R. Maile, F.R. Dani, G.R. Jones, E.D. Morgan and D. Ortius, J. Chromatogr. A, 816
(1998) 169.
124 G. Monetti, F.R. Dani, G. Pieraccini and S. Turillazzi, Rapid Comm. Mass Spec., 11
(1997) 857.
125 B. Frbrot, C. Malosse and A.H. Cain, J. High Res. Chromatogr.20 (1997) 340.
126 A. Peppuy, A. Robert, E. Semon, C. Ginies, M. Lettere, O. Bonnard and C. Bordereau,
J. Insect Physiol., 47 (2001) 445.
127 J.M. Bland, W. Osbrink, D. Woodson and C.B. Vigo, Abstracts of Papers, 218th
NationalMeeting Am. Chem. Soc., 218 (1999) AGRO 130. New Orleans, LA, August
22-26 1999.
128 T. Monnin, C. Malosse and C. Peeters, J. Chem. Ecol., 24 (1998) 473.
129 C. Peeters, T. Monnin and C. Malosse, Proc. R. Soc. Biol. Sci. B, 266 (1999) 1323.
130 J. Liebig, C. Peeters, N.J. Oldham, C. Markstaedter and B. Hoelldobler, Proc. Nat.
Acad. Sci. USA, 97 (2000) 4124.
131 M.F. Sledge, G. Moneti, G. Pieraccini and S. Turillazzi, J. Chromatogr. A, 873 (2000)
73.
132 S. Turillazzi, M.F. Sledge and G. Moneti, Ethol. Ecol. Evol., 10 (1998) 293.
133 S. Turillazzi, M.F. Sledge, F.R. Dani, R. Cervo, A. Massolo and L. Fondelli,
Naturwissenschaften, 87 (2000) 172.
134 K. Dettner, G. Schwinger and P. Wunderle, J. Chem. Ecol., 11 (1985) 859.
135 C. D6scoins and M. Gallois, Ann. Zool. Ecol. Anim., 11 (1979) 521.
136 R. Deml and K. Dettner, J. Chem. Ecol., 20 (1994) 2127.
137 B.V. Burger, Z. Munro, M. Roth, H.S.C. Spies, V. Truter, H. Geertsma and A. Habich,
J. Chem. Ecol., 11 (1985) 1093.
138 E.D. Morgan, Anal. Chim. Acta, 236 (1990) 227.
696
139 A.A. Cosse, J.J. Endris, J.G. Millar and T.C. Baker, Entomol. Exp. Appl., 72 (1994)
233.
140 A.A. Coss6, J.L. Todd, J.G. Millar, L.A. Martinez and T.C. Baker, J. Chem. Ecol., 21
(1995) 1823.
141 P.L. Phelan and H. Lin, J. Chem. Ecol., 17 (1991) 1253.
142 T.C. Leskey, R.K. Prokopy, S.E. Wright, P.L. Phelan and L.W. Haynes, J. Chem.
Ecol., 27 (2001) 1.
143 J.M. Patt, D.F. Rhoades and J.A. Corkill, Phytochemistry, 27 (1988) 91.
144 A. Calogirou, B.R. Larsen, C. Brussol, M. Duane and D. Kotzias, Anal. Chem., 68
(1996) 1499.
145 R.E. Charlton and R.T. Carde, J. Insect Physiol., 28 (1982) 423.
146 K.F. Haynes and R.E. Hunt, J. Chem. Ecol., 16 (1990) 509.
147 K.F. Haynes, L.K. Gaston, M. M. Pope and T.C. Baker, Environ. Entomol., 12 (1983)
1597.
148 T.C. Baker, L.K. Gaston, M.M. Pope, L.P.S. Kuenen and R.S. Vetter, J. Chem. Ecol.,
7 (1981) 961.
149 P. Witzgall and B. Frerot, J. Chem. Ecol., 15 (1989) 707.
150 P. Witzgall, M. Bengtsson, G. Karg, A.-C. Backman, L. Streinz, P.A. Kirsch, Z. Blum
and J. Lofqvist, J. Chem. Ecol., 22 (1996) 191.
151 A. Shani, J. Chem. Ecol., 16 (1990) 959.
152 M.J. Lacey and C.J. Sanders, J. Chem. Ecol., 18 (1992) 1421.
153 W. Boland, P. Ney, L. Jaenicke and G. Gassmann, in: P. Schreier (ed.), Analysis of
Volatiles. Walter de Gruyter, Berlin, New York, 1984.
154 K. Stratmann, W. Boland and D.G. Muller, Angew. Chem. Int. Ed., 31 (1992) 1246.
155 D. Wirth, I. Fischer-Liu, W. Boland, D. Icheln, T. Runge, W.A. Konig, J. Phillips and
M. Clayton, Helv. Chim. Acta, 75 (1992) 734.
156 D. Wirth and W. Boland, Helv. Chim Acta, 73 (1990) 916.
157 H.J. Bestmann, J. Erler and 0. Vostrowsky, Experientia, 44 (1988) 797.
158 K. Seidelmann, K. Luber and H.-J. Ferenz, J. Chem Ecol., 26 (2000) 1897.
159 R. Wegener, S. Schulz, T. Meiners, K. Hadwich and M. Hilker, J. Chem. Ecol., 27
(2001) 499.
160 B.E. Hibbard and L.B. Bjostad, J. Chem. Ecol., 14 (1988) 1523.
161 B.E. Hibbard and L.B. Bjostad, J. Econ. Entomol., 82 (1989) 773.
697
Chapter21
21.1 INTRODUCTION
The term "odor" refers to the biological, physical and psychological effects
caused by the interaction between chemical stimulants-aromas and fra-
grances-and the olfactory systems of living creatures [1]. The study of the
chemical composition of fragrances and aromas is relevant to several fields. In
food science, chemicals or their blends associated with odors discerned as
desirable or agreeable are related to the multivariate set of sensorial responses
known as "flavor" [2]. They are the predominant elements in the perception of
food and beverage quality [3] and, therefore, its acceptance by consumers. Also,
"off-flavors"-connected to volatile substances with unpleasant odors-may be
caused by microbial contamination of foodstuffs [4]. The investigation of these
substances can be an important tool for food safety studies.
Odor can be a significant environmental parameter in some situations, since
it can be related to the human perception of comfort and to the sanitary
conditions of an indoor atmosphere [5], to the contamination of air resulting
from agricultural [6] and industrial [7] activities, as well as to the quality of
natural and treated water [8,9].
Chemical characterization of fragrances released by vegetables and animals
is important to several branches of industry and science. Natural essential oils
from plants are the traditional base ingredients for perfumes and related toiletry
products. However, the rarity of several plants and possible dermatotoxic effects
caused by some of the natural products cause a demand for synthetic alternative
mixtures. The designing of these "synthetic" essential ingredients depends
fundamentally on reliable analytical data on the composition of their natural
correspondents [10]. Moreover, knowledge about the composition and emission
dynamics of floral scents and similar biogenic mixtures of volatile organic
compounds is fundamental in several biological studies due to their multiple
roles in plant reproductive processes, defense against predators and in intra-
species communication [11].
700
example, the composition of samples obtained from detached parts of plants may
not correspond to the actual mixture released by undisturbed live organisms
[21-23], which causes a demand for methods allowing in vivo sampling. Also, the
fact that the production and emission of plant volatiles can be affected or
triggered by factors such as light, environmental temperature, stress and the
presence of trace atmospheric pollutants [24,25], can introduce further
complications.
(d) Some odorants have limited chemical stability, due to photolysis, oxida-
tion and other reactions; e.g., the atmospheric chemical lifetime of mono-
terpenes under daylight conditions was estimated as ranging from less than 5
min (a-terpinene) to 3 h (a- and P-pinene, sabinene) [24].
In consequence, chemical characterization of fragrances and aromas (F&A)
normally demands state-of-the-art techniques for sampling and sample prepara-
tion, analyte separation, detection and quantitation. Usually, the application of
an analytical separation technique in the final steps [2] is involved. Gas chroma-
tography coupled to mass spectrometry (GC-MS) and other similar detection
schemes, such as infrared absorption spectrometry [26], are the conventional
devices employed for F&A chemical analyses. The coupling of olfactometric
detection to gas chromatography (GC-O) is also extremely relevant for qualita-
tive and quantitative chemical analysis of fragrances [27]; several instrumental
and methodological variants of GC-O are described in the literature, such as
CHARM-Analysis and aroma extraction dilution analysis (AEDA) [28]. Along
with the chromatographic techniques, use of "electronic noses" (arrays of elec-
trochemical sensors that generate an electric signal which emulates the expected
response from the human olfactory system) has been growing in recent years.
The development of these fascinating devices involves the arrival of adequate
new materials, such as piezoelectric crystals and synthetic conductive polymers,
as well as sophisticated data processing and chemometric techniques [29]. Their
more remarkable features, from the analytical standpoint, are the possibility of
fast direct qualitative and quantitative evaluation of fragrances and aromas
with limited or no preliminary sample preparation procedures, although the low
sensitivities provided by the devices presently available still prevent their use in
solving most F&A analytical problems.
Except for applications involving measurement with the above-mentioned
electronic noses and similar techniques, adequate isolation and pre-concentra-
tion of the odor-active analytes are mandatory and critical steps in the methodol-
ogies for chemical characterization of fragrances and aromas [2]. Several
sorbent-extraction approaches-classical liquid-liquid extraction (LLE), solid
phase extraction (SPE), solid phase microextraction (SPME) and supercritical
fluid extraction (SFE)-have been employed for sample preparation in F&A
analyses, as well as dynamic and static headspace methods and procedures based
on analytical distillation. In the following sections, the application of some of
these techniques to procedures for chemical characterization of aromas, flavors
and fragrances will be addressed.
701
21.2 ANALYTICAL DISTILLATION PROCEDURES
Distillation is usually carried out in two slightly different ways: in the first, the
matrix to be extracted is mixed or suspended with water in a suitable vessel
fitted with a condenser and, while the mixture is boiled, a condensate phase is
collected (hydrodistillation). After the process, an organic, water-insoluble
fraction (for vegetable materials, the essential oil) can be separated from the
water. In the second procedure, steam is passed through a vessel containing the
matrix-water mixture (steam distillation) to yield a similar condensate. In
either case, an important parameter affecting the composition of the condensate
is the pH of the water [30].
702
Fig. 21.1. Likens-Nickerson simultaneous distillation-extraction
apparatus (modified from Perpbte et al. [38]): 1, body; 2, sample flask; 3, 3
extracting solvent flask; 4, cold finger; 5, inlet for purge gas; 6, cold water
inlet and outlet.
body after their condensation and return to the corresponding flasks, allowing
continuous reflux. A modified apparatus allows use of solvents lighter than
water, such as pentane or ethyl acetate, as extractors. Among others, this device
was employed by Schlotzhauer et al. [35] to study insect attractors in the
fragrance of flowers from Japanese honeysuckle (Lonicera japonica). Large
amounts of linalool and a-farnesene were found in samples collected during
daylight periods. Perpete et al. [36] combined liquid-liquid extraction and SDE
to find the key compound responsible for the worty off-flavor in alcohol-free
beers (1-methylpropionaldehyde), which was not previously found in reports
using other sample preparation techniques such as dynamic headspace analysis.
Fragrances from fresh and processed plums [37], hops [38] and y-ray irradiated
mushrooms [39] were also studied using SDE followed by GC analysis of the
extracts.
703
After GC-MS and GC-O analysis of the distillates/extracts, 2-nonanone,
1-octen-3-ol, 2-heptanol, ethyl hexanoate, methylanisole and 2-heptanone were
determined to be the most relevant odorants on the aromas of natural and
creamy Gorgonzola cheeses. Other interesting application of low-pressure distil-
lation was described by Bouchilloux et al. [43]: using high-vacuum distillation
(temperatures between 35 and 45°C and pressures between 0.5 and 1.5 Pa) and
simultaneous derivatization with p-hydroxymercuribenzoic acid, the authors
were able to determine trace levels of some powerful odorous thiols in Bordeaux
red wines. The analytes were identified and quantified in the samples as their
volatile organomercury derivatives, in concentrations ranging from 0.25-10
-
ggl-l (3-mercapto-2-methylpropanol), 10 ngl-l to 5 tgl1l (mercaptohexanol) and
1-200 ngl- (3-mercaptohexyl acetate). Another alternative is microwave-
assisted distillation, which has been described recently for analysis of the
off-flavors geosmin and methylisoborneol in catfish tissue, in combination with
either adsorbent trapping [44] or SPME [45].
The simplest way to assess the chemical composition of an aroma is direct analy-
sis (by GC or another convenient technique) of a portion of the air in contact
with the odor source, without any other sample treatment step. However, since
little or no pre-concentration of the analytes is involved and, considering the typ-
ical trace or ultratrace levels of the components, usually it is not feasible to apply
SHS to chemical fragrance analysis [48]. For example, out of 118 references
indexed in an extensive literature review on headspace analysis of floral fra-
grances [49], only nine employed SHS as the sampling or sample preparation
procedure. Nonetheless, when techniques and devices with adequate detection
limits are available, and depending on the concentration of the analytes, SHS
can be particularly suitable due to its inherent simplicity. For example, Clarkson
and Cooke [50] employed GC-MS and GC-AED (Atomic Emission Detection) sys-
tems equipped with high-volume injection ports to analyze aroma compounds in
the headspace of commercial cigarettes. Injecting 1 ml of the HS air in contact
with the samples, the authors pointed out that, for these samples, both analysis
time and detection limits were better than those obtained using a commercial
dynamic headspace analyzer.
704
21.3.2 Dynamic headspace analysis (DHS)
A large number of reports describing the use of DHS methods can be found in the
literature regarding the chemical characterization of F&A. Due to the constant
depletion of the analytes from the sample or from the adjacent atmosphere, the
general approach on DHS potentially provides improved analytical sensitivity
when compared to SHS and other equilibrium extraction procedures [2].
Desorption of trapped analytes for subsequent analysis can be performed either
with small volumes of adequate solvents [51] or using on-line automated therm-
al desorption devices [52], the later being eminently suitable for routine
procedures.
Experimental set-ups such as that depicted in Fig. 21.2 [53] can be used,
either in the laboratory or under field conditions, to collect fragrance compounds
emitted by live plants in dynamic conditions. A vase containing a live plant or
some of its parts can be isolated from the environment by a glass bell or similar
device. A controlled flow of purified air is passed through the chamber, carrying
the aroma compounds to a tube containing a sorbent or to a cold finger, where
the volatiles are accumulated for further desorption and analysis. Using a
similar apparatus with polyvinylacetate bags as the isolating device instead of
the glass bell, to emulate the usual procedure employed in field sampling of floral
scents, Raguso and Pellmyr [54] performed an extensive study on the method-
ological aspects of sampling and pre-concentration of floral fragrances by DHS
with solvent desorption and chromatographic analysis. Live flowers of Clarkia
breweri were used as fragrance source, as well as synthetic mixtures of typical
flower odorants. It was found that Porapak Q was more efficient for than Tenax
TA or Tenax TA/charcoal mixtures as sorbent for the evaluated fragrances.
However, the relative abundance of the detected compounds-a useful
parameter in biochemical and ecochemical studies-were not dependent with
the composition of the sorbent. The trapped compounds were desorbed using
either diethyl ether, dichloromethane or hexane; the last solvent provided better
recoveries. In the range studied (0.5-1.0 min-'), the effect of the stripping air
flow rate on the recovery of the trapped analytes generally was not significant
At
Fig. 21.2. Chamber for DHS isolation and collec-
tion of volatiles emitted by living plants (modified
from Baltussen [53]).
705
when compared to other sources of variation on the results between experi-
ments, such as variation of the composition of the volatile emission between
different specimens of the same plant. Sample collection and pre-concentration
by DHS with solvent desorption and GC-MS analysis was also the approach
adopted by Helsper et al. [25] to assess the circadian emission profiles of volatile
compounds by live "Honesty" roses (Rosa hybrida). A similar approach was
adopted by Christensen et alii [55] in a study of the fragrance ofjasmine flowers
(Jasminumpolyanthum) irradiated with y-rays in an attempt to produce floral
fragrances with reduced dermatotoxic or allergic properties. Live flowers were
enclosed in glass bulbs and the volatiles released in 2 h periods carried to
Porapak Q cartridges by a 200 ml-min 1 air stream. The entrapped compounds
were desorbed with 2 ml of purified pentane and the extracts analysed by
GC-MS, GC-FID and GC-O. It was found that the major odorants in the head-
space of non-irradiated flowers were linalool, benzyl acetate, p-cresol, iso-
eugenol, eugenol, 2-methoxy-p-cresol, phenethyl acetate; benzyl alcohol,
(Z)-3-hexenol, (Z)-3-hexenyl acetate and butyrate also have some impact on the
fragrance. No qualitative difference was observed between the fragrance
composition of irradiated and non-irradiated specimens; however, the relative
concentrations of benzenoid and phenylpropanoid odorants such as iso-eugenol
and 2-methoxy-p-cresol were reduced, as well as the overall production of
volatile compounds per plant after irradiation.
Although it can demand more complex and expensive hardware (prices up to
US$15,000) and cannot be applied to fragrances containing temperature-sensi-
tive compounds, DHS with thermal desorption (DHS-TD) has several advan-
tages over solvent desorption: simpler procedure, improved detection limits and
no interference of solvent peaks in the chromatograms. Ageloupoulos et al. [56]
monitored the temporal emission pattern of the volatile compounds (unsatu-
rated C 6 -alyphatic alcohols and aldehydes) released by single intact and damaged
leaves of potato (Solanum tuberosum) and broad bean (Vicia faba) by DHS
(using a specially designed sampling chamber) coupled to GC-MS. The volatiles
were trapped in Tenax TA cartridges and desorbed either thermally (inserting
the cartridges inside the programmed-temperature injector of the GC-MS sys-
tem) or by a solvent (diethyl ether). It was determined that TD, which provided
better detectability and therefore demanded reduced sampling times, allowed
assessing the time emission profiles for the monitored compounds. Vercammen
et al. [57] also employed a commercial automated on-line thermal desorption
apparatus (ATD) coupled to a GC-MS to compare sorption tubes packed with
Tenax TA and with pure grinded polydimethylsiloxane (PDMS), as well as static
headspace SPME, for DHS collection of fragrance compounds released by live
roses and jasmine. PDMS was found to be more efficient for the isolation of vola-
tile odorants released by these plants, resulting also in cleaner chromatograms.
Use of DHS with ATD devices is also predominant for analysis of the fragrances
of detached parts of plants. Mazza and Cottrell [58] determined that the preva-
lent HS components in the aerial parts of three species of Echinacea tradition-
706
ally employed as medicinal plants by North-American native peoples (E.
angustifolia, E. pallida, and E. purpurea) were -myrcene, a- and 3-pinene,
limonene, camphene, trans-ocimene, 3-hexen-l-ol and 2-methyl-4-pentenal. For
the roots, the HS contained mainly a-phellandrene, dimethyl sulfide, 2-methyl-
butanal, 3-methylbutanal, 2-methylpropanal, acetaldehyde, camphene,
2-propenal, and limonene. The volatiles were trapped in Tenax TA cartridges
and analyzed by GC-MS after thermal desorption. Tenax TA was also deter-
mined as adequate for collection of organic volatiles in milk by DHS-TD [59].
Analytical methodologies involving DHS-TD for characterization of the
aroma of foods and beverages are also common. To study the release of volatile
flavor compounds by leaves of Japanese pepper (Xanthoxylum piperitum), a
spice employed in traditional oriental cuisine, Jiang and Kubota [60] collected
the volatiles in the headspace of crushed, mechanically disturbed or intact leaves
in Tenax TA cartridges. The analytes were thermally desorbed in the injector
port of a GC-MS, where they were cryofocused, separated, detected and quanti-
fied (Fig. 21.3). It was determined that the number of detected volatile
compounds increased from 12 (from intact leaves) to 22 (mechanically disturbed
5
s4 intact leaves
c
10 20 30 40 50 60
4 disturbed leaves
a
3
C
2
C
0C
IC
U
10
I . 1.
26
I).i
I
30
. J 40 56 60
I . crushed leaves
C I
3
oJ
1t
I
;
20
w
30
. 11
-t
40
Time / min
-----I-L-
Ali...
50 60
707
leaves) and then to 36 (crushed leaves). For the disturbed leaves, it was found
that the main compounds responsible for the characteristic aroma were limo-
nene (0.56 g.ggl) and (Z)-3-hexen-l-ol (0.63 bug.gl). For crushed leaves, higher
concentrations of C6 -aldehydes, especially (Z)-3-hexenal (17.88 tggl), and
hexanal (6.77 ug.g-), imparted an undesirable grassy odor to the samples.
The combination of DHS, thermal desorption and other sample preparation
methods is occasionally described. Bylait6 et al. [61] studied the composition of
the volatile fraction from different parts of lovage (Levisticum officinale, Koch),
a plant widely employed in the perfume industry. Essential oils from different
parts of the plant (leaves, stems, flowers, seeds, and roots) were prepared by
hydrodistillation and the volatile fractions of these oils were isolated by
DHS-TD, using Tenax TA as sorbent. They found 98 compounds in the volatile
part of the oils, including 41 unreported substances for this plant (e.g.,
2-methyl-2-propenal and carvacrol); the most concentrated compound in all
samples was a-phellandrone.
Introduction of artifacts produced by degradation of the sorbent can be a
major difficulty in DHS-TD; e.g., Canac-Arteaga et al. [62] found out that the
interaction of water vapor, volatile compounds and the trapping material (Tenax
TA) can produce artifacts during the analysis of aromas of dehydrated cheese
and Parmesan cheese. Retention of water by the trapping material can be also a
potential problem, especially when it is necessary to cryofocus the analytes in
the chromatographic column before separation (moisture can freeze and clog the
column). Carbon molecular sieves (Carbosieve, Carboxen) can retain large
amounts of water, as opposed to non-polar polymeric sorbents such as Tenax and
graphitized carbon blacks. Polar polymers as Porapak T and N can also hold
appreciable amounts of water which, however, can be easily removed prior to
analyte desorption by a current of dry air [63].
Introduction of products from the thermal degradation of adsorptive
materials can be avoided in DHS procedures by using cryotrapping. StrAnsky
and Valterovd [64] determined the emission profile of compounds related to the
putrid fragrance from Hydrosme rivieriflowers during their flowering period. A
single live specimen was enclosed in a glass cylinder, which was purged by 15
ml-min - ' of pure air. The compounds released were trapped in a U-shaped tube
containing methanol at -78°C. It was found that emission of dimethyl di- and
trisulphides started after 70 h of flower enclosement, peaking at 118 h and
ending after 142 h (Fig. 21.4). Before 70 h of enclosure, only n-alkanes were
detected. To eliminate water co-volatilized with the analytes, Kolb et al. [65]
proposed a modification in a commercial on-line DHS-cryotrapping device by
incorporation of a water trap (glass-lined steel tube packed with 65% LiCl on
Chromosorb W/AW). The authors reported good results in the application of this
technique to several processes, including the characterization of aroma from
ground coffee. An interesting variation of DHS with cryotrapping is Short-Path
Thermal Desorption (SPTD) [66]: a small amount of solid sample (food or plant
part) is placed in a glass-lined stainless-steel tube, which is directly connected to
708
2-
1-
0-
the cooled tip of the chromatographic column. The sample inside the tube is
carefully heated and the volatile F&A compounds released are directly trapped
and cryofocused at the entrance of the column.
709
23
8
these samples. Zini et al. [70] employed HS-SPME and GC-ITMS (ion trap mass
spectrometry) to assess the emission profiles from intact and mechanically
damaged leaves of living Eucalyptus citriodora trees, using the sampling
chamber shown in Fig. 21.5 and extractions with PDMS/DVB fibers performed
every 30 min for continuous periods between 8 and 10 h. The main compounds
identified were isoprene, citronellal, citronellol and P-caryophyllene. Different
patterns of dependence between extracted amounts and leaf enclosure times
were observed; e.g., for rose oxide (cis-4-methyl-2-(2-methyl-l-propenyl)-tetra-
hydropyran) a maximum in the area versus time curves appears after 300-400
min of leaf enclosure while for citronellal the peak areas decay exponentially
from the start of the experiments. A similar study was conducted on flowering
Boroniamegastigma plants by MacTavish et al. [71]. Volatiles from specimens
in vases placed inside glass or steel vessels were extracted for 30 min with 100
/Am PDMS fibers and analyzed by GC-MS and GC-FID. The principal substances
identified were a-pinene, 5-acetoxylinalool, dodecyl acetate, Z-n-heptadec-8-ene
and f-ionone isomers. From the examination of emission profiles under varied
illumination conditions during 26 h periods of plant enclosure (Fig. 21.6), it was
determined that the pattern is both endogenously and environmentally
controlled.
HS-SPME also has been extensively employed in recent applications related
to food flavors. Roberts et alii [72] studied methodological and operational
aspects of application of HS-SPME to analysis of volatile flavor compounds. For
coffee aroma, best overall sensitivity was achieved using fibers coated with 65
/gm PDMS-DVB; this fiber was also found as the most adequate for heavier polar
odorants such as vanillin. Carbowax-DVB fibers were found to be suitable for
-
organic acids. Non-polar odorants were detectable at the /g.1 levels, were polar
compounds could generally be detected only at mg.l-l levels; e.g., in the extrac-
tion of coffee aroma with PDMS-DVB fiber for 5 min followed by GC-MS analy-
-
sis, 7.7 mgl 1 l vanillin, 116 mgl-1 furaneol, 2.3 mgl l ethylguaiacol, 5.7 mgl-l
-
guaiacol, 40 mgl-l 4-vinylguaiacol, 0.2 mgl-l P-damascenone, 0.4 mg-l ' 2,3-di-
ethyl-5-methylpyrazine and 1 mg l-' 2-ethyl-3,5-dimethylpyrazine were found.
In a similar study conducted by Augusto et al. [73], different SPME fibers were
compared for the characterization of the aromas of industrialized pulps of Bra-
710
A
U
m
a.
U
U
U
Time / min
Fig. 21.6. Dependence between corrected peak areas (peak area per plant) and plant enclosement
time for Boronia megastigma flowering plants (modified from MacTavish et alii [71]). A,
5-acetoxy-linalool; B, a-pinene; 0, continuous darkness; *, continuous illumination; A, A,
alternating darkness/illumination.
711
odorant compounds. For subsequent experiments, the extracted compounds
were thermally desorbed from the fiber and directly introduced into the ioniza-
tion chamber of a mass spectrometer without chromatographic separation. More
than 60 compounds were identified in the HS-SPME extract by GC-MS;
Carboxen-PDMS fiber was found to be especially efficient for collecting sulfur
compounds (e.g., dimethyl sulfide, methyl thioacetate), which are important
impact odorants for this cheese variety. The HS-SPME-MS data obtained
without chromatographic separation was statistically processed and found to be
valuable for qualitative chemometric classification of the samples.
Some drawbacks of HS-SPME should be considered. Since the extracted
amounts are limited by the low mass of sorbent present in the commercial fibers
(less than 1 mg coating on a 100 gm PDMS SPME fiber), the quantification
limits may be higher when compared to DHS methods [57]. However, significant
increases in analyte recovery can be achieved using schemes such as that
proposed by Zhang and Pawliszyn [84]. Using SPME fibers modified to be cooled
by an internal flow of liquid CO2 (which improves the efficiency of the transfer-
ence of analyte from the headspace to the fiber), these authors report near
quantitative extraction of aromatic hydrocarbons by HS-SPME and results
similar to those achieved with a DHS-ATD device. Another problem, related to
the use of fibers with mixed adsorptive coatings (Carboxen-PDMS and
PDMS-DVB) is the lack of analytical accuracy caused by competition between
the analytes for the active sites available on the fiber [85]. When these materials
are employed, extractions with very short exposures of the fiber to the sample
should be adopted to circumvent these problems [86].
Despite its simplicity, the modern tendency is to replace LLE by other tech-
niques, due to the necessity for high purity solvents for trace analysis and to
reduce the environmental and health risks associated with their manipulation.
Also, in the case of biological F&A, they cannot be applied to live samples.
However, there are several contemporary reports on the use of this technique for
F&A analysis, especially for collection of preliminary data [87]. L6pez and
712
G6mez [88] addressed some operational parameters on the application of LLE to
extract aroma compounds from wines. Several solvents-diethyl ether, n-pent-
ane, Freon-11, n-hexane, 1:1 ether/pentane, 1:1 ether/hexane and dichloro-
methane-were compared for extractions of odoriferous terpenic substances
from "artificial wine" (12% v/v ethanol in water). It was concluded that dichloro-
methane and 1:1 ether/pentane are the best solvents for extracting these
analytes from wine. In a related study, continuous LLE was employed by Rocha
et al. [89] to extract free and releasable odor volatiles from Portuguese Bairrada
white grapes employed in the local wine industry. A volume of 250 ml of raw or
enzymatically treated must was continuously extracted for 25 h with 75 ml of
dichloromethane at 50°C. The organic phase was dried and concentrated by
vacuum distillation prior to the GC-MS analysis. Chemometric analysis of the
data allowed differentiation of several grape varieties and of the enzyme treat-
ment studied. Conventional batch LLE is still also occasionally employed in F&A
analysis; recent examples include the identification of odorants in sugar beet
juice [90] and in Grenache Noir fortified wines [91].
SPE can be directly applied to isolate odorants from liquid or liquefiable odorif-
erous samples such as beverages, fruit pulps and tissues. A typical contemporary
application of SPE to aroma analysis was presented by Wada and Shibamoto
[92], who studied the direct extraction of odorants from Chablis red wine using
Porapak Q columns. Samples spiked with selected aroma analytes (2-methyl--
1-propanol, 3-methyl-l-butanol, 2-phenylethanol, ethyl hexanoate, ethyl octano-
ate, diethyl butanedioate and hexanoic, octanoic and decanoic acids) were passed
through 30 cm glass columns packed with the sorbent. Diethyl ether, dichloro-
methane and pentane were evaluated to desorb the extracted materials, which
were analyzed by GC-FID and GC-MS. Dichloromethane was found to be the best
desorbing solvent, recovering up to (103.0 + 3.7) % of the spiked analytes. Analy-
sis of unspiked samples revealed 67 volatile odor compounds in this variety of red
wine. SPE has also been applied to characterize butter aroma: Adahchour et al.
[93] assessed the use of SPE cartridges packed with different sorbents- C,,, C,
NH2 and CN-bonded silica, as well as SDB-1 (PS-DVB copolymer)-to extract
aroma compounds from this material. Butter samples were melted and the aque-
ous fraction, after separation of the fat by centrifugation, was passed through the
SPE cartridges. The cartridges were eluted by methyl acetate, which was dried
and analyzed by GC-MS. Best overall recoveries were obtained with SDB-1 car-
tridges (average recovery ca. 80%); detection limits for impact odorants such as
vanillin and diacetyl were in the low-pg range. For samples such as fruit pulps,
SPE can be combined with isolation techniques such as distillation, as exempli-
fied by Boulanger et al. [94]. After distillation of cupuassu pulp and extraction
with XAD-2 resin, the main odorants were identified as linalool, a-terpineol,
2-phenylethanol, myrcene, limonene, ethyl 2-methylbutanoate, ethyl hexanoate,
713
butyl butanoate, 2,6-dimethyl-oct-7-en-2,6-diol, (E)- and (Z)-2,6-dimethyl-octa-
2,7-dien-1,6-diol and methoxy-2,5-dimethyi-3(2H)-furanone.
Stanfill and Ashley [95] presented an interesting application of SPE to
analysis of flavor-related alkylbenzenes in cigarette smoke. Tobacco from
disassembled cigarettes was "smoked" in a smoking machine and the particulate
material produced collected with filter pads. The collected particulate was then
suspended in hexane and the target analytes extracted from these solutions with
CN-silica SPE cartridges. After elution with a hexane/toluene/tetrahydrofuran
mixture and concentration of the extract by vacuum evaporation, alkylbenzene
odorants were detected and quantified by GC-MS. Detection limits per cigarette
ranged from 1.1 ng (methyleugenol) to 27.8 ng (eugenol). This method was
tested to quantify aroma alkylbenzenes in commercial brands sold in the USA,
and levels of target analytes up to several micrograms per cigarette were found.
Compared with the direct extraction methods discussed above, SFE presents
some advantages for fragrance analysis. For example, since the critical point for
CO 2 (the most popular and most convenient fluid for SFE) is 31.1° C at 7.38 MPa,
extractions can be carried out under milder temperature conditions and without
need of further aggressive procedures such as distillation of excess solvent, as is
common in LLE and SPE [96].
Due to its widespread application on industrial scale to process vegetable
materials, most of the analytical applications of SFE are related to fragrance
analysis of plant materials, using leaves, barks or flowers as samples [97].
Volatile compounds from flowers, leaves and stems of guaca (Spilanthes
americana, Mutis-a South American plant with medicinal and insecticide
properties) were isolated by Stashenko et al. [98] using both SFE and SDE, and
analyzed by GC-MS, GC-FID and GC-NPD. SFE was performed with CO 2 at
40-45°C and 7.24-7.53 MPa, using up to 30 g of vegetable sample. Extraction
time was 2 h and, after depressurization, the collected analytes were dissolved in
2 ml CH 2C12 . Compared to SDE, a larger number of compounds could be detected
at levels above 100 ppb in the SFE extracts (e.g., 67 analytes with SFE and 43
analytes with SDE, for floral extractions); the overall extraction yield for SFE
was up to twice as large of that for SDE. SFE was found to be especially effective
to isolate sesquiterpenes and nitrogenated compounds from these samples.
There are also several SFE applications reported for food flavor analysis.
Leunissen et al. [99] developed a SFE-GC-MS method for fast analysis of roasted
peanut aroma. For aroma isolation, 2.00 g to 2.50 g of ground roasted peanuts
were extracted for 10 min with CO2 at 50°C and 9.6 MPa. The conditions were
adjusted to minimize the extraction of greasy materials. The CO2 was passed
through a silica trap at -5°C to collect the extracted materials; the trapped
analytes were desorbed by dichloromethane and analysed by GC-MS. The char-
acteristic aroma compounds were found to be methylpyrrole, hexanol, hexanal,
714
2-furanmethanol, 2-furancarboxyaldehyde, benzeneacetaldehyde and several
alkylpyrazines. Quantification limits for these analytes were estimated to be in
the range between 80 ng.g (methylpyrrole) and 0.4 ng g (ethyl- and 2,3,5-tri-
methylpyrazine). Supercritical CO 2 was also employed by Calvey et al. [100] to
study organosulfur lacrimatory odorants in ramp/wild leek (Allium tricoccum,
Ait.) a native North American variety of garlic. Ramp bulb homogenates were
mixed with Hydromatrix (diatomaceous earth plus crystalline silica) and
extracted with 2 ml/min CO2 at 35°C and 24.3 MPa for 30 min. The extracted
materials were cold-trapped on glass beads, dissolved in methanol and analyzed
by LC-MS. A total of 24 sulfur compounds (thiosulfinates and cepaenes) were
detected in the samples.
Supercritical CO 2 is a poor extractor for polar substances; therefore, for such
analytes addition of modifiers or use of other fluids, such as supercritical water,
is advisable [2]. Jimdnez-Carmona et al. [101] isolated the aromatic essential oils
from marjoram leaves using supercritical water (2 mlmin-l of water at 150°C
and -5 MPa, for 0.4 g of ground sample), with further analysis by GC-FID.
Compared to hydrodistillation, the essential oils produced after supercritical
water extraction had a higher content of oxygenated monoterpenes, which are
the most significative odorants for this spice (Fig. 21.7). Kubatova et al. [102]
also compared water-SFE and CO2-SFE with hydrodistilation to separate
aromatic essential oils from savory (Satureja hortensis) and peppermint
(Mentha piperita). Extraction of savory with supercritical water for 40 min
removed almost 100% more thymol and carvacrol and 150% more borneol and
linalool than hydrodistillation. For CO2 extractions, yields of borneol and lina-
lool were similar than those for hydrodistillation; however, for thymol and
carvacrol the efficiency of CO,-SFE was only half of that for hydrodistillation.
Water-SFE was determined as highly selective for polar oxygenated flavor
compounds, when compared to CO,-SFE or hydrodistillation.
715
(A)
75
e 50
8
,>
2 9 10
25 IS
6
3 !1
7
10 Time (min)
8 (RI
S-1
Fig. 21.7. GC-FID chromato-
grams of marjoram essential oil 75
extracted by hydrodistillation - 9
(A) and water/SFE (B) (from
Jim6nez-Carmona et al. [101]). .
Peak identification: 1, a-pinene; e 50o
2, -pinene; 3, ]-myrcene; e- ! 11
4, eucalyptol; 5, linalool; 6, 10
2-methyl-6-methylen-7-octen-2-
ol; 7, terpinen-4-ol; 8, a-terp- 25, 6
IS
ineol; 9, geraniol; 10, geranyl i
acetate; 11, 4-ethenyl-a,a,4-
trimethyl-3-(1-methylethenyl)-
cyclohexanemethanol; IS,
2 ;!? i ; i~~~-'
-
'-------'- L'-'.-'-...--' In 7n
716
disturbance. Therefore, both adequate portable or semi-portable instruments
(such as GC-MS) and sampling procedures based on small and rugged hardware
are needed. On the sample preparation side, SPME seems to be the technique
that most closely fits these demands.
REFERENCES
717
J.F. Jackson (Eds.), Modern Methods of Plant Analysis. Springer-Verlag, Berlin,
1997, Vol. 19, p. 141.
31 Y. Saritas, M.M Sonwa, H. Iznaguen, W. A. Knig, H. Muhle and R. Mues,
Phytochemistry, 57 (2001) 443.
32 J.A. Pino, A. Rosado and V. Fuentes, J. Essent. Oil Res., 13 (2001) 21.
33 C.D. Frizzo, A.C. Santos, N. Paroul, L.A. Serafini, E. Dellacassa, D. Lorenzo and P.
Moyna, Braz. Arch. Biol. Technol., 43 (2000) 313.
34 A. Elamrani, S. Zrira, B. Benjilali and M. Berrada, J. Essent. Oil. Res., 12 (2000) 487.
35 W.S. Schlotzhauer, S.D. Pair and R. J. Horvat, J. Agric. Food Chem., 44 (1996) 206.
36 P. Perpete and S. Collin, J. Agric. Food Chem., 47 (1999) 2374.
37 L.F. di Cesare, G. Cortellino and M. Proietti, Ind. Aliment. (Italy), 37 (1998) 14.
38 P. Perp6te, L. Ml1otte, S. Dupire and S. Collin, J. Am. Soc. Brew. Chem., 56 (1998)
104.
39 J.L. Mau and S.J. Hwang, J. Agric. Food Chem., 45 (1997) 1849.
40 P.A. Tarantilis and M.G. Polissiou, J. Agric. Food Chem., 45 (1997) 459.
41 B. Siegmund, E. Leitner, I. Mayer, W. Pfannhauser, P. Farkas, J. Sadecka and M.
Kovac, Food Res. Technol., 205 (1997) 73.
42 L. Moio, P. Piombino and F. Addeo, J. DairyRes., 67 (2000) 273.
43 P. Bouchilloux, P. Darriet, R. Henry, V. Lavigne-Cruege and D. Dubourdieu, J.
Agric. Food Chem., 46 (1998) 3095.
44 E.D. Conte, C.Y. Shen, D.W Miller and P.W. Perschbacher, Anal. Chem., 68 (1996)
2713.
45 S.W. Lloyd and C.C. Grimm, J. Agric. Food Chem., 47 (1999) 164.
46 H. Dobson, in: H. Linskens and J.F. Jackson (Eds.), Modern Methods of Plant
Analysis. Springer-Verlag, Berlin, 1997, Vol. 19, p. 231.
47 H. Steinhart, A. Stephan and M. Bficking, J. High Resolut. Chromatogr., 23 (2000)
489.
48 R.J. Stevenson, X.D. Chen and O.E. Millls, Food Res. Int., 29 (1996) 265
49 J.T. Knudsen, L. Tollsten and L.G. Bergstrom, Phytochemistry, 33 (1993) 253.
50 P. Clarkson and M. Cooke, Anal. Chim. Acta, 335 (1996) 253.
51 S.A. Rankin and F.W. Bodyfelt, J. Food Sci., 60 (1995) 1205.
52 E. Valero, E. Miranda, J. Sanz and I. Martinez-Castro, Chromatographia,44 (1997) 59.
53 H.A. Baltussen, New concepts in sorption based sample preparation for chromato-
graphy (Ph.D. Thesis), Technical University of Eindhoven, Eindhoven, 2000, p. 97.
54 R.A. Raguso and 0. Pellmyr, Oikos, 81 (1998) 238.
55 L.P. Christensen, H.B. Jakobsen, K. Kristiansen and J. Miller, J. Agric. Food Chem.,
45 (1997) 2199.
56 N.G. Agelopoulos, A.M. Hooper, S.P. Maniar, J.A. Pickett and L.J. Wadhams, J.
Chem. Ecol., 25 (1999) 1411.
57 J. Vercammen, P. Sandra, E. Baltussen, T. Sandra and F. David, J. High Resolut.
Chromatogr., 23 (2000) 547.
58 G. Mazza and T. Cottrell, J. Agric. Food Chem., 47 (1999) 3081.
59 B. Vallejo-Cordoba and S. Nakai, J. Agric. Food Chem., 41 (1993) 2378.
60 L. Jiang and K. Kubota, J. Agric. Food Chem., 49 (2001) 1353.
61 E. Bylaite, J.P. Roozen, A. Legger, R.P. Venskutonis and M.A. Posthumus, J. Agric.
Food Chem., 48 (2000) 6183.
62 D. Canac-Arteaga, C. Viallon and J.L Berdague, Analusis, 28 (2000) 550.
63 J. Gawlowski, T. Gierczak, A. Jezo and J. Niedzielski, Analyst, 124 (1999) 1553.
64 K. Stransky and I. Valterova, Phytochemistry, 52 (1999) 1387.
65 B. Kolbe, G. Zwick and M. Auer, J. High Resolut. Chromatogr., 19 (1996) 37.
66 J.G. Wilkes, E.D. Conte, Y. Kim, M. Holcomb, J.B. Sutherland and D.W. Miller, J.
Chromatogr. A, 880 (2000) 3.
718
67 J. Pawliszyn (Ed.), Applications of Solid Phase Microextraction. RSC, Cambridge,
1999.
68 A.J. Matich, in J. Pawliszyn (Ed.), Applications of Solid PhaseMicroextraction. RSC,
Cambridge, 1999. p.3 4 9 .
69 D.A. Vereen, J.P. McCall and D.J. Butcher, Microchem. J., 65 (2000) 269.
70 C.A. Zini, F. Augusto, E. Christensen, B.P. Smith, E.B. Caramao and J. Pawliszyn,
Anal. Chem., 73 (2001) 4729.
71 H. S. MacTavish, N.W. Davies and R.C. Menary, Ann. Bot., 82 (2000) 347.
72 D. D. Roberts, P. Pollien and C. Milo, J. Agric. Food Chem., 48 (2000) 2430.
73 F. Augusto, A.L.P. Valente, E.S. Tada and S.R. Rivellino, J. Chromatogr. A, 873
(2000) 117.
74 X.M. Wan, R.J. Stevenson, X.D. Chen and L.D. Melton, FoodRes. Int., 32 (1999) 175.
75 B. Ancos, E. Ibanez, G. Reglero and M.P. Cano, J. Agric. Food Chem., 48 (2000) 873.
76 S.G. Wyllie and J.K. Fellman, J. Agric. Food Chem., 48 (2000) 3493.
77 H.T. Sabarez, W.E. Price and J. Korth, J. Agric. Food Chem., 48 (2000) 1838.
78 J.C. Beaulieu and C.C. Grimm, J. Agric. Food Chem., 49 (2001) 1345.
79 C. Sala, M. Mestres, M.P. Marti, O. Busto and J. Guasch, J. Chromatogr.A, 880
(2000) 93.
80 J. Weber, M. Beeg, C. Bartzsch, K.H. Feller, D.D. Garcia, M. Reichenbacher and K.
Danzer, J. High Resolut. Chromatogr., 22 (1999) 322.
81 E.A. Nonato, F. Carazza, F.C. Silva, C.R. Carvalho and Z.L. Cardeal, J. Agric. Food
Chem., 49 (2001) 3533.
82 C.J. Scarlata and S.E. Ebeler, J. Agric. Food Chem., 47 (1999) 2505.
83 C. P6res, C. Viallon, J.L. Berdagu6, Anal. Chem., 73 (2001) 1030.
84 Z. Zhang and J. Pawliszyn, Anal. Chem., 67 (1995) 34.
85 R.A. Murray, Anal. Chem., 73 (2001) 1646.
86 J. Koziel, M.Y. Jia and J. Pawliszyn, Anal. Chem., 72 (2000) 5178.
87 M. Mestres, O. Busto and J. Guasch, J. Chromatogr. A, 881 (2000) 569.
88 E.F. L6pez and E.F. G6mez, Chromatographia,52 (2000) 798.
89 S. Rocha, P. Coutinho, A. Barros, M.A. Coimbra, I. Delgadillo and A. D. Cardoso, J.
Agric. Food Chem., 48 (2000) 4802.
90 P. Pihlsgard, M. Larsson, A. Leufv6n and H. Lingnert, J. Agric. Food Chem., 48
(2000) 4844.
91 R. Schneider, R. Baumes, C. Bayonove and A. Razungles, J. Agric. Food Chem., 46
(1998) 3230.
92 K. Wada and T. Shibamoto, J. Agric. Food Chem., 45 (1997) 4362.
93 M. Adahchour, R.J.J. Vreuls, A. van der Heijden and U.A.Th. Brinkman, J.
Chromatogr.A, 844 (1999) 295.
94 R. Boulanger and J. Crouzet, FlavourFrag.J., 15 (2000) 251.
95 S.B. Stanfill and D.L. Ashley, J. Agric. Food Chem., 48-(2000) 1298.
96 Q. Lang and C.M. Wai, Talanta, 53 (2001) 771.
97 R.M. Smith, J. Chromatogr.A, 856 (1999) 83.
98 E.E. Stashenko, M.A. Puertas and M.Y. Combariza, J. Chromatogr.A, 752 (1996)
223.
99 M. Leunissen, V.J. Davidson and Y. Kakuda, J. Agric. Food Chem., 44 (1996) 2694.
100 E.M Calvey, K.D White, J.E. Matusik, D. Sha and E. Block, Phytochemistry, 49 (1998)
359.
101 M.M. Jim6nez-Carmona, J.L. Ubera and M.D. Luque de Castro, J. Chromatogr.A,
855 (1999) 625.
102 A. Kubatova, A.J.M. Lagadec, D.J. Miller and S.B. Hawthorne, FlavourFragr.J., 16
(2001) 64.
719
Chapter22
22.1 INTRODUCTION
Water is the most frequent chemical compound covering 71% of the earth's
surface and more vital to the survival of humankind than food. Thus, to preserve
what is indispensable to life and to protect human health and the environment,
one of man's most important goals is to provide high quality water in sufficient
quantity at affordable costs, while maintaining the various functional roles of
the ecosystems. To achieve this objective, critical limits for various pollutants
have been established in many countries. The surveillance of these critical limits
and of water pollution in general calls for precise and accurate methods for the
analysis of contaminants in water.
Water analysis deals with very different types of aqueous samples, such as
ground and surface water used directly or indirectly for drinking water produc-
tion, rain water, municipal and industrial waste water and process water. These
various kinds of water may differ not only with respect to the types of pollutants
encountered, but in particular with respect to the pollution level, which has to be
considered during sampling, sample preparation and instrumental analysis.
Sample preparation of ground and surface water with a low pollutant level is
usually less cumbersome than, e.g., that of soil and biological samples (including
most food samples). Thus, in organic analysis sample preparation is often
restricted to the extraction of pollutants from the aqueous sample, while a
further cleanup is only required with highly polluted samples or in ultra-trace
analysis. Except for the basic matrix, the water itself, other matrix components
are less dominant in aqueous samples than in soil and biological samples. Humic
compounds and inorganic salts are the major matrix components which have to
be considered in the analysis of aqueous samples. Humic substances interfere in
particular with the analysis of other organic pollutants if analysis is done by LC
with UV detection, as they lead to a pronounced "hump" at the beginning of the
LC chromatogram. The use of a more specific instrument such as a mass
spectrometer will reduce matrix interferences and thus the need for sample
cleanup. However, even if matrix components such as humic compounds and
Comprehensive Analytical Chemistry XXXVII
J. Pawliszyn (Ed.)
© 2002 Elsevier Science B.V. All rights reserved
inorganic salts do not directly interfere with the instrumental analysis, they may
deteriorate the performance of a chromatographic column and thus the robust-
ness of an analytical method or influence the instrumental detection (e.g.
leading to an ion suppression in LC/MS with electrospray ionization); however,
these effects are again less pronounced than in soil or biological samples. Thus,
in water analysis the analytical chemist will have to balance the extent of sample
preparation (in particular sample cleanup) against the resulting labor costs.
Water samples do not only consist of the aqueous, but also of a solid phase in
form of suspended particles. Hydrophobic compounds may preferentially be
adsorbed onto these particles. Many standardized methods of water analysis
include a filtration step using, e.g., a 0.45 pim membrane filter, which removes
the larger fraction of these suspended particles. Further sample preparation of
compounds adsorbed on these suspended particles is not considered in this
chapter. Moreover, the discussion is restricted to organic pollutants. (Note that
many inorganic compounds.in aqueous samples can be subjected to an instru-
mental analysis without any further sample preparation).
722
In particular, analytical data are provided rapidly allowing fast decision making,
which is particularly important if an accidental discharge of (toxic) chemicals
into the aquatic environment has occurred.
In this chapter, various sample preparation methods used for water analysis
are also assessed with respect to their potential for field analysis.
723
higher recoveries are only achieved with exhaustive extraction methods such as
SPE, liquid-liquid extraction (LLE) and purge-and-trap (dynamic headspace),
while with other sample preparation methods, such as static headspace or
SPME, an equilibration (or pre-equilibration) between the aqueous sample and
the extraction phase is attempted, which leads to an only partial recovery of the
analytes.
Limit of detection (LOD). With respect to sample preparation, only the limit
of detection of the entire analytical method (sample preparation and instrumen-
tal analysis) is of relevance; the instrumental detection limit is not. The LOD
should be sufficiently low to verify critical limits. In the European Union, the
critical limit for pesticides in drinking water was set at 0.1+0.05 ug/l. For a
reliable verification of this limit, a LOD of approx. 0.02 /ug/l is necessary. Even
lower method detection limits are desirable, not primarily to detect even lower
concentrations of analytes, but particularly to use smaller sample volumes, as
realized, for instance, in on-line SPE-GC or on-line-SPE-LC and in SPME.
Further criteria used for assessing sample extraction methods for water
analysis are (a) solvent consumption, (b) speed, (c) commercial availability of
equipment and material, (d) ease of automation, (e) suitability for field analysis.
These criteria are discussed below while describing the various sample prepara-
tion methods of water analysis.
724
the kinetics of the extraction (in particular the diffusion of the analytes in the
aqueous phase, through the boundary layer water/extraction phase and within
the extraction phase) and the thermodynamics, i.e. partitioning of the analytes
between the aqueous and the extraction phase, as discussed below.
In this chapter, the methods of extraction and cleanup of aqueous samples
are discussed. Some of these methods are described in separate chapters where
the fundamental aspects are discussed in more detail. To provide a general
overview of sample preparation for water analysis, a certain overlap with previ-
ous chapters is accepted. Emphasis is placed on new methods, while established
methods such as LLE are only briefly discussed. Applications of established
methods such as headspace, LLE and SPE in water analysis are numerous and
will not be summarized here. Instead, applications of new extraction methods
for water analysis are presented in a tabular form. The potential of the various
extraction methods for field analysis is also discussed.
725
If weak acids or bases are to be extracted, the pH has a significant influence
on the actual chemical form of the analyte in the aqueous and the organic phase
and thus the extraction. In this case, the distribution ratio D is used to describe
the partitioning, where D is the concentration ratio of the analytes in all
chemical forms in the organic and the aqueous phase, i.e., for a weak acid HA:
22.2.2 Procedures
For LLE the selection of the optimum solvent is of particular importance, where
the general principle "like dissolves like" applies. Strategies for solvent selection
have been reviewed recently [6]. In this review, a classification of a large number
of solvents according to the polarity and the solvatochromic parameters is given.
Further solvent properties to be considered are summarized in Ref. [3], e.g.
distribution coefficient, selectivity, solubility of the solvent in the aqueous
phase, density, viscosity and vapor pressure. As mentioned above, weak acids or
bases are transformed into their neutral form by a proper choice of the pH value,
i.e., low pH values for acids and high pH values for bases. However, for water
analysis it should be kept in mind that at low pH values more interfering humic
substances are coextracted. If the analyte is ionic over the entire pH range, the
method of ion-pair extraction may be used, i.e. a counter ion is added to the
aqueous sample and the newly formed neutral complex extracted into a non-
polar solvent. Addition of inorganic salts may improve the recovery in LLE
(salting-out effect).
LLE is usually performed as discontinuous extraction either manually or in a
shaking apparatus using a separating funnel, as described by Holden [3]. After
repeated extraction, the organic phases are combined, usually dried over non-
hydrous sodium sulfate, if analyzed by GC (MS), and reduced in volume, as
described below.
If Kd is very small or if the sample is very large, continuous LLE should be
favored [2], where fresh solvent is boiled, condensed and left to percolate
repetitively through the aqueous sample. These continuous liquid-liquid extrac-
tors are commercially available and use organic solvents that are either lighter
than or heavier than water. Several extractors can be operated in parallel
unattended, e.g., overnight, and may provide very high enrichment factors of up
to 10 5 [2].
The formation of emulsions often renders phase separation difficult. Among
other methods [2], filtration through a glass wool plug or centrifugation is used
to break up the emulsion.
726
Liquid-liquid extraction procedures are very simple and straightforward,
and continue to play an important role in water analysis, especially if a small
number of samples are to be analyzed. Thus, many EPA methods are still based
on LLE. However, these methods are time-consuming and difficult to automate
and thus hardly suitable for routine analysis of a large number of water samples.
The consumption of large quantities of solvents and the environmental and
health hazards of these solvents (in particular the frequently used dichloro-
methane) are further severe disadvantages of LLE. LLE is hardly suited for field
analysis of aqueous samples.
Although LLE is more difficult to automate than the other extraction methods
presented below, there have been several approaches to an automatic LLE.
Here, only the LLE coupled on-line to GC is discussed.
In on-line LLE-GC [7-9], the aqueous sample is periodically injected via a
T-piece (segmentor) into the organic phase. The combined streams flow through
an extraction coil out of glass, fused silica or PTFE in which a segmented flow is
formed and extraction occurs. Phase separation is achieved using a semiperme-
able membrane or a sandwich-type phase separator [10] before the organic
phase is injected via a valve into the GC. As the volume ratio of the organic phase
(Vs) to the aqueous phase (Vw) is near unity, the distribution coefficient Kd
between both phases should be > 10 to achieve near quantitative extraction (see
Eq. (22.4)), a prerequisite not easily met in real-life water samples where
pollutants cover a wide range of polarities. Moreover, for such real-life samples,
the formation of emulsions poses a problem. The range of analytes amenable to
this technique may be extended if the on-line LLE-GC system involves simulta-
neous extraction and derivatization [11].
On-line LLE-GC coupling can also be used for automated monitoring of
pollutants in water samples in the field, as shown for the monitoring of volatile
organic compounds [12].
Applications of LLE in water analysis are so numerous that they will not be
presented here. Rather, the applications of on-line LLE-GC to the analysis of
aqueous samples are summarized. The on-line LLE-GC system described above
was used for monitoring volatile organic pollutants in aqueous samples [7] and
for determining halocarbons in seawater [8], organochlorine pesticides in
ground water [9] and pesticide intermediates in water [13]. The approach of
simultaneous extraction and derivatization using on-line LLE-GC was applied to
the analysis of organic acids after ion-pair extraction [11], phenols [14, 15], and
N-methylcarbamates [16].
727
22.2.5 Solvent evaporation
LLE usually leads to a relatively large volume of an organic extract and thus low
analyte concentration, so that a reduction of the extract volume by evaporation
becomes necessary. Thus, solvent evaporation techniques will be briefly
described here. The same techniques are also applicable to other extraction
methods, such as solid-phase extraction, but they will only be discussed here.
Solvent evaporation is a critical part of sample preparation of aqueous
samples as volatile analytes may be readily lost. Thus, special skill and attention
of the operator are necessary. By adding a keeper, i.e., 0.5-1 ml of a high boiling
solvent in which the analytes readily dissolve, volatile losses may be reduced.
In a rotary evaporator, the solvent is removed under reduced pressure by
mechanical rotation of a flask containing the organic extract (or eluate) in a
water bath at both controlled pressure and controlled temperature.
In a Kuderna-Danish(K-D)evaporator,the solvent is distilled and condensed
in a Snyder column. The returning condensate assists in the process of
recondensing volatile analytes [2].
In the automated evaporative concentration system (EVACS), a distillation
column is also used [2]. Distillation appears to be better controlled, e.g., by
sensing the sample level. The final concentration of 10 to 1 ml is achieved by
nitrogen blow-down.
Nitrogen blow-down is often performed manually, e.g., for solid-phase
extraction eluates. With this technique thermal stress is avoided, but special
attention has to be paid. Thus, automatic devices that allow the blow-down of
large solvent volumes and incorporate a sensing device for the final solvent
volume (e.g., 1 ml) should be preferred.
A headspace sample is a gas sample which has been previously in contact with a
liquid or solid sample from which volatile compounds were released into the gas
phase. The headspace sample is subsequently analyzed by gas chromatography
(HS-GC). In this chapter, only the analysis of water samples by HS-GC is
discussed. The water sample is placed in a closed container in contact with the
gas phase (the extracting gas). The gas extraction can be carried out as a
one-step extraction (static or equilibrium headspace), similar to solvent extrac-
tion, or as a continuous extraction by stripping off the volatile compounds in a
continuous flow of an inert gas (dynamic headspace). As with most sample
preparation methods, headspace sampling is used for the extraction of the
analytes from the water sample, and simultaneous sample cleanup as interfer-
ences from non-volatile matrix components (humic substances, inorganic salts)
during the subsequent analysis are minimized. If headspace sampling replaces
728
LLE, the numerous problems with the use of solvents in the subsequent chro-
matographic separation and analyte detection are avoided.
Static headspace sampling is usually a one-step procedure, where an aliquot
of the vapor phase is directly transferred to the gas chromatograph. Dynamic
headspace is usually a two-step procedure, where volatile analytes stripped off
the aqueous phase by an inert purge gas are separated from the matrix or
headspace gas in a trap. From there, they are released into a stream of the GC
carrier gas by thermal desorption and transferred to the column. Thus, dynamic
headspace sampling leads to an enrichment of the analytes and is the method of
choice for low analyte concentrations. Headspace solid-phase microextraction
(SPME) represents an alternative to the classical dynamic headspace and will be
discussed in a separate section.
While headspace sampling is straightforward and appears to be simple,
problems may arise during the subsequent transfer to the GC and the chromato-
graphic separation. Headspace sampling is also described in Chapter 10 and has
recently been reported in another book [17] and a review [18], where the
problems arising from the coupling of headspace sampling and subsequent GC
analysis are discussed in detail. For routine analysis in water laboratories, a
deciding criterion is the automation of HS-GC.
In static headspace, the liquid phase of the water sample (in a vial closed tightly
by a septum) is in equilibrium with the gas phase. The partial pressure of the
analyte in the gas phase is described by Henry's law which for ideal dilute
solutions is given by
Pi = Hxi (22.7)
wherepi is the partial pressure of the component i in the vapor phase, xi its mole
fraction in the aqueous phase (i.e., if n represents the number of moles present,
x i = ni/n,,t,) and H is Henry's constant, which is the basis for a quantification by
HS-GC. Henry's constant and thus the analyte response vary significantly from
analyte to analyte. In addition, Henry's constant depends on the temperature
and increases significantly with temperature. Thus, for reproducible headspace
analyses the water sample has to be thermostatted where enhanced tempera-
tures (e.g., 80°C) significantly increase the partial pressure of the analytes and
thus the response.
In water analysis, the partial pressure of the water in the headspace gas
dominates by far over the partial pressures of the analytes, which, as discussed
below, may cause problems in the subsequent GC analysis.
In static headspace, sampling of the gas phase can be done using a gas-tight
syringe or a sample loop as used in gas analysis by GC. The latter technique is
easier to automate. Filling the loop with headspace is achieved by first pressuriz-
ing the vial with an inert gas (e.g. the GC carrier gas). Alternatively, instead of
729
carrier gas
Fig. 22.1. Schematic set-up of a static
headspace gas chromatograph with bal-
ance pressure headspace system, a cryo-
genic refocusing unit (cryogenic trap)
and a water trap. The system is shown in
the standby mode with backflushing of
the water trap for regeneration. V1-V4 =
solenoid valves. The whole sampling
cycle comprises the additional steps:
pressurizing (needle down, V1, V2, V3
open, V4 open to V1), sampling (needle
down, V1,V2 closed, V3 open, V4 closed
to cryogenic trap) [17] (courtesy of
Elsevier, Amsterdam, Netherlands).
first filling the loop, the pressurized gas can be expanded directly into the GC col-
umn. This method is called "balanced pressure sampling" [191. A commercially
available instrumental set-up which is completely automated is shown in Fig.
22.1. In a first step, the carrier gas pressurizes the sample vial by a proper actua-
tion of the four valves (usually up to the column head-pressure). In a second step,
sample transfer is performed for a short time, while disconnecting the carrier
gas flow. The pressurized headspace in the vial expands directly into the column
and no headspace gas is wasted due to an unnecessary expansion to atmosphere.
730
contain a stationary phase (see Ref. [17]), or better by cryogenic focusing, where
the analytes are trapped in the liquid phase of a GC column at low temperatures.
Cryogenic focusing is readily achieved using the technique of whole-column
cryotrapping (WCC) [26-29], in which the entire GC including the column is
cooled (e.g., to -100°C). Most gas chromatographs are equipped with a so-called
subambient accessory, where, e.g., liquid nitrogen is introduced into the oven
through a valve and cryofocusing can be performed automatically [30,31].
However, in order to generate a narrow starting band of the analytes, it is
sufficient to cool only the beginning of the coated capillary column. A very
effective band concentration is achieved if cryogenic focusing is combined with
an additional focusing by a negative temperature gradient (i.e. the front of the
moving zone is cooler than its rear) [32-34]. This can be achieved with the
instrumental set-up shown in Fig. 22.1. During the sample introduction, cold
nitrogen gas (prepared outside the oven by passing nitrogen through a liquid
nitrogen bath) flows through a PTFE tube outside the fused silica pre-column,
but counter to the warm headspace gas. This set-up is commercially available
and leads to a double focusing of the analyte and to a very narrow band of the
analytes leaving the trap.
Both static and dynamic headspace sampling have to solve the "water
problem", in particular in water analysis. The water in the headspace vapor may
block the GC capillary and cause a peak distortion. The problem is more severe
in dynamic headspace (e.g., with purge and trap) than in static headspace. Water
in dynamic headspace can be removed with a condenser cooled to +5 to -10°C
and placed in front of the Tenax trap [35,36]. Alternatively, the water can be
removed using a Nafion dryer [37-42]. In static headspace, the water is mostly
removed using a water trap with a sorbent coated with LiCl [43]. The water trap
is regenerated in the backflush mode at the end of the GC run by increasing the
oven temperature above + 120°C. Such a water trap is also shown in Fig. 22.1.
Static cryogenic headspace has the main advantage of running fully auto-
mated under computer control. The commercially available set-up shown in Fig.
22.1 can automatically handle 40 vials each containing 2.0 ml of an aqueous sam-
ple. Such automated systems have a good long-term reproducibility. With the
analysis of BTX in water, for instance, a precision of <3% is achieved [17].
Headspace sampling (both static and dynamic, i.e. purge and trap) have been
established methods of water analysis for many years and are well suited to the
routine analysis of volatile compounds in aqueous samples (see e.g., EPA method
624 [76]). Thus, applications of this method are not reported here.
731
tion (SPE), a liquid is passed over a sorbent packed in a cartridge or embedded in
a disk. As a result of the strong attraction, the analytes are retained on the
sorbent. In a second step, the analytes are eluted by passing a small volume of
organic solvent through the cartridge or disk which disrupts the bonds between
analytes and sorbents.
SLE was introduced in the mid 1970s [44]. In these early applications,
enrichment was achieved by adsorption on polystyrene-divinylbenzene
(PS-DVB) polymeric resins (e.g., XAD-2) packed in glass cartridges followed by
elution with a small quantity of suitable organic solvents. The term "solid-phase
extraction (SPE)" was introduced for this sample preparation technique and is
still predominantly used today. Although this term is neither suited to a system-
atic classification of the various sample enrichment methods nor does it describe
the physical process underlying this sample preparation, it will also be used here.
In these early applications, good recoveries and an acceptable precision were
achieved for several analytes; however, the procedure was still time-consuming
as the cartridge had to be packed by the analyst and the sorbents had to be
precleaned thoroughly.
The breakthrough of the SPE method was achieved at the advent of commer-
cially available, prepacked, disposable cartridges with chemically modified
reversed phase silica (such as C8 and C,,). However, in these early days the
recoveries using these modified silica often varied from batch to batch. Today,
different formats such as cartridges or disks and a large variety of sorbents with
different chemical matrices and reproducible recoveries from many suppliers are
on the market (vide infra). They permit an enrichment of organics not only from
environmental but, in particular, from biological samples. In water analysis,
SPE is probably the most widely used sample preparation technique which is
applicable not only to non-polar (hydrophobic) analytes but, more recently, also
to polar (more hydrophilic) compounds. The method is described in several
books [4,5,45-46] and reviews [47-51]. The reader is particularly referred to
Chapter 12 and a recent review by Hennion [48]. The major advantages of
solid-phase extraction are (a) decreased use of (toxic) organic solvents, (b) ease of
automation (allowing a high sample throughput with minimum labor involve-
ment) and (c)its suitability for field analysis (which has not been fully explored).
Method development may be more tedious and the reproducibility poorer with
SPE when compared to LLE.
Solid-phase extraction as a sample preparation step in water analysis is
mainly used for the enrichment of analytes. As with other sample preparation
techniques, a partial sample cleanup is achieved simultaneously so that in
general (at least for the analysis of drinking, ground and surface water), no
additional cleanup is required, in particular if sample preparation is followed by
a selective and specific instrumental analysis as GC/MS and LC/MS. Sample
cleanup in water analysis by SPE is discussed in Section 22.7, while specific
sorbents with enhanced selectivity are described in Section 22.4.6.
The term solid-phase microextraction (SPME), as introduced by Pawliszyn
732
in 1990 [52], suggests that this new method is a miniaturized solid-phase
extraction. Although with some fibre coatings analytes are extracted from the
aqueous phase by adsorption as with SPE, the extraction principle differs
(non-exhaustive extraction with batch equilibrium or preequilibrium with
SPME when compared to exhaustive extraction with flow-through equilibrium
or preequilibrium with SPE) and will thus be presented in Section 22.5.
As with most analytical techniques, the commercial availability of the equip-
ment or material and the incorporation into official standardised methods are
necessary prerequisites for a wide-spread use in routine analysis. Thus, for the
promotion of SPE it has been of advantage that many supplier (more than 50
companies [48,53]) can make products for SPE because of lacking severe patent
restrictions. Moreover, in Germany as early as 1994 several official standard
methods of water analysis (DIN methods) were exclusively based on SPE. More
recently, the technique has also been included in the European CEN methods
and the US EPA methods.
Formats [48]: Cartridges are still the most popular format. They are typically
filled with 40-60 Aum dp packing material and in general made out of polypropy-
lene, polyethylene or glass. (The larger particle size is, e.g., necessary to achieve
sufficiently high flow rates.) Originally, sample contamination by extraction of
polymer additives or impurities represented a problem, in particular for water
analysis, but the use of high purity medical grade polymers for the production of
the cartridge bodies and the frits reduced such contamination. For water analy-
sis, additional reservoirs can be adapted to the cartridge to increase the sample
volume. Vacuum manifolds which allow a parallel processing of several (e.g., 12)
water samples are supplied by several manufacturers. They may include devices
for the drying of the cartridge with vacuum or nitrogen and for the volume
reduction of the eluate by nitrogen blowdown. If suspended particles are not
removed from water samples by filtration, plugging of the cartridge or exces-
sively long percolation times may represent a problem. Thus, nowadays filters
are often integrated into the SPE cartridge. For on-line SPE-GC/MS and
SPE-LC/MS special cartridges are manufactured (vide infra).
With the second most popular format, the disk, much higher flow rates can
be realized due to its large cross-sectional area and thin adsorber bed. The
sorbent (on a polymer or silica substrate) is embedded in a web of PTFE or glass
fibre. Glass fibres are thicker and more rigid and therefore provide higher flow
rates. The sorbent particles are smaller than those used in cartridges (approx. 8
Am instead of 40 Am dp). As mentioned above, the shorter sampling rate and the
smaller particle size allow a more efficient trapping of analytes at higher flow
rates when compared to cartridges. Moreover, disks are more suited to the
extraction of large water sample volumes containing suspended particles. They
are available in several different diameters, the most frequently used size being
733
47 mm. Small (4.6 mm) disks were used for on-line coupling of SPE with LC [54].
While disks are particularly suited to water analysis, it should be mentioned that
other formats have been introduced for a (high throughput) analysis of biological
samples such as the 96-well-plate format [55] or the SPE pipette tip.
Procedures for water analysis: When conventional SPE is used for water
analysis, samples of up to 1 1 are typically handled (in contrast to biological
samples where the sample volume is limited to a few ml). In water analysis, the
SPE procedure (e.g., with C1, sorbents) consists of five steps: (1) conditioning of
the sorbent with one or several organic solvents to reduce background
contaminants and to open the hydrophobic chains to increase the effective
surface area; (2) application of the water sample; (3) washing of the adsorbed
sample by passing a small volume of pure water through the cartridge or disk; (4)
drying of the sorbent bed by passing nitrogen through the bed or by applying a
vacuum; and (5) desorption and recovery of the sorbent with suitable solvents.
Many official standardised SPE methods of water analysis include pre-
filtration through a 0.45 /m PTFE filter. This avoids a plugging of the sorbent
(vide supra) and ensures that only the water-soluble fraction of the analyte is
enriched and analyzed. In particular, hydrophobic analytes may be adsorbed on
suspended particles. Although such suspended particles may be trapped in the
adsorber beds, it is not the primary aim of SPE to extract and enrich the analytes
in both the aqueous and the particulate phase (suspended particles).
The procedure described above is simple and straightforward. Nevertheless,
depending e.g., on the analyte, a thorough method development and validation
are necessary. It has been stated [48] that most SPE methods published to date
have largely been developed empirically in a labor-intensive and tirne-consum-
ing trial and error process with little consideration of the chemistry and physics
involved.
Various approaches to SPE method development that have been described
are based on the analogy between SPE and LC. The following parameters are to
be determined during the method development, as indicated in the SPE
sequence outlined above: (a) selection of the type and amount of sorbent; (b)
determination of the sample volume which can be applied without loss in
recovery (the so-called safe sample volume); (c) composition and volume of the
washing solution which can be applied without loss of analyte and (d) composi-
tion and volume of the elution solution. Some knowledge of the breakthrough
volume is of particular importance.
734
desorption (elution) step [48]. Thus, in a first approximation, SPE can be
described as a simple chromatographic process where the sorbent is the station-
ary phase and (in water analysis) water the mobile phase during extraction and
the appropriate solvent during the desorption (elution) step.
For sample enrichment by SPE in water analysis, reversed phase sorbents
(and sorbents with comparable properties) are frequently selected.
When an aqueous sample is spiked with a solute and percolated through a
SPE cartridge, a breakthrough curve can be observed (e.g., by measuring the UV
absorption), as shown in Fig. 22.2 [48]. The volumes Vb and Vm are defined at 1%
of the initial absorbance A, and at 99% of the initial absorbance A, respectively.
Under ideal conditions, this curve has a bilogarithmic shape, where the inflec-
tion point corresponds to the retention volume of the analyte Vr. The break-
through volume represents the maximum sample volume which can be applied
to a SPE cartridge or disk with a 100% recovery. Thus, knowledge of the
breakthrough volume for a given analyte and sorbent is important for the
development of a SPE method.
The breakthrough volume can be determined by using a spiked water sample
and by monitoring the UV signal continuously and discretely at the outlet of a
pre-column or cartridge [5,56-61] (provided the spike level is lower than the
sorbent capacity). However, such measurements are time-consuming and not
very accurate, in particular for analytes with strong retention (leading to a sig-
nificant spread of the breakthrough curve). Hence, various methods have been
developed to estimate or predict the values of breakthrough volumes, as
described in detail in Chapter 12 and by Hennion [48]. Thus, as a result of the
analogy between SPE and LC, breakthrough volumes may be predicted using
retention factor values [5,48,58]. Vb (at the 1% level, Fig. 22.2) is related to the
retention volume by
735
rates of 5 ml/min [62]. For more detailed information, the reader is referred to
Chapter 12 and to Refs. [48,61-65,92].
Moreover, for SPE cartridges with C,8 silica, k, can be estimated based on
octanol/water partitioning coefficients [66-68]. Since the retention mechanism
is primarily governed by hydrophobic interactions between analyte and the alkyl
chains bond to silica, a relation has been observed between the retention factors
of the analytes and their octanol-water partitioning coefficient (Kw). In particu-
lar, a linear relationship was found between log k, using methanol/water as the
mobile phase and log K,,. (Such a linear relationship was even found for
polystyrene-divinylbenzene polymers, but not for porous graphitic carbon [68],
vide infra). However, it has been shown for very polar analytes (log Ko < 1.5)
that these log Ko values are of limited help for predicting SPE breakthrough
volumes [48]. Since the breakthrough volume is small with such a compound, it
is more reliable to determine k s with measurements of a short C1, column using
methanol/water as the mobile phase in different ratios and by an extrapolation
to zero methanol content.
It should be kept in mind that a 100% recovery of the analyte (as assumed in
the discussion above) is not a necessary prerequisite for water analysis. (The US
EPA accepts methods with recoveries ranging from 70-130%). Thus, as dis-
cussed by Hennion [48], recovery curves (i.e., the recovery as a function of
sample volume) are of more relevance to SPE method development than break-
through curves. Such recovery curves demonstrate that sample volumes higher
than the breakthrough volume can be used in SPE of compounds with high log ks
with only small losses in recovery.
Apart from an efficient retention of the analyte on the sorbent (as discussed
above), a quantitative desorption (elution) of the analyte from the sorbent is the
second important step in SPE method development. To optimise the step, a good
knowledge of the interactions between analyte and sorbent is required. These
interactions are the same as in LC and include mainly (a) hydrophobic interac-
tions (dipole-dipole, dipole-induced dipole and dispersive interactions with
binding energies ranging from 1-10 kcal/mol (approx. 10-40 kJ/mol)), (b)
hydrogen bonding with binding energies ranging from 5-10 kcal/mol and (c)
ionic interactions with binding energies ranging from 50-200 kcal/mol [48,69].
In the past, a large variety of sorbents have been developed and marketed which
allow the extraction of both non-polar and/or polar organic compounds from
water. In addition, sorbents with enhanced selectivity have been developed
which allow sample extraction and cleanup to be conducted in one step. These
sorbents are based on size-exclusion (restricted access materials, RAMs) or
736
molecular recognition (immunosorbents, ISs; molecular imprinted polymers,
MIPs). They are described in Section 22.4.6.
Conventional sorbents include chemically bonded reversed phase silica,
normal phase sorbents, styrene-divinylbenzene polymers, carbon-based sorb-
ents, ion pair and ion-exchange sorbents and mixed mode sorbents. The various
types of sorbents are described in detail in Chapter 12 and Ref. [48].
Chemically bonded reversedphase silica. Initially, chemically bonded revers-
ed phase silica (C8, Cl 8) were mainly used for SPE of water samples. They were
made out of the same stationary phase as those in LC columns, but with a larger
particle diameter (vide supra). These chemically modified silica are well suited to
the extraction of non-polar to moderately polar analytes such as many classes of
pesticides including organophosphorous, triazines, phenylurea and carbamate
pesticides. Poor recoveries were observed with more polar compounds (log Kow <
1.5), in particular polar pesticide metabolites. Phenol (log Kow = 1.5) and
deisopropylatrazine (DIA, log Kw = 1.1) are often considered as representatives
of polar water pollutants and are used by analysts and manufacturers to charac-
terize SPE sorbents [48]. Using conventional SPE cartridges with 500 mg Cl8
packing and a 1 1 water sample, the recovery for phenol is low [70]. It can be
increased by using higher amounts of sorbents or lower sample volumes and by
adding salt (salting-out effect).
The trend in LC is to produce totally non-polar chemically modified silica by
reducing the number of residual silanol groups of the original silica. To achieve
this, trifunctional silane is used for bonding the n-alkyl chains and end-capping
is carried out with trimethylsilane after bonding (although some residual silanol
groups will always remain) [71-73]. In contrast, it was realized that the presence
of residual silanol groups is of advantage for SPE to enhance the interaction with
more polar analytes. Thus, basic analytes are retained by an ion exchange mech-
anism. A second requirement for a good analyte retention is a high coverage of
the surface by alkyl chains, which is often expressed as carbon content percent-
age [48]. These requirements (high percentage of alkyl chains and simultaneous
presence of residual silanol groups) led to the design and development of specific
C,8 silica for SPE. A high carbon content percentage is achieved using silica with
a high specific surface area (500-600 m2/g) as a starting material, while the sur-
face modification by monofunctional silane and no end-capping was used to
enhance the number of residual OH groups and to also provide, in addition to
hydrophobic interactions, hydrogen bonding and ionic interactions. This allows
more polar compounds to be extracted even if large sample volumes are used.
Both properties (high surface coverage by alkyl chains and a high number of
residual silanol groups) are difficult to realize simultaneously. An increase in the
number of residual OH-groups increases the hydrogen bonding interaction,
while the use of monofunctional silanes for surface modification reduces the
hydrophobic interactions. The choice of the optimum sorbent depends on the
individual analyte (in particular on the range of analytes) to be trapped and the
sample volume necessary to reach the desired detection limit.
737
Manufacturers often provide information on the silane function (mono-
functional, trifunctional, end-capping and percent carbon content). A list of
commercially available Cl 8 sorbents and their characteristic properties is avail-
able in Ref. [48].
There is another difference between LC and SPE with C18 silica in water
analysis: the mobile phase in LC contains an organic solvent which, when
adsorbed to the stationary phase, provides a good contact between the solute and
the hydrophobic sorbent surface, while this organic solvent is absent in the
aqueous sample to be enriched. Therefore, the addition of a small amount of
organic solvent (0.5%) to the water sample is recommended.
In addition to C, and C18 silica, other modified silica are available on the
market, namely C2 silica with carbon content in the range 3-6% (most of them
end-capped) and cyclohexyl and phenyl silica. With the latter an increase in
retention of aromatic analytes was reported [105]. Cyanopropyl and amino-
propyl silica represent polar phases which exhibit both polar and non-polar
interactions. Hydrophobic interactions are achieved by using a high carbon
loading of typically 8-9% for cyanopropyl silica.
Finally, normal phase SPE sorbents, e.g., bare silica, alumina, Florisil (a
synthetic magnesium silicate), and silica modified by amino, cyano and diol
groups are available. For these sorbents, SPE is described as adsorption chroma-
tography where only organic solvents are used as mobile phase. Therefore, in
water analysis they are not used for sample enrichment, but rather for sample
cleanup, as described in Section 22.7.3.
Poly (styrene-divinyl benzene) copolymers. The principles and new develop-
ments in SPE based on polymer sorbents have recently been reviewed [77]. As
mentioned above, polystyrene-divinyl benzene (PS-DVB) copolymer resins as
Amberlite XAD resins were among the first sorbents used for SPE [44] and have
been used in many analytical laboratories in spite of the tedious cleaning and
packing of the cartridge. Later, disposable pre-columns designed for on-line
extraction became commercially available. It was only with the advent of PS-
DVB resins with high specific surface areas (700-1200 m2/g), prepacked and
precleaned in disposable cartridges, that the use of these SPE sorbents became
wide-spread in routine water analysis. They are available in cartridge or disk
format [75]1. The higher degree of cross-linking in these polymeric resins leads to
a higher porosity material. The resulting increased specific surface area allows
enhanced F-e interactions and leads to a significant increase in retention of the
polar compounds. For such polar compounds, the retention power is increased
by a factor of 20-60 when compared to the original PS-DVB resins with lower
specific surface area and by a factor of approx. 100 when compared to C18 silica
[5,74,77-89]. For deisopropylatrazine, DIA, in aqueous samples, for instance, a
retention factor logk, = 2.3 + 0.1 was observed for C, 8 material. Log h, increased
to 3.2 + 0.2 for PS-DVB resins with specific surface areas of 350 m2/g and even to
4.4 + 0.3 for a similar resin with a specific surface area of 1060 m2 /g [90]. (Log k,
for porous graphitized carbon, discussed below, is >3.5 for this compound.)
738
In other words, for analytes with a polarity similar to that of DIA and for
large sample volumes, high surface PS-DVB copolymers appear to be the
sorbents of choice. Another advantage over C18 silica is their stability over a pH
range of 1-14, while silica-based sorbents are only stable over a pH range of 2-8.
Unfortunately, there are only few chromatographic data available for these
resins which are not yet routinely used as LC packings [77,91,93,94].
More recently, chemically modified resins have been developed [95-101] and
have become commercially available. The PS-DVB polymer was, e.g., modified
by introduction of acetyl groups or by sulfonation. Even more attractive are
polystyrene-divinylbenzene-N-vinylpyrrolidone (PS-DVB-NVP) copolymers
which have simultaneously hydrophilic and lipophilic properties and are advert-
ised as "universal" sorbents. They are able to enrich acidic, basic and neutral
compounds varying substantially in polarity. For these new sorbents, a
significant increase in retention was reported for several very polar analytes
(such as catechol), which is explained by both the high specific surface area (800
m2/g) and the presence of a pyrrolidone group which acts as hydrogen acceptor
[101]. These hydrophilic/lipophilic sorbents appear to be particularly attractive
for the extraction of drugs and their metabolites from various body fluids. For
water analysis they have been used for the extraction of polar pesticides [291]
and multiresidue analysis [292]. A list of commercially available SPE sorbents
based on PS-DVB or PS-DVB-NVP resins is given in Ref. [48].
Carbon-basedsorbents [102]. The most popular carbon-based sorbent used in
SPE is graphitized carbon black (GCB), a non-porous sorbent with a relatively
low specific surface area (100-200 m2 /g). It is produced by heating carbon black
to high temperatures (2700-3000°C). With this material, chemisorption of oxy-
gen leads to various functional groups which are important for the interaction
with the analyte. GCB sorbents are commercially available on SPE cartridges
and disks [103] and are also produced as porous material (porous graphitic
carbon). The retention mechanism differs from those of C18 silica and PS-DVB
copolymers [104]: compounds are retained by both hydrophobic and electronic
interactions (ion exchange mechanisms). Thus, non-polar, hydrophobic and very
polar analytes in water are well retained which makes the sorbent particularly
attractive for a multi-residue analysis of, e.g., pesticides (including the polar
metabolites), as extensively shown by Di Corcia and others [103-113]. A particu-
larly high retention is achieved for planar molecules containing several
functional groups and for systems which are rich in polarizable electrons [137].
Ion exchange and ion pair sorbents. Ionic or ionizable analytes can be
extracted and enriched from water samples using ion-exchange sorbents which
are mainly produced by chemical modification of PS-DVB copolymers. Weak
cation exchangers normally have carboxylic groups and strong cation
exchangers sulfonic acid groups, while weak anion exchangers have primary or
secondary amino groups and strong anion exchangers quaternary amino groups.
SPE of ionizable analytes is straightforward: retention occurs at a sample pH
where the analyte is present in its ionic form and desorption at a pH where the
739
analyte is present in its neutral form. If the analyte is ionic over the entire pH
range, desorption is achieved with a mobile phase of higher ionic strength. With
water samples (environmental samples), the main problem is caused by the
inorganic ions present in high amounts, as they readily adsorb to the ion
exchange sorbent and cause an overload of the sorbent. This problem can be
partly overcome by a pretreatment of the water sample to remove the major
inorganic ions, e.g., by precipitation or complexation [114]. When the organic
ions are more hydrophobic, an additional hydrophobic interaction occurs with
the matrix of the ion exchanger leading to a better retention.
Another important approach to the extraction of ionic organic analytes from
water represents the method of ion-pair extraction. With organic anions such as
aromatic sulfonic acids, for instance, ion pairs are formed with quaternary
ammonium salts (tetramethyl-, tetrabutyl- or cetyltrimethyl ammonium) which
are then enriched on a conventional C or Cs sorbent. Typical applications are
aromatic sulfonates such as naphthalene sulfonates and linear alkylbenzene-
sulfonates (LAS) where the latter are used as anionic surfactants [115-119].
Mixed mode sorbents. Mixed mode sorbents contain reversed phase alkyl
chains and cation exchangers bonded to the same matrix (e.g., octyl chain and
cation exchange groups). They are particularly used for a simultaneous extrac-
tion and cleanup of drugs and their metabolites in body fluids [120]. Drugs
containing nitrogen are transformed into their protonated forms at low pH and
bond to ion exchange sorbents, while all other analytes and interferences that
first bound to the reversed phase mode of the sorbents by hydrophilic interac-
tions are eluted with methanol. Finally, the analytes are desorbed with a basic
solution to break the ionic interactions. Thus, these mixed mode sorbents allow a
simultaneous extraction and cleanup. Apart from drug analysis, they have been
applied to the analysis of aqueous samples, e.g., the extraction of triazines from
water [120].
740
substances prior to the subsequent chromatographic analysis, e.g., for the
determination of acidic herbicides [130] or triazines [131].
Immunoaffinity sorbents (ISs). In trace and ultra-trace analysis of organic
pollutants in water, an unequivocal identification and quantification of target
analytes may be difficult because of many interferences present at higher
concentrations which may coelute with the target analyte. This is particularly
the case if the chromatographic unit (GC or LC) is not coupled with a specific
detector such as a MS. In water analysis of polar compounds by LC with UV
detection, there is the additional problem of coelution with humic substances
which are observed as a large unresolved hump in the first part of the chromato-
gram. Such interferences can be avoided if highly selective SPE sorbents, such as
immunoaffinity sorbents (ISs) or molecular imprinted polymers (MIPs), are
used for sample enrichment. These techniques are based on molecular recogni-
tion. Immunoaffinity sorbents are described in detail in Chapter 33. Here, only
the basic principles and the application of these techniques to water analysis are
briefly repeated.
In the case of immunoaffinity sorbents, the molecular recognition is
achieved by antigen-antibody interaction. The antibodies are covalently bound
to an appropriate sorbent to form the immunosorbent, which is packed into a
SPE cartridge or pre-column. The binding of the analyte is the result of good
spatial complementation and specific intermolecular interactions between
analyte and antibody. Binding can also occur to analytes with structures similar
to those of the target molecule, leading to a so-called cross reactivity of antibod-
ies. This cross reactivity, which is a negative property of immunoassays, can be
used to extract structurally related analytes.
Immunosorbents have mainly been applied to the analysis of biological
samples, as discussed in Chapter 33. In this field, commercial ISs have been
introduced in the last decade. In environmental analysis, they have been pre-
dominantly used for a selective extraction of target pesticides and their major
metabolites [132-136] and steroid estrogens [137]. In addition, they have been
developed for the extraction of a whole compound class by exploiting the above-
mentioned cross reactivity or by mixing two antibodies in the same cartridge.
Therefore, ISs were again developed for the extraction of classes of pesticides
[138,139] as well as for BTEX, polycyclic aromatic hydrocarbons (PAH) and
azodyes [140-150].
Molecularly imprinted polymers (MIPs). The preparation of antibodies is
still cumbersome and involves animals. Therefore, it has been attempted to
synthesise polymers with molecular recognition sites analogous to those in
antibodies. In recent years, this has led to the development of molecularly
imprinted polymers (MIPs). The synthesis is achieved using suitable monomers
assembled around the target molecule (the "template molecule") and a cross-
linker for subsequent polymerisation. In a second step, template molecules are
washed from the porous polymer with a suitable solvent resulting in polymers
with cavities which are "imprints" of the target molecule. These cavities act as
741
recognition sites if the MIP is used for the selective extraction of target
molecules from the sample. The selectivity is based on both the shape of the
cavity and the specific interactions between the target molecule and the polymer
(usually hydrophobic, hydrogen or electronic interactions). When compared to
antibodies, such MIPs can be prepared more rapidly and easily. In addition, they
are stable at high temperatures and over a large pH range. The use of MIPs as
SPE sorbent was first described by Sellergren in 1994 [151]. This technique is
also termed "molecular imprinted SPE" (MISPE) and has so far mainly been
applied to the selective extraction of drugs from body fluids and other biological
samples [152,153]. Synthesis of MIPs and their applications have been reviewed
repeatedly [154-158].
A specific problem of the use of MIPs as selective SPE sorbents in trace
analysis results from residual template molecules which were not completely
removed during the washing step and which may elute together with the target
molecule during the subsequent sample extraction/elution. Such interferences
are avoided if closely structural analogues are used as template molecules during
MIP synthesis.
MIPs are usually prepared in an apolar (non-aqueous) solvent. There are
some difficulties using MIPs for water analysis since the hydrogen bonding and
other interactions are weakened in a highly polar aqueous environment which
can decrease the interaction between the analyte and the MIP and thus the
selectivity. This is one reason why applications to water analysis are still scarce.
So far, MISPE has mainly been applied to the extraction of pesticides [159-164,
294,295] as well as 4-nitrophenol [165] from different aqueous samples. The
selectivity of the sample cleanup can be further enhanced by combing RAM and
MIP sorbents as shown for the analysis of triazines in river water [296].
The commercial availability of MIPs as SPE sorbents will be a necessary
prerequisite for their wide-spread use for water analysis.
742
during the day, night or weekend, time savings are still achieved. The currently
fastest serial processing systems extract 25-50 samples per hour [166,167]. A
significant increase in sample throughput, however, is only achieved if solid-
phase extraction is done in parallel [167-169], i.e., if large numbers of samples
are extracted simultaneously videe infra). Parallel processing systems can han-
dle up to 400 samples per hour [167]. Although with automated SPE a higher
precision (and accuracy) is generally achieved, systematic errors (resulting, for
instance, from blocked cartridges) may occur undetected. Another problem with
automated systems is analyte carryover which may lead to erroneously elevated
analyte responses at low levels. Moreover, sample stability may become a prob-
lem, in particular if a large number of samples is processed over a long period,
such as the weekend.
For automated solid-phase extraction to be worthwhile, a minimum number
of samples is required. It has been stated that this break-even number is about
10 [166]. Automated solid-phase extraction can be performed on-line or off-line
as described below.
743
compared to conventional SPE cartridges. Breakthrough is avoided by using
smaller sample volumes (e.g., 10-100 ml). As discussed above, breakthrough of
more polar compounds is reduced if polystyrene-divinyl benzene (PS-DVB) is
used instead of the conventional C8 or C,,. As the entire analytes trapped on the
sorbent are transferred onto the analytical column, comparable or even better
detection limits are achieved with on-line SPE-LC using 10-100 ml sample
volumes when compared with conventional SPE extraction of a 1 1water sample.
Thus, the use of on-line SPE-LC is particularly attractive if only small sample
volumes are available.
(b) The sorbents of the pre-column and the analytical column have to be
compatible. Ideally, the sorbents of the pre-column and the analytical column
should be identical. This is easily realised if C18 sorbents are used for SPE.
However, an on-line coupling of SPE with LC using PS-DVB pre-columns and
C, 8 analytical columns [99,171-183,270] or graphitized carbon [184-186] is also
readily achieved.
While the accuracies are comparable [189-191], a better precision is often
achieved with on-line SPE-LC than with off-line extraction [187,188].
Apart from commercial equipment specially designed for on-line SPE-LC,
conventional LC instruments with autosamplers may be modified and prog-
rammed for on-line extraction [192]. In this case, the aqueous sample is injected
onto a pre-column in several 1 ml fractions for sample enrichment before elution
onto the analytical column follows.
Sample enrichment using a restricted access material (RAM), as discussed in
Section 22.4.6, is usually only done on-line, as disposable RAM cartridges are too
expensive and automation is desirable for routine analysis of a large number of
similar samples. Sample volumes are usually restricted to < 50 Al. After each
injection, the RAM pre-column is regenerated by washing and the regeneration
is done in parallel to the chromatographic separation of the next sample.
In water analysis, the use of RAM sorbents is particularly attractive for the
removal of humic substances, as mentioned above [133,134]. In addition,
immunosorbents (ISs) [140,141,193-197] and molecular imprinted polymers
[MIPs] [163-165,198-201] have also been used successfully in on-line SPE-LC
applications for water analysis.
SPE-GC. For the analysis of volatile compounds amenable to GC, in water
analysis on-line coupling of SPE with GC is more attractive than on-line coup-
ling with LC, as the higher separation efficiency of GC columns and the ease of
GC/MS couplings lead to a better selectivity and specificity of the overall analyti-
cal method. On-line SPE-GC coupling has been reported and reviewed in detail
[170,202]. Sample enrichment on the SPE cartridge is similar to that for SPE-LC
(pre-column conditioning, sample application, washing with a small amount of
water). However, for a better compatibility with the GC, typical sample volumes
for water analysis are 1-10 ml with SPE-GC compared to 10-100 ml with
SPE-LC. In contrast to SPE-LC, in SPE-GC residual water on the pre-column is
not compatible with the GC and has to be removed in a drying step (vide infra).
744
Desorption (elution) is achieved with 50-100 pL of a suitable organic solvent. In
order to transfer the entire desorption solution to the GC, the GC system has to
be equipped with a "large-volume-injection-system". A schematic set-up of an
on-line SPE-GC system is shown in Fig. 22.3. The system has at least three
switching valves: one to switch between aqueous (conditioning and sampling)
and organic (elution) solvents, a second to supply drying gas, and a third to direct
the desorption solution to the GC system. As shown in the figure, for on-column
large volume injection (e.g., 100 l) a system of two pre-columns and an analyti-
cal column is used. The first uncoated pre-column (e.g., 5 m long) acts as a
retention gap using partially concurrent solvent evaporation (PCSE) [203-205],
while the analytes are focused on a second coated pre-column (e.g., 2 m long), the
so-called retaining column. After this retaining column the solvent is split off via
a T-splitter (solvent vapour exit, SVE) [202,206,207].
For analyte retention, PS-DVB sorbents are best suited because of their
higher breakthrough volume (when compared to C or C,,). Non-polar to
medium-polar compounds are quantitatively retained on a 10 x 2-4 mm i.d.
pre-column and with sample volumes of < 10 ml (or even up to 100 ml). For
desorption (elution) of typical water pollutants, methyl- or ethylacetate are used
[202,208,209]. Originally, the cartridge was dried by purging with nitrogen. This
procedure is rather time-consuming. Thus, a short drying cartridge (e.g., packed
with silica) inserted between the SPE pre-column and the GC is recommended
[170,210-212]. The drying agent is reconditioned by heating after the GC run
and can be re-used for up to 100 times.
Devices for on-line SPE-GC are commercially available and similar to those
used for on-line SPE-LC (see Fig. 22.3). With these devices all steps are carried
out automatically under computer control. One of these systems allows > 100
cartridges to be automatically loaded into the instrument. However, for the
analysis of water with a low pollutant level (e.g., surface or tap water), cartridges
can be re-used at least 100 times [202], while it is recommended to replace the
retention gap after 50-100 GC runs [202].
It has been stated that this approach to on-line SPE-GC is applicable to all
analytes which can be trapped on the (polymeric) pre-column and dissolved by
745
methyl- or ethylacetate and which are amenable to GC. The potential of on-line
SPE for on-site field analysis is evident and will be reported in Section 22.4.8.
A potential problem of on-line SPE-GC is the loss of volatile analytes due to
the approach of partially concurrent solvent evaporation (PCSE) and the instal-
lation of solvent vapor exit [170,213,214]. With an optimized system, analytes as
volatile as monochlorobenzene and xylene are recovered.
In water analysis, sampling is predominantly still done in the field, while sample
preparation and instrumental analysis are carried out in a central laboratory.
Rapid screening tests and various sensors have been developed and are, in part,
commercially available for many pollutants. They are mainly used to monitor,
e.g., ground and surface water near hazardous waste sites and surface water
after an accidental release of pollutants into the aqueous environment. How-
ever, these tests are usually at best semi-quantitative and give information on a
very limited number of target pollutants. For a detailed qualitative and
quantitative analysis of water samples in the field, chromatographic instru-
ments such as GC or LC coupled with selective detectors (in particular with a
mass spectrometer) are still needed and here SPE plays an important role.
Water extraction and instrumental analysis in the field can be done off-line
or on-line. Off-line methods are important. After solid-phase extraction of water
samples using, e.g., a cartridge, it is much easier and safer to transport the
cartridge to the central laboratory than a 1 L water sample in a glass bottle. This
holds in particular for samples collected at remote sites [216-222]. Moreover,
using this approach sample stability is usually improved. The stability of various
analytes sorbed on disposable cartridges has been investigated as a function of
the storage time and temperature, type of sorbent and cartridge, pH and matrix.
While in general an acceptable stability was observed with all analytes, this was
not the case for some non-stabile organophosphorous pesticides [173]. Polymeric
746
sorbents appear to be better suited to long-term storage. Polar phenolic com-
pounds, for instance, could be stored at -20°C for two months [81] while a good
stability over a seven-week-period was reported for several groups of pesticides
[224].
For an on-site monitoring of individual pollutants, the fully automated
on-line SPE-GC or SPE-LC systems are particularly suitable. Thus, under the
"Rhine Basin programme", a System for Automated Monitoring of Organic
Compounds in Surface Waters (SAMOS) based on on-line SPE-LC was devel-
oped and tested [188]. This system was considered as an early warning system.
At analyte concentrations of 1 lg/l, a good precision ranging from 1-15% was
observed, with the exception of polar compounds (due to matrix interference and
partial breakthrough) [225]. The robustness of the system was studied in two
laboratories during five and seven months long periods. No major problem was
encountered in over 1000 analyses. An even better performance is achieved
when the water sample is filtered [226].
For more volatile compounds, the on-line SPE-GC coupled with a mass
spectrometer is probably better suited (although most likely less robust than
on-line SPE-LC systems). If combined with a sample delivery unit (as, for
instance, available on commercial on-line SPE-GC/MS systems), an unattended
run is possible [202]. Non-target analysis is done with the MS running in the
scanning mode, while selected ion monitoring (SIM) is used for target analysis.
During the last two decades numerous applications of solid-phase extraction for
water analysis have been reported. At present more than 200 papers appear
every year in which the use of SPE for water analysis is reported. Thus, an
exhaustive overview of the application of SPE in this area is not possible here.
Various applications of special SPE techniques such as the use of polymeric and
carbon based sorbents, restricted access materials, immunosorbents and molec-
ular imprinted polymers were reported in Sections 22.4.5 and 22.4.6. Many
further applications are found in a review by Hennion [48].
SPE is particularly suited to the extraction of moderately polar to polar and
semivolatile compounds from aqueous matrices. Pesticides are typical analytes
which belong to this polarity range. Thus method developments for pesticides
dominate the SPE applications as summarised in a recent review [49] in which
the monitoring of triazines and their degradation products in natural water is
discussed in detail. Moreover solid-phase extraction of carbamate pesticides
[227], acidic herbicides [228] and quaternary ammonium herbicides [229] has
been reviewed. An overview of multiresidue analysis based on SPE was present-
ed [230]. In further reviews the application of SPE to the extraction and analysis
of polycyclic aromatic hydrocarbons [231], phenols [232], antifouling compounds
[224] and aromatic sulfonates [117, 231] has been summarised. In addition SPE
was extensively applied to the extraction of surfactants and estrogens. More
747
TABLE 22.1
Recent applications of on-line solid-phase extraction for water analysis (coupled to liquid
chromatography (LC), capillary electrophoresis (CE), supercritical fluid chromatography (SFC)
or gas chromatography (GC))
Compound class (type of agent) Instrumental Special methods Ref.
analysis
Pesticides (metabolites) LC 49,227,271-280
(quaternary ammonium) LC Ion pair extr. 274,277
(carbamates) LC Post-column deriv. 278
(acidic herbicides) LC 275
CE 287
SFC 288
Polar pollutants LC 279,280
LC Dual pre-column 293
Endocrine disruptors (estrogenic LC 246,253
hormones)
(alkylphenols, bisphenol A) LC 259
Antifouling compounds LC 281,282
GC 290
Naphthalene sulfonates LC Ion-pair extr. 283
Phenols LC 279,284,285
4-nitrophenol LC Mol. imprinted polymer 165
Anilines LC 285
Cyanobacterial toxins LC 286
Organic pollutants GC 289
748
This relatively new sample preparation method was invented in 1990 by
Pawliszyn [52] and rapidly spread into various areas of analytical chemistry with
more than 700 papers published today. The method is described in several books
and reviews [297-314]. Aqueous samples were among the first matrices studied
and water analysis is still one of the main application fields of SPME.
SPME is a very simple solventless sample preparation method which
combines sampling, sample enrichment and sample cleanup in one step. With
SPME, the extracting polymer phase is either exposed directly to the aqueous
sample or to the headspace above the sample. The theory of SPME is presented
in detail in Chapter 13. Here, only the basic equations describing the extraction
process from an aqueous sample and its headspace are repeated. If the extracting
polymer phase is exposed to the headspace of an aqueous sample (and the
humidity in the headspace is neglected), the equilibrium between the three
phases is described by two distribution coefficients, the distribution coefficient
between the gas and the aqueous phase, Kgw, and the one between the polymer
and the gas phase, Kpg, where the overall distribution coefficient between the
polymer phase and the aqueous phase, Kpw, is
KpW = Kpg. Kgw (22.9)
If extraction occurs by absorption (rather than adsorption) by the coating, the
extracted mass of the analyte is
Kpw Vp CoVW (22.10)
n- (22.10)
Kpw .V. +Kgw .Vg +Vw
Vp, Vg, V, are the volumes of the polymer, gas and aqueous phases, respectively,
and C0 the initial concentration of the analyte in the aqueous sample. Equation
(22.10) states that the amount of analyte extracted is independentof the location
of the extractingpolymerphase, which may be placed in the headspaceor directly
in the aqueous sample. If there is no headspace, Eq. (22.10) simplifies to
n= (22.11)
Kw Vp +Vw
which applies to direct SPME.
Usually, the volume of the polymer coating is much smaller than the volume
of the aqueous sample (Vp < < Vw) so that Eq. (22.11) further simplifies to
N = Kpw Vp. Co (22.12)
i.e., the extracted analyte mass is independent of the volume of the aqueous
sample which is important for field sampling.
22.5.2 Procedures
SPME devices. The original and still most popular procedure for SPME of
aqueous samples uses a small fused silica fibre coated with a thin polymer phase
749
attached to a syringe-like holder (see Chapter 14). (As mentioned above, the
volume of the extracting phase is very small, usually less than 1 L, when
compared to the sample volume, therefore "microextraction"). The syringe-like
holder provides protection of the fibre during transport and allows a piercing of
the rubber septum of the sample vial and the GC or LC injector. For GC
applications in water analysis (which by far prevail), the fibre is first exposed
either directly to the aqueous sample (direct SPME) or to the headspace (head-
space SPME) until an equilibrium (or pre-equilibrium) is reached and then
introduced (manually or by autosampler) into the hot injector, where thermal
desorption occurs. A polymer-coated fibre can also be used for LC application.
After analyte extraction, the fibre is exposed directly to the mobile phase in a
special interface [315], where solvent elution of the analytes occur. Other SPME
devices have also been reported for water analysis: for GC analysis a stir bar can
be coated with a polymer phase [316,474]) leading to a significant increase in the
extracting phase volume and thus to lower detection limits for water analysis.
For LC applications, the inner surface of a capillary can be coated with the
extracting phase (in-tube SPME) or the coating of a GC open tubular column can
be used for extraction [317-323]. The extraction capillary is placed between the
sample injector loop and the injection needle of a commercial LC instrument. To
reach an equilibrium or preequilibrium in in-tube SPME-LC, the sample or an
aliquot of the sample is repeatedly drawn through the extraction capillary by the
autosampler.
Direct SPME versus headspaceSPME. For water analysis, both direct SPME
and/or headspace SPME [300,325-329] can be used. Headspace SPME is the
method of choice for dirty water samples, such as waste water, if the analytes are
sufficiently volatile. The extraction phase (fibre) is protected from adverse
effects by the matrix, e.g., precipitation of non-volatile and high molecular
weight compounds such as humic substances. It appears that most compounds
amenable to GC can also be extracted by headspace SPME [327]. As mentioned
above, the amount of extracted compounds does not depend on the location of
the fibre. However, the choice of the sampling mode has a significant impact on
the extraction kinetics. When the fibre is exposed to the headspace, the analytes
are first removed from the headspace, followed by an indirect extraction from
the aqueous phase. Thus, volatile analytes are faster extracted than semi-
volatiles. In general, the equilibration times for volatile compounds are shorter
for headspace SPME than for direct SPME. This is i.a. due to the higher
diffusion coefficient in the gas phase when compared to the aqueous phase.
Headspace analysis of semi-volatiles can be accelerated by increasing the
temperature and by agitating the sample (e.g., by stirring). With increasing
temperature the vapor pressure of the analytes increases, but the distribution
coefficient Kw decreases. Thus, under equilibrium conditions, the extraction
yield as a function of the temperature passes through a maximum when the fibre
is not cooled [330].
750
Sample volume. The sample volume can be neglected if the distribution
coefficient between polymer and aqueous phase is not very high [331,332]. It is a
significant advantage for water analysis that small sample volumes (e.g., 2 ml)
are sufficient for quantitative analysis. Thus, the aqueous samples can be placed
directly into an autosampler, provided the sample is sufficiently agitated
[333-338]. Thus, SPME is particularly attractive if only small volumes of water
samples are available. In spite of the small sample volume and although in
general only a small fraction of the analytes is extracted by SPME, low method
detection limits are still achieved as the entire extracted analytes are transferred
to the chromatographic system.
For fibre SPME in particular a variety of extraction phases with different polari-
ties and thus different selectivities are commercially available [339]. The selec-
tion of a suitable extraction phase is often based on the principle "like solves
like". The extraction phases can be differentiated according to the underlying
extraction principle: Liquid polymeric phases such as polydimethylsiloxane
(PDMS) and polyacrylate (PA) extract by absorption and have been widely used
for water analysis. Solid sorbents such as carboxene or polystyrene-divinyl-
benzene (PS-DVB) extract by adsorption [341]. If extraction occurs by absorp-
tion into the liquid polymeric film (until a constant concentration in the bulk of
the polymer is reached), the extraction equilibrium of a given compound is only
described by its distribution coefficient and not influenced by other analytes.
Thus, even if an analyte is present in large excess it will not displace minor com-
ponents [52,327,343-346]. In contrast, SPME fibres coated with porous solid
sorbents extract by adsorption. They are sometimes preferred because of their
high sensitivity (and selectivity). As only a limited surface area is available for
adsorption, analytes will compete at higher concentrations for the adsorption
sites. This causes a displacement of compounds with poor affinities towards the
sorbent at longer extraction times. In addition, a non-linear extraction yield will
be observed at higher concentrations. Porous particles such as carboxene and
PS-DVB are often embedded in other partially cross-linked phases, such as
PDMS or carbowax (CW). A carbowax-templated resin is used for SPME-LC
applications [339].
22.5.4 Calibration
751
last method the distribution coefficient Kpg between the polymeric and the gas
phase can be determined based on GC retention times [347,348], while the
coefficients for the polymer-water phase (Henry's constant) can be obtained
from physicochemical tables [349]. For solid extraction phases, these calibration
modes are only applicable at low concentrations, i.e., within the linear range of
the adsorption isotherm. At higher concentrations, a calibration based on the
diffusion constant is possible [324].
22.5.5 Derivatization
A coupling of SPME with GC is easier than with LC, but restricted to compounds
amenable to GC (i.e. compounds of sufficient thermal stability and volatility).
The range of compounds which can be analyzed by SPME-GC can be extended if
non-volatile compounds are derivatized. Derivatization can be done in the
aqueous sample (i.e. prior to extraction), in the injector, and directly in the fibre
coating [350-360].
752
Inorganic salts. It was observed repeatedly that salt addition to aqueous
samples enhances the amount of analyte extracted by SPME [327,343-346,
362,364,366,367] as it is also observed in other extraction methods (LLE, SPE).
In general, the effect of salt addition increases with the polarity of the compound
[343]. The effect is observed both for direct and for headspace SPME [327].
Organic solvents, lower alcohols. For method validation, aqueous samples
were often spiked with reference compounds dissolved in polar volatile solvents.
It was demonstrated for methanol both for direct SPME and for headspace
SPME that the extraction yield only decreases significantly at methanol concen-
trations of >5% [327,343,345,368-370].
Natural organic matter. Humic compounds (natural organic matter, NOM)
are often present in groundwater of concentrations of > 1 mg/l, i.e., in substan-
tial excess to anthropogenic pollutants. Thus, displacement reactions during the
absorption of analytes may be anticipated. However, studies with pesticides and
other compounds in aqueous samples did not reveal a pronounced effect of
humic compounds on the extraction of other pollutants, provided the concentra-
tion did not exceed 10 mg/l [343,345,371,372]. The impact of natural organic
matter in aqueous samples on the extraction can be completely eliminated by
using headspace SPME.
22.5.9 Applications
753
Fig. 22.4. Schematic set-up of automatic analyzer based on solid-phase microextraction: 1,
sample preparation vessel; 2, peristaltic pump; 3, magnetic stirrer; 4, pH electrode; 5, level
switch; 6, programmable burette (titration); 7, progammable burette (internal standard); 8 and
9, peristaltic pumps; 10, extraction vessel) [2571.
were reported for pesticides that are amenable to GC. Phenoxy acid herbicides
are derivatized prior to GC analysis. Moreover, polar pesticides were analyzed by
LC after SPME. Many studies have been published on the SPME-GC analysis of
volatile organic compounds (VOCs), including halogenated hydrocarbons and
aromatic hydrocarbons (fuel related compounds), where headspace SPME is
often preferred to direct SPME.
754
TABLE 22.2
Applications of solid-phase microextraction (SPME) to the analysis of aqueous samples
Compound class (type of Instrumental Special methods Ref.
agent) analysis
Pesticides Review 374
GC 343, 366, 370,
375-400, 473, 478
GC Derivatization 401-403
LC 404,405
Phenols GC 354, 362, 406, 408
GC Headspace 353, 365, 407
Polycyclic aromatic GC 300, 328, 337,
hydrocarbons 409-411,413, 415,474
LC 412
Aromatic hydrocarbons, fuel GC 414-421
related compounds GC Headspace 300, 328, 329, 422-425
IR 426
Volatile organic compounds, GC 427-431, 456, 477
chlorinated hydrocarbons GC Headspace 361, 432-435
IR 436
Volatile polar compounds GC (headspace) 437, 438
(solvents)
Amines GC 439, 440
Derivatization 355
Nitroaromatic compounds, GC 441
explosives IR 442
Off-odour compounds GC 443, 444
Organometallic compounds GC (AED) 358, 359, 445-447
(Hg, Mn, Pb, Sn) GC (AED, ICP) Derivatization, 448-450
headspace
LC 451
Inorganic ions GC 452
Complexation 453
Polychlorinated biphenyls GC 410,411,454, 455
Headspace 338
Surfactants LC (Derivatization) 457, 458
Carbonyl compounds GC Derivatization, 459, 460
headspace
Hydroxyaromatics LC 461
GC 462
Miscellaneous GC 463-467, 469-471
LC 468, 472
Biologically active substances GC 475
Phthalates GC 476
two (liquid) phases are in direct physical contact through the pores. Separation
with these membranes is only based on size exclusion, where sufficiently small
molecules can permeate through the pores and larger ones cannot. Non-porous
membranes consist of a liquid or polymer film. Analyte molecules must dissolve
755
into this film to be able to pass through. Thus, separation with non-porous
membranes is based on the same principles as in LLE [480].
With porous membranes all small molecules may pass the membranes. With
non-porous membranes, however, one exploits the different physico-chemical
properties of the varying (small) analytes to achieve a separation between the
different compounds in much the similar way as in LLE (even if the analyte
molecules are equal in size).
Thus, while porous membranes are mainly used for sample cleanup to, e.g.,
remove proteins from biological samples, non-porous membranes allow the
extraction and enrichment of a sample and, in addition, a cleanup.
Depending on the driving force, three sample preparation techniques use
porous membrane: (a) dialysis (a concentration-driven process), (b) electro-
dialysis (an electric potential-driven process) and (c) filtration (a pressure-
driven process) [480]. Filtration is an important step in water analysis, as many
standard methods require that water samples be filtered before extraction to
exclude particles (and, unfortunately, also hydrophobic organics bound to these
particles). The other two techniques have been mainly applied to biological and
food samples. They are not further discussed here; the reader is referred to a
review 480] where these methods (including filtration) are described and
discussed in detail. Dialysis may in principle be used for cleanup of water
samples, i.e., to remove humic substances from environmental samples. For this
purpose, membranes with much smaller pores than those used for biological
samples have to be applied [481]. This approach may be promising in particular
if analyte enrichment is achieved in a second step, i.e., by solid-phase extraction,
but does not seem to have been used so far.
756
particularly suited to the extraction of ionizable organic compounds and uses the
following principle: in the donor phase (the sample), the analytes are first
transformed into their neutral extractable form using a pH adjustment. They
are then transported through the liquid membrane into the acceptor phase,
where they are transformed into their ionic form by further proper pH adjust-
ment and thus prevented from re-entering the membrane, i.e., the analytes are
trapped in the acceptor phase. For a complete trapping, the pH of the acceptor
phase should be at least three units lower than the pKa of the analyte. Neutral
compounds may be extracted by SLM, but will never be trapped in the acceptor
phase and no enrichment is achieved.
If organic compounds are present in their ionic form over a large range of pH
values, an addition of an ion pairing agent allows the extraction in the neutral
form, as shown for anionic surfactants [488]. As the extract is aqueous, the SLM
technique is best compatible with reversed phase LC or ion chromatography
(IC).
Microporous membrane liquid-liquid extraction (MMLLE). With the
MMLLE technique, the acceptor phase is an organic solvent and the same
solvent forms the liquid membrane by filling the pores of a porous hydrophobic
membrane [482,483]. In the case of water analysis the aqueous sample repre-
sents the donor phase. MMLLE is particularly suited to the extraction of highly
hydrophobic compounds which are readily extracted from water into an organic
solvent. In contrast to the SLM approach, they are not back-extracted into a
second aqueous phase. Thus, MMLLE is based on the same principles as LLE
but usually performed in a flowing system. In contrast to SLM, the extract ends
up in an organic phase and is thus best compatible with GC analysis. The
extraction efficiency in MMLLE is determined by the distribution coefficient
between water and the organic solvent.
Polymeric membrane extraction (PME). Moreover, a non-porous polymeric
membrane (such as a silicon rubber) may also be used for extraction. In water
analysis, the donor phase is the aqueous sample and the acceptor phase in
general an organic solvent (although a system, i.e., aqueous phase/membrane/
aqueous phase, is also conceivable). As the organic solvent penetrates into the
polymer causing it to swell considerably, the situation is similar to MMLLE. The
extraction efficiency is determined by the analyte-membrane distribution coeffi-
cient. As diffusion coefficients in polymers are lower than in liquids the mass
transfer is lower than in MMLLE [482].
Two further special membrane extraction methods are described in Section
22.6.4.
757
E = CA/Cw (22.13)
where CA and Cw are the concentrations in the acceptor phase and the aqueous
sample, respectively [482]. In MMLLE and PME, the maximum enrichment
factor for an aqueous sample, Ee(,,, is equal to the distribution constant, K5, of a
given analyte between water and the organic solvent used as acceptor phase, i.e.,
the equilibrium concentration C(eq) of the analyte in both phases is given as
758
22.6.4 Special membrane techniques
Membrane extraction with sorbent interface (MESI). While the membrane tech-
niques discussed so far use both a liquid donor and a liquid acceptor phase, the
MESI technique uses a gaseous acceptor phase. This technique was developed by
Pawliszyn and coworkers [311,490-495] and is described in detail in Chapter 14.
MESI systems consist of a membrane (usually a silicon hollow fibre) in which the
analytes are extracted from the surrounding liquid or gaseous sample. A flow of
inert gas transports the analytes through a cool sorbent trap where they are
enriched. In a subsequent step, they are eluted by thermal desorption and trans-
ferred to the GC. This technique is particularly well suited to the enrichment of
trace organics in air, but also for the analysis of volatiles in water. As discussed
below, for water analysis sampling of the headspace above the water is preferred
[494, 495] (see Section 22.6.6).
Membrane inlet mass spectrometry (MIMS). In membrane inlet mass spectro-
metry, the organic compounds diffuse from the aqueous sample through a non-
porous hydrophobic polymer membrane (usually a silicon membrane) into the
vacuum of the ion source of a mass spectrometer. This hydrophobic membrane
favors non-polar organic compounds which dissolve well in such a membrane.
The MIMS technique was pioneered by Cooks [496] who introduced the
concept of a direct insertion membrane probe. With this technique the mem-
brane is inserted directly into the mass spectrometer and the aqueous sample
transported to the membrane inlet. The technique was reviewed [497-506] and
is described in detail in Chapter 16 in which the theory of membrane extraction
is also presented. As discussed in Chapter 16, the entire extraction process
consists of three steps: (a) solvation of the analyte in the membrane, (b) trans-
port through the membrane and, finally, (c) evaporation from the membrane
surface into the mass spectrometer. Further parameters determining the
analyte transport through the membrane are the diffusion constant in the
membrane and the distribution coefficient between the aqueous and the
membrane phase, whereby the distribution coefficient is the more important
parameter. This distribution coefficient varies significantly from compound to
compound depending on the solubility in the membrane phase. The MIMS
technique has a high selectivity for hydrophobic compounds.
Applications of the MIMS technique were recently reviewed [500,501]. The
MIMS technique has mainly been applied to water analysis where measure-
ments of volatile organic compounds are standard applications. More recently is
has been shown that also semi-volatile organic compounds (SVOC) can be
determined (see Table 22.3).
Many of these VOCs can be measured at the ng/1 level directly from the aque-
ous phase while detection limits for SVOCs with standard MIMS techniques are
much higher.
It should be emphasised that MIMS does not include any chromatographic
separation steps which is a limiting factor if complex mixtures of pollutants,
759
TABLE 22.3
Applications of membrane techniques to the analysis of organic compounds in water
Compound class (type of agent) Extraction method Instrumental Ref.
analysis
Volatile organic compounds (VOCs) MIMS MS 496-506, 517,
518, 522-526
Phenols MIMS MS 527
Dicarboxylic acids MIMS DCI-MS 528
Steroid hormones MIMS DCI-MS 529
Hormones, biomolecules MIMS DCI-MS 530
Polycyclic aromatic hydrocarbons, MIMS DCI-MS 531
estrogenic comp., pesticides
Semivolatile compounds MIMS (trap,release) MS 532
Volatile, semivol. compounds MIMS (trap,release) MS 533
Volatile, semivol. compounds MIMS GC-MS 534
Volatile organic compounds MESI GC-MS 535
Pesticides (triazines) SLM LC 536-541
(phenoxy acids) SLM LC 542-544
(sulfonylurea) SLM LC 545-546
Phenols SLM LC 547
Anionic surfactants SLM LC 548
Anilines SLM LC 549
Pesticides (triazines) MMLLE FIA 550
Organotin MMLLE GC-MS 551
Cationic surfactants MMLLE LC 552
Other pesticides MMLLE LC 553, 554
Chlorinated organics PME GC 555
Neutral organics PME GC 556
Semivolatile organics PME LC 557
SLM = supported liquid membrane; MMLLE = microporous membrane liquid-liquid extraction;
PME = polymeric membrane extraction; MIMS = membrane inlet mass spectrometry; DCI =
desorption chemical ionization; MESI = membrane extraction with sorbent interface.
22.6.5 Automation
760
sis: (a) sampling and analyte enrichment are done in the field (i.e. at a remote
site) while further cleanup (if necessary) and instrumental analysis are carried
out in a central laboratory; (b) sampling, enrichment and instrumental analysis
are performed on-site in the field.
The first approach is realised with passive membrane-based sampling
devices which have been developed for ultra-trace enrichment of hydrophobic
pollutants, i.e., for air, soil and water analysis. Here, only aqueous samples are
considered. They use the approach of polymeric membrane extraction (PME).
The samplers have been termed semi-permeable membrane devices (SPMD) and
were developed by Petty and coworkers [508-514] and have been reviewed [507].
They mimic the bioaccumulation of hydrophobic micropollutants by organisms
containing lipid compartments. Standard SPMD devices are commercially avail-
able and consist of a lay flat non-porous low density polyethylene membrane
(standard size tube: 106 cm long, 2.4 cm wide, 75-90 m thick), which contains 1
ml triolein (glyceroltrioleate) as acceptor phase. Only neutral compounds with a
molecular weight of < 600 Da diffuse through the membrane, thus precluding, in
the case of water samples, the sampling of humic substances and compounds
bound to suspended particles. The amount of accumulated analytes is propor-
tional to the concentration of the dissolved (and thus bioavailable) fraction in
water. The capacity is related to the compound's octanol-water partition coeffi-
cients, K,: the higher the K,,, the greater the capacity of the SPMD for that
chemical, where the device is only attractive for compounds with log Ko > 3, i.e.
for compounds with high bioaccumulation potential. They are typically employ-
ed in environmental water (such as river or lake water) for a period of 14-30
days, i.e. they sample water-bound contaminants integratively. The subsequent
ultra-trace analysis in a central laboratory involves a substantial cleanup, i.e.
organic solvent dialysis, size exclusion chromatography, fractionation on
Florisil, silica gel or alumina and, if necessary, additional reversed-phase chro-
matography before instrumental analysis by, e.g., GC/MS. The concentration of
water-borne pollutants is given by
Cw = CL VL/R t (22.16)
where C, is the concentration of the analyte in water, CL that in the lipid triolein,
VL is the volume of triolein, R is the sampling rate and t the sampling time.
Further details for the estimation of the ambient water concentrations have
been presented [508-514]. Moreover, pitfalls (such as biofouling) and interfer-
ences (coextraction of polyethylene oligomers, oleic acid and methyloleate) have
been discussed [507].
In a second approach, membrane extraction techniques have also been
employed for direct on-site monitoring in the field. Such on-site monitoring is
readily achieved if these techniques are on-line coupled with LC (MS) or GC
(MS) or direct with a MS. Mobile or even portable GC and GC/MS systems are
commercially available and used for such on-site monitoring. Polymeric mem-
brane extraction (PME) with an inert gas as acceptor phase is particularly suited
761
to air analysis because of the higher diffusion coefficients of organic compounds
in the gas phase when compared to the liquid one. However, it is also well suited
to on-site monitoring of organic pollutants in water (in particular volatile
compounds). In this case, it is advantageous to sample the headspace of water
samples, as proposed for the MESI technique using a cup immersed in the water
[494,495]. The MIMS technique is also particularly well suited to on-site and
on-line monitoring of organic pollutants in water [515]. For these on-site mea-
surements, a MIMS was placed in a mobile laboratory [516] or a boat [517] for
on-line measurements of VOCs in river water. The method is also suited to
on-site waste water monitoring in chemical plants [518]. In addition, membrane
inlet systems coupled with GC/MS are employed [519,520], e.g., for below
surface measurements of marine pollution. For a direct field analysis of a water
stream a membrane extraction flow cell (hollow fibre) was connected to a mobile
gas chromatograph after enrichment onto a sorption tube [521] in a similar way
to that realised with the MESI technique.
22.6.7 Applications
Natural organic matter (NOM), such as humic and fulvic acids, may influence
the extraction process, as discussed in the preceding sections, and may also
interfere with the LC analysis with UV detection, leading to a broad unresolved
hump at the beginning of the chromatogram. In particular, with LLE and SPE
using conventional sorbents (e.g., C sorbents), a substantial fraction of the
NOM is coextracted, especially if low pH values are used for extraction, while
with solid-phase microextraction or headspace extraction, only a small part or
none of this NOM is extracted. With non-porous polymer membrane techniques,
762
high molecular weight NOM cannot diffuse through the membrane, while with
microporous membrane extraction methods, a size exclusion of NOM is achieved
along with a proper selection of the pore size. MMLLE methods combine size
exclusion with a (selective) extraction of smaller analytes. This approach is also
realized in solid-phase extraction using restricted access material as discussed
above.
If a size exclusion is not part of the extraction process itself, high molecular
weight NOM can be removed from the aqueous sample prior to extraction by
conventional size exclusion chromatography (SEC).
Column switching methods using the heart cut principle are used for automated
sample cleanup of many different types of samples including water samples.
Thus, in coupled column LC (LC-LC), after sample extraction, e.g., by LLE a
coarse separation of the analytes is achieved on a first column. Using switching
valves the fraction containing the analytes of interest is separated and automati-
cally transferred onto a second column (e.g., of different polarity) for final
analysis. A similar approach is used in LC-GC coupling [558]. With normal phase
LC, this approach is straightforward and robust [559-561], as an organic solvent
is finally injected into the GC. The direct coupling of reversed phase LC with GC
is more difficult as the LC heart cut contains a high fraction of water which is
problematic for the subsequent GC analysis, as normal retention gaps cannot
handle this water. The method of normal phase LC-GC, as applied to aqueous
samples, has been critically reviewed by Louter et al. [170], who concluded that
serious limitations exist with this approach.
Although column switching modes are well established in, e.g., pesticide
residue analysis of food, they are less well established in routine analysis of
water samples. Coupled column LC has recently been used for pesticide analysis
in natural water [562-567].
763
chromatography. Fractionation of these neutral compounds is achieved by elu-
tion with solvents or solvent mixtures of increasing polarity. In environmental
analysis, traditional column chromatography, well known from natural
compound isolation, can be replaced by using prepacked normal phase SPE
cartridges.
This sample cleanup is particularly suited to non-target analysis of moder-
ately polluted aqueous samples. However, it is rather time-consuming (even if
the normal phase fractionation is done automatically, e.g., with automated
fraction collection, or by applying preparative LC). Highly selective sorbents for
SPE such as immunoaffinity sorbents and molecular imprinted polymers will
facilitate this sample cleanup in target analysis (see Section 22.4.6).
REFERENCES
764
24 Th.A. Brettell and R.L. Grob, Part One, Int. Lab., Nov./Dec. (1985) 14.
25 Th.A. Brettell and R.L. Grob, Part Two, Int. Lab., April (1986) 30.
26 J.F. Pankow, J. High Resolut. Chromatogr., Chromatogr. Commun., 6 (1983) 292.
27 J.F. Pankow and M.E. Rosen, J. High Resolut. Chromatogr., Chromatogr. Commun.,
7 (1984) 504.
28 J.F. Pankow, J. High Resolut. Chromatogr., Chromatogr. Commun., 9 (1986) 18.
29 J.F. Pankow, J. High Resolut. Chromatogr., Chromatogr. Commun., 10 (1987) 409.
30 X.-P. Lee, T. Kumazawa, K. Sato, K. Watanabe, H. Seno and 0. Suzuki, Analyst, 123
(1998) 147.
31 K. Watanabe, H. Seno, A. Ishii and 0. Suzuki, Anal. Chem., 69 (1997) 5178.
32 R. E. Kaiser, Anal. Chem., 45 (1973) 965.
33 J.A. Rijks, J. Drozd and J. Novak, J. Chromatogr., 186 (1979) 167.
34 B. Kolb, B. Liebhard and L.S. Ettre, Chromatographia,21 (1986) 305.
35 P. Werkhoff and W. Bretschneider, J. Chromatogr., 405 (1987) 87.
36 J.F. Pankow, Environ. Sci. Technol., 25 (1991) 123.
37 J.W. Cochran and J.M. Henson, J. High Resolut. Chromatogr., Chromatogr.
Commun., 11 (1988) 869.
38 P.G. Simmonds, J. Chromatogr., 289 (1984) 117.
39 M. Shimoda and T. Shibamoto, J. High Resolut. Chromatogr., Chromatogr.
Commun., 13 (1990) 518.
40 Th. Noij, A. van Es, C. Cramers and J. Rijks, J. High Resolut. Chromatogr., Chroma-
togr. Commun., 10 (1987) 60.
41 B.B. Baker, Jr., Am. Ind. Hyg. Assoc. J., (1974) 735.
42 J.W. Cochran, J. High Resolut. Chromatogr., Chromatogr.Commun., 11 (1988) 663.
43 B. Kolb, G. Zwick and M. Auer, J. High Resolut. Chromatogr., 19 (1996) 37.
44 G.A. Junk, J.J. Richard, M.D. Grieser, D. Witiak, J.L. Witiak, M.D. Angello, R. Wick,
H.J. Svec, J.S. Fritz and G.V. Calder, J. Chromatogr., 99 (1974) 715.
45 E.M. Thurman and M.S. Mills, Solid-Phase Extraction-Principlesand Practice.
Wiley, New York, 1998.
46 N.J.K. Simpson, Solid Phase Extraction-Principles, Strategies and Applications.
Marcel Dekker, New York, 1998.
47 I. Liska, J Chromatogr. A, 885 (2000), 3.
48 M.C. Hennion, J. Chromatogr.A, 856 (1999)3.
49 H. Sabik, R. Jeannot and B. Rondeau, J. Chromatogr.A, 885 (2000) 217.
50 A. Balinova, J. Chromatogr.A, 754 (1996) 125.
51 M.E. Leon-Gonzalez and L.V. Peres-Arribas, J. Chromatogr.A, 90 2621.
52 C. Arthur and J. Pawliszyn, Anal. Chem., 62 (1990), 2145.
53 M.C. Hennion, C. Cau-Dit-Coumes and V. Pichon, J. Chromatogr.A, 823 (1998) 147.
54 D. Barcelo, G. Durand, 40 v. Bouvot and M. Neilen, Environ. Sci. Technol., 27 (1993)
271.
55 D.D. Blevins and D.O. Hall, LC-GC Int., September (1998) 17-30.
56 R.E. Majors, LC-GC Int., September (1988) 8.
57 P. Subra, M.-C. Hennion, R. Rosset and R.W. Frei, J. Chromatogr., 456 (1988) 121.
58 C.E. Werkhoven-Goewie, U.A.Th. Brinkman and R.W. Frei, Anal. Chem., 53 (1981)
2072.
59 W. Golkiewicz, C.E. Werkhoven-Goewie, U.A.Th. Brinkman, R.W. Frei, H. Colin and
G. Guiochon, J. Chromatogr. Sci., 21 (1983) 27.
60 R. Ferrer, J.L. Beltran and J. Guiteras, Anal. Chim. Acta, 346 (1997) 253.
61 M.L. Larrivee and C.F. Poole, Anal. Chem., 66 (1994) 139.
62 K.G. Miller and C.F. Poole, J. High Resolut. Chromatogr., 176 (1994) 125.
63 C.F. Poole, S.K. Poole, D.S. Seibert and C.M. Chapman, J. Chromatogr.B, 689 (1997)
245.
765
64 D. Seibert and C.F. Poole, Chromatographia,41 (1995) 245.
65 C.F. Poole, A.D. Gunatilleka and R. Sethuraman, J. Chromatogr.A, 885 (2000), 17.
66 T. Braumann, G. Weber and L.H. Grimme, J. Chromatogr., 261 (1983) 329.
67 T. Braumann, J. Chromatogr., 373 (1986) 191.
68 G. Machtalere, V. Pichon and M.C. Hennion, J. High Res. Chromatogr., 23 (2000),
437.
69 R.D. McDowall, LC-GC Int., 7 (1994) 638.
70 P. Mussmann, K. Levsen and W. Radeck, FreseniusJ. Anal. Chem., 348 (1994) 654.
71 A.B. Scholten, H.A. Claessens, J.W. de Haan and C.A. Cramers, J. Chromatogr. A,
759 (1997) 37.
72 J. Nawrocki, J. Chromatogr.A, 779 (1997) 29.
73 R.J.M. Vervoort, M.W.J. Derksen and A.J.J. Debets, J. Chromatogr. A, 765 (1997)
157.
74 D. Puig and D. Barcelo, Chromatographia,40 (1995) 435.
75 D. Barcelo, S. Chiron, S. Lacorte, E. Martinez, J.S. Salau and M.-C. Hennion, Trends
Anal. Chem., 13 (1994) 352.
76 L.H. Keith (Ed.), Compilation of EPA's Sampling and Analysis Methods, 2nd edn.
Lewis Publishers, Boca Raton, FL, 1996.
77 C.W. Huck and G.K. Bonn, J. Chromatogr.A, 885 (2000), 51.
78 S. Guenu and M.-C. Hennion, J. Chromatogr.A, 737 (1996) 15.
79 V. Pichon, Analusis, 26 (1998) 91.
80 C. Molina, P. Grasso, E. Benfenati and D. Barcelo, J. Chromatogr.A, 737 (1996) 47.
81 M. Castillo, D. Puig and D. Barcelo, J. Chromatogr.A, 778 (1997) 301.
82 A. Junker-Buchheit and M. Witzenbacher, J. Chromatogr. A, 737 (1996) 67.
83 C. Crespo, R.M. Mare6 and F. Burrull, J. Chromatogr. A, 670 (1994) 135.
84 F. Hernandez, C. Hidalgo, JV. Sancho and F.J. Lopez, Anal. Chem., 70 (1998) 3322.
85 J, Hadgeson, J. Collins and W. Bashe, J. Chromatogr.A, 659 (1994) 395.
86 A.M. Kvistad, E. Lundanes and T. Greibrokk, Chromatographia,48 (1998) 707.
87 C. Aguilar, F. Borrull and R.M. Marc6, J. Chromatogr.A, 771 (1997) 221.
88 T.A. Albanis, D.G. Hela, T.M. Sakellarides and I.K. Konstantinou, J. Chromatogr.A,
823 (1998) 59.
89 I. Tolosa, J.W. Readman and L.D. Mee, J. Chromatogr.A, 725 (1996) 93.
90 M.C. Hennion, C. Cau-Dit-Coumes and V. Pichon, J. Chromatogr. A, 823 (1998) 93.
91 M.-C. Hennion and V. Pichon, Environ. Sci. Technol., 28 (1994) 576 A.
92 M.L. Mayer, S.K. Poole and C.F. Poole, J. Chromatogr.A, 697 (1995) 89.
93 M.-C. Hennion and V. Coquart, J. Chromatogr., 642 (1993) 211.
94 D. Bolliet and C.F. Poole, Chromatographia,46 (1997) 381
95 N.Masque, R.M, Marce and F. Borrull, Trends Anal. Chem., 17 (1998) 384.
96 L. Schmidt, J.J. Sun, J.S. Fritz, D.F. Hagen, C.G. Markell and E.E. Wisted, J.
Chromatogr., 641 (1993) 57
97 T.K. Chambers and J.S. Fritz, J. Chromatogr.A, 797 (1998) 139.
98 N. Masqu6, M. Galia, R.M. Marc6 and F. Borrull, J. Chromatogr.A, 803 (1998) 147.
99 N. Masqu6, R.M. Marc6 and F. Borrull, J. Chromatogr.A, 793 (1998) 257.
100 N. Masqu6, M. Galia, R.M. Marc6 and F. Borrull, J. Chromatogr.A, 771 (1997) 55.
101 E.S.P. Bouvier, P.C. Iraneta, U.D. Neue, P.D. McDonald, D.J. Phillips, M. Capparella
and Y.F. Cheng, LC-GC int., September (1998) 35.
102 M.C. Hennion, J. Chromatogr.A, 885 (2000), 73
103 A. Di Corcia, C. Crescenzi, A. Marcomini and R. Samperi, Environ. Sci. Technol., 32
(1998) 711.
104 M.-C. Hennion, V. Coquart, S. Guenu and C. Sella, J. Chromatogr.A, 712 (1995) 287.
105 A. Di Corcia and M. Marchetti, Environ. Sci. Technol., 26 (1992) 66.
106 A. Di Corcia, S. Marchese and R. Samperi, J. Chromatogr., 642 (1993) 175.
766
107 A. Di Corcia, S. Marchese and R. Samperi, J. Chromatogr., 642 (1993) 163.
108 H. Sabik and R. Jeannot, J. Chromatogr.A, 818 (1998) 197.
109 A. Lagana, G. Fago and A. Marino, J. Chromatogr. A, 796 (1998) 309.
110 C. Crescenzi, A. Di Corcia, R. Samperi and M.A. Marcomini, Anal. Chem., 67 (1995)
1797.
111 A. Di Corcia and M. Marchetti, Anal. Chem., 63 (1991) 580.
112 G. D'Ascenzo, A. Gentili, S. Marchese, A. Marino and D. Perret, Chromatographia,
48 (1998) 497.
113 C. Crescenzi, A. Di Corcia, E. Guerriero and R. Samperi, Environ. Sci. Technol., 31
(1997) 479.
114 M.W.F. Nielen, U.A.Th. Brinkman and R.W. Frei, Anal. Chem., 57 (1985) 806.
115 E.R. Brouwer, J. Slobodnik, H. Lingeman and U.A.Th. Brinkman, Analusis, 20
(1992) 121.
116 F.T. Lange, M. Wenz and H.J. Brauch, J. High Resolut. Chromatogr., 18 (1995) 243.
117 T. Reemtsma, J. Chromatogr.A, 733 (1996) 473.
118 C. Sarzanini, M.C. Bruzzoniti, G. Sacchero and E. Mentasti, J. Chromatogr.A, 739
(1996) 63.
119 R.A. Gimeno, J.L. Beltran, R.M. Marce and F. Borrull, J. Chromatogr. A,., 890
(2000), 289.
120 M.S. Mills, E.M. Thurman and M.J. Pedersen, J. Chromatogr., 629 (1993) 11.
121 I.H. Hagestam and T.C. Pinkerton, Anal. Chem., 57 (1985) 1757.
122 K.K. Unger, Chromatographia,31 (1991) 507.
123 S. Vielhauer, A. Rudolphi, K.S. Boos, D. Seidel, J. Chromatogr.B, 666 (1995) 315.
124 A. Rudolphi, K.S. Boos and D. Seidel, Chromatographia,41 (1995) 645.
125 B.J. Gurley, M. Marx and K. Olsen, J. Chromatogr.B, 670 (1995) 358.
126 Z. Yu and D. Westerlund, J. Chromatogr. A, 725 (1996) 137.
127 Z. Yu and D. Westerlund, J. Chromatogr. A, 725 (1996) 149.
128 Z. Yu, D. Westerlund and K.S. Boos, J. Chromatogr. B, 698 (1997) 379.
129 K.S. Boos and A. Rudolphi, LC-GC Int., 15 (1997) 602.
130 P. Onnerfjord, D. Barcelo, J. Emnens, L. Gorton and G. Marko-Varga, J.
Chromatogr.A, 737 (1996) 35.
131 E. Hogendoorn and P. van Zoonen, J. Chromatogr. A, 892 (2000) 435.
132 A. Marx, T. Giersch and B. Hock, Anal. Lett., 28 (1995) 267.
133 D.H. Thomas, M. Beck-Westermeyer and D.S. Haga, Anal. Chem., 66 (1994) 3823.
134 S.J. Shahtaheri, M.F. Katmeh, P. Kwasowski and D. Stevenson, J. Chromatogr. A,
697 (1995) 131.
135 S.J. Shataheri, P.W. Kwasowski and D. Stevenson, Chromatographia,47 (1998) 453.
136 K.A. Bean and J.D. Henion, J. Chromatogr. A, 791 (1997) 119.
137 E. Matisova and S. Skrabakova, J. Chromatogr.A, 707 (1995) 145.
138 E. Schoenzetter, V. Pichon, D. Thiebaut, A. Fernandez-Alba and M.C. Hennion, J.
Microcolumn Sep., 12 (2000) 316.
139 S. Ben Rejeb, N.F. Durand, A. Martel, B. Le Poullennec, J.F. Lawrence, M.C.
Hennion and F. Le Goffic, Anal. Chim. Acta, 379 (1998) 41.
140 V. Pichon, L. Chen, M.-C. Hennion, R. Daniel, A. Martel, F. Le Goffic, J. Abian and D.
Barcelo, Anal. Chem., 67 (1995) 2451.
141 V. Pichon, L. Chen and M.-C. Hennion, Anal. Chim. Acta, 311 (1995) 429.
142 V. Pichon, L. Chen, N. Durand, F. le Goffic and M.-C. Hennion, J. Chromatogr.A, 725
(1996) 107.
143 V. Pichon, H. Rogniaux, N. Fischer-Durand, S. Ben Rejeb, F. Le Goffic and M.-C.
Hennion, Chromatographia,45 (1997) 289.
144 S. Ouyang, Y. Xu and Y.H. Chen, Anal. Chem., 70 (1998) 931.
145 M. Bouzige, V. Pichon and M.-C. Hennion, J. Chromatogr.A, 823 (1998) 197.
767
146 M. Bouzige, V. Pichon and M.-C. Hennion, Environ. Sci. Technol., 33 (1999) 1916.
147 M. Cichna, D. Knopp, R. Niessner, Anal. Chim. Acta, 339 (1997) 241.
148 A. Martin-Esteban, P. Kwasowski and D. Stevenson, Chromatographia,45 (1997)
364.
149 V. Pichon, M. Bouzige and M.-C. Hennion, Anal. Chim. Acta, 376 (1998) 21.
150 V. Pichon, M. Bouzige, C. Midge and M.C. Hennion, Trends Anal. Chem., 18 (1999)
219.
151 B. Sellergren, Anal. Chem., 66 (1994) 1578.
152 P. Martin, I.D. Wilson, D.E. Morgan and G.R. Jonas and K. Jones, Anal. Commun.,
34 (1997) 45.
153 M.T. Muldoon and L.H. Stanker, Anal. Chem., 69 (1997) 803.
154 J. Steinke, D.C. Sherrington and I.R. Dunkin, Adv. Polym. Sci., 123 (1995) 81.
155 B. Sellergren, Trends Anal. Chem., 18 (1999) 164.
156 D. Stevenson, Trends Anal. Chem., 18 (1999) 154
157 B. Sellergren, Trends Anal. Chem., 16 (1997) 310.
158 N. Masqu6, R.M. Marc and F. Borrull Trends Anal. Chem., 20 (2001) 477.
159 C. Baggiani, F. Trotta, G. Giraudi, C. Giovannoll and A. Vanni, Anal. Commun., 36
(1999) 263.
160 J. Matsui, M. Okada, M. Tsuruoka and T. Takeuchi, Anal. Commun., 34 (1997) 85.
161 J. Matsui, K. Fujiwara, S. Ugata and T. Takeuchi, J. Chromatogr.A, 889 (2000) 25.
162 I. Ferrer, P. Lanza, A. Tolokan, V. Horvath, B. Sellergren, O. Horval and D. Barcelo,
Anal. Chem., 72 (2000) 3934.
163 B. Bjarnason, L. Chimuka and 0. Ramstrom, Anal. Chem., 71 (1999) 2152.
164 I. Ferrer and D. Barcel6, Trends Anal. Chem., 18 (1999) 180.
165 N. Masque, R.M. Marc6, F. Borrull, P.A.G. Cormack and D.C. Sherrington, Anal.
Chem., 72 (2000) 4122.
166 D.T. Rossi and N. Zhang, J. Chromatogr. A, 885 (2000) 97.
167 G.A. Smith and T.L. Lloyd, LC-GC, 5 (1998) S22.
168 F.X. Diamond, W.E. Vickery and J. de Kanel, J. Anal. Toxicol., 20 (1997) 587.
169 E.L. Johnson, K.L. Hoffman and L.A. Pachla, in: J.R. Strimaitis and G. Hawk (Eds.),
Advances in LaboratoryAutomation Robotics. Zymark, Hopington, MA, 1988, p. 111.
170 A.J.H. Louter, J.J. Vreuls and U.A.Th. Brinkman, J. Chromatogr.A, 842 (1999) 391.
171 D. Barcelo and M.C. Hennion, Anal. Chim. Acta, 318 (1995) 1.
172 D. Barcelo and M.-C. Hennion, in: DeterminationofPesticides and theirDegradation
Productsin Water. Elsevier, Amsterdam, 1997, pp. 357-428.
173 I. Ferrer and D. Barcelo, J. Chromatogr. A, 778 (1997) 161.
174 J. Slobodnik, A.J.H. Louter, J.J. Vreuls, I. Liska and U.A.Th Brinkman, J. Chroma-
togr. A, 768 (1997) 239.
175 G.R. Mills, J. Chromatogr. A, 813 (1998) 63.
176 C. Hidalgo, J.V.C. Sancho, F.J. Lopez and F. Hernandez, J. Chromatogr. A, 823
(1998) 121.
177 R. El Harrak, M. Calull, R.M. Marce and F. Borrull, J. High Resolut. Chromatogr.,21
(1998) 667.
178 R.J. Vreeken, P. Speksnijder, I. Bobeldijk-Pastorova and Th.H.M. Noij, J. Chroma-
togr. A, 794 (1998) 187.
179 S. Guenu and M.C. Hennion, Anal. Methods Instrum., 2 (1995) 247.
180 D. Puig and D. Barcelo, J. Chromatogr.A, 778 (1997) 313.
181 R.W. Fedeniuk and P.J. Shand, J. Chromatogr.A, 812 (1998) 3.
182 N. Masqu6, E. Pocurull, R.M. Marc6 and F. Borrull, Chromatographia,47 (1998) 176.
183 E. Pocurull, R.M. Marce and F. Borrull, Chromatographia,41 (1995) 521.
184 J.M. Soriano, B. Jimenez, M.J. Redondo and J.C. Molto, J. Chromatogr.A, 822 (1998)
67.
768
185 M. Ibanez, Y. Pico and J. Manes, Chromatographia,45 (1997) 402.
186 S. Dupas, S. Guenu, V. Pichon, A. Montiel, B. Welte and M.C. Hennion, Int. J.
Environ. Anal. Chem., 65 (1996) 53.
187 V. Pichon and M.-C. Hennion, J. Chromatogr.A, 665 (1994) 269.
188 U.A.Th. Brinkman, J. Slobodnik and J.J. Vreuls, Trends Anal. Chem., 13 (1994) 373.
189 S. Lacorte and D. Barcelo, Anal. Chim. Acta, 296 (1994) 223.
190 S. Lacorte and D. Barcelo, J. Chromatogr.A, 725 (1996) 85.
191 S. Lacorte, N. Ehresmann and D. Barcelo, Environ. Sci. Technol., 29 (1995) 2834.
192 R. Reupert, G. Brausen and R. Schuster, Acta Hydrochem. Hydrobiol. 26 (1998) 318.
193 A. Farjam, G.J. de Jong, R.W. Frei, U.A.Th. Brinkman, W. Haasnoot, A.R.M.
Hamers, R. Schilt and F.A. Huf, J. Chromatogr., 452 (1988) 419.
194 A. Farjam, N.C. van de Merbel, H. Lingeman, R.W. Frei and U.A.Th. Brinkman, Int.
J. Environ. Anal. Chem., 45 (1991) 73.
195 I. Ferrer, V. Pichon, M.-C. Hennion and D. Barcelo, J. Chromatogr.A, 777 (1997) 91.
196 I. Ferrer, M.-C. Hennion and D. Barcelo, Anal. Chem., 69 (1997) 4508.
197 J. Henion, E. Brewer and G. Rule, Anal. Chem., 70 (1998) 650A.
198 K.-S. Boos, J. High Resolut. Chromatogr., 23 (2000) 272.
199 C.T. Fleischer, K.-S. Boos, F. Lanza and B. Sellergren, J. High Resolut. Chromatogr.
23 (2000) 284.
200 C.T. Fleischer, P. Chiap and K.-S. Boos, 1s t International Workshop on Molecularly
Imprinted Polymers, Cardiff, 2000.
201 R. Koeber, C.T. Fleischer, K.-S. Boos, D. Barcel6, F. Lanza, B. Sellergren and Mica
network, 1 t International Workshop on Molecularly Imprinted Polymers, Cardiff,
2000.
202 A.J.H. Louter, C.A. van Beckvelt, P. Cid Montanes, J. Seobodnik, J.J. Vreues and
U.A. Th. Brinkman, J. Chromatogr. A, 725 (1996) 67.
203 H.J. Cortes, C.D. Pfeiffer, G.L. Jewett and B.E. Richter, J. Microcol. Sep., 1 (1989)
28.
204 K. Grob Jr. and Z. Li, J. Chromatogr., 473 (1989) 423.
205 K. Grob, J. Chromatogr.A, 473 (1989) 73.
206 E. Noroozian, F.A. Maris, M.W.F. Nielen, R.W. Frei, G.J. de Jong and U.A.Th.
Brinkman, J. High Resolut. Chromatogr. Commun., 11 (1987) 17.
207 Th.H.M. Noij, E. Weiss, T. Herps, H. Van Cruchten and J. Rijks, J. High Resolut.
Chromatogr. Chromatogr. Commun., 11 (1988) 181.
208 J.J. Vreuls, W.J.G.M. Cuppen, E. Dolecka, F.A. Maris, G.J. de Jong and U.A.Th.
Brinkman, J. High Resolut. Chromatogr., 12 (1989) 807.
209 J.J. Vreuls, W.J.G.M. Cupen, G.J. de Jong and U.A.Th. Brinkman, J. High Resolut.
Chromatogr., 13 (1990) 157.
210 J.J. Vreuls, R.T. Ghijsen, G.J. de Jong and U.A.Th. Brinkman, J. Chromatogr., 625
(1992) 237.
211 Y. Pic6, J.J. Vreuls, R.T. Ghijsen and U.A.Th. Brinkman, Chromatographia, 38
(1994)461.
212 Y. Pic6, A.J.H. Louter, J.J. Vreuls and U.A.Th. Brinkman, Analyst 119 (1994) 2025.
213 E. Boselli, B. Grolimund, K. Grob, G. Lercker and R. Amadb, J. High Resolut.
Chromatogr., 21 (1998) 355.
214 B. Grolimund, E. Boselli, K. Grob, R. Amadb and G. Lercker, J. High Resolut.
Chromatogr., 21 (1998) 378.
215 H. Huang, J.R. Kagel and D.T. Rossi, J. Pharm.Biomed. Anal., 19 (1999) 613.
216 D. Barcelo, S. Chiron, S. Lacorte, E. Martinez, J.S. and M.C. Hennion, Trends Anal.
Chem., 13 (1994) 352.
217 S.A. Senseman, T.L. Lavy and J.D. Mattice, Anal. Chem., 67 (1995) 3064.
218 E. Martinez and D. Barcelo, Chromatographia,42 (1996) 72.
769
219 C. Crescenzi, A. Di Corcia, M.D. Madbouly and R. Samperi, Environ. Sci. Technol., 29
(1995) 2185.
220 G.A. Penuela and D. Barcelo, J. Chromatogr. A, 823 (1998) 81.
221 G. Carrera, P. Fernandez, R. Vilanova and J.O. Grimalt, J. Chromatogr. A, 823
(1998) 189.
222 G.A. Penuela and D. Barcelo, J. Chromatogr. A, 795 (1998) 93.
223 S. Lacorte, N. Ehresmann and D. Barcelo, Environ. Sci. Technol., 29 (1995) 2834.
224 E.Pocurull, L. Brossa, F. Borrull and R.M. Marce, J. Chromatogr.A, 885 (2000) 361.
225 J. Slobodnik, M.G.M. Groenewegen, E.R. Brouwer, H. Lingeman and U.A.Th.
Brinkman, J. Chromatogr., 642 (1993) 359.
226 S. Lacorte, J.J. Vreuls, J.S. Salau, F. Ventura and D. Barcelo, J. Chromatogr. A, 795
(1998) 71.
227 J.M. Soriano, B. Jimenez, G. Font and J.C. Molto, Crit. Rev. Anal. Chem., 31 (2001)
19.
228 M.J.M. Wells and L.Z. Yu, J. Chromatogr.A, 885 (2000) 237.
229 Y. Pico, G. Font, J.C. Molto and J. Manes., J. Chromatogr.A, 885 (2000) 251.
230 V. Pichon, J. Chromatogr.A, 885 (2000) 195.
231 R.M. Marce and F. Borrull, J. Chromatogr.A, 885 (2000) 273.
232 I.Rodriguez, M.P. Llompart and R. Cela, J. Chromatogr.A, 885 (2000) 291.
233 P.L. Ferguson, C.R. Iden and B.J. Brownawell, J. Chromatogr.A, 938 (2001), 79.
234 M. Petrovic and D. Barcelo, J. Mass Spectrom., 36 (2001) 1173.
235 S.H. Benomar, M.R. Clench and D.W. Allen, Anal. Chim. Acta, 445 (2001) 255.
236 M. Castillo, G. Penuela and D. Barcelo, Fresenius J. Anal. Chem., 369 (2001) 620.
237 P. de Voogt, O. Kwast, R. Hendriks and N. Jonkers, Analusis, 28 (2000) 776.
238 M. Petrovic and D. Barcelo, FreseniusJ. Anal. Chem., 368 (2000) 676.
239 P. Eichhorn, M.E. Flavier, M.L. Paje and T.P. Knepper, Sci. Total. Environ., 269
(2001) 75.
240 J. Riu and D. Barcelo, Analyst, 126 (2001) 825.
241 J. Riu, E. Martinez, D. Barcelo, A. Ginebreda and L. Tirapu, Fresenius J. Anal.
Chem., 371 (2001) 448.
242 C.A. Moody, W.C. Kwan, J.W. Martin, D.C.G. Muir and S.A. Mabury, Anal. Chem., 73
(2001) 2200.
243 W.H. Ding and Y.H. Liao, Anal. Chem., 73 (2001) 36.
244 I. Ferrer and E.T. Furlong, Environ. Sci. Technol., 35 (2001) 2583.
245 S. Hui-Feng, T. Hase, N. Hata, I. Kasahara and S. Taguchi, Anal. Sci., 17 (2001)
1291.
246 M.J.L. de Alda and D. Barcelo, J. Chromatogr.A, 938 (2001) 145.
247 M. Katayama, T. Sasaki, Y. Matsuda, S. Kaneko, T. Iwamoto and M. Tanaka,
Biomed. Chromatogr., 15 (2001) 403.
248 S.A. Snyder, D.L. Villeneuve, E.M. Snyder and J.P. Giesy, Environ. Sci. Technol., 35
(2001) 3620.
249 H.M. Kuch and K. Ballschmiter, Environ. Sci. Technol., 35 (2001) 3201.
250 X.Y. Xiao, D.V. McCalley and J. McEvoy. J. Chromatogr.A, 923, (2001) 195.
251 A. Lagana, G. Fabo, A. Marino and D. Santarelli, Anal. Lett., 34 (2001) 913.
252 S. Nakamura, T.H. Sian and S. Daishima, J. Chromatogr.A, 919 (2001) 275.
253 M.J.L. de Alda and D. Barcelo, J. Chromatogr. A, 911 (2001) 203.
254 Y. Ishii, S. Okita, M. Torigai and S.J. Yun, Bunseki Kagaku, 49 (2000) 753.
255 M.J.L. de Ala and D. Barcelo, J. Chromatogr.A, 892 (2000) 391.
256 H.M. Kuch and K. Ballschmiter, Fresenius J. Anal. Chem., 366 (2000) 391.
257 A. Lagana, A. Bacaloni, G. Fago and A. Marino, Rapid Commun Mass Spectrom., 14
(2000) 401.
258 S.A. Snyder, T.L. Keith, D.A. Verbrugge, E.M. Snyder, T.S. Gross, K. Kannan and
770
J.P. Giesy, Environ. Sci. Technol., 33 (1999) 2814.
259 M. Careri, L. Elviri and A. Mangia, J. AOAC Int., 84 (2001) 1383.
260 M.C. Alonso and D. Barcelo, J. Chromatogr.A, 889, (2000) 231.
261 M.C. Alonso and D. Barcelo, Anal. Chim. Acta, 400 (1999) 211.
262 M.C. Alonso, M. Castillo and D. Barcelo, Anal. Chem., 71 (1999) 2586.
263 R. Loos, J. Riu, M.C. Alonso and D. Barcelo, J. Mass Spectrom., 35 (2000) 1197.
264 E. Marengo, M.C. Gennaro and V. Gianotti, Chemometr. Intell. Lab. Syst., 53 (2000)
57.
265 C. Wolf, T. Storm, F.T. Lange, T. Reemtsma, H.J. Brauch, S.H. Eberle and M. Jekel,
Anal. Chem., 72 (2000) 5466.
266 R. Loos, M.C. Alonso and D. Barcelo, J. Chromatogr.A, 890 (2000) 225.
267 R.A. Gimeno, R.M. Marce and F. Borrull, Chromatographia,53 (2001) 22.
268 C.H. Liu and W.H. Ding, J. Chromatogr.A, 926 (2001) 341.
269 C. Creszenzi, A. Di Corcia, A. Marcomini, G. Pojana and R. Samperi, J. Chromatogr.
A, (2001) 97.
270 J. Kruppa, A. Preiss, K. Levsn and H.P. Kalbus, Acta Hydrochim. Hydrobiol., 24
(1996) 226.
271 C. Aguilar, I. Ferrer, F. Borrull, R.M. Marce and D. Barcelo, Anal. Chim. Acta, 386
(1999) 237.
272 A.C. Hogenboom, W.M.A. Niessen and U.A.T. Brinkman, J. Chromatogr. A, 841
(1999) 33.
273 H. Bagheri, J. Slobodnik and U.A.T. Brinkman, Anal. Lett., 33 (2000) 249.
274 R. Castro, E. Moyano and M.T. Garceran, J. Chromatogr. A, 869 (2000) 441.
275 T.C.R. Santos, J.C. Rocha and D. Barcelo, J. Chromatogr.A, 879 (2000), 3.
276 A.C. Hogenboom, W.M.A. Niessen and U.A.T. Brinkman, Rapid Commun Mass.
Spectrom., 14 (2000) 1914.
277 R. Castro, E. Moyano and M.T. Galceran, J. Chromatogr.A, 914 (2001), 111.
278 M.P.G. De Llasera and M. Bernal-Gonzalez, Water Res., 35 (2001) 1933.
279 N. Masque, M. Galia, R.M. Marce and F. Borrull, J. High Res. Chromatogr., 22 (1999)
547.
280 J. Patsias and E. Papadopoulou-Mourkidou, J. AOAC Int., 82 (1999) 968.
281 I. Ferrer and D. Barcelo, J. Chromatogr. A, 854 (1999) 197.
282 R.A. Gimeno, C. Aguilar, R.M. Marce and F. Borrull, J. Chromatogr.A, 915 (2001),
139.
283 E. Pocurull, C. Aguilar, M.C. Alonso, D. Barcelo, F. Borrull and R.M. Marce, J.
Chromatogr. A, 854 (1999) 187.
284 R. Wissiack, E. Rosenberg and M. Grasserbauer, J. Chromatogr.A, 896 (2000) 159.
285 J. Patsias and E. Papadopoulou-Mourkidou, J. Chromatogr.A, 904 (2000), 171.
286 C. Rivasseau, G. Vanhoenacker, P. Sandra and M.C. Hennion, J. Microcolumn Sep.,
12 (2000) 323.
287 P. Hinsmann, L. Arce, A. Rios and M. Valcarcel, J. Chromatogr.A, 866 (2000) 137.
288 A. Medvedovici, V. David and P. Sandra, Rev. Roum Chim., 45 (2000) 827.
289 J. Slobodnik, S. Ramalho, b.M. van Baar, A.J.H. Louter and U.A.T. Brinkman,
Chemosphere, 41 (2000) 1469.
290 E. Pocurull, L. Brossa, F. Borrull and R.M. Marce, J. Chromatogr.A, 885 (2000) 361.
291 M. Peruzzi, G. Bartolucci and F. Cioni, J. Chromatogr. A, 867 (2000) 169.
292 S. Lacorte, I. Guiffard, D. Fraisse and D. Barcelo, Anal. Chem., 72 (2000) 1430.
293 A.C. Hogenboom, M.P. Hofman, D.A. Jolly, W.M.A. Niessen and U.A.T. Brinkman, J.
Chromatogr. A, 885 (2000) 377.
294 E. Turiel, A Martin-Esteban, P. Fernandez, C. Perez-Conde and C. Camara, Anal.
Chem., 73 (2001) 5133.
295 A Martin-Esteban E. Turiel and D. Stevenson, Chromatographia,53 (2001) S 434.
771
296 R. Koeber, C. Fleischer, F. Lanza, K.S. Boos, B. Sellergren and D. Barcelo, Anal.
Chem., 73 (2001) 2437.
297 J. Pawliszyn, Solid-PhaseMicroextraction:Theory and PracticeVCH, New York, NY,
1997.
298 J. Pawliszyn (Ed.), Applications of Solid Phase Microextraction. Royal Society of
Chemistry, Cambridge, 1999.
299 A. Boyd-Boland, M. Chai, L. Luo, Z. Zhang, M. Yang, J. Pawliszyn and T. Gorecki,
Environ. Sci. Technol., 28 (1994) 569A.
300 Z. Zhang and J. Pawliszyn, Anal. Chem., 65 (1993), 1843.
301 C.L. Arthur, L. Killam, K.D. Buchholz, D. Potter, M. Chai, Z. Zhang and J. Pawliszyn,
Environ. Lab., 11 (1992) 10.
302 C.L. Arthur, D. Potter, K. Buchholz, S. Motlagh and J. Pawliszyn, LC-GC, 10 (1992)
656.
303 H. Lord and J. Pawliszyn, J. Chromatogr.A, 885 (2000) 153.
304 H. Prosen and L. Zupancic-Kralj, Trends Anal. Chem., 18 (199) 272.
305 J. Pawliszyn, Trends Anal. Chem., 14 (1995) 113.
306 Z. Zhang, M. Yang and J. Pawliszyn, Anal. Chem., 66 (1994) 844A.
307 T. Gorecki, A. Boyd-Boland, Z. Zhang and J. Pawliszyn, Can. J. Chem., 74 (1996)
1297.
308 B. Zygmunt, A. Jastrzebska and J. Namiesnik, Crit. Rev. Anal. Chem., 31 (2001) 1.
309 R. Eisert and K. Levsen, J. Chromatogr. A, 733 (1996) 143.
310 R. Eisert and J. Pawliszyn, Crit. Rev. Anal. Chem., 27, (1997) 103.
311 Y. Luo and J. Pawliszyn in: A.J. Handley (Ed.), Extraction Methods in Organic
Analysis, Sheffield Academic Press, Sheffield, 1998, Ch. 4, 75.
312 M.D. Alpendurada, J. Chromatogr.A, 889 (2000), 3.
313 N. Prosen and L. Zupancic-Kralig, Trends Anal. Chem., 18 (1999) 272.
314 A. Penalver, E. Pocurull, F. Borrull and R.M. Marce, Trends Anal. Chem., 18 (1999)
557.
315 J. Chen and J. Pawliszyn, Anal. Chem., 67 (1995) 2530.
316 P. Sandra, B. Tienpont, J. Vercammen, A. Tredoux, T. Sandra and F. David, J.
Chromatogr. A, 928 (2001) 117.
317 R. Eisert and J. Pawliszyn, Anal. Chem., 69 (1997) 69.
318 H. Kataoka, S. Narimatsu, H. Lord and J. Pawliszyn, Anal. Chem., 71 (1999) 4237.
319 J. Wu, H.J. Lord, J. Pawliszyn and H. Kataoka, J. Microcolumn Sep., 12 (2000), 255.
320 J. Wu, X.M. Yu, H.L. Lord and J. Pawliszyn, Analyst, 125 (2000) 391.
321 J.C. Wu, H. Lord and J. Pawliszyn, Talanta, 54 (2001) 655.
322 W.M. Mullett, P. Martin and J. Pawliszyn, Anal. Chem., 73 (2001) 2383.
323 Y. Saito, Y. Nakao, M. Imaizumi, T. Takeichi, Y. Kiso and K. Jinno, Fresenius J.
Anal. Chem., 368 (2000) 641.
324 S.N. Semenov, J.A. Koziel and J. Pawliszyn, J. Chromatogr.A, 873 (2000) 39.
325 J. Ai, Anal. Chem., 69 (1997) 3260.
326 J. Ai, Anal. Chem., 70 (1998) 4822.
327 E. Belau, C. Grote, M. Spiekermann and K. Levsen, Field Anal. Chem. Technol., 5
(2001) 37.
328 Z. Zhang and J. Pawliszyn, J. High Resolut. Chromatogr.16 (1993) 689.
329 B. MacGillivray, J. Pawliszyn, P. Fowlie and C. Sagara, J. Chromatogr. Sci., 32
(1994) 317.
330 Z. Zhang and J. Pawliszyn, Anal. Chem., 67 (1995) 34.
331 T. Gorecki and J. Pawliszyn, Analyst, 122 (1997) 1079.
332 T. Gorecki, A. Khaled and J. Pawliszyn, Analyst, 123 (1999) 2819.
333 D. Louch, S. Motlagh and J. Pawliszyn, Anal. Chem., 64 (1992) 1187.
334 S. Motlagh and J. Pawliszyn, Anal. Chim. Acta, 284 (1993) 265.
772
335 Z. Penton, H. Geppert and V. Betz, GIT Spez. Chromatogr., 2 (1996) 112.
336 R. Eisert and J. Pawliszyn, J. Chromatogr.A, 776 (1997) 293.
337 Z.Y. Zhang and J. Pawliszyn, Anal. Chem., 65 (1993), 1843.
338 B.D. Page and G. Lacroix, J. Chromatogr.A, 757 (1997), 173.
339 V. Mani in: J. Pawliszyn (Ed.), Applicationof Solid Phase Extraction. Royal Society of
Chemistry, Cambridge, Ch. 5, 57.
340 S.L. Chong, D. Wang, J.D. Hayes, B.W. Wilhite and A. Malik, Anal. Chem., 69 (1997)
3889.
341 F. Mangani and R. Cenciarini. Chromatographia,41 (1995) 678.
342 Z.Y. Wang, C.H. Xiao, C.Y. Wu and H.M. Han, J. Chromatogr. A, 893 (2000) 157.
343 R. Eisert and K. Levsen, J. Am. Soc. Mass Spectrom., 6 (1995), 1119.
344 P. Popp, S. Mothes and L. Briiggemann, Vomn Wasser, 85 (1995) 229.
345 C. Grote, E. Belau, K. Levsen and G. Wiinsch, Acta Hydrochim. Hydrobiol., 27 (1999)
193.
346 C. Grote and K. Levsen in: J. Pawliszyn (Ed.), Application of Solid Phase Extraction.
Royal Society of Chemistry, Cambridge, 1999, Ch. 12, 1.
347 B. Schafer, P. Hennig and W. Engewald, J. High Resolut. Chromatogr., 20 (1997),
217.
348 P.A. Martos, A. Saraullo and J. Pawliszyn, Anal. Chem., 69 (1997) 402.
349 B. Shurmer and J. Pawliszyn, Anal. Chem., 72 (2000) 3660.
350 L. Pan and J. Pawliszyn, Anal. Chem., 69 (1997) 196.
351 P.A. Martos and J. Pawliszyn, ACS Symp. Ser., 705 (1998), 92.
352 L. Pan, M. Adams and J. Pawliszyn, Anal. Chem., 67 (1995) 4396.
353 P. Bartak and L. Cap, J. Chromatogr.A, 767 (1997) 171.
354 K.D. Buchholz and J. Pawliszyn, Enuiron. Sci. Technol., 27 (1993) 2844.
355 L. Pan, J.M. Chong and J. Pawliszyn, J. Chromatogr.A, 773 (1997) 249.
356 R. Eisert and K. Levsen, GIT Fachz. Lab., 40 (1996) 581.
357 Y. Cai and J.M. Bayona, J. Chromatogr.A, 696 (1995) 113.
358 T. Gorecki and J. Pawliszyn, Anal. Chem., 68 (1996), 3008.
359 Y. Morcillo, Y. Cai and J.M. Bayona, J. High Resolut. Chromatogr., 18 (1995), 767.
360 M. Guidotti and M. Vitali, Ann. Chim., 87 (1997) 497.
361 P. Popp and A. Paschke, Chromatographia,46 (1997) 419.
362 M. Mder, S. Schrader, U. Franck and P. Popp, Fresenius J. Anal. Chem., 357 (1997)
326.
363 B. Schafer and W. Engewald, Fresenius'J. Anal. Chem., 352 (1995) 535.
364 L. Miller, E. Fattore and E. Benfenati, J. Chromatogr.A, 791 (1997) 221.
365 K. D. Buchholz and J. Pawliszyn, Anal. Chem., 66 (1994) 160.
366 A.A. Boyd-Boland, S. Magdic and J. Pawliszyn, Analyst, 121 (1996) 929.
367 J. Dewulf, H. van Langenhove and M. Everaert, J. Chromatogr.A, 761 (1997) 205.
368 C.L. Arthur, L.M. Killam, K.D. Buchholz, J. Pawliszyn and J.R. Berg, Anal. Chem.,
64 (1992) 1960.
369 R. Eisert and K. Levsen, G. Wiinsch, Vom Wasser, 86 (1996) 1.
370 L. Urruty and M. Montury, J. Agric. Food Chem., 44 (1996) 3871.
371 K.K. Chee, M.K. Wong and H.K. Lee, J. Microcolumn Sep., 8 (1996) 131.
372 S.D. Huang, C.P. Cheng and Y.H. Sung, Anal. Chim. Acta, 343 (1997) 101.
373 C. Grote, K. Levsen and G. Winsch, Anal. Chem., 71 (1999) 4513.
374 J. Beltran, FJ Lopez and F. Hernandez, J. Chromatogr.A, 885 (2000) 389.
375 R. Hu, D. Elia, JM Berthion and S. Poliak, Chromatographia,53 (2001) 306.
376 D. Lambropoulou, T. Sakellarides and T. Albanis, Fresenius J. Anal. Chem., 368
(2000) 616.
377 M.C Sampedro, O. Martin, C.Lopez de Armentia, M.A. Goicolea, E. Rodriguez, Z.
Gomez de Balugera, J. Costa-Moreira and R.J. Barrio, J. Chromatogr. A, 893 (2000)
773
347.
378 J. Lipinski, FreseniusJ. Anal. Chem., 367 (2000) 445.
379 D.A. Lambropoulou, I.K Konstantinou and T.A. Albanis, J. Chromatogr. A, 893
(2000) 143.
380 A.Penalver, E. Pocurull and R.M. Marce, J. Chromatogr.A, 839 (1999) 253.
381 R. Eisert and K. Levsen, Fresenius J. Anal. Chem., 351 (1995) 555.
382 I. J. Barnabas, J.R. Dean, I.A. Fowlis and S.P. Owen, J. Chromatogr. A, 705 (1995)
305.
383 A.A. Boyd-Boland and J. Pawliszyn, J. Chromatogr.A, 704 (1995) 163.
384 S. Magdic, A.A. Boyd-Boland, K. Jinno and J. Pawliszyn, J. Chromatogr.A, 736
(1996) 219.
385 R. Eisert, K. Levsen and G. Wiinsch, J. Chromatogr. A, 683 (1994) 175.
386 T. Gorecki, R. Mindrup and J. Pawliszyn, Analyst, 121 (1996) 1381.
387 T.K. Choudhury, K.O. Gerhardt and T.P. Mawhinney, Environ. Sci. Technol., 30
(1996) 3259.
388 R. Young, V. Lopez-Avila and W.F. Beckert, J. High Resolut. Chromatogr., 19 (1996)
247.
389 K.N. Graham, L.P. Sarna, G.R.B. Webster, J.D. Gaynor, and H.Y.F. Ng, J. Chroma-
togr. A, 725 (1996) 129.
390 M.T. Sng, F.K. Lee and H.A. Lakso, J. Chromatogr. A, 759 (1997) 225.
391 I. Valor, J.C. Molto, D. Apraiz and G. Font, J. Chromatogr.A, 767 (1997) 195.
392 V. Lopez-Avila, R. Young and W.F. Beckert, J. High Resolut. Chromatogr., 20 (1997)
487.
393 R. Ferrari, T. Nilsson, R. Arena, P. Arlati, G. Bartolucci, R. Basla, F. Cioni, G. Del
Carlo, P. Dellavedova, E. Fattore, M. Funi, C. Grote, M. Guidotti, S. Morgillo, L.
Miller and M. Volante, J. Chromatogr. A, 795 (1998) 371.
394 J. Dugay, C. Miege and M.C. Hennion, J. Chromatogr.A, 795 (1998) 27.
395 C. Aguilar, S. Penalver, E. Pocurull, F. Borrull and R.M. Marce, J. Chromatogr.A,
795 (1998) 105.
396 J. Beltran, F.J. Lopez, O. Cepria and F. Hernandez, J. Chromatogr.A, 808 (1998)
257.
397 G.P. Jackson and A.R.J. Andrews, Analyst, 123 (1998) 1085.
398 C. Miege and J. Dugay, Analusis, 26 (1998) M137.
399 A. Wenner and M. Wortberg, J. High Resolut. Chromatogr., 21 (1998) 661.
400 B.V Stolyarov and V.V Boiko, Russian J. Appl. Chem., 73 (2000) 1963.
401 T. Nilsson, D. Baglio, I. Galdo-Miguez, J.O. Madsen and S. Facchetti. J. Chromatogr.
A, 826 (1998) 211.
402 M.R. Lee, R.-J. Lee, Y.W. Lin, C.M. Chen and B.H. Hwang, Anal. Chem., 70 (1998)
1963.
403 T.Nilsson, D. Baglio, I. Galdo-Miguez, J.O. Madsen and S. Facchetti, J. Chromatogr.
A, 826 (1998) 211.
372 D.C Stahl and D.C Tilotta, Environ. Sci. Technol., 35 (2001) 3507.
404 S.H. Salleh, Y. Saito, Y. Kiso and K. Jinno, Anal. Chim. Acta, 433 (2001) 207.
405 K. Jinno, T. Muramatsu, Y. Saito, Y. Kiso, S. Magdic and J. Pawliszyn, J. Chroma-
togr. A, 754 (1996) 137.
406 M.R. Lee, Y.C. Yeh, W.S. Hsiang and B.H. Hwang, J. Chromatogr.A, 806 (1998) 317.
407 J. Porschmann, F.D. Kopinke and J. Pawliszyn, J. Chromatogr.A, 816 (1998) 159.
408 M. Guidotti and G. Ravaioli, Ann. Chim. (Rome), 88 (1998) 629.
409 R.A. Doong, S.M. Chang and Y.C. Sun, J. Chromatogr. Sci., 38 (2000) 528.
410 D.W. Potter and J. Pawiiszyn, Environ. Sci. Technol., 28 (1994) 298.
411 Y. Liu, M.L. Lee, K.J. Hageman, Y. Yang and S.B. Hawthorne, Anal. Chem., 69 (1997)
5001.
774
412 M.R. Negrao and M.F. Alpendurada, J. Chromatogr. A, 823 (1998) 211
413 A. Paschke, P. Popp and G. Schiiiirmann, FreseniusJ. Anal. Chem., 363 (1999) 426.
414 D.W. Potter and J. Pawliszyn, J. Chromatogr., 625 (1992) 247.
415 J.J. Langenfeld, S.B. Hawthorne and D.J. Miller, Anal. Chem., 68 (1996) 144.
416 C.L. Arthur, L.M. Killam, S. Motlagh, M. Lim, D. Potter and J. Pawliszyn, J.
Environ. Sci. Technol., 26 (1992) 979.
417 I. Valor, C. Cortada and J.C. Molto, J. High Resolut. Chromatogr., 19 (1996) 472.
418 S.P. Thomas, R. Sri Ranjan, G.R.B. Webster and L.P. Sarna, Environ. Sci. Technol.,
30 (1996) 1521.
419 M. Eriksson, A. Swartling and G. Dalhammar, Appl. Microbiol. Biotechnol., 50 (1998)
129.
420 A.A.M. Hassan, E. Benfenati, G. Facchini and R. Fanelli, Toxicol. Environ. Chem., 55
(1996) 73.
421 B. Christall, UFZ-Ber (1997) 12. Optimierung Umweltvertriglicher Analysen-
verfahren fr Mineralolkohlenwasserftoffe im Boden, 14.
422 E. Javorszky, E. Molnar and K. Torkos and J. Borossay, Chromatographia,51 (2000)
328.
423 J. Ritter, V.K. Stromquist, H.T. Mayfield, M.V. Henley and B.K. Lavine, Microchem.
J., 54 (1996) 59.
424 D. Djozan and Y. Assadi, Chromatographia,45 (1997) 183.
425 M. Llompart, K. Li and M. Fingas, J. Chromatogr.A, 824 (1998) 53.
426 D.C. Stahl and D.C. Tilotta, Environ. Sci. Technol., 33 (1999) 814.
427 F.J. Santos, M.T. Galceran and D. Fraisse, J. Chromatogr. A, 742 (1996) 181.
428 J. Czerwinski, B. Zygmunt and J. Namiesnik, FreseniusEnviron. Bull., 5 (1996) 55.
429 M.L. Chen, I.F. Mao and C.C. Hsu, Toxicol. Environ. Chem., 60 (1997) 39.
430 R. Mindrup, R. Shirey and V. Mani, Proc. Water Qual. Technol. Conf. 1995 (PT. 1)
(1996) 57.
431 T. Nilsson, L. Montanarella, D. Baglio, R. Tilio, G. Bidoglio and S. Facchetti, Int. J.
Environ. Anal. Chem., 69 (1998) 217.
432 M. Chai, C.L. Arthur, J. Pawliszyn, R.P. Belardi and K.F. Pratt, Analyst, 118 (1993)
1501
433 T. Nilsson, F. Pelusio, L. Montanarella, B. Larsen, S. Facchetti and J.O. Madsen, J.
High Resolut. Chromatogr., 18 (1995) 617.
434 K.J. James and M.A. Stack, FreseniusJ. Anal. Chem., 358 (1997) 833.
435 T. Nilsson, R. Ferrari and S. Facchetti, Anal. Chim. Acta, 356 (1997) 113.
436 D.L. Heglund and D.C. Tilotta, Environ. Sci. Technol., 30 (1996) 1212.
437 D.A. Cassada, Y. Zhang, D.D. Snow and R.F Spalding, Anal. Chem., 72 (2000) 4654.
438 E. Matisova, J. Sedlakova, M. Slezackova and T. Welsch. J. High Resolut.
Chromatogr., (1999) 109.
439 H. van Doorn, C.B. Grabanski, D.J. Miller and S.B. Hawthorne. J. Chromatogr.A,
829 (1999) 223.
440 L. Miller, E. Fattore and E. Benfenati, J. Chromatogr.A, 791 (1997) 221.
441 S-A. Barshick and W.H. Gries, Anal. Chem., 70 (1998) 3015.
442 D.C Stahl and D.C Tilotta, Environ. Sci. Technol., 35 (2001) 3507.
443 S.W. Lloyd, J.M. Lea, P.V. Zinba and C.C. Grimm, Water Res., 32 (1998) 2140.
444 R. McCallum, P. Pendleton, R. Schumann and M.-U. Trinh, Analyst, 123 (1998)
2155.
445 S. Tutschku, S. Mothes and R. Wennrich, FreseniusJ. Anal. Chem., 354 (1996) 587.
446 Y. Cai, S. Monsalud, K.G. Furton, R. Jaffe and R.D. Jones, Appl. Organomet. Chem.,
12 (1998) 565.
447 F. Yang and Y.K. Chau, Analyst, 124 (1999) 71.
448 T. De Smaele, L. Moens, P. Sandra and R. Dams, Microchim. Acta, 30 (1999) 241.
775
449 Y. Cai and J.M. Bayona, J. Chromatogr.A, 696 (1995) 113.
450 S. Mothes and R. Wennrich, J. High Resolut. Chromatogr., 22 (1999) 181
451 J. Wu, Z. Mester and J. Pawliszyn, J. Anal. Atomic Spectrom., 16 (2001) 159.
452 M. Guidotti, J. AOAC Int., 83 (2000) 1082.
453 C. Jia, Y. Luo and J. Pawliszyn, J. Microcolumn Sep., 10 (1998) 167.
454 M. Llompart, K. Li and M. Fingas, Anal. Chem., 70 (1998) 2510.
455 A. Paschke, P. Popp and G. Schueuermann, FreseniusJ. Anal. Chem., 360 (1998) 52.
456 Z. Takats and K. Torkos., Chromatographia,48 (1999) 74.
457 R. Aranda and R.C. Burk, J. Chromatogr.A, 829 (1998) 401.
458 A.A. Boyd-Boland, J.B. Pawliszyn, Anal. Chem., 68 (1996) 1521.
459 M-L. Bao, F. Pantani, O. Griffini, D. Burrini, D. Santianni and K. Barbieri, J.
Chromatogr.A, 809 (1998) 75.
460 P.A. Martos and J. Pawliszyn, Anal. Chem., 70 (1998) 2311.
461 Y.C. Wu and S.D. Huang, J. Chromatogr.A, 835 (1999) 127.
462 S-D. Huang, C-P. Cheng and Y.H. Sung, Anal. Chim. Acta, 343 (1997) 101.
463 M. Winkler, J.V. Headley and K.M. Peru, J. Chromatogr., (2000) 203.
464 K.K. Chee, M.K. Wong and H.K. Lee, J. Microcolumn Sep., 8 (1996) 131.
465 E. Benfenati, P. Pierucci, R. Fanelli, A. Preiss, M. Godejohann, M. Astratov, K.
Levsen and D. Barcelo, J. Chromatogr. A, 831 (1999) 243.
466 B. Aikawa and R.C. Burk, Int. J. Environ. Anal. Chem., 66 (1997) 215.
467 S-D. Huang, C. Yi ting and C-S. Lin, J. Chromatogr.A, 769 (1997) 239
468 R. Battle, C. Sanchez and C. Nerin, J. AOAC Int., 84 (2001) 431.
469 H.R. Rogers and S.D.W. Comber, Chemosphere, 37 (1998) 1413.
470 N.Huppert, M. Wuertele and H.H. Hahn, FreseniusJ. Anal. Chem., 362 (1998) 529.
471 H. Lakso, and WF. Ng. Anal. Chem., 69 (1997) 1866.
472 M. Moder, P. Popp and J. Pawliszyn, J. Microcolumn Sep., 10 (1998) 225.
473 P. Popp, C. Bauer, M. Moder and A. Paschke, J. Chromatrogr.A, 897 (2000) 153.
474 P. Popp, C. Bauer and L. Wennrich, Anal. Chim. Acta, 436 (2001) 1.
475 M. Moder, S. Schrader, M. Winkler and P. Popp, J. Chromatogr.A, 873 (2000) 95.
476 K. Luks-Betlej, P. Popp, B. Janoszka and H. Paschke, J. Chromatogr.A, 938 (2001) 93.
477 L. Wennrich, W. Engewald and P. Popp, Acta Hydrochim. Hydrobiol., 25 (1997) 329.
478 D.A. Volmer and J.P.M. Hui, Arch. Environ. Contam. Toxicol., 35 (1998) 1.
479 M.H.V. Mulder, BasisPrinciplesof Membrane Technology. Kluwer, Dordrecht, 1991.
480 N.C. van de Merbel, J. Chromatogr.A, 856 (1999) 55.
481 N.C. van de Merbel, F.M. Lagerwerf, H. Lingeman and U.A.Th. Brinkman, Int. J.
Environ. Anal. Chem., 54 (1994) 105.
482 J.A. J6nsson and L. Mathiasson, J. Chromatogr. A, 902 (2000) 205.
483 J.A. Jonsson and L. Mathiasson, Trends Anal. Chem., 18 (1999) 318.
484 J.A. Jonsson and L. Mathiasson, Trends Anal. Chem., 11 (1992) 106.
485 S. Palmarsdottir, E. Thordarson, L.-E. Edholm, J.A. Jonsson and L. Mathiasson,
Anal. Chemr., 69 (1997) 1732.
487 J.A. Jonsson and L. Mathiasson, Trends Anal. Chem., 18 (1999) 325.
488 P. Dzygiel, P. Wieczarek, L. Mathiasson and J.A. Jonsson, Anal. Lett., 31 (1998) 1261.
489 NT.Megersa, T. Solomon and J.A. Jdnsson, J. Chromatogr.A, 830 (1999) 203.
490 K.F. Pratt and J. Parliszyn, Anal. Chem., 64 (1992) 2101.
491 M. Yang, S. Harms, Y. Lou and J. Pawliszyn, Anal. Chem., 66, (1994) 1339.
492 M. Yang, M. Adams and J. Pawliszyn, Anal. Chem., 68 (1996) 2782.
493 Y. Luo, M. Adams and J. Pawliszyn, Anal. Chem., 70 (1998) 19.
494 A. Segal, T. Gorecki, P. Mussche, J. Lips and J. Pawliszyn, J. Chromatogr., 873A
(2000) 13.
495 Y. Luo and J. Pawliszyn, Anal. Chem., 72 (2000) 1058.
496 M.E. Bier and R.G. Cooks, Anal. Chem., 59 (1987) 597.
776
497 T. Kotiaho, J. Mass Spectrom., 31 (1996) 1.
498 F.R. Lauritsen, T. Kotiaho, T.K. Choudhury and R.G. Cooks, Anal. Chem., 64 (1992)
1205.
499 S. Bauer, Trends Anal. Chem., 14 (1995) 202.
500 N. Srinivasan, R.C. Johnson, N. Kasthurikrishnan, P.W. Wong and R.G. Cooks,
Anal. Chim. Acta, 350 (1997) 257..
501 R.C. Johnson, R.G. Cooks, T.M. Allen, M.E. Cisper and P.H. Hemberger, Mass Spec.
Rev., 19 (2000) 1.
502 T. Kotiaho, R.A. Ketola, M. Ojala, T. Mansikka and R. Kostiainen, Am. Env. Lab.,
3/97 (1997) 19.
503 T. Kotiaho, F.R. Lauritsen, T.K. Choudhury, R.G. Cooks and G.T. Tsao,Anal. Chem.,
63 (1991) 875A.
504 M. Ojala, R.A. Ketola and T. Kotiaho, LC-GC, 16 (1998) 1026.
505 T. Kotiaho, R. Kostiainen, R.A. Ketola, M. Ojala, I. Mattila and T. Mansikka, in: E.J.
Karjalainen, A.E. Hesso, J.E. Jalonen and U.P. Karjalainen (Eds.), Advances in Mass
Spectrometry, Vol. 14. Elsevier, 1998, pp. 501.
506 F.R. Lauritsen and T. Kotiaho, Rev. Anal. Chem., 15 (1996) 237.
507 J.D. Petty, C.E. Orazio, J.N. Huckins,R.W. Gale, J.A. Lebo, K.R. Echols and W.L.
Cranor, J. Chromatogr. A, 879 (2000) 83.
508 T. Colborn, F.S. vom Sall and A.M. Soto, Environ. Health Perspect., 101 (1993) 378
509 J.N. Huckins, G.k. Manuweera, J.D. Petty, D. MacKay and J.A. Lebo, Environ. Sci.
Technol., 27 (1993) 2489.
510 J.D. Petty, J.N. Huckins, C.E. Orazio, J.A. Lebo, B.C. Poulton and R.W. Gale,
Environ. Sci. Technol., 29 (1995) 2561.
511 J.A. Lebo, R.W. Gale, J.D. Petty, D.E. Tillitt, J.N. Huckins and J.C. Meadows,
Enuiron. Sci. Technol., 29 (1995) 2886.
512 J.D. Petty, J.N. Huckins and J.L. Zajicek, Chemosphere, 27 (1993) 1609.
513 J.A. Lebo, J.L. Zajicek, C.E. Orazio, J.D. Petty and J.N. Huckins, Polycyclic Arom.
Comp., 8 (1996) 53.
514 J.D. Petty, B.C. Poulton, C.S. Charbonneau, J.N. Huckins, S.B. Jones and J.T.
Cameron, Environ. Sci. Technol., 32 (1998) 837.
515 P.S.H. Wong, R.G. Cooks, M.E. Cisper and P.H. Hemberger, Environ. Sci. Technol.,
29 (1995) 215A.
516 V.T. Virkki, R.A. Ketola, M. Ojala, T. Kotiaho, V. Komppa, A. Grove and S. Facchetti,
Anal. Chem., 67 (1995) 1421.
517 B.J. Harland and P.J. Nicholson, Sci. Total Environ., 135 (1993) 37.
518 T. Kotiaho, R. Kostiainen, R.A. Ketola, T. Mansikka, I. Mattila, V. Komppa, T.
Hankanen, K. Wickstrom, J. Waldvogel and 0. PilviS, Proc. Contr. Qual. 11 (1998) 71.
519 G. Matz, W. SchrBder, A. Harder, A. Schillings and P. Rechenbach, Field. Anal.
Chem. Technol., 1 (1997) 181.
520 G. Matz and G. Kibelka, in: Proceedingsof PITTCON '98, New Orleans, Lousiana,
March 1-5, 1998, 1864P.
521 B. Hauser and P. Popp, J. Chromatogr.A, 909 (2001) 3.
522 M. Soni, S. Bauer, J.W. Amy, P. Wong and R.G. Cooks, Anal. Chem., 67 (1995) 1409.
523 R.A. Ketola, V.T. Virkki, M. Ojala, V. Komppa and T. Kotiaho, Talanta, 44 (1997)
373.
524 M.A. LaPack, J.C. Tou and C.G. Enke, Anal. Chem., 62 (1990) 1265.
525 L.E. Slivon, M.R. Bauer, J.S. Ho and W.L. Budde, Anal. Chem., 63 (1991) 1335.
526 R. Kostiainen, T Kotiaho, I. Mattila, T. Mansikka and R.A. Ketola, Anal. Chem., 59
(1987) 597.
527 R.M. Alberici, R. Sparrapan, W.F. Jardim and M.N. Eberlin, Environ. Sci. Technol.,
35 (2001) 2084.
777
528 R.A. Ketola and F.R. Lauritsen, Rapid Commun. Mass Spectrom., 13 (1999) 749.
529 F.R. Lauritsen and J. Rose, Analyst, 125 (2000) 1577.
530 F.R. Lauritsen, M.A. Mendes and T. Aggerholm, Analyst, 125 (2000) 211.
531 T. Aggerholm and F.R. Lauritsen, Rapid Commun. Mass Spectrom., 15 (2001) 1826.
532 F.R. Lauritsen and R.A. Ketola, Anal. Chem., 69 (1997) 4917.
533 M.A. Mendes and M.N. Eberlin, Analyst, 125 (2000) 21.
534 G. Matz, M. Loogk and F. Lennemann, J. Chromatogr.A, 819 (1998) 51.
535 G. Matz, G. Kibelka, J. Dahl and F. Lennemann, J. Chromatogr. A, 830 (1999)
365-376.
536 J. Trocewicz, J. Chromatogr.A, 725 (1996) 121.
537 L. Chimuka, M.M. Nindi and J.A. J6nsson, Int. J. Environ. Anal. Chem., 68 (1997), 429.
538 N. Megersa and J.A. J6nsson, Analyst, 123 (1998) 225.
539 N. Megersa, T. Solomon and J.A. Jonsson, J. Chromatogr.A,., 830 (1999) 203.
540 N. Megersa, T. Solomon, B.S. Chandravanshi and J.A. Jnsson, Bull. Chem. Soc.
Ethiopia, 14 (2000) 9.
541 N. Megersa, L. Chimuka, T. Solomon and J.A. J6nsson, J. Sep. Sci., 24 (2001)567.
542 G. Nilv6, G. Audunsson and J.A. Jonsson, J. Chromatogr., 471 (1989) 151.
543 L Mathiasson, G. Nilv6 and B. U16n, Int. J. Environ. Anal. Chem., 45 (1991) 117.
544 M. Knutsson, G. Nilv6, L. Mathiasson and J.A. J5nsson, J. Agric. Food Chem., 40
(1992) 2413.
545 G. Nilve and R. Stebbins, Chromatographia,32 (1991) 269.
546 G. Nilve, M. Knutsson and J.A. Jonsson, J. Chromatogr.A, 668 (1994) 75.
547 M. Knutsson, L. Mathiasson and J.A. Jonsson, Chromatographia,42 (1996) 165.
548 T. Miliotis, M. Knutsson, J.A.Jonsson and L.Mathiasson, Int. J. Environ. Anal.
Chem., 64 (1996) 35.
549 J. Norberg, A. Zander and J.A. Jonsson, Chromatographia,46 (1997) 483
550 R. Carabias Martinez, E. Rodriguez Gonzalo, M.P. Santiago Toribio and J.
Hernandez Mendez, Anal. Chim. Acta, 321 (1996) 147.
551 K. Ndung'u and L. Mathiasson, Anal. Chim. Acta, 404 (2000) 319.
552 J.Norberg, E. Thordarson, L. Mathiasson and J.A. Jonsson, J. Chromatogr.A, 869
(2000) 523.
553 M. Sandahl, L. Mathiasson and J.A. Jonsson, J. Chromatogr. A, 893 (2000) 123.
554 M. Sandahl, E. Ulfsson and L. Mathiasson, Anal. Chim. Acta, 424 (2000) 1.
555 R.G. Melcher and P.L. Morabito, Anal. Chem., 62 (1990) 2183
556 P.L. Morabito and R.G. Melcher, Proc. Contr. Qual., 3 (1992) 35.
557 X. Guo and S. Mitra, J. Chromatogr.A, 904 (2000) 189.
558 K. Grob, On-line coupled LC-GC. Hithig, Heidelberg, 1991.
559 K. Grob, J. Chromatogr. A, 703 (1995) 265.
560 M.-L. Riekkola, J. Chromatogr. A, 473 (1989) 315.
561 I.L. Davies, K.E. Markides, M.L. Lee, M.W. Raynor and K.D. Bartle, J. High Resolut.
Chromatogr., 12 (1989) 193.
562 E. Dijkman, D. Mooibroek, R. Hoogerbrugge, E. Hogendoorn, J.V. Sancho, O. Pozo
and F. Hernandez, J. Chromatogr. A, 926 (2001) 113.
563 F. Hernandez, C. Hidalgo, S. Grimalt and J.V. Sancho, Quim. Anal., 20 (2001) 81.
564 E.A. Hogendoorn, K. Westhuis, E. Dijkman, E.A.G. Heusinkveld, P. Chamraskul, P.
Biadul, R.A. Baumann, A.A. Cornelese and M.A. van der Linden, Int. J. Environ.
Anal. Chem., 78 (2000) 67.
565 J.L.M. Vidal, P.P. Vazquez and J.M. Fernandez, Chromatographia,51 (2000) 187.
566 A. Motoyama, A. Suzuki, O. Shirota and R. Namba, Rapid Commun. Mass Spectrom.,
13 (1999) 2204.
567 L.N. Konda, M.B. Barroso, G. Morovjan and P. Csokan, J. Chromatogr. Sci., 37
(1999), 71.
778
Chapter23
23.1 INTRODUCTION
TABLE 23.1
Checklist for carrying out an effective sampling strategy in drug analysis
1. What are your data quality objectives?
2. Is specialized sampling equipment needed and/or available?
3. Are the samplers experienced in the type of sampling required/available?
4. Have all analytes been listed? Have detection level and methods been specified for each
analyte?
5. Are all analytes stable in the sample? Is a special preservation method needed after the
sampling?
6. Are there specific types of quality control samples? Does the instrument require
optimization of its operating parameters?
7. What type of sampling approach will be used? Random, systematic, judgemental or a
combination of these?
8. Will the type of sampling meet your data quality objectives? Is the sampling approach
compatible with data analysis method?
9. How many samples are needed? How many methods are specified? How many test
samples are needed for each method?
10. What types of quality control samples are needed? How many exploratory samples are
needed? How many supplementary samples will be taken?
780
The sample preparation step is necessary to isolate the components of inter-
est from a sample matrix because most analytical instruments cannot handle the
matrix directly. Sample preparation can include clean-up procedures for very
complex (dirty) samples. This step must also bring the analytes to a suitable con-
centration level. During a separation the isolated mixture of analytes is divided
into its constituents. Separation techniques such as chromatography and elec-
trophoresis are well suited for the analysis of complex multi-component samples.
The direct injection of complex matrixes such as plasma, serum, whole blood, tis-
sue homogenates, saliva or urine into these separation systems was studied in
the past with moderate success [20,21]. Typically, however, injection of samples
with large amounts of protein results in a rapid deterioration of the column's
separation performance, an increase in background interferences, and a dra-
matic increase in column backpressure. In addition, the chromatographic sorb-
ent selectivity may be altered by irreversible adsorption of matrix compounds
such as proteins. Therefore, the analysis of drugs present in these biological
samples requires well-designed sample preparation procedures. The isolation
and measurement of organic compounds present in a biological matrix,
especially at low concentration, presents a significant analytical challenge.
Objectives of sample preparation for drug analysis in biological samples and
pharmaceutical products are summarized in Table 23.2. The objectives of the
analytical method will indicate how much effort will be necessary to put into a
sample preparation step. For example, TDM usually requires specificity to dis-
tinguish the drug to be monitored from similar compounds, metabolites or
co-administered drugs. In contrast, a pharmacokinetic study of a potential drug
candidate requires a specific and sensitive analytical method. Therefore, it is
important to obtain the required analyte at a stable and high recovery rate for
each individual application.
TABLE 23.2
Objectives of sample preparation for drug analysis
1. Removal of unwanted macromolecular contaminants (mainly proteins) that would
interfere with analyte determination.
2. Removal of selected analyte components if the resolving power of the chromatographic
and electrophoretic column is insufficient to separate all the components in the sample
completely or in a reasonably practical time.
3. Removal of material that would affect chromatographic and electrophoretic resolution or
reproducibility.
4. Removal of material that could block the chromatograph and electrophoresis tubing,
valve, column or frits.
5. Determination of total analyte present in a complex.
6. Solubilization of analytes to enable injection under the initial chromatographic and
electrophoretic conditions.
7. Dilution to reduce solvent strength or avoid solvent incompatibility.
8. Concentration of the analyte within the detection limits of the analytical instrument.
9. Stabilization of the analyte to avoid hydrolytic or enzymatic degradation.
10. Derivatization of analytes for enhancement of sensitivity and selectivity.
781
23.1.2 Classification of sampling and sample preparation for drug
analysis
As shown in Table 23.3, the sampling and sample preparation for drug analysis
can be mainly classified into five separate operations: sample collection, release
of drugs from matrix, liquid handling, removal of endogenous compounds and
enhancement of sensitivity and selectivity [26]. For the sample collection, proce-
dures are adopted to ensure the integrity of the determinant during collection,
transport and storage. Various sampling techniques have been used for different
samples such as blood, urine, milk, hair and tissues. These samples must be col-
lected without errors by contamination from sampling equipment (e.g., syringe,
pipette, cannulas, tube and vessel) or by use of the wrong sampling technique.
Biological matrices are complex mixtures of compounds and usually have
high protein content. High proportions of the drugs may be bound to protein,
glucuronide and sulphate, and be present as free drug only at low concentra-
tions. Therefore, for analysis of total drug content, the analyst must first release
drugs from the matrix by hydrolysis. The hydrolysis can be performed either by
enzymes or by acid or base. Sonication, dilution or heating are also possible
means of improving the recovery of drugs from biological samples. In the
opposite case, where the goal is to assess the free and therefore therapeutically
relevant concentration, the analyst must ensure that the equilibrium between
free and bound drug is not perturbed during sampling and sample preparation.
Liquid handling procedures mainly provide the links between each opera-
tional step. Procedures for liquid handling involve the addition, mixing, separa-
tion or removal of liquids. They can often be the rate-limiting steps in a sample
preparation process because they are typically labour-intensive and tedious
procedures.
A biological matrix may be solid, particulate or a mixed composition of
organic compounds in an aqueous solution, and consists of many components
such as macromolecules and small molecules. Ultrafiltration and protein precip-
itation can be used to remove protein from biological samples. The membrane
filters used for ultrafiltration are efficient, although care must be taken to avoid
binding of the drug to the membrane. Precipitation using organic solvents and
inorganic acids or salts is effective for removal of proteins. Dialysis can be used to
separate an analyte from the matrix by diffusion through a semi-permeable
membrane rather than centrifugal force with ultrafiltration. On the other hand,
liquid-liquid extraction LLE), solid-phase extraction (SPE) and solid-phase
microextraction (SPME) are useful sample preparation techniques, and can
produce clean extracts for analysis very efficiently. Supercritical fluid extraction
(SFE), immunoaffinity extraction (IAE), molecularly imprinted polymer (MIP)-
based extraction and membrane-based extraction are also used as specific and
efficient sample preparation techniques. These techniques, which are based on
the adsorption or partitioning of analyte, are responsible for removing the
majority of the biological material from the sample matrix prior to analysis.
782
TABLE 23.3
Classification of sampling and sample preparation
Sample collection Suction
Drawing
Cutting
Pipetting
Head-space sampling
Release of drugs from matrix Hydrolysis Acid
Base
Enzyme Proteases
Lipase
[-Glucosidase
Arylsulphatase
Sonication
Liquid handling Aspiration
Centrifugation
Dilution
Evaporation
Filtering
Mixing
Salting-out
Separation
Removal of endogenous Precipitation Organic solvents
compounds Inorganic acids and salts
Ammonium sulphate
Ultrafiltration
Dialysis
Membrane extraction
Liquid-liquid extraction
Supercritical fluid extraction
Solid-phase extraction
Solid-phase microextraction
Immunoaffinity extraction
Liquid chromatography
Procedures for enhancement of Pre-column derivatization
sensitivity and selectivity Post-column derivatization
Use of selective detectors
783
23.1.3 Objectives and scope of this chapter
The process of interaction between drugs and human organs is divided into three
phases: the pharmaceutical (exposure) phase, the pharmacokinetic (toxicokin-
etic) phase and the pharmacodynamic (toxicodynamic) phase. Pharmacodyma-
nics is concerned with the study of interactions between the drug (or its active
metabolites) and the receptors in target tissues. While this subject is of great
importance in understanding the mechanism of toxicological effects of drugs, it
is normally not a major concern to drug analysts. On the other hand, many
processes that take place during the pharmaceutical and pharmacokinetics
phases are highly relevant to drug analysis. The manner in which a drug is
introduced into a biological system, that is, the route of administration, dictates
the set of digestive and enzymatic conditions to which the drug is subjected and,
therefore, the metabolic fate of the drug and its distribution in various tissues.
Common routes of introduction include oral ingestion, intravenous and intra-
muscular injections, inhalation, and dermal application. Once a drug is intro-
duced into a biological system, it is subjected to the processes of absorption,
distribution, metabolism, and excretion, which are subjects of pharmacokin-
etics. Some basic principles governing the interactions between drugs and the
human biological system with respect to drug introduction, absorption, distribu-
tion, and excretion will be addressed. Such knowledge facilitates the proper
selection of specimens for analysis, sample pretreatment, and analytical data
interpretation. Various biological matrices used for drug analysis in clinical and
forensic applications are summarized in Table 23.4. In this section, characteris-
tics, collection and handling of plasma, serum, blood, urine, saliva, milk, sweat,
feces, tissues and breath are described. Other matrices in the table are not
commonly used in conventional clinical and toxicological analysis, but they may
provide valuable information otherwise not available in clinical and forensic
applications. The details of these matrices for sampling and sample preparation
784
TABLE 23.4
Biological matrices used for drug analysis in clinical and forensic applications
Liquids and gases Blood: Whole blood, plasma, serum
Urine
Secretion and effusion: Saliva, tears, sweat, milk, cerebrospinal fluid,
gastric juice, bile, semen, liquor amnii
Breath
are also described in the books [1,19,50] and well-documented reviews [39,42,
49,51-56].
Blood is perhaps the most useful sample for the identification and quantification
of drugs and for the interpretation of data of toxicological significance. Blood is a
complex fluid containing solubilized proteins, dissolved fats and salts, and
suspended cells. The major constituents-the red blood cells (erythrocytes)-
can be separated from the clear fluid (plasma) by centrifugation, if the blood is
mixed with a suitable anticoagulant during the draw. If blood is allowed to stand
without the addition of anticoagulating agents, the red cells will eventually clot
and the resulting straw-coloured serum can be decanted. In plastic tubes,
clotting may be delayed for more than one hour and some hemolysis can take
place. Hemolysis may also occur during the blood draw if the needle is positioned
too close to the wall of the vessel, or if it is drawn too quickly and the red cells
impact the wall of the sampling tube too hard. Hemolysis may or may not impact
the analytical results, depending on the specific test used. As matrices for drug
analysis, the significant difference between serum and plasma is that serum does
not contain fibrinogen and some clotting factors (2.5-5% of proteins). In general,
both plasma and serum are routinely used for drug analysis. A method developed
for plasma can normally be applied without modification to serum. On the other
hand, whole blood may be the sample of choice when trying to detect the
presence of a drug, or when quantitative information is needed on a deteriorated
blood sample where complete separation of plasma or serum from red cells is not
possible. When anticoagulation is used there is a potential for analytical interfer-
ence. Whole blood is often extracted with no pre-treatment. The blood sample is
added to a sample vial, and extraction is initiated. It may however be preferable
to extract from blood collected with heparin or another anti-coagulation agent.
As heparin is a mucopolysaccharide and may interact with some analytes,
785
ethylenediamine tetraacetic acid (EDTA) or one of the other anticoagulants may
be preferable depending on circumstances. EDTA should not be used where
metallics or organometallics are to be analyzed. The containers or their stoppers
may also cause contamination but these problems are well recognized. A com-
mon error in blood collection is sample haemolysis which causes analyte dilution
and interference in a serum assay. Because serum and plasma consist of a large
amount of protein, drugs will bind to proteins to varying degrees depending on
their individual physicochemical properties. In general, acid and neutral drugs
bind primarily to albumin, and basic drugs bind primarily to acid glycoprotein.
Only free drug is available for extravascular distribution and elimination, and is
able to cross cellular membranes and interact with drug receptors. Where the
goal is analysis of total drug concentration, direct analysis of drugs in protein-
free serum or plasma may not reflect the total drug concentration. Therefore,
protein is often denatured prior to extraction with an organic solvent. Alterna-
tively, drugs strongly bound to protein may be freed with appropriate adjust-
ment of pH.
23.2.2 Urine
Urine is one of the most commonly studied biological matrices for drug analysis,
particularly because of its relative ease of collection. However, samples should
not be taken from women during menstruation, and men should be warned that
there is a risk of contamination by sperm remains. For pharmacokinetic studies,
urine is usually sampled according to detailed instructions over 24 hours into
pre-cleaned bottles. Preservatives may be required to stabilize the analyte
particularly if it is susceptible to bacterial degradation or the urine may be
chilled during the collection period. To guarantee the stability of the urine, the
time that elapses between sampling and analysis should be kept as short as
possible. To avoid the risk of contamination the urine is not acidified. After
through shaking, the urine is divided and the subsamples are frozen immedi-
ately. As a matrix, urine has moderate complexity, and typically contains both
organic and inorganic constituents, as well as a relatively high salt content. In
addition, urine suffers from high variability. Results are typically normalized by
expressing drug concentration per gram creatinine, in order to circumvent the
highly variable dilution that results from the body's attempts to maintain water
balance [57,58]. Compared to blood, plasma, or serum, urine is relatively free of
protein, thus making it possible for direct extraction with an organic solvent.
However, the interpretation of results obtained from urine samples is compli-
cated by several factors, such as the amount of excretion, the variation of pH
values, the variation of ionic strength, and the time lapse after the intake of the
drug. Variations in urine pH effects both elimination of many drugs, as well as
extraction efficiency when effects on the drug's acid/base balance are significant.
Variations in ionic strength can also have a significant effect on drug extraction.
The additional components of the sample may also provide for competitive
786
sorption, further reducing the amount extracted. While this provides an oppor-
tunity for studying distribution of a drug between the various components of a
urine sample, it is often total urine concentration that must be reported. This is
accomplished by using an appropriate internal standard, or standard addition
for quantification. The internal standard selected should be structurally related
to the other compounds. If urine samples must be significantly diluted prior to
extraction, because of very high drug concentrations, variability in the matrix
becomes of less significance for extractions.
23.2.3 Saliva
Saliva is a colorless fluid excreted into the oral cavity from three principle
glands: (1) the parotid gland, exiting at the top of the mouth, secretes saliva
derived mainly from blood plasma (serous fluid); (2) the sublingual glands, exit-
ing at the sides of the mouth, excrete both serous fluid and mucin; and (3) the
submandibular glands, exiting at the base of the tongue, also excrete both serous
fluid and mucin. Saliva is approximately 99% water, 0.3% protein (mostly
enzymes) and 0.3% mucin with the balance salts. The mucin gives saliva its
sticky character. The low protein concentration in saliva makes drug binding
minimal compared to that observed in plasma. Unstimulated saliva pH is in the
range of 5.6-7.0 and increases with stimulation (to more approximate the pH of
blood, i.e., 7.4) to a maximum of 8.0. Therefore, as discussed above, drug concen-
trations in saliva partially depend on the pH of the saliva and the degree of stim-
ulation. Compared to urine collection, saliva sample-collection procedures are
less invasive and cause less concern about violation of privacy and adulteration
of samples. Because saliva can be used to estimate the actual, non-protein
bound, circulating concentration of some drugs and their metabolites at the time
of collection, test results may have potential for correlation with performance
impairment [59]. Compared to blood, saliva samples can be directly analyzed
without a prior extraction step. These advantages have promoted numerous
investigations and several review articles on the suitability of saliva in TDM and
in forensic applications [59-61]. Limited amounts of mixed saliva may be col-
lected by spitting. Larger amounts of saliva may be collected by stimulating
saliva flow by mastication of rubber bands, wax, Teflon tape or gum; sucking on
pebbles, marbles or candy; by placing citric acid on the tongue; adsorption on
cotton rolls or administration of pilocarpine. Selection of material for saliva sim-
ulation must be carefully chosen because lipophilic drugs may be adsorbed into
the material. For identification purposes, adsorption (similar to SPE or SPME)
could be employed to directly extract the drug from the saliva while inside the
mouth. Little research has been done in this area, as the sample size would be
limited and the adsorption process variable (depending on the cooperation of the
individual, i.e. the more sucking/chewing, the greater the adsorption). Some
devices (OraSure®) both stimulate saliva flow and collect the saliva on an absor-
bent pad [62-64]. Other devices have been developed to acquire saliva from
787
selected glands [65]. In general, these devices acquire saliva by placing the end of
the device over the gland and applying suction [66]. Although selected gland
secretions have an advantage in reducing saliva/plasma ratios and minimizing
oral contamination, most studies collect mixed saliva specimens because it is less
invasive. An ultrafiltration device (SalivaSac®) has been developed to reduce the
viscosity of saliva for easier analysis, and is similar to that collected by suction
from the parotidal gland. It consists of a dialysis membrane enclosing sucrose
crystals and is approximately 3.5 cm in diameter and a few millimeters thick
[67,68]. The device is placed in the mouth and massaged with the tongue for a
few minutes until all of the crystals are dissolved, collecting 1-2 ml of saliva
ultrafiltrate. Both the sucking on the device and the sweet taste stimulate saliva
production. The SalivaSac® may also be externally coated with citric acid to
stimulate saliva production. The dialysis membrane is chosen to exclude most
higher-molecular-mass substances, therefore, the mucopolysaccharides, food
particles and bacteria are not collected. The SalivaSac® may also have a handle
for easier insertion and removal. However, two problems concerning the correla-
tion of drug levels of the ultrafiltrate with saliva levels are apparent when quan-
titative information is needed. The sucrose in the SalivaSac® takes up an
appreciable amount of the molar volume of water inside the device. Therefore,
measurement of the density of the fluid is necessary by careful weighing and a
correction factor must be calculated [69]. Furthermore, diffusion through the
dialysis membrane is related to molecular mass and, therefore, water is prefer-
entially collected relative to drugs [69]. Thus, the concentration of drugs in the
SalivaSac® is lower than in the external saliva and this ratio may vary depend-
ing on how the user moves the device during the saliva collection. If one only
wishes to have qualitative information on drug use, then these concerns are not
applicable.
23.2.4 Milk
The main constituents of human milk are water (88%), proteins (3%), lipids (ca.
3%), carbohydrate in the form of lactose (6.8%) and other solids (minerals,
vitamins etc.). The lipids are in the form of fat droplets suspended in the watery
matrix. The colostrum that appears in early lactation differs significantly from
true milk in that it contains less lactose and virtually no fat. The non-water
constituents are present in different physical forms; dissolved (lactose), colloid-
ally dispersed (protein) and emulsified in water (lipids or fats) [70]. Human milk
can serve both as a means of exposure of a newborn to compounds the mother
has been previously exposed to, and provide a relatively convenient means of
biomonitoring for toxicant exposure. Human milk is quite constant in composi-
tion from the third week on, but the colostrum from early lactation is signifi-
cantly different as noted above. Such variation could impact analyses if not
taken into account. The sampling of human milk is always done with the
assistance of a trained person. The breast, in particular the nipple, is cleaned
788
with sterile compress and bidistilled water, in order to remove possible contami-
nants (e.g., breast cosmetics). The sampling equipment, which includes an
electric milk pump is unpacked and installed at the place where the sampling is
done. The breast adaptor is pressed on to one breast, so that no air is drawn in
and a light vacuum is created. The pump is started and controlled by the test
person herself. As soon as the desired volume has been reached, the pump is
stopped and the filled milk bottle is removed from the breast. To minimize
contact with the air, the sample is rapidly subdivided into prepared tubes. If
possible this procedure should be done under clean-room conditions in the
laboratory. If the let-down of the milk dose not take place, even though the
breast is full, the test person may use a nasal spray of oxytocin in order to
stimulate the milk flow. Milk samples are commonly used for the trace analysis
of pesticides, heavy metals, antibiotics and some drugs.
23.2.5 Hair
789
Hair is mainly collected from the area at the back of the head, called the
posterior vertex. This area has less variability in hair growth rate than other
areas. Moreover, the hair is less subject to age and sex-related influences [71].
The sample size collected should be 200-250 mg at a minimum because of the
relatively large interindividual variability. When head hair is not available or is
too short, pubic or axilla hairs may be collected. Other than for drugs of abuse
testing (violation of laws on narcotics, probation, drug abstinence in detoxifica-
tion treatments, child custody, divorce), testing hair for pharmaceuticals in
forensic science is most frequently performed on deceased persons. Thus, the
quantity of the collected hair could be significantly higher than for living people,
except indeed for newborns, the very young or the very old. The problem of
external contamination is very important for drugs of abuse as people can be
exposed to a smokey or dusty environment. Papers dealing with the problem,
which also impacts the establishment of cut-off values, are numerous and
various studies have been published [72-76]. The approach is very different for
pharmaceutical analyses as the drugs are normally to be taken by mouth or by
another therapeutic route under normal pharmaceutical presentation. Another
important serious concern is the change in the drug concentration induced by
cosmetic treatments of hair. The strong bases used for such treatments may
cause hair damage resulting in loss from the hair matrix or alternatively under
favorable environmental contamination conditions, a higher incorporating ratio
[76]. In fact, the history of hair analysis is the history of extraction procedures,
because after the drug has been extracted it can be handled as if it had been
extracted from urine or blood samples. Various extraction procedures have been
proposed including chronologically [77]: methanolic sonication, 0.1 M sodium
hydroxide, 0.1 M hydrochloric acid, water, buffers, 1 M sodium hydroxide, 0.1%
sodium dodecylsulfate, acetone, pronase, performic acid-pronase, methanol, 5 M
hydrochloric acid, proteinase K, dithiothreitol, phosphate buffer pH 7.4, -glucu-
ronidase-arylsulfatase and SFE [78,79]. With the exception of SFE and methan-
olic sonication, the resulting mixture from the pretreatment step must be
further purified prior to analytical examination. It could be as simple as LLE
with common solvents or as complex as a multi-step trace enrichment procedure
as that devoted to the analysis of 17-keto steroids and their esters [80]. The fields
in which hair analysis has so far been applied are mainly forensic toxicology and
drug abuse studies, followed by clinical toxicology and clinical chemistry. The
largest number of papers on hair analysis has dealt with cocaine, opiates and
amphetamines. These drugs are the subject of almost 50% of all papers on hair
analysis. However, recent hair analysis studies have addressed other kinds of
drugs, for example doping agents like clenbuterol, therapeutic drugs like benzo-
diazepines, methadone and carbamazepine.
23.2.6 Sweat
Sweat is an alternative target for the determination of drug-excretion patterns,
and drug compositions in sweat may partially contribute to the observation of
790
drugs in hair [81,82]. Sweat is excreted from eccrine and apocrine glands and
skin is bathed with aqueous and sebaceous secretions, especially on the face and
scalp. The sebaceous secretions are primarily lipids that may transport and
absorb many drugs. Sebum is excreted more on the scalp and forehead than on
other areas of the body. Therefore, different concentrations may be expected,
depending upon the area of the body in which the sample is taken, because
fat-soluble drugs may be sequestered or secreted in sebum. Almost all studies
obtain mixed secretions of sweat and sebum, which is incorrectly referred to as
sweat. Sweating may be induced by exercise and several milliliters of sweat may
be collected in conjunction with an occlusive bandages, consisting of one to three
layers of filter paper or pieces of cotton, gauze or towel. Significant advances
have been made during the past years with the development a "sweat patch"
technology, which was proposed by PharmChem Labs. (Menlo Park, USA).
"Sweat patches" used for collecting perspiration samples from human skin have
improved from an earlier occlusive design to the current dressing technology
using very thin polyurethane and acrylate adhesives. This new nonocclusive
approach allows the moisture content of perspiration to evaporate and makes it
possible to apply a single dressing to skin and maintain sterile attachment for
several days with rare dermatological reactions. The final content of the collec-
tion pad is an accumulation or integration of that contributed to it by skin [83]. A
patch was developed that included a chemical binding layer in the absorbent pad
to prevent back-diffusion of the drug through the skin. This design has been
used to monitor theophylline in monkeys [84] and caffeine in infants [85]. Both
of these early patches used aqueous media and an occlusive covering to stimulate
sweat and prevent evaporation of the drug. A device has also been developed that
has a covering that allows the passage of sweat from the skin through the device
and prevents external water and other molecules from back-diffusing into the
absorptive pad [86]. This device is being marketed as the PharmChek® sweat
patch [87]. A potential problem with the PharmChek® patch is the absence of a
layer between the skin and the absorptive pad, to prevent bacterial transfer into
the pad and, therefore, the possibility of bacterial growth and drug degradation.
Careful preparation of the skin prior to application of the patch should kill or
remove bacteria and prevent these problems, and the absence of substantial
moisture in the pad decreases the possibility of bacterial growth. Even with
these many devices, sweat is more difficult to collect non-invasively than is
saliva, due to the lower amounts/unit area secreted in a given time. As an
alternative collection of sweat, wiping the skin with a cotton pad moistened with
alcohol was developed [88]. This procedure allows rapid collection of drugs that
may arise both from sweat evaporating on the surface of the skin and from
external contamination. The exact composition of the fluid analyzed is not
known because both the aqueous secretions (true sweat) as well as sebum are
collected. Whether or not the presence of drugs on skin wipes indicates use only
or use and exposure is unknown. In a survey of a university population, skin
wipes detected more cocaine use/exposure than did hair analysis [89].
791
23.2.7 Other matrices
TABLE 23.5
Characteristics of illicit pharmaceuticals
1. Various forms of carriers or containers are used.
2. Various unintentional impurities, such as degradation products, synthesis by-products,
and residual solvents are present.
3. Intentional multiple-component drug combinations and the addition of adulterants and
excipients are common.
4. The contents are often unexpected.
5. Many of these drugs may be sold as counterfeits of pharmaceutical products.
792
The nature of the active ingredient is only one of several important consider-
ations in the design of pharmaceutical products. It is widely recognized that
pharmaceutical products from different manufacturers, containing the same
active ingredient and dosage, can produce markedly different circulating con-
centrations of the active ingredient. Optimal drug formulation is a crucial step in
the design of products that produce optimal therapeutic effect. In addition to
meeting regulatory standards for active ingredient quality, identity, purity and
stability, products are typically optimized for utility and bioavailability to assure
the best competitive advantage. The nature of the formulation procedures
and/or auxiliary (non-medical) components of a product can for example, make a
product easier to handle, limit degradation, maintain homogeneity, enhance
bioavailability and targeting of the active ingredient to receptor sites, provide for
additional routes of administration, and target the product for additional thera-
peutic indications for use. These auxiliary inactive ingredients, known collec-
tively as excipients, can also solubilize the active ingredient, suppress the growth
of microorganisms, provide bulk or a coating for a tablet or capsule, provide color
or mask an unpleasant taste or odor. A good working definition of excipients is
"agents in a medicinal preparation regardless of the nature, purpose or quanti-
ties employed, other than the components intended as the active ingredients".
With so many tasks to perform, it is obvious therefore that the nature and num-
bers of excipients in various pharmaceutical products can vary widely. It is
essential therefore, that for analysis of pharmaceutical products, the nature of
these excipients is taken into account, for their potential impact on the process
of analysis of the active ingredient. Excipients can be divided based on either for-
mulation type or excipient action. Table 23.6 provides an overview of common
excipients for various drug formulations. In this section, characteristics, collec-
tion and handling of illicit preparations and pharmaceutical products for foren-
sic and pharmaceutical analysis are described.
793
TABLE 23.6
Common excipients for pharmaceutical preparations
794
similar abilities in differentiating between closely related but different drug sei-
zures. Interestingly, for ecstasy, isosafrole was identified as a precursor for syn-
thesis. With LLE, this compound was difficult to analyze, due to its high
volatility. There has also been a report on the use of SPME for profiling of manu-
facturing by-products and impurities from an illicit drug seizure of 4-methoxy-
amphetamine [93]. The results indicated the synthetic route likely used in the
preparation of the compound, and the method was found to be rapid, non-
destructive and give results complementary to those from conventional LLE.
795
Liquid formulations: e.g. lotionsemulsionslsuspensions. Water as a bulk
filler is probably the most common excipient for liquid topicals. Excipients to
control pH and suppress microbial growth are also common. Isopropyl and
larger alcohols are commonly used to improve consistency, feel and solubiliza-
tion. Oils are used to produce emulsions and suspensions, and other compounds
provide humectant (water-retaining) and emollient (softening or soothing) prop-
erties. Topicals for ophthalmologic applications require excipients to control
tonicity, pH and sterility as well.
23.3.2.4 Injectables
Injectables are prepared for intramuscular, intravenous and subcutaneous
administration. Critical factors for these preparations are their sterility and
tolerance of the tissue at the site of injection. Water is the predominant solvent,
although non-aqueous solvents can also be used to improve the solubility of the
active ingredient. Glucose (5%) and saline (0.9%) are generally used, although
some preparations may be hypertonic. Some preparations may also incorporate
fixed oils, and nonglyceryl esters of fatty acids.
23.3.2.5 Suppositories
Excipients for these are generally fatty preparations with melting points below
body temperature and solidification points above room temperature.
23.3.2.6 Inhalants
The most obvious excipient in these preparations is the propellant. These may be
hydrocarbons, fluorocarbons, other halocarbons or compressed ambient gases
such as nitrogen and carbon dioxide. Other major excipients in inhalants can
also include solid or liquid components, for example, a filler such as lactose is
often used in dry powder inhalers.
796
biological origin, the nature of which can vary. Some matrix variation may
therefore result between lots for either of these reasons. The list of compounds
that can potentially be used as fillers in pharmaceutical preparations is substan-
tial (see Table 23.6). In many cases fillers will not dissolve in a sample and will
therefore be present as a discontinuous phase that may compete with the
extraction phase for the drug.
23.3.3.2 Capsules
These small single-dose containers are pre-manufactured and then filled with
either a solid or liquid medication. They are manufactured primarily from
gelatin. Common additives are sucrose, to increase hardness, and glycerin or
sorbitol to increase softness. As a protein, it would be expected that gelatin could
bind to some drug compounds. Gelatin content should not be permitted to vary
significantly in samples for analysis, without the use of a suitable internal
standard.
23.3.3.3 Coatings
Of the wide variety of tablet coatings used, the proteins and polymers would be
expected to compete with the extraction phase for the analyte of interest, while
the waxes and paraffins have the potential to change the sample polarity, and
hence impact partition coefficients. If the total amount of coating present is low,
the effect may not be significant.
23.3.3.4 Flavorings
Pharmaceuticals have for the most part, left the era where nasty-tasting medi-
cine was considered strong, and therefore a positive characteristic. The public
now expects oral pharmaceuticals to be palatable, particularly in pharma-
ceuticals for children. Flavor preference and consistency for pharmaceuticals are
areas of significant concern. Sucrose and glucose are commonly used sweeteners,
and glycerin may be used as both a sweetener and a solvent. Other sweeteners
may be present because of the use of natural flavorings. Sorbitol, xylitol and arti-
ficial sweeteners may be used where there is concern over disturbing the carbo-
hydrate control of diabetics. In addition to its use to provide desired tonicity or
solubility, sodium chloride may also be used to improve palatability. Many fruit
flavors prepared as syrups from natural sources are used in flavorings. However,
with the significant variability that exists in these preparations it is difficult to
precisely predict their impact on extraction, where their concentrations are
high. Synthetic flavorings by contrast, are more consistent and cost effective.
There is however a common consensus, whether founded or not, that natural fla-
vors taste better and are safer than synthetics, a fact that may limit the use of
synthetics. Where significant variation in flavoring composition exists between
samples extraction efficiency may be variable and hence external calibration
impractical. Standard addition or the use of an internal standard may be
required. Essential oils are also commonly used in flavoring, and like natural
797
fruit flavors, can be a rather ill-defined mixture of many compounds. Because
they consist mainly of compounds with low boiling points, they could interfere
with headspace extraction of pharmaceuticals by solid sorbents.
23.3.3.5 Colorings
While color in pharmaceuticals may be looked upon as a matter of aesthetics
rather than of necessity, the adding of color for psychological or aesthetic rea-
sons has been shown to influence therapeutic response. Coloring also enhances
the effect of flavoring. Most importantly, color (and tablet shape) can also give a
unique identity to a product, reducing confusion, errors and noncompliance. Col-
orings used in pharmaceuticals can be divided into two classes. Dyes are soluble
colored materials that go into true solution, generally in water. Pigments on the
other hand are colored particulate material. A subclass of the pigments is the
'lakes', which are prepared by the tight adsorption of dyes to otherwise un-
colored insoluble particulate materials such as alumina. Particulate coloring
agents, if present in significant amounts, may compete with the extraction phase
for affinity for the active ingredient. Where amounts vary from one sample to
another, some consideration will have to be given to this in analytical method
development.
23.3.3.7. Buffers
The pH of a pharmaceutical product typically must be controlled, either to
optimize solubility, bioavailability or stability, or for compatibility with the site
of administration. A salt of a drug compound is more water-soluble than the free
acid or free base form, and upon dissolution is less likely to result in a strongly
acidic or basic solution. Non-ionic species however are more lipid-soluble and
better absorbed. It is therefore critical in many drug formulations that buffering
components are included to achieve an appropriate balance between water and
lipid solubility so that a compound reaches the required absorptive surface, with
an adequate concentration of the lipid-soluble moiety to permit absorption. In
the case of injectable medications, solutions for intramuscular and subcutane-
ous injection must be close to the pH of the tissue into which they are injected.
Modest deviation from physiologic pH is better tolerated for solutions for slow
intravenous infusion, particularly where the molarity (buffering capacity) is low.
Topicals for skin application are slightly acidic to avoid skin irritations, intra-
vaginal preparations typically have a pH of 4-5 to maintain normal vaginal flora,
798
and ophthalmologic topicals require physiologic pH. The pH appropriate for
optimum formulation and/or bioavailability may not be appropriate for optimal
extraction, and significant pH adjustment may be required to counter the
buffering capacity of the excipients, and provide for adequate extraction
efficiency.
23.3.3.8 Suspending, emulsifying and stiffening agents
These excipients are used in both liquid and solid dosage forms, typically in a
small quantity. Many of these compounds are surfactants that act as suspending
and emulsifying agents to aid in the development of solutions and topical
formulations of materials that may otherwise be difficult to prepare in an
acceptable form. They may also be used in aqueous solutions to be inhaled, to
hydrate and liquify bronchiopulmonary secretions or to help clear debris. They
also have cleansing and antimicrobial action and may be used as preservatives.
The surfactants are either ionic or non-ionic. Other compounds that may be used
include mucilages and gums, methylcellulose, paraffin, fatty alcohols and lubri-
cants such as talc and edible oils. All of these compounds could potentially
change the polarity of a sample, and the partitioning of an analyte between a
sample and an extraction phase. If only a very low concentration of one of these
excipients is employed, the impact on extraction efficiency/precision would be
low. Where the concentration is higher, the primary concern would be in
ensuring consistency between samples, and between samples and standards.
Poor precision would result if there were variation in the nature and/or concen-
tration of these excipients.
23.3.3.9 Preservatives
There are three primary classes of preservatives, antimicrobials, antioxidants
and chelating agents. The surfactants discussed previously may also be employ-
ed as preservatives. Antimicrobials include such things as phenol and thymol,
thimerosal, parabens, chlorobutanol and benzyl alcohol. Concentrations range
from about 0.1-10%. Compounds present at the high end of this range may
interfere with extraction. Antioxidants include butylated hydroxy anisol and
butylated hydroxy toluene at about 0.02% for non-aqueous formulations, and
sulfites at the ppm level for aqueous formulations. At these levels, interference
with extraction is unlikely. The primary chelating agent used is ethylenediamine
tetraacetic acid (EDTA). It is typically the disodium salt that is employed. Unless
the compound being extracted is a metal species, interference with extraction is
unlikely.
799
analytes often exist at low concentration in these samples. Thus, sample prepa-
ration is usually necessary in order to extract and isolate the analytes of interest
from these complex matrices. Many sample preparation approaches such as LLE
[25-30,94], SFE [31], SPE [26-37,94-100], SPME [36-46,94,101-106], affinity
sorbent extraction [97,107,108], membrane-based extraction [34,35,98,
109-111] and others are used for drug analysis. Such steps may be the only
sample preparation or quite often they are used in combination with other
methods. In this section, the characteristics of these techniques and their
applications in clinical and pharmaceutical analysis are discussed. Applications
of these sample preparation techniques reported during the past five or six years
are summarized in Tables 23.7-10.
Protein in biological fluids or tissues must often be removed from the sample
matrix before a suitable extraction process is performed. If the analytes have
high protein-binding capacities, the adopted protein-removal protocol should
provide for release of these analytes from protein so that they may not be lost.
Deproteinization approaches include the use of (1) extreme temperatures,
osmotic or ionic strength, or pH conditions; (2) organic precipitation agents; (3)
proteolytic enzymes. Protein precipitation, ultrafiltration and dialysis can be
used to remove protein from biological samples. Common deproteinization
procedures involve protein denaturation by acids such as trichloroacetic acid
and sulfosalicylic acid, the use of extreme temperature of 90-100°C or freezing-
thawing cycles, and co-precipitation with bases such as zinc hydroxide. However,
methods using such precipitants rarely exceed 80% recovery, frequently less.
On the other hand, the use of water-miscible organic solvents has several
advantages. The mild conditions used under the precipitation process minimize
the possibility of decomposition of labile drugs. Furthermore, appropriate
solvents may be selected so that they are compatible with the mobile phases used
in HPLC analysis. Commonly used solvents include methanol, ethanol, acetone
and acetonitrile. Methanol is often preferred because it produces a flocculent
precipitate to give a clearer supernatant suitable for direct injection. Aceto-
nitrile is the most popular, but there can be problems of poor recovery and also
late eluting peaks. There are other problems that should be considered. Protein
precipitation may not be complete. Some globulins can remain in solution, and
react in the analytical system. Furthermore, such procedures by their nature are
dilutional and are best utilized for high concentration determinants. The pre-
cipitants can react in the analytical system, eluting as a peak in HPLC or
interfering in a spectrophotometric assay. The membrane filters used for ultra-
filtration are efficient, although the analyst should ensure that binding of the
drug to the membrane is not significant. The technique of ultrafiltration may be
used to measure total drug in serum if it is only marginally protein bound.
Ultrafiltration works best for acidic and neutral drugs; basic drugs tend to bind
800
to the membrane. In addition, care must be taken to not compromise the
membrane integrity during centrifugation.
Dialysis can be used to separate an analyte from the matrix by diffusion
through a semi-permeable membrane rather than centrifugal force with ultra-
filtration. Dialysis is the classic procedure for protein removal from small
molecular weight analytes; the automation of this technique in continuous flow
analyzers revolutionized clinical chemistry enabling an explosion of throughput
to occur in that area. The combination of this technique with a trace enrichment
column is available commercially for linkage to an HPLC (e.g., Gibson ASTED®
system). These membrane-based techniques are also described in Section 23.4.7.
801
TABLE 23.7
Selected applications for drug analysis using liquid-liquid extraction
Analyte Matrix Extraction solvent Analytical Ref.
method*
Thiopentane Plasma Pentane HPLC-UV 112
Sameridine Plasma Hexane GC-NPD 113
Eliprodil Plasma, urine Hexane HPLC-FL 114
Methylenedioxy Serum, tablet Hexane HPLC-FL 115
amphetamine
Tricyclic antidepressants Serum, plasma Hexane HPLC-UV 116
Candidates &metabolite Plasma Hexane LC/MS/MS 117
Propranolol, furosemide Plasma Hexane/diethyl ether HPLC-UV 118
HIV protease inhibitors Plasma Hexane/ethyl acetate HPLC-UV 119
Deramciclane &metabolite Plasma Hexane/diethyl ether LC/MS 120
Atovaquone Plasma Hexane/isoamyl alcohol HPLC-UV 121
Amitriptyline, Serum Hexane/isoamyl alcohol GC-NPD 122
noramitriptyline
Citalopram Plasma Hexane/isoamyl alcohol HPLC-FL 123
Clozapine &metabolite Plasma Hexane/isoamyl alcohol HPLC-UV 124
Clemastine Plasma Toluene GC-NPD 125
Disopyramide &metabolite Plasma Toluene CE-UV 126
Dextrophane, guaifenesin Plasma Chloroform HPLC-FL 127
Lamotrigine Serum Chloroform HPLC-UV 128
ThioTEPA, Plasma Chloroform GC-NPD 129
cyclophosphamide
Flunitrazepam Hair Chloroform/ diethyl ether GC/MS 130
Nalbuphine Plasma Chloroform/isopropanol HPLC-ECD 131
Bezitramide &metabolite Urine Chloroform/isopropanol GC-NPD 132
Ranitidine Serum Dichloromethane HPLC-UV 133
Cetirizine Plasma Dichloromethane HPLC-UV 134
Albendazole &metabolite Plasma Dichloromethane CE-UV 135
Disopyramide &metabolite Plasma, urine Dichloromethane HPLC-UV 136
Anabolic steroids Urine Dichloromethane HPLC-DAD 137
Phenmetrazine Urine 1-Chlorobutane GC/MS 138
Benzodiazepines Whole blood 1-Chlorobutane HPLC-DAD 139
Arbidol Plasma tert.-Butyl methyl ether HPLC-UV 140
Losigamone Plasma tert.-Butyl methyl ether HPLC-DAD 141
HIV protease inhibitors Plasma tert.-Butyl methyl ether HPLC-UV 142
Staurosporine Plasma Diisopropyl ether HPLC-FL 143
Triazolam Muscle Diethyl ether GC/MS 144
Efavirenz Plasma Diethyl ether HPLC-UV 145
Docetaxel Plasma Diethyl ether HPLC-UV 146
Vinorelbine Plasma, blood Diethyl ether HPLC-FL 147
Rapamycin Whole blood Diethyl ether/butyl chloride HPLC-UV 148
Cimetidine Plasma Ethyl acetate HPLC-UV 149
Nitrofurantoin Plasma, urine Ethyl acetate HPLC-UV 150
Penem antibiotics Plasma, urine Ethyl acetate HPLC-UV 151
Captopril Plasma Ethyl acetate HPLC-UV 152
continued
802
TABLE 23.7 (continuation)
acetate, and their solvent mixes are used for the LLE of various drugs. There is
little to distinguish these solvents in terms of their extraction power, although
the polar solvents will often also give a higher background. A serious drawback is
the large consumption of pure solvents. There is usually a considerable cost for
the acquisition and disposal of these solvents and many classical methods
demand chlorinated and/or fluorinated solvents, the use of which is at odds with
current environmental awareness and legislation. Macek et al. [123] used hex-
ane-isoamyl alcohol (98:2 v/v) for the LLE of citalopram from human plasma,
with analysis by HPLC with FL after back-extraction to 0.02 M hydrochloric
acid. Bouley et al. [142] developed a rapid, sensitive, and specific HPLC method
for the simultaneous TDM of HIV-protease inhibitors (indinavir, neldinavir,
ritonavir and saquinavir) in human plasma, after LLE with tert-butyl methyl
ether and a sequential washing of the reconstituted sample with hexane. A
number of flow-system LLE approaches have been presented based on mixing
the aqueous and organic phases in a tube coil and their subsequent separation.
These approaches are not widely used in sample preparation for chromato-
graphic analysis. On the other hand, semi-automated LLE methods using
96-deep-well plates have been developed for drug analysis [161,162]. All liquid
transfers during the sample preparation were automated, and the extraction
time was relatively short. The details of other LLE techniques are also described
in well-documented reviews [13,25-30,94].
803
contributes to efficient penetration of porous materials. An attractive property
is that the solvent power may be tuned by varying the temperature or pressure of
the media. Most commonly the pressure is varied, allowing the temperature to
be kept low (<40°C), ensuring that no analyte degradation occurs during extrac-
tion. Carbon dioxide or CO2-rich mixtures have been used almost exclusively as
the extraction media, because of ease of manipulation, good solvent strength and
compatibility with solutes. Carbon dioxide (with critical temperature and
pressure of 31°C and 73.8 bar, respectively) is inert, non-toxic, non-flammable,
non-corrosive, odorless and inexpensive [163]. The diffusion properties of super-
critical CO 2 allow the rapid penetration and extraction of samples, and the low
critical temperature allows good stability of thermally labile materials [163].
Hence, the use of hazardous organic solvents can be kept to a minimum in the
SFE procedure [164]. The use of ammonia or hydrocarbons as supercritical
fluids is less convenient and safe than CO2 . The solvating power of supercritical
CO2, particularly for the extraction of polar compounds, may be increased by the
addition of modifier substances, which are usually organic solvents that are
added to the source of compressed fluid before the pump or to the extraction gas
after it is compressed (e.g., CO2 mixed with methanol) [164-166. Gradient
elution, as opposed to the use of pre-mixed solvents gives shorter analysis times
[166,167]. The use of modifiers does limit the choice of detectors although there
are a large number to choose from. For example UV detectors, flame ionization
detectors, nitrogen-phosphorous detectors, electron capture detectors and mass
spectrometers are all commonly employed for these analyses. Adjusting parame-
ters such as pressure, temperature and the amount/composition of modifier can
readily alter the selectivity of the extraction [163] as can the humidity of the
sample. Generally a partial dehydration allows for a faster extraction [167]. The
density of the extracting solvent can also be altered by pressure changes, giving
the advantage of being able to alter the solubility parameter (selectivity). Also
the nature of the matrix material affects the extraction. Generally a rapid and
complete extraction depends upon the matrix particles being small. An increase
in the porosity of the matrix will also lead to improved extraction [167].
To date, most studies involving the SFE of illicit drugs and pharmaceuticals
have been applied to the extraction of biological materials, solid and semisolid
dosage forms and bulk drug substances. Opiates [168], cocaine and its metabo-
lite benzoylecgonine [169] and amphetamines [170] were extracted from hair
samples using CO 2 as the supercritical fluid modified with the polar mixture of
tetraethylamine, methanol and water. Simmons and Stewart [1711 extracted
benzodiazepine, an anabolic agent and a non-steroidal anti-inflammatory from
water and serum using a supercritical CO2 mobile phase. Scott and Oliver [172]
used SFE for the extraction of temazepam from whole blood samples. SFE has
also been used for the extraction of various pharmaceuticals from tablets,
capsules, creams, ointments, aqueous matrices and infusions [173]. Lawrence et
al. [174] employed SFE in extracting benzodiazepines from the matrices of
standard dosage forms. Scalia et al. [175] reported on the potential of SFE for the
804
isolation of vitamins A and E and their esters from tablet matrices. Tena et al.
[176] demonstrated an improved SFE of sulphonamides from solid supports
using supercritical CO 2 and methanol modified CO2 as mobile phases. Howard et
al. [177] used SFE for the sample preparation of sustained release felodipine
tablets. Masuda et al. [178] designed an SFE-supercritical chromatography-UV
system for quantitative analysis of retinal palmitate and tocopherol acetate in a
hydrophobic ointment. Recently, Wang et al. [179] reported an in situ SFE and
chemical derivatization procedure followed by GC/MS for the determination of
amphetamines in urine. Brewer et al. [180] also demonstrated SFE using CO 2
modified with methanol for the extraction of cocaine benzoylecgonine, codeine
and morphine from human hair sample in combination with GC/MS. This
method was faster and gave higher recoveries than the conventional acid
hydrolysis method. The details of other SFE techniques are also described in
well-documented reviews [31,164,173,181].
805
from a 500-mg cartridge is about 0.6-1.2 ml. Using a solvent that is too strong
will result in the elution of unnecessary sample components that are more
strongly retained than the analytes, whereas a solvent that is too weak will
result in excessive elution solvent volumes, which negate the advantage of
reducing solvent consumption with SPE cartridges. Sometimes, a desired
solvent strength may have to be obtained by blending appropriate amounts of
miscible solvents. If a water-immiscible eluant is selected and the final analysis
is to be performed with GC, a "drying" procedure should be applied between the
rinsing and the eluting steps. It has been reported that the combination of
vacuum and a small amount of methanol can produce a "dry" eluate without
causing a substantial loss of drugs.
The fundamental mechanisms of the retention are based on: (1) nonpolar
interaction, (2) polar interaction and (3) ionic interaction, and a wide selection of
SPE products is possible in the combination of these interactions using commer-
cial SPE sorbents. The primary decision for analysts in SPE method develop-
ment is the selection of the type of sorbent to obtain optimal extraction. Various
SPE sorbents, diatomaceous earth Extrelut®, Chem Elut®, Bond Elut Certify®
and Chromabond® mixed-mode columns, from different manufactures are now
widely used. Mixed-mode sorbents and restricted access matrix sorbents have
also been introduced as well as the emerging selective sorbents such as immuno-
sorbents or molecularly imprinted polymers (see Section 23.4.6). The mixed-
phase extraction columns (Bond-Elut Certify®, Chromabond®, Isolute HCX®
and TSC®) show good recoveries and allow retention of all functional groups
and differing polarities. Bond-Elut Certify® is a mixed phase of C8 and SCX,
with the functions of both hydrophobicity and ion exchange, and it is suitable for
the extraction of basic drugs that have been dissociated, in blood and urine
samples. SPE discs are also a useful and rapid way to extract drugs from liquid
specimens. In the conventional packed type solid phase, it is difficult to elute the
analyte of interest using minial solvent, if the organic solvent composition is not
raised to around 100%. Therefore, it is necessary to evaporate to dryness and
redissolve in the HPLC mobile phase. The Empore® disk cartridge has a
membrane structure, and the elution can be achieved with HPLC mobile phase.
It is then possible to directly inject into the HPLC system. New formats have also
been introduced, e.g., the 96-well SPE plates [182-184] and the microfibers for
SPME (see Section 23.4.5) for drug analysis. Further details of the SPE tech-
nique are also described by Poole in Chapter 12.
SPE can be performed off-line, with the sample preparation being separated
from the subsequent chromatographic analysis, or on-line by direct connection
to the chromatographic system. On-line techniques do not require further
handling of the samples between the trace-enrichment and the separation step
and, therefore are highly suitable for fully automated techniques. With on-line
fully-automatic solid phase extraction equipment, a column switching system is
incorporated, a precolumn is fixed in the sample loop of the injector, and the
pre-concentration is carried out through the load of a comparatively large
806
sample and the retention of the analyte of interest in the solid phase. Next, the
solute is eluted and introduced into the separation column, by switching the
injection valve to the injection position. Special attention for hyphenation has
been given to on-line SPE-HPLC followed with various detection modes, which
represents a fast, modern and reliable approach of trace analysis. Important
advantages are a decrease in the risk of sample contamination, the removal of
analyte losses by evaporation and finally the transfer and analysis of the entirety
of the extracted species. In contrast to off-line SPE where only an aliquot of the
extract is injected into the chromatograph, the analysis of the complete sample
allows the sample volume to be dramatically reduced. It has been surprising to
see that gas chromatography (GC) has been the preferred method for analytical
chemists for a long time but that the on-line coupling of SPE with LC became the
first robust on-line technique. This is explained by the good compatibility of the
LC aqueous mobile phases with the SPE of biological or environmental samples
that are mainly aqueous. On-line coupling of SPE with GC is more difficult
because of the inherent incompatibility between the aqueous nature of the SPE
step and the requirement for dryness in the GC system. Much work has been
done in this area and automatic devices have recently become available. D. Wells
and T.L. Lloyd also describe the details of an automated technique for clinical
and pharmaceutical analysis in Chapter 24.
SPE is applied for isolation, concentration, and clean up of a target substance
from a biological specimen, which may be a complicated matrix like serum,
urine, saliva, tissues, etc. Selected applications of SPE techniques reported
recently are summarized in Table 23.8. C18 is the most popular sorbent for drug
analysis. Prieto et al. [193] used C8 and C18 SPE cartridges for the extraction of
the antihypertensive drug cilazapril and its active metabolite cilazaprilat from
pharmaceuticals and urine samples, with analysis by HPLC-DAD or ED. Fluni-
trazepam (Rohypnol), and its metabolites were extracted with RP-18 [196] and
Bond-Elut Certify® [242] from plasma and urine samples, respectively. Bevalot
et al. [221] reported a LC/MS method for the analysis of doping agents and
corticosteroids in hairs, by methanolic extraction followed by SPE on a C18
cartridge. Cocaine and other opiates were also extracted from hair [239], plasma
and urine samples [238] using Bond-Elute Certify®. In addition, SPE was
applied for the clean-up of milk samples [206,220,240]. The details of other SPE
techniques are also described in well-documented reviews [26-37,94-106].
807
TABLE 23.8
Selected applications for drug analysis using solid-phase extraction
Analyte Matrix SPE sorbent* Hyphenated Ref.
analysis*
Almokalant Plasma C2 HPLC-FL 185
Haloperidol, perphenazine Serum C2 HPLC-UV 186
Citalopram Plasma C4 HPLC-FL 187
Pyrimethamine, Blood, urine C8 HPLC-UV 188
sulphadoxine
Modafinil Plasma C8 HPLC-UV 189
Cilagapril &metabolite Pharmaceuti, cal, urine C8 HPLC-ED 190
Atovaquone Plasma, who]le blood C8 HPLC-UV 191
Olanzapine &metabolite Plasma C8 HPLC-ED 192
Cilazapril &metabolite Pharmaceuti cals, C8 / C18 HPLC-ED, 193
urine DAD
Verapamil, norverapamil Urine C8 / C18 HPLC-FL 136
Gilbendamide Plasma RP-8 / RP-18 HPLC-UV 194
Flunitrazepam & Plasma RP-18 HPLC-UV 195
metabolite
Anabolic steroids Urine C18 HPLC-UV 196
Benzodiazepines Plasma, serum C18 HPLC-DAD 197
Illicit drugs Urine C18 LCMS 198
Fluconazole Plasma C18 MECC-UV 199
P-Blockers Blood, urine C18 GC/MS 200
Drugs Hair C18 HPLC-DAD 201
Methotrexate &metabolite Urine C18 HPLC-UV 202
Doxorubicin, Plasma C18 HPLC-FL 203
prochlorperazine
Fluticasone propionate Plasma C18 LC/MS 204
Iornoxicam Plasma C18 HPLC-UV 205
Phenytoin Plasma, milk C18 HPLC-UV 206
HIV-protease inhibitors Serum C18 HPLC-UV 207
Budesonide Plasma C18 LC/MS/MS 208
Cefotaxime Pharmaceutica]Is, C18 HPLC-UV 209
urine
Sirolimus Blood C18 LC/MS/MS 210
Opiate agonists Serum, blood, urine C18 HPLC-UV 211
HIV-protease inhibitors Plasma C18 HPLC-UV 212,
213
Mefloquine &metabolite Plasma. Serum, blood C18 HPLC-UV 214
Chlorambucil Plasma, serum C18 LCMS 215
Olanzapine Serum C18 LC/MS 216
Omeprazole Serum C18 HPLC-UV 217
Caffeine Medicinal prescription C18 HPLC-UV 218
Vitamins Pharmaceuticals C18 HPLC-UV 219
Carbamazepine & Plasma, milk C18 HPLC-UV 220
metabolite
Corticosteroids Hair C18 LC/MS 221
Retinoids Pharmaceuticals C18 HPLC-FL 222
continued
808
TABLE 23.8 (continuation)
preparation time, solvent purchase and disposal cost, and can improve the
detection limits. It has been used routinely in combination with gas chromatog-
raphy (GC) and GC/mass spectrometry (GC/MS), and successfully applied to a
wide variety of compounds, especially for the extraction of volatile and semi-
volatile organic pollutants from water samples. SPME was also introduced for
direct coupling with high performance liquid chromatography (HPLC) and
LC/MS in order to analyze weakly volatile or thermally labile compounds not
amenable GC or GC/MS. An SPME/HPLC interface equipped with a special
desorption chamber is utilized for solvent desorption prior to HPLC analysis in
place of thermal desorption in the injection port of a GC. Moreover, a new
SPME/HPLC system known as in-tube SPME, was recently developed using an
open tubular fused silica capillary as the SPME device instead of the SPME fiber.
The main advantages of SPME are simplicity, rapidity, solvent elimination,
high sensitivity, small sample volume, lower cost and simple automation. Since
809
1995, a number of SPME methods have been developed to extract drugs from
various biological samples such as urine, serum, plasma, whole blood, saliva and
hair (Table 23.9). The fiber SPME device consists of a fiber assembly with a
built-in extraction fiber inside the needle and an assembly holder. In fiber
SPME, analytes are extracted directly from the sample onto a polymeric station-
ary phase coated on the fiber. When the fiber is inserted into the sample, the
target analytes partition from the sample matrix into the stationary phase until
equilibrium is reached. Two types of fiber SPME techniques can be used to
extract analytes: headspace (HS) SPME and direct immersion (DI) SPME. In
HS-SPME, the fiber is exposed in the headspace of gaseous, liquid or solid
samples. In DI-SPME, the fiber is directly immersed in liquid samples. Minor
variants in the method depend on whether or not derivatization is applied and in
which phase, the type of sample agitation and whether or not additives are
required to optimize extraction. The fiber with concentrated analytes is trans-
ferred to an instrument for desorption, followed by separation and quantitation.
HS- and DI-SPME techniques can be used in combination with any GC, GC/MS,
HPLC and LC/MS systems. The affinity of the fiber coating for an analyte is the
most important factor in SPME. As shown in Table 23.9, fiber coatings of
different polarities and thicknesses were selected for different drugs. Most drugs
in biological samples were extracted with 100 gm PDMS for nonpolar drugs and
either 85 gm PA or a porous polymer DVB fibre for polar drugs.
In-tube SPME using an open tubular capillary as the SPME device has been
used for coupling with HPLC or LC/MS. It is suitable for automation, and extrac-
tion, desorption and injection are performed continuously using a standard
autosampler. Automated sample handling procedures not only shorten the total
analysis time but also usually provide better accuracy and precision relative to
manual techniques. With the in-tube SPME technique, organic compounds in
aqueous samples are directly extracted from the sample into the internally
coated stationary phase of a capillary and then desorbed by introducing a moving
stream of mobile phase or static desorption solvent when the analytes are more
strongly absorbed to the capillary coating. Desorbed compounds are finally
injected onto the column for analysis.
Although the theories of fiber and in-tube SPME methods are similar, the
significant difference between these methods is that the extraction of analytes is
performed on the outer surface of the fiber for fiber SPME and in the inner sur-
face of the capillary for in-tube SPME. Commercially available SPME fibers for
drug analysis are limited, but GC capillary columns with a vast array of station-
ary phases are commercially available for in-tube SPME. Headspace fiber SPME
is suitable for the extraction of drugs in gaseous, liquid and solid samples, and
has the advantage of avoiding contact with an aggressive matrix incompatible
with the fiber. Direct immersion fiber SPME can be used to extract drugs from
both clear and cloudy liquid samples, whereas in-tube SPME is limited to the
extraction of clear liquid samples. Therefore, the headspace SPME technique is
suitable for direct extraction from whole blood samples, but direct immersion
810
TABLE 23.9
Selected applications for drug analysis using solid-phase microextraction
Analyte Matrix SPME device Extractn. Hyphenated Ref.
modea analysis*
Anorectics Urine 30 m PDMS DI GC/MS 246
Cannabinoids Hair 30 gm PDMS DI GC/MS 247
Cannabinoids Saliva 30 m PDMS DI GC/MS 248
Amphetamines Blood 100 m PDMS HS GC/MS 249
Amphetamines Urine 100 gm PDMS HS GC/MS 250
Amphetamines Urine 100 m PDMS HS GC-FID 251
Amphetamines Hair 100 gm PDMS HS GC-NPD 252
Amphetamines Hair 100/Am PDMS HS GC/MS 253
Amphetamines Hair 100 m PDMS D + HS GC/MS 254
Amphetamines, MDA, Urine 100 m PDMS HS GC/MS 255-
MDMA, MDEA 257
Amphetamines Urine 100 m PDMS D + DI GC-NPD, 258
GC/MS
Amphetamines, Urine 100 m PDMS DI GC/MS 259
MDMA
Amphetamines, MDA, Urine 100 m PDMS D + DI GC-NPD, 260
MDMA, MDEA GC/MS
Amphetamines, Whole blood 100 m PDMS D + HS GC/MS 261
fenfluramine
Methamphetamine, Urine 100,um PDMS DI GC-NPD, 262
cocaine, diazepam GC/MS
Anaesthetics Blood 100 m PDMS HS GC-FID 263
Local anaesthetics Blood 100 m PDMS HS GC/MS 264
Anaesthetics Blood 100 m PDMS DI GC-FID 265
Phencyclidine Blood, urine 100 m PDMS HS GC-SID 266
Halothane Blood 100,um PDMS HS GC/MS 267
Lidocaine Plasma 100 m PDMS DI GC-FID 268
Lidocaine Plasma 100 m PDMS DI GC-FID 269
Lidocaine Urine 100 m PDMS DI HPLC-UV 270
Tricyclic antidepress. Blood 100 am PDMS HS GC-FID 271
Tricyclic antidepress. Blood 100 am PDMS HS GC/MS 272
Tricyclic antidepress. Plasma 100 m PDMS DI GC-NPD 273
Tricyclic antidepress. Urine 100 m PDMS HS GC-FID 274
Amitriptyline Urine 100 am PDMS DI HPLC-UV 275
Valproic acid Plasma 100 m PDMS DI GC-FID 276
Diphenylmethane Blood, urine 100 m PDMS HS GC-FID 277
antihistaminics
Phenothiazines Blood, urine 100 m PDMS HS GC-FID 278
Clozapine Plasma 100 m PDMS DI GC-NPD 279
Cocaine Urine 100 gm PDMS DI GC-NPD 280
Meperidine Blood, urine 100 m PDMS HS GC-FID 281
Benzoylecgonine Urine 100 am PDMS D + HS GC/MS 282
Cannabinoids Saliva 100 am PDMS DI GC/MS 283
Methadone Hair 100zm PDMS DI GC/MS 284
Nicotine, cotinine Urine 100gAm PDMS HS GC/MS 285
continued
811
TABLE 23.9 (continuation)
812
centration. For in-tube SPME, number, volume and speed of draw/eject cycles,
and sample pH are important factors for efficient extraction. On the other hand,
the desorption of analyte from a fiber or capillary coating depends on the tem-
perature of injection port and exposure time in combination with GC or GC/MS,
or component and volume of solvent when used in combination with HPLC or
LC/MS. Therefore, these SPME parameters should be optimized when develop-
ing a new SPME method for drug analysis. SPME principles and optimization
strategies are also described by J. Pawliszyn in Chapter 13.
There are numerous references to the SPME analysis of drugs from biologi-
cal fluids and matrices. A summary of references is given in Table 23.9. Most of
the applications have been related to forensic and toxicological analysis; how-
ever, they demonstrate the potential of SPME for other drug analyses, such as
clinical, metabolic and pharmaceutical. Most of the methods shown to date have
involved HS techniques for volatile drugs. Some methods have employed DI and
derivatization for less volatile analytes. It has provided low detection limits and
excellent quantitation. Especially in the HS mode, SPME extractions offer the
potential for very clean analyses, with little to no interference from non-volatile
compounds. Because of the relatively low partition coefficients between polar
drugs and the commercially available HPLC/SPME fibres, the application of
SPME for the assay of low volatility drugs and metabolites in plasma may be
limited to those drugs with high therapeutic concentrations, in the range of 1 to
100 ,ug/ml. The situation can be improved with the use of current tandem
quadrupole LC/MS instrumentation, where concentrations approaching 1 ng/ml
in plasma can be analysed [245]. The analysis of drugs in plasma for pharmaco-
kinetic studies and TDM is an important area of biomedical analysis. Currently
however there are few practical SPME/HPLC methods available. The bulk of the
SPME applications in drug analysis to date have involved GC analysis. Amphet-
amines were extracted from urine [250,251,255-257], hair [252-254] and blood
[249] by HS-SPME with PDMS fiber under alkaline conditions. Lidocaine was
extracted by DI-SPME with PDMS fiber from plasma and urine, and analyzed by
GC-FID [268,269] and HPLC-UV [270]. PDMS/DVB [289], PA [293,294] and
CW/DVB [303] fibers were used for the extraction of benzodiazepines in plasma,
serum and urine samples by DI-SPME. Benzodiazepines were also analyzed in
combination with DI-SPME with CW/TPR and HPLC-UV [305]. Recently Yuan
et al. [307] developed an immunoaffinity (IE)-SPME technique for the specific
extraction of theophylline in serum sample. The IE-SPME fiber, covalently
immobilized a theophylline antiserum on the surface of a fused silica fiber, was
used as a selective and sensitive extraction medium. For automated SPME/
HPLC analysis, Kataoka et al. [308-311] reported automated on-line methods
for the analysis of amphetamines, ranitidine and 3-blockers by in-tube SPME
coupled with LC/MS. In these methods, Omegawax 250 capillary (Supelco,
Bellefonte, PA) was used as the extraction device. Wu et al. [312] later demon-
strated improvements obtained for the extraction of a series of P-blocker drugs
in urine and serum by the use of novel polypyrrole (PPY) polymers as compared
813
to the conventional Omegawax GC phase with in-tube. Saito et al. [315] used a
DB-1 capillary for the in-tube SPME of tricyclic antidepressants in urine, and
extraction efficiency was raised by insertion of stainless steel wire into the
capillary. Mullet et al. [316] recently developed a new in-tube SPME technique
using a polyether ether ketone (PEEK) capillary packed with propranolol MIP
particles. The details of other SPME techniques are also described in books [17]
and well-documented reviews [36-46,94,101-106].
814
to the selectivity of IAE, an important parameter is the extraction recovery,
which is linked to the breakthrough volume and to the capacity. In order to
achieve low detection limits the extraction sorbent must allow the precon-
centration of analytes from large sample volumes without breakthrough of the
analyte. Since the breakthrough volume depends on the total number of accessi-
ble binding sites and on the strength of the interaction between the analytes and
the antibodies, the selection of the materials used for the covalent bonding has
become of prime importance. These should allow a high capacity and avoid any
kind of non-specific interactions with the analytes. Immunosorbents can be used
in off-line procedures in disposable cartridges. A minimum desorption volume
should be used in order to obtain a high enrichment factor. This cannot be
achieved using the conventional aqueous elution solutions of affinity chromatog-
raphy. Organic solvents are required for the elution of small molecules. There is
also interest in coupling extraction on-line with LC [321,322]. Pichon et al. [318]
developed several silica-based immunosorbents and demonstrated that a single
immunoprecolumn could be directly desorbed on-line using a water-organic sol-
vent mixture, and was easily regenerated for re-use [318,320,321,324]. It has
been demonstrated that the coupling of immunoaffinity SPE in off-line or
on-line procedures is a powerful technique for the determination and easy quan-
tification at low levels of many organic compounds in various complex samples.
The sample handling step is greatly simplified. The selectivity in the extraction
step also allows easier identification of analytes of interest and their confirma-
tion is reinforced since trapping occurs on the basis of structure recognition.
Moreover, a clear baseline enhances the overall sensitivity of the analysis and
allows the handling of a smaller sample in comparison to non-selective extrac-
tion procedures. Hennion and Pichon describe the details of IAE in Chapter 33.
IAE coupled on-line with LC/MS has been described [325-327]. Cai and
Henion [326] have reported an automated on-line clean-up and enrichment
procedure using a commercially available workstation to perform the complete
process. Three different columns were used in the column switching system: an
IAE column packed with the bound antibodies, a trapping column and an
analytical column. The work-up process consisted of the following steps: precon-
ditioning of the IAE column, loading of the urine sample, elution of the unbound
matrix to waste, equilibration of the trapping and analytical columns with
mobile phase, elution of the bound analytes and trapping in the trapping
column, back-flushing of the trapped analytes onto the analytical column,
separation and introduction into the MS. The advantage of such a procedure is
not only the full automation but also the highly selective and sensitive detection.
For example, the LOD for the IAE-LC-LC-MS-MS detection of lysergic acid
diethylamide (LSD) in urine was 20-fold below that using off-line SPE and
LC-MS-MS [325]. IAE was also applied to LSD analysis in whole blood [328] and
hair [329] samples. IAE units (LSD ImmunoElute®) are commercially available
for the analysis [329]. Rashid et al. [330] developed morphine-specific immuno-
sorbents to concentrate morphine from urine samples prior to HPLC analysis
815
with electrochemical detection. IAE columns demonstrated remarkably high
specificity towards morphine, showing minimal binding with other opiates such
as codeine, normorphine, norcodeine, morphine-glucuronide. Recently, Clarke
et al. [331] reported ultrafast IAE of warfarin by a 2.1-mm-i.d. sandwich micro-
column that contained a 1.1-mm layer of an anti-warfarin antibody support. In
addition, the IAE technique was applied to specific clean-up of anabolic steroids
[332], P-2 agonists [333] and fluoroquinolines [334] from biological samples.
More recently, IA-SPME for theophylline analysis [307] was also developed as
unique IAE application. The details of other IAE techniques are also summa-
rized in well-documented reviews [97,107,108,335,336].
816
MIP-SPE sorbents, leading to efficient sample clean-up. The selectivity of the
extraction leads to distinctly cleaner chromatographic traces and the ability to
improve sensitivity by extraction from larger sample volumes. The analytical
performance of the MIP-SPE based method was found to be equivalent to or
better than that of a standard method based on the use of LLE for sample
clean-up [337].
As a disadvantage, MIPs are made in the presence of large amounts of
template molecules and small amounts of the imprint molecule remaining in the
resultant polymer may later leak during extraction. This has been observed in
several cases. Hence, method development must include a confirmation that
leakage of remaining traces of the imprint species does not interfere with the
assay, and increase uncertainty in the concentration determination. This is
particularly important when dealing with trace analysis. One approach to com-
pletely avoid this risk is the use of a close structural analogue of the analyte(s) of
interest for the preparation of the MIP. Provided the imprint species and the
analyte(s) can be separated by the subsequent LC or GC, which in most
instances is a valid assumption, the leakage appears as a separate peak and
presents no problem. A limitation of MIP based SPE is somewhat of a lack of
knowledge in the use of MIPs for biological samples. MIP preparation entails the
use of organic solvents and, in consequence, most studies on rebinding to
imprints have been conducted using organic solvents as the incubation medium.
A key success factor for aqueous rebinding is the ability to balance specific
binding to the imprints and non-specific binding to the polymer by hydrophobic
interactions. For each compound, (both analyte and all other components of the
sample), the observed retention is due to the sum of specific and nonspecific
binding. Hence, if the non-specific element dominates, any selectivity shown by
the imprints will be obscured. Problems with non-specific adsorption to the
polymer can be reduced by the use of small amounts of MIP, thereby reducing
the polymer surface area available for non-specific adsorption. Another means of
reducing problems with non-specific adsorption is the use of proper washing
schemes prior to elution. The rationale is that the selective imprint analyte
binding increases in strength and non-specific adsorption of hydrophobic nature
is weakened. This leads to redistribution of non-specifically bound analyte to
imprint sites and washing off of non-related structures.
MIPs are capable of molecular recognition and are stable for long-term
storage, easy to prepare and inexpensive, thus they may be considered as new
artificial affinity media. Different modes of MIP based SPE have been demon-
strated, including various modes of off-line and on-line SPE for pre-concentra-
tion or pre-treatment of analytes, conventional SPE where the MIP is packed
into columns or cartridges, and batch mode SPE where the MIP is incubated
with the sample. Several applications for the use of MIP for biological samples
have now appeared. The first MIP-SPE was reported for pentamidine, which is
used for the treatment of AIDS-related pneumonia [338]. The high selectivity of
the polymer was demonstrated with on-line sample enrichment of a urine
817
sample spiked with pentamidine. In this system, 20% (v/v) phosphate buffer in
acetonitrile was used for both loading and elution, and the adsorption (pH 5) and
desorption (pH 3) were accomplished by changing the pH of the buffer. Thus
electrostatic interaction is the main factor for the specific MIP extraction of
pentamidine. MIP-SPE of sameridine [339] and bupivacaine [340], which are
provided for anesthesia during surgery and prolonged post-operative analgesia,
was reported for batch-wise pre-concentration from plasma samples prior to GC
analysis. A non-steroidal estrogen antagonist, tamoxifen, was extracted from
biological samples by MIP-SPE [341]. As another example, this technique has
been extended to the extraction of theophylline in serum [342,343], phenytoin in
plasma [344], clenbuterol in urine [345-347] and propranolol in biological
samples [348,349]. More recently, magnetic MIP beads for drug radioligand
binding assay [350], MIP-membrane SPE for terbumeton extraction [351], and
MIP-SPME for propranolol analyses [316] were also developed as unique MIP
applications. The details of other MI techniques are also summarized in well-
documented reviews [337,352-354].
818
driven process and filtration, a pressure-driven process. The fourth technique
uses nonporous membranes and will be referred to as membrane extraction.
In porous membrane techniques the liquids on each side of the membrane
are physically connected through the pores. These membranes are used in
Donnan dialysis to separate low-molecular-mass analytes from high-molecular-
mass matrix components, leading to an efficient clean-up but no discrimination
between different small molecules. No enrichment of the small molecules is
possible, instead the analytes are diluted as the driving force of the mass transfer
process is a simple concentration difference across the membrane. Dialysis is
widely used for protein concentration, etc., in biochemistry, but has not seen
much use in sample preparation for chromatography. As one example, in the
ASTED® (automated sequential trace enrichment of dialysates) process, both
clean-up and enrichment can be performed by a combination of dialysis and SPE
and is applied to drug analysis in blood plasma. Other variations of porous
membrane techniques are microdialysis [356], which is extensively used in
neuroscience research for in vivo sampling, and electrodialysis, where an electric
field over a dialysis membrane promotes selective transport of charged com-
pounds. Lange et al. [357] reported the intracerebral microdialysis technique for
monitoring of free drug concentrations in brain extracellular fluid, in which the
knowledge of associated free drug concentrations provide information on blood-
brain barrier drug transport. Furthermore, the dialysis sampling technique was
applied to the pharmacokinetic studies of unbound cefepime in rat bile [358] and
unbound cephaloridine in rat blood [359]. In addition, a number of micro- and
nanofiltration techniques belong to the field of porous membrane techniques. In
contrast, a nonporous membrane is a liquid or solid (e.g., polymeric) phase that
is placed between two other phases, usually liquid but sometimes gaseous. One
of these phases is the sample to be processed, the donor phase. On the other side
of the membrane is the acceptor phase, where the extracted analytes are
collected and transferred to the analytical instrument. This arrangement
permits the versatile chemistry of LLE to be used and extended, which can
provide a highly effective clean-up as well as high enrichment factors. The
operations can normally be automated and in most cases there is no, or an
insignificant, use of organic solvents. The details of membrane techniques are
also described by J. Pawliszyn (Chapter 14), J.A. Jonsson (Chapter 15), T.
Kotiaho and F.R. Lauritsen (Chapter 16), and T. Matsuura (Chapter 30).
Membrane techniques have been applied to the determination of various
drugs in biological fluids. Selected applications of membrane-based sample
preparation techniques reported recently are summarized in Table 23.10. On-
line dialysis has been applied in a fully automated HPLC method for analysis of
the non-ionic X-ray contrast agent iodixanol in plasma [360]. Continuous moni-
toring of free (i.e., non protein bound) drug concentrations in different body
compartments by microdialysis represents an important step towards correct
interpretation of data in pharmacokinetic or toxicokinetic studies. This samp-
ling methodology is undoubtedly an in vivo experimental alternative to the
819
TABLE 23.10
Selected applications for drug analysis using membrane-based separation techniques
Analyte Matrix Membrane Hyphenated analysis* Ref.
Antiepileptics Plasma DM HPLC-UV 360
Benzodiazepines Plasma DM GC-FID, GC-NPD, GC/MS 361
Clozapine Plasma DM HPLC-UV 362
Sildenofil &metabolite Plasma DM HPLC-UV 363
Levosimendan Plasma DM HPLC-UV 364
Velapamil Plasma DM HPLC-UV 365
Antidepressants Plasma DM HPLC-UV 366
Anticonvulsants Plasma DM HPLC-UV 367
Dopamine, cocaine Brain DM HPLC-UV 368
Amphetamines Plasma, serum DM HPLC-FL 369
Sulphonamides Serum, urine DM CE-UV 370
5-Fluorourasil Plasma DM CE-UV 371
Iodixanol Plasma DM HPLC-UV 372
Albendazole &metabolite Plasma DM HPLC-UV 373
Cefazolin Blood, brain DM HPLC-UV 374
Propranolol Blood DM HPLC-FL 375
Drugs Blood DM LC/MS/MS 376
Cephaloridine Blood DM HPLC-UV 377
Cefsulodin Blood DM HPLC-UV 378
Cefepime Bile DM HPLC-UV 379
Tricyclic antidepressants Serum, urine DM CE-UV 370
Bambuterol Plasma SLM CE-UV 381
Basic drugs Plasma SLM CE-UV 382
Ziprasidone Plasma SLM LC.MS 383
Lamivudine, fluticasone Plasma, serum, SLM LCMS 384
urine
Local anesthetics Plasma MMLLE GC-FID 385
Ropivacaine &metabolite Urine SLM HPLC-UV 386
Methotrexate &metabolite Serum, urine SLM LC/MS/MS 387
aDM: dialysis membrane; SLM: supported liquid membrane; MMLLE: microporous membrane
LLE. *See list of abbreviations.
820
which the process is driven by difference in pH between the donor and acceptor
solutions. Then the whole volume of the acceptor solution is injected into the CE
capillary by using the double stacking procedure for large volume injection. In
SLM-CE and other drug analysis applications using microporous membrane liq-
uid-liquid extraction (MMLLE)-GC [384], all pertaining to blood plasma sam-
ples, similar high degrees of selectivity have been demonstrated. In these
applications, enrichment factors of 30-70 times were obtained. The available
volume of a blood plasma sample is limited, so the systems were optimized so
that the entire extract from each sample (<1 ml) was injected, resulting in one
chromatographic run. Recently, Rule et al. [387] reported a new membrane-
based extraction technique for LC/MS/MS determination of methotrexate and
its metabolite in human plasma and urine samples using 384-well plates con-
taining C18 particles incorporated into a glass-fiber membrane. The details of
other membrane-based extraction techniques are also summarized in well-
documented reviews [34,35,109-111, 388-395].
821
23.4.8.2 Microwave-assisted extraction
Microwave-assisted extraction (MAE) is a process for using microwave energy to
heat solvents in contact with a sample in order to partition analytes from the
sample matrix into the solvent. The ability to rapidly heat the sample solvent
mixture is inherent to MAE and is the main advantage of this technique. By
using closed vessels the extraction can be performed at elevated temperatures,
accelerating the mass transfer of target compounds from the sample matrix. A
typical extraction procedure takes 15-30 min and uses solvent volumes in the
range of 10-30 ml. In addition, sample throughput is increased as several
samples can be extracted simultaneously. In most cases analyte recovery and
reproducibility are improved compared to conventional techniques. Although
sample preparation for chemical analysis is a major activity for the pharmaceuti-
cal industry, relatively few papers have been published in this area using MAE.
MAE has been used for extraction of antibiotics such as chloramphenicol in eggs
[404], sulphamethazine in swine tissue [405], vitamins from foodstuffs [406] and
puerarin in Chinese herbal medicine [407] using a household microwave oven.
Felodipine and one of its degradation products have been extracted from tablets
[408]. By optimizing the extracting solvent (5% methanol in acetonitrile) whole
tablets could be extracted with full recovery without grinding the tablet prior to
extraction. The details of other MAE techniques are also summarized in a
well-documented review [47].
23.5 CONCLUSIONS
822
sules, coatings, flavorings, colorings, buffers, suspendings and preservatives.
Clean-up of these samples for the analysis of target drugs are commonly
performed by LLE, SPE, SPME and their combination after precipitation,
centrifugation and dialysis.
Because most analytical instruments cannot handle the matrix directly sam-
ple preparation is necessary to isolate the components of interest from the sam-
ple matrix. This however has always been a rather neglected part of analytical
chemistry, even though it is also one of the most significant sources of inaccura-
cies in the entire analytical procedure. Suitable sample preparation is an impor-
tant prerequisite for analysis of biological samples. Many of the techniques
currently used for the preparation of complex samples prior to chromatographic
or electrophoretic analysis, such as filtration, precipitation and extraction with
organic solvents, have been in use for several decades with essentially no modifi-
cation over the years. Universal LLE procedures are preferable for general
screening procedures, because substances with very different physico-chemical
properties may be isolated from heterogeneous matrices. The acceptance of new
technologies has been slow, both by the users themselves and by regulatory
agencies. In general, sample preparation is still regarded as "low tech" and in
many laboratories the associated tasks are assigned to the least trained staff,
who may often be reluctant to accept new technologies. This is one of the main
reasons why the implementation of new sample preparation techniques has been
slow and thereby sample preparation in many cases is still the most time-
consuming and often rate-limiting step of the total analytical procedure.
As described in this chapter, SFE, SPE, SPME, IAE, MD, ME, LPME and
MAE have been developed as newer sample preparation techniques. Among
these techniques, SPE is the most popular sample preparation technique for
drug analysis and has come to play a truly crucial role in laboratories all over the
world. It has also replaced older techniques to a relatively large extent. The
increased development of SPE has occurred during the past decade, with many
improvements in formats, automation and the introduction of new phases. One
reason for the general acceptance of SPE has been the pressure to decrease
organic solvent usage in laboratories, which has encouraged the requirement for
solvent-free procedures and has greatly contributed to the growth of SPE at the
expense of LLE procedures. Sample pretreatment for SPE depends on the
sample type: whole blood and tissue need deproteinization and filtration/centri-
fugation steps before application to the SPE columns, whereas for urine usually
a simple dilution step and/or centrifugation is satisfactory. However, clogging,
channeling and percolation are typical problems of SPE encountered in every-
day laboratory work. Some promising approaches in SPE are based on special
packings such as restricted access materials (RAMs) and MIPs [410]. Other
methods of sample preparation are SFE, on-column sample preparation with
column-switching techniques and on-line SPE in LC using RAMs [410,411],
LC-GC coupling [412] and membrane-based sample preparation such as
dialysis, electrodialysis, ultrafiltration. Although these methods have their own
823
merits, most of them are only found in isolated applications, and often they do
not achieve the sensitivity and selectivity of LLE and SPE. Finally, some
methods require the use of expensive equipment. Other problems are fouling of
membranes in membrane-based sample preparation and irreversible binding of
some high-molecular-mass material in on-line SPE.
SPME is becoming an attractive alternative to SPE and LLE for some
applications. SPME is a solvent-free and concentrating extraction technique and
applied for the analysis of drugs and metabolites in biological fluids. The most
striking attribute of SPME is its low total recovery as reported for many
methods. This is not unexpected because SPME is an equilibrium extraction
technique rather than an exhaustive extraction. However, several SPME meth-
ods have also been presented with high total recovery and the performance of the
majority of methods was good with regard to sensitivity, linearity, precision and
accuracy. As evidenced with SPE, immunoaffinity sorbent [307] and MIP [316]
are also expected to be applied as new sorbent materials for highly efficient
extraction of drugs from various biological samples by SPME. With the develop-
ment of more sensitive phases it may be possible to further miniaturize the
technique. As the market for SPME increases in the future this could lead to the
introduction of disposable low-cost extraction fibers (e.g. in the form of a
carousel) or tubes such as in other areas of sample preparation, e.g., SPE
multiwell plates. A more approach to the design and concept of SPME has been
recently proposed in the form of Twister®, available from Gerstel [413--415]. In
this technique, known as stir bar sorptive extraction, a coated magnetic stirring
bar is used, with the same PDMS phase as for SPME but in a thicker layer
(0.3-1.0 mm) and it is also compatible with both GC and HPLC desorption
procedures. It can theoretically improve sensitivity compared to a SPME fiber
for certain applications, but it requires the special desorption unit and is difficult
to automate. To date the Twister has only been applied to environmental
samples.
The trends are clearly to simplify the labor of sample preparation, increasing
its reliability and precision, and eliminating the clean-up step by using more
selective extraction procedures. The development of more selective sorbents will
remain an active area of research, as it is today for IAE sorbents and MIPs. IAE
has been shown as another feasible procedure for trace analysis of biological
samples. As more antibodies become available and as procedures to develop
them become more sophisticated it is likely that more methods will be developed.
MIP antibody mimics are much easier to synthesize and consequently much less
expensive. MIPs have provided a new tool in affinity separation-based SPE,
because conventional affinity sorbents are expensive. The development of SPE
based on inexpensive MIPs allows a disposable type of affinity-SPE, potentially
making the availability of affinity extraction with excellent selectivity wide-
spread. Many applications can be expected in clinical, pharmaceutical and
biochemical analyses for molecular recognition techniques such as affinity
extraction media using immobilized antibodies, substitution of the immuno-
824
assay for antibody extraction, sensor material which may replace the biochip,
novel carriers for capillary electrophoresis, selective membranes, and other new
intelligent materials for molecular recognition.
Dialysis is less used as a sample preparation technique in this field and is
being succeeded by batch-like techniques such as 96-well plate SPE, which
allows several hundreds of samples to be processed per day. It can therefore be
expected that the importance of on-line dialysis for biomedical samples will
decrease in the next few years. Non-porous membrane extraction has been
shown to be compatible with HPLC, GC and CE and applicable to many different
sample types. Its robustness appears to be comparable to that of on-line dialysis,
although the leakage of membrane-incorporated solvent could be problematic
for routine use. Method development should not be more difficult than for liquid
extraction (followed by a back-extraction). However, despite its favorable chara-
cteristics, membrane extraction has seen little use outside the university
laboratory. An important reason could be that the technique has not been com-
mercialized so far, while most users do not want to manufacture analytical sys-
tems themselves, but rely upon commercially available equipment. As long as
this does not change, there is little chance that membrane extraction will have a
similar breakthrough as dialysis when it was commercialized some 10 years ago.
Users have had an increasing interest in the automation of sample prepara-
tion for faster sample-preparation procedures, improved precision and more
cost-effective analyses. The key attractive features of automated sample prepa-
ration techniques include miniaturization, throughput, reproducibility and
traceability. In the last decade, new concepts have been developed, which allow
the on-line coupling of sample preparation devices to separation and detection
systems, and which are especially designed for automation. In the future,
advances in better integration of sampling/sample preparation and instrumen-
tal analysis will allow wider use of automated on-line analysis in clinical and
pharmaceutical analysis.
REFERENCES
1 R.H. Liu and D.E. Gadzala (Eds.), Handbook of Drug Analysis: Applications in
Forensic and Clinical Laboratory. American Chemical Society, Washington, DC,
1997.
2 K. Valko (Ed.), SeparationMethods in Drug Synthesis and Purification.Handbook of
Analytical Separation, Vol. 1. Elsevier, Amsterdam, 2000.
3 J.H. Lin and A.Y. Lu, Pharmacol.Rev., 49 (1997) 403.
4 A. Marzo, Pharmacol.Res., 36 (1997) 425.
5 K.A. Jackson and S.E. Rosenbaum, Drug Dev. Ind. Pharm., 24 (1998) 1155.
6 S. Ulrich, C. Wurthmann, M. Brosz and F.P. Meyer, Clin. Pharmacokinet., 34 (1998)
227.
7 M.H.H. Ensom, G.A. Davis, C.D. Cropp and R.J. Ensom, Clin. Pharmacokinet., 34
(1998) 265.
8 S. Ulrich, I. Schrodter, G. Partscht and P. Baumann, Psychopharmakotherapie,7
(2000) 2.
825
9 V. Iyengar, J. Kumpulainen, K. Okamoto, M. Morita, S. Hirai and S. Nomoto, Prog.
Clin. Biol. Res., 380 (1993) 329.
10 H.H. Maurer, J. Chromatogr.B, 580 (1992) 3.
11 H.H. Maurer, J. Chromatogr. B, 713 (1998) 3.
12 H.H. Maurer, J. Chromatogr.B, 733 (1999) 3.
13 O.H. Drummer, J. Chromatogr. B, 733 (1999) 27.
14 A. Polettini, J. Chromatogr. B, 733 (1999) 47.
15 P. Marquet and G. Lachiatre, J. Chromatogr.B, 733 (1999) 93.
16 M.J. Bogusz, Forensic Science. Handbook of Analytical Separation, Vol. 2. Elsevier,
Amsterdam, 2000.
17 J. Pawliszyn, Solid Phase Microextraction: Theory and Practice. Wiley-VCH, New
York, 1997.
18 N.T. Crosby and I. Patel, General Principles of Good Sampling Practice. The Royal
Society of Chemistry, Cambridge, UK, 1995.
19 M. Stoeppler (Ed.), Sampling and Sample Preparation. Springer-Verlag, Berlin,
Germany, 1997.
20 E. Riva, R. Merati and L. Cavenaghi, J. Chromatogr., 553 (1991) 35.
21 A. Haque and J.T. Stewart, Biomed. Chromatogr., 13 (1999) 51.
22 D. Stevenson and I.D. Wilson (Eds.), Sample Preparation for Biomedical and
EnvironmentalAnalysis. Plenum Press, New York, 1994.
23 E. Thurman and M. Mills, Solid Phase Extraction: Principles and Practice.
Wiley-VCH, New York, 1998.
24 J. Fritz, Analytical Solid Phase Extraction. Wiley-VCH, New York, 1999.
25 S.P. Bjergaard, K.E. Rasmussen and T.G. Halvorsen, J. Chromatogr.A, 902 (2000)
91.
26 R.D. McDowall, J. Chromatogr.B, 492 (1989) 3.
27 W. Thormann, S. Lienhard and P. Wernly, J. Chromatogr., 636 (1993) 137.
28 D.K. Lloyd, J. Chromatogr.A, 735 (1996) 29.
29 O.H. Drummer, J. Chromatogr. B 713 (1998) 201.
30 T. Kumazawa and 0. Suzuki, J. Chromatogr.B, 747 (2000) 241.
31 R.E. Majors, LC-GC, 13 (1995) 547.
32 R.A. de Zeeuw, J. Chromatogr.B, 689 (1997) 71.
33 J.P. Franke and R.A. de Zeeuw, J. Chromatogr.B, 713 (1998) 51.
34 J.R. Veraart, H. Lingeman and U.A.Th. Brinkman, J. Chromatogr.A, 856 (1999)
483.
35 M. Gilar, E.S.P. Bouvier and B.J. Compton, J. Chromatogr.A, 909 (2001) 111.
36 F. Degel, Clin. Biochem., 29 (1996) 529
37 Z.E. Penton, Adv. Chromatogr., 37 (1997) 205.
38 L. Junting, C. Peng and 0. Suzuki, Forensic Sci. Int., 97 (1998) 93.
39 F. Sporkert and F. Pragst, ForensicSci. Int., 107 (2000) 129.
40 G. Theodoridis, E.H.M. Koster and G.J. de Jong, J. Chromatogr.B Biomed. Sci. 745
(2000) 49.
41 N.H. Snow, J. Chromatogr.A, 885 (2000) 445.
42 H.L. Lord and J. Pawliszyn, J. Chromatogr. A, 902 (2000) 17.
43 S. Ulrich, J. Chromatogr. A, 902 (2000) 167.
44 G.A. Mills and V. Walker, J. Chromatogr. A, 902 (2000) 267.
45 K.G. Furton, J. Wang, Y-L. Hsu, J. Walton and J.R. Almirall, J. Chromatogr.Sci., 38
(2000) 297.
46 H. Kataoka, H.L. Lord and J. Pawliszyn, in: I.D. Wilson, T.D. Adlard and C.F. Poole
and M. Cook (Eds.), Encyclopedia of Separation Science. Academic Press, London,
2000, p. 4153.
47 C.S. Eskilsson and E. Bjorklund, J. Chromatogr.A, 902 (2000) 227.
826
48 J. Segura, R. Ventura and C. Jurado, J. Chromatogr. B, 713 (1998) 61.
49 T. Inoue and S. Seta, ForensicSci. Rev., 4 (1992) 89.
50 I.D. Watson, in: I.D. Wilson (Ed.), Sample Preparation for Biomedical and
Environmental Analysis. Plenum Press, New York, p.71, 1994.
51 D.A. Kidwell, J.C. Holland and S. Athanaselis, J. Chromatogr. B, 713 (1998) 111.
52 C. Staub, J. Chromatogr. B 733 (1999) 119.
53 P. Kintz and N. Samyn, J. Chromatogr. B, 733 (1999) 137.
54 Y. Nakayama, J. Chromatogr.B, 733 (1999) 161.
55 Y. Gaillard and G. Pepin, J. Chromatogr.B, 733 (1999) 231.
56 M.A. Huestis, J.M. Oyler, E.J. Cone, A.T. Wstadik, D. Schoendorfer and R.E. Joseph,
J. Chromatogr. B, 733 (1999) 247.
57 M. Yashiki, N. Nagasawa, T. Kojima, T. Miyazaki and Y. Iwasaki, Jpn. J. Forensic
Toxicol., 13 (1995) 17.
58 B. Dehon, L. Humbert, L. Devisme, M. Stievenart, D. Mathieu and N. Houdret, J.
Anal. Toxicol., 24 (2000) 22.
59 D.B. Menkes, R.C. Howard, G.F.S. Spears and E.R. Cairns, Psychopharmacology,103
(1991) 277.
60 E.J. Cone, J. Anal. Toxicol., 14 (1990) 1.
61 W. Schramm, R.H. Smith, P.A. Craig and D.A. Kidwell, J. Anal. Toxicol., 16 (1992) 1.
62 L.M. North, N.D. Gaudette, M.L. Cordeiro, J.H. Fitchen, S.L. Davidson and M.S.
Hindahl, Ann. N.Y. Acad. Sci., 694 (1993) 332.
63 M.L. Cordeiro, C.S. Turpin and S.A. McAdams, Ann. N.Y. Acad. Sci., 694 (1993) 330.
64 T. Thieme, J. Fitchen, F. Bartos, M. Salinsky, T. Green and W. Holden, Ann. N.Y.
Acad. Sci., 694 (1993) 337.
65 M. Navazesh, Ann. N.Y. Acad. Sci., 694 (1993) 72.
66 M. Slavik, J. Wu, N. Brown, A. Williams, S. Wagner and J. Slavik, Ann. N.Y. Acad.
Sci., 694 (1993) 317.
67 W. Schramm, R.H. Smith and P.A. Craig, Ann. N.Y. Acad. Sci., 694 (1993) 311.
68 W. Schramm and R.H. Smith, Clin. Chem., 37 (1991) 114.
69 W. Schramm, O.F. Pomerleau, C.S. Pomerleau and H.E. Grates, Prevent. Med., 21
(1992) 63.
70 F. Harding (Ed.), Milk Quality. Blackie, Glasgow, 1995, 75.
71 M.R. Harkey, Forensic Sci. Int., 63 (1993) 9.
72 T. Mieczkowski, Microgram, 28 (1995) 193.
73 D.L. Blank and D.A. Kidwell, ForensicSci. Int., 70 (1995) 13.
74 P. Kintz and P. Mangin, ForensicSci. Int., 70 (1995) 3.
75 D.A. Kidwell and D.L. Blank, Forensic Sci. Int., 63 (1993) 137.
76 C. Jurado, P. Kintz, M. Menendez and M. Repetto, Int. J. Leg. Med., 110 (1997) 159.
77 H. Sachs, ForensicSci. Int., 84 (1997) 7.
78 C. Staub, ForensicSci. Int., 84 (1997) 295.
79 M. Chiarotti, ForensicSci. Int., 63 (1993) 161.
80 Y. Gaillard, F. Vaysette, A. Balland, G. PBpin, J. Chromatogr.B, 735 (1999) 189.
81 W.L. Wang and E.J. Cone, ForensicSci. Int., 70 (1995) 39.
82 D.L. Blank and D.A. Kidwell, Forensic Sci. Int., 63 (1993) 145.
83 M. Burns and R.C. Baselt, J. Anal. Toxicol., 19 (1995) 41.
84 D.P. Conner, R.G. Almirez, P. Rhyne, K. Zamani, B.J. Bolden and C.C. Peck, Skin
Pharmacol., 2 (1989) 155.
85 D. Conner, E. Millora, K. Zamani, D. Nix, R. Almirez, P. Rhyne-Kirsch and C. Peck,
J. Invest. Dermatol., 2 (1990) 186.
86 G. Skopp, L. Potsch, H.-P. Eser and M.R. Miller, J. Forensic Sci. 41 6 (1996) 933.
87 I. Sunshine and J.P. Sutliff, in: S.H.Y. Wong and I. Sunshine (Eds.), Handbook of
Analytical Therapeutic Drug Monitoring and Toxicology. CRC Press, Boca Raton,
827
FL, 1997, p. 253.
88 F.P. Smith and D.A. Kidwell, Forensic Sci. Int., 83 (1996) 179.
89 D.A. Kidwell, M.A. Blanco and F.P. Smith, Forensic Sci. Int., 84 (1997) 75.
90 C. Grote and J. Pawliszyn, Anal. Chem., 69 (1997) 587.
91 R. Hypler, Crhova, J. Gasparic, Z. Zaddk, M. Cikova and V. Balasova, J. Chromatogr.
B, 739 (2000) 183.
92 K.E. Kongshaug, S. Pedersen-Bjergaard, K.E. Rasmussen and M. Krogh, Chroma-
tographia, 50 (1999) 247.
93 J.C. Coumbaros, K.P. Kirkbride and G. Klass, J. Forensic Sci., 44 (1999) 1237.
94 J.J. Vreuls, A.J.H. Louter and U.A.Th. Brinkman, J. Chromatogr.A, 856 (1999) 279.
95 D. Barcelo and M.-C. Hennion, Anal. Chim. Acta, 318 (1995) 1.
96 R.W. Fedeniuk and P.J. Shand, J. Chromatogr.A, 812 (1998) 3.
97 M.-C. Hennion, J. Chromatogr.A, 856 (1999) 3.
98 P.R. Haddad, P. Doble and M. Macka, J. Chromatogr.A, 856 (1999) 145.
99 M.E.L. Gonzalez and L.V.P. Arribas, J. Chromatogr. A, 902 (2000) 3.
100 D. Martinez, M.J. Cugat, F. Borrull and M. Calull, J. Chromatogr.A, 902 (2000) 65.
101 J. Pawliszyn, J. Chromatogr.Sci., 38 (2000) 270.
102 H. Kataoka, H. Lord and J. Pawliszyn, J. Chromatogr. A, 880 (2000) 35.
103 H. Lord and J. Pawliszyn, J. Chromatogr. A, 885 (2000) 153.
104 J. Beltran, F.J. Lopez and F. Hernandez, J. Chromatogr.A, 885 (2000) 389.
105 M. de F. Alpendurada, J. Chromatogr. A, 889 (2000) 3.
106 H. Kataoka, Anal. Bioanal. Chem., 373 (2002) 31.
107 N. Delaunay, V. Pichon and M.-C. Hennion, J. Chromatogr. B, 745 (2000) 15.
108 D. Stevenson, J. Chromatogr.B, 745 (2000) 39.
109 N.C. van de Merbel, J. Chromatogr. A, 856 (1999) 55.
110 B.M. Cordero, J.L. Perez, C.G. Pinto, M.E.F. Laespada, R.C. Martinez and E.R.
Gonzalo, J. Chronatogr.A, 902 (2000) 195.
111 J.A. Jonsson and L. Mathiasson, J. Chromatogr.A, 902 (2000) 205.
112 D.J. Jones, K.T. Nguyen, M.J. McLeish, D.P. Crankshaw and D.J. Morgan, J.
Chromatogr.B, 675 (1996) 174.
113 C. N.-Hoog, P. Neidenstrom and T. Arvidsson, J. Chromatogr.B, 760 (2001) 123.
114 B. Malavasi, M. Ripamonti, A. Rouchouse and V. Ascalone, J. Chromatogr.A, 729
(1996) 323.
115 F. Sadeghipour and J.L. Veuthey, J. Chromatogr. A, 787 (1997) 137.
116 R. Theurillat and W. Thormann, J. Pharm.Biomed. Anal., 18 (1998) 751.
117 Y.Q. Xia, D.B. Whigan and M. Jemal, Rapid Commun. Mass Spectrom., 13 (1999)
1611.
118 M. Walshe, M.T. Kelly and M.R. Smyth, J. Pharm. Biomed. Anal., 14 (1996) 475.
119 K.C. Marsh, E. Eiden and E. McDonald, J. Chromatogr.B, 704 (1997) 307.
120 A. Tolokan, V. Horvath, G. Horvai, A. Egresi, K.B. Nemes and I. Kiebovich, J.
Chromatogr. A, 896 (2000) 279.
121 S.L. Hannan, G.A. Ridout and A.E. Jones, J. Chromatogr. B, 678 (1996) 297.
122 S. Ulrich, T. Isensee and U. Pester, J. Chromatogr. B, 685 (1996) 81.
123 J. Macek, P. Ptacek and J. Klima, J. Chromatogr. B, 755 (2001) 279.
124 A. Llerena, R. Berecz, M.J. Norberto and A. dela Rubia, J. Chromatogr. B, 755 (2001)
349.
125 N.N. Davydova, S.U. Yasuda, R.L. Woosley and I.W. Wainer, J. Chromatogr. B, 744
(2000) 177.
126 V.A. Jabor, V.L. Lanchote and P.L. Bonato, Electrophoresis,22 (2001) 1406.
127 S. Stavchansky, S. Demirbas, L. Reyderman and C.K. Chai, J. Pharm. Biomed. Anal.,
13 (1995) 919.
128 A.P. Hart, S. M.-Proo, W. Blackwell andA. Dasgupta, Ther. DrugMonit., 19 (1997) 431.
828
129 A.D. Huitma, M.M. Tibben, T. Kerbusch, J.W. Zwikker, S. Rodenhuis and J.H.
Beijnen, J. Chromatogr.B, 716 (1998) 177.
130 V. Cirimeie, P. Kintz, C. Staub and P. Mangin, Forensic Sci. Int., 84 (1997) 189.
131 F. de Cazanove, J.M. Kinowski, M. Audran, A. Rochette and F. Bressolle, J. Chroma-
togr.B, 690 (1997) 203.
132 S.M. De Baere, W.E. Lambert and A.P. De Leenheer, J. Anal. Toxicol., 22 (1998) 18.
133 C.L.-Calull, L.G.-Capdevila, C. Arroyo and J. Bonal, J. Chromatogr.B, 693 (1997)
228.
134 J. Macek, P. Ptacek and J. Klima, J. Chromatogr. B, 736 (1999) 231.
135 A. Prochazkova, M. Chouki, R. Theurillat and W. Thormann, Electrophoresis, 21
(2000) 729.
136 R. Bortocan, V.L. Lanchote, E.J. Cesarino and P.S. Bonato, J. Chromatogr.B, 744
(2000) 299.
137 R.G. Lumbreras, D.P. Trapero and R.I. Hornillos, J. Chromatogr. B, 754 (2001) 419.
138 A. Dasgupta and C.E. Mahle, J. Forensic. Sci., 42 (1997) 937.
139 A. El Mahjoub, C. Staub, J. Pharm.Biomed. Anal., 23 (2000) 1057.
140 R. Metz, P. Muth, M. Ferger, V.G. Kukes and H. Vergin, J. Chromatogr.A, 810 (1998)
63.
141 C.D. Torchin, W.D. Yonekawa, I.M. Kapetanovic and H.J. Kupferberg, J. Chroma-
togr. B, 724 (1999) 101.
142 M. Bouley, C. Briere, C. Padoin and 0. Petitjean and M. Tod, Ther. Drug Monit., 23
(2001) 56.
143 R. van Gijn, 0. van Tellingen, J.J. de Clippeleir, M.J. Hillebrand, E. Boven, J.B.
Vermorken, W.W. ten Bokkel Huinink, S. Schwertz, P. Graf and J.H. Beijnen, J.
Chromatogr. B, 667 (1995) 269.
144 H. Inoue, Y. Maeno, M. Iwasa, J. Monma and R. Matoba, J. Chromatogr. B, 701
(1997) 47.
145 P. Villani, M. Pregnolato, S. Banfo, M. Rettani, D. Burroni, E. Seminari, R. Maserati
and M.B. Regazzi, Ther. Drug Monit., 21 (1999) 346.
146 J. Ciccolini, J. Catalin, M.F. Blachon and A. Durand, J. Chromatogr.B, 759 (2001)
299.
147 A. Gauvin, F. Pinguet, S. Poujol. C. Astre and F. Bressolle, J. Chromatogr.B, 748
(2000) 389.
148 J.O. Svensson, C. Brattstrom and J. Sawe, Ther. Drug Monit., 19 (1997) 112.
149 M.T. Kelly, D. McGuirk and F.J. Bloomfield, J. Chromatogr.B, 668 (1995) 117.
150 P. Muth, R. Metz, B. Siems, W.W. Bolten and H. Vergin, J. Chromatogr.A, 679 (1996)
169.
151 M. Matsuoka, T. Maki, K. Banno and T. Sato, J. Chromatogr.B, 685 (1996) 237.
152 C. Arroyo, C. L.-Calull, L. G.-Capodevila, I. Gich, M. Barbanoj and J. Bonal, J.
Chromatogr. B, 688 (1997) 339.
153 G.P. Kaijser, J.H. Beijnen, A. Bult, H.J. Keizer and W.J. Underberg, J. Chromatogr.
B, 690 (1997) 131.
154 T. Kerbusch, M.J. Jeuken, J. Derraz, J.W. van Putten, A.D. Huitema and J.H.
Beijnen, Ther. Drug Monit., 22 (2000) 613.
155 C. Guitton, J.M. Kinowski, R. Aznar and F. Bressolle, J. Chromatogr.B, 690 (1997)
211.
156 R. Ohta, T. Amano, K. Yamashita and M. Motohashi, J. Chromatogr.B, 704 (1997)
325.
157 C. Sottani, R. Turci, L. Perbellini and C. Minoia, Rapid Commun. Mass Spectrom., 12
(1998) 1063.
158 H.J. Mascher, J. Chromatogr.A, 812 (1998) 339.
159 B.B. Rasmussen and K. Brosen, J. Chromatogr.B, 676 (1996) 169.
829
160 A. Sparreboom, P. de Bruijn, K. Nooter, W.J. Loos, G. Stoter and J. Verweij, J.
Chromatogr. B, 705 (1998) 159.
161 S. Steinbormer and J. Henion, Anal. Chem., 71 (1999) 2340.
162 N. Zhang, K.L. Hoffman, W. Li and D.T. Rossi, J. Pharm. Biomed. Anal., 22 (2000)
131.
163 C. Staub, Forensic Sci. Int., 84 (1997) 295.
164 J.W. King and J.E. France, in: B. Wenclawiak (Ed.), Analysis with Supercritical
Fluids: Extraction and Chromatography.Springer-Verlag, Bellin, 1992, p. 32.
165 B.J. Perrigo and B.P. Joynt, Can. Soc. Forens. Sci. J., 25 (1992) 70.
166 R.F. Cross, J.L. Ezzell and B.E. Richter, J. Chromatogr. Sci., 31 (1993) 162.
167 C. Staub, P. Edder and J.L. Veuthey, in: P. Kintz (Ed.) Drug Testing in Hair. CRC
Press, Boca Raton, FL, 1996 p. 1 2 2 .
168 V. Cirimele, V. Kintz, R. Majdalani and P. Mangin, J. Chromatogr.B, 673 (1995) 173.
169 J.F. Morrison, S.N. Chesler, W.J. Yoo and C.M. Selavka, Anal. Chem., 70 (1998) 163.
170 D.L. Allen and J.S. Oliver, Forensic Sci. Int., 107 (2000) 191.
171 B.R. Simmons and J.T. Stewart, J. Chromatogr. B, 688 (1997) 291.
172 K.S. Scott and J.S. Oliver, J. Anal. Toxicol., 70 (1997) 297.
173 L. Karlsson, A. Torstensson and L.T. Taylor, J. Pharm. Biomed. Anal., 15 (1997) 601.
174 J.K. Lawrence, A.K. Larson and I.R. Lebbett, Anal. Chim. Acta, 288 (1994) 123.
175 S. Scalia, G. Ruberto and F. Bonina, J. Pharm. Sci., 84 (1995) 433.
176 M.T. Tena, M.D. Luque de Castro and M. Valcarcel, Chromatographia,40 (1995) 197.
177 A.L. Howard, M.C. Shah, M.A. Brooks, J.T. Strode and L.T. Taylor, J. Pharm. Sci.,
83 (1994) 1537.
178 M. Masuda, S. Koike, M. Handa, K. Sagara and T. Mizutani, Anal. Sci., 9 (1993) 29.
179 S.M. Wang, Y.S. Giang and Y.C. Ling, J. Chromatogr. B, 759 (2001) 17.
180 W.E. Brewer, R.C. Galipo, K.W. Sellers and S.L. Morgan, Anal. Chem., 73 (2001)
2371.
181 C. Radcliffe, K. Maguire and B. Lockwood, J. Biochem. Biophys. Meth., 43 (2000) 261.
182 M. Jemal, M. Huang, Y. Mao, D. Whigan and A. Schuster, Rapid Commun. Mass
Spectrom., 14 (2000) 1023.
183 G. Rule, M. Chapple and J. Henion, Anal. Chem., 73 (2001) 439.
184 M.C. Rouan, C. Buffet, L. Masson, F. Marfil, H. Humbert, G. Maurer, J. Chromatogr.
B, 754 (2001) 45.
185 H. Svennberg and P.O. Lagerstrom, J. Chromatogr. B, 689 (1997) 371.
186 H.R. Angelo and A. Petersen, Ther. Drug Monit., 23 (2001) 157.
187 D. Ohman, B. Carlsson and B. Norlander, J. Chromatogr. B, 753 (2001) 365.
188 H. Astier, C. Renard, V. Cheminel, O. Soares, C. Mounier, F. Peyron and J.F.
Chaulet, J. Chromatogr. B, 698 (1997) 217.
189 P. Burmat and F. Do B. Robles, J. Chromatogr. B, 706 (1998) 295.
190 J.A. Prietro, R.M. Jimenez and R.M. Alonso, J. Chromatogr. B, 714 (1998) 285.
191 N. Lindegardh and Y. Bergqvist, J. Chromatogr. B, 744 (2000) 9.
192 M.A. Raggi, G. Casamenti, R. Mandrioli and V. Volterra, J. Chromatogr. B, 750
(2001) 137.
193 J.A. Prietro, R.M. Jimenez, R.M. Alonso and E. Oritiz, J. Chromatogr. B, 754 (2001)
23.
194 E.B.A. Adjaye and G.K. Shiu, J. Chromatogr. B, 707 (1998) 161.
195 J.R. Valdes Santurio and E. Gonzalez Porto, J. Chromatogr. B, 682 (1996) 364.
196 A.E.I. Mahjoub and C. Staub, J. Chromatogr. B, 754 (2001) 271.
197 K.K. Akerman, J. Jolkkonen, M. Parvainen and I. Penttila, Clin. Chem., 42 (1996)
1412.
198 M. Tatsuno, M. Nishikawa, M. Katagi and H. Tsuchihashi, J. Anal. Toxicol., 20
(1996) 281.
830
199 F. von Heeren, R. Tanner, R. Theurillat and W. Thormann, J. Chromatogr.A, 745
(1996) 165.
200 S.B. Black, A.M. Stenhouse and R.C. Hansson, J. Chromatogr.B, 685 (1996) 67.
201 Y. Gaillard and G. Pepin, J. Chromatogr.A, 762 (1997) 251.
202 T. Hirai, S. Matsumoto and I. Kishi, J. Chromatogr. B, 690 (1997) 267.
203 C. Mou, N. Ganju, K.S. Sridhar and A. Krishan, J. Chromatogr. B, 703 (1997) 217.
204 Y.N. Li, B.N. Tattam, K.F. Brown and J.P. Seale, J. Pharm.Biomed. Anal., 16 (1997)
447.
205 S. R.-Welte and P. Dittrich, J. Chromatogr.B, 707 (1998) 151.
206 R. Shimoyama, T. Ohkubo, K. Sugawara, T. Ogasawara, T. Ozaki, A. Kagiya and Y.
Saito, J. Pharm. Biomed. Anal., 17 (1998) 863.
207 S. Frappier, D. Breilh, E. Diarte, B. Ba, D. Ducint, J.L. Pellegrin and M.C. Saux, J.
Chromatogr. B, 714 (1998) 384.
208 K. Kronkvist, M. Gustavsson, A.K. Wendel and H. Jaegfeldt, J. Chromatogr.A, 823
(1998) 401.
209 L.G. Martinez, P. C.-Falco, A. S.-Cabeza and F. B.-Reig, J. Chromatogr.B, 718 (1998)
143.
210 P.J. Taylor and A.G. Johnson, J. Chromatogr.B, 718 (1998) 251.
211 M.J. Bogusz, R.D. Maier, K.D. Kruger and U. Kohls, J. Anal. Toxicol., 22 (1998) 549.
212 M.L. Foisy and J.P. Sommadossi, J. Chromatogr.B, 721 (1999) 239.
213 C. Marzolini, A. Telenti, T. Buclin, J. Biollaz and L.A. Decosterd, J. Chromatogr.B,
740 (2000) 43.
214 M.D. Green, Y. Bergqvist, D.L. Mount, S. Corbett and M.J. D'Souza, J. Chromatogr.
B, 727 (1999) 159.
215 I.D. Davies, J.P. Allanson and R.C. Causon, J. Chromatogr.B, 732 (1999) 173.
216 M.J. Bogusz, K.D. Kruger, R.D. Maier, R. Erkwoh and F. Tuchtenhagen, J. Chroma-
togr. B, 732 (1999) 257.
217 G.G. Encina, R. Farran and L. Martinez, J. Pharm.Biomed. Anal., 21 (1999) 371.
218 Y.Y. Ku, K.C. Wen, L.K. Ho and Y.S. Chang, J. Pharm.Biomed. Anal., 20 (1999) 351.
219 P. Moreno and V. Salvado, J. Chromatogr.A, 870 (2000) 207.
220 R. Shimoyama, T. Ohkubo and K. Sugawara, Ann. Clin. Biochem., 37 (2000) 210.
221 F. Bevalot, Y. Gaillard and M.A. Lhermitte and G. Pepin, J. Chromatogr.B, 740
(2000) 227.
222 R. Gatti, M.G. Gioia and V. Cavrini, J. Pharm. Biomed. Anal., 23 (2000) 147.
223 P.J. Taylor, P. Salm, S.V. Lynch and P.I. Pillans, Ther. Drug Monit., 22 (2000) 608.
224 S.H. Gan and R. Ismail, J. Chromatogr. B, 759 (2001) 325.
225 P. Koopmans, C.E. Gidding, S.S. de Graaf and D.R. Uges, Ther. Drug Monit., 23
(2001) 406.
226 Z. Wang, R.M. Dsida and M.J. Avram, J. Chromatogr. B, 755 (2001) 383.
227 K.R. Kim and H.R. Yoon, J. Chromatogr. B, 682 (1996) 55.
228 W. Weinmann and M. Svoboda, J. Anal. Toxicol., 22 (1998) 319.
229 A. Marchese, C. McHugh, J. Bi and H. Kehler, J. Mass Spectrom., 33 (1998) 1071.
230 J.M. Morales, C.H. Jung, A. Alarcon and A. Barreda, J. Chromatogr. B, 746 (2000)
133.
231 H. Ohta, Y. Seto and N. Tsunoda, J. Chromatogr.B, 691 (1997) 351.
232 E. Kramer and K.A. Kovar, J. Chromatogr. B, 731 (1999) 167.
233 S. Paterson, R. Cordero, S. McCulloch and P. Houldsworth, Ann. Clin. Biochem., 37
(2000) 690.
234 M. Kobylinska and K. Kobylinska, J. Chromatogr. B, 744 (2000) 207.
235 V. R.-Gutierrez, M.E. Juan, A. Cert and J.M. Planas, Anal. Chem., 72 (2000) 4458.
236 E. Mikami, T. Goto, T. Ohno, H. Matsumoto, K. Inagaki, H. Ishihara and M. Nishida,
J. Chromatogr. B, 744 (2000) 81.
831
237 J.O. Svensson, A. Bruchfeld, R. Schvarcz and L. Stahle, Ther. DrugMonit., 22 (2000)
215.
238 P. Femandez, N. Lafuente, A.M. Bermejo, M. L.-Rivadulla and A. Cruz, J. Anal.
Toxicol., 20 (1996) 224.
239 K.M. Hold, D.G. Wilkins, D.E. Rollins, R.E. Joseph and E.J. Cone, J. Chromatogr.
Sci., 36 (1998) 125.
240 S. Dodd, A. Buist, G.D. Burrows, K.P. Maguire and T.R. Norman, J. Chromatogr.B,
730 (1999) 249.
241 S.A. Reuschel, S.E. Percey, S. Liu, D.M. Eades and R.L. Foltz, J. Anal. Toxicol., 23
(1999) 306.
242 H. Nguyen and D.R. Nau, J. Anal. Toxicol., 24 (2000) 37.
243 B.H. Forngren, N. Tyrefors, K.E. Markides and B. Langstrom, J. Chromatogr.B, 748
(2000) 189.
244 R. Swart, P. Koivisto and K.E. Markides, J. Chromatogr.B, 736 (1999) 247.
245 H. Lord, R. Grant and J. Pawliszyn, unpublished data.
246 M. Chiarotti, S. Strano-Rossi and R. Marsili, J. Microcolumn Sep., 9 (1997) 249.
247 S. Strano-Rossi and M. Chiarotti, J. Anal. Toxicol., 23 (1999) 7.
248 N. Dfucci, N. De Giovanni, M. Chiarotti and S. Scarlata, ForensicSci. Int., 119 (2001)
318.
249 N. Nagasawa, M. Yashiki, Y. Iwasaki, K. Hara and T. Kojima, Forensic Sci. Int., 78
(1996) 95.
250 M. Yashiki, T. Kojima, T. Miyazaki, N. Nagasawa, Y. Iwasaki and K. Hara, Forensic
Sci. Int., 76 (1995) 169.
251 H.L. Lord and J. Pawliszyn, Anal. Chem., 69 (1997) 3899.
252 I. Koide, O. Noguchi, K. Okada, A. Yokoyama, H. Oda, S. Yamamoto and H. Kataoka,
J. Chromatogr.B, 707 (1998) 99.
253 J. Liu, K. Hara, S. Kashimura, M. Kashiwagi and M. Kageura, J. Chromatogr.B, 758
(2001) 95.
254 J. Liu, K. Hara, S. Kashimura, M. Kashiwagi and M. Kageura, J. Chromatogr.B, 758
(2001) 95.
255 F. Centini, A. Masti and I.B. Comparini, ForensicSci. Int., 83 (1996) 161.
256 C. Battu, P. Marquet, A.L. Fauconnet, E. Lacassie and G. Lachatre, J. Chromatogr.
Sci., 36 (1998) 1.
257 C. Jurado, M.P. Gimenez, T. Soriano, M. Menedez and M. Repetto, J. Anal. Toxicol.,
24 (2000) 11.
258 H.G. Ugland, M. Krogh and K.E. Rasmussen, J. Chromatogr.B, 701 (1997) 29.
259 S.-W. Myung, H.-K. Min, S. Kim, M. Kim, J.-B. Cho, K. Taek-Jae, J. Chromatogr.B,
716 (1998) 359.
260 H.G. Ugland, M. Krogh and K.E. Rasmussen, J. Pharm. Biomed. Anal., 19 (1999)
463.
261 A. Namera, M. Yashiki, J. Liu, K. Okajima, K. Hara, T. Imamura and T. Kojima,
Forensic Sci. Int., 109 (2000) 215.
262 Y. Makino, T. Takagi, S. Ohta and M. Hirobe, Chromatography, 18 (1997) 195.
263 T. Kumazawa, X.-P. Lee, K. Sato, H. Seno, A. Ishii and 0. Suzuki, Jpn. J. Forensic
Toxiol., 13 (1995) 182.
264 T. Watanabe, A. Namera, M. Yashiki, Y. Iwasaki, T. Kojima, J. Chromatogr.B, 709
(1998) 225.
265 T. Kumazawa, K. Sato, H. Seno, A. Ishii and 0. Suzuki, Chromatographia,43 (1996)
59.
266 A. Ishii, H. Seno, T. Kumazawa, K. Watanabe, H. Hattori and 0. Suzuki, Chroma-
tographia, 43 (1996) 331.
267 F. Musshoff, H. Junker and B. Madea, J. Anal. Toxicol., 24 (2000) 372.
832
268 E.H.M. Koster, C. Wemes, J.B. Morsink and G.J. de Jong, J. Chromatogr. B, 739
(2000) 175.
269 E.H. Koster, C. Wemes, J.B. Morsink and G.J. de Jong, J. Chromatogr. B, 739 (2000)
175.
270 E.H.M. Koster, N.S.K. Hofman and G.J. de Jong, Chromatographia,47 (1999) 678.
271 X.-P. Lee, T. Kumazawa, K. Sato and 0. Suzuki, J. Chromatogr. Sci., 35 (1997) 302.
272 A. Namera, T. Watanabe, M. Yashiki, Y. Iwasaki, T. Kojima and J. Anal. Toxicol., 22
(1998) 396.
273 S. Ulrich, J. Martens, J. Chromatogr. B, 696 (1997) 217.
274 T. Kumazawa, X.-P. Lee, M.-C. Tsai, H. Seno, A. Ishii and K. Sato, Jpn. J. Forensic
Toxiol., 13 (1995) 25.
275 K. Jinno, M. Kawazoe and M. Hayashida, Chromatographia,52 (2000) 309.
276 M. Krogh, K. Johansen, F. Tonnesen and K.E. Rasmussen, J. Chromatogr. B, 673
(1995) 299.
277 M. Nishikawa, H. Seno, A. Ishii, O. Suzuki, T. Kumazawa, K. Watanabe and H.
Hattori, J. Chromatogr. Sci., 35 (1997) 275.
278 H. Seno, T. Kumazawa, A. Ishii, M. Nishikawa, K. Watanabe, H. Hattori and 0.
Suzuki, Jpn. J. ForensicToxicol., 14 (1996) 30.
279 S. Ulrich, S. Kruggel, H. Weigmann and C. Hiemke, J. Chromatogr.B, 731 (1999)
231.
280 T. Kumazawa, K. Watanabe, K. Sato, H. Seno, A. Ishii and 0. Suzuki, Jpn. J.
ForensicToxicol., 13 (1995) 207.
281 H. Seno, T. Kumazawa, A. Ishii, M. Nishikawa, H. Hattori, O. Suzuki, Jpn. J.
ForensicToxicol., 13 (1995) 211.
282 B.J. Hall, A.R. Parikh and J.S. Brodbelt, J. ForensicSci., 44 (1999) 527.
283 B.J. Hall, M. Satterfield-Doerr, A.R. Parikh and J.S. Brodbelt, Anal. Chem., 70
(1998) 1788.
284 A.C.S. Lucas, A.M. Bermejo, M.J. Tabernero, P. Fernandez and S. S.-Rossi, Forensic
Sci. Int., 107 (2000) 225.
285 M. Yashiki, N. Nagasawa, T. Kojima, T. Miyazaki and Y. Iwasaki, Jpn. J. Forensic
Toxicol., 13 (1995) 17.
286 F.Y. Guan, H. Seno, A. Ishii, K. Watanabe, T. Kumazawa, H. Hattori and 0. Suzuki,
J. Anal. Toxicol., 23 (1999) 54.
287 A. Ishii, H. Seno, T. Kumazawa, M. Nishikawa, K. Watanabe, H. Hattori and 0.
Suzuki, Jpn. J. Forensic Toxicol., 14 (1996) 228.
288 K.E. Kongshaug, S.P. Bjergaard, K.E. Rasmussen and M. Krogh, Chromatographia,
50 (1999) 247.
289 P. Okeyo, S.M. Rentz and N.H. Snow, J. High Resolut. Chromatogr., 20 (1997) 171.
290 F. Sporkert and F. Pragst, J Chromatogr.B, 746 (2000) 255.
291 D. Poli, E. Bergamaschi, P. Manini, R. Andreoli and A. Mutti, J. Chromatogr.B, 732
(1999) 115.
292 U. Staerk and W.R. Kulpmann, J. Chromatogr.B, 745 (2000) 399.
293 K.J. Reubsaet, H.R. Norli, P. Hemmersbach and K.E. Rasmussen, J. Pharm.Biomed.
Anal., 18 (1998) 667.
294 K. Jinno, T. Taniguchi and M. Hayashida, J. Pharm.Biomed. Anal., 17 (1998) 1081.
295 M. Krogh, H. Grefslie and K.E. Rasmussen, J. Chromatogr.B, 689 (1997) 357.
296 P. Okeyo, S.M. Rentz and N.H. Snow, J. High Resol. Chromatogr., 20 (1997) 171.
297 T. Kumazawa, H. Seno, K. Watanabe, H. Hattori, A. Ishii, K. Sato and 0. Suzuki, J.
Mass Spectrom., 35 (2000) 1091.
298 P.D. Okeyo and N.H. Snow, J. Microcol. Sep., 10 (1998) 551.
299 T. Felix, B.J. Hall and J.S. Brodbelt, Anal. Chim. Acta, 371 (1998) 195.
300 F. Sporkert and F. Pragst, J. Anal. Toxicol., 24 (2000) 316.
833
301 B.J. Hall and J.S. Brodbelt, J. Chromatogr. A, 772 (1997) 275.
302 D.A. Volmer and J.P.M. Hui, Rapid Commun. Mass Spectrom., 11 (1997) 1926.
303 Y. Luo, L. Pan and J. Pawliszyn, J. Microcolumn Sep., 10 (1998) 193.
304 C.M. Lock, L. Chen and D.A. Volmer, Rapid Commun. Mass Spectrom., 13 (1999)
1744.
305 K. Jinno, M. Taniguchi, H. Sawada and M. Hayashida, Analusis, 26 (1998) M27.
306 S. Li and S.G. Weber, Anal. Chem., 69 (1997) 1217.
307 H. Yuan, W.M. Mullett and J. Pawliszyn, Analyst, 126 (2001) 1456.
308 H. Kataoka, H.L. Lord and J. Pawliszyn, J. Anal. Toxicol., 24 (2000) 257.
309 H. Kataoka, H.L. Lord and J. Pawliszyn, J. Chromatogr. B, 731 (1999) 353.
310 H. Kataoka, H.L. Lord and J. Pawliszyn, Anal. Chem., 71 (1999) 4237.
311 H. Kataoka, H.L. Lord, S. Yamamoto, S. Narimatsu and J. Pawliszyn, J. Microcol.
Sep., 12 (2000) 493.
312 J. Wu, H.L. Lord, J. Pawliszyn and H. Kataoka, J. Microcol. Sep., 12 (2000) 255.
313 J. Wu, H.L. Lord and J. Pawliszyn, Talanta, 54 (2001) 655.
314 H. Yuan, Z. Mester, H.L. Lord and J. Pawliszyn, J. Anal. Toxicol., 24 (2000) 718.
315 Y. Saito, M. Kawazoe, M. Hayashida and K. Jinno, Analyst, 125 (2000) 807.
316 W. Mullett, P. Martin and J. Pawliszyn, Anal. Chem., 73 (2001) 2383.
317 D.S. Hage, J. Chromatogr. B, 715 (1998) 3.
318 V. Pichon, M. Bouzige and M.-C. Hennion, Anal. Chim. Acta, 376 (1998) 21.
319 J.M. Van Emon, C.L. Gerlack and K. Bowman, J. Chromatogr.B, 715 (1998) 211.
320 V. Pichon, M. Bouzige, C. Miege and M.-C. Hennion, Trends Anal. Chem., 18 (1999)
219.
321 M.-C. Hennion, J. Chromatogr.A, 856 (1999) 3.
322 M.-C. Hennion, C. C.-D.-Coumes and V. Pichon, J. Chromatogr.A, 823 (1998) 147.
323 V. Pichon, L. Chen and M.-C. Hennion, Anal. Chim. Acta, 311 (1995) 429.
324 V. Pichon, L. Chen, N. Durand, F. Le Goffic and M.-C. Hennion, J. Chromatogr.A,
725 (1996) 107.
325 J. Cai and J. Henion, Anal. Chem., 68 (1996) 72.
326 J. Cai and J. Henion, J. Chromatogr.B, 691 (1997) 357.
327 M.L. Nedved, S. H.-Goudarzi, B. Ganem and J.D. Henion, Anal. Chem., 68 (1996)
4228.
328 S. Kerrigan and D.E. Brooks, J. Immunol. Meth., 224 (1999) 11.
329 J. Rohrich, S. Zorntlein and J. Becker, Forensic Sci. Int., 107 (2000) 181.
330 B.A. Rashid, G.W. Aheme, M.F. Katmeh, P. Kwasowski and D. Stevenson, J.
Chromatogr. A, 797 (1998) 245.
331 W. Clarke, A.R. Chowdhuri and D.S. Hage, Anal. Chem., 73 (2001) 2157.
332 M. Dubois, X. Taillieu, Y. Colemonts, B. Lansival, J. De Graeve and P. Delahaut,
Analyst, 123 (1998) 2611.
333 A. Koole, J. Bosman, J.P. Franke and R.A. de Zeeuw, J. Chromatogr.B, 726 (1999) 149.
334 C.K. Holtzapple, S.A. Buckley and L.H. Stanker, J. Chromatogr.B, 754 (2001) 1.
335 S.A. Reuschel, D. Eades and R.L. Foltz, J. Chromatogr.B, 733 (1999) 145.
336 D.S. Hage, Clin. Chem., 45 (1999) 593.
337 L.I. Andersson, J. Chromatogr.B, 739 (2000) 163.
338 B. Sellergren, Anal. Chem., 66 (1994) 1578.
339 L.I. Andersson, A. Paprica and T. Arvidsson, Chromatographia,46 (1997) 57.
340 L.I. Andersson, Analyst, 125 (2000) 1515.
341 B.A. Rashid, R.J. Briggs, J.N. Hay and D. Stevenson, Anal. Commun., 34 (1997) 303.
342 W.M. Mullett and E.P.C. Lai, Anal. Chem., 70 (1998) 3636.
343 W.M. Mullett and E.P. Lai, J. Pharm.Biomed. Anal., 21 (1999) 835.
344 A. Bereczki, A. Tolokan, G. Horvail, V. Horvath, F. Lanza, A.J. Hall and B.
Sellergren, J. Chromatogr. A, 930 (2001) 31.
834
345 C. Berggren, S. Bayoudh, D. Sherrington and K. Ensing, J. Chromatogr. A, 889
(2000) 105.
346 G. Brambilla, M. Fiori, B. Rizzo, V. Crescenzi and G. Masci, J. Chromatogr. B, 759
(2001) 27.
347 C. Crescenzi, S. Bayoudh, P.A.G. Cormack, T. Klein and K. Ensing, Anal. Chem., 73
(2001) 2171.
348 P. Martin, I.D. Wilson, D.E. Morgan, G.R. Jones and K. Jones, Anal. Commun. 34
(1997) 45.
349 P. Martin, I.D. Wilson and G.R. Jones, J. Chromatogr.A, 889 (2000) 143.
350 R.J. Ansell and K. Mosbach, Analyst, 123 (1998) 1611.
351 T.A. Sergeyev, H. Matuschewski, S.A. Piletsky, J. Bendig, U. Schedler and M.
Ulbricht, J. Chromatogr.A, 907 (2001) 89.
352 T. Takeuchi and J. Haginaka, J. Chromatogr.B, 728 (1999) 1.
353 B. Sellergren and L.I. Andersson, Methods: A Companion to Methods in Enzymology,
22 (2000) 92.
354 L.I. Andersson, J. Chromatogr.B, 745 (2000) 3.
355 R.E. Majors, LC-GC Int., 8 (1995) 375.
356 S.M. Lunte and C.E. Lunte, Adv. Chromatogr., (1995) 383.
357 E.C. de Lange, B.A. de Boer and D.D. Breimer, Adv. DrugDeliv.Rev., 36 (1999) 211.
358 Y.L. Chang, M.H. Chou, M.F. Lin, C.F. Chen and T.H. Tsai, J. Chromatogr.A, 914
(2001) 77.
359 T.H. Tsai, H.Y. Kao and C.F. Chen, Biomed. Chromatogr., 15 (2001) 79.
360 K. Johansen, M. Krogh, A.T. Andresen, A.S. Christophersen, G. Lehne and K.E.
Rasmussen, J. Chromatogr. B 669 (1995) 281.
361 R.H. Hernandez, A.J.H. Louter, N.C. van de Merbel, U.A.Th. Brinkman, J. Pharm.
Biomed. Anal., 14 (1996) 1077.
362 K. Johansen, M. Krogh and K.E. Rasmussen, J. Chromatogr.B, 690 (1997) 223.
363 J.D.H. Cooper, D.C. Muirhead, J.E. Taylor and P.R. Baker, J. Chromatogr. B, 701
(1997) 281.
364 M. Karlsson, T. Korkolainen and T. Wikberg, Biomed. Chromatogr., 11 (1997) 54.
365 A. Ceccato, P. Chiap, Ph. Hubert, B. Toussaint and J. Crommen, J. Chromatogr. A,
750 (1996) 351.
366 K. Johanssen and K.E. Rasmussen, J. Pharm.Biomed. Anal., 16 (1998) 1159.
367 G. Higgins, A. Hughes, J. Dutton and N.B. Roberts, Ann. Clin. Biochem., 35 (1998)
534.
368 L.H. Parsons, T.M. Kerr and F. Weiss, J. Chromatogr.B, 709 (1998) 35.
369 F. Sadeghipour and J.L. Veuthey, J. Pharm.Biomed. Anal., 17 (1998) 801.
370 J.R. Veraart, M.C.E. Groot, C. Gooijer, H. Lingeman, N.H. Velthorst and U.A.Th.
Brinkman, Electrophoresis, 19 (1998) 2944.
371 R.M. Mader, M. Brunner, B. Rizovski, C. Mensik, G.G. Steger, H.G. Eichler and M.
Muller, Electrophoresis, 19 (1998) 2981.
372 P.B. Jacobsen, J. Chromatogr.B, 749 (2000) 135.
373 P. Chiap, B. Evrard, M.A. Bimazubute, P. de Tullio, Ph. Hubert, L. Delattre and J.
Crommen, J. Chromatogr. A, 870 (2000) 121.
374 T.H. Tsai and Y.F. Chen, Biomed. Chromatogr., 14 (2000) 274.
375 C. Mislanova, A. Stefancova, J. Oraveova, J. Horecky, T. Tmovec and W. Lindner, J.
Chromatogr. B, 739 (2000) 151.
376 N. Kobayashi, M. Kazui and T. Ikeda, J. Pharm.Biomed. Anal., 21 (2000) 1233.
377 T.H. Tsai, H.Y. Kao and C.F. Chen, Biomed. Chromatogr., 15 (2001) 79.
378 T.H. Tsai, F.C. Cheng, Y.F. Chen and C.F. Chen, J. Chromatogr.A, 914 (2001) 83.
379 Y.L. Chang, M.H. Chou, M.F. Lin, C.F. Chen and T.H. Tsai, J. Chromatogr.A, 914
(2001) 77.
835
380 J.R. Veraart and U.A.Th. Brinkman, J. Chromatogr. A, 922 (2001) 339.
381 S. Palmarsdottir, L. Mathiasson, J.A. Jonsson and L.-E. Edholm, J. Chromatogr. B,
688 (1997) 127.
382 S. Palmarsdottir, E. Thrdarson, L.-E. Edholm and J.A. Jonsson, L. Mathiasson,
Anal. Chem., 69 (1997) 1732.
383 J. Janiszewski, P. Schneider, K. Hoffmaster, M. Swyden, D. Wells and H. Fouda,
Rapid Commun. Mass Spectrom., 11 (1997) 1033.
384 R.S. Plumb, R.D. Gray and C.M. Jones, J. Chromatogr.B, 694 (1997) 123.
385 Y. Shen, J.A. Jonnson and L. Mathiasson, Anal. Chem., 70 (1998) 946.
386 J.A. Jonsson, M. Andersson, C. Melander, J. Norberg, E. Thordarson and L.
Mathiasson, J. Chromatogr.A, 870 (2000) 151.
387 G. Rule, M. Chapple and J. Henion, Anal. Chem., 73 (2001) 439.
388 H. Lingeman and S.J. H.-Oussoren, J. Chromatogr.B, 689 (1997) 221.
389 L.A. Dawson, J. Chromatogr.B, 697 (1997) 89.
390 L. Denoroy, L. Bert, S. Parrot, F. Robert and B. Renaud, Electrophoresis, 19 (1998)
2841.
391 D.K. Hansen, M.L. Davies, S.M. Lunte and C.E. Lunte, J. Pharm. Sci., 88 (1999) 14.
392 D.J. Weiss, C.E. Lunte and S.M. Lunte, Trends Anal. Chem., 19 (2000) 606.
393 L. Stable, Adv. Drug Deliv. Rev., 45 (2000) 149.
394 M.I. Davies, J.D. Cooper, S.S. Desmond, C.E. Lunte and S.M. Lunte, Adv. DrugDeliv.
Rev., 45 (2000) 169.
395 J.A. Jonsson and L. Mathiasson, J. Chromatogr.A, 902 (2000) 205.
396 K.E. Rasmussen, S.P. Bjergaard, M. Krogh, H.G. Ugland, T. Gronhaug, J.
Chromatogr. A, 873 (2000) 11.
397 H.G. Ugland, M. Krogh and K.E. Rasmussen, J. Chromatogr.B, 749 (2000) 85.
398 T.G. Halvorsen, S.P. Bjergaard and K.E. Rasmussen, J. Chromatogr.A, 909 (2001)
87.
399 M.A. Jeannot and F.F. Cantwell, Anal. Chem., 68 (1996) 2236.
400 M.A. Jeannot and F.F. Cantwell, Anal. Chem., 69 (1997) 235.
401 M.A. Jeannot and F.F. Cantwell, Anal. Chem., 69 (1997) 2935.
402 Y. He and H.K. Lee, Anal. Chem., 69 (1997) 4634.
403 L. Zhao and H.K. Lee, J. Chromatogr.A, 919 (2001) 381.
404 M.H. Akhtar, L.G. Croteau, C. Dani and K.A. Elsooud, Spectroscopy, 13 (1997) 33.
405 M.H. Akhtar, M. Wong, S.R.H. Crooks and A. Sauve, Food Addit. Contam., 15 (1998)
542.
406 G.M. Greenway and N. Kometa, Analyst, 119 (1994) 929.
407 Z. Guo, Q. Jin, G. Fan, Y. Duan, C. Qin and M. Wen, Anal. Chem., 436 (2001) 41.
408 C.S. Eskilsson, E. Bjorklund, L. Mathiasson, L. Karlsson and A. Torstensson, J.
Chromatogr. A, 840 (1999) 59.
409 H. Sachs and P. Kintz, J. Chromatogr.B, 713 (1998) 147.
410 K.S. Boos and C.H. Grimm, Trends Anal. Chem., 18 (1999) 175.
411 H. Weigmann, J. Bierbrauer, S. Hartter and C. Hiemke, Ther. DrugMonit., 19 (1997)
480.
412 J.J. Vreuls, A.J.H. Louter and U.A.T. Brinkman, J. Chromatogr.A, 856 (1999) 279.
413 E. Baltussen, P. Sandra, F. David and C. Cramers, J. Microcol. Sep., 11 (1999) 737.
414 T. Benijts, J. Vercammen, R. Dams, H.P. Tuan, W. Lambert and P. Sandra, J.
Chromatogr. B, 755 (2001) 137.
415 J. Vercammen, C. Peres, C. Devos, P. Sandra, F. Vanhaecke and L. Moens, Anal.
Chem., 73 (2001) 1509.
836
Chapter24
24.1 INTRODUCTION
The use of automation has been shown to be very successful for the many varied
sample preparation techniques performed in pharmaceutical and clinical
analysis. The choices for automation range in complexity according to the
particular function to be performed. While traditionally the term "sample prep"
refers to concentration of analyte, exchange of solvent, and/or removal of inter-
fering substances prior to analysis, automation processes exist for a multitude of
supporting functions as well, such as solvent delivery, sample dissolution,
sample aspiration and dispensing, sample reformatting from tubes to plates,
homogenization, capping/uncapping, sealing, and delivery of sample to the
detection system. These individual processes can be combined to form a semi-
automated or fully automated method. Also, the manner in which tasks can be
combined varies from a single benchtop workstation approach to multiple
modules linked with communications software and a robotic arm to shuttle
components, as will be described in this chapter.
There are many motivating factors for automating a specific sample prepara-
tion procedure. Most commonly, processes are automated to achieve higher
throughput. Although throughput considerations often meet or exceed goals,
sometimes they fall short of expectations. Automation is not always faster than
manual operation, but it can reduce hands-on time significantly, thereby freeing
a scientist to perform other tasks. Unattended operation may also allow over-
night runs, in some cases. Freeing an individual from hazardous (e.g., radioactiv-
ity or toxic chemical usage) and/or mundane tasks is another valid reason for
implementing automation. The automation of most processes has been shown to
bring a degree of reproducibility and quality to the results that cannot be real-
ized among different workers each performing the method manually. This com-
monality in the approach lends itself to easier assay duplication or transfer.
Automation can often facilitate troubleshooting by removing any manual vari-
ability from processing. By then varying the process parameters in a controlled
838
automatically, offering walk-away automation solutions. In addition, clinical
laboratories are moving toward modular automation in an effort to improve cost
efficiency. Clearly, the role of automation for sample preparation is wide-
ranging and encompasses many different functions within both pharmaceutical
and clinical analysis.
The objectives for this chapter are to: (a) discuss strategies and attributes to
consider when implementing automated processes into analytical laboratories;
(b) overview specific applications demonstrating the extensive use of automation
in pharmaceutical bioanalysis and clinical analysis; and (c) assess and compare
automated systems, as well as offer a glimpse into future automation trends.
839
(d) robotic workstations having a gripper arm that can pick and place
labware;
(e) full robotic systems linking multiple workstations and/or modules with
other peripheral devices.
Although discrete classifications are offered for the purpose of noting closely
related instrument equivalents, one should view this attribute more as a
continuum.
100%
. 75 %
E
o 0%
25 %
840
need to be analyzed? How many different assays need to be accommodated and
how frequently will changeovers occur? How similar are the different assays?
How challenging are the analytical requirements of the different assays? Will
the system be used for method development? What level of process documenta-
tion is required and how much of that information should be captured by the
automated system? These are all fundamental questions to be posed when
considering specific needs with regard to an automated system.
(4) Functionality
Consider the extent to which a system automates a process (e.g., semi-
automated versus fully automated). Determine the exact elements or steps
which are to be performed on a particular instrument. For example, the process-
ing steps that can be used to evaluate a system's functionality might include:
capping/uncapping samples, weighing, loading or exchanging consumables,
conditioning, sampling, washing, eluting, vortexing or mixing, evaporating,
filtering, derivatizing, incubating, reagent additions, vacuum manifold control,
centrifugation, bar code labeling and reading, sealing and injecting.
841
(7) Degree of Difficulty
The learning curve associated with operating an automated sample preparation
system will vary depending on both the skill level of staff and the complexity of
the equipment and/or its software. Simpler systems usually begin to pay small
dividends almost immediately, whereas complex robotic systems may take
months or even years to fully implement with the realization of greater returns.
Equipment installation, operation, maintenance and continuing system devel-
opment roles and timeframes should be clearly defined. An evaluation of the
in-house skills and resources (chemistry, automation, programming and engi-
neering) required for each automation option should be made. This assessment
should then be matched with existing talent available, training options, and
contract support options. The more complicated workstations and robotics
systems usually require the identification of a champion or system administra-
tor early in the process.
(9) Ruggedness
It is difficult to determine a system's reliability prior to purchasing and using the
equipment in support of daily lab activities. Ask the manufacturer to supply
references of users who may give a candid review of their experiences. Some
companies are willing to offer the use of a demonstration unit for a hands-on
evaluation prior to making a purchase. The process of acquiring and installing
the demonstration unit will often provide insight into the quality and availabil-
ity of the vendor's technical and service support personnel for this system. As
each individual step of the automated process is evaluated, keep in mind that the
system will only be as reliable as the weakest link in the process. For more
complicated or customized systems, it is wise to agree on performance specifica-
tions as part of the purchase agreement. This performance expectation shared
up-front serves as a good, and fair, source of protection for both the customer
and the manufacturer.
842
samples per hour, depending on the system. Sometimes, too much attention is
placed on the speed of the instrument, rather than its functionality. In such
cases, it is not uncommon for laboratories to increase the speed of a single link in
a process only to realize that they have merely shifted the bottleneck and had
little impact on the overall productivity of the application.
(11) ProcessingMode
Automated systems also vary in the manner in which samples are processed,
which can influence overall system speed and throughput. Some systems process
each sample individually in a serial or staggered-serial fashion. This mode of
operation may afford reproducibility advantages such as gravimetric
confirmation of liquid transfers, unique control and monitoring of each sample
processed, and uniform sample history or exposure. Other systems process
samples in parallel using multiple tip liquid handling transfers. The processing
mode frequently involves a tradeoff of giving up some control of sample
processing in return for greater speed. A good example of increasing parallel
processing speed involves moving from a single sample 1-tip serial mode to a
96-sample parallel mode with the implementation of a 96-well microplate format
and 96-tip liquid handler for sample preparation. Other combinations offering
improved throughput also exist where a system concurrently extracts samples
serially on multiple, distinct processing modules.
843
(13) ProcessingFlow Control
Another important variable to consider is the means by which solvents are
passed through an extraction or filtration device. Processing flow control can
involve using either vacuum, positive pressure or centrifugation, controlled
either manually or by the automated system. Many automated systems use
vacuum manifolds since they are relatively simple, inexpensive and their small
size allows easy integration with the automation deck or its hardware. Other
automated systems complement or substitute for vacuum processing by using
positive pressure flow control or centrifugation. There is ample evidence of
improved analytical precision using positive pressure as opposed to vacuum flow
control for solid-phase extraction applications [2,3]. Centrifugation also offers
varying control and has been applied for a number of techniques. Different
automated systems offer varying degrees of flow control ranging from requiring
manual intervention to simple on/off control and to finer, varying control with
pulsing and multiple stepped levels of applied force. Some systems also offer a
means for attempting to identify sorbent bed blockages by liquid level detection.
24.3.1 Background
844
a staggered serial manner, especially for on-line applications. At this stage, there
were liquid handlers with multiple tips (e.g., 4 or 8) for performing at least
certain steps in parallel.
In the early 1990s, the attention of pharmaceutical automation efforts was
redirected to early drug discovery applications by the introduction of systematic
research approaches such as high throughput screening and combinatorial
synthesis. These approaches sought to harness iterative, automated procedures
executed in parallel, microplate array format. By the mid-1990s, the concept of
parallel sample preparation and the automated tools created for microplate
formats were beginning to be adapted for applications further along in the drug
development cycle. The first solid-phase extraction microplates became
commercially available in 1996 and scientists continued to explore performing
other types of sample prep in parallel microplate format as well. Such principles
even spread to dosing animals in parallel with multiple test compounds "n-in-1"
[4] to reduce animal handling or, alternatively, pooling of analytical samples [5].
Automation was further driven by advances in detection systems, namely
liquid chromatography interfaced with tandem mass spectrometry (LC/MS/MS),
and their more widespread affordability within the last five years. These systems
have allowed for the processing of a greater number of samples than ever before
[6]. Although mass spectrometry has not obviated the need for matrix removal
and chromatographic separation, it has loosened the requirements for resolving
multiple analytes prior to detection. These LC/MSMS systems have become
prevalent in the pharmaceutical industry and have recently begun to appear in
the larger clinical reference laboratories for some esoteric analyses. As a result,
laboratories are now using higher throughput sample preparation schemes
combined with automation to better realize the potential of their LC/MS/MS
systems. Because of the cost and size of such analytical instruments, more recent
automated systems have tended to be off-line. Serial processing systems in many
cases are no longer able to keep pace with the faster analysis times. However, in
many instances, sample prep platforms based on parallel processing have again
reached a point where they are capable of supplying samples for analysis on
multiple LC/MS/MS systems.
845
techniques used for both clinical and pharmaceutical analysis, as described in
the chapter by Kataoka and Lord, there are a variety of automation tools
available. The bioanalytical sample preparation techniques in common use
today include:
- protein precipitation;
- liquid-liquid extraction;
- off-line solid-phase extraction using cartridges and microplates;
- on-line solid-phase extraction techniques using microplates, cartridges and
small diameter LC columns;
- on-line sample preparation techniques such as turbulent-flow chromato-
graphy, restricted access media and immunoaffinity extraction.
Automated liquid handling workstations using 4 or 8 tips are commonly used for
bioanalytical sample preparation tasks. Examples of these workstations include
the MultiPROBE IIM (Packard Bioscience Company, Meriden, Conn., USA)
M
(Fig. 24.2), Genesis" (Tecan Group Ltd., Mannedorf, Switzerland), Biomek"'
2000 (Beckman Coulter, Fullerton, Calif., USA), Speedy" (Zinsser Analytic,
Frankfurt, Germany), and the Model 215"' (Gilson, Middleton, Wisc., USA).
Note that the MicrolabM Series of liquid handling instruments (Hamilton
Instruments, Reno, Nev., USA) is available in a 1-tip version as well as in
multiple tip versions for liquid handling tasks. All of these instruments are
interfaced to a computer and are software-controlled for precise adjustment of
parameters. These workstations are also used for addition of reagents to adjust
pH, addition of internal standard, dilutions and liquid transfers (e.g., from tube
to tube, plate to plate, tube to plate, and tube or plate to autosampler vials).
Higher throughput liquid aspiration and dispensing is achieved by 96-tip
pipettors, such as the Quadra96" (Tomtec, Hamden, Conn., USA), Multimek"'
(Beckman Coulter) (Fig. 24.3), Personal PipettorTM 96 (Apricot Designs,
Monrovia, Calif., USA), Sciclone'" (Zymark, Hopkinton, Mass., USA) or
Fig. 24.2. Example of a 4-tip liquid handling workstation, the MultiPROBE II (Packard
Bioscience Company). Photo reprinted with permission from Packard Bioscience.
846
Fig. 24.3. Example of a 96-tip liquid handling workstation, the Multimek96'" (Beckman
Coulter). Photo reprinted with permission from Beckman Coulter, Inc. Copyright 2001.
Proteinprecipitation
A fast and simple method of sample preparation is protein precipitation. This
non-selective technique involves adding a water-miscible organic solvent (e.g.,
acetonitrile) to the biological matrix (usually in a 3:1 ratio, v/v) to denature and
precipitate the proteins. The proteins are removed by centrifuging or filtering
and an aliquot of the diluted supernatant is injected for analysis. Protein
precipitation is often used in drug discovery support when an analytical method
needs to be developed quickly and concentrations of analyte are relatively high.
In this situation, the generic extraction approach, with its associated disadvan-
tage of matrix carryover causing ion suppression, is a more important criterion
than achieving greater sensitivity by using a more selective technique. Typical
sample matrices seen in drug discovery support that are used with protein
precipitation are plasma, serum and tissue homogenates.
Bioanalytical sample prep using protein precipitation is usually a semi-
automated procedure, accomplished in part by having a liquid handling worksta-
tion aspirate and dispense volumes of organic solvent to samples contained in
test tubes or microplates. The liquid handler, if desired, can also aspirate fixed or
varying volumes of sample from source tubes into the tube or well used for the
precipitation; disposable tips are frequently used to avoid carryover. Additional
847
tasks that can be automated include preparation of calibration and quality
control standards, and internal standard additions. The tubes or plates contain-
ing organic solvent and sample then are capped/sealed, vortexed and then
centrifuged to pellet the protein. An aliquot of the supernatant can be aspirated
using the same liquid handling instrument. The volume aspirated should not
disrupt the protein pellet upon removal, which can be facilitated by choosing an
instrument with liquid level detection and tracking. This aliquot is transferred
to a clean vial or microplate, often containing a small volume of buffer that is
capped, vortexed and ready for injection into the analytical system for separation
and detection. Note that another option instead of dilution of supernatant
(depending on volume used) is evaporation, followed by reconstitution. Although
this procedure is simple and can be done manually, automation is introduced
when the number of samples to be extracted becomes large and the procedure is
to be performed routinely. An application reported by Watt details a protein
precipitation procedure using the Biomek 2000 in which throughput increased
to 400 samples per day in conjunction with LC/MS/MS analysis [7].
Filtration is another approach to protein precipitation and it can provide a
more easily automated sample prep procedure from start to finish. In this
application, a filtration 96-well plate is assembled on top of a vacuum manifold,
and a clean collection plate is placed inside the manifold to capture the filtrate. A
liquid handling workstation delivers acetonitrile and plasma to the wells of the
filtration plate. Precipitation occurs instantly and vacuum is applied to collect
the filtrate. Proteins remain on top of the filter and are discarded with the plate.
The filtrate contained in the underlying microplate can be evaporated before
reconstitution and analysis, or an aliquot of the diluted supernatant can be
injected directly.
It has been demonstrated that the order of addition of reagent and sample
can influence the efficiency of protein precipitation in filter plates, as well as the
degree of mixing or vortexing occurring when these two items are added
together. Also of note is that not all filter varieties available in plates possess the
necessary porosity to capture all proteins and the necessary resistance to organic
solvents to retain the organic solvent until vacuum or centrifugation is applied.
Thus, two common problems seen using filter plates are the breakthrough of
proteins into the filtrate and organic solvent that drips through prematurely.
A resourceful approach to obtain adequate efficiency of mixing is to first aspirate
acetonitrile, followed by an air gap, and then aspirate plasma in the same
disposable tip, as described by Biddlecombe [8]. This mixture is then forcibly
dispensed into wells of the filtration plate, thus avoiding a manual intervention
to vortex the plate. A condition of using the liquid handler for this filtration
approach is that it be able to work with a vacuum manifold on its deck and,
optionally, be able to sequentially aspirate solvents separated by an air gap. An
alternative to vacuum is to use centrifugation to filter the precipitated mixture.
This approach can be realized by placing a filter plate, mated with a collection
plate underneath, onto the deck of the liquid handler. Liquids are delivered
848
directly into the wells as described above, but then the plate combination is
manually inserted into a centrifuge for the filtration step.
Liquid-liquid extraction
Liquid-liquid extraction (LLE) involves mixing an immiscible organic solvent
with an aqueous biological sample (e.g., plasma) to selectively extract the
analyte(s) into the organic phase. The organic layer is transferred, evaporated to
dryness, and reconstituted prior to analysis. While this methodology provides
clean extracts, it has traditionally been difficult to automate, until recently.
There are three common approaches to automating LLE in a fully or semi-
automated setting: LLE performed in tubes, LLE in microplates, and LLE using
solid-supported media in flow-through columns or wells.
Typical 4- and 8-tip workstations can semi-automate LLE in test tubes but
manual intervention is often required for capping/vortexing. The detection of
the interface between the aqueous and organic layers is important in order to
aspirate the organic phase while leaving the aqueous phase behind in the
extraction tube. Removal of the organic layer from on top of the aqueous layer in
a tube can be performed by aspirating from a fixed depth in the tube. Freezing of
the aqueous layer at the tube bottom can aid in this transfer. The depth required
is determined by trial and error or, upon freezing, the upper organic layer can be
manually poured off into a clean tube. Such an application utilizing a liquid
handling instrument (MultiPROBE II) with liquid dispensing into test tubes is
reported by Jemal [9].
The technology now exists to automatically detect this phase boundary, and
an example of a single probe workstation performing automated LLE in this
manner is the Myriad ALLEX (Mettler Toledo Bohdan, Mundelein, Ill., USA).
The ALLEX is able to use different sizes of test tubes and automatically detect
the liquid-liquid phase boundaries, performing up to 60 separations per hour. It
also displays multiple and back extraction capability. A special note is that if the
organic layer is denser than the aqueous layer and remains at the well bottom,
the aqueous layer must be aspirated from the sample at a fixed depth, leaving the
organic layer in the well bottom for evaporation, reconstitution and analysis. In
this situation, the Myriad ALLEX workstation provides an efficient and superior
solution.
A higher throughput approach to performing LLE is to use microplates
instead of test tubes. Since the volume capacity of microplates (2 ml or less) is
smaller than tubes, sample and solvent volumes are reduced (e.g., 100 Al plasma
to 800 1I organic solvent). Multiple probe liquid handlers can be utilized for
sample and solvent aspiration/dispensing, and are especially convenient to
reformat samples from source tubes into a microplate. The microplates are
sealed and then manually vortexed to perform the mixing step. Liquid handlers
with 4-, 8- and 96-tips can all perform this action using microplates. The Myriad
ALLEX cannot use microplates, so aspiration of the organic layer is performed
using fixed depth techniques as described above, followed by evaporation and
849
reconstitution of the organic layer, all in the microplate format. The 96-tip liquid
handling instruments, e.g., Quadra96 (Tomtec), provide the highest throughput
for LLE in microplates and there are published applications of their use [10,11].
An important safety consideration for LLE applications is to vent the volatile
organic solvent vapors by either placing the instrument in a fume hood or
locating trunks from the ceiling near the work area to remove the vapors.
A more completely automated approach to LLE uses flow-through 96-well
microplates filled with inert diatomaceous earth particles of high surface area
(hydromatrix), available from numerous vendors (e.g., Varian, Harbor City,
Calif., USA; Orochem Technologies, Westmont, Ill., USA; and International
Sorbent Technology Ltd., South Wales, UK). When using the solid-supported
particles to perform LLE, plasma with buffer solution is added to the plate wells
and allowed to partition for a few minutes on the particle surface. A hydrophobic
filter on the bottom of each well prevents the aqueous phase from breaking
through into the collection plate. Organic solvent is then added to the wells, the
analyte partitions into the organic solvent as it flows through, then analyte is
eluted through the device, collecting in a container underneath. Multiple probe
and 96-tip liquid handlers can automate this entire extraction procedure for
limited volumes using microplates, as detailed in the report by Peng [121.
Solutions do flow through the particle bed via gravity, which maximizes the time
for interaction and partitioning of analyte from the aqueous into the organic
phase, so little or no vacuum is necessary. Advantages of using solid-supported
LLE include the avoidance of capping/mixing steps and the potential formation
of emulsions. A special note with the use of these particle beds is that the
aqueous phase is not retrievable.
Solid-supported LLE products are also available as individual cartridges,
essentially syringe barrels with a frit on the bottom and containing the diatom-
aceous earth particles. The cartridge format of diatomaceous earth is usually
used with gravity flow. One example of an application that efficiently uses these
solid-supported LLE cartridges in a fully automated mode is a custom configured
Zymate XP robotics system (Zymark). Zymark provides the Zymate"' robot
system, in which a central robotic arm is the focal point for a range of activities
custom configured about its perimeter (Fig. 24.4). The XP robot is now the third
generation laboratory robot for such an automation system. The robot
incorporates technology for interchangeable hands to carry various tubes and
containers, and to perform functions such as gripping, weighing, pipetting,
vortexing, centrifuging and evaporating. Tactile sensing capability is included in
the robot hand and all robot axes have optical encoders to verify successful
completion of functions. An application described by Alianti [13] uses a unique
tapered cartridge format ("LRC" configuration) of diatomaceous earth. The
robotics system is equipped with an analytical balance, an extraction station, a
1 ml pipet hand, a general purpose hand, a test tube dispenser, a vortex station,
an evaporating station, and an LC sipping section interfaced to an autosampler.
The robot transfers serum or plasma to an unconditioned cartridge, internal
850
Fig. 24.4. Example of a Zymate XP robotic system (Zymark) configured for performing LLE
using solid-supported cartridges. Photo courtesy of DuPont Pharmaceuticals.
standard is added, ethyl acetate is added and elution occurs using two 3 ml
aliquots, eluant is evaporated to dryness and the sample is reconstituted and
transferred to an autosampler vial for injection.
851
instruments produced specifically for SPE sorbent-filled tubes include the
single-tip ASPEC XL (Gilson) and its higher throughput version ASPEC XL4
that processes cartridges using positive pressure with four tips in parallel. The
ASPEC XL can be configured with one or two Rheodyne injection valves
(Rheodyne, Cotati, Calif., USA) to inject the prepared eluate on-line. An applica-
tion using the ASPEC XL that illustrates fully automated solid-phase extraction
coupled on-line with injection is reported by Streel [14].
A more recent and higher throughput Gilson product for SPE is the "SPE
215" system, which eliminates the need for caps on top of the cartridges via use
of an integrated silicone sealing foot. This design seals the tops of cartridges and
applies positive displacement for liquid processing of up to 8 samples at a time
via its 8-tips. An injection module can be added to perform injections of up to 8
samples in parallel (when interfaced with 8 LC systems). In terms of product for-
mat capabilities, the SPE 215 (like the ASPEC XL4) accepts the 1 and 3 ml car-
tridge sizes and is compatible with SPE microplates. When "tab-less" cartridges
are used, a greater density of cartridges can be placed in the available footprint
to retain a 9 mm center to center spacing as in microplates; e.g., 96 "tab-less"
1 ml cartridges can be placed on the rack and elution can be accomplished in a
microplate instead of test tubes.
Another example of automation providing high throughput in the syringe
barrel format for SPE is available via serial processing among multiple stations
(e.g., 10 modules placed side by side) using an instrument called the Rapid-
Trace" (Zymark), shown in Fig. 24.5. Each module processes liquids through
SPE cartridges via positive displacement. This modular design allows for
capacity to be added as demand requires. In terms of throughput, about 50
Fig. 24.5. Example of modular automation for SPE using cartridges, the RapidTrace T'(Zymark).
Photo reprinted with permission from Zymark.
852
samples per hour have been reported by Huang for a quantitative bioanalytical
LC/MS/MS application [15]. An additional application utilizing the
RapidTraceTm approach to automating SPE is illustrated by a forensic urine drug
testing method [16].
With today's rapid pace of drug development, the time required for method
development has been reduced from several weeks to just a few days, so the pri-
ority to automate this necessary task has become much greater. The develop-
ment of solid-phase extraction methods using cartridges was traditionally
performed manually and, once optimized, the methods were transferred to auto-
mation. However, with the advent of the RapidTrace format, method develop-
ment can now be routinely automated. The modular approach of the RapidTrace
makes it ideal for performing method development since the extraction condi-
tions can be manipulated one at a time among different modules, and even
among different samples within the same module. A discussion of its versatility
for accelerated bioanalytical methods development is provided by Xen [17].
A second workstation design using SPE cartridges and configured for meth-
od development purposes is the Spe-ed Wiz (Applied Separations, Allentown,
Penn., USA). This instrument uses positive pressure to process up to 60 samples
per hour. It uses standard cartridges and the software controls the drying times,
flow rates and solvents delivered to each cartridge. By using different sorbent
chemistries in the particle beds, many variables of solid-phase extraction can be
examined at once in an effort to develop and optimize an extraction method. A
good example of method development and optimization using SPE is provided by
Bakklai [18]. A more comprehensive automation solution for bioanalysis, in
terms of both capability and flexibility, is exemplified by a full robotics system.
Such systems frequently offer additional process controls such as a balance to
confirm liquid transfers and afford the opportunity to correct processes on the
fly. An application illustrating the range of activities of a custom configured
Zymate robotics system for performing solid-phase extraction method develop-
ment in bioanalysis is described by Parker [19].
A particular benchtop workstation called duo-PREP' (Pharmaceutical
Technology Ltd., Cambridgeshire, UK) processes liquids through SPE
cartridges in a unique manner. In contrast with conventional liquid processing
through SPE cartridges, in which solutions are passed through a stationary
solid-phase bed from top to bottom and out, duo-PREP operates by the move-
ment of the sorbent bed through each solution. The process is performed as
follows. An array of up to 48 cartridges (specifically duo-PREP SPE cartridges,
as regular SPE cartridges are not compatible) is lowered into an array of
multiwells containing the process solutions. As the cartridge descends into each
well below it, an outer flexible seal engages with the inner wall of the well,
generating a pocket of air above the solution. Once the cartridge inlet enters into
the solution, the liquid is forced upward, through the solid sorbent bed and into
the cartridge. On withdrawal of the cartridge, the arrangement of the seal with
the well creates a partial vacuum on the underside that pulls the solution back
853
out of the cartridge and into the well underneath. Thus, controlled reversible
flow is achieved. This procedure is performed for the condition, load, wash and
elution steps. For elution, clean multiwells are used and eluate remains in the
wells for injection by an XYZ autosampler [20].
854
i%
'-i
Fig. 24.6. Example of a 96-tip liquid handling workstation configured for microplate solid-phase
extraction using a vacuum manifold, the Quadra96TM (Tomtec). Photo reprinted with permission
from Tomtec.
solvent troughs (when necessary) during plate processing. Time required for an
extraction varies depending on the solvent volumes used and number of steps,
but can be as fast as 10 min; in practice, as many as four plates (384 samples) can
conveniently be processed per hour. When considering the total time require-
ment, however, the time to reformat samples from tubes to plate before use on
the Quadra96 must be considered. Several applications demonstrate the
versatility and speed of this 96-tip liquid handling approach to SPE [23-25].
There are also 96-tip pipettors available that can speed up common pipetting
operations, such as diluting samples, adding internal standard solution, and
sample reconstitution in microplates, as required. Such instruments are exem-
plified by the Personal Pipettor M 96 (Apricot Designs), Hydra96 (Robbins Sci-
TM
entific), Multimek 96 (Beckman Coulter), GenMateTM (Tecan) and
T M
RapidPlate (Zymark). The volume limitations vary among tip selected and
instrument, e.g., Multimek uses either 50 or 200 g1 disposable tips; thus, for dis-
pensing 500 gl volumes, multiple aspirations and dispenses are required. Many
of these instruments have an interchangeable pipetting head system so that they
can utilize 384-well labware on their deck.
The common XYZ liquid handlers described previously also can perform
solid-phase extractions; although not 96-tip based, they commonly offer 4-tip or
8-tip processing capabilities. The first few semi-automated 96-well SPE applica-
tions published used the MultiPROBE workstation [26,27]. The software offered
total control of the vacuum for processing via a manifold, and useful features of
the innovator Packard model included liquid level sensing, variable span tips
TM
and detection of blocked SPE wells. The Packard WinPREP software also
855
TM
Fig. 24.7. Example ofa gripper arm on a Biomek workstation used to assemble and disassemble
components and move items around the deck (Beckman Coulter). Photo reprinted with
permission from Beckman Coulter, Inc. Copyright 2001.
856
capabilities to a deck platform to meet the goals of the method, e.g., weighing,
pipetting, mixing, homogenizing/grinding, LLE, SPE, centrifuging, evaporation
and filtration. Cyberlab is another example of a vendor that has built a variety of
custom workstations for automating sample preparation. Capping and
uncapping is another function that is automatable; workstations or modules for
capping/uncapping are available, e.g., from Mettler Toledo Bohdan and Zymark.
The previously mentioned ASPEC XL4 and SPE 215 systems from Gilson,
which accommodate discrete SPE cartridges, also process most types of 96-well
extraction plates. The XL4 is a 4-tip system which requires special caps be placed
on top of the microplate wells prior to use; a capper unit is available to quickly
add these caps in the form of strips to an extraction plate. In contrast, the SPE
215 uses positive pressure via a silicone sealing mat and does not require caps; it
is also faster using its 8 tips. Both these systems are fully automated and can
elute directly into deep well microplates for subsequent injection by a microplate
compatible autosampler.
Full automation for SPE microplates using a custom configured Zymate
robotics system is exemplified by the work of Biddlecombe, Smith and Lloyd
[30-33]. The systems described are built around a Zymate XP robot designed to
handle microplates in conjunction with various modules, e.g., a liquid handler
(Tecan RSP 8051) for sample transfers, a reagent addition station (RAS)
equipped with dual 12-port solvent selection valves for all liquid additions, a
microplate vortex station, a 96-well evaporation station, a plate sealer, a temper-
ature controlled carousel for the storage of all microplates and extraction
devices, and a custom microplate centrifuge (Fig. 24.8). In addition to SPE,
lI
25 PORT
TOLV
4T
SELECT
DRY - t~~flp
ix I A T
`'Gv
VORTEX BALANCE
MAND
CENTRIFUGE .>/
SEALER fu
Fig. 24.8. Schematic diagram of a robotics system configured to perform protein precipitation,
LLE and SPE in microplates. Photo courtesy of Glaxo SmithKline.
857
systems such as this one can also perform protein precipitation and LLE in
microplate format [34,35]. Note that full automation for sample prep in micro-
plates using a robotics system is not limited to the Zymark XP series of robots.
CRS Robotics Corporation (Burlington, Ontario, Canada) has its own line of
robotic arms upon which full workstations can be custom configured.
858
Fig. 24.9. On-line SPE can be accomplished with the Prospekt which includes a solvent delivery
unit, cartridge transport and sealing module, and an autosampler; (a) extraction mode (b)
elution mode. Photo reprinted with permission from Spark Holland.
859
molecules and diverted to waste. Column switching with the introduction of a
strong organic solvent flow then elutes the analyte from the turbulent flow
column onto the mass spectrometer. The Turboflow T column can be back-
flushed and re-equilibrated for repeated use, up to a maximum number of
injections which varies based on matrix material and volumes injected. An entire
system for turbulent flow chromatography is configured by Cohesive
Technologies (Franklin, Mass., USA) to minimize carry-over and provide for
optimum integration of components with the mass spectrometer. Numerous
applications have been demonstrated using this technique [38-401.
Another on-line sample preparation approach is reported that uses a
1x50 mm LC column packed with 30-50 [km stationary phases. The large
particle size allows a higher flow rate than normal to be used (e.g., 4 ml/min),
leading to a short analysis cycle time. These conditions do not represent a
condition of true turbulent flow, thus the terms "ultra-high flow-rate liquid
chromatography" and "direct on-line bioanalysis" have been used to describe
this technique [41]. A splitter is placed after the extraction column, permitting
about 0.2 ml/min to enter the mass spectrometer. Although only minimal
separation resolution can be achieved, it affords a rapid extraction to keep pace
with fast LC/MS/MS run times. While this example illustrates a generic
procedure, it has been adopted to provide for rapid method development for
SPE. An optimized SPE method can be achieved within one working day by
manipulating pH and the percentage of organic in the mobile phase to affect the
retention of analyte and matrix interferences 42]. This method can then be
transferred to a SPE microplate by matching sorbent chemistries.
860
between injections, and required mobile phases are not always compatible with
some ionization techniques used in LC/MS/MS. The overall approach is auto-
mated, however, with low capital outlay for automation equipment since it uses
common HPLC hardware configurations, and is another choice in bioanalysis.
861
Fig. 24.10. A pre-analytical automated sample preparation system for a clinical laboratory, the
Power Processor M (Beckman Coulter). Photo reprinted with permission from Beckman Coulter,
Inc. Copyright 2001.
862
Larger clinical laboratories have Laboratory Information Systems (LIS) that
are interfaced with separate information systems for billing, and to patient
information systems for sample test requests and display of results. In the
laboratory, the testing requested for a specific sample is accessed by the worksta-
tion through "Host Query" functions and the test request processed in real time
by the analyzer. Once the bar-coded sample identification is read, the instru-
ment receives information on the tests to be performed directly from the LIS.
The instrument's "review by exception" software evaluates results and the
acceptable results are forwarded directly to the LIS. After a further review
process, these results can be released to the physician for review via the patient
information system.
The clinical laboratory uses test tubes as preferred sample containers
(not microplates as in pharmaceutical analysis). Safety concerns from
potentially hazardous or infected blood from patients can be satisfied using
"through the stopper" aspirations for many sample types. Tests performed are
most often immunoassay-based and detection endpoints include fluorescence,
colorimetric, kinetic rate, and/or chemiluminescence (not LC/MS/MS as in
pharmaceutical). This format approach permits the use of dedicated benchtop
workstations and analyzers that perform clinical chemistry tests automatically,
offering walk-away automation solutions (via large on-board capacity for
reagents and supplies) and integration with LIS. Examples of benchtop work-
stations offering immunoassay detection include the Access®2 (Beckman
Coulter; Fig. 24.11), the ADx System (Abbott Diagnostics, Chicago, Ill., USA)
and the ACS 180 (Bayer Diagnostics, Tarrytown, New York, USA).
A large, floor model modular automation approach is the AccelNet' labora-
tory automation network from Beckman Coulter. The AccelNet seamlessly com-
bines two SYNCHRON CX®9 ALX clinical chemistry systems, a specimen
manipulator, an automated sample processor, a centrifuge, and a data manager
Fig. 24.11. A benchtop automated immunoassay workstation for clinical analyses, the Access'
(Beckman Coulter). Photo reprinted with permission from Beckman Coulter, Inc. Copyright
2001.
863
to create a versatile workstation. The SYNCHRON series of clinical chemistry
analyzers comprise immunoassay systems in which up to 150 tests are available.
These tests include critical care and general chemistries, urine chemistries, eso-
teric chemistries, as well as tests for drugs of abuse, proteins, serologies, and
therapeutic drug monitoring. In terms of throughput, the CX500 Pro offers 53
on-board assays and throughput up to 700 tests per hour, the CX1000 Pro offers
57 assays at 1000 tests per hour and LX2000 Pro offers 65 assays at 1540 tests
per hour. Two additional examples of a scalable, modular automation system are
the Advia® LabCell Bayer Diagnostics [49] and the Lab-Frame® SELECT
(Ortho Clinical Diagnostics, Raritan, New Jersey, USA). Laboratories can adapt
an upgrade path with this equipment through the addition of conveyer belt mod-
ules that allow the addition of other work cells (e.g., chemistry, hematology).
Therapeutic drug monitoring (TDM) laboratories perform bioanalysis to
relate the plasma concentrations of a drug to a patient's therapeutic response
and/or adjust drug dosage to minimize adverse effects or toxicity. Typical
analytes for which TDM is performed using immunoassay detection include
digoxin, acetaminophen, phenobarbital, gentamicin, phenytoin, theophylline
and valproic acid. Lower volume, highly specific drug testing requires special
handling and sample preparation techniques. Immunoassay detection cannot be
used in cases where there is metabolite cross-reactivity (e.g., cyclosporine)
and/or when multiple analytes must be resolved and measured simultaneously
(e.g., antidepressant drugs), thus the more specific procedure LC/UV is used.
Sample preparation for LC/UV analysis most commonly involves solid-phase
extraction using individual columns or cartridges, and to a lesser extent involves
on-line SPE and LLE. Some common choices for automating the use of solid-
phase extraction cartridges in the clinical laboratory are the ASPEC XL/XL4 and
SPE Model 215 (Gilson) and the RapidTrace (Zymark), as discussed previously
for pharmaceutical analysis. The Prospekt (Spark Holland) has also been
demonstrated to perform reliably for automated on-line SPE in the clinical
laboratory setting. Note that although LC/UV is commonly used for detection,
there are growing instances where LC with mass spectrometry detection (LC/MS
and LC/MS/MS) is warranted (e.g., tacrolimus and sirolimus, two anti-rejection
drugs used in transplant cases; some of the newer anti-AIDS drugs; for newborn
screening in detecting inborn errors of metabolism).
Although dedicated workstations and modular automation for clinical labo-
ratory testing are widely used and continue to proliferate, there is the occasional
need for a more flexible and affordable automation solution, particularly for
molecular diagnostic applications performed in clinical research laboratories. An
example of this approach is the automation of the polymerase chain reaction
(PCR) amplification response to detect specific RNA and DNA sequences. The
PCR test is best performed by highly skilled technologists because of the
complexity of the assay and the potential for laboratory contamination. Thus,
automating this labor-intensive procedure is a desired goal and applications
have been demonstrated using an XYZ liquid handling workstation, e.g., the
864
MultiPROBE II (Packard Bioscience Company) and the Biomek 2000 (Beckman
Coulter). Integral components for this PCR assay are mounted on the worksta-
tion deck, namely a DNA engine with a remote alpha dock system and an
automated thermal cycling device. PCR protocols require the mixture of
multiple reagents in specific ratios for the amplification process. The exact
volumes of the reagents and buffers vary based on the total number of amplifica-
tions to be run and the desired final sample concentration and volume. An XYZ
liquid handling workstation has been found to provide a cost efficient alternative
to molecular diagnostic assays while demonstrating minimal inter-sample
contamination [50].
Automation in the clinical laboratory continues to advance in sophistication
and the full treatment of the subject is outside the scope of this text. Cost-
effective modular systems are common, and evolve into integrated laboratory
automation systems. Automation of the sample preparation step is but one small
component of the focus of hospital and clinical laboratories. Full robotics appli-
cations using robotic arms on tracks and the interfacing of multiple analyzer
components and processing systems have been demonstrated for high through-
put to accomplish "total laboratory automation", practised by some of the
largest laboratories in the world. There exist already even mobile robots that
travel across hospital floors and ride in elevators to pick up and deliver samples,
part of the pre-analytical automation component. Many disciplines of engineer-
ing and robotics are thus involved in this rapidly advancing approach to
automate all laboratory processes in the clinical laboratory and hospital setting,
and the reader is referred to other texts for their further description [51,52].
865
cycle times again. Examples include parallel LC systems switching onto a mass
spectrometry source [53] or the combination of a reduced separation time by
implementing a monolithic column coupled with a parallel MS (MUX®) inter-
face [54]. Meanwhile, direct injection techniques (discussed previously) per-
formed in parallel continue to improve and challenge the need for sample prep at
all in certain cases [55].
Miniaturization is a continuing theme as detection limits improve, through-
put increases, and therefore sample sizes decrease. The density of microplate
formats continues to increase from 96-well to now 384-well and applications
using 384-well solid-phase extraction have been reported [56,57]. Also, liquid
handler working volumes continue to decrease and approach the point in many
instances where automation is no longer optional, but necessary to process sam-
ples. For example, the development and production of DNA/RNA biochip arrays
for diagnostic applications [58] presents unique liquid handling requirements
for nanoliter volumes. Tecan has developed a novel nano-pipetting system as an
option for the Genesis workstation to dispense volumes as small as 0.5 nl with a
spatial resolution of 100 Am.
In the meantime, many product lines from automation vendors have
diversified, through either expansion or acquisition, to the point of overlap and
broader competition. Companies that once concentrated on fully automated,
robotic systems now offer workstations and liquid handlers, while those that
began producing liquid handlers now have added plate stacking capability,
robotic arms, and a variety of other peripheral devices (e.g., carousels, incuba-
tors, centrifuges). A critical component to the integration and functional success
of linking peripherals and modular devices is software, which continues to
improve in utilizing different standards from multiple vendors.
REFERENCES
866
15 N. Helen Huang, John R. Kagel and David T. Rossi, J. Pharm. Biomed. Anal., 19
((1999) 613.
16 F.X. Diamond, W.E. Vickery and J. de Kanel, J. Anal. Toxicol., 20 (1996) 587.
17 X. Ren and A. Witkowski, in: Proceedings of the International Symposium on
Laboratory Automation and Robotics, 1996.
18 A. Bakkali, L.A. Berrueta, B. Gallo and F. Vicente, J. Chromatogr.B, 729 (1999) 139.
19 D. Parker, D.T. Rossi and D.S. Wright, Anal. Chem., 68 (1996) 2437.
20 J. Bates, UKLaboratory, February (1998) 12.
21 D.A. Wells, LC/GC, 17(7) (1999) 600.
22 C.Z. Matthews, E.J. Woolf, L. Lin, W. Fang, J. Hsieh, S. Ha, R. Simpson and B.K.
Matuszewski, J. Chromatogr.B, 751 (2001) 237.
23 J. Janiszewski, R.P. Schneider, K. Hoffmaster, M. Swyden, D. Wells and H. Fouda,
Rapid Comm. Mass Spectrom., 11 (1997) 1033.
24 G. Rule and J. Henion, J. Am. Soc. Mass Spectrom., 10 (1999) 1322.
25 S.X. Peng, S.L. King, D.M. Bornes, D.J. Foltz, T.R. Baker and M.G. Natchus, Anal.
Chem., 72 (2000) 1913.
26 J.P. Allanson, R. A. Biddlecombe, A.E. Jones and S. Pleasance, Rapid Comm. Mass
Spectrom., 10 (1996) 811.
27 H. Simpson, A. Berthemy, D. Buhrman, R. Burton, J. Newton, M. Kealy, D. Wells and
D. Wu, Rapid Comm. Mass Spectrom., 12 (1998) 75.
28 D. Schutze, B. Boss and J. Schmid, J. Chromatogr.B, 748 (2000) 55.
29 P. Ahrweiler, T. Sitko, S. Glass, K. Knotts and K. Ruterbories, in: Proceedings of Lab
Automation Conference, 2001.
30 S.L. Callejas, R.A. Biddlecombe, A.E. Jones, K.B. Joyce, A.I. Pereira and S.
Pleasance, J. Chromatogr.B, 718 (1998) 243.
31 R.A. Biddlecombe and S. Pleasance, in: Proceedings of the International Symposium
on Laboratory Automation and Robotics, 1997.
32 T.L. Lloyd, E. Lynch, J. Short and A. Wernicki, in: Proceedings of the International
Symposium on Laboratory Automation and Robotics, 2001.
33 G. Smith, J. Bruner and J. Alianti, in: Proceedings of the International Symposium
on Laboratory Automation and Robotics, 1997.
34 R.A. Biddlecombe and S. Pleasance, in: Proceedings of the International Symposium
on Laboratory Automation and Robotics, 1998.
35 R.A. Biddlecombe and S. Pleasance, in: Proceedings of the International Symposium
on Laboratory Automation and Robotics, 1999.
36 M.W.J. van Hout, C.M. Hofland, H.A.G. Niederlander and G.J. de Jong, Rapid
Commun. Mass Spectrom., 14 (2000) 2103.
37 A. Schellen, B. Ooms, M. van Gils, O. Halmingh, E. van der Vlis, D. van de Lagemaat
and E. Verheij, Rapid Commun. Mass Spectrom., 14 (2000) 230.
38 C.J. Oberhauser et al., LC-GC, 18(7) (2000) 716.
39 D. Zimmer, V. Pickard, W. Czembor and C. Muller, J. Chromatogr.A, 854 (1999) 23.
40 N. Brignol, R. Bakhtiar, L. Dou, T. Majumdar and F.L.S. Tse, Rapid Commun. Mass
Spectrom., 14 (2000) 141.
41 J. Ayrton, G.J. Dear, W.J. Leavens, D.N. Mallett and R.S. Plumb, J. Chromatogr.A,
828 (1998) 199.
42 J. Ding and U. Neue, Rapid Commun. Mass Spectrom., 13 (1999) 2151.
43 A. Haque and J.T. Stewart, Biomed. Chromatogr., 13 (1999) 51.
44 M.L. Nedved, S. Habibi-Goudarzi, B. Ganem and J.D. Henion, Anal. Chem., 68
(1996) 4228.
45 C.S. Creaser, S.J. Feely, E. Houghton and M. Seymour, J. Chromatogr.A, 794 (1998)
37.
46 D.C. Turnell and J.D.H. Cooper, J. Chromatogr., 395 (1987) 613.
867
47 P.-H. Hsyu and T.L. Lloyd, J. Chromatogr.B, 655 (1994) 253.
48 R.A. Felder, Clin. Chim. Acta, 278 (1998) 257.
49 M. Campanelli, J. Autom. Lab. Robotics, 3(3) (1998) 46.
50 T.E. Mifflin, C.A. Estey and R.A. Felder, Clin. Chim. Acta, 290 (2000) 199.
51 R.A. Felder, J. Clin. Lig. Assay, 22(1) (1999) 13.
52 S. Bauer and C. Teplitz, Med. Lab Observ., 7(9) (1995) 44.
53 L. Yang, T.D. Mann, D. Little, N. Wu, RP. Clement and P.J. Rudewicz, Anal. Chem.,
73 (2001) 1740.
54 Y. Deng, J.-T. Wu, T.L. Lloyd, C.L. Chi, T.V. Olah and S.E. Unger, Rapid Commun.
Mass Spectrom., 16 (2002) 1116.
55 C. Cameron, R.P. Grant, S. Mackenzie and M. Young, in: American Society for Mass
Spectrometry Annual Meeting, Chicago, 2001.
56 G. Rule, M. Chapple and J. Henion, Anal. Chem., 73 (2001) 439.
57 R.A. Biddlecombe, C. Benevides and S. Pleasance, Rapid Commun. Mass Spectrom.,
15 (2001) 33.
58 M.R. Knapp et al., Am. Lab., November (1998) 22.
868
Chapter25
25.1.1 Considerations
The term "food" refers to the broad range of edible materials that comprise the
essential body nutrients required for life and growth, such as proteins, carbohy-
drates, fats, vitamins, or minerals. Foodstuffs are described variously as "liquid"
or "solid", and "wet" or "dry", depending on the amounts of water and fat they
contain. Samples of plant origin are classified for analytical purposes as having a
high or medium water content and a lower content of saccharides (from 5% to
15%), very low water content (dry), or a high content of oils [1]. Similarly, food
samples can be divided into four main groups based on water and fat content [2].
Food samples of biological origin (liquid or solid) have been divided generally
into the five categories described in Table 25.1. This coarse division is important
when considering the choice of isolation technique, extraction solvent, and
sample clean-up method during an analytical procedure [3].
Moisture content is an important consideration during sampling procedures,
in part because it affects the extent of sample heterogeneity. Virtually all foods
are heterogeneous, and the analyst should be familiar with their variability in
composition and structure. In general, fresh foods of plant origin are more vari-
able in composition than fresh foods of animal origin. The analyst should be also
aware of the postmortem or postharvest physiological changes that can occur
after a fresh food is sampled and which can affect sample heterogeneity. A com-
bination of cold storage and chemical preservation may be required to maintain
sample integrity in the event of prolonged storage.
Although the chemical and physical properties of foods are inherently
variable, even between samples that originate from the same breed or strain, the
variability in composition of a single food sample can be minimized with proper
sampling and sample pretreatment techniques. Two approaches can be used for
sampling a food mass that is larger than the amount required for analysis in the
ComprehensiveAnalytical ChemistryXXXVII
J. Pawliszyn (Ed.)
© 2002 Elsevier Science B.V. All rights reserved 869
TABLE 25.1
General classification of food samples according to their content (with permission from Ref. [3])
Sample Character Typical analytes
Milk Aqueous, proteins, lipids Veterinary drugs, toxic elements,
pesticides, industrial contaminants
Eggs High lipids and albumin content Veterinary drugs, industrial
contaminants, pesticides
Other samples of Various fat, proteins, or water Drugs, industrial contaminants,
animal origin (e.g. pesticides
muscle, liver, fat)
Plant material (e.g. Various water, plant pigments, Pesticides, toxic elements,
fruits, vegetables, lipids, proteins, essential oils or industrial contaminants
seeds) waxes
Food (e.g. meat, fish, Various fat, oils, lipids, proteins, Pesticides, industrial
milk, cereals, wine, sugar, starch, water, or pigments contaminants, synthetic colorants,
juices, plant oils, additives, synthetic sweeteners,
sugar) antioxidants
870
The sample size is also dependent on the relationship that exists between the
mass required to adequately represent a sample and the characteristics of that
sample [6]. If a foodstuff consists of some mixture of different-sized particles,
enough sample mass needs to be collected in order to adequately represent all of
the particles. Because large particles are more difficult to represent than smaller
ones, a mass that is large enough to represent the larger particles will also be
representative of the smaller ones. The segregation of finer, denser particles to
the bottom of the sample container must be recognized during the sampling
process to ensure that all particles are represented and to avoid large sampling
errors. The theory of sampling along with solutions for correct sampling are
well-described in other texts [7-9].
25.1.2 Techniques
871
methods for obtaining representative and replicate samples, and for estimating
the error involved in sampling. Measures of the precision of the mean results are
also required, in addition to statistical analysis and interpretation of the data
obtained [12]. The reader is referred to standard text and reference sources in
the field for these purposes [7-9,12,13].
Once a food sample has been collected using the sampling techniques discussed
in Section 25.1.2, a suitable method is required to make the material less
heterogeneous. Various approaches may be utilized for reducing the particle
weight and size in a primary sample, so that smaller subsamples can be taken for
a representative analysis of the whole [4]. Finely divided materials also dissolve
faster and are easier to extract because of their greater surface area.
Methods for reducing solid or semi-solid foods include mechanical grinding,
mixing, rolling, agitating, stirring, chopping, crushing, macerating, mincing,
pressing, pulverizing, or any other reasonable means of comminuting the sam-
ple. Sample reduction can also be achieved with a Wiley or ball mill, mortar and
pestle, mechanical high-speed beaters or blenders (for soft or wet foods), and
meat grinders. Liquid samples can be mixed using magnetic stirrers or sonic
oscillators. Figure 25.1 demonstrates the importance of selecting the appropri-
ate hardware for sample mixing, and of blending the sample for a sufficient
period of time [15].
There are several other factors to consider when reducing a food sample.
Food choppers, blenders, and mixers should be constructed of metal alloys that
resist corrosion or erosion, and that are inert enough to prevent contamination
872
BUichi B-400 MixerA Mixer B Mixer C
U
0
47
n
W
U
e
Fig. 25.1. Effect of choice of mixing equipment and blending time on sample heterogeneity for
sunflower seeds. Reprinted with permission from Ref. [15].
of the product. Aeration of the product during the blending process should be
avoided since this can result in appreciable changes in oxidizable components. It
is also important to avoid heating the material during the grinding step since
this can accelerate chemical changes in the foodstuff. The surfaces of all mixing
equipment should be clean and dry, since changes in sample moisture content
can change the chemical and physical nature of the foodstuff. Care should also be
taken to prevent the release of volatile constituents during grinding, if this is of
concern [14].
The analyst should be aware of the enzymatic changes that can rapidly occur
in crushed plant and animal tissues. In animal tissue, rapid enzymatic changes
may result in appreciable changes of certain food components, particularly in
the case of carbohydrate and nitrogenous compounds [14]. It may also be
necessary to inactivate food enzymes, for example by denaturation in boiling
methanol-water or ethanol-water mixtures [16].
In conclusion, it is imperative for the analyst to be familiar with the food
matrix that is being analyzed. Since it is not feasible to discuss all possible cases
here, it is important that the analyst to consult the appropriate sources for
information before beginning a new sampling procedure.
25.2.3 Moisture
873
or stored, since the material may lose moisture during blending, or deteriorate
during storage. Determining the moisture content through sample drying may
be necessary in order to calculate the nutritive value of a food product or to
express analytical results on a uniform scale, for example in the determination of
the dry matter in flour [17]. Moisture is also important in terms of food quality,
since it affects food freshness, preservation, and resistance to deterioration.
Water is present in food samples in three forms [11]: as a solvent or
dispersing media; adsorbed on the internal or external surfaces, or as fine
capillaries by capillary condensation; and as water of hydration.
Solvent or free water is most easily removed. The rate at which moisture is
removed from foods is affected by drying temperature, particle size, vacuum,
crust formations on the surface, and surface area of the sample [11]. Bound
water is quite difficult to remove and normally requires a vacuum process.
Vacuum drying is preferred nonetheless, since this technique significantly
reduces the deterioration of samples during heating. For example, plant tissues,
which are often dried prior to the analysis step, might undergo extensive
enzymatic changes during the drying process, especially when they are exposed
to air. Utilizing vacuum drying also accelerates the drying time, which can take
up to 16 h under ideal conditions.
Other precautions need to be considered when drying foods at elevated
temperatures, since chemical reactions such as hydrolysis can occur and chemi-
cal reactions can be accelerated. Moisture determinations can be erroneous if
hydrolysis has occurred, since the water of hydrolysis has not been released from
the sample. On the other hand, very dry samples may absorb water from the air
before the moisture determination has been completed.
A general rule of thumb for sample drying is that it should be as rapid and at
as low a temperature as possible. Vacuum methods that can used to dry a sample
include vacuum ovens and lyophilization, or freeze-drying. Other methods that
can be employed are distillation, microwave drying, and the Fischer titration
method. The titration method is particularly applicable to low-moisture foods
that give erratic results when heated or under vacuum [11].
Finally, when drying a sample, the analyst should be aware that a certain
level of moisture might be required for prolonged food storage, since chemical
reactions such as oxidative deterioration can occur when moisture levels are too
low, for example in vegetables such as carrots and potatoes, which will develop
oxidized flavours or become rancid in two to three weeks at a 2 or 3% moisture
content. Oxidative deterioration of these foods is inhibited for several months
when they have a 8-10% moisture content [14].
874
extraction (or any other type of extraction) of non-polar compounds from animal
and vegetable matrices [18-22]. The presence of matrix interferences in sample
extracts can result in a multitude of problems, including the generation of
emulsions, sample turbidity, contamination or plugging of equipment, and,
perhaps most importantly, the masking of the analytical signal for the target
analyte and the consequent increase in the method limit of detection.
Co-extractives are frequently removed during a post-extraction clean-up
step that requires passing the liquid extract through a clean-up column for sorp-
tion or filtration of the interferences. Commonly used clean-up materials include
Florisil, alumina, silica gel, in addition to gel permeation chromatography,
solid-phase extraction materials, etc. Solid-phase materials can also be used to
exclude co-extractives from the analyte concentration step, that is, the material
may only retain the target analyte and not the interferences. This step is also
referred to as analyte enrichment, since the analyte concentration is increased
over that of the matrix background signal, if indeed any occurs at all. The factors
that affect the choice of clean-up material are similar to those considered when
choosing a solid-phase for the extraction of liquid food samples. Overviews of
both types of applications will therefore be presented together in Section 25.3.
Of particular interest are analytical methods that incorporate an in situ or
on-line clean-up technique. Sample clean-up in this case can be achieved in situ,
for example, with a simple and elegant extraction technique called matrix
solid-phase dispersion (MSPD) [23,24]. The advantage to MSPD is that it
combines sample blending, clean-up, and extraction into one technique. During
an MSPD procedure, the sample matrix is mixed with an appropriate polymer
resin, such as the reverse-phase chromatographic sorbent, C,,. The solid or
semi-solid sample is prepared for extraction by grinding it in the presence of the
sorbent using a mortar and pestle, which facilitates disruption of the sample
matrix. Total disruption is achieved once the cell components are disrupted and
the sample is evenly dispersed over the polymer material [23]. The end result is
that the entire dispersed sample becomes a unique chromatographic phase from
which either the analyte or matrix components can be selectively eluted using an
organic solvent or solvent mixture with the appropriate eluent strength. The
solvent mixture is usually water-immiscible.
The MSPD technique was originally applied to the isolation of drug residues
from animal tissues [25]. It has since been successfully applied to the wide
variety of food matrices shown in Table 25.2, including dairy or medical
products, animal tissues, vegetables, fruits, and aquatic species. It should be
noted that adsorbent consumption can be high for samples with high lipid
content, and that an additional clean-up step may be required for an extract
obtained from a complex biological matrix.
A particularly interesting application uses a miniaturized and automated
MSPD extraction method for the isolation of pesticides from fruit samples [40].
This method was optimized for a variety of organophosphorous pesticides and a
pyrethroid from oranges, but satisfactory recoveries were also obtained from
875
TABLE 25.2
MSPD clean-up and extraction of food matrices
Sample Analyte MSPD material Ref.
Citrus fruit Various pesticides C8 26
Oranges, grape, onions, Carbamate pesticides Cl, C8, cyano, amine and 27
tomatoes phenyl solid phases
Milk Organochlorine and ClS 28
organophosphorus pesticides
Milk Veterinary drugs C18 29
Medical foods Vitamin K, C18 30
Meat, milk, cheese Tetracyclines ClS 31
Fish Surfactants C18 32-34
Fish Triazine pesticides C8s 35
Beef fat Chlorinated pesticides C18 36
Liver P-agonists C8 , C18 37
Liver Clenbuterol ClS 38
Chicken muscle Sulfonamides C8S 39
pears and grapes. The method requires only 25 mg of sample and 100 A of
organic solvent. Solid-phase C8 was determined to be the optimum dispersion
material for this application.
The technique of matrix-solid phase dispersion can be adjusted to retain
particular compounds by choosing an appropriate dispersion material in addi-
tion to using a specific eluent. Most applications have utilized the reverse-phase
material C,,, in part because the solid silica support facilitates sample disruption
while silanol groups on the silica surface may associate with polar components in
the sample matrix [23]. However, a recent application has demonstrated that
clean-up from kidney tissue can be achieved with a cross-linked acrylic polymer
[41]. In this case, the acrylic polymer XAD-7 HP was able to retain lipid
components, such as fatty acids, sterols, and triglycerides, in addition to protein
matter in the presence of an ethanol-modified water eluent at 100°C. Figure 25.2
demonstrates how the kidney sample clean-up was achieved with the XAD-7 HP
resin.
A slightly different in situ sample clean-up technique has been employed for
the selective extraction of polychlorinated biphenyls (PCBs) from lard, fish, fish
meal, and cod-liver oil [42,43]. In these cases, a fat retainer was placed in
sequence inside an extraction thimble, rather than being dispersed through the
sample. The packing of the extraction thimble shown in Fig. 25.3 demonstrates
how matrix interferences are initially co-extracted from the sample, but are then
trapped by the fat-retainer inside the thimble. Several fat retainers have been
investigated, including sulfuric acid, Florisil, and basic, neutral, and acidic alu-
mina, for static extractions performed at 100°C, followed by elution with n-hex-
ane. Figure 25.4 demonstrates that the magnitude of fat retention is similar for
876
1 filter paper
flow SFE support
direction
Sand +Na2 SO4
Matrix +
Na 2SO 4 +sand
1 filter paper
Fat retainer
r
Silica gel
2 filter papers
Left: Fig. 25.2. Extract from kidney samples pretreated with or without the acrylic polymer
XAD-7 HP prior to pressurized liquid extraction with 30% ethanol in water at 100°C and 50 atm.
Samples: 0.5 g beef kidney + 2 g diatomaceous earth.
Right: Fig. 25.3. Packing of an extraction thimble with fat retainer. Reproduced with permission
from Ref. [42].
iB,
05"
I-
Fig. 25.4. Amount of fat retained for five fat retainers using different fat/fat retainer ratios.
Reproduced with permission from Ref. [42].
877
each retainer, but is dependent on the fat/fat retainer ratio. Fat retainers have
been utilized in a similar mode when conducting supercritical fluid extraction
(SFE) with CO 2 [44-48].
The analyses of liquid food samples have an advantage over those associated
with solid samples in that they usually require one less pretreatment step, due to
their liquid form. In some cases, very little sample preparation may be required
if the liquid is sufficiently free of matrix interferences. Straightforward tech-
niques that may used to prepare "clean" liquid samples prior to the analysis step
include sample dilution, evaporation, distillation, microdialysis, lyophilization,
or liquid-liquid extraction (LLE) [49]. Sample drying by lyophilization was
discussed in the previous section, and is particularly useful for the analysis of
nonvolatile organics. The technique of microdialysis is further discussed in
Section 25.3.4.
The technique of LLE is included in this list of "conventional" or straightfor-
ward methods since it is well-described in standard texts and references, and has
also been described in some recent reviews [50,51]. Further information on LLE
methods is also found in Chapter 11. The LLE technique is frequently utilized in
the analysis of toxicants, but can also be applied to food components, for example
in the extraction of low relative molecular mass compounds from food samples,
such as milk, soft drinks, wine, or beer. The extraction procedure generally
results in the separation of hydrophilic and lipophilic compounds, such as fat and
proteins, following a protein denaturation step with an acid or organic solvent,
or following solvent extraction under gentler conditions [16].
Other major techniques for the isolation or purification of liquid food
samples are solid-phase extraction (SPE), including immunoaffinity extraction
(IAE) and molecularly-imprinted polymers (MIPs); microextraction techniques,
including solid-phase microextraction (SPME) and spin bar sorptive extraction
(SBSE); and membrane extraction techniques, including dialysis. These meth-
ods are characterized by a reduced use of organic solvents, and the associated
toxic effects to the laboratory worker and the environment.
878
TABLE 25.3
Solid-phase extraction and clean-up of liquid foods and solvent extracts
Sample Anlalyte Solid-phase Ref.
SPE clean-up of solvent extracts
Fruits and vegetables Pe sticides C,,, CN 54
Apples, pears Be nzoylurea insecticides Silica 55
Meat extract Heeterocyclic amines Cs8 56
Liver, kidney, muscle Pe nicillin antibiotics Ion-exchange 57
SPE of liquid food
Red wine Pii gments Silica, C, CN, alumina, NH2, 58
Florisil, carbon black
Wine Various pesticides C8s, PS-DVB, Oasis cartridge 59
Alcoholic beverages Sy nthetic colours NH2 60
Orange juice Ca .rotenoids C1
, 61
Soft drinks Caffeine Cl 8 62
Fruit juices Phlenolic acids Cl 8 63
Butter (liquid) FiE avour compounds C,,, Cs, NH, CN, PS-DVB 64
Immunoaffinity SPE and clean-lup
Orange juice s-t Lriazine pesticides Immunoaffinity 65
Fruit and vegetables PhLenylurea herbidices Immunoaffinity 66
Fruit and vegetables Tr iazine herbicides Immunoaffinity 67
Fish Mi icrocystins Immunoaffinity 68
Grains M)icotoxin Immunoaffinity 69
Herbs Ni trated polycyclic Immunoaffinity 70
ariomatic hydrocarbon
Peanut butter, paprika, Af latoxins Immunoaffinity 71
pistachios
MIP SPE and clean-up
Liver Cl1enbuterol MIP 72
Chewing Gum Ni cotine and analogs MIP 73
Liver Tr iazine pesticides MIP 74
PS-DVB = Polystyrenedivinyl-benzene.
879
s-triazines in orange juice, where the cartridge containing the immobilized anti-
bodies is coupled on-line to a gas chromatograph via a reversed-phase cartridge
[65].
Table 25.3 also cites examples in which molecularly imprinted polymers
were used as solid-phase sorbents for the enrichment of analytes from liquid
foods and solvent extracts. MIPs are highly stable polymers that possess recogni-
tion sites within the polymer matrix that are specific for the three-dimensional
shape and functionalities of the analyte of interest [75]. For example, an MIP
material was utilized as part of a two-tier, on-line sample clean-up method
performed concurrently with sample extraction for the determination of clen-
buterol in liver. The liver samples were first blended with C,, in a MSPD
clean-up procedure, then a molecularly imprinted solid-phase extraction
cartridge was placed in-line after the MSPD cartridge to selectively trap the
analyte during elution with acetonitrile [72].
SPE methods may also use ion-exchange materials. For example, anion
exchange membranes have been utilized for the determination of glucosinolates
in canola and mustard seeds. The analytes in this case were isolated by
immersing the membranes in an aqueous suspension of the ground seeds. The
membrane was then removed from the suspension, washed, and submerged in
an appropriate solvent for elution inside of a shaken vial [76,77].
880
TABLE 25.4
SPME sampling of liquid foods and solvent extracts
Sample Analyte Fiber Ref.
Liquid food, headspace
Beer Alcohols and esters PA 79
Wine Flavours PDMS 80
Milk Fatty acids PA 81
Alcoholic beverages Flavours PA 82
Vanilla extracts Volatiles PA 83
Orange juice Flavours PDMS 84
Vegetable oils Volatiles DVB/Carboxene/PDMS 85
Wine Sulphides Carboxen/PDMS 86
Wine Diacetyl PDMS, CW-DVB 87
with a 100 tm PDMS SPME fiber [93]. However, such a SBSE technique does
not have the same selectivity as SPME.
Although the SBSE technique has only been recently developed, it has
already seen modest use in the food industry. Stir bar sorptive extraction has
been applied to the determination of dicarboximide fungicides in wine [94],
organochlorine pesticides and chlorobenzenes in fruit and vegetables [95,96],
benzoic acid in lemon-flavoured beverages [97], and flavour compounds in
strawberries [98].
881
to the analyte, and only partial sample clean-up is achieved by membrane
separation of lower molecular weight species from higher molecular weight
matrix components. A dialysis clean-up step is therefore often combined with a
subsequent enrichment technique, for example on an automated trace
enrichment of dialysates system, also known as ASTED.
Dialysis techniques are strictly not extraction techniques, unlike the non-
porous membrane extraction methods. However, they will be discussed here
nonetheless, since they are highly effective for the clean-up of liquid foods and
solvent extracts. Several on-line microdialysis methodologies have been devel-
oped for this purpose. For example, on-line microdialysis clean-up has been
coupled with liquid chromatography and programmable fluorescence detection
for the analysis of chicken liver fortified with fluoroquinolone antibacterials
[101]. Fluoroquinolones in eggs [102] and chicken liver and muscle [103], in
addition to sarafloxacin residues in fortified and incurred eggs [104], can be
determined in a similar manner.
On-line microdialysis has also been utilized to improve the sensitivity of a
disposable lactate biosensor used in the flow-injection mode for the analysis of
L-lactate in milk and yoghurt [105]. The measurement of lactulose in milk did
not require pre-treatment when a microdialysis probe was used as the sampling
system [106]. Glucose determination in milk and juice samples can be achieved
with little or no sample pretreatment using a microsystem that integrates a
microdialysis probe with a glucose oxidase bioreactor [107].
During a nonporous membrane extraction technique, a liquid or solid (e.g.
polymeric) phase is placed between two other phases, which are usually liquid
but sometimes gaseous [100]. The sample to be processed may be viewed as part
of the donor phase, while an acceptor phase on the other side of the membrane
collects the analyte for transfer to the analytical instrument. In this fashion,
unparalleled sample clean-up and analyte enrichment can be achieved when
compared with classical liquid-liquid extractions.
Nonporous membrane techniques have tremendous potential for the food
industry, although they as yet have seen limited use in this field. For example,
Vitamin E has been determined in butter samples after dissolution of the butter
in a micellar medium. Following on-line saponification, the nutrient was
enriched across a silicone membrane and taken up in acetonitrile prior to liquid
chromatographic analysis using an electrochemical detector [108]. Continuous
extraction of Vitamin E isomers from vegetable oils has been achieved in a
similar manner [109]. Another membrane separation device has been coupled to
a liquid chromatograph for the enrichment of pesticide multiresidues from egg
extracts generated via Soxhlet extraction [1101.
The supported liquid membrane (SLM) extraction principle can also be
extended to solid or semi-solid samples by incorporating a donor channel unit
that permits close contact between the sample and the membrane. Using such a
configuration, it has been possible to extract and quantify vanillin in food
samples (e.g. chocolate) [111] and caffeine in coffee and tea [112].
882
25.4 EXTRACTION OF SOLID SAMPLES
883
heat the sample with the application of microwaves. This type of heating is fast
and temperature gradients are kept to a minimum. A drawback to the technique
is the requirement for an extraction solvent that is able to absorb microwaves. In
addition, a subsequent clean-up step is usually required once the microwave
vessel has cooled sufficiently for handling.
Microwave techniques have been applied to biological and food samples quite
extensively. The first use of the microwave domestic oven in the laboratory was
for the determination of trace metals in biological samples [124]. This was
followed by the extraction of crude fat and nutrients from food [125], and such
solutes as pyrimidine-glucoside from seeds and fava beans [126]. A patented
variation of MAE is the microwave-assisted process [127], or MAP, which was
first applied to the extraction of essential oils from plant products [128]. MAP
methods mainly concern biological applications ranging from analytical to
processing scale.
All the applications cited thus far have utilized closed-vessel systems.
Recently, closed-vessel MAE has been used for a number of marine tissue
applications. For example, organic contaminants such as polychlorinated bi-
phenyl (PCB) congeners, chlorinated pesticides, and polycyclic aromatic hydro-
carbons (PAHs) have been extracted from standard reference materials [129]. In
this case, it was determined that the moisture content in the samples greatly
influenced analyte recovery, and it was necessary to standardize the moisture
content in all samples in a batch prior to extraction. Closed vessel MAE has also
been utilized for the determination of PCBs in freeze-dried mussels [130],
organochlorine compounds in cod liver and fish fillets [131], xenoestrogens in
liver and muscle tissue from rainbow trout [132], and ionic arsenic species in
oyster tissue [133], as well as sulphamethazine in swine tissue [134] and fat in
chocolate [135].
Several authors have investigated novel microwave extraction applications.
For example, a simultaneous extraction-derivatization procedure was developed
for the analysis of methylmercury in biological samples [136]. In addition, the
determination of the release of dimethyl sulfide from cereals and canola was
facilitate by MAE, producing a gaseous sample in the headspace above the
microwave-extracted food sample which could be injected to a gas chromato-
graph [137].
A microwave extraction procedure may also make use of an open vessel, in
what is called focused open-vessel microwave-assisted extraction (FOV-MAE).
In a closed vessel procedure, an additional sample treatment step may be
required to remove co-extracted water from a wet sample. However, in an open
vessel procedure, the water may be removed from the sample via azeotropic
distillation. This was shown to be the case during the determination of organo-
chlorine compounds in cod liver and fish fillets [131]. Open vessel procedures
have also been applied to the determination of polychlorinated biphenyls and
chlorinated pesticides in standard reference materials of cod liver oil and freeze-
dried mussel tissue [138], and mercury in fish [139].
884
Most of the MAE applications have concerned the determination of exoge-
nous species. Microwave extraction procedures are also useful for the character-
ization of food components. The MAE technique has been applied to the
determination of the fatty acid profile of mackerel and cod [140], trace element
analysis in plant materials [141], and the extraction of free amino acids from
foods [142]. During the extraction of lipids from milk samples, it was determined
that triglycerides are more stable using an MAE procedure compared with the
conventional Weibull-Berntrop extraction procedure, since there is less chemic-
al transformation of the triglycerides via hydrolysis [143].
A second technique that has been employed for the rapid extraction of food
samples at elevated temperatures is pressurized liquid extraction (PLE). PLE
methods frequently utilize the Accelerated Solvent Extraction (ASE) system
developed by Dionex, or any other system that performs static or dynamic
solvent extractions at elevated temperatures and pressures. The advantage to
performing extractions under pressurized conditions is that the upper extrac-
tion temperature is not limited by the boiling point of the solvent, as is the case
with the traditional Soxhlet system. A flow-through system such as the ASE is
also particularly beneficial in food analysis. Static extractions are performed
inside steel extraction vessels that have ample capacity for food samples, from
11-100 ml. The static extraction period is followed by elution of the extraction
solvent into a collection vial. PLE extracts are usually cleaner than microwave
extracts, since matrix components that do not dissolve in the extraction solvent
may be retained inside of the vessel. In addition, in situ clean-up methods can be
employed during PLE methods. While most food samples are blended with only
diatomaceous earth or sand prior to the PLE extraction step, further in situ
clean-up can be achieved by utilizing the matrix solid-phase dispersion tech-
nique (MSPD) [24-41] or by placing fat retainers in series inside of the vessel
[42,43]. Further details on these applications have been discussed in Section
25.2.4.
PLE methods have been utilized for both the extraction of food components
and the isolation of food contaminants. With regards to food composition, PLE
has been particularly useful in the determination of fats and lipids in food
samples such as meats [144] and egg-containing food [145]. The PLE technique
has also been applied to the determination of the fatty acid composition in cereal,
egg yolk, and chicken breast muscle [146]. PLE techniques have been employed
for the extraction and speciation of arsenic in freeze-dried carrots [147] and to
assess the levels of vitamin K-1 in medical foods [30].
Pressurized liquid extraction has also seen widespread use in the isolation of
contaminants from foodstuffs. There has been particular interest in applying
the technology to the analysis of lipid-containing foods. For example, organo-
chlorine compounds have been isolated from cod liver and fish fillets [131], and
885
various pesticides have been removed from baby foods [148]. PLE has also been
utilized for the determination of polycyclic aromatic hydrocarbons in smoked
fish and pork [149], and polychlorinated biphenyls in cod-liver oil and milk
powder [150], as well as fish tissue [151]. Corticosteroid residues in bovine liver
have been quantified using a two-step extraction method. Fat components were
first removed from the sample with hexane in a defatting step. This was followed
by the actual extraction of the analytes with a 1:1 hexane-ethyl acetate mixture
[152].
Pressurized liquid extraction has also been useful for the rapid analysis of
toxicants in fruits and vegetables. The technique has been applied to the
determination of fungicides in oranges and bananas [153], organophosphorous
pesticides in apples and carrots [154], and various pesticide residues in fresh
pear, cantaloupe, white potato, and cabbage [155].
A discussion on the utilization of hot, aqueous extraction solvents during the
analysis of food samples is included in this section. Hot, pressurized water, also
called subcritical water, is a novel extraction solvent that has been discussed at
length in an earlier chapter. The benefit of utilizing subcritical water for
analytical extractions is that the solvent strength can be tuned by varying the
extraction temperature and/or through the addition of a cosolvent. Water as a
solvent is easily obtained and disposed of, being benign to the laboratory worker
and the environment. Aqueous extractions of food samples are also convenient,
since the sample matrix does not need to be dried prior to the extraction step.
The application of this relatively new technology to the analysis of foods has
been limited thus far. The technique has been largely restricted to the isolation
of food components or contaminants from foods of plant origin. This is likely due
to the fact that hot water extracts from animal tissues can be quite turbid and
highly coloured [41]. Applications to foods of plant origin have included the
selective extraction of oxygenates from savory and peppermint [156], essential
oils from oregano [157], fennel [158], and marjoram [159], fungicides from
various vegetables [160], and organochlorine compounds from strawberries
[95,96], kohlrabi, lettuce, and tomatoes [96].
In our laboratory, the problem of coextractives from foods of animal origin
was overcome by incorporating an in situ MSPD clean-up technique into the
extraction procedure. In this case, the pesticide atrazine was isolated from beef
kidney that was dispersed with an acrylic polymer [41]. It was necessary to
modify the water with another benign solvent (ethanol at 30% v/v) in order to
attain the solvent strength necessary to achieve complete recovery of the target
analyte from the dispersed matrix.
The use of supercritical fluids in food analysis has grown tremendously in the
past decade. Two applications that have had significant development in recent
years have been the determination of fat and associated nutrients, as well as the
886
isolation of pesticides from food matrices. Supercritical fluid extraction (SFE)
technology has also been applied to the determination of PAHs and PCBs, drugs,
and other food toxicants. Many of these applications are summarized in Table
25.6, all of which utilize supercritical carbon dioxide (SC CO2). SC CO2 continues
to be the fluid of choice, since its critical parameters (31.1°C, 72.8 bar) are easily
achieved with high pressure instrumentation. Further, it is non-toxic and easy
to obtain. Some of the SF-based methodologies utilize suitable modifiers to
enhance analyte recovery. Protocols have been developed that encompass
sample clean-up, derivatization, and automation in the methodology. The
clean-up of lipids has been of critical importance during the extraction of
non-polar compounds such as pesticides, PAHs and PCBs, due to the propensity
of these toxicants to accumulate in the lipid phase.
It is probably fair to say that analytical SFE will be a method of choice for the
fat determination of foods in the future. Toward that end, several collaborative
and peer-verified methods have been published utilizing SFE for the determina-
tion of lipid levels in oilseeds, meats, and food products.
There has been limited use of alternative fluids for the analysis of food
samples. For example, three fluids were examined for the removal of ethoxyquin
from lean beef and beef fat: carbon dioxide, trifluoromethane, and 1,1,1,2-tetra-
fluoroethane. While CO2 appeared to react with the analyte during the
extraction, methanol-modified hydrofluorocarbons provided more complete ex-
tractions than the pure fluids. Quantitative extraction was achieved at the 0.5
ppm level [197]. Binary mixtures of CO 2 and nitrogen have also been utilized for
the selective extraction of organochlorine and organophosphate pesticides from
poultry. The binary mixtures provided quantitative recoveries of the analytes
while significantly reducing lipid solubility and coextraction [198].
REFERENCES
887
TABLE 25.6
Supercritical fluid extraction of food samples
Sample Analyte Ref.
Fatand Nutrients
Ground beef Total fat 161-163
Ground cumin Volatile oils 164
Egg-containing foods Cholesterol 165
Chocolate Triacylglycerols 135
Infant formula Fat 166
Milk products Total fat 167
Milk products Total fat 168
Deep-fried chicken and potato Lipids 169
Oilseeds Total fat 162,163
Bakery samples Total fat 163
Milk powder, infant formula Fat-soluble vitamins 170,171
Tomato skin Lycopene 172
Raisins 5-Hydroxymethyl-2-furaldehyde 173
Pesticides
Butter fat, corn oil Organochlorine and organophosphate 174
Wheat, maize Organophosphate 175
Eggs Triazine 176
Eggs Organochlorine 177
Sugar cane, oranges Diuron 178
Garlic Organochlorine 179
Cereals Various 180
Beef and chicken tissue Carbamate 181
Apples Fenpyroximate 182
Apples Various 183
Chicken fat, ground beef, lard Organochlorine and organophosphate 184
Grains Organochlorine and organophosphate 185
Poultry tissue Organophosphate 186
Drugs
Bovine liver 3-agonists 187
Poultry eggs and muscle Nicarbazin 188
Animal tissues Steroids 189
Eggs Sulfamethazine 190
Pig fat Androstenone 191
PAHs and PCBs
Smoked meat PAHs 18
Toasted bread PAHs 19
Liver PAHs 20
Fish PCBs 21
Fat PCBs 22
Miscellaneous
Irradiated foods Hydrocarbons and 2-alkylcyclobutanones 192,193
Cured meats Nitrosamines 194
Beef Liver Aflatoxin Ml 195
Corn Aftatoxin B1 196
888
6 C.A. Ramsey and J. Suggs, Environ. Test. Anal., 10(2) (2001) 13.
7 F.F. Pitard, Pierre Gy's Sampling Theory and Sampling Practice, 2 d Edn. CRC
Press, Boca Raton, 1993.
8 P.M. Gy, Sampling of Heterogeneous and Dynamic Material Systems. Elsevier,
Amsterdam, 1992.
9 P.M. Gy, Sampling of Particulate Materials, Theory and Practice. Elsevier,
Amsterdam, 1982.
10 R.F. Cross, LC/GC, 18 (2000) 468.
11 L.W. Aurand, A.E. Woods and M.R. Wells, Food Composition and Analysis. Van
Nostrand Reinhold, New York, 1987, Chapter 2.
12 M.A. Joslyn, Methods in Food Analysis, 2nd Edn. Academic Press, New York, 1970,
Chapter 2.
13 W.G. Cochran, Sampling Techniques. Wiley, New York, 1959.
14 M.A. Joslyn, Methods in Food Analysis, 2nd Edn. Academic Press, New York, 1970,
Chapter 3.
15 B. Berilter, X. Giard, C.J. Zizek and J. Fssler, Am. Lab., 33(21) (2001) 10.
16 H. S0rensen, S. Sorensen, C. Bjergegaard and S. Michaelsen, Chromatographyand
Capillary Electrophoresisin Food Analysis. Royal Society of Chemistry, Cambridge,
UK, 1999, p. 73.
17 Ibid., p. 71
18 M.Y. Ali and R.B. Cole, J. Agric. Food Chem., 49 (2001) 4192.
19 M.N. Kayali-Sayadi, S. Rubio-Barraso, R. Garcia-Iranzo and L.M. Polo-Diez, J. Liq.
Chrom. Rel. Technol., 23 (2000) 1913.
20 S.G. Amigo, M.S.G. Falcon, M.A.L. Yusty and J.S. Lozano, Fres. J. Anal. Chem., 367
(2000) 572.
21 K. Yasumura, E. Kitamura, M. Uno and M. Tamaki, J. FoodHyg. Soc. Jpn., 42 (2001)
1.
22 E. Bjtrklund, M. JBremo and L. Mathiasson, J. Liq. Chrom. Rel. Technol., 23 (2000)
2337.
23 S.A. Barker, J. Chromatogr.A, 885 (2000) 115.
24 S.A. Barker, J. Chromatogr.A, 880 (2000) 63.
25 S.A. Barker, A.R. Long and C.R. Short, J. Chromatogr., 475 (1989) 353.
26 A.I. Valenzuela, R. Lorenzini, M.J. Redondo and G. Font, J. Chromatogr.A, 839
(1999) 101.
27 M. Fernandez, Y. Pic6 and J. Mafies, J. Chromatogr.A, 871 (2000) 43.
28 C. Yagie, S. Bayarri, R. Ldzaro, P. Conchello, A. Arifo and A. Herrera, J. AOAC Int.,
84 (2001) 1561.
29 S.A. Barker and A.R. Long, J. AOAC Int., 77 (1994) 848.
30 G.W. Chase, Jr. and B. Thompson, J. AOAC Int., 83 (2000) 407.
31 E. Brandsteterova, P. Kubalec, L. BovanovA, P. Simko, A. Bednarikovd and L.
Machackovd, Z. Lebensm. Unters. Forsch. A, 205 (1997) 311.
32 J. Tolls, M. Haller and D.T.H.M. Sijm, Anal. Chem., 71 (1999) 5242.
33 J. Tolls, M. Haller and D.T.H.M. Sijm, J. Chromatogr.A, 839 (1999) 109.
34 M. Zhao, F. van der Wielen and P. de Voogt, J. Chromatogr. A, 837 (1999) 129.
35 P. Guant and S.A. Barker, Int. J. Environ. Pollut., 13 (2000) 284.
36 A.R. Long, M.M. Soliman and S.A. Barker, J. AOAC Int., 74 (1991) 493.
37 S. Collins, M. O'Keefe, R. Calverley and M.R. Smyth, Proceedingsof EuroresidueIII
(1996) 340.
38 E. Horne, M. O'Keefe, C. Desbrow and A. Howells, Analyst, 123 (1998) 2517.
39 K. Kishida and N. Furusawa, J. Chromatogr.A, 937 (2001) 49.
40 E.M. Kristenson, E.G.J. Haverkate, C.J. Slooten, L. Ramos, R.J.J. Vreuls and
U.A.Th. Brinkman, J. Chromatogr. A, 917 (2001) 277.
889
41 M.S.S. Curren and J.W. King, J. Agric. Food Chem., 49 (2001) 2175.
42 E. Bjorklund, A. Miller and C. von Holst, Anal. Chem., 73 (2001) 4050.
43 Dionex, Application Note ASE 322, Dionex Corporation, Sunnyvale, CA, 1996.
44 M. Jaremo, E. BjSrklund, M. Milsson, L. Karlsson and L. Mathiasson, J. Chromatogr.
A, 877 (2000) 167.
45 J.E. France, J.W. King and J.M. Snyder, J. Agric. Food Chem., 39 (1991) 1871.
46 R.C. Hale and M.O. Gaylor, Environ. Sci. Technol., 29 (1995) 1043.
47 E.G. Alley and G. Lu, J. AOAC Int., 78 (1995) 1051.
48 E. Bjorklund, M. Jiaremo and L. Mathiasson, J. Liq. Chrom. Rel. Technol., 23 (2000)
2337.
49 R.E. Majors, LC/GC, 14 (1996) 936.
50 J.R. Dean, Extraction Methods for Environmental Analysis, .iley, Chichester, UK,
1998.
51 A.J. Holden, in: A.J. Handley (Ed.), Extraction Methods in Organic Analysis.
Sheffield Academic Press, Sheffield, UK, 1999, p. 5.
52 V. Ruiz-Gutierrez and M.C. Perez-Camino, J. Chromatogr.A, 885 (2000) 321.
53 C. Dominguez, D.A. Guillen and C.G. Barroso, J. Chromatogr. A, 918 (2001) 303.
54 K. Nordmeyer and H.-P. Thier, Z. Lebensm. Unters. Forsch. A, 208 (1999) 259.
55 N.G. Tsiropoulos, P.G. Aplada-Sarlis and G.E. Miliadis, J. AOAC Int., 82 (1999) 213.
56 F. Toribio, E. Moyano, L. Puignou and M.T. Galceran, J. Chromatogr.A, 880 (2000)
101.
57 Y. Ito, Y. Ikai, H. Oka, H. Matsumoto, Y. Miyazaki, K. Takeba and H. Nagase, J.
Chromatogr. A, 911 (2001) 217.
58 G.A. Csiktusnddi Kiss, E. Forgacs, T. Cserhdti, M. Candeias, L. Vilas-Boas, R. Bronze
and I. Spranger, J. Chromatogr. A, 889 (2000) 51.
59 J.J. Jimbnez, J.L. Bernal, M.J. del Nozal, L. Toribio and E. Arias, J. Chromatogr.A,
919 (2001) 147.
60 S.M. Dugar, J.N. Leibowitz and R.H. Dyer, J. AOAC Int., 77 (1994) 1335.
61 J.F. Fisher and R.L Rouseff, J. Agric. Food Chem., 34 (1986) 985.
62 Y. Daghbouche, S. Garrigues, M.T. Vidal and M. de la Guardia, Anal. Chem., 69
(1997) 1086.
63 Y. Amakura, M. Okada, S. Tusji and Y. Tonogai, J. Chromatogr.A, 891 (2000) 183.
64 M. Adahchour, R.J.J. Vreuls, A. van der Heijden and U.A.Th. Brinkman, J.
Chromatogr. A, 844 (1999) 295.
65 J. Dallfige, T. Hankemeier, R.J.J. Vreuls and U.A.Th. Brinkman, J. Chromatogr.A,
830 (1999) 377.
66 J.F. Lawrence, C. M6nard, M.-C. Hennion, V. Pichon, F. Le Goffic and N. Durand, J.
Chromatogr. A, 732 (1996) 277.
67 J.F. Lawrence, C. M6nard, M.-C. Hennion, V. Pichon, F. Le Goffic and N. Durand, J.
Chromatogr. A, 752 (1996) 147.
68 J.F. Lawrence and C. M6nard, J. Chromatogr.A, 922 (2001) 111.
69 P. Zillner, J. Jodlbauer and W. Lindner, J. Chromatogr. A, 858 (1999) 167.
70 B. Spitzer, M. Cichna, P. Markl, G. Sontag, D. Knopp and R. Niessner, J.
Chromatogr.A, 880 (2000) 113.
71 J. Stroka, R. van Otterdijk and E. Anklam, J. Chromatogr.A, 904 (2000) 251.
72 C. Crescenzi, S. Bayoudh, P.A.G. Cormack, T. Klein and K. Ensing, Anal. Chem., 73
(2001) 2171.
73 A. Zander, P. Findlay, T. Renner, B. Sellergren and A. Swietlow, Anal. Chem., 70
(1998) 3304.
74 M.T. Muldoon and L.H. Stanker, Anal. Chem., 69 (1997) 803.
75 0. Ramstrom, K. Skudar, J. Haines, P. Patel and 0. Briiggeman, J. Agric. Food
Chem., 49 (2001) 2105.
890
76 A.M. Szmigielska, J.J. Schoenau and V. Levers, J. Agric. Food Chem., 48 (2000) 4487.
77 A.M. Szmigielska and J.J. Schoenau, J. Agric. Food Chem., 48 (2000) 5190.
78 J.C. Wu, W. Xie and J. Pawliszyn, Analyst, 125 (2000) 2216.
79 H.H. Jelen, K. Wlazly, E. Wasowicz and E. Kaminski, J. Agric. Food Chem., 46 (1998)
1469.
80 C. Sala, M. Mestres, M.P. Marti, O. Busto and J. Guasch, J. Chromatogr. A, 880
(2000) 93.
81 A.F. Gonzdlez-C6rdova and B. Vallejo-Cordoba, J. Agric. Food Chem., 49 (2001)
4603.
82 S.E. Ebeler, G.M. Sun, M. Datta, P. Stremple and A.K. Vickers, J. AOAC Int., 84
(2001) 479.
83 T. Sostaric, M.C. Boyce and E.E. Spickett, J. Agric. Food Chem., 48 (2000) 5802.
84 M. Jia, Q.H. Zhang and D.B. Min, J. Agric. Food Chem., 46 (1998) 2744.
85 H.H. Jelen, M. Obuchowska, R. Zawirska-Wojtasiak and E. Wasowicz, J. Agric. Food
Chem., 48 (2000) 2360.
86 M. Mestres, C. Sala, M.P. Marti, O. Busto and J. Guasch, J. Chromatogr. A, 835
(1999) 137.
87 Y. Hayasaka and E.J. Bartowsky, J. Agric. Food Chem., 47 (1999) 612.
88 A.L. Simplicio and L.V. Boas, J. Chromatogr.A, 833 (1999) 35.
89 J.J. Jimenez, J.L. Bernal, M.J. del Nozal, M.T. Martin and A.L. Mayorga, J. Chroma-
togr. A, 829 (1999) 269.
90 S.B. Hawthorne, D.J. Miller, J. Pawliszyn and C.L. Arthur, J. Chromatogr., 603
(1992) 185.
91 Z. Wang, B. Hennion, L. Urruty and M. Montury, FoodAdditives Contam., 17 (2000)
915.
92 C.G. Zambonin, L. Monaci and A. Aresta, Food Chem., 75 (2001) 249.
93 E. Baltussen, P. Sandra, F. David and C. Cramers, J. Microcolumn Sep., 11 (1999)
737.
94 P. Sandra, B. Tienpont, J. Vercammen, A. Tredoux, T. Sandra and F. David, J.
Chromatogr.A, 928 (2001) 117.
95 L. Wennrich, P. Popp, G. Koller and J. Breuste, J. AOAC Int., 84 (2001) 1194.
96 L. Wennrich, P. Popp and J. Breuste, Chromatographia,Suppl. 53 (2001) S-380.
97 A.G.J. Tredoux, H.H. Lauer, T. Heideman and P. Sandra, J. High Resolut. Chroma-
togr., 23 (2000) 644.
98 M. Kreck, A. Scharrer, S. Bilke and A. Mosandl, Eur.FoodRes. Tech., 213 (2001) 389.
99 J.A. Jonsson and L. Mathiasson, in: P.R. Brown and E. Grushka (Ed.), Advances in
Chromatography. Marcel Dekker, New York, 2001, Chapter 2.
100 J.A. J6nsson and L. Mathiasson, J. Chromatogr. A, 902 (2000) 205.
101 R.J. Maxwell and E. Cohen, J. High Resolut. Chromatogr., 21 (1998) 241.
102 M.J. Schneider and D.J. Donoghue, J. AOAC Int., 83 (2000) 1306.
103 M.J. Schneider, J. Chromatogr. Sci., 39 (2001) 351.
104 R.J. Maxwell, E. Cohen and D.J. Donoghue, J. Agric. Food Chem., 47 (1999) 1563.
105 F. Palmisano, M. Quinto, R. Rizzi and P.G. Zambonin, Analyst, 126 (2001) 866.
106 D. Moscone, R.A. Bernardo, E. Marconi, A. Amine and G. Palleschi, Analyst, 124
(1999) 325.
107 S. Mannino, O. Brenna, S. Buratti and M.S. Cosio, Electroanalysis9 (1997) 1337.
108 M.M.D. Zamarreno, A.S. Perez, M.B.Rangel and J.H. Mendez, Anal. Chim. Acta, 386
(1999) 99.
109 A. Sanchez-Perez, M.M. Delgado-Zamarreno, M. Bustamante-Rangel, and J.
Hernandez-Mendez, J. Chromatogr. A, 881 (2000) 229.
110 R. Carabias-Martinez, E. Rodriguez-Gonzalo, P.H. Paniagua-Marcos and J.
Hernandez-Mendez, J. Chromatogr. A, 869 (2000) 427.
891
111 M. Luque, E. Luque-Perez, A. Rios and M. Valcdrcel, Anal. Chim. Acta, 410 (2000)
127.
112 E. Luque-PBrez, A. Rios, M. Valcarcel, L.-G. Danielsson and F. Ingman, Lab. Autom.
Inform. Manage., 34 (1999) 131.
113 J.B. Cai, B.H. Liu and Q.D. Su, J. Chromatogr.A, 930 (2001) 1.
114 C. Palma-Harris, R.F. McFeeters and H.P. Fleming, J. Agric. Food. Chem., 49 (2001)
4203.
115 D. Shooter, N. Jayatissa and N. Renner, J. Dairy Res., 66 (1999) 115.
116 J. Sung, B.D Gardner, J.F. Holland and R.M. Beaudry, J. Agric. Food Chem., 45
(1997) 1801.
117 E.P. Jarvenpaa, Z. Zhang, R. Huopalahti and J.W. King, Z. Lebensm. Unters. Forsch
A, 207 (1998) 39.
118 H.W. Chin, R.A. Bernhard and M. Rosenberg, J. Food Sci., 61 (1996) 1118.
119 N.P. Sen, S.W. Seaman and B.D. Page, J. Chromatogr.A, 788 (1997) 131.
120 M. Zhu, F.J. Aviles, E.D. Conte, D.W. Miller and P.W. Perschbacher, J. Chromatogr.
A, 833 (1999) 223.
121 J. Song, L. Fan and R.M. Beaudry, J. Agric. Food Chem., 46 (1998) 3721.
122 J.W. Arnold and S.D. Senter, J. Sci. Food Agric., 78 (1998) 343.
123 C.S. Eskilsson and E. Bjorklund, J. Chromatogr.A, 902 (2000) 227.
124 A. Abu-Samra, J.S. Morris and S.R. Koirtyohann, Anal. Chem., 47 (1975) 1475.
125 K. Ganzler, A. Salg6 and K. Valk6, J. Chromatogr., 371 (1986) 299.
126 K. Ganzler, A. Salg6, Z. Lebensm. Unters. Forsch., 184 (1987) 274.
127 J.R.J. Par6, J.M.R. BBlanger and S.S. Stafford, Trends Anal. Chem., 13 (1994) 176.
128 J.R.J. Pare, US Patent 5 002 784 (1991).
129 S. Jayaraman, R.J. Pruell and R. McKinney, Chemosphere, 44 (2001) 181.
130 N. Carro, I. Garcia and M. Llompart, Analusis, 28 (2000) 720.
131 M. Weichbrodt, W. Vetter and B. Luckas, J. AOACInt., 83 (2000) 1334.
132 S.N. Pedersen and C. Lindholst, J. Chromatogr. A, 864 (1999) 17.
133 A. Chatterjee, Talanta, 51 (2000) 303.
134 M.H. Akhtar, M.L. Wong, S.R.H. Crooks and A. Sauve, Food Additives Contam., 15
(1998) 542.
135 C. Simoneau, C. Naudin, P. Hannaert and E. Anklam, Food Res. Int., 33 (2000) 733.
136 M. Abuin, A.M. Carro and R.A. Lorenzo, J. Chromatogr. A, 889 (2000) 185.
137 Y.L. Ren, J. Agric. Food Chem., 49 (2001) 1737.
138 S. Thompson and H. Budzinski, Int. J. Env. Anal. Chem., 76 (2000) 49.
139 C.S. Chiou, S.J. Jiang and K.S.K. Danadurai, Spectrochim. Acta B, 56 (2001) 1133.
140 A. Batista, W. Vetter and B. Luckas, Eur. Food Res. Technol., 212 (2001) 377.
141 J. Borkowska-Burnecka, Fres. J. Anal. Chem., 368 (2000) 633.
142 A. Kovacs, K. Ganzler and L. Simon-Sarkadi, Z. Lebensm. Unters. Forsch., 207 (1998)
26.
143 L.E. Garcia-Ayuso, J. Velasco, M.C. Dobarganes and M.D.L. de Castro, Int. Dairy J.,
9 (1999) 667.
144 Dionex, Application Note ASE 334, Dionex Corporation, Sunnyvale, CA, 1999.
145 E. Boselli, V. Velazco, M.F. Caboni and G. Lercker, J. Chromatogr. A, 917 (2001) 239.
146 K. Schafer, Anal. Chim. Acta, 358 (1998) 69.
147 N.P. Vela, D.T. Heitkemper and K.R. Stewart, Analyst, 126 (2001) 1011.
148 J.C. Chuang, K. Hart, J.S. Chang, L.E. Boman, J.M. Van Emon and A.W. Reed, Anal.
Chim. Acta, 444 (2001) 87.
149 G.D. Wang, A.S. Lee, M. Lewis, B. Kamath and R.K. Archer, J. Agric. Food Chem., 47
(1999) 1062.
150 A. Miiller, E. Bjbrklund and C. von Holst, J. Chromatogr. A, 925 (2001) 197.
151 Dionex, Application Note ASE 322, Dionex Corporation, Sunnyvale, CA, 1996.
892
152 R. Draisci, C. Marchiafava, L. Palleschi, P. Cammarata and S. Cavalli, J. Chroma-
togr. B, 753 (2001) 217.
153 S. Kakimoto, H. Obana, M. Okihashi and S. Hori, J. Food Hyg. Soc. Japan,38 (1997)
358.
154 B.E. Richter, F. Hoefler and M. Linkerhaegner, LC/GC, 19 (2001) 408.
155 K. Adou, W.R. Bontoyan and P.J. Sweeney, J. Agric. Food Chem., 49 (2001) 4153.
156 A. Kubdtovb, A.J.M. Lagadec, D.J. Miller and S.B. Hawthorne, Flav. Frag. J., 19
(2001) 64.
157 R.S. Ayala and M.D.L. de Castro, Food Chem., 75 (2001) 109.
158 L. Gamiz-Garcia and M.D.L de Castro, Talanta, 51 (2000) 1179.
159 M.M. Jimenez-Carmona, J.L. Ubera and M.D.L. de Castro, J. Chromatogr.A, 855
(1999) 625.
160 T.M. Pawlowski and C.F. Poole, J. Agric. Food Chem., 46 (1998) 3124.
161 F.J. Eller and J.W. King, J. Agric. Food Chem., 49 (2001) 4609.
162 S.L. Taylor, F.J. Eller and J.W. King, Food. Res. Int., 30 (1997) 365.
163 F.J. Eller and J.W. King, J. Agric. Food Chem., 46 (1998) 3657.
164 D.L. Heikes, B. Scott and N.A. Gorzovalitis, J. AOAC Int., 84 (2001) 1130.
165 E. Boselli, M.F. Caboni and G. Lercker, Eur. Food Res. Technol., 212 (2001) 244.
166 M. Ashraf-Khorassani, R. Hellmer, L.T. Taylor and D.C. Messer, Am. Lab., 32 (2000)
40.
167 L. Manganiello, A. Rios and M. Valcarcel, J. Chromatogr.A, 874 (2000) 265.
168 F. Dionisi, B. Hug, J.M. Aeschlimann and A. Houllemar, J. FoodSci., 64 (1999) 612.
169 N. Devineni, P. Mallikarjunan, M.S. Chinnan and R.D. Phillips, J. Am. Oil. Chem.
Soc., 74 (1997) 1517.
170 C. Turner, M. Persson, L. Mathiasson, P. Adlercreutz and J.W. King, Enzyme Microb.
Technol., 29 (2001) 111.
171 C. Turner and L. Mathiasson, J. Chromatogr.A, 874 (2000) 275.
172 M. Ollanketo, K. Hartonen, M.L. Riekkola, Y. Holm and R. Hiltunen, Eur. Food Res.
Technol., 212 (2001) 561.
173 M. Palma and L.T. Taylor, J. Agric. Food Chem., 49 (2001) 628.
174 M.L. Hopper, J. Chromatogr. A, 840 (1999) 93.
175 K.N.T. Norman and S.H.W. Panton, J. Chromatogr.A, 907 (2001) 247.
176 J.W. Pensabene, W. Fiddler and D.J. Donoghue, J. Agric. Food Chem., 48 (2000)
1668.
177 W. Fiddler, J.W. Pensabene, R.A. Gates and D.J. Donoghue, J. Agric. Food Chem., 47
(1999) 206.
178 F.M. Lancas and S.R. Rissato, J. Microcolumn. Sep., 10 (1998) 473.
179 J.H. Wang, Q.A. Xu and K. Jiao, J. Chromatogr. A, 818 (1998) 138.
180 K. Yoshii, Y. Tonogai, Y. Tsumura, Y. Nakamura and T. Shibata, J. Food Hyg. Soc.
Japan,39 (1998) 184.
181 K.J. Voorhees, A.A. Gharaibeh and B. Murugaverl, J. Agric. Food Chem., 46 (1998)
2353.
182 B.L. Halvorsen, C. Thomsen, T. Greibrokk and E. Lundanes, J. Chromatogr. A, 880
(2000) 121.
183 R. Stefani, M. Buzzi and R. Grazzi, J. Chromatogr. A, 782 (1997) 123.
184 K.-S. Nam and J.W. King, J. High Resolut. Chromatogr., 17 (1994) 577.
185 J.W. King, M.L. Hopper, R.G. Luchtefeld, S.L. Taylor and W.L. Orton, J. AOAC Int.,
76 (1993) 857.
186 J.M. Snyder, J.W. King, L.D. Rowe and J.A. Woerner, J. AOAC Int., 76 (1993) 888.
187 M.J. O'Keefe, M. O'Keefe and J.D. Glennon, Analyst, 124 (1999) 1355.
188 D.K. Matabudul, N.T. Crosby and S. Sumar, Analyst, 124 (1999) 499.
189 A.A.M. Stolker, P.W. Zoontjes and L.A. van Ginkel, Analyst, 123 (1998) 2671.
893
190 J.W. Pensabene, W. Fiddler and D.J. Donoghue, J. Food Sci., 63 (1998) 25.
191 M.A. Magard, H.E.B. Berg, V. Tagesson, M.L.G. Jremo, L.L.H. Karlsson, L.J.E.
Mathiasson, M. Bonneau and J. Hansen-Moller, J. Agric. Food Chem., 43 (1995) 114.
192 P. Horvatovich, M. Miesch, C. Hasselmann and E. Marchioni, J. Chromatogr.A, 897
(2000) 259.
193 E.M. Stewart, W.C. McRoberts, J.T.G. Hamilton and W.D. Graham, J. AOAC Int., 84
(2001) 976.
194 J.W. Pensabene and W. Fiddler, Supercrit. Fluid Meth. Protoc., 13 (2000) 23.
195 S.L. Taylor, J.W. King, J.I. Greer and J.L. Richard, J. Food Protect., 60 (1997) 698.
196 S.L Taylor, J.W. King, J.L. Richard and J.I Greer, J. Agric. Food Chem., 41 (1993)
910.
197 D.R. Brannegan, M. Ashraf-Khorassani and L.T. Taylor, Chromatographia, 54
(2001) 399.
198 J.W. King and Z. Zhang, Anal. Chem., 70 (1998) 1431.
894
Chapter26
26.1 INTRODUCTION
896
26.2 DESIRED DELIVERABLES OF ENVIRONMENTAL ANALYTICAL
CHEMISTRY
897
studies related to natural attenuation [7,26]. Current on-site characterization
methods based on immunoassay [27-29] and colorimetry [30] suffer from severe
interference from other soil organic constituents. The immunoassay is based on
a reaction between a targeted chemical and a specific antibody to produce a
response that can be quantitated using radioactivity or fluorescence detection.
Colorimetric detection is based on reacting the pollutants with a chemical
reagent to produce a color that can be monitored by UV/VIS detection. In the
case of cyclic nitramine explosives such as RDX and HMX the compounds are
allowed to react with Zn/HCl to produce a pink to yellow color which can then be
quantified using a portable spectrophotometer at X 607 nm [29,30].
Analytical chemistry must also be able to provide an accurate and precise
assessment of the performance of soil (bio)remediation technologies. At present,
most field-scale remediation technologies are assessed based on the removal of
the pollutant while the degradation products and their fate generally remain
unknown [26,31]. The need for fast analytical techniques capable of performing
on-line analysis of soil during remediation technologies is a challenge that
should attract exploration by analytical chemists.
Last but not least, analysis of environmental samples, particularly soil,
should focus more on solvent-free sample preparation techniques. Solventless
sample preparation should be a legitimate strategy for soil analysis since the
ultimate objective of identifying the type and concentration of pollutants is to
safely remove them from the contaminated soil. Many solventless sample prepa-
ration techniques have already been developed and are available commercially.
Some of these techniques are based on SFE, SPME and headspace techniques as
summarized in Fig. 26.1. Mester et al. [32] have reviewed most currently
practised solventless sample preparation techniques for the analysis of several
environmental matrices including biological, sediment and plants. The following
discussion will describe some of these solvent-free sample preparation tech-
niques and compare their performance with other solvent-based traditional
methods such as the century-old Soxhlet [33].
898
26.3 SAMPLE PREPARATION PROTOCOLS FOR SOIL ANALYSIS
The most difficult step during soil analysis is sample preparation, which requires
that the targeted analyte(s) be effectively extracted from the soil matrix suffi-
ciently pure for subsequent detection and quantification. Several types of sam-
ple extraction techniques are currently available for soil analysis (Table 26.2).
Selection among these techniques depends on the type of soil and the type(s) of
analyte(s) to be extracted. To achieve optimal analyte recovery from soil, a
proper solvent with a particular polarity is routinely used in either a Soxhlet or a
sonicator. For instance, hydrophobic aliphatic hydrocarbons (oil and grease) can
easily be extracted from soil with a nonpolar solvent such as hexane (EPA 9071
[4]), whereas the more polar PAHs require a more polar solvent such as methyl-
ene dichloride. Also, polynitroorganic compounds such as explosives (EPA 8330
[4]) and polychlorinated compounds such as PCBs [34] are best extracted with
polar solvents such as acetonitrile or a mixture of hexane and a polar solvent
such as acetone, respectively. Where the soil is contaminated with VOCs such as
BTEX, gasoline, TCE, PCE, VC and metal hydrides, a headspace technique is
used for direct injection into a GC. Following analyte extraction from soil,
extracts are subject to several other manipulations including concentration and
chromatographic separation prior to detection. As mentioned earlier, for envi-
ronmental consideration the trend in soil analysis is now focusing increasingly
on solventless sample preparation techniques as shown in Fig. 26.1.
In the following sections we will describe specific examples of analyte extrac-
tion from soil and other related biological matrices followed by proper clean-up
procedures prior to analyte detection with a suitable detector for identification
and quantification.
Generally, more attention has been given to PHCs than any other family of
compounds due to their vast abundance and wide spread use. Over 2 billion
metric tons of petroleum products are used annually in the chemical, petrochem-
icals and transportation industries, resulting in severe accidental spills and
contamination in the terrestrial and marine environments [35]. Due to the
toxicity and carcinogenicity of some of these products, particularly PAHs, consi-
derable effort and capital are allocated to monitor their fate in the environment
[36]. For years the Environmental Protection Agency (EPA) Methods 418.1,
413.2 [37], 3560/8440 and 9071 [4] have been widely used to analyze PHCs in
soils, but recently these methods have been found to be inappropriate for certain
types of soil [38]. For example, discrepancies as high as 300% are obtained when
PHCs are analyzed in sediments [39]. Douglas et al. [35] recently reviewed the
limitations associated with some of the most prominent Standard Methods in
this field. The standardized EPA methods for the analysis of petroleum hydro-
carbons are derived from the American Public Health Association Method
899
TABLE 26.2
Soil sample preparation methods
Extraction type Analytes Extraction solvents Characteristics/Limitation
Soxhlet extraction Nonvolatile and Methylene chloride, EPA certified # 3540/
[4]; automated semivolatile organic
hexane, cyclohexane, 3541. Large volumes of
Soxhlet extraction compounds (TPHs, acetone/hexane (1:1, organic solvents (500 ml)
[4] PCBs, PAHs) v/v), toluene/methanol and long extraction times
(10:1, v/v) are needed
Pressurized fluid Water insoluble Acetone/hexane (1:1) EPA certified # 3545. Uses
extraction (PFE) semivolatile organics for organochlorine less solvent and less time
[4,8,9] (organophosphorus pesticides or PCBs. than Soxhlet.
and organochlorine Methylene chloride for
pesticides, chlorin- organophosphorus
ated herbicides, and pesticides. Acetone/
PCBs). methylene chloride/
phosphoric acid for
chlorinated herbicides.
Ultrasonic Nonvolatile and Similar to PFE method EPA certified # 3550/
extraction [4] semivolatile organic without organophos- 8330. Unsuitable for low
compounds (not phorus compounds. concentrations. Normally
validated for Acetonitrile for rapid but takes 16 h for
organophosphorus explosives. explosives.
compounds [10]).
Supercritical fluid TPHs, PAHs. Supercritical carbon EPA certified # 3560/
extraction (SFE) Organochlorine and dioxide (SC-CO2) for 3561. Can extract a wide
[4] organophosphate THPs [11] and PAHs. range of nonvolatile and
pesticides. Metals. SC-CO 2 with methanol semi-volatile organics and
Explosives [14]. co-solvent for pesticides metals. Not suitable for
[12] and metals [13]. volatiles such as gasoline.
SC-CO 2 with acetonitrile Co-solvent enhances
co-solvent for explosives. recoveries.
Equilibrium Volatile organic Solvent-free. EPA certified # 5021.
headspace [4,15] compounds (VOCs; Suitable for the extraction
BTEX, VC, TCE, of a wide range of volatile
PCE) organics with bp lower
than 180°C.
Vacuum Volatile organic Solvent-free. EPA certified # 5032.
distillation [4,16] compounds (VOCs)
Headspace SPME VOCs, HAPs, Solvent-free. Suitable for automatic
[17] chlorinated on-line SPME/GC and
benzenes, heavy SPME-ICP-MS analysis.
metals [18]
Purge-and-trap Volatile organic Solvent-free. EPA certified # 5035.
[4] compounds (VOCs) Suitable for on-line
injection and analysis.
Thermal Semi-volatile Solvent free. EPA certified # 8275.
extraction (TE) organics. Suitable for on-line
[4] injection and analysis.
Acid digestion [4]; Total recoverable Strong acids. EPA certified # 3050/
microwave metals. 3051. Corrosive.
assisted acid
digestion
900
(APHA [40]), originally developed to analyze mineral oil and grease in water and
sludge. According to the APHA method, any extractable chemical with a C-H
bond that absorbs in the IR region 2700-3200 cm 1 is measured as mineral oil
and grease.
In general, current understanding of the environmental fate of PHCs is
limited due to the lack of uniform analytical methods for the identification and
quantitation of complex mixtures of PHCs, particularly in soil. Sorption of PHCs
onto soil particles and the co-extraction of other organic material from soil are
some of the major problems encountered during soil analysis for PHCs. In
addition petroleum consists of a complex mixture of products ranging from
straight chain and branched aliphatics (paraffins) to simple and complicated
PAHs, necessitating a compound class separation prior to analysis. A successful
analytical protocol for the determination of PHCs in soil must thus take the
above limitations into consideration. Recently, there has been a genuine interest
from several researchers and agencies to improve and optimize standard
methods that have been developed earlier for the analysis of PHCs in soil [35,41].
At present, more specific techniques based on fingerprinting are considered for
the determination of TPHs in several soil matrices for the identification of the
source of contamination. Source determination is important for companies
facing liability claims [42]. Fingerprinting tools are based on extensive and
elaborate chromatographic and mass spectrometric techniques.
Extraction of semi-volatile petroleum hydrocarbons with an organic solvent
In a typical example, a soil sample from a refinery site with at least 10-15 years'
history of contamination with bunker crude oil is sieved (10 mesh), mixed with
MgSO4 and homogenized. Hexadecane and p-terphenyl are used as recovery
standards for aliphatic hydrocarbons and PAHs, respectively. The sample is
extracted as described in EPA Method 3450 [4] with an organic solvent such as
methylene chloride (Freon was used earlier but banned following Montreal
protocol in 1995 [43]) in Soxhlet for a total of 80 cycles. In some cases the Soxhlet
method was replaced with sonication using a suitable solvent such as hexane (30
min). Final extracts are dried over sodium sulfate, concentrated and eluted
through a silica column using solvents with different polarities, starting with the
least polar (hexane) and ending with the most polar one (methanol). Hydropho-
bic aliphatic compounds are first eluted with hexane whereas PAHs are eluted
with carbon tetrachloride or hexane mixed with methylene chloride (4:1, v/v).
Purified extracts are then analyzed for PHCs with IR spectrometry at 2700-
3200 cm- ' (vH, 2932 cm'), GC/MS or GC/FID. Figure 26.2 represents a typical
GC/MS chromatogram of various fractions collected during the analysis of TPH
in soil. Alternatively PHCs extracts can be quantified gravimetrically after
completely removing the solvent. A mass balance can then be drawn between
crude soil extracts and the collected fractions.
Solventless extraction of PHCs from soil
To reduce the amount of solvents and the time required for analysis, researchers
901
Aliphatic
L
Polar
Time (min)
Fig. 26.2. Typical GC/MS chromatogram of various fractions collected during silica clean-up of
Soxhlet extracts of petroleum contaminated soil.
902
4+5 A
1 31
.1
r"@+1-
i
',Q'S1N
I6 ,,,, ,-
I lr-n~~n nn M~~~~lirI1
4+5
36
___4'J ,I ..; 1 - ,
4+5 C
6
3
I
Time (min.)
Fig. 26.3. A headspace GC/FID chromatogram of BTEX in a contaminated soil. (A) BTEX
standard: 1, benzene; 2, toluene; 3, ethylbenzene; 4+5, m,p-xylenes; 6, o-xylene). (B) Gasoline
standard. (C) contaminated soil sample.
903
29.6 min.
A
710
74 110
92 150
50 2 1 I 1 128 14 _ 4
' i! ' i I j / 1194 '148
2Fs 28 .13 327340
20' 40 60 80 io0 ii20 40 0 ibo0 0o
i 22o 240 20 20O 360 320 340
m/z
Fig. 26.4. A GCJMS Selective Ion Monitoring (SIM) chromatogram of: (A) soil extract contam-
inated with Aroclor 1248; and (B) the mass spectrum of 2,3',5,5'-tetrachloro-1,1'-biphenyl
congener at 29.6 min.
904
2 3
Time (min.)
Fig. 26.5. Headspace GC/FID chromatograms of: (A) VOCs standard; 1, chloroethane; 2,
1,1-dichloroethene; 3, 1,1-dichloroethane; 4, 1,2-dichloroethane; 5, 1,1,1-trichloroethane); and
(B) VOCs obtained from the headspace of a contaminated soil.
min. The extracts are treated with acid and copper and fractionated on an
Alumina column using methylene chloride as eluent. The collected methylene
chloride fractions are concentrated and analyzed with a GC/ECD or GC/MSD.
In cases of volatile chlorinated solvents such as PCE, TCE and VC, the soil
sample can be analyzed directly with a suitable GC using a headspace technique
as described earlier with BTEX (Fig. 26.5).
905
At present, explosives are extracted from soil by sonication with acetonitrile
by EPA Method # 8330 [4]. Briefly, soil samples are air-dried in a fume hood in
the dark to a constant weight at room temperature then saturated with acetone
to achieve homogeneous distribution of the contaminant in the soil. The soil is
ground with a mortar, then thoroughly sonicated with acetonitrile for 18 h for
subsequent analysis by HPLC/UV (X54,m) Detection limits for TNT, RDX and
HMX are < 1 mg/kg. Alternatively, the soil can be directly extracted by shaking
the soil sample in acetone for 30 min. However, acetone is a powerful polar
solvent that can extract several other polar organic compounds from soil in
addition to the explosive thus causing interference with analysis. Also, acetone
absorbs at 2 54 nm, which is the wavelength used in EPA Method # 8330 [4] to
detect and quantify explosives. As discussed below, it is possible to minimize
interference caused by co-extracted chemicals from soil using supercritical fluid
extraction with carbon dioxide (SC-CO,). Furthermore, techniques that are
more specific than IR such as GC-MS and LC/MS can be used for explosives
detection and quantification.
The recent discovery of solid phase microextraction (SPME) [17,59] has
reduced the analysis time required for analysis and lowered the detection limit of
analytes. SPME is a solventless and rapid sample preparation technique that
uses a polymer-coated fiber for the adsorption of organic compounds from an
aqueous solution or the headspace of a soil sample followed by direct thermal
desorption into a GC or HPLC for subsequent detection and quantification. The
technique is known for its sensitivity, which enables detection in the ug 1 ' range.
Further details on the theory, applications and limitations of the SPME tech-
nique are available elsewhere [17,59]. Table 26.3 shows the concentration of
TNT and its derivatives as determined by SPME/GC-MS in a wet soil sample
obtained from a ditch near a TNT manufacturing plant. The analysis was done
by inserting a polydimethylsiloxane coated fiber in the water phase followed by
desorption inside a GC injector connected to a MS detector. Table 26.3 shows
TNT and several of its derivatives including 2-NT (2-nitrotoluene), 3-NT (3-
nitrotoluene), 4-NT (4-nitrotoluene), 2,6-DNT (2,6-dinitrotoluene), 2,5-DNT
(2,5-dinitrotoluene), 2,3-DNT (2,3-dinitrotoluene), 2,4-DNT (2,4-dinitrotolu-
ene), 3,5-DNT (3,5-dinitrotoluene), 3,4-DNT (3,4-dinitrotoluene), TNT (2,4,6-
trinitrotoluene), 4-ADNT (4-aminodinitrotoluene) and 2-ADNT (2-aminodi-
nitrotoluene). Table 26.3 shows the concentration of these products as found by
the standard HPLC based EPA method 8330 and by SPME/GC-MS [4]. In gen-
eral, a correlation factor of 90-100% was obtained between the two methods,
although the limit of detection (DL) by both techniques was found to be 10 ppb
and 0.5 ppm, respectively.
26.3.4 Agrochemicals
906
TABLE 26.3
Determination of TNT and its derivatives (mg/kg) in water taken from a contaminated area
nearby a ditch at a TNT manufacturing plant: SPME/GC-MS versus EPA Method 8330
Compounds SPME/GC-MS (RSD) HPLC/UV (RSD)*
2-Nitrotoluene 3.78 (2.26) 2.0 (1.50)
3-Nitrotoluene 1.11 (2.74) nd
4-Nitrotoluene 0.29 (2.10) nd
2,6-Dinitrotoluene 97.27 (5.40) 100.0 (0.21)
2,5-Dinitrotoluene 8.60 (5.85) nd
2,3-Dinitrotoluene a 12.6 (0.17)
2,4-Dinitrotoluene 100.49 (4.93) 80.4 (0.03)
3,5-Dinitrotoluene 2.39 (4.87) nd
3,4-Dinitrotoluene 35.1 (4.56) b
2,4,6-Trinitrotoluene 50.16 (3.25) 60.5 (0.09)
TNT-isomer 1.06 (7.47) nd
TNT-isomer 0.10 (6.18) nd
TNT-isomer 0.44 (18.9) nd
4-Amino 2,6-dinitrotoluene 0.34 (6.85) nd
2-Amino 4,6-dinitrotoluene 0.58 (10.1) nd
*Detected at 254 nm, RSD was calculated based on triplicate analysis. nd: Not detected.
a: 2,3-DNT overlapped with 2,4-DNT; b: 3,4-DNT overlapped with 2,6-DNT.
907
with tetrabutylammonium hydroxide or methyl iodide [671) to less polar and
more soluble forms in CO2 .
Alternatively bound pesticide residues may be determined by immunochemi-
cal methods such ELISA that can recognize the pesticide directly bound to humic
acid materials without bond cleavage [68]. Other extraction methods such as
MAE and PFE have been developed to minimize the use of solvents or to reduce
the extraction time [69].
908
their degradation products in aqueous and solid matrices using in situ deriva-
tization or acid-modified SC-CO2 . For a wider application of the SFE technique
for the extraction and determination of heavy metals, the reader can refer to
another review by Chester et al. [44]. With the ability to extract a wide range of
heavy metals, this technique is important for the study of nuclear solid waste
remediation and other soils heavily contaminated with metals.
Extraction of metals from soil can be accomplished using the previously
described SPME. Mester et al. [74] reported a coupled SPME-ICP/MS technique
for the determination of volatile metal hydrides such as Se, Sn, As and Sb
(generated in situ) or methylated species in case of tin and mercury [75]. For a
comprehensive review of some of the most recent sample preparation techniques
used for the analysis of metals in soils, sediments and plants including SPME,
the reader is referred to two reviews by Mester et al. [74] and Sturgeon [76].
Finally, some workers have adopted a sequential extraction technique to
enhance the recovery of metals. F6rstner [77] reported sequential application of
several steps in the determination of metals distribution in soil according to
their matrix interaction. These steps include a cation exchange with 1 M
ammonium acetate followed by extraction with a reducible phase with 0.1 M
hydroxyl amine-HCl at pH = 2, and extraction of oxidizable phases with 30%
hydroxy minerals.
909
is 30 mg/kg (dry weight basis) [90]. LC/MS analyses confirm the absence of any
HMX degradation in the plant tissue, emphasizing that the explosive HMX is
bioextracted rather than biodegraded.
The current sonication method, which involves sonication of plant tissue
with an organic solvent in a refrigerated bath for 16 h, may not be suitable for
the determination of short-lived intermediate products. As described above, to
reduce the amount of solvents and time required for analysis, researchers have
recently focused their attention on solventless sample preparation methods such
as SFE [92]. Furthermore, the low supercritical temperature of CO2 also makes
it possible to solubilize thermally labile non-volatile chemicals such as the
explosive HMX, allowing their extraction without thermal decomposition. How-
ever, the non-polar character of CO2 limits its affinity for the extraction of polar
molecules such as polynitroorganic explosives, but this difficulty can be over-
come by the addition of a suitable co-solvent or by choosing an alternate
supercritical fluid for extraction.
Plant tissue samples from either the leaves, stem or roots are freeze-dried
and mixed with sodium sulfate or inert sand (particle diameters less than 0.5
mm) to sufficiently fill the extraction cell (5 ml) for the SFE extractor. After
passing through the extraction cell, SC-CO2 is allowed to pass through a
restrictor wherein SC-CO2 is depressurized to allow CO 2 gas to escape leaving
behind the analytes in the collection vial. The extract is then analyzed for
explosives content with either HPLC/UV (254 nm) or GC connected with ECD or
MSD. Both sonication with acetonitrile and SC-CO 2 were found to be compatible
in their extraction efficiency of HMX from the plant tissue, although at higher
concentration (> 300 mg/kg), sonication is found to be slightly better [93].
Previously we demonstrated that SC-CO2 recovery of TNT from soil [14] is
compatible with that of EPA Method # 8330 [4] which requires sonication with
acetonitrile (Table 26.3). The extractability of explosives from soil by SC-CO 2
can be improved by adding a suitable co-solvent either directly to the sample
(static solvent addition) or to CO2 (dynamic addition) prior to extraction [14].
For instance, the relatively poor SC-CO2 recoveries of RDX, HMX and TNT (26,
0 and 70%, respectively) obtained from a contaminated dry sandy soil (100
mg/kg) are improved drastically (90, 70 and 87%, respectively) when the soil is
first treated with 5-15% v/v of acetonitrile [14].
Interestingly, the SC-CO2 recoveries of TNT amine derivatives 2-ADNT and
4-ADNT are found to be lower than that of TNT itself (Table 26.4). The
extracted amounts correlate with their Freundlich soil sorption isotherms [14].
Amino-TNT derivatives, frequently found with TNT as degradation products,
with higher water-soil partition coefficients, Kd, are poorly extracted [14].
Amines are basic compounds and are thus expected to interact strongly with soil
humic acids to produce irreversibly bound complexes. The efficiency of SC-CO2
extractability of energetic chemicals from soil is also found to depend on several
other instrumental parameters such as pressure, temperature and time of
extraction. In general, any increase in pressure or temperature that does not
910
TABLE 26.4
Extraction of explosives and related compounds from several soil matrices using either SC-CO 2
extraction or sonication with acetonitrile (EPA # 8330)
Soil source Analyte Method of extraction (mg/kg) %SC-CO,/EPA
SC-CO2* (RSD) EPA** (RSD)
Sand RDX 75.0 (5.4) 76.6 (4.0) 97.9
HMX 57.4 (6.5) 81.3 (3.3) 70.6
Burning pit RDX 3310 (7.5) 3620 (3.8) 91.4
HMX 1150 (6.9) 1710 (4.0) 67.3
Top soil TNT 205.3 (5.1) 220.5 (13.4) 93.1
Disposal site TNT 5841 (4.5) 5265 (5.2) 110
Fungi treated TNT 0.75 0.78 96.1
Contaminated soil 2-ADNT 3.31 6.33 52.3
4-ADNT 2.92 5.04 57.9
*Samples (2 g) were each mixed with acetonitrile (AcN) (10% v/w) before extraction with SC-CO 2
(300 atm, 280 ml/min gas, oven and restrictor temperatures were 60°C and 150°C, respectively).
MeOH (5% v/v) was used as a modifier in the extraction of the sample from the burning pit.
Extraction time was 30 min. Analytes were collected in acetonitrile (10 ml).
**Samples (2 g) were each extracted with acetonitrile (5 ml) using sonication for 16 h. Extracts
from both methods were analyzed by HPLC (254 nm).
911
fitted with a small test tube containing KOH (0.5 M) to trap liberated carbon
dioxide ( 4 CO,). Uniformly ring labeled [UL- 1 N] RDX is used to confirm the
identity of RDX breakdown products such as N2 , N2 0O,and other ring cleavage
products. Full details on biodegradation experiments can be found in Refs.
[94,95]. Degradation products in the soil-slurry mixture are either directly
analyzed with an SPME/GC-MS system or with a LC-MS. Figure 26.6 is a typical
LC/MS chromatogram of products detected during RDX biodegradation with
sludge. The chromatogram shows the three nitroso derivatives of RDX as
determined by their deprotonated molecular mass ions [M-H]. These ions are
detected at m/z 206, 190 and 174 Da, matching molecular mass formulas of
C,HNO,,
3 CH 6N6,04 and CH 6N0,,, respectively. The peaks are identified as the
mononitroso hexahydro-l-nitroso-3,5-dinitro-1,3,5-triazine (MNX), hexa-
hydro-1,3-dinitroso-5-nitro-1,3,5-triazine (DNX) and hexahydro-1,3,5-tri-
nitroso-1,3,5-triazine (TNX). Figure 26.6 also shows a second set of LC-MS
peaks at the beginning of the chromatogram with lower retention times (<4
min) with [M-H] appearing at m/z 135 Da. Using reference standards and
uniformly ring labeled [N]-RDX, the peak is identified as the ring cleavage
products methylenedinitramine (O2NNHCH2NHNO2) [94]. The observation of
the latter metabolites confirm the occurrence of ring cleavage which is a prereq-
uisite step prior to mineralization (liberation of CO,). A combined LC/MS and
SPME/GC-MS time course shows that none of the above RDX degradation
intermediate accumulates indefinitely. Instead, they all (bio)transform further
to eventually produce HCHO, CO 2 and nitrous oxide, N2 0. The formation of N2 0O
from RDX biodegradation is seemingly universal since both N2 0O(detected by
GC/ECD) and HCHO (detected by SPME/GC-MS) also can be generated during
biodegradation of both RDX under aerobic conditions using the fungus
Phanaerocheate chrysosporium [96]. Furthermore chemical reduction and
hydrolysis have been found to lead to N2 0 and HCHO. The universal transfor-
mation of the cyclic nitramine explosive to eventually produce nitrous oxide,
100
TNX
[M-H]:135
2.5
1~~ V
5.0 75
I
10.0
,MNX
~~~~I.
12.5
RDX
15.0 17.5 20.0
Retention time (min)
Fig. 26.6. Typical LC/MS chromatogram of products detected during RDX biodegradation with
sludge in a soil. The figure shows the initial metabolites MNX, DNX and TNX and the ring
cleavage product methylenedinitramine, which ultimately decomposes in water to produce
HCHO and N2 O.
912
HCHO and CO 2 under abiotic and biotic conditions could be exploited for the
development of chemical or biochemical sensor for on-site monitoring of cyclic
nitramine explosives in the environment.
913
(detected as its benzoate derivative 1-methylethylbenzoate). A time course study
for the formation and disappearance of these metabolites is used to construct a
degradation pathway which involves the formation of 2-OH-BP and 2,3-di-
OH-BP that later undergo ring cleavage, ultimately resulting in mineralization
(CO, formation) [104].
Acknowledgements
The author would like to thank the Department of National Defence, Canada for
their partial support on conducting the characterization part on energetic
chemicals. We thank Mme. C. Beaulieu and Mr. S. Deschamps for providing the
figures on TPHs and VOCs. We would like to thank Dr. An-Lac Nguyen for
revising the manuscript. Finally, we would like to thank the Strategic Environ-
mental Research and Development Program (SERDP # CU1213), USA, for
partial funding of the present work on RDX.
REFERENCES
1 T.F. Jenkins, M.E. Walsh, P.G. Thorne, P.H. Miyares, T.A. Ranny, C.L. Grant and
J.R. Esparza, Site characterization for explosives contamination at a military firing
range impact area, Special Report 98-9. US Army Corps of Engineers, CRREL,
Hanover, NH, 1998.
2 S. Thiboutot, G. Ampleman, T.F. Jenkins, C.L. Grant, P.G. Thee, MF. Walsh, T.A.
Ranney, J. Esparza and M.H. Stutz, Proceedings of the Third Tri-Service Environ-
mental Technology Workshop, San-Diego, CA, 18-20 August, 1998.
3 J. Crepin, and R.L. Johnson, in: M.R. Carter (Ed.), Soil Sampling and Methods of
Analysis. Lewis Publishers, Boca Raton, FL, 1993, pp. 5.
4 EPA SW-846 update 3rd edn., Methods 3051, 3540, 3541, 3545, 3550, 3560, 3561,
5021, 5032, 5035, 8330, 8275, Office of Solid Waste, Washington, DC, 1997.
5 C. Carlon, A. Critto, A. Marcomini and P. Nathanail, Environ. Pollut., 111 (2001)
417.
6 J.C. Pennington, J.M. Brannon, D. Gunnison, D.W. Harrelson, M. Zakikhani, P.
Miyares, T.F. Jenkins, J. Clarke, C. Hayes, D. Ringleberg, E. Perkins and H.
Fredrickson, Soil Sediment Contain., 10 (2001) 45.
7 T.F. Jenkins, C.L. Grant, G.S. Brar, P.G. Thorne, P.W Schumacher,. and T.A
Ranney, FieldAnal. Chem. Technol., 1 (1997) 151.
8 B. Richter, J. Ezzell and D. Felix, Single Laboratory Method Validation Report.
Extraction of TCL/PPL (Target Compound List/Priory Pollutant List) BNAs and
Pesticides using Accelerated Solvent Extraction (ASE) with Analytical Validation by
GC/MS and GC/ECD. Document 116064.A, Dionex Corporation, 1994.
9 B. Richter, J. Ezzell and D. Felix, Single Laboratory Method Validation Report.
Extraction of TCL/PPL (Target Compound List/Priory Pollutant List) Chlorinated
herbicides and PCBs using Accelerated Solvent Extraction (ASE). Document
101124.A, Dionex Corporation, 1994.
10 C.S. Hein, P.J. Marsden and A.S. Shurtleff, Evaluation of Methods 3540 (Soxhlet)
and 3550 (Sonication) for Evaluation of Appendix IX from Solid Samples. S-CUBEC,
Report for EPA, Work Assignment No 03, Doc. No. SSS-R-88-9436, 1988.
11 J.L. Snyder, R.L. Grob, M.E. McNally and T.S. Oostdyk, Anal. Chem., 64 (1992)
1940.
914
12 S.B. Hawthorne and D.J. Miller, Anal. Chem., 66 (1994) 4005.
13 Y. Liu, V. Lopez-Avila, M. Alcazar, W.F. Beckert and E.M. Heithmar. J. Chromatogr.
Sci., 31 (1993) 310.
14 J. Hawari, A. Halasz, L. Dusseault, J. Kumita, E. Zhou, L. Paquet, G. Ampleman and
S. Thiboutot, Proceedings of the 5th Meeting On Supercritical Fluids, Nice, France,
1998. pp. 161.
15 P. Flores and T. Bellar, Determination of Volatile Organic Compounds in Soils using
Equilibrium Headspace Analysis and Capillary Column Gas Chromatography/Mass
Spectrometry. U.S. Environmental Protection Agency, Office of Research and Devel-
opment, Cincinnati, OH, 1992.
16 M.H. Hiatt, Anal. Chem., 53 (1981) 1541.
17 J. Pawliszyn, Solid Phase Microextraction: Theory and Practice. Wiley-VCH, New
York, NY. 1997.
18 J.R. Dean and P. Hancock, in: J. Pawliszyn (Ed.), Application of Solid Phase
Microextraction.The Royal Society of Chemistry, Cambrige, UK, 1999. pp. 248.
19 J.M. Brannon, P. Deliman, C. Ruiz, C. Price, M. Qasim, J.A. Gerald, C. Hayes and S.
Yost, Conceptual model and process descriptor formulations for fate and transport of
UXO, Technical Report IRRP-99-1, U.S. Army Corps of Engineers, Waterways
experimental station, Vicksburg, MS, 1999.
20 S.D. Comfort, P.J. Shea, L.S. Hundal, Z. Li, B.L. Woodbury, J.L. Martin and W.L.
Powers, J. Eviron. Qual., 24 (1995) 1174.
21 H.M. Selim, S.K. Xue and I.K. Iskandar, Soil Sci., 160 (1995) 328.
22 J. Singh, S.D. Comfort and P.J. Shea, J. Environ. Qual., 27 (1998) 1240.
23 T.W. Sheremata and J. Hawari, Environ. Sci. Technol., 34 (2000) 3462.
24 HSDB, TNT. Hazardous substances data bank, National library of medicine,
National toxicology information program, Bethesda, MD, 1994.
25 R.J. Spanggord, W.R. Mabey, T.W. Chou and J.H. Smidth, in: D.E. Rickert (Ed.),
Chemical Industry Institute of Toxicology Series: Toxicity of Nitroaromatic Com-
pounds. Hemisphere, Washington, DC, 1985, pp. 15.
26 D.L. Kaplan, in: S.K. Sikdar and R.L. Irvine (Eds.), Bioremediation:Principlesand
Practice, Vol. II, Biodegradation Technology Developments. Technomic, Lancaster-
Basel, 1998, pp. 549.
27 G.B. Teany and R.T. Hudak, Proceedings of the 87th Annual Air and Waste Manage-
ment Association, paper # 94-RP 143.05, Cincinnati, Ohio, 1994.
28 G.B. Teany, R.T. Hudak and J.M. Melby, Proceedings of Field Screening Methods for
Hazardous Wastes and Toxic Chemicals, Las-Vegas, Nevada, 1995, p. 965.
29 P.R. Cauger, D.B. Holt, C.H. Patterson Jr., P.T. Charles, L. Shriver-Lake and A.W.
Kusterbeck, J. Hazard. Materials, 83 (2001) 51.
30 K.F. Myers, E.F. McCormick, A.B. Strong, P.G. Thorne and T.F. Jenkins, Com-
parison of Commercial Colorimetric and Enzyme Immunoassay Field Screening
Methods for TNT in Soil, Technical Report IRRP-94-4, Waterways Experiment
Station and U.S. Army CREEL, 1994.
31 J.A. Sundquist and S.S. Sisodia, Proceedings of 28 th Mid-Atlantic Industrial and
Hazardous Waste Conference, Vol. 28, 1996, p. 93.
32 Z. Mester, R. Sturgeon and J. Pawlyszin, Spectrochim. Acta B, 56 (2001) 233.
33 F. Soxhlet, Dingers Polytech. J., 232 (1879) 461
34 M.D. Erikson, Analytical Chemistry of PCBs. Butterworth, Stoneham, MA, 1986.
35 G.S. Douglas, K.J. McCarty, D.T. Dahlen, J.A. Seavey, W.G. Steihauer, R.C Prince
and D.L. Elmendorf, in: P.T. Kostecki and E.J. Calabrese (Eds.), ContaminatedSoils:
Diesel Fuel Contamination. Lewis Publishers, Chelsa, MI, 1992.
36 M.L. Lee, M.V. Novotny and K.D. Bartle, Analytical Chemistry of Polycyclic Aromatic
Compounds. Academic Press, New York, 1981, pp. 17.
915
37 Test Methods for Chemical Analysis of Water and Wastes Environmental Monitoring
and Support Lab., Cincinnati, Ohio, PB-297 686, Method 418.1, 413.2, 1978.
38 J.H. Martin, Jr., A.J. Siebert and R.C. Loehr, J. Environ. Eng., 117 (1991) 291.
39 L.H. Hilpert, W.E. May, S.A. Wise, S.N. Chesler and H.S. Hertz, Anal Chem., 50
(1987) 458.
40 APHA, 19th edn. Standard Methods 5520, 1995.
41 G.S. Douglas and A.D. Uhler, Environ. Test. Anal., 2 (1993) 46.
42 G.S. Douglas and J.S. Brown, Advanced Chemical Fingerprinting of Petroleum
Contaminated Soils and Water. 10 th Annual Conference on Contaminated Soils,
October 23-26, 1995.
43 Montreal protocol on Substances that Deplete the Ozone Layer, Montreal Protocol
Unit (MPU), ESDG/UNDP, New York, NY., Sep., 1987.
44 T.L. Chester, J.D. Pinkston and D.E. Raynie, Anal. Chem., 68 (1996) 487R.
45 J.A. Hawari, C. Beaulieu, D. Ouellette, A. Halasz and H. Van Tra, Int. J. Environ.
Anal. Chem., 60 (1995) 123.
46 J.L. Snyder, R.L. Grob, M.E. McNally and T.S. Oostdyk, J. Chromatogr. Sci., 31
(1993) 183.
47 M.T. Fahmy, M.E. Paulaitis, D.M. Johnson and M.E. McNally, Anal. Chem., 65
(1993) 1462.
48 J.W. Hills and H.H. Hill, J. Chromatogr. Sci., 31 (1993) 6.
49 J.J. Langenfeld, S.B. Hawthorne, D.J. Miller and J. Pawliszyn, Anal. Chem., 66
(1994) 909.
50 S.B. Hawthorne, in: J. Pawliszyn (Ed.), Sample Preparation in Field and Laboratory,
Elsevier, Amsterdam, 2002 (this book and references therein).
51 G. King, in: J. Pawliszyn (Editor), Sample Preparation in Field and Laboratory,
Elsevier, Amsterdam, 2002 (this book and references therein).
52 T.C. Voice and B. Kolb, Environ. Sci. Technol., 27 (1993) 709.
53 T. Urbanski, in: M. Jurecki (Trans.) and S. Laverton (Ed.), Chemistry and Tech-
nology of Explosives. Pergamon, Oxford, 1967, Vol. III, pp. 17.
54 R. Haas, I. Schreiber, E. v. Low and G. Stork, FreseniusJ. Anal. Chem., 338 (1990)
41.
55 A. Myler and W. Sisk, in: G.S. Sayler, R. Fox and J.W. Blackburn (Eds.), Bio-
remediation of Explosives Contaminated Soils (Scientific Questions/Engineering
Realities), Environmental Biotechnology for Waste Treatment. Plenum, New York,
1991, pp. 137.
56 M. Jackson, J.M. Green, R.L. Hash, D.C. Lindsten and A.F. Tatyrek, Nitramine
(RDX and HMX) wastewater treatment at the Holston Army Ammunition Plant,
Report ARLCD-77013, US Army Armament Research and Development Command,
Dover, NJ, 1978.
57 J. Yinon, Forensic and Environmental Detection of Explosives. John Wiley, Chich-
ester, UK, 1999.
58 J. Akhavan, The Chemistry of Explosives. The Royal Society of Chemistry,
Cambridge, UK, 1998.
59 J. Pawliszyn, in: R.M. Smith (Ed.), RSC Chromatography Monographs Series.
Loughborough University of Technology, The Royal Society of Chemistry, UK, 1999.
60 S.S.Yang, A.I. Goldsmith and I. Smetena, J. Chromatogr. A, 754 (1996) 3.
61 A. Di Corcia, A.B. Caracciolo, C. Crescenzi, G. Giuliano, S. Murtas and R. Samperi,
Environ. Sci. Technol., 33 (1999) 3271.
62 J.R. Dean, I.J. Barnabas and S.P. Owen, Analyst, 121 (1996) 465.
63 A.M. Robertson and J.N. Lester, Environ. Sci. Technol., 28 (1994) 346.
64 S. Paillound and W. Haerdi, Chromatographia,38 (1994) 514.
65 R. Alzaga, J.M. Bayona and D. Barcel6, J. Agric. Food Chem., 43 (1995) 395.
916
66 R. Alzaga, J.M. Bayona and D. Barcel6, J. High Resol. Chromatogr., 19 (1996) 23.
67 V. Lopez-Avila, N.S. Dodhiwala and W.F. Beckert, J. Agric. Food Chem., 41 (1993)
2038.
68 M.G. Weller, A. Zeck, P. Pfortner, E. Simon and R. Niessner, Int. J. Environ. Anal.
Chem., 75 (1999) 201.
69 V. Camel, Analyst, 126 (2001) 1182.
70 L. Yang, J.W.H. Lam, R.E. Sturgeon and J.W. McLaren, J. Anal. Atom. Spectrom., 13
(1998) 1245.
71 Y. Lin, R.D. Brauer, K.E. Laintz and C.M. Wai, Anal. Chem., 65 (1993) 2549.
72 J.M. Bayona and Y. Cai. TrAc, Trends Anal. Chem., 13 (1994) 327.
73 J.M. Bayona, Tech. Instrum. Anal. Chem., 17 (1995) 465
74 Z. Mester, R. Stergeon and J.W. Lam, J. Anal. At. Spectrom., 15 (2000) 1461.
75 Y. Morcillo, Y. Cai and J.M. Bayona, J. High Resol. Chromatogr., 18 (1995) 767.
76 R. Sturgeon, Commun. Soil Sci. Plant Anal., 31 (2000) 1479.
77 U. Forstner, Contaminated Sediments. Springer, Berlin, 1989. pp. 157.
78 R.B Meagher, Curr. Opin. PlantBiol., 3 (2000) 153.
79 A. Guivarch, P. Hinsinger and S. Staunton, PlantSoil, 211 (1999) 131.
80 W.H.O. Ernst, New Phytol., 146 (2000) 357.
81 K.V. Nedunuri, R.S. Govindaraju, M.K. Banks, A.P. Schwab and Z. Chen, J. Environ.
Eng., 126 (2000) 335.
82 S.P. Pradhan, J.R. Conrad, J.R. Paterek and V.J. Srivastava, J. Soil Contamin., 7
(1998) 467.
83 R.L. Brigmon, N.C. Bell, D.L. Freedman and C.J. Berry, J. Soil Contam., 7 (1998) 433.
84 J.G. Burken, J.V. Shanks and P.L. Thompson, in: J.C. Spain, J.B. Hughes and H.-J.
Knackmuss (Eds.), Biodegradation of Nitroaromatic Compounds and Explosives.
Lewis Publishers, Boca Raton, FL, 2000, pp. 239.
85 P.L. Thompson, L.A. Ramer and J.L Schnoor, Environ. Sci. Technol., 32 (1998) 975.
86 S.D. Harvey, R.J. Fellows, D.A. Cataldo and R.M. Bean, Environ. Toxicol. Chem., 10
(1991) 845.
87 P.L. Thompson, L.A. Ramer and J.L. Schnoor, Environ. Toxicol. Chem., 18 (1999)
279.
88 S.D. Harvey, R.J. Fellows, D.A. Cataldo and R.M.J. Bean, J. Chromatogr. A, 518
(1990) 361.
89 S.L. Larson, Ann. N.Y. Acad. Sci., 829 (1997) 195.
90 C. Groom, S. Beaudet, A. Halasz, L. Paquet and J.A. Hawari, J. Chromatogr.A, 909
(2001) 53.
91 S.L. Larson, A.B. Strong, S.L. Yost, B.L. Escalon and D. Parker, Tech. Report
IRRP-98-5 US Army Corps of Engineers, Washington, DC, 1998.
92 S.B. Hawthorne, Anal. Chem., 62 (1990) 633A.
93 A. Halasz, C. Groom, L. Paquet, C. Beaulieu, S. Deschamps, A. Corriveau, S.
Thiboutot, G. Ampleman and J. Hawari, J. Chromatography(2002), in press.
94 J. Hawari, A. Halasz, T. Sheremata, S. Beaudet, C. Groom, L. Paquet, C. Rhofir, G.
Ampleman and S. Thiboutot, Appl. Environ. Microbiol., 66 (2000) 2652.
95 C.F. Shen, S.R. Guiot, S. Thiboutot, G. Ampleman and J. Hawari, Biodegradation,8
(1998) 339.
96 T. Sheremata, A. Halasz, L. Paquet, S. Thiboutot, G. Ampleman and J. Hawari,
Environ. Sci. Technol., 35 (2001) 1037.
97 J. Poerschmann, Z. Zhang, F.-D. Kopinke and J. Pawliszyn, Anal. Chem., 69 (1997)
597.
98 F.-D. Kopinke, J. Porschmann and M. Remmler, Naturwissenschaften,82 (1995) 28.
99 C. Denis-Larose, D. Labbe, H. Bergeron, A.M. Jones, C.W. Greer, J. Hawari, M.J.
Grossman, B.M. Sankey and P.C.K Lau, Appl. Environ. Microbiol., 63 (1997) 2915.
917
100 T. MacPherson, C. Greer, E. Zhou, A.M. Jones, G.Wisse, P.C.K. Lau, B. Sankey, M.J
Grossman and J.A. Hawari, Environ. Sci. Technol., 32 (1998) 421.
101 E.S. Olson, D.C. Stanley and J.R. Gallagher, Energy Fuels, 7 (1993) 159.
102 M.K. Lee, J.D. Senius and M.J. Grossman, Appl. Environ. Microbiol., 61 (1995) 4362.
103 J.A. Hawari, A. Demeter and R. Samson, Environ. Sci. Tech., 26 (1992) 2022.
104 C. Rhofir and J. Hawari, J. Chromatogr.A, 873 (2000) 53.
918
Chapter27
27.1 INTRODUCTION
E x t ra io n
Solven c Acid Digestion Soxhlet extraction
o |gTLC
TLC ~ 1~Utrasonication
SFEASE
Fig. 27.1. Recent advances in sample preparation methods for common forensic evidence.
80
T ,'
60 I'
1~~~~~~~~.
. 40
L 20
.·-./". ,'"
u)
0
0 20 40 60 80
LLEISPEIGC (mg/l phenobarbital)
analysis by ICPMS [1] to On-Chip technologies for DNA analysis, the focus of
this chapter is on recently developed methods employing SPME including ignit-
able liquid residues analyses, explosive traces analyses, and the extraction and
identification of drugs and poisons from biological specimens, amongst other
forensic applications. A recent review of sample preparation and analysis meth-
ods has shown that SPME has proven to be an important sample preparation
technique for the analysis of forensic specimens due to the many advantages the
920
Odors I Water I Debris
]
Fig. 27.3. Some forensic applications employing SPME under various different modes.
technique offers when it is applied to these types of samples [2]. SPME allows for
multiple sampling, preservation of the sample and minimizes the risk of sample
contamination. It is often faster than traditional techniques and can be readily
automated. The lower detection limits generally afforded by SPME allow for
confirmation of positive samples that previously went undetected. An additional
benefit of SPME is the reduction of solvents in order to minimize the risk of ana-
lysts being exposed to toxic substances and also to save forensic science laborato-
ries money. This chapter describes the application of SPME for the analysis of
forensic specimens including ignitable liquid residues, often referred to as
accelerants, explosive traces, drugs and poisons from biological specimens and
other forensic applications. Recently developed SPME methods are also pre-
sented, including novel methods for the analysis of ignitable liquids on human
skin, odor signatures, and several drug applications such as the analysis of
free-fraction antipsychotic drugs levels and GHB identification without the need
for derivitization.
Figure 27.3 illustrates some applications of the SPME sampling and sample
preparation strategy within the forensic sciences. The major areas discussed in
this chapter include: drug analysis for toxicology; bulk drug analysis; trace anal-
ysis including high explosives and ignitable liquids; and sampling from various
matrices and a number of SPME sampling modes. A recent review paper [2] on
the application of SPME to forensic samples has also addressed this topic while
adding some new results by the authors.
921
investigators are trained to determine the possible point (location) of origin of
the fire and often collect the debris in that area in the hope that a small amount
of ignitable liquid residue has remained. The most widely used method for
sampling and pre-concentration is the multi-step process involving pre-
concentration of the volatile compounds in the debris onto an activated charcoal
strip. The investigators at the scene collect and package the debris into a new
metal paint can in such a way that the volatile species are contained within the
can. The can is transported to the laboratory for analysis some time later (the
time frame between collection and analysis may be several weeks). A small strip
of activated charcoal is placed in the heated headspace of the sealed can (at 80°C)
for 14 h (overnight) and the strip is then eluted of the analytes of interest with a
solvent such as carbon disulfide. This process is time- and labor-intensive and is
reported to detect approximately 0.1 i1 of an ignitable liquid residue from a
sample [3] of a gasoline spike into a 1 gal can when the typical extraction
volumes (50 A1)and GC-MS conditions are used.
The same year Pawliszyn et al. introduced SPME as a sampling and sample
preparation alternative [4], a coated wire adsorption technique was being
applied to the recovery of ignitable liquid residues by Tranthim-Fryer [5]. This
early technique involved the heated headspace (70-80°C) adsorption of ignitable
liquids onto carbon-coated aluminum or copper wire followed by n-pentane
elution with ultrasonic vibration [5]. The first report of SPME applied to the
recovery of ignitable liquid residues in 1994 [6] demonstrated improved sensitiv-
ity for the recovery of ignitable liquid residues with reduced analysis times and
the elimination of toxic solvents compared to the established activated charcoal
strip/solvent elution method [7]. The SPME analyses of gasoline and kerosene
has been compared to headspace, cold trap and solvent extraction methods and
shown to provide accurate information with less interference peaks [8].
Headspace SPME has also been applied to the analysis of flammable and
combustible substances in human body fluids [9,10]. A detailed study of the
recovery of gasoline residues confirmed the utility of the SPME technique
including lack of interference problems in the presence of wood or plastic
pyrolysis products and the ability of SPME to provide reproducible multiple
analyses from a single sample [11]. The recovery of ignitable liquids directly
from aqueous solvents has also been demonstrated, with SPME proving to be
more than an order of magnitude more sensitive than the conventional solvent
extraction method on 500 ppb preparations and allowed for positive identifica-
tion of diesel fuel in aqueous samples [121. SPME has also been used to identify
the presence of gasoline in a real arson-suspected fire debris sample, while
conventional methods lacked adequate sensitivity for identification (13).
SPME methods have been optimized for a variety of ignitable liquids and
conditions [14,15] and applied to the recovery and identification of ignitable
liquids from human skin [16]. The SPME method developed for recovery of
ignitable liquid residues from human skin used 100 m polydimethylsiloxane
(PDMS) fibers with gentle heating for 5 min followed by 10 min sampling from a
922
plastic bag shrouding the suspected hand. The recovery was found to be depend-
ent on the initial amount, environmental conditions, ignitable liquid type and
time since application [16]. Figure 27.4 shows chromatograms of a blank hand
(top) and hands contaminated with 10 Al of gasoline (middle) and diesel fuel
10000
', 8000
c 6000
4000
2000
0
ji
Innn
0 5 10 15 20 25 30 min.
14000
Gasoline residues on hand after 1.5 hours
12000
.10000
,Z
8000
6000
2000
-LUUU ······
_/ mSJ-<L.._YI!
0 10
bI 4Villll~\l
15 20 25
i i
30 min.
80000
pristane
_ 60000
40000 tetradecane
20000
0 10 20 30 40 50
Fig. 27.4. Headspace SPME/GC/FID of blank and hands contaminated with accelerants and
sampled for 15 min using headspace SPME.
923
(bottom) several hours after exposure. A proposed comprehensive SPME scheme
for detecting ignitable liquid residues in suspected arson cases used a low tem-
perature carboxen/PDMS fiber followed by an elevated temperature PDMS fiber
extraction demonstrating simplified sample preparation and high recovery effi-
ciency compared to headspace charcoal adsorption and solvent extraction
methods [17]. A review of contemporary sample preparation methods, including
SPME, for the detection of ignitable liquids in suspected arson cases was
recently published [18]. The E30 committee on Forensic Sciences of the Ameri-
can Society for Testing and Materials (ASTM) approved a standard method
entitled E 2154-Standard Practice for the Separation and Concentration of
Ignitable Liquid Residues from Fire Debris Samples by Passive Headspace
Concentration with Solid Phase Microextraction (SPME) in 2001 for publication
in the Annual Book of ASTM Standards, Vol 14.01 of 2002.
The detection and identification of very low levels of the organic explosives
commonly referred to as the high explosives can be a very important aspect of a
bombing investigation. Explosives detection before the detonation (pre-blast)
can provide information that could prevent a bombing and lead to the arrest of a
perpetrator and the identification of an explosives reside following a bombing
(post-blast) can provide investigative leads to solving the crime. The analysis of
organic explosives from soil found around storage facilities is also of importance
to environmental scientists as these compounds are monitored due to their
carcinogenic properties. The application of SPME to the analysis of these
compounds offers definite advantages over solvent extraction protocols.
Analysis of the semi-volatile explosive compounds, including nitrobenzene and
dinitrotoluenes in water has been reported using a PDMS fiber and GC/FID
analysis with detection limits reported at 9-15 ng/ml [19]. Headspace and direct
aqueous immersion SPME using PDMS/DVB fibers followed by GC/MS and
LC/UV were used to recover 14 explosives with varying results depending on the
extracted explosive [201. An optimized headspace SPME/GC/MS method has
been published using polyacrylate resin with 30-min adsorptions at 100°C and
desorptions at 200°C with reported detection limits of 0.5-10 ng/750 ml head-
space for EGDN, NG, PETN, TNT and RDX [21]. Trace explosive signatures
have been determined from World War II unexploded undersea ordinance using
direct immersion SPME/GC/MS and SPME/GC/Reversal Electron Attachment
Detection (READ) yielding improved extractions over solid phase extraction
with sensitivities of 10 parts per trillion for DNT and TNT for 15 min extractions
using PDMS/DVB fibers [22]. Direct immersion SPME using a Carbowax/DVB
fiber with 10 min sampling followed by a GC/ITMS analysis yielded limits of
detection of 10-325 ppt for TNT, RDX, and amino-DNTs in seawater [23]. Direct
immersion SPME/GC, SPME/LC and SPME/CZE methods have been optimized,
compared and successfully applied to actual post-explosion debris samples
924
Analytical Pump(2)
len-port Valve at "Sample Loading" positionn
tion
I)
I
Analytical Pump(2)
Ten-Port Valve at "Sampie lajectiol" Positio
[24,25]. Headspace SPME has also been applied to the characterization of odor
signatures, including explosive odor signatures [26]. The relative effects of major
controllable variables for the SPME recovery of explosives and ignitable liquid
residues have been reviewed including analyte chemistry, sampling mode
(direct, headspace, and partial headspace), fiber chemistry, adsorption times,
adsorption temperatures, desorption temperature, desorption time, and matrix
effects including water content relative to sample container size [27].
An improved HPLC interface (see Fig. 27.5) utilizing a 10 port valve and a
pre-focusing column was reported [27] to improve the resolution of the high
explosives analysis to the point where all fourteen [14] compounds in the EPA
8330 method standard mixture are readily resolved using a single solvent system
as the mobile phase. A comparison of the detection limits calculated for direct
injection vs SPME extraction for MECC/PDA, GC/ECD and LC/UV methods is
summarized in Table 27.1 [27].
925
TABLE 27.1
Detection limits (5 replicates, S/N > 3) of explosives by SPME/GC, SPME/HPLC and SPME/
MECC (measured at 254 nm, *220 nm; *'200 nm)
Explosives LOD (ug/ml) LOD (ng/ml)
MECC/PDA GC/ECD LC/UV (254 nm)
MECC SPME/ GC SPME/ GC HPLC SPME/
MECC HPLC
NB 0.74 (6.8) 0.69 (5.3) 53 (3.3) 0.24 (3.5) 3.9 (2.7) 1.2 (2.7)
1,3-DNB 0.41 (7.6) 0.38 (5.8) 44 (2.7) 0.22 (2.7) 2.3 (1.2) 0.8 (1.2)
1,3,5-TNB 0.62 (7.6) 0.62 (11.8) 39 (3.7) 0.18 (2.1) 2.6 (1.1) 0.6 (1.2)
2-NT 0.92 (4.3) 0.53 (11.7) 55 (3.2) ' 0.24 (3.4) 2 (3.7) 1.8 (3.8)
4-NT 1.12 (5.8) 0.51 (10.2) 47 (2.7) 0.24 (2.8) 7.4 (4.1) 1.9 (4.2)
3-NT 0.87 (4.3) 0.42 (11.8) 64 (2.7) 0.24 (2.9) 7.3 (4.0) 1.7 (4.0)
4-A-2,6-DNT 0.6 (7.4) 0.21 (4.0) 30 (1.7) 0.1 (1.6) 2.9 (1.6) 1.2 (1.6)
2-A-4,6-DNT 0.49 (9.2) 0.18 (3.8) 29 (1.4) 0.09 (1.4) 3.4 (2.1) 1.2 (19)
2,6-DNT 0.63 (6.3) 0.28 (5.3) 27 (1.9) 0.09 (1.5) 4.4 (3.0) 1.3 (2.9)
2,4-DNT 0.41 (7.5) 0.19 (3.4) 29 (1.5) 0.1 (1.7) 3.4 (1.9) 1.2 (1.7)
TNT 0.54 (7.8) 0.25 (4.2) 22 (1.7) 0.09 (1.7) 2.9 (1.2) 11 (1.3)
HMX 1.61 (6.6) 0.73 (9.8) - - 3.8 (2.6) 1.2 (2.7)
RDX 1.24 (7.4) 1.01 (17.4) 130 0.61 (5.2) 2.9 (2.3) 1.1 (2.5)
Tetryl 0.67 (6.9) 0.24 (4.1) 43 (4.0) 0.25 (4.2) 3.2 (2.8) 1.3 (2.9)
EGDN 2.96 (8.5)* 2.35 (18.2) 51 (3.7) 0.22 (3.8) 550 120 (4.6)*
NG 1.71 (4.5)* 0.37 (7.8)* 89 (2.8) 0.58 (2.9) 500 110 (4.8)*
PETN 1.89 (6.7)* 0.20 (6.4)* 94 (3.7) 0.61 (3.8) 380 80 (4.3)*
The extraction of poisons from biological specimens, including urine, serum and
blood, continues to be a difficult and time-consuming task. Increasingly, tradi-
tional liquid-liquid extraction techniques have been improved on and replaced
by newer methods including solid-phase extraction (SPE) and supercritical fluid
extraction (SFE) [28]. Most recently, SPME has emerged as a promising alterna-
tive for the rapid recovery of drugs from biological fluids such as urine or blood
and of poisons from biological matrices [14,29] with recent applications detailed
below.
A number of recent review papers have described the use of SPME within
toxicology applications including a review [30] that provides a good overview of
the theoretical and practical understanding of the use of microextraction tech-
nologies for drug analysis and three reviews [31-33] focusing on the extraction of
926
drugs from biological matrices. A fourth review [34] in the same year would also
cover the theory and application of SPME in biomedical analysis. Recent reports
of SPME-GC-MS for drug analysis [35], the use of SPME with HPLC and CE
[36] and specifically dealing with the analysis of drugs from hair using a
headspace sampling method [37] are further evidence of the wider acceptance of
SPME as a sampling tool in toxicology laboratories.
A large number of SPME applications to toxicology have involved the
recovery of drugs from urine or the recovery of drugs from blood and serum.
SPME/GC has been successfully applied to the recovery of methadone, benzo-
diazepines, cannabinoids, phencyclidine, methaqualone, amphetamines and
their metabolites from urine, but was not as successful for the recovery of
cocaine or opiates and their metabolites at pH 12 using 20 min direct immersion
SPME at 40°C with PDMS and polyacrylate fibers [38]. A direct immersion
SPME method for the detection of cocaine down to 6 ng/0.5 ml human urine has
been reported using a PDMS fiber with 30 min sampling from urine with added
NaF followed by GC/nitrogen-phosphorous detection (NPD) [39]. A SPME
method has been developed for the determination of benzoylecgonine in urine
with hexyl chloroformate derivitization using a 100 Aum PDMS fiber for 10 min at
55°C and ion trap MS detection [40].
The determination of amphetamine, methamphetamine (MA) and dimeth-
amphetamine in urine has been performed by 100 Atm PDMS SPME/GC/MS
sampled for 30 min at pH 12.4 with 30% salt added [41]. Methamphetamine and
amphetamine were detected down to ca. 10 ng/ml using PDMS/DVB fibers to
extract the drugs from urine treated with 1 g/ml sodium carbonate (Na 2CO3 ) for
30 min at 65°C followed by GC/NPD analysis [42]. Another method for the
recovery of methamphetamine and amphetamine used 100 Am PDMS fibers
from urine treated with NaCl (0.5 g/ml) adjusted to pH 12 and extracted for 20
min followed by a NaOH-H3 BO4 buffer wash prior to GC analysis with reported
sensitivity 10-100 times greater than headspace methods including SPME [43].
Amphetamines and inflammable compounds in urine and blood have been
analyzed using a 100 Am PDMS fiber immersed in basic solutions for 5 min at
80°C with analysis by GC/MS [44]. The automated determination of amphet-
amines and ecstasy in urine has been reported using 100 m PDMS fibers
immersed at pH 10 for 16 min followed by analysis by GC/NPD or GC/MS [45].
Although sometimes less sensitive, depending on the volatility of the drug,
headspace SPME has the advantage of yielding cleaner extracts from biological
samples. A 5 min headspace SPME/GC/CI-MS procedure with 100 m PDMS
fibers for amphetamines from urine containing potassium carbonate at 80°C was
shown to be 20 times more sensitive than the conventional headspace method
[46]. Another headspace SPME/GC/MS procedure using a 100 Am PDMS fibers
extraction for 15 min over a solution containing NaCl at 75°C was shown to be
sensitive enough for routine confirmation of positive EMIT and RIA results for
amphetamine and its analogs (MA, MDMA, MDEA) [47]. A headspace SPME/
GC/FID method has also been reported for the analysis of amphetamines in clini-
927
cal urine samples using 100 um PDMS fibers for 15 min at 60°C with further
method development and derivatization suggested for conclusive differentiation
between the various types of amphetamines [48]. A screening procedure for 21
amphetamine related compounds has been developed using 100 m PDMS
headspace SPME/GC/MS using a 10 min extraction at 80°C with salt added [49].
Finally, the rapid analysis of amphetamine, MA, methylenedioxyethylamphet-
amine (MDA) and methylenedioxymethamphetamine (MDMA) has been report-
ed using 100 Am PDMS headspace SPME/GC/MS using a 10 min extraction at
100°C with 2 N NaOH followed by direct on-fiber derivatization for 20 min at
60°C using trifluoroacetic anhydride [50].
Methadone has been quantified in urine using 100 Am PDMS fiber for 15 min
at pH 7.7 followed by GC/MS quantitation [51]. Nicotine and cotinine in urine
have been analyzed down to 5 and 300 ng/ml, respectively, using 100 Am PDMS
headspace SPME/GC/MS using a 5 min extraction at 80°C with the addition of
potassium carbonate [52]. Tricyclic antidepressants in urine have been analyzed
down to 24-38 ng/ml using 100 Am PDMS headspace SPME/GC-FID using a 15
min extraction at 100°C with the addition of sodium hydroxide [53]. Meperidine
in urine and blood have been analyzed down to 20 and 100 ng/ml, respectively,
using 100 pm PDMS headspace SPME/GC-FID using a 30 min extraction at
100°C with the addition of sodium hydroxide and sodium chloride [54]. There
have recently been several reports of the determination of benzodiazepines and
metabolites using direct immersion methods. A method using 65 am PDMS/DVB
sampling for 30 min followed by GC-FID analysis was applied to the detection of
benzodiazepines in urine [55]. Another method for determining benzodiazepines
in urine used 65 gm Carbowax/DVB sampling for 60 min at 45°C with added salt
followed by GC-FID and GC-MS analysis [56]. The analysis of benzophenones
for the detection of benzodiazepines in urine has been reported using 100 Am
PDMS SPME/GC/ECD using a 30 min extraction at pH 9.4 with the addition of
potassium hydroxide [57]. The analysis of benzophenone and metabolites in
water and human urine has been reported using 65 Am Carbowax/DVB SPME/
GC/M$ with a 45 min immersion time [58]. Barbiturates have been determined
in urine using 65 pm Carbowax/DVB SPME/GC/MS using a 20 min immersion
time [59].
Non-routine volatiles including methylene chloride and petroleum products
were confirmed in urine and in a gastric sample using 100 Am PDMS headspace
SPME/GC/MS using a 10 min extraction at 60°C with the addition of sodium
chloride [60]. The absence of an air peak in the GC/MS afforded by the SPME
procedure offered a tremendous advantage in the identification of the unknown
volatiles, which proved to be crucial evidence in the investigation of traffic
fatalities [61]. Headspace SPME using 100 am PDMS fibers for 15 min at 60°C
has been used to determine benzene, toluene, ethylbenzene and xylenes (BTEX)
in urine with GC/MS [62]. An immersion method using 100 /m PDMS fibers for
5 min with added NaCl has been used to analyze urinary toluene and xylene from
workers using organic solvents with GC/FID [63]. The analysis of chlorophenols
928
in urine has been accomplished using 85 m polyacrylate SPME/GC/MS using a
50 min extraction at pH 1 [63]. Another method for the analysis of chlorophenols
in urine of subjects exposed to chlorobenzene used 85 gm polyacrylate SPME/
GC/MS using a 30 min extraction with added salt [64].
The free concentration of valproic acid in human plasma has been determined at
1000 ng/ml using 100 m PDMS direct immersion SPME/GC-FID using a 3 min
extraction from a sample previously dialyzed for 25 min and adjusted to pH 2.5
[65]. A calibration curve for valproic acid recovered from buffered standards and
plasma samples centrifugally purified to remove proteins followed by direct
immersion SPME/GC-MS with a 65 gm Carbowax/DVB fiber for 10 min is shown
in Fig. 27.6. A 7.5 gg/ml caprylic acid internal standard was added to plasma
solutions containing valproic acid (1.25-12.5 g/ml) at pH 6.86. Two milliliters of
each plasma sample was pipetted into a 2 ml Centricon YM-30 cell, and then
centrifuged 30 min to get 1 ml filtrate followed by SPME and desorption into the
GC inlet for 5 min at 250°C. Figure 27.6 illustrates the excellent linearity of this
SPME method for quantifying the free fraction concentration of this drug.
Antidepressant drugs and metabolites in human plasma have been analyzed
down to 90-200 ng/ml using 100 m PDMS headspace SPME/GC-NPD using a
10 min extraction at 22°C with high protein binding appearing to be the limiting
mechanism for better extractions [66]. Four tricyclic antidepressants in blood
have been analyzed down to 61-2000 ng/ml using 100 gm PDMS headspace
SPME/GC-FID using a 60 min extraction at 100°C with the addition of sodium
hydroxide [67]. The analysis of antidepressants in blood and urine have been
analyzed, including an analysis in a fatal intoxication case using 100 gm PDMS
headspace SPME/GC-MS using a 45 min extraction at 120°C with the addition of
sodium hydroxide [68,69]. Methamphetamine and amphetamine in blood have
been analyzed down to 10 ng/ml using 100 gm PDMS headspace SPME/GC/MS
using a 5 min extraction at 80°C with the addition of sodium hydroxide [70].
Phencyclidine has been recovered from whole blood and urine analyzed down to
1.0 ng/ml and 0.25 ng/ml, respectively, using 100 gm PDMS headspace SPME/
GC/surface ionization detection (SID) using a 30 min extraction at 90°C with the
addition of sodium hydroxide and K2 C0 [71]. Methadone and metabolites have
been recovered from plasma via 100 gm PDMS direct immersion SPME/GC-MS
using a 30 min extraction at a pH 7.7 [72]. Diphenylmethane in blood and urine
has been analyzed by 100 gm PDMS headspace SPME/GC-FID using a 10 min
extraction at 98°C with 10 N NaOH added [73]. Diazepam in human plasma has
been determined using 85 gm polyacrylate direct immersion SPME/GC-FPD
using a 4 min extraction at pH 5.5 with 1-octanol added [74]. The authors
applied a method using 100 gm PDMS SPME/GC/MS with 5 min immersion at
25°C to confirm a suspected drink tampering case with the equivalent of ten
tablets containing 2 mg of diazepam each added to a whisky bottle.
929
1
q
0.9
0.8
0.7
0.6
'. 0.5
v 0.4
o 0.3
X 0.2
0.1
0
-0.1 0 2000 4000 6000 8000 10000 12000 14000
Fig. 27.6. Aqueous calibration curve for valproic acid in a pH 6.8 buffer using caprylic acid as an
internal standard analyzed by SPME/GC-MS.
Steroids have been analyzed from human serum using 85 tLm polyacrylate
direct immersion SPME/GC/MS using a 30 min extraction followed by headspace
derivitization with BSTFA [75,76]. Caffeine metabolites have been recovered
from blood and urine via 65 Am carbowax/DVB SPME/GC-NPD using a 60 min
extraction at 40°C with added acid [77]. SPME has been successfully applied to
the detection of ethanol in human body fluids, including urine and blood, down
to 10 and 20 jLg/ml, respectively, using 65 A8m Carbowax/DVB headspace SPME/
GC-FID using a 15 min extraction at 70°C with the addition of (NH4 )2SO4 [78].
An automated headspace SPME method for the analysis of blood alcohol concen-
tration using 65 Am Carbowax/divinylbenzene headspace SPME/GC-FID using 3
min exposures showed excellent precision and linearity with results in close
agreement with the conventional static headspace method [79]. An improved
method for the recovery of ethanol from blood and urine employed a 75 Atm
carboxen/PDMS headspace SPME/GC/FID using a 15 min extraction at 60°C
with the addition of (NH4 )2SO4 and sodium dithionite, improving sensitivity by
one to three orders of magnitude compared to previous methods [80]. The analy-
sis of methanol in whole blood has also been reported using headspace SPME/
GC-FID with a carboxen/PDMS fiber using 10 min exposures at 60°C [81]. An
example of SPME analysis of a blood alcohol casework sample analyzed by 65 m
Carbowax/DVB headspace automated SPME/GC-MS using 3 min headspace
adsorption at 25°C.
Although the most common toxicological analysis involve drug extraction and
recovery, drugs actually represent only ca. 60% of the more than 1500 possible
poisons [41]. The second largest class of poisons are the pesticides representing
ca. 30% of possible compounds followed by anions, metals, gases and volatiles,
930
which comprise a total of less than 10% of possible poisons [41]. Organo-
phosphate pesticides in blood and urine have been analyzed down to 5-80 ng/ml
and 2-24 ng/ml, respectively, using 100 Am PDMS headspace SPME/GC/NPD
using a 20 min extraction at 100°C with the addition of HC1 only for blood and
HC1, NaCl and (NH 4 )2 SO4 for urine [82]. The common organophosphorous
pesticide, malathion, in blood has been detected down to 1000 ng/ml using 100
tm PDMS headspace SPME/GC/MS using a 5 min extraction at 90°C with the
addition of (NH 4 )2 SO4 and H2 SO4 [83]. Six carbamate pesticides in blood and
urine have been analyzed down to 100-500 ng/ml and 10-50 ng/ml, respectively,
using 100 tm PDMS headspace SPME/GC/NPD using a 30 min extraction at
70°C with the addition of NaC1 [84]. Phenothiazines have been detected using
100 um PDMS headspace SPME/GC-FID with a 40 min extraction at 140°C with
added NaC1 [85]. Dinitroaniline herbicides have been analyzed using 100 tm
PDMS SPME/GC-ECD with 30 min extractions at 70°C for water and 90°C for
blood [86]. The analysis of parathion in blood has also been reported using
headspace SPME/GC-MS [87]. The natural insecticide, nereistoxin, has been
analyzed using 65 gm PDMS/DVB headspace SPME/GC-MS with a 30 min
extraction at 70°C with added NaCI [88].
A headspace SPME/GC/MS method has been developed for the analysis of
toluene, xylenes and hydrocarbons at 100-1000 ng/ml in human blood and
applied to the medico-legal autopsy of a fire victim with kerosene substances
detected [89]. The analysis of thinner components in whole blood and urine
down to 1 ng/ml has been performed using 100 Am PDMS headspace SPME/
GC-FID using a 5 min extraction at 80°C [90]. Cresol and phenols have been
detected using 85 Am polyacrylate headspace SPME/GC-FID with a 30 min
extraction at 100°C with added NaC1 [91]. Benzene and toluene in human blood
have also been monitored using 65 m carboxen/PDMS headspace SPME/
GC-MS with a 30 min extraction at 20°C [92]. Headspace sampling of BTEX in
blood and urine of cyclists exposed to air pollutants employed 75 Am carboxen/
PDMS SPME/GC-MS with a 30 min extraction at 50°C [93]. Tetrachloroethylene
and trichloroethylene have been analyzed in serum tissue and urine in a fatality
using 100 m PDMS headspace SPME/GC-ECD with a 1 min extraction at 60°C
[94].
Local anesthetics have been analyzed down to 58-830 ng/ml using 100 gm
PDMS headspace SPME/GC-FID using a 40 min extraction at 100°C from
human blood deproteinized with perchloric acid and sodium hydroxide and
(NH4 )2SO4 added [95]. The recovery of local anesthetics from human blood has
also been accomplished with a direct immersion using 100 Mm PDMS SPME/
GC-FID with a 40 min extraction [96]. Another recent method for the recovery of
local anesthetics employed 100 Mm PDMS headspace SPME/GC-MS using sel-
ected ion monitoring (SIM) with a 45 min extraction at 120°C with added NaOH
[97]. The analysis of the highly toxic methylmercury in aqueous and tissue
samples has been reported using 100 Am PDMS headspace SPME/GC-atomic
fluorescence spectrometry (AFS) using simultaneous extraction/derivitization
931
with sodium tetraethylborate at 25°C and pH 4.5 with an acetate buffer [98]. The
simultaneous determination of mercury, alkylmercury, lead and tin in human
body fluids was accomplished using 100 Am PDMS headspace SPME/GC-MS-MS
also using derivitization with sodium tetraethylborate for 10 min at pH 5.3 [99].
Methylmercury in biological samples has also been reported using headspace
SPME/GC-atomic absorption spectrometry (AAS) using hydride derivitization
[100]. Cyanide in blood has been analyzed using 65 tkm carbowax/DVB head-
space SPME/GC-NPD using a 45 min extraction at 50°C with Na2 SO4 added
[101]. Lead in blood and urine has been analyzed below 10 ppb using 65 ~m
PDMS/DVB headspace SPME/GC-FID using a 15 min extraction/derivitization
time with sodium tetraethylborate at 25°C and pH 4.0 with an acetate buffer
[102].
A modified SPME device has been developed which is capable of detecting down
to 5.8 nmol/l ethanol, 1.8 nmol/l acetone and 0.3 nmol/l isoprene in human
breath using a 65 m PDMS/DVB fiber and 1 min sampling followed by GC/MS
analysis [103]. SPME has been shown to be a useful tool to determine phospho-
lipid/water partition coefficients and free (bioavailable) concentrations in vitro
systems, which may make experimental data more meaningful for quantitative
in vivo extrapolations [104]. SPME has been used to determine residual organic
solvents in pharmaceutical samples with an optimized method using 65 Jum
PDMS/DVB with 30 min headspace extraction with NaCl added [105]. A method
for profiling confiscated ecstasy and amphetamine uses 65 tm PDMS/DVB
headspace SPME/GC-NPD with a 30 min extraction at 90°C with a pH of 5.0
using an acetate buffer [106]. SPME has also been used to characterize street
narcotic odors. The room-temperature recoveries of cocaine odor chemicals,
including methyl benzoate, required optimized headspace extraction times of up
to 12 h with the optimum fiber found to be 65/ m carbowax/DVB [107]. Although
it has been established that most US currency in circulation is contaminated
with gg quantities of cocaine, the corresponding odor chemical (methyl benzo-
ate) levels are shown to be well below the threshold levels of drug detector dogs
[107,108]. Recently, odor signature chemicals for other controlled substances
have been identified using SPME/GC/MS including street MDMA (Ecstasy)
samples as shown in Fig. 27.7.
A method for determining cannabinoids in water and saliva used 75 Aum
carboxen/PDMS direct immersion SPME/GC-MS with a 10 min extraction [109].
SPME has also been applied to the analysis of cannabinoids and amphetamines
in hair [110,111] and has been used to characterize items including cinnamon,
vodka, coffee and tobacco [112-115] which could prove useful investigative infor-
mation. Recently, a fast and simple SPME method has been developed for the
identification of the y-hydroxybutyric acid (GHB) from aqueous samples [116].
The direct injection of a GHB extraction into a GC port under normal tempera-
932
I
6500000 1. Methanphetamie
6000000 2. Piperonal
2 4
5500000 3. 3,4 methyedioxyphenyl
5000000 acetone (MD-P2P) I
4500000 4. 1-(3,4 methylenedioxyphenyl-2-
. 4000000 propanol)
2 3500000 5. Bylated Hydroxytoluene
3000000
(B3Hl)
2500000
5
2000000
1500000
1000000
500000
1 - II .I ,
2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00 22.00 24.00 26.00 28.00 30.00
Time
Fig. 27.7. Odor compounds commonly identified in MDMA tablets with PDMS fiber sampled for
3 h.
One obvious advantage of the SPME sampling and sample concentration tech-
nique is the adaptability of the technique for field sampling. The apparatus is
portable and amenable to remote sampling schemes. Possible future applica-
tions include the collection of explosives at the scene of a post-blast bombing
scene and the scene of a suspected arson. One of the major drawbacks of current
sampling schemes for these types of evidence is the time lag between the
occurrence of the event and the analysis of the analytes of interest. The advent of
portable analytical instrumentation such as GC/MS systems could resolve some
of the problems associated with this time lag including the deterioration of
sample, possible contamination and evidence integrity. A remote sampling
scheme involving SPME with subsequent analysis would improve the quality of
these types of evidence.
933
1(00% 42
75%
87
0%-
25%
56
0%
73
00%
75%
50%
25% 45 89
59 . 105
0%
50 75 160o 125 150
m/z
Fig. 27.8. GC/EIMS of y-hydroxybutyric acid (GHB) (top) yielding an inconclusive identification
of this drug compared to the SPME/GC/CIMS (bottom) providing improved fragmentation and a
molecular ion + 1 peak.
27.7 CONCLUSIONS
REFERENCES
934
3 ASTM E 1412-00, 2001 Annual Book of ASTM Standards, 14.02, (2001) pp. 431-433.
4 C.L. Arthur and J. Pawliszyn, Anal. Chem., 62 (1990) 2145.
5 D.J. Tranthim-Fryer, J. Forensic Sci., 35 (1990) 271.
6 J.R. Almirall, K.G. Furton and J.C. Bruna, Southern Association of Forensic
Scientists Fall 1994 Meeting, Orlando, Florida, September 7-10, 1994.
7 K.G. Furton, J.R. Almirall and J. Bruna, J. High. Resolut. Chromatogr., 18 (1995) 625.
8 T. Kaneko and M. Nakada, Reports of the National Research Institute of Police
Science: Research on ForensicScience, 48 (1995) 107.
9 Y. Iwasaki, M. Yashiki, N. Nagasawa, T. Miyazaki and T. Kojima, Jpn. J. Forensic
Toxicol., 13 (1995) 189.
10 X.P. Lee, T. Kumazawa and K. Sato, Int. J. Legal. Med., 107 (1995) 310.
11 K.G. Furton, J.R. Almirall and J. Bruna, J. ForensicSci., 41 (1996) 12.
12 J.R. Almirall, J. Bruna and K.G. Furton, Sci. Justice, 36 (1996) 283.
13 Steffen and J. Pawliszyn, Anal. Communications, 33 (1996) 129.
14 J.R. Almirall and K.G. Furton, in: S.A. Wercinski (Ed.), Applications in Solid Phase
Microextraction: A Practical Guide. Marcel Dekker, New York, 1999, Ch. 7, pp.
203-216.
15 K. Higgins (Ed.), Proceedings of the SPIE Conference on Investigation and Forensic
Science Technologies, Boston, 1998, 3576, pp. 136-141.
16 J.R. Almirall, J. Wang, K. Lothridge and K.G. Furton, J. ForensicSci., 45 (2000) 461.
17 Q. Ren and W. Bertsch, J. ForensicSci., 44 (1999) 504.
18 W. Bertsch and Q. Ren, ForensicSci. Rev., 11 (1999) 141.
19 J.Y. Horng and S.D. Huang, J. Chromatogr.,678 (1994) 313.
20 M. Bi, M.S. thesis, Florida International University, Miami, FL, 1998.
21 K.P. Kirkbride, G. Klass and P.E. Pigou, J. ForensicSci., 43 (1998) 76.
22 M.R. Darrach, A. Chutjian and G.A. Plett, Environ. Sci. Technol., 32 (1998) 1354.
23 S.A. Barshick and W.H. Griest, Anal. Chem., 70 (1998) 3015.
24 K. Higgins (Ed.), Proceedings of the SPIE Conference on Investigation and Forensic
Science Technologies, Boston, 1998, 3576, pp. 18-23.
25 K.G. Furton, L. Wu and J.R. Almirall, J. Forensic Sci., 45 (2000) 845.
26 K.G. Furton and L.J. Myers, Talanta, 54 (2001) 487.
27 K.G. Furton, J.R. Almirall, M. Bi, J. Wang and L. Wu, J. Chromatogr.A, 885 (2000)
419.
28 K.G. Furton and J. Rein, Anal. Chim. Acta, 236 (1990) 99.
29 L. Junting, C. Peng and 0. Suzuki., ForensicSci. Int., 97 (1998) 93.
30 H. Lord and J. Pawliszyn, J. Chromatogr.A, 902 (2000) 17.
31 N.H. Snow, J. Chromatogr.A, 885 (2000) 445.
32 G.A. Mills and V. Walker, J. Chromatogr.A, 902 (2000) 267.
33 G. Theodoridis, E.H.M. Koster and G.J. de Jong, J. Chromatogr. B, 745 (2000) 49.
34 S. Ulrich, J. Chromatogr.A, 902 (2000) 167.
35 U. Staerk and W.R. Kiilpmann, J. Chromatogr. B, 745 (2000) 399.
36 K.E. Rasmussen, S. Pederson-Bjergaard, M. Krogh, H.G. Ugland and T. Gronhaug,
J. Chromatogr.A, 873 (2000) 3.
37 F. Sporkert and F. Pragst, ForensicSci. Int., 107 (2000) 129.
38 K. Singer, B. Wenz, V. Seefeld and U. Speer, Labor-Med. (German), 18 (1995) 112.
39 T. Kumazawa, K. Watanabe, K. Sato, H. Seno, A. Ishii and 0. Suzuki, Jpn. J.
Forensic Toxicol., 13(3) (1995) 207.
40 B.J. Hall, A.R. Parikh and J.S. Brodbelt, J. Forensic Sci., 44 (1999) 527.
41 S.W. Myung, H.K. Min, S. Kim, M. Kim, J.B. Cho and T.J. Kim, J. Chromatogr., 716
(1998) 359.
42 A. Ishii, H. Seno, T. Kumazawa, M. Nishikawa, K. Watanabe, H. Hattori and 0.
Suzuki, Jpn. J. Forensic Toxicol., 14 (1996) 2228.
935
43 K. Ameno, C. Fuke, S. Ameno, H. Kinoshita and I. Ijiri, Can. Soc. Forens. Sci. J., 29
(1996) 43.
44 M.Yashiki, T. Kojima and T. Miyazaki, Jpn. J. Forensic Toxicol., 12 (1994) 120.
45 H. Ugland, M. Kogh and K. Rasmussen, J. Pharm. Bomed. Anal., 19 (1999) 463.
46 M. Yashiki, T. Kojima, T. Miyazaki, N. Nagasawa, Y. Iwasaki and K. Hara, Forensic
Sci. Int., 76 (1995) 169.
47 F. Centini, A. Masti and I.B. Comparini, Forensic Sci. Int., 83 (1996) 161.
48 H. Lord and J. Pawliszyn, Anal. Chem., 69 (1997) 3899.
49 C. Battu, P. Marquet, A. Fauconnet, E. Lacassie and G. Lachatre, J. Chromatogr.
Sci., 36 (1998) 1.
50 C. Jurado, M. Gimenez, T. Soriano, M. Menedez and M. Repetto, J. Anal. Toxicol., 24
(2000) 11.
51 M. Chiarotti and R. Marsili, J. Microcolumn, 6 (1994) 577.
52 M. Yashiki, N. Nagasawa, T. Kojima, T. Miyazaki and Y. Iwasaki, Jpn. J. Forensic
Toxicol., 13 (1995) 1724.
53 T. Kumazawa, X.P. Lee, M.C. Tsai, H. Seno, A. Ishi and K. Sato, Jpn. J. Forensic
Toxicol., 13 (1995) 25.
54 H. Seno, T. Kumazawa, A. Ishii, M. Nishikawa, H. Hattori and 0. Suzuki, Jpn. J.
Forensic Toxicol., 13 (1995) 211.
55 H. Seno, T. Kumazawa, A. Ishii, K. Watanbe, H. Hattori and 0. Suzuki, Jpn. J.
ForensicToxicol., 15 (1997) 16.
56 Y. Luo, L. Pan and J. Pawliszyn, J. Microcolumn, 10 (1998) 193.
57 F. Prosen, H. Seno, A. Ishii, K. Watanabe, T. Kumazawa, H. Hattori and 0. Suzuki,
J. Anal. Toxicol., 23 (1999) 54.
58 T. Felix, B. Hall and J. Brodbelt, Anal. Chim. Acta, 371 (1998) 195.
59 B. Hall and J. Brodbelt, J. Chromatogr.A, 777 (1997) 275.
60 W.E. Brewer, R.C. Galipo, S.L. Morgan and K.H. Habben, J. Anal. Toxicol., 21 (1997)
286.
61 S. Fustinoni, R. Giampiccolo, S. Pulvirenti, M. Buratti and A. Colombi, J.
Chromatogr. B, 723 (1999) 105.
62 F. Asakawa, F. Jitsunari, J. Choi, S. Suna, N. Takeda and T. Kitamado, Bull.
Environ. Contamin. Toxicol., 62 (1999) 109.
63 M. Lee, Y. Yeh, W. Hsiang and C. Chen, J. Chromatogr. B, 707 (1998) 91.
64 M. Guidotti, G. Ravaioli and M. Vitali, J. High Resolut. Chromatogr., 22 (1999) 427.
65 M. Krogh, K. Johansen, F. Tonnesen and K.E. Rasmussen, J. Chromatogr.B, 673
(1995) 299.
66 S. Ulrich and J. Martens, J. Chromatogr.B, 696 (1997) 217.
67 X.P. Lee, T. Kumazawa, K. Sato and 0. Suzuki, J. Chromatogr.Sci., 35 (1997) 302.
68 T. Watanabe, A. Namera, M. Yashiki, Y. Iwasaki and T. Kojima, Jpn. J. Legal. Med.,
52 (1998) 69.
69 A. Namera, T. Watanabe, M. Yahiki, Y. Iwasaki and T. Kojima, J. Anal. Toxicol., 22
(1998) 396.
70 N. Nagasawa, M. Yashiki, Y. Iwasaki, K. Hara and T. Kojima, Forensic Sci. Int., 78
(1996) 95.
71 A. Ishii, T, Kumazawa, K. Watanabe, H. Hattori, O. Suzuki, Chromatographia,43
(1996) 331.
72 A. Bermejo, R. Seara, A. dos Santos Lucas, M. Tabernero, P. Fernandez and R.
Marsili, J. Anal. Toxicol., 24 (2000) 66.
73 M. Nishikawa, H. Seno, A. Ishii, O. Suzuki, T. Kumazawa, K. Watanabe and H.
Hattori, J. Chromatogr. Sci., 35 (1997) 275.
74 M. Krough, H. Grefslie and K. Rasmussen, J. Chromatogr.B 689 (1997) 357.
75 P. Okeyo, S. Rentz and N. Snow, J. High Res. Chromatogr., 20 (1997) 171.
936
76 P. Okeyo and N. Snow, J. Microcolumn Sep., 10 (1998) 551.
77 T. Kumazawa, H. Seno, X. Lee and A. Ishii, Anal. Chim. Acta, 387 (1999) 53.
78 T. Kumazawa, H.Seno, X.P Lee, A. Ishii, O. Suzuki and K. Sato, Chromatographia,
43 (1996) 393.
79 Z. Penton, Can. Soc. Forens. Sci. J., 30 (1997) 7.
80 Z. Lee, T. Kumazawa, K. Sato, H. Seno, A. Ishii and 0. Suzuki, Chromatographia,47
(1998) 593.
81 X. Lee, T. Kumazawa, T. Jurosawa, K. Akiya, Y. Akiya, S. Fruta and K. Sato, Jpn. J.
Forensic Toxicol., 16 (1998) 64.
82 X.P Lee, T. Kumazawa, K. Sato and 0. Suzuki, Chromatographia,42 (1996) 135.
83 A. Namera, M. Yashiki, N. Nagasawa, Y. Iwasaki and T. Kojima, ForensicSci. Int., 88
(1997) 125.
84 H. Seno, T. Kumazawa, A. Ishii, M. Nishika, K. Watanabe, H. Hattori and 0. Suzuki,
Jpn. J. ForensicToxicol., 14 (1996) 199.
85 H. Seno, T. Kumazawa, A. Ishii, H. Hattori, M. Nishikawa, K. Watanabe and 0.
Suzuki, Jpn. J. ForensicToxicol., 14 (1996) 30.
86 F. Prosen, K. Watanabe, A. Ishii, H. Seno and 0. Suzuki, Jpn. J. ForensicToxicol., 15
(1997) 151.
87 F. Musshoff, H. Junker and B. Madea, Clin. Chem. Lab. Med., 37 (1999) 639.
88 A. Namera, T. Watabe, Y. Yashiki, T. Kojima and T. Urabe, J. Chromatogr. Sci., 37
(1999) 77.
89 Y. Iwasaki, M. Yashiki, N. Nagasawa, T. Miyazaki and T. Kojima, Jpn. J. Forensic
Toxicol., 13 (1995) 189.
90 X.P Lee, T. Kumazawa and K. Sato, Int. J. Legal. Med., 107 (1995) 310.
91 X. Lee, T. Kumazawa, S. Furuta, T. Kurosawa, K. Akiya, I. Skiya, K. Sato, Jpn. J.
ForensicToxicol., 15 (1996) 21.
92 E. Schimmiing, K. Levsen, C. Koehme and W. Schuermann, Fresenius' J. Anal.
Chem., 363 (1999) 88.
93 R. Andreoli, P. Manini, E. Bergamaschi, A. Brustolin and A. Mutti, Chroma-
tographia,50 (1999) 167.
94 B. Dehon, L. Humbert, L. Devisme, M. Stievenart, D. Mathieu, N. Houdret and M.
L'hermitte, J. Anal. Toxicol., 24 (2000) 22.
95 T. Kumazawa, X. P Lee, K. Sato, H. Seno, A.Ishii and 0. Suzuki, Jpn. J. Forensic
Toxicol., 13 (1995) 182.
96 T. Kumazawa, K. Sato, H. Seno, A. Ishii and 0. Suzuki, Chromatographia,43 (1996)
59.
97 T. Watanabe, A. Namera, M. Yaskiki, Y. Iwasaki and T. Kojima, J. Chromatogr.B,
709 (1998) 225.
98 Y. Cai, S. Monsalud, K.G.Furton, R. Jaffe and R.D. Jones, Appl. Organomet. Chem.,
12 (1998) 565.
99 L. Dunemann, H. Hajimiragha and J. Begerow, Fresenius'J.Anal. Chem., 363 (1999)
466.
100 B. He, G. Jiang and Z. Ni, J. Anal. Atom. Spectrom. 13 (1998) 1141.
101 K. Takekawa, M. Oya, A. Kido and 0. Suzuki, Chromatographia,47 (1998) 209.
102 X. Yu, H. Yuan, T. Gorecki and J. Pawliszyn, Anal. Chem., 71 (1999) 2998.
103 C. Grote and J. Pawliszyn, Anal. Chem., 69 (1997) 587.
104 W.H.J. Vaes, E.U. Ramos, C. Hamwijk, I.V. Holsteijn, B.J. Blaauboer, W. Seinen,
H.J.M. Verhaar and J.L.M. Hermens, Chem. Res. Toxicol., 10 (1997) 1067.
105 C. Camarasu, M. Mezei and A. Szabo, Acta Pharm.Hung., 69 (1999) 77.
106 K. Kongshaug, S. Pedersen-Bjergard, M. Krogh and K. Rasmussen, Chroma-
tographia,50 (1999) 247.
107 K.G. Furton, Y.-C. Hong, Y.-L. Hsu, T. Luo, S. Rose and J. Walton, J. Chromatogr.
937
Sci., 40 (2002) 147.
108 K. Higgins (Ed.), Proceedings of the SPIE Conference on Investigation and Forensic
Science Technologies, Boston, 1998, 3576, p. 41-46.
109 B. Hall, M. Satterfield-Doeerr, A. Parikh and J. Brodbelt, Anal. Chem., 70 (1998)
1788.
110 S. Yamamoto and H. Kataoka, J. Chromatogr.B, 707 (1998) 99.
111 S. Strano-Rossi and M. Chiarotti, J. Anal. Toxicol., 23 (1999) 7.
112 K. Miller, C. Poole, T. Pawlowski, Chromatographia,42 (1996) 639.
113 L. Ng, M. Hupe, J. Harnois and D. Moccia, J. Sci. Food Agric., 70 (1996) 380.
114 C. Bicchi, O., Panero, G. Pellegrino and A. Vanni, J. Agric. Food Chem., 45 (1997)
4680.
115 J. Clark and J. Bunch, J. Agric. Food Chem., 45 (1997) 844.
116 J.R. Almirall, A.D. Garcia, J. High Resolut. Chromatogr., in preparation
117 K.G. Furton, A.J. Sabucedo, J. Rein and W.L Hearn, J. Chromatogr., in press.
938
Chapter28
28.1 INTRODUCTION
It is currently recognized that the determination of the total trace metal burden
in most biological and environmental samples provides insufficient information
about lability, i.e., bioavailability and toxicity, or for use in risk assessment
scenarios. The separation and quantification of identifiable forms of an element
constitute the field of speciation and atomic spectrometry detectors, when
coupled to an appropriate separation technique, currently play a major role in
the development of this discipline, with potential impact in areas relating to
nutrition, toxicology, geochemistry and environmental chemistry.
Speciation of trace elements provides more specific information about the
real status and impact of a given element in an environmental or biological
system. The term "chemical species" was originally used to refer to a specific
form (monoatomic or molecular) or configuration in which an element can occur,
or to a distinct group of atoms consistently present in different compounds or
matrices [1]. More recently, the International Union of Pure and Applied
Chemistry (IUPAC) has proposed definitions of terms relating to speciation [2]:
"the analytical activity of identifying and measuring the quantities of one or
more individual chemical species, where the latter are defined as specific forms
of an element defined as to molecular, complex, or nuclear structure or oxidation
state."
To understand the impact and availability of target elements in any system
requires different approaches, which comprise the extended directions of
speciation analysis:
(i) Determination of the exact chemical form of a given element includes not
only its oxidation state, but also the structure of the molecule, if the metal is
covalently bonded. It is well known that the different chemical forms of an
element may exhibit significant differences in toxicity, mobility or reaction
kinetics.
940
of confidence [18]. Chemical measurement traceability is the property of the
result of a measurement, either physical or chemical, or the value of a standard
whereby it can be related to stated references, usually national or international
standards, through an unbroken chain of comparisons [18]. The use of CRMs is
necessary for all analytical laboratories undertaking methods development to
ensure the validity of a proposed new approach. During the past decade, a
number of CRMs have become available for methylmercury, trimethyllead [19],
organoarsenic [20-22], alkyltin [23,24] and chromium oxidation states [25].
Table 28.1 summarizes currently available CRMs and their sources for organo-
metallic species. Most reference materials are based on matrices of environmen-
tal interest, mirroring the increase in environmental security issues which arose
during the 1980s and 1990s. Only very few biological/clinical CRMs are available
for trace element speciation; the biotech revolution will likely induce production
of a number of health related speciation standards in the future.
An important issue for the speciation field is the provision of fully character-
ized CRMs, wherein not only the amount of a given species has been defined, but
the total content of the element along with information values for macro (N, C,
0) and micro constituents (trace elements) is also available. This is important
from the viewpoint of methods development, to know precisely the nature of the
material in question, and what type of interferences may be encountered. This
approach has been adopted by the National Institute of Standards and Technol-
ogy (NIST) and the National Research Council of Canada (NRCC).
Several articles [26-29], and recently an entire book [29], have been devoted
to questions associated with the production of reference materials for speciation
purposes.
28.1.3 Instrumentation
941
TABLE 28.1
Commercially available CRMs for organometallic speciation
Arsenic Arsenobetaine: Solution BCR CRM 626 1031 tmol/kg
Tuna fish tissue BCR CRM 627 52 mol/kg
Dogfish muscle NRC CRM DORM-2 16.4 mg/kg
Dimethylarsenic acid: Tuna fish tissue BCR CRM 627 2 gmol/kg
Tetramethylarsonium: Dogfish Muscle NRC CRM DORM-2 0.248 mg/kg
942
spectroscopic detector, and very low background due to the high purity inert gas
used as the mobile phase. Typical drawbacks of such approaches are that most of
the target organometallic compounds are usually not directly suited to GC
separation because of their thermal instability, high reactivity or ionic nature in
the environment. Therefore, it is necessary to apply some form of derivatization
prior to analysis. Several reviews have been published on the application of GC
for speciation analysis [12,31,32].
28.1.3.2 HPLC and CE
As most of the species of interest are ionic, polar compounds and are generally
found in the aqueous environment, liquid chromatography and capillary electro-
phoresis are a natural choice for separation. The matrix load on the atomic or
ionic source used for detection can present a significant limiting factor. Typical
LC flow-rates (1-2 ml/min) can be introduced only in conjunction with low
efficiency nebulization to split matrix load about 100-fold. The situation is only
worsened if the eluent systems contain organic solvents, such as methanol or
acetonitrile, because the tolerance of plasma- or flame-based detection systems
is much lower for these solvents than for water. As a result, detection limits for
HPLC- or CE-based methods are several orders of magnitude worse than GC-
based methods. Miniaturization of the aqueous phase separation system will
likely overcome many of the limitations presented by the mobile phase. Use of
aqueous phase separation is very popular for arsenic [33], selenium [34-36],
chromium [37,38] and platinum [39,40] speciation, partly because there are no
readily available gas chromatographic derivatization methods for these species.
Condensed phase separation is gaining momentum in the speciation field, also
due to the increasing interest in metal containing biopolymers such as proteins
and DNA [9,41-46].
28.1.3.3 Detection
The very low detection limits required to quantitate trace element species relies,
in most cases, on powerful atomic spectroscopic detectors. Atomic absorption
[47], atomic fluorescence [48-51], atomic emission with inductively coupled
plasma (ICP) sources [52-54], glow-discharges [55,56] or microwave induced
plasma (MIP) sources and, recently, electrospray ionization mass spectrometry
[57], are most commonly used. Additionally, several of the conventional gas
chromatographic detectors are finding direct application, principally the flame,
pulsed flame photometric and ionization detectors [58]. However, the detection
system of choice in most cases is currently ICP-MS [59-65]. These systems
utilize an argon plasma, which serves as the elemental ion source for the mass
spectrometer. The identification of the analyte is based entirely on chromato-
graphic retention times. Since no molecular information is available from
conventional plasma sources, structural information is lost and identification
becomes impossible. Recent work with multidimensional instrumentation has
helped to rectify this problem [17]. A combination of liquid chromatography with
electrospray/ionspray sources can be used for measurement of fragment ion
943
patterns, generating (potentially near simultaneous) structural information
which can be used to deduce species identity in the absence of a standard.
Additionally, "dual mode" analysis, achieved by judicious selection of operating
conditions in the atmospheric pressure ionization source interface, permits
elemental atomic and molecular information to be obtained on the same sample.
"Soft" ionization sources can also provide for a dual mode of operation to permit
atomic and molecular information to be achieved with glow discharges [66,67],
furnace atomization plasma ionization mass spectrometry (FAPIMS) [68], low
pressure ICPs [69] and low power microwave induced plasmas (MIPs) [70].
The ideal analytical instrument would, without human intervention, perform all
analytical steps, from sampling to data analysis, and even undertake decision
making based on the results obtained. Although today's sophisticated instru-
ments can analyze complex samples and evaluate results, most sampling and
sample preparation practices are based on 19th century technologies, such as the
common Soxhlet extraction method. Traditional sample preparation methods
are typically time consuming, employ multi-step procedures having high risk for
loss of analytes, and use extensive amounts of organic solvents. These character-
istics make such methods very difficult to automate and integrate into modern
sampling/separation systems. Consequently, most of the analysis time is
944
consumed with sampling and sample preparation, which contribute significantly
to the cost of analysis.
Extensive use of organic solvents in analytical laboratories is no longer
tolerated because of the associated health risks and disposal concerns. As a
result, several different extraction approaches have been studied in the past
decade aimed at enhancing the extraction efficiency. The major goals for
improvement include reducing solvent usage, decreasing the extraction time,
decreasing the amount of sample required for analysis, and automation of the
extraction procedure, preferably online with separation and detection. Historic-
ally, the first extraction methods used were conventional solvent extraction
approaches. In the last ten years, however, a significant scientific effort has been
made to enhance the efficiency of classical solvent leaching procedures by using
microwave energy, ultrasonic waves, supercritical fluids and pressurized
solvents. Table 28.2 summarizes some typical working conditions for these new
extraction methods compared to the conventional Soxhlet process. A brief
review of these techniques is presented below.
Several books [88-90] and review articles [91-94] have been published on the
theoretical and practical approaches to the use of microwave chemistry and
microwave assisted extraction. The microwave source is a non-ionizing radiation
945
-els~~ 5" ~~~~~~~10.
.5'
o ~ gt
=0 ao 0
C) -
S
O 0£
"o ~
.".B ~ IE $5 0C)00~ed
0
vF a: 0"1 m
ae P - i=W.= 0 cl . z
4C ) ..
5 ®
o 0
Cd
0 o
0"0 11~~.
aO) C-
0 0"~~~~ o 4°
C)
'o C)
w
F
VA
0 M:
L Si .
o0 d
001 m oz
'ji.Q e E e iL i , 2-
a000
aC) ,bX
, 0
a4) i
0 O ~~ ~ ~ ~ t .
a)4
'd a: .)
o 0
.00 0" At
co 0 9u
C)
:t
01)
kF - w E r0BC- )OC)0 e
R la
C -- 1<
.-
,0
t
Oa
0 .
CID 0
00
a
U, O5
. o
.,
*.-
0. I
0
10
-a
¢ 01cv
a0
946
TABLE 28.3
Selected physical parameters for some of the most commonly used solvents. (All parameters
obtained from Refs. [89] and [94].)
Solvent Dielectric Dipole Dissipation Boiling point' Closed vessel
constant (e') momentb factor (tan 8) (°C) temperatured (°C)
-
(x 10 4)
field in the frequency range 300-300,000 MHz that induces motion of molecules
by ion migration and rotation of dipoles. Dipole rotation can occur not only in the
selected solvent, but also in the sample itself. The alignment and relaxation of
the molecular dipoles in the microwave field occurs 109 times per second
resulting in extremely quick heating. Commercial microwave extraction devices
operate at 2.45 GHz.
The ability of a solvent to absorb microwave energy and pass it on to other
molecules in the form of heat will partly depend on the dissipation factor (tan 6).
The dissipation factor is given by the following equation [88]:
tan 6 = E"/c' (28.1)
where £" is the dielectric loss (a measure of the efficiency of converting micro-
wave energy into heat) and ' is the dielectric constant (a measure of the
polarizibility of a molecule in an electric field). Polar molecules and ionic
solutions will strongly absorb microwave energy because they have a permanent
dipole moment that will be affected by the microwaves. However, non-polar
solvents, such as hexane, will not heat up when exposed to microwaves. Table
28.3 presents selected physical parameters, including dielectric constants and
dissipation factors, for solvents that are used in most applications [94].
Because of the high dielectric constant of water, the free water content of the
sample induces a very rapid and concentrated heating during the microwave
extraction process. The resulting pressure increase may result in the rupture of
membranes in biological samples, enabling various constituents to be released. A
similar mechanism is suspected to occur in sediments and soils. Microwave heat-
ing of the water in the matrix leads to boiling, bubble formation and increased
pressure, resulting in disintegration of the microstructure of the matrix, thereby
increasing the available surface for extraction by the solvent [91].
Microwave assisted extraction can be performed either with closed vessels
(under controlled pressure and temperature), or with open vessels (at atmos-
pheric pressure). Both systems are illustrated schematically in Fig. 28.1.
947
Diffused microwave system
I '/- -A
-I
I hnrntablc,
MLtgnetrol ter
I vessel
Solvent
Iocsed | t2··
Fig. 28.1. Closed and open vessel microwave units. Reprinted from Trends in Analytical
Chemistry, Vol. 19, V. Camel: Microwave-assisted solvent extraction of environmental samples,
pp. 229, Copyright (2000), with permission from Elsevier Science.
In closed vessels, the solvent can be heated above its standard boiling point,
thus enhancing both extraction speed and efficiency. Such systems permit tem-
perature control of the extraction process. In addition, they usually lead to an
increase in sample throughput because several vessels can used simultaneously.
Because of the non-homogeneous electric field in the microwave cavity, sample
vessels are usually placed on a turntable. In so-called open systems, extractions
proceed under atmospheric pressure. As a consequence, the maximum possible
temperature is determined by the boiling point of the solvent at ambient pres-
sure. Such systems use focused microwaves, so that the heating of the sample is
homogeneous and very efficient. Vapour loss is prevented by the presence of a
reflux system fitted to the top of the extraction vessel. Organometallic com-
pounds are typically extracted using such open vessel systems, as the close con-
trol of the energy delivered to the sample prevents destruction of carbon-metal
bonds [95].
Microwave assisted extraction has been popular for leaching of organotin
from biological and environmental samples such as fish-tissue and sediment
[95-101]. The polar nature of the butyltin species permits it to easily absorb
microwave energy, thereby facilitating solubilization of the species and
minimizing the problem of re-adsorption. Microwave assisted extraction has also
been used for efficient recovery of methylmercury [102-107].
948
28.2.3 Supercritical fluid extraction
The properties of fluids under supercritical conditions, i.e., between those of liq-
uids and gases, render salvation power (similar to liquids), while the low viscos-
ity and high diffusivity (similar to gases) facilitate transport phenomena, thus
making extraction faster and more effective. The most common SF extractant
used at present is CO2 because of its low toxicity and inflammability, reasonable
critical conditions and chemical inertness. The most significant aspect of SC-
CO 2 is its weak interaction with both analytes and matrices which, when pure,
provide poor efficiencies in the extraction of environmentally persistent pollut-
ants. This problem can be overcome either by the addition of co-solvents (metha-
nol, pentane, toluene) to increase the polarity of the solvent system, or by
reducing the polarity of the target analytes by complexation or ion pair forma-
tion, esterification and reverse-micelle formation. The advantages of SFE vs
both Soxhlet and other conventional leaching techniques include: reduced
extraction time, a compatibility for on-line coupling to either detectors or
chromatographs, reduced need for cleanup due to the high selectivity achieved
through manipulation of both pressure and temperature and, most significantly,
elimination of solvent removal steps as the extractant is released from the
leached species after depressurization. Trapping techniques employed for collec-
tion of analyte after extraction include: liquid collection, cryogenic trapping and
solid-phase trapping, the latter being the most effective as it allows for simulta-
neous collection, cleanup and concentration prior to either individual chromato-
graphic separation or direct detection. Despite the advantages of SFE and the
wide variety of commercially available apparatus and instruments, which have
stimulated the development of new applications, this technique has so far not
fulfilled the expectations of suppliers and users owing to the following: (a) the
large discrepancies in extraction efficiency between spiked and endogenous
analytes in samples (b) the number of methods reported in the literature for
which the efficiency is lower than that provided by classical methods; (c) the lack
of information about both the means of overcoming analyte-matrix interactions
and the selection of the most appropriate modifier.
Recently, an excellent review article by Bayona [108] has appeared outlining
SFE applications in the speciation field. Most compounds of interest exhibit high
polarity and/or are of an ionic nature wherein direct extraction by non-polar
SF-CO2 is not, in most cases, a suitable approach. As noted earlier, to overcome
the issue of high polarity of the analyte, complexation/derivatization may be
undertaken or the polarity of the extraction phase may be increased using
matrix modifiers. Both approaches have been examined in a number of studies.
The typical approach used for butyltin leaching involves application of metha-
nol/acetic acid/formic acid/HCl modifiers. The use of complexing agents, such as
ammonium pyrrolidine dithiocarbamate (APDC) in the presence of the above-
mentioned polar modifiers, does not improve the extraction efficiency. Most
likely, the resulting complexes are not stable under the supercritical conditions
949
[109]. Grignard derivatization has also been used in an attempt to convert ionic
compounds into non polar species which may be more soluble in supercritical
carbon dioxide [110]. Several combinations of modifiers and different matrices
have been described for the extraction of organotin species [108,109,111-116].
Supercritical fluid CO 2 has also been used for the extraction of di- and trialkyl-
lead species. Determination of alkyllead following SFE was successfully used in a
CRM certification campaign, although the material, an urban dust sample, was
spiked, which may have conferred a simplified leaching process [19]. SFE has
also been applied for methylmercury extraction [105,117,118-1211.
28.2.4 Sonication
950
Solvent-free techniques are particularly desirable for trace analyses so as to
avoid the cumbersome problems associated with solvent impurities. However, a
significant portion of the work related to speciation employs solvent trapping/
preconcentration. This normally involves trapping the analyte in an organic
solvent. If the volatility of the analyte is significantly lower than that of the
trapping solvent, further preconcentration can be achieved by evaporation of the
solvent. Organic solvent extraction/trapping is typically used in combination
with Grignard [130] derivatization or ethylation with sodium tetraethylborate
(NaBEt 4) [131].
The theory and practice of SPME has been presented elsewhere in this book and
can also be found in a recent text [132]. Therefore, the focus here is on some
typical elemental speciation applications of this technique. Mester et al. [133]
recently reviewed the determination of organometallic species using SPME
sampling with gas chromatographic separation/detection. Rapid advances in this
field have served to expand the scope of applications, interface development,
instrument technique and fiber preparation.
Most elemental speciation applications of SPME are based on their coupling
with classical gas chromatographic instrumentation. Methods have been
described for mercury, tin, lead, arsenic and selenium. Both headspace and
direct extraction methods have been used for the sampling of organometallic
compounds and, generally, some type of derivatization process is necessary for
their GC separation. The typical derivatization method used for tin, lead and
mercury species is ethylation using NaBEt 4 reagent. Moens et al. [134] reported
a comparison of sensitivity between 'conventional' liquid/liquid- and headspace
SPME-extraction of butyltin and organolead compounds. They found that head-
space SPME provided an approximately 300-fold better sensitivity for butyltin
compounds and about a 35-fold enhancement for trimethyllead. As a result,
analysis of butyltin species based on SPME can thus be accomplished using a
conventional Flame Ionization Detector (FID) instead of the more expensive
mass spectrometric systems.
Several researchers have reported results on the headspace sampling/deter-
mination of metal hydrides using solid phase microextraction techniques. Jiang
and coworkers [135,136] developed a sampling method for organomercury
species based on hydride generation using KBH 4 reagent. Mester et al. [137]
tested the compatibility of two different fibers (and two different extraction
phenomena) for sampling volatile metal hydrides coupled with ICP-MS detec-
tion. An adsorption-based carboxen coating provided better sensitivity than an
absorption-based extraction with a liquid-type polydimethylsiloxane polymeric
coating (PDMS). The success of absorptive sampling confirmed the relatively
high stability of the studied metal hydrides (As, Se, Sn and Sb) because they
survive diffusion into and release from the polymeric liquid.
951
Guidotti et al. [138,139] described the determination of Se(IV) by headspace
SPME-GC/MS following derivatization based on either piaselenol formation or
ethylation. The method can also be applied to Se(IV) and Se(VI) speciation by
determining the Se(IV) content first, followed by the total selenium concentra-
tion after converting all selenium species to the Se(IV) oxidation state.
Mester and Pawliszyn [140] reported a method for the speciation of di-
methylarsinic acid (DMA) and monomethylarsonic acid (MMA) by SPME-
GC/MS. Thioglycolmethylate (TGM) was used for derivatization, having as its
basis the known "affinity" of the arsenic compound for the thiol-group [30,141].
Organometallic species which are normally saturated (non-ionic) and suffi-
ciently volatile can be sampled by SPME and determined by GC without deriva-
tization. Gorecki and Pawliszyn [142] described a simple sampling procedure for
tetraethyllead. A GC-MS was used for quantitation. Similar studies were per-
formed by Snell et al. [143] for determination of dimethylmercury in the head-
space of natural gas condensates. This method was characterized by a very short
sampling time (30 s). The extremely complex matrix, comprising volatile organic
material (which is also extracted with the dimethylmercury), presented a serious
problem even for the extremely selective microwave induced plasma atomic
emission detection system used.
Mester et al. [144] recently described a SPME method for methylmercury
determination based on the relatively high vapour pressure of methylmercury
chloride. Headspace SPME sampling was performed above a methylmercury
solution which was previously saturated with sodium chloride. A slightly polar
solid coating (PDMS/DVB) was used for extraction. Sample introduction into an
ICP-MS was achieved with a unique thermal desorption interface consisting of a
heated glass-lined splitless type GC injector placed directly at the base of the
torch to minimize the length of transfer-line. This arrangement provided for fast
desorption and high sample introduction efficiency. A schematic diagram of the
thermal desorption ICP-MS interface is shown in Fig. 28.2. This unit provides a
very versatile interface, which can be connected to virtually any atomic spectro-
scopic source.
SPME can also be used for sampling of non-volatile/ionic metal species from
aqueous media. Jia et al. [145] studied the extraction of mercury species from
INDUCTIVELY COUPLED
PLASMA MASS SPECTROMETER
952
aqueous solutions. The SPME fiber was modified for sampling in that different
crown ethers were adsorbed onto the non-polar SPME coating by simply dipping
the fiber into a solution of the crown ether. An HPLC system was subsequently
employed for the separation of free complexing agent from the metal complex.
Because there are no commercially available ion-exchange or specific metal
selective coatings available, the study of new coating materials and extraction
principles is of interest. Caruso and coworkers [146] recently reported on
development of SPME sol-gel coatings for organometal determination. Sol-gel
coatings are very mechanically, thermally and chemically stable in organic
solvents as well as acidic and basic solutions where commercial SPME coatings
cannot be used. The fibers were evaluated for SPME-HPLC applications using
diphenylmercury, trimethylphenyltin and triphenylarsine. The effect of organic
solvents and acids on the fibers was also studied.
In a recent paper, Wu et al. [147] described the application of a polypyrrole
polymer coating for SPME extraction of small anionic analytes such as C1-, F-,
Br, NO,- P04 3 , S042-, SeO4 2- , SeO32- and AsO43- . Extraction was based on the ion
exchange characteristics of the polypyrrole coating. The same coating has been
used for sampling arsenobetaine from aqueous samples [126].
A different approach to the use of liquid phase microextraction is to substi-
tute an open tubular capillary for the sample introduction loop in an HPLC
system. Such an in-tube SPME system has been described for the extraction of
trimethyllead and triethyllead [148,149] arsenobetaine [126], and tributyltin
[125] from biological samples.
Cold traps are used for two main reasons: enrichment purposes and solute band
concentration. Most cold traps are home-made and are operated manually,
requiring skill and experience of the operator. In keeping with the general trend
in analytical instrumentation, however, any technique will be successful only if
it can be operated automatically and unattended.
A distinction has to be made between 'cryogenic condensation' and 'cryogenic
focusing', although the common term 'cryogenic trapping' is used for both. The
term cryogenic condensation is used here for techniques in which the volatile com-
pounds are trapped simply by condensation in traps, which usually contain no sta-
tionary phase. Condensation is also the prevailing mechanism if a coated capillary
is cooled down to such a low temperature that the liquid phase will solidify and
lose its properties as a chromatographic phase. The term cryogenic focusing is
used if the volatile compounds are trapped in the liquid phase of a column at a low
temperature which, however, is still above its glass transition temperature. Dur-
ing sample introduction, the compounds dissolve in the liquid phase and immedi-
ately but slowly migrate downstream on the cold column.
In his recent article, Kolb [150] reviewed headspace sampling methods,
including cryotrapping: among many others, it is an excellent starting point for
953
headspace sampling techniques. Donard's research group has pioneered the
adaptation of cryotrapping technology for speciation studies [102,151-163].
Cryotrapping is widely used in the speciation field, especially for precon-
centration of very volatile metal(oid) species following hydride generation/
ethylation derivatization. The typical trapping temperature is in the -150 to
-196°C range obtained with liquid nitrogen cooling. Cryotrapping can also be
used as a means of sampling volatile metal(oid) species in the environment. The
trapped volatile species can be introduced into a separation system (typically
GC) after a fast heat-up, or the cryotrap can be applied as a 'crude' separation
device where, with gradual heating, the analytes leave the trap according to their
boiling points.
Cryogenic sample collection has been used for the determination of organotin
species after hydride generation [164-167], ethylation with sodium tetraethyl-
borate [168] and for the sampling of saturated alkyltin species from sea water and
air [161]. It has also been extensively applied in speciation studies of arsenic
[169-174] and mercury [151,154,155,158,163,175-178]. Volatile metal(oid) spe-
cies have been detected in air [77,156], water [153,157,161,162] and in landfill gas
samples [179,180] using cryogenic sampling. Most of these studies employed
detectors having multielement capabilities (manly ICP-MS) for simultaneous
monitoring of the volatile species of different elements, such as arsenic, antimony,
tin, selenium and mercury. Volatile metal carbonyl species (Ni(CO) 4 Mo(CO)6 and
W(CO) 6) have been detected in fermentation gases from a municipal sewage treat-
ment plant using cryogenic gas sampling/trapping [181,182].
One of the main advantages of cryogenic trapping compared to all other
preconcentration methods is that unstable/reactive volatile organometallic
species do not come into contact with any liquid or solid sorbent material,
significantly reducing the possibility of alteration/decomposition of analyte and
also reducing the possibility of introducing contaminants via the solvent used in
other techniques. At same time, cryotrapping technology, especially for field
application-despite its clear advantages-remains not very appealing because
of its bulkiness and the requirement for liquid nitrogen.
954
,o SS
o
o, 3
0
= ;-8
' a a) --4 .-
aa)
o
S
C- .a_
0
0 Cr,
- 0Cr5
' 6 S
-S 8¼
Cr~
S~*S o 6
G1E
~ .2 ,2
o.2
Cr m
o. .B . -° at
*5r
M
-C
O .
~
o ~, O
S.®
C
.2
w
o,
0 a)2 o O~r t
,o a .s .eCoa0._ O
~
at C)
-I 2 2
C
0
oC a)
9s
C)
; ~ _O ..oO
ata
r -6 ~ ~ O X
0 0
S o
o '
E
>4
~'~ ~o
R r
I
52 a~-l
Ca 7
a)
a) Op'at C-
.2 52aCr-g
-0 baa-' e Mf:,Qi
a) oa
9 S
ag Cr a) Ca C.C
a)
aC a_a)
C 2Cr 5 a))
O-S
C) o Od. ,.o 3
o. - . 255
o, o)C.
S
a m a C
Cr- Sm o, a>~ a 3 )
7H
m s
0 D
0 a
*2 C,
Cr
C
c0 a P r
r a) 2
D
6 .0 o
a Cr
Sd >4 i S tE- 5
Cr
0 ~o
Cr
Ca)
D
955
[A 1IA VIllA
-
H He
IiA IVA VA
"' VIA
-" v1
Li Be B C N 0 F Ne
A
Na M, Al osi P S Cl Ar
Ai A A IA A A
iO
Rb Sr .Y Zr Nb Mo Tc Ru Rh Pd Ag Cd In . A
Sn Sb Xe
] A A A- A A A A
Cs Ba La Hf Ta W Re Os Ir Pt Au Hg
A DA
Pb F'
Pi,
At Rn
* A A A ** A
Fr Ra Ac
Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
Th Pa U Np Pu Am Cm Bk Cf Es Fm Md No L.
Fig. 28.3. Periodic Table of the elements indicating elements which can be volatilized at
atmospheric pressure and room temperature as: O, ethylated species (with sodium tetra-
ethylborate); A, hydride with conventional hydride generation (sodium tetrahydroborate); A,
newly discovered hydride forming elements; *, carbonyl formation (using carbon monoxide); 0,
halide generation. *Osmium tetraoxide. **Elemental vapour (Hg° and Cd°). Molybdenum and
tungsten carbonyl species have been detected in the environment [182].
956
During hydride generation, alkyl group(s) bonded to the metall(oid) center
remain unchanged, providing an excellent means of differentiating between
alkyl and inorganic species. Hydride generation combined with GC has been
used for the determination of inorganic, mono and dimethylarsenic [184,185].
Because of the specific requirements regarding the acidity of the sample matrix
during hydride generation of different species, the process can be fine-tuned to
selectively derivatize certain species [186-188]. Generally, the hydride genera-
tion process is rapid, shows high efficiency, and is inexpensive. On the down side,
the gaseous products are difficult to transfer to a GC and the efficiency of
hydride generation can be heavily influenced by the presence of transition
metals.
Hydride generation can also be used as a gaseous sample introduction/inter-
face unit between aqueous phase separation (HPLC, CE) and atomic spectroscopic
detectors. The clear advantage of this approach is the high sample introduction
efficiency compared with conventional nebulization processes. However, the
hydride interface tends to further complicate the already complex hyphenated
systems, introducing additional uncertainty factors. It remains a popular inter-
face/sample introduction technique between chromatography and atomic spectro-
scopic detectors, especially when the detectors are less sensitive and/or robust and
the trade-off for better detection limits is worthwhile [189-193].
28.4.2 Ethylation
957
compounds. The application of the propyl-[216,217] and phenylborate [218,219]
reagent also appears to be a promising approach.
The transformation of analyte into a more volatile compound can also be achieved
by reaction with a Grignard reagent (R-MgX) in a suitable solvent. The R alkyl
group can be methyl, ethyl, propyl, butyl, pentyl, or hexyl. Alkylation via the Grig-
nard reaction is a widely used derivatization technique for organotin [220-222],
alkyllead [223], antimony [224] and germanium [225] prior to their determina-
tion. The Grignard reaction is generally performed after an organic solvent
extraction because the Grignard reagent is not compatible with water. Grignard
derivatization permits the determination of different organometallic species in
various environmental matrices (water, sediment, biota) with high derivatiza-
tion yields and excellent reproducibility. However, the derivatization procedure is
quite lengthy, requires several sample manipulation steps which increase the risk
of contamination, decomposition and losses, and significantly increases the analy-
sis cost. In a recent study, Morabito et al. [213] compared Grignard pentylation
and ethylation with ethylation using sodium tetraethylborate for determination
of alkyltin in mussel tissue (Fig. 28.4). The two Grignard derivatization methods
provided very similar results, however, sodium tetrethylborate-based ethylation
was always biased somewhat lower for the studied species.
958
introduced into a detector. Alternatively, they can be pre-concentrated on
sorbent material. Inorganic arsenic, mono- and dimethylarsenic [227],
germanium [228], tributyltin [229], methylmercury [144], silicon [230,231] and
gallium have been reported. The high boiling point and the thermal instability of
these species makes them unsuitable for gas chromatographic analysis. At the
same time, the high boiling points facilitate room temperature trapping as
opposed to cryotrapping necessary for hydrides. This process is also free from
classical transition metal interferences typical of hydride generation methods.
959
complete the measurement with an extremely selective atomic spectroscopic
detector. This approach can be entertained for methylmercury/inorganic
mercury and alkyllead/inorganic lead.
Because serious matrix interferences may arise in many cases during speci-
ation analysis, this has to be seriously considered when deciding on the calibra-
tion approach to be taken. Analytical methods are sufficiently robust and the
sample matrix is sufficiently well-characterized to employ external calibration in
only a limited number of cases. The method of standard additions or isotope
dilution is thus strongly encouraged. Isotopic spiking and isotope dilution mea-
surements provide the most advanced calibration/quantification approaches,
based on changes in the known isotopic distribution of an element before and
after spiking with a single isotope. When used for speciation analysis, isotope
dilution requires advanced mass spectrometric facilities to enable accurate and
precise isotope ratio measurements to be achieved characterizing a single chro-
matographic peak. Many laboratories are equipped with such instruments but,
unfortunately, isotopically enriched species-specific standards are not commer-
cially available.
REFERENCES
960
20 A. Chatterjee, Y. Shibata, J. Yoshinaga and M. Morita, Appl. Organomet. Chem., 15
(2001) 306.
21 J. Yoshinaga, A. Chatterjee, Y. Shibata, M. Morita and J.S. Edmonds, Clin. Chem., 46
(2000) 1781.
22 F. Lagarde, M.B. Amran, M.J.F. Leroy, C. Demesmay, M. Olle, A. Lamotte, H.
Muntau, P. Michel, P. Thomas, S. Caroli, E. Larsen, P. Bonner, G. Rauret, M.
Foulkes, A. Howard, B. Griepink and E.A. Maier, Fresen. J. Anal. Chem., 363 (1999)
18.
23 R. Morabito, H. Muntau, W. Cofino and P. Quevauviller, J. Environ. Monit., 1 (1999)
75.
24 J. Yoshinaga, H. Kon, T. Horiguchi, M. Morita and K. Okamoto, Anal. Sci., 14 (1998)
1121.
25 K. Vercoutere, R. Cornelis, L. Mees and P. Quevauviller, Analyst, 123 (1998) 965.
26 P. Quevauviller, Fresen. J. Anal. Chem., 354 (1996) 515.
27 B. Michalke, Fresen. J. Anal. Chem., 363 (1999) 439.
28 P. Quevauviller, Spectrochim. Acta B, 53 (1998) 1261.
29 P. Quevauviller, Method Performance Studies for Speciation Analysis. RSC,
Cambridge, UK, 1998.
30 Z. Mester, G. Vitanyi, R. Morabito and P. Fodor, J. Chromatogr.A, 832 (1999) 183.
31 Y.S. Drugov, J. Anal. Chem., 53 (1998) 606.
32 H. Tao, Bunseki Kagaku, 46 (1997) 239.
33 T. Guerin, A. Astruc and M. Astruc, Talanta, 50 (1999) 1.
34 K. Pyrzynska, Chemia Analityczna, 40 (1995) 677.
35 K. Pyrzynska, Analyst, 121 (1996) R77.
36 K. Pyrzynska, Anal. Sci., 14 (1998) 479.
37 H. Ding, L.K. Olson and J.A. Caruso, Spectrochim. Acta B, 51 (1996) 1801.
38 K. Vercoutere and R. Cornelis, Qual. Assurance Environ. Anal., 17 (1995) 195.
39 W.R.L. Cairns, L. Ebdon and S.J. Hill, Fresen. J. Anal. Chem., 355 (1996) 202.
40 S. Lustig, J. De Kimpe, R. Cornelis, P. Schramel and B. Michalke, Electrophoresis,20
(1999) 1627.
41 M.M. Bayon, A.B.S. Cabezuelo, E.B. Gonzalez, J.I.G. Alonso and A. Sanz Medel, J.
Anal. Atom. Spectrom., 14 (1999) 947.
42 H. Koyama, K. Omura, A. Ejima, Y. Kasanuma, C. Watanabe and H. Satoh, Anal.
Biochem., 267 (1999) 84.
43 K.T. Suzuki, Analusis, 26 (1998) M57.
44 A.T. Lombardi and A.A.H. Vieira, Phycologia, 37 (1998) 34.
45 C.N. Ferrarello, M.M. Bayon, R.F. de la Campa and A. Sanz Medel, J. Anal. Atom.
Spectrom., 15 (2000) 1558.
46 K. Polec, O. Garcia Arribas, M. Perez Calvo, J. Szpunar, B. Ribas Ozonas and R.
Lobinski, J. Anal. Atom. Spectrom., 15 (2000) 1363.
47 I. Havezov, Fresen. J. Anal. Chem., 355 (1996) 452.
48 M. Vilano, A. Padro and R. Rubio, Anal. Chim. Acta, 411 (2000) 71.
49 J.L. Gomez Ariza, D. Sanchez Rodas, R. Beltran, W. Corns and P. Stockwel, Appl.
Organomet. Chem., 12 (1998) 439.
50 A. D'Ulivo and S. Rapsomanikis, Anal. Lett., 30 (1997) 2109.
51 Z. Mester and P. Fodor, Anal. Chim. Acta, 386 (1999) 89.
52 Y.L. Feng, H. Narasaki, H.Y. Chen and L.C. Tian, Anal. Chim. Acta, 386 (1999) 297.
53 I.R. Pereiro, A. Wasik and R. Lobinski, Fresen. J. Anal. Chem., 363 (1999) 460.
54 P.C. Uden, S.M. Bird, M. Kotrebai, P. Nolibos, J.F. Tyson, E. Block and E. Denoyer,
Fresen. J. Anal. Chem., 362 (1998) 447.
55 M.A. Dempster and R.K. Marcus, J. Anal. Atom. Spectrom., 15 (1999) 43.
56 N.G.O. Velado, R. Pereiro and A. Sanz Medel, J. Anal. Atom. Spectrom., 15 (1999) 49.
961
57 Stewart, II, Spectrochim. Acta B, 54 (1999) 1649.
58 M.B. DelaCalleGuntinas, C. Brunori, R. Scerbo, S. Chiavarini, P. Quevauviller, F.
Adams and R. Morabito, J. Anal. Atom. Spectrom., 12 (1997) 1041.
59 K. Sutton, R.M.C. Sutton and J.A. Caruso, J. Chromatogr.A, 789 (1997) 85.
60 J.S. Becker and H.J. Dietze, Spectrochim. Acta B, 53 (1998) 1475.
61 A.E. Croslyn, B.W. Smith and J.D. Winefordner, Crit. Rev. Anal. Chem., 27 (1997)
199.
62 J.A.C. Broekaert, Mikrochim. Acta, 120 (1995) 21.
63 S.J. Hill, L.J. Pitts and A.S. Fisher, Trends Anal. Chem., 19 (2000) 120.
64 K.L. Sutton and J.A. Caruso, J. Chromatogr.A, 856 (1999) 243.
65 G.K. Zoorob, J.W. McKiernan and J.A. Caruso, Mikrochim. Acta, 128 (1998) 145.
66 J.P. Guzowski and G.M. Hieftje, Anal. Chem., 72 (2000) 3812.
67 C. Lewis, S.K. Doom, D.M. Wayne, F.L. King and V. Majidi, Appl. Spectrosc., 54
(2000) 1236.
68 R. Guevremont and R.E. Sturgeon, J. Anal. Atom. Spectrom., 15 (2000) 37.
69 G. O'Connor, L. Ebdon, E.H. Evans, H. Ding, L.K. Olson and J.A. Caruso, J. Anal.
Atom. Spectrom., 11 (1996) 1151.
70 N.P. Vela, J.A. Caruso and R.D. Satzger, Appl. Spectrosc., 51 (1997) 1500.
71 J.L.G. Ariza, E. Morales, D. Sanchez Rodas and I. Giraldez, Trends Anal. Chem., 19
(2000) 200.
72 C. Bancon Montigny, G. Lespes and M. Potin Gautier, Water Res., 35 (2001) 224.
73 J.L. Gomez Ariza, I. Giraldez, E. Morales, F. Ariese, W. Cofino and P. Quevauviller,
J. Environ. Monit., 1 (1999) 197.
74 J. Feldmann, V.W.M. Lai, W.R. Cullen, M.S. Ma, X.F. Lu and X.C. Le, Clin. Chem., 45
(1999) 1988.
75 M.M. Gomez, T. Gasparic, M.A. Palacios and C. Camara, Anal. Chim. Acta, 374
(1998) 241.
76 J.L. Gomez Ariza, J.A. Pozas, I. Giraldez and E. Morales, Int. J. Environ. Anal.
Chem., 74 (1999) 215.
77 K. Haas and J. Feldmann, Anal. Chem., 72 (2000) 4205.
78 G.E.M. Hall, J.C. Pelchat and G. Gauthier, J. Anal. Atom. Spectrom., 14 (1999) 205.
79 R.M. Olivas, P. Quevauviller and O.F.X. Donard, Fresen. J. Anal. Chem., 360 (1998)
512.
80 Y. Wang, R.M. Harrison and P. Quevauviller, Appl. Organomet. Chem., 10 (1996) 69.
81 P. Quevauviller and O.F.X. Donard, J. Environ. Monit., 1 (1999) 503.
82 M.T. Perez Corona, Y. Madrid Albarran and C. Camara, Fresen. J. Anal. Chem., 368
(2000) 471.
83 R. Draisci, C. Marchiafava, L. Palleschi, P. Cammarata and S. Cavalli, J.
Chromatogr. B, 753 (2001) 217.
84 P.A. Gallagher, J.A. Shoemaker, X.Y. Wei, C.A. Brockhoff Schwegel and J.T. Creed,
Fresen. J. Anal. Chem., 369 (2001) 71.
85 P.A. Gallagher, X.Y. Wei, J.A. Shoemaker, C.A. Brockhoff and J.T. Creed, J. Anal.
Atom. Spectrom., 14 (1999) 1829.
86 J.W. McKiernan, J.T. Creed, C.A. Brockhoff, J.A. Caruso and R.M. Lorenzana, J.
Anal. Atom. Spectrom., 14 (1999) 607.
87 Dionex Application Note 339, 1999.
88 H.M. Kingston and L.B. Jassie, Introduction to Microwave Sample Preparation.
American Chemical Society, Washington, DC, 1988.
89 H.M. Kingston and S.J. Haswell, Microwave-Enhanced Chemistry. American Chem-
ical Society, Washington, DC, 1997.
90 H.M.S. Kingston and P.J. Walter, in: A. Montaser (Ed.), Inductively Coupled Plasma
Mass Spectrometry. Wiley, New York, NY, 1998, p. 33.
962
91 V. Camel, Trends Anal. Chem., 19 (2000) 229.
92 Q.H. Jin, F. Liang, H.Q. Zhang, L.W. Zhao, Y.F. Huan and D.Q. Song, Trends Anal.
Chem., 18 (1999) 479.
93 K.J. Lamble and S.J. Hill, Analyst, 123 (1998) R103.
94 C.S. Eskilsson and E. Bjorklund, J. Chromatogr.A, 902 (2000) 227.
95 J. Szpunar, V.O. Schmitt, O.F.X. Donard and R. Lobinski, Trends Anal. Chem. 15
(1996) 181.
96 B.S. Tselentis, M. Maroulakou, J.F. Lascourreges, J. Szpunar, V. Smith and O.F.X.
Donard, Mar. Pollut. Bull., 38 (1999) 146.
97 V.O. Schmitt, J. Szpunar, O.F.X. Donard and R. Lobinski, Can. J. Anal. Sci. Spect.,
42 (1997) 41.
98 I.R. Pereiro, V.O. Schmitt, J. Szpunar, O.F.X. Donard and R. Lobinski, Anal. Chem.,
68 (1996) 4135.
99 J. Szpunar, J. Bettmer, M. Robert, H. Chassaigne, K. Cammann, R. Lobinski and
O.F.X. Donard, Talanta, 44 (1997) 1389.
100 J.M. Ruiz, J. Szpunar and O.F.X. Donard, Sci. Total Environ., 198 (1997) 225.
101 H. Garraud, M. Robert, C.R. Quetel, J. Szpunar and O.F.X. Donard, Atom. Spectrosc.,
17 (1996) 183.
102 C.M. Tseng, A. deDiego, F.M. Martin and O.F.X. Donard, J. Anal. Atom. Spectrom.,
12 (1997) 629.
103 M.J. Vazquez, A.M. Carro, R.A. Lorenzo and R. Cela, Anal. Chem., 69 (1997) 221.
104 I.R. Pereiro, A. Wasik and R. Lobinski, J. Anal. Atom. Spectrom., 13 (1998) 743.
105 R.A. Lorenzo, M.J. Vazquez, A.M. Carro and R. Cela, Trends Anal. Chem., 18 (1999)
410.
106 M.J. Vazquez, M. Abuin, A.M. Carro, R.A. Lorenzo and R. Cela, Chemosphere, 39
(1999) 1211.
107 M. Abuin, A.M. Carro and R.A. Lorenzo, J. Chromatogr. A, 889 (2000) 185.
108 J.M. Bayona, Trends Anal. Chem., 19 (2000) 107.
109 Y. Cai, M. Abalos and J.M. Bayona, Appl. Organomet. Chem., 12 (1998) 577.
110 Y. Cai, R. Alzaga and J.M. Bayona, Anal. Chem., 66 (1994) 1161.
111 I. Fernandez Escobar and J.M. Bayona, Anal. Chim. Acta, 355 (1997) 269.
112 V. Lopez Avila, Y. Liu and W.F. Beckert, J. Chromatogr.A, 785 (1997) 279.
113 N.P. Vela and J.A. Caruso, J. Anal. Atom. Spectrom., 11 (1996) 1129.
114 J.M. Bayona, Qual. Assurance Environ. Anal., 17 (1995) 465.
115 Y. Liu, V. Lopez Avila, M. Alcaraz and W.F. Beckert, J. AOACInt., 78 (1995) 1275.
116 K.M. Attar, Appl. Organomet. Chem., 10 (1996) 317.
117 H. Emteborg, E. Bjorklund, F. Odman, L. Karlsson, L. Mathiasson, W. Frech and
D.C. Baxter, Analyst, 121 (1996) 19.
118 W. Holak, J. AOAC Int., 78 (1995) 1124.
119 R. Cela Torrijos, M. Miguens Rodriguez, A.M. Carro Diaz and R.A. Lorenzo Ferreira,
J. Chromatogr.A, 750 (1996) 191.
120 Y.C. Sun, J. Mierzwa, Y.T. Chung and M.H. Yang, Anal. Commun., 34 (1997) 333.
121 N.S. Simon, Int. J. Environ. Anal. Chem., 68 (1997) 313.
122 http://www.tmasc.com, 2001.
123 M.Abalos, J.M. Bayona and P. Quevauviller, Appl. Organomet. Chem., 12 (1998) 541.
124 S. Shawky, H. Emons and H.W. Durbeck, Anal. Commun., 33 (1996) 107.
125 J.C. Wu, Z. Mester and J. Pawliszyn, J. Anal. Atom. Spectrom., 16 (2001) 159.
126 J.C. Wu, Z. Mester and J. Pawliszyn, Anal. Chim. Acta, 424 (2000) 211.
127 J.A. Caruso, D.T. Heitkemper and C. B'Hymer, Analyst, 126 (2001) 136.
128 J.L. Gomez Ariza, D. Sanchez Rodas, I. Giraldez and E. Morales, Analyst, 125 (2000)
401.
129 G.L. Hu and X.R. Wang, Anal. Lett., 31 (1998) 1445.
963
130 Y.K. Chau, F. Yang and M. Brown, Anal. Chim. Acta, 338 (1997) 51.
131 M.B. delaCalleGuntinas, R. Scerbo, S. Chiavarini, P. Quevauviller and R. Morabito,
Appl. Organomet. Chem., 11 (1997) 693.
132 J. Pawliszyn, Solid Phase Microextraction: Theory and Practice. Wiley, New York
(NY), 1997.
133 Z. Mester, R. Sturgeon and J. Pawliszyn, Spectrochim. Acta B, 56 (2001) 233.
134 L. Moens, T. De Smaele, R. Dams, P. Van den Broeck and P. Sandra, Anal. Chem., 69
(1997) 1604.
135 B. He, G.B. Jiang and Z.M. Ni, J. Anal. Atom. Spectrom., 13 (1998) 1141.
136 B. He and G.B. Jiang, Fresen. J. Anal. Chem., 365 (1999) 615.
137 Z. Mester, R.E. Sturgeon and J.W. Lam, J. Anal. Atom. Spectrom., 15 (2000) 1461.
138 M. Guidotti, J. AOAC Int., 83 (2000) 1082.
139 M. Guidotti, G. Ravaioli and M. Vitali, HRCJ. High Resolut. Chromatogr.,22 (1999)
414.
140 Z. Mester and J. Pawliszyn, J. Chromatogr.A, 873 (2000) 129.
141 Z. Mester, G. Horvath, G. Vitanyi, L. Lelik and P. Fodor, Rapid Commun. Mass.
Spectrom., 13 (1999) 350.
142 T. Gorecki and J. Pawliszyn, Anal. Chem., 68 (1996) 3008.
143 J.P. Snell, W. Frech and Y. Thomassen, Analyst, 121 (1996) 1055.
144 Z. Mester, J. Lam, R. Sturgeon and J. Pawliszyn, J. Anal. Atom. Spectrom., 15 (2000)
837.
145 C.R. Jia, Y.Z. Luo and J. Pawliszyn, J. Microcolumn Sep., 10 (1998) 167.
146 T.P. Gbatu, K.L. Sutton and J.A. Caruso, Anal. Chim. Acta, 402 (1999) 67.
147 J.C. Wu, X.M. Yu, H. Lord and J. Pawliszyn, Analyst, 125 (2000) 391.
148 Z. Mester and J. Pawliszyn, Rapid Commun. Mass. Spectrom. 13 (1999) 1999.
149 Z. Mester, H. Lord and J. Pawliszyn, J. Anal. Atom. Spectrom., 15 (2000) 595.
150 B. Kolb, J. Chromatogr. A, 842 (1999) 163.
151 C.M. Tseng, A. DeDiego, F.M. Martin, D. Amouroux and O.F.X. Donard, J. Anal.
Atom. Spectrom., 12 (1997) 743.
152 F. Martin, F.M. Corrigan, O.F.X. Donard, J. Kelly, J.A.O. Besson and D.F. Horrobin,
Human. Exp. Toxicol., 16 (1997) 512.
153 D. Amouroux, E. Tessier, C. Pecheyran and O.F.X. Donard, Anal. Chim. Acta, 377
(1998) 241.
154 C.M. Tseng, A. deDiego, H. Pinaly, D. Amouraoux and O.F.X. Donard, J. Anal. Atom.
Spectrom., 13 (1998) 755.
155 A. deDiego, C.M. Tseng, T. Stoichev, D. Amouroux and O.F.X. Donard, J. Anal. Atom.
Spectrom., 13 (1998) 623.
156 C. Pecheyran, C.R. Quetel, F.M.M. Lecuyer and O.F.X. Donard, Anal. Chem., 70
(1998) 2639.
157 C. Pecheyran, D. Amouroux and O.F.X. Donard, J. Anal. Atom. Spectrom., 13 (1998)
615.
158 C.M. Tseng, A. De Diego, J.C. Wasserman, D. Amouroux and O.F.X. Donard,
Chemosphere, 39 (1999) 1119.
159 A. de Diego, C. Pecheyran, C.M. Tseng and O.F.X. Donard, in: A. Sanz-Medel (Ed.),
Flow Analysis with Atomic Spectrometric Detectors. Elsevier, Amsterdam, 1999, p.
375.
160 C.M. Tseng, D. Amouroux, I.D. Brindle and O.F.X. Donard, J. Envuiron. Monit., 2
(2000) 603.
161 D. Amouroux, E. Tessier and O.F.X. Donard, Environ. Sci. Technol., 34 (2000) 988.
162 C. Pecheyran, B. Lalere and O.F.X. Donard, Environ. Sci. Technol., 34 (2000) 27.
163 A. de Diego, C.M. Tseng, N. Dimov, D. Amouroux and O.F.X. Donard, Appl.
Organomet. Chem., 15 (2001) 490.
964
164 J. Sanz, M. Perez, M.T. Martinez and M. Plaza, Talanta, 50 (1999) 149.
165 J. Sanz, M. Perez, M.T. Martinez and M. Plaza, Talanta, 51 (2000) 849.
166 J. Sanz Asensio, M.T. Martinez Soria, M. Plaza Medina and M. Perez Clavijo, Anal.
Chim. Acta, 409 (2000) 171.
167 S. Cabredo, J. Galban and J. Sanz, Talanta, 46 (1998) 631.
168 R. Eiden, H.F. Scholer and M. Gastner, J. Chromatogr. A, 809 (1998) 151.
169 T. Kaise, M. Ogura, T. Nozaki, K. Saitoh, T. Sakurai, C. Matsubara, C. Watanabe and
K. Hanaoka, Appl. Organomet. Chem., 11 (1997) 297.
170 A.G. Howard and C. Salou, Anal. Chim. Acta, 333 (1996) 89.
171 J.L. Burguera, M. Burguera, C. Rivas and P. Carrero, Talanta, 45 (1998) 531.
172 A.G. Howard and C. Salou, J. Anal. Atom. Spectrom., 13 (1998) 683.
173 U. Pyell, A. Dworschak, E. Nitschke and B. Neidhart, Fresen. J. Anal. Chem., 363
(1999) 495.
174 J.Y. Cabon and N. Cabon, Fresen. J. Anal. Chem., 368 (2000) 484.
175 I.R. Pereiro, A. Wasik and R. Lobinski, Anal. Chem., 70 (1998) 4063.
176 M. Ceulemans and F.C. Adams, J. Anal. Atom. Spectrom., 11 (1996) 201.
177 A. Wasik, I.R. Pereiro, C. Dietz, J. Szpunar and R. Lobinski, Anal. Commun., 35
(1998) 331.
178 C. Dietz, Y. Madrid, C. Camara and P. Quevauviller, Anal. Chem., 72 (2000) 4178.
179 J. Feldmann, I. Koch and W.R. Cullen, Analyst, 123 (1998) 815.
180 J. Feldmann, E.M. Krupp, D. Glindemann, A.V. Hirner and W.R. Cullen, Appl.
Organomet. Chem., 13 (1999) 739.
181 J. Feldmann, J. Environ. Monit., 1 (1999) 33.
182 J. Feldmann and W.R. Cullen, Environ. Sci. Technol., 31 (1997) 2125.
183 J. Dedina, in: A. Sanz-Medel (Ed.), Flow Analysis with Atomic Spectrometric
Detectors. Elsevier, Amsterdam, 1999.
184 Z. Mester and P. Fodor, Spectrochim. Acta B, 52 (1997) 1763.
185 S.A. Pergantis, W. Winnik, E.M. Heithmar and W.R. Cullen, Talanta,44 (1997) 1941.
186 A.G. Howard, J. Anal. Atom. Spectrom., 12 (1997) 267.
187 J.H. Weber, Trends Anal. Chem., 16 (1997) 73.
188 S.N. Willie, Spectrochim. Acta B, 51 (1996) 1781.
189 Z. Mester and P. Fodor, J. Chromatogr.A, 756 (1996) 292.
190 Z. Mester, A. Woller and P. Fodor, Microchem. J., 54 (1996) 184.
191 S. Yalcin and X.C. Le, Talanta, 47 (1998) 787.
192 Z. Slejkovec, J.T. van Elteren and A.R. Byrne, Talanta, 49 (1999) 619.
193 J.L. Gomez Ariza, D. Sanchez Rodas, E. Morales, O. Herrgott and I.L. Marr, Appl.
Organomet. Chem., 13 (1999) 783.
194 N. Mikac, M. Branica, Y. Wang and R.M. Harrison, Environ. Sci. Technol., 30 (1996)
499.
195 X.M. Yu and J. Pawliszyn, Anal. Chem., 72 (2000) 1788.
196 P.W. Looser, M. Berg, K. Fent, J. Muhlemann and R.P. Schwarzenbach, Anal. Chem.,
72 (2000) 5136.
197 J.R. Encinar, J. Alonso and A. Sanz Medel, J. Anal. Atom. Spectrom., 15 (2000) 1233.
198 B.S. Tselentis and E.S. Tzannatos, Fresen. Environ. Bull., 9 (2000) 499.
199 S. Aguerre, C. Bancon Montigny, G. Lespes and M. Potin Gautier, Analyst, 125
(2000) 263.
200 H. Tao, R.B. Rajendran, C.R. Quetel, T. Nakazeto, M. Tominaga and A. Miyazaki,
Anal. Chem., 71 (1999) 4208.
201 A.A. Ansari, I.B. Singh and H.J. Tobschall, Sci. Total Environ., 223 (1998) 157.
202 C. Montigny, G. Lespes and M. PotinGautier, J. Chromatogr.A, 819 (1998) 221.
203 C.G. Arnold, M. Berg, S.R. Muller, U. Dommann and R.P. Schwarzenbach, Anal.
Chem., 70 (1998) 3094.
965
204 R. Ritsema, T. deSmaele, L. Moens, A.S. deJong and O.F.X. Donard, Environ. Pollut.,
99 (1998) 271.
205 C. CarlierPinasseau, G. Lespes and M. Astruc, Environ. Technol., 18 (1997) 1179.
206 D. Milde, Z. Plzak and M. Suchanek, Collection Czech. Chem. Commun., 62 (1997)
1403.
207 C. Gerbersmann, M. Heisterkamp, F.C. Adams and J.A.C. Broekaert, Anal. Chim.
Acta, 350 (1997) 273.
208 S. Girousi, E. Rosenberg, A. Voulgaropoulos and M. Grasserbauer, Fresen. J. Anal.
Chem., 358 (1997) 828.
209 C. CarlierPinasseau, G. Lespes and M. Astruc, Appl. Organomet. Chem., 10 (1996)
505.
210 S. Tutschku, S. Mothes and R. Wennrich, Fresen.J. Anal. Chem., 354 (1996) 587.
211 R. Allabashi, J. Rendl and M. Grasserbauer, Fresen. J. Anal. Chem., 360 (1998) 723.
212 R.G. Fernandez, M.M. Bayon, J.I.G. Alonso and A. Sanz Medel, J. Mass Spectrom., 35
(2000) 639.
213 R. Morabito, P. Massanisso and P. Quevauviller, TrendsAnal. Chem., 19 (2000) 113.
214 R. Zufiaurre, B. Pons and C. Nerin, J. Chromatogr. A, 779 (1997) 299.
215 U.M. Gruter, J. Kresimon and A.V. Hirner, Fresen. J. Anal. Chem., 368 (2000) 67.
216 K. Bergmann and B. Neidhart, J. Sep. Sci., 24 (2001) 221.
217 M. Heisterkamp and F.C. Adams, J. Anal. Atom. Spectrom., 14 (1999) 1307.
218 Y. Cai, S. Monsalud and K.G. Furton, Chromatographia,52 (2000) 82.
219 Y. Cai, S. Monsalud, R. Jaffe and R.D. Jones, J. Chromatogr.A, 876 (2000) 147.
220 S. Diez, L. Ortiz and J.M. Bayona, Chromatographia,52 (2000) 657.
221 G. Binato, G. Biancotto, R. Piro and R. Angeletti, Fresen. J. Anal. Chem., 361 (1998)
333.
222 1. FernandezEscobar, M. Gibert, A. Messeguer and J.M. Bayona, Anal. Chem., 70
(1998) 3703.
223 J.R. Baena, S. Cardenas, M. Gallego and M. Valcarcel,Anal. Chem., 72 (2000) 1510.
224 M.B. delaCalleGuntinas and F.C. Adams, J. Chromatogr.A, 764 (1997) 169.
225 G.B. Jiang and F.C. Adams, Anal. Chim. Acta, 337 (1997) 83.
226 P. Smichowski and S.V. Farias, Microchem. J., 67 (2000) 147.
227 Z. Mester and R.E. Sturgeon, J. Anal. Atom. Spectrom., 16 (2001) 470.
228 X.W. Guo and X.M. Guo, Anal. Chim. Acta, 330 (1996) 237.
229 Z. Mester, R.E. Sturgeon, J.W. Lam, P.S. Maxwell and L. Peter, J. Anal. Atom.
Spectrom., 16 (2001) 1313.
230 A.L. Molinero, L. Martinez, A. Villareal and J.R. Castillo, Talanta, 45 (1998) 1211.
231 A.L. Molinero, A. Morales, A. Villareal and J.R. Castillo, Fresen. J. Anal. Chem., 358
(1997) 599.
966
Chapter29
29.1 INTRODUCTION
As shown in Fig. 29.1, a sample can be treated with conventional sample cleanup
and preconcentration methods such as solid phase extraction (SPE) and liquid-
liquid extraction (LLC). Then, the prepared sample (usually greatly concen-
trated in comparison with the original sample) is injected into a capillary by
either hydrodynamic or electrokinetic injection methods, and CE separation and
determination are followed. This off-line sample cleanup and preconcentra-
tion-CE has been used for many practical samples. For example, different struc-
tural types of industrial dye samples can be preconcentrated on a SAX Bond Elut
cartridge [10]. Then the dyes are eluted by mixture of HCl and MeOH. After
evaporating the solvent, the dyes are re-dissolved in 0.5 ml running buffer of CE
and are analyzed by CE. This off-line SPE concentration could lower the detec-
tion limit by 50-fold [10]. Various phenols in wastewater included in the 8041
U.S. Environmental Protection Agency (EPA) method can be preconcentrated in
a styrene-divinylbenzene SPM cartridge, then separated and determined by
non-aqueous CE using a buffer of ammonium acetate dissolved in N-methyl-
formamide-acetonitrile [11]. Quinolone antibiotics in plasma samples can also
be separated and quantitatively determined by CE with SPE pre-treatment [12].
The plasma matrix is removed, and the quinolone antibiotics are preconcen-
trated in a SPE cartridge [12]. In addition to the SPE cartridges, a SPE disk has
also been used for extraction ethylenediaminetetraacetic acid (EDTA) in envi-
ronmental water as nickel EDTA chelate. Then, the preconcentrated nickel
EDTA chelate is quantitatively determined by CE-MS [13].
1 Clean up and 2 Injection 3 CE
Preconcentration
(Hydrodynamic injection,
Electroldnetic injection)
+Uroj
968
Liquid-liquid extraction has also been used for the off-line cleanup and
preconcentration of CE. Fluorescently labelled amino acids and peptides are
extracted and concentrated into an organic phase such as acetophenone, then
they are separated and determined by CE with fluorescence detection [14]. A
membrane, gel, or hollow fiber also can be used for the off-line preconcentration
of CE [15-17]. Water in a hollow fiber filled with sample solution can be removed
by evaporation or Donnan effect; thus, sample in the fiber is concentrated into a
zone with a length of only a few mm [17]. Then, the concentrated sample zone
can be injected into a capillary for following CE determination. The method can
provide a high concentration factor up to 1000-fold for protein samples [17].
Isotachophoretic (ITP) has also been used for the off-line sample pre-treatment
of CE [18].
Although the above off-line preconcentration methods can lower the concen-
tration detection limit of CE -100-fold or even -1000-fold, the main disadvan-
tage is inefficient use of the concentrated sample because usually less than 1% of
the concentrated sample is injected into the capillary for further CE determina-
tion. Also, sample loss and contamination may occur during the preconcentra-
tion process, especially for preparation of a minute volume of sample. Therefore,
on-line sample cleanup and preconcentration methods are desirable for CE.
969
A Frit or filter
X - -o -L - - ww 9 Ha w
-,vaa
v ww "
B
A).]1nM terbutaline with enrich'rhi
.
<
0 1 2 3
Migration time (min)
Fig. 29.2. Schematic of the compact on-line coupling of SPE-CE (A) and CE electropherograms of
terbutaline (B). In (B), A and B are electropherograms of 1 nM and 10 nM terbutaline with
on-line SPE concentration, and C is an electropherogram of 10 /M terbutaline without on-line
SPE concentration, respectively. For details see Ref. [21]. (B) is reprinted with permission from
Ref. [21].
terbutaline while detection limit is only 4.4 AM without the SPE enrichment
[21]. A similar SPE-CE has also been used for peptide separation [23]. Peptides
from a sample matrix are adsorbed on a reversed phase resin (C-8 or C-18)
sorbent and subsequently released for CE separation by injection of organic
elutant. This technique allows the peptide separation on the level of -10 ng/ml
[23]. Membrane materials with cationic-exchange or hydrophobic characteris-
tics also have been used for on-line preconcentration and cleanup of CE for pro-
tein and peptide samples [24,25]. This compact SPE-CE lowers the detection
limit of CE by a factor of several hundreds or even thousands.
Solid phase microextraction (SPME) has also been coupled to CE [26]. Figure
29.3A shows a coupling attachment for SPME to CE. A 1.5 cm capillary segment
(180 m i.d., 380 m o.d.) is closely attached to a CE capillary in a heat shrink-
able tube. A glass fiber of 150 /Am coated with poly(dimethylsiloxane) is used for
concentrating polycyclic aromatic hydrocarbons (PHAs). The glass fiber with
absorbed analytes is immersed into the capillary segment, and the absorbed
analytes are directly released into the CE buffer stream, and then enter into the
capillary. Neutral PHAs are separated by CE with borate CE running buffer con-
taining cyclodextrin and its derivatives. Figure 29.3B shows one example of the
CE separation. It can be seen that SPME-CE greatly lowers the detection limits
of CE for PHAs. The detection limit of pyrene is 8 ppb [26].
Furthermore, a SPME-CE interface that realizes a zero-dead volume con-
nection between SPME and CE has also been proposed recently (Fig. 29.3C) [27].
A 40 Am fiber coated with polyacrylate film is used for extraction and concentra-
970
A E
Separation capillary
50 pm id, 350 pm od
c
Heat shrinkable
tubing 2 cm a
a
a
Glassfiber150pm
diameter with
absorbed analytes,
A__
--
Nylon coating ' 6'' (b
on glass fiber
4
2 7 1 t
12 14 16 I 2 22
Time (min)
C -HV
Fig. 29.3. Schematics of on-column SPME-CE (A, C) and CE separation results of PAHs (B). In
(B), (a) and (b) are electropherograms obtained without and with on-column SPME concentra-
tion (by system shown as A), respectively. Standard sample solution of EPA610 is 40-fold (a) and
4000-fold (b) diluted. Peaks 1-15 are dibenz[a,h]anthracene, acenaphthylene, acenaphthene,
naphthalene, fluorene, anthracene, phenanthrene, chrysene, benz[a]anthracene, benzo(k)fluor-
anthene, fluoranthene, benzo[a]pyrene, pyrene, benzo[b]fluoranthene, indeno[1,2,3-cd]pyrene,
respectively. For details see Ref. [26]. (A) and (B) are reprinted with permission from Ref. [26].
tion of phenol samples. Then, the fiber is directly immersed into the inlet end of
a 75 C/m CE capillary. It has been used for the determination of various phenol
compounds.
Recently, SPME-CE has been further applied to investigate the chemical
composition of small living objects such as portions of plants and organs of
insects by combining with on-fiber fluorescence derivatization techniques and
laser-induced fluorescence detection [28]. Firstly, an SPME fiber was dipped
into a vial containing fluorescent derivatization reagent 4-fluoro-7-nitro-2,1,3-
benzoxadiazole (NBD-F, 2-3 mg/ml in ethanol) solution for 10 min with mag-
netic stirring at 1000 rpm. The fluorescent derivatization reagent was adsorbed
on the fiber. Secondly, the fiber was immersed into a small living object such as a
grape for 20 s. Amino acids such as glutamic acid in the grape was extracted and
concentrated on the fiber. Thirdly, the fiber was inserted into the headspace of a
971
4-ml amber vial containing 1 ml of triethylamine (TEA) at 60°C water bath, so
that the fluorescent derivatization reaction of amino acids occurred easily.
Finally, the fiber was immersed into the zero-dead volume SPME-CE interface,
followed by desorption with methanol and micellar electrokinetic capillary chro-
matography (MECC) separation with laser-induced fluorescence detection.
where L, Vof, Veps, and Ver b are length of sample solution plug, electroosmotic
flow, electrophoretic velocities in sample plug and background buffer, respect-
ively. Then, the concentration factor F is expressed as follows:
F = (Veo f + Veps)/(Veof + Vepb) (29.2)
Equation (29.2) shows that a high concentration factor F can be obtained with a
small Vepb and large Vep,. In order to achieve large Veps and small Vepb, sample is
usually dissolved in a much lower conductivity buffer than the running buffer or
even in pure water. In some cases, organic solvent is added to the sample to
decrease the conductivity of the sample plug.
972
One merit of sample stacking is that it can inject a long sample plug into the
capillary. In some experiments, even more than half of the capillary is filled with
sample plug [32]. Either electroosmotic flow is adjusted by adding a modifier [33]
into the buffer or polarity switching [34] is employed after finishing the sample
stacking so that the long water plug will not affect the separation efficiency.
Sample stacking can also be accomplished during the electrokinetic injection
process when the inlet capillary is immersed into a sample vial with a low con-
ductivity matrix. This stacking mode is also called field amplification sample
injection [35]. Injection of a short plug of water before the sample is reported to
be able to enhance the electric field at the injection point and thus increase the
stacking effect [36].
The stacking technique has been further extended to concentration of neu-
tral analytes in MEKC separation [36,37]. A concentration factor up to 106 has
been reported [38]. This technique has been well reviewed [39]. Similar to sam-
ple stacking, other concentration methods such as sample self stacking [40],
acetonitrile stacking [41], sweeping [42], and the use of pH junction [43] are also
based on the moving velocity difference of ions across the boundary of the sample
plug and the running buffer.
29.3.2.2 CITP-CE
Capillary isotachophoresis is a unique electrophoresis mode where a sample plug
is sandwiched between a leading electrolyte (with a faster mobility than any ion
in the sample) and a terminating electrolyte (with a lower mobility than any ion
in the sample) in a capillary. When an electric field is applied across the capillary,
the resulting field is not homogeneous throughout the capillary. Separation
occurs between the boundaries based on the individual ion mobilities. As a
result, sample ions are separated in consecutive zones according to their mobili-
ties, and all zones migrate at the same velocity. An ion diffusing out of its zone
speeds up or slows down, depending on the velocity of the neighbouring zone it
encounters, thereby rejoining its focused zone. This CITP can also be on-line
coupled to CE.
Basically, there are two approaches to coupling CITP to CE. A simple
approach is to carry out CITP and CE in a single capillary [44]; this approach is
also called transient CITP-CE [9]. After filling the running buffer into the capil-
lary, a leading buffer (the running buffer can be used as the leading buffer in
some case), sample solution, and terminating buffer are subsequently injected
into the capillary from the inlet end. After applying a voltage across the two ends
of the capillary, CITP occurs in the first part including the inlet end of the capil-
lary. Then, concentrated sample zones by the CITP are separated by CE in the
remaining part of the capillary. This CITP-CE has been used for separation of
sugars [45], proteins [46] and other biochemical samples; concentration factor
can be up to 1000-fold [47]. The drawback of this technique is that CE resolution
might be decreased because of overloading in some cases.
Another approach is to carry out CITP and CE in two different capillaries
973
HighVoiage Power Supply
I ,I
LE
TE EIzE ILE 3 TE =t
CZED
'tD G
LE
2 E iLE 4 TE LE
Fig. 29.5. Schematics of instrumentation for comprehensive CITP-CE (A) and its operation
principle (B). In (A), BR1, BR2, and BR3 represent buffer reservoirs 1, 2, and 3, respectively. V1
and V2 represent valves 1 and 2, respectively. In (B), CZED is CE detector; TE, LE, s represent
terminating electrolyte, leading electrolyte, and sample, respectively. For details see Ref. [51].
Reprinted with permission from Ref. [51].
that are connected either by a T-junction [48,49] or by immersing the inlet end of
the CE capillary (with small diameter) into the CITP capillary (with large inter-
nal diameter) [50]. Firstly, a voltage is applied across the CITP capillary for car-
rying out ITP concentration of sample ions. When the concentrated analyte
zones move to the inlet end of the CE capillary, a voltage is applied across the
CITP and CE capillaries for injecting the concentrated zones into the CE capil-
lary. Then CE separation is carried out. This CITP-CE has been used for separa-
tion of small anions [49], amino acids [51], and so on, a concentration factor as
high as 10 5 -10 5 has been reported [49]. However, this CITP-CE can only analyze
a few concentrated zones at a time to prevent CE overloading. Moreover, two
detectors are required (one for CITP and the other for CE). The CITP detector is
used to determine when the concentrated bands reach the injection end of the
capillary. The timing for injection of the concentrated zone into the CE capillary
is important [52].
Recently, a comprehensive CITP-CE coupling configuration (Fig. 29.5) has
been reported [52]. Firstly, the sample is focused in the CITP capillary. Second-
ly, the focused sample zone is injected into the CE capillary by electrokinetic
injection. Thirdly, the remaining sample zones are forced to move back away
from the injection end of the CE capillary by injecting leading electrolyte from
the syringe pump. Fourthly, CE separation is carried out. Comprehensive
974
CITP-CE involves repetitive runs of the above four steps. Since only small por-
tions of the concentrated zones are sequentially injected for CE separation, the
problem of overloading does not exist. The system has been evaluated using a
mixture of angiotensins. Under optimized conditions, a detection limit of ap-
proximately 5 nM can be achieved by injecting 10/ l of angiotensin solution [52].
~B ~ ~~ 2 32
Sample concentration :0.05 mg/ml I Abs
Without concentration
Fig. 29.6. Schematic of the on-column electrophoretic concentration with a hollow fiber (A) and
CE results of proteins (B). Peaks 1, 2, and 3 are cytochrome c, lysozyme, and ribonuclease A,
respectively.
975
buffers
A buffer C aay side
porous membrane 3
0
analyte side channel 100 P I
injection tee
of
: IaIwaste
separation
channel
Waste r
i
l.. /
/D/ SE
....
analyte main channel
(I f/:1X
rI
.side.channel
separation channel
Top view
PM
" sodium silicate
bonding interface
cover plat
substrate
50
Cross setionview time [Is]
Fig. 29.7. Schematics of a porous membrane microchip for DNA concentration and electro-
phoretic analysis (A) and porous membrane structure (B). (C) CCD images of analyte
concentrated for different times and injection after concentration. (D) Successive electro-
pherograms of a DNA PCR marker (50, 150, 300, 500, 750, and 1000 bp) after 150 s, 200 s, and
250 s of preconcentration. For details see Ref. [54]. Reprinted with permission from Ref. [54].
channel are separated from each other by a distance of 3-12 mm and connected
by a thin porous silicate layer. The porous silicate layer allows the passage of cur-
rent to establish an electrical connection between the separated channels but
prevents large molecules (e.g., DNA) from traversing the porous silicate layer.
Therefore, when a voltage is applied between the analyte and side reservoirs,
DNA samples will be concentrated in front of the porous membrane (Fig. 29.7C).
Then, the concentrated DNA samples are injected in the separation channel by
switching the voltage. Figure 29.7D shows that the concentration factor depends
on the concentration time.
Comparing with stacking and CITP, one merit of the on-column electropho-
retic concentration method is that it does not need to use discontinuous buffers,
i.e., the sample can be directly prepared in the CE running buffer.
976
29.3.3 On-line membrane cleanup for CE
A porous membrane or a hollow fiber can also be used for cleanup of analytes
from the sample matrix to prevent particulate material or high molecular mass
matter from entering the CE capillary. Figure 29.8 shows one configuration of
the coupling between capillary and tubular membrane. The tubular membrane
can be used as sampling interface. Sample solution flowed into the jacket is
introduced by diffusion and/or permeation through the membrane. The jacket
Sample Waste
-' \Membe
Capillary Capillay
F . l out
Dram Capilln
a r
I nA F 5 5 5 0 5
0 5 0 s 0 5 0o 5 6 D 5
Tite (anin)
Fig. 29.9. Schematic of the on-line microdialysis-CE-electrochemical detection system (A) and
electropherograms of transdermally delivered nicotine in microdialysate after administration of
nicotine patch (B). Details see Ref. [60]. Reprinted with permission from Ref. [60].
977
can be connected to a valve that allows easy replacement of the solution in the
jacket. This membrane sampling allows CE direct determination of gaseous
samples, low molecular weight constituents in blood plasma, ionisable/non-ionic
solutes in wastewater [561.
A dialysis system can also be on-line connected to CE by a flow-gating
approach [57,58] or via a micro-injection valve-based interface [59,601. Figure
29.9 shows one example of the on-line coupling of in ivo microdialysis with
CE/electrochemicaldetection [61]. As shown in Fig. 29.9A, a microdialysis probe
is immersed into a rabbit body. Outlet of the microdialysis probe is connected to
a micro-injection valve (60 nl) using a capillary. Injection of analytes into the CE
capillary is accomplished electrokinetically by applying the separation voltage at
the gap junction interface. Figure 29.9B shows monitoring results of nicotine
concentration in the animal after administration of the nicotine patch.
Figure 29.10 shows one example of on-line dialysis-CE coupling with a
FIA-CE interface for determination of small anions [62]. The sample is pumped
into the dialysis unit, and small anions are dialysed into an acceptor stream of
water. The acceptor steam is led into an injector loop. When injection takes place
the sample is transferred to the electrolyte. The injected sample then passes by
the end of the capillary situated in the FIA interface and a small fraction is
electrokinetically introduced into the capillary where separation takes place.
Experimental results show that this system can be used for determination of
small anions in various drinks such as milk with good reproducibility.
Capillary isoelectric focusing (CIEF) is known to be an effective tool with
high separation resolution in biomolecular analysis [63]. In particular, for the
recently developed whole column imaged CIEF [64,65], it greatly decreases anal-
ysis time, and enables direct observation of the isoelectric focusing and separa-
tion process. However, the protein sample is usually required to desalt before the
CIEF experiment. Figure 29,11 shows an on-line desalting for CIEF with a hol-
low fiber [66]. The protein sample is injected into the hollow fiber, which is
immersed into an ampholyte solution for desalting. After desalting, the protein
sample is flowed into the capillary and whole column imaging CIEF is carried
out [661.
978
a) (
PTFE tubing
- ·
b)
3 4 5 6 7 8 9 rmin
Fig. 29.10. Schematics of the on-line dialysis-FIA-CE system (a) and CE results (b). In (a) A is a
cross-sectional view of the FIA-CE interface. S, A, E, M, D, V1, V2, W, C, Pt, and HV represent
sample, acceptor stream, electrolyte, dialysis membrane, UV detector, injection valve in filling
position, injection valve in injection position, waste, capillary, platinum electrodes, and high-
voltage supply, respectively. In (b), A and B are electropherograms of 24 anion standard mixture
without and with dialysis, respectively. Peaks 1-20 are S2032 Cl-, SO,', NO 3 , oxalate, HS,
SO3, citrate, maleinate, fumarate, F-, tartrate/succinate/formate/malate, HPO4,% HCO3 ,
acetate, OH, propionate, lactate, butyrate, and benzoate, respectively. For details see Ref. [61].
Reprinted with permission from Ref. [61].
979
ailaries
Ampholytes '
solution i
Fig. 29.11. Schematic of on-line desalting with a dialysis hollow fiber for CIEF. For details see
Ref. [65]. Reprinted with permission from Ref. [65].
29.5 SUMMARY
REFERENCES
980
6 X. Xi and E.S. Yeung, Appl. Spectrosc., 45 (1991) 1199.
7 X.-Z. Wu, A. Hosaka, E. Kobayashi and T. Hobo, J. Chromatogr.A, 726 (1996) 205.
8 Dankova, D. Kaniansky, S. Fanali and F. Ivanyi, J. Chromatogr.A, 838 (1999) 31.
9 J.R. Veraart, H. Lingeman and U.A.Th. Brinkman, J. Chromatogr. A, 856 (1999)
483.
10 L. Farry, D.A. Oxspring, W.F. Smyth and R. Marchant, Anal. Chim. Acta, 349 (1997)
221.
11 S. Morales and R. Cela, J. Chromatogr.A, 896 (2000) 95.
12 M. Hernandez, F. Borrull and M. Calull, J. Chromatogr., B: Biomed. Sci. Appl., 742
(2000) 255.
13 R.L. Sheppard and J. Henion, Electrophoresis, 18 (1997) 287.
14 W. Zhan, T. Wang and S.F.Y. Li, Electrophoresis,21 (2000) 3593.
15 R. Zhang and S. Hjerten, Anal. Chem., 69 (1997) 1585.
16 S. Hjerten, J.-L. Liao and R. Zhang, J. Chromatogr.A, 676 (1994) 409.
17 J.-L. Liao, R. Zhang and S. Hjerten, J. Chromatogr.A, 676 (1994) 421.
18 K. Dusan, K. Eva, M. Vlasta and M. Marian, Electrophoresis,18 (1997) 260.
19 A.J.J. Debets, M. Mazereeuw, W.H. Voogt, D.J. van Iperen, H. Lingeman, K.-P. Hupe
and U.A.Th. Brinkman, J. Chromatogr.A, 608 (1992) 151.
20 I. Morita and J.I Sawada, J. Chromatogr.A, 641 (1993) 375.
21 M. Petersson, K. Wahlund and S. Nilsson, J. Chromatogr.A, 841 (1999) 249.
22 A.J.J. Debets, M. Mazereeuw, W.H. Voogt, D.J. van Iperen, H. Lingeman, K.-P. Hupe
and U.A.Th. Brinkman, J. Chromatogr.A, 608 (1992) 151.
23 M.A. Strausbauch, J.P. Landers and P.J.Wettstein, Anal. Chem., 68 (1996) 306.
24 S. Naylor and A.J. Tomlinson, Biomed. Chromatogr., 10 (1996) 325.
25 E. Rohde, A.J. Tomlinson, D.H. Johnson and S. Naylor, J. Chromatogr.B, 713 (1998)
301.
26 A.L. Nguyen and J.H.T. Luong, Anal. Chem., 69 (1997) 1726.
27 C-W. Whang and J. Pawliszyn, Anal. Commun., 35 (1998) 353.
28 A. So, Sampling Small Object with Solid Phase Microextraction. Master Thesis, 2000,
University of Waterloo, Ontario, Canada.
29 R.L. Chien and D.S. Burgi, Anal. Chem., 64 (1992) 489A.
30 D.S. Burgi and R.L. Chien, Anal. Chem., 63 (1991) 2042.
31 R.L. Chien, Anal. Chem., 63 (1991) 2866.
32 Y. He and H. K. Lee, Anal. Chem., 71 (1999) 995.
33 J.P. Quirino and S. Terabe, Electrophoresis,21 (2000) 355.
34 M. Albert, L. Debusschere, C. Demesmay and J.L. Rocca, J. Chromatogr. A, 757
(1997) 281.
35 R.L. Chien and D.S. Burgi, J. Chromatogr., 559 (1991) 141.
36 Z. Liu, P. Sam, S. Sirimanne, P. McClure, J. Grainger and D. Patterson, J. Chroma-
togr. A, 673 (1994) 126.
37 J.P. Quirino and S. Terabe, J. Chromatogr.A, 781 (1997) 119.
38 J.P. Quirino and S. Terabe, Anal. Chem., 72 (2000) 1023.
39 J.P. Quirino and S. Terabe, J. Cap. Electrophor., 4 (1997) 233.
40 P. Gebauer, W. Thormann and P. Bocek, J. Chromatogr., 608 (1992) 47.
41 Z.K. Shihabi, J. Cap. Electrophor., 2 (1995) 267.
42 J.P. Quirino and S. Terabe, Science, 282 (1998) 465.
43 P. BritzMckibbin, J. Wong and D.D.Y. Chen, J. Chromatogr.A, 853 (1999) 535.
44 F. Foret, E. Szoko and B.L. Karger, J. Chromatogr., 608,(1992) 3.
45 C.-C. Wang, W.P. McCann and S.C. Beale, J. Chromatogr.B, 676 (1996) 19.
46 J. Bergmann, U. Jaehde, M. Mazereeuw, U.R. Tjaden and W. Schunack, J. Chroma-
togr. A, 734 (1996) 381.
47 J.-L. Liao, R. Zhaang and S. Hjerten, J. Chromatogr.A, 676 (1994) 421.
981
48 A.W. Moore Jr. and J.W. Jorgenson, Anal. Chem., 67 (1995) 3456.
49 D. Kaniansky, I. Zelensky, A. Hybenova and F.I. Onuska, Anal. Chem., 66 (1994)
4258.
50 N.J. Reinhoud, A.P. Tinke, U.R. Tjaden, W.M.A. Niessen and J. Van der Greef, J.
Chromatogr., 627 (1992) 263.
51 D. Kaniansky and J. Marak, J. Chromatogr., 498 (1990) 191.
52 S. Chen and M.L. Lee, Anal. Chem., 72(2000) 816.
53 X.-Z. Wu, A. Hosaka and T. Toshiyuki, Anal. Chem., 70 (1998) 2081.
54 X.-Z. Wu and 'T. Kasashima, Anal. Sci., 16 (2000) 329.
55 J. Khadurina, S.C. Jacobson, L.C. Waters, R.S. Foote and J.M. Ramsey, Anal. Chem.,
71 (1999) 1815.
56 L. Bao and P.K. Dasgupta, Anal. Chem., 64 (1992) 991.
57 M.W. Lada and R.T. Kennedy, Anal. Chem., 68 (1996) 2790.
58 M.W. Lada, G. Schaller, M.H. Carriger, T.W. Vickroy and R.T. Kennedy, Anal. Chim.
Acta, 307 (1995) 217.
59 B.L. Hogan, S.M. Lunte, J.F. Stobaugh and C.E.Lunte, Anal. Chem., 66 (1994) 596.
60 C.E. Lunte, D.O. Scott and P.T. Kissinger, Anal. Chem., 63 (1991) 773A.
61 J. Zhou, D.M. Heckert, H. Zuo, C.E. Lunte and S.M. Lunte, Anal. Chim. Acta, 379
(1999) 307.
62 P. Kuban and B. Krlberg, Anal. Chem., 69 (1997) 1169.
63 S. Hjerten and M. Zhu, J. Chromatogr., 346 (1985) 265.
64 J. Wu and J. Pawliszyn, Am. Lab., 26 (1994) 48.
65 J. Wu and J. Pawliszyn, Anal. Chem., 64 (1992) 224.
66 J. Wu and J. Pawliszyn, Anal. Chem., 67 (1995) 2010.
67 K. Gevaert and J. Vandekerckhove, Electrophoresis,21 (2000) 1145.
68 H. Zhang, P. Andren and R. Caprioli, J. Mass Spectrom., 30 (1995) 1768.
69 P. Courchesne and S. Patterson, BioTechniques, 22 (1997) 244.
70 K. Gevaert, H. Demol, M. Puype, D. Broekaert, S. Deboeck, T. Houthaeve and J.
Vandekerckhove, Electrophoresis, 18 (1997) 2950.
71 C. Bratt, C. Lindberg and G. Marko-Varga, J. Chromatogr., 909 (2001) 279.
72 J. Gobom, M. Schuerenberg, M. Mueller, D. Theiss, H. Lehrach and E. Nordhoff,
Anal. Chem., 73 (2001) 434.
73 F. Han and S.J. Lillard, Anal. Chem., 72 (2000) 4073.
74 S.N. Krylov, E. Arriaga, Z. Zhang, N.W.C. Chan, M.M. Palcic and N.J. Dovichi, J.
Chromatogr. B, 741 (2000) 31.
75 E.S. Yeung, J. Chromatogr.A, 830 (1999) 243.
76 S. Vaidyanathan, J. Rowland, D. Kell and R. Goodancre, Anal. Chem., 73 (2001) 4134
982
Chapter30
30.1 INTRODUCTION
It is not possible to cover the entire scope of progress in membrane technologies
in a single chapter since, nowadays, they touch almost all aspects of human life:
e.g., membrane technologies are penetrating rapidly into biotechnology and bio-
medical applications. Serious attempts are currently being made to develop
membranes that allow enhancement in efficiencies of fuel cells and batteries.
The author's experience is limited to research in the area of conventional mem-
branes and membrane separation processes, hence this article is focused on
progress made in reverse osmosis, ultrafiltration, nanofiltration, microfiltra-
tion, gas and vapor separation and pervaporation. However, even in this rather
limited area of membrane applications, remarkable progress has been achieved;
i.e. recovery of drinkable water from seawater has been increased by developing
high pressure resistant membranes; membranes operable at ultra-low operating
pressures have been developed; performance of gas separation membranes has
been significantly improved; inorganic membrane module for pervaporation has
been commercialised; a number of new techniques and instruments are being
used daily to characterize the membrane surface, which will have impact not
only on separation membrane performance but also on biotechnology and bio-
medical applications of membrane surface in the future; new methods for the
measurement of nano-pores have been developed and used; transport equations
of charged membranes have enabled the prediction of nanofiltration mem-
branes; and a new hybrid system is presently used for the treatment of heavily
polluted wastewater. This chapter also includes a brief outline of a new extrac-
tion method using reverse micelles that seems to have potential in applications
in biotechnology.
984
A
0
E
E
a
.
.
no
Fig. 30.1. Flux versus operating time for conventional and newly developed high pressure RO
membrane (from Ref. [1]).
COCI COCI
TMC A TPC
0 0 0, 0\ Organic Layer
Highly rosslinked Aromatic olyamide Membrane
at the operating pressure of 9.0 MPa. On the other hand, the compaction of the
novel NTR-70SWH membrane was moderate even at an operating pressure of
9.0 MPa.
Reverse osmosis membranes that exhibit high fluxes at low operating pres-
sures are also prepared by interfacial poly-condensation [2]. As shown in Fig.
30.2, a mixture of aromatic dicarboxylic acid chloride and tricarboxylic acid chlo-
ride dissolved in an organic solvent, which is immiscible to water, is brought into
contact with an aromatic diamine dissolved in water. An interfacial poly-
condensation will take place in situ, forming a very thin layer of aromatic poly-
amide at the top of the support layer. This is a well established method of prepar-
ing a thin film composite membrane. In a new approach, some additives were
985
Low Pressure RO-I Low Pressure RO-i Ultra Low Pressure RO Super UftraLow Pressure RO
Fig. 30.3. Cross-sectional pictures of newly developed ultra-low pressure RO membranes (from
Ref. [2]).
99.9
99.8
1 99.5
t 99.0
95.0
90.0
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Normalized water flux, m3/m2 day MPa
Fig. 30.4. Progress in the performance of RO membrane for seawater desalination (from Ref.
[2]).
brought into the monomer solution to increase the interfacial area. Figure 30.3
shows cross-sectional pictures taken near the surfaces of advanced low-pressure
reverse osmosis membranes. The increase in the effective surface area can be
observed in the figure, which resulted in an increased membrane flux. Figure
30.4 illustrates the progress made in the performance of reverse osmosis mem-
brane during the past three decades. Remarkable progress has been seen both in
selectivity and water flux since the 1970s. In particular, an order of magnitude
increase has been achieved in water flux of reverse osmosis membranes, which
enabled RO operation under ultra-low pressures. Currently, no limit is seen in
the horizon for the improvement of the RO membrane flux.
There is another interesting report in which the performance of a thin film
composite membrane was enhanced by surface modification [3]. Mukherjee et al.
applied surface chemical reaction on top of an aromatic polyamide thin film com-
posite membrane (FT-30). Membranes were soaked at controlled temperatures
986
in solutions of various concentrations of hydrofluoric acid (HF), fluorosililic acid
(FSA) and their mixtures. The membranes were taken out after various inter-
vals of time, rinsed with deionized water and then tested for their RO perfor-
mance. A dramatic effect of the surface modification on the membrane
performance was observed when the membrane was soaked in 15 wt.% HF solu-
tion for 7 days. The flux increased from 3.5 1/m 2 h to 18 1/m 2 h while NaCl separa-
tion increased from 94.5 to 95.3% under the same operating pressure of 250 psig
(1724 kPag). According to X-ray photoelectron spectroscopy (XPS), the fluorine
ratio (F/C) at the membrane surface increased from zero of fresh FT-30 mem-
brane to 0.012 of a membrane immersed in 15 wt.% HF solution for 4 days. It fur-
ther increased to 0.044 when the membrane was immersed in 15 wt.% HF
solution for 75 days. The SEM picture revealed the thinning down of the polymer
network of the selective skin layer due to the etching of the surface by HF. Thus,
both the chemical modification of the surface and the thinning of the selective
skin layer contributed to the dramatic increase of the flux, while NaCl separa-
tion was practically unchanged.
987
100
: :
":.']'T;' tradeoff f
T'"1997 "'// , ...
...
for OMS membl no
.. ....
: ::..... o si
0u
0
1
104 102 10' 10° 10' 10' 10' 10o
Permeability of 02 [Barrer]
Fig. 30.5. Oxygen/nitrogen selectivity versus oxygen permeability (from Ref. [41).
988
TABLE 30.1
Permeance data for fluorinated polyimide membranes (from Ref. [5])
Membrane Permeance (GPU a ) Permeance ratio
CO2 02 CO,/CH,4 O
2 N2
6FDA-6FAP Asymmetric 1490 345 43 5.2
6FDA-6FAP Denseb 0.75 0.16 36 4.5
6FDA-DDS Asymmetric 0.57 0.15 144 12
6FDA-DDS Dense b 0.0032 0.0048 116 8.5
al1GPU = 1 x 10 - 6 cm 3 STP/cm 2 s cmHg = 3.349 x 10-10 mol/m2 s Pa.
bMembrane thickness = 50 xm.
350 GPU (1172 x 10 -1° mol/m 2 s Pa) was obtained with CO2/CH 4 selectivity of 40.
This high selectivity indicates that the thin skin layer, the thickness of which
was calculated to be 25 nm, was defect-free. Moreover, the selectivity of the
asymmetric membrane was higher than that of the dense homogeneous
membrane with a thickness of 50 gm. Higher selectivity of the thinner film is
probably due to the formation of a more highly ordered polymer structure as the
film thickness decreases. Table 30.1 shows the results of gas permeation experi-
ments for asymmetric membranes as well as homogeneous membranes prepared
from other polyimides. The table also shows that the permeability ratio in-
creased from 36 (which corresponds to the dense homogeneous membrane) to
43, as the thickness of the top skin layer decreased. They have also reported an
asymmetric membrane with CO, permeance of 1490 GPU (4990 x 10 -1° mol/m2 s
Pa) with CO)CH 4 permeance ratio of 43. It may be difficult to make a membrane
of a large area with sufficiently high pressure resistance according to their
method, but Kawakami's work indicates that there is practically no limit for the
permeation flux of gas separation membranes.
Zeolite membranes have been developed on a small laboratory scale for gas
separation, separation of organic solvents and for pervaporation. Commercial
zeolite membranes are now available for pervaporation. Kita et al. reported
recently on the preparation of X- and Y-type zeolite membranes and their
pervaporation performance [6]. Zeolite membranes were grown hydrothermally
on the surface of a porous cylindrical mullite support. The aluminosilicate gel
was prepared by mixing an aqueous sol (NaOH 14% and SiO, 27%) and an
alkaline aluminate solution prepared by dissolving sodium aluminate (Al/NaOH
= 0.81) and sodium hydroxide in distilled water. After gel formation and aging
(for NaY only), the gel was poured into a reaction vessel fitted with a condenser
and a heater, and then a porous support tube, the surface of which was treated
with seed crystals of zeolite, was immersed in the gel. After hydrothermal
treatment at 100°C for a specific period the coated tube was washed with water
989
TABLE 30.2
Some pervaporation data pertinent to NaX and NaY zeolite membranes (from Ref. [6])
Zeolite Feed mixture (A/B) 10 wt.% of A T Separation factor Flux
type (IC) (A/B) (kg/m2 h)
NaX Water/ethanol 75 360 0.89
Methanol/benzene 50 24 1.25
Methanol/MTBE 50 320 0.26
NaY Water/ethanol 75 130 1.59
Methanol/benzene 50 3800 0.93
Methanol/MTBE 50 3300 1.11
Methanol/MTBE 60 5300 1.70
Ethanol/ETBE 65 1300 0.48
Ethanol/benzene 60 930 0.22
Ethanol/cyclohexane 60 1000 0.27
and dried. The pervaporation data of NaX and NaY zeolite based membranes are
given in Table 30.2. It is shown that membranes are water selective for the
separation of water/alcohol mixtures and both flux and selectivity of these
membranes are very high. It is thought that mass transfer occurs through
microporous cavity in zeolite material and inter-crystalline pores. Table 30.3
summarizes the pervaporation data of NaY zeolite membrane together with
those of polymeric membranes. Methanol and MTBE were separated in the
above experiments. It is obvious that the flux of the zeolite membrane is
extremely high compared with the polymeric membranes. Commercial zeolite
membrane is now available as a product of The Smart Chemical Company. The
membrane combines both a well-defined pore size (0.42 nm) and a high porosity
(47%) and is used for dehydration of organic solvents at a temperature as high as
100°C. The volume of organic solvent treated is from 500 ml to 3 1.
990
TABLE 30.3
Pervaporation performance of various membranes (from Ref. [61)
Membrane MeOH in feed T Separation factor Flux
(wt.%) (°C) (MeOH/MTBE) (kg/m2 h)
Cellulose acetate 0.83-6.9 23-49 14-454 0.05-1.41
Poly(vinyl alcohol) 5-30 45 3-4 0.12-0.16
BPDA polyimide 4.1 60 1400 0.6
Poly(phenylene oxide) 1-20 22 8-24 0.18-0.65
Nafion 417 3.2-5.3 50 35 0.64
Poly(styrene-co- styrene-sulfonic 5.0-14.3 25 25,000-35,000 0.001-0.02
acid) in Mg form
Silicalite/ stainless steel 5-50 (vol.%) 30-50 4-9 0.08-0.12
NaY/alumina 1-50 35-60 2,000-24,000 1.0-2.3
991
OO00000000 SURFACE O0- 00~* 0**
O000O00000 00-0000000
00000000 BULK Q.QQ*QQQQ*
0000000000000
0000000 000
000000004 0000000000
*0-000-00 '000000000
(A) (C)
*0000 Q 00 SURFACE
0000
0000000
Fig. 30.6. A schematic diagram illustrating SMM migration. Closed circles, SMM molecules; open
circles, base polymer molecules; A, time zero; B, time in-between; C, time infinity (from Ref.
[101).
molecular structure of a typical SMM. The solubility in the base polymer can
depend on the additive's average molecular weight, its molecular structure and
its low polar component, which can be controlled by varying factors such as the
main chain length, the number of C-F bond and the type of chemicals used to
synthesize the main chain [7].
SMM was blended in polyethersulfone (PES) membrane. It was found that
the contact angle of the surface increased from 76 ° of PES to 116 ° by blending
SMM, which value is nearly equal to that of Teflon. XPS analysis also confirmed
that SMM was concentrated at the surface of the membrane. These SMM
blended PES membranes were found to remove chloroform effectively from
water by pervaporation [7]. They were found to be less fouled in the treatment of
cutting machine oil/water emulsion by ultrafiltration [8]. It was also found that
the mechanical strength of PES membranes is generally enhanced by blending
SMM into PES [9]. The time required for the SMM to migrate to the membrane
surface was measured and the kinetics of the surface migration was established
[10]. The change in fluorine content at the surface is given in Fig. 30.8 as the
functions of evaporation time and cast membrane thickness.
An attempt is currently being made to develop microfiltration membranes
with hydrophobic surface. These membranes are expected to be useful for the
purpose of membrane distillation. A similar technique was used to render the
HO CH 10H
F-(CFz ) W-(CH,)- - CH 2 0H-o-)
N-C-O-(CH2 3 -I CH
o z C:0(Ch2)z (CF 2 ) F
992
40 ¢(
A
35
30
A
S
C 25
20
o
105 A
.r
To 10 A
A,
A 0.24mm PES/SMM film
5 A A 0.12mm PES/SMM film
0 -
I T r I I I n . Ir i.i \I
0 1 2 3 4 5 6 7 8 9 10 2000
Evaporation time in oven at 1300 [min]
Fig. 30.8. Florine content versus evaporation time (from Ref. [10]).
membrane surface hydrophilic [11]. Recently, Uragami et al. and Miyata et al.
blended surface modifying macromolecules they synthesized into poly(1-tri-
methylsilyl-l-propyne) and polydimethylsiloxane membranes and tested the
performance of the membranes in pervaporation [12,13].
Pore sizes down to 1 nm were measured by a modified bubble point method. The
bubble point method to measure the pore size and the pore size distribution is
based on Cantor's equation:
Ap = 4( cos0/d
The pressure that allows a gas bubble to penetrate through a water-filled pore of
diameter, d, is given as Ap. Using the surface tension, , of water and zero
contact angle (0), the diameter of the pore measurable by the above method is
calculated to be 64 nm. A lower surface tension, or interfacial tension, is
necessary, when a pore size as small as 1 nm is measured. The interfacial tension
of two immiscible liquids A and B can be used for this purpose.
Suppose two liquids A and B are in contact with each other in the membrane
pore. Suppose also A does not wet the membrane surface, whereas B does. When
pressure is applied on liquid B that can wet the membrane surface, the contact
993
16
14
12
10
8
0 1 2 3 4 5 6 7 8
Diameter (nm)
Fig. 30.9. Pore size distributions of some nanofiltration membranes obtained by modified bubble
point method (from Ref. [14]).
Scanning electron microscopy (SEM) has been a powerful tool for investigating
the morphology of membranes from the onset of reverse osmosis. Atomic force
microscopy (AFM) became popular in the 1990s and many AFM pictures of
reverse osmosis membrane surfaces have been taken. There are two examples in
which has AFM contributed to the development of high flux reverse osmosis
membranes. Hirose et al. found a relationship between the flux of reverse
osmosis membranes and their roughness parameters measured by AFM. Accord-
ing to their experiments, an increase in surface roughness resulted in a higher
water permeation flux. As an extension of this observation, they prepared an
aromatic polyamide composite membrane with an extremely rough surface. The
roughness of the surface was confirmed by transmission electron microscope.
The membrane so prepared, called ES 20 membrane, maintained a high salt
994
rejection while the flux was doubled from the conventional low pressure RO
membrane [16]. Fusaoka also found a similar rough structure on the surface of
their super ultra low pressure composite reverse osmosis membrane [1].
Bessieres et al. observed that the actual surface area of an ultrafiltration
membrane (molecular weight cut off = 200 kDa) is almost 40% more than the
geometric surface area (based on smooth surface) of the membrane [17].
The surface roughness of a microscopic scale also has an effect on the fouling
of reverse osmosis membranes by colloidal particles. Elimelech et al., by
comparing the fouling of cellulose acetate and aromatic polyamide composite
membranes, concluded that the effects of surface roughness on the interaction
force between particles and membrane surface resulted in enhanced attachment
of colloids onto the membrane surface, and hence more severe fouling [18]. The
important role of surface roughness in enhancing the attachment rate between
particles or between a particle and a surface has been pointed out by several
investigators. It has also been shown that an increase in the macroscopic surface
roughness in hollow fibers decreased the mass transfer coefficient and thus
increased concentration polarization [19].
The effect of surface roughness on the membrane transport and fouling will
be known more in detail as progresses will be made in the instrument to observe
the membrane surface. The development of hydrodynamics in a microscopic
scale is also called for.
The AFM method also played an important role in studying the morphology
of gas separation membranes. Integrally skinned asymmetric membranes were
prepared from poly(2,6-dimethyl-1,4-phenylene) oxide (PPO) using different
nonsolvents added to the casting solution. These nonsolvent additives consisted
of branched and linear alcohols ranging from C3 to C,,. Permeation data for pure
gases of CO2, CH 4, 02 and N2 were obtained for the membranes so prepared. The
membranes were further characterized by atomic force microscopy [20]. Figure
30.10 shows the permeance ratio of CO2/CH 4 versus mean diameter of nodules
observed under AFM. It is found that the membrane prepared from a solution
that contained 2-ethyl-l-hexanol as its additive (m in Fig. 30.10) had the
highest permeance ratio of 24.2 and the lowest mean diameter, 27.8 nm. There-
fore, the smallest nodule resulted in the highest permeance ratio. A possible
explanation for the tendency of decreasing permeance ratio with an increase in
nodule size is that the disentanglement of polymer molecules between nodules
occurs faster than the polymer vitrification. The smaller capillary force that
works between larger nodules also decreases the tendency for nodules to be
closed. An important observation in this study was the physical appearance of
nodules. There were two types of nodules identified as merged and discrete
(given in Fig. 30.10 as m and d). The discrete nodules are defined by the regularly
occurring and distinct depressions of the AFM image. The merged nodules, on
the other hand, show a flat surface. Another criteria for the discrete and merged
surface are the roughness parameter, Ra. If the roughness parameter is above
0.62 nm, the surface is defined as discrete while it is defined as merged if the
995
28
26 -
Im_m
, 24-
X 4m
c 20-
c 18- 3di
c.) 6d
d 16
10 1 i r . r C- I
20 40 60 80 100 120
Nodule diameter (m)
Fig. 30.10. Pure CO 2/CH, permeance ratio as a function of nodule diameter (from Ref. [20]): (1)
2-ethyl-l-hexanol; (2) 1-octanol; (3) 2-propanol; (4) 2-decanol; (5) 3,5,5-trimethyl-l-hexanol; (6)
2,4-dimethyl-3-pentanol; (7) 2,4,4-trimethyl-1-pentanol; (8) 2,2-dimethyl-3-pentanol; (9)
3-ethyl-2,2-dimethyl-3-pentanol; m, merged; d, discrete.
roughness is below 0.62 nm. Figure 30.10 shows that membranes with merged
nodules form the upper boundary of the data, while those with discrete nodules
form the lower boundary, with exceptions of 5m and 9m. Therefore, membranes
with smooth surfaces formed by merged nodules potentially exhibit high
selectivity. This information is important for the choice of nonsolvent additives
to be used in the preparation of casting solutions. A similar conclusion was
obtained also for the choice of solvent.
996
to know the degree in freedom of the motion of spin probes in a given
environment by analysing ESR signals obtained when the spin probes are
subjected to a magnetic field; i.e., the signal will consist of symmetric peaks
when the spin probes are freely mobile, while the asymmetricity of the peak
shape will increase as the motion of the spin probe becomes more restricted. In
this work, spin probes were incorporated into membranes in two ways. In one,
spin probes were brought into the skin layer of a reverse osmosis membrane
under a reverse osmosis condition, where an aqueous solution of the spin probe
is used as feed. In the other, spin probes were incorporated into a homogeneous
dense film by dissolving the spin probes into a polymer solution from which the
film was cast. Then, both the asymmetric membrane and the homogeneous film
were subjected to ESR analysis.
Figure 30.11 shows the ESR spectra of an aqueous solution of the radical
probe. The spectra consist of three symmetric peaks, since the NO radical of the
probe is freely mobile in the solution. The ratio of the peak heights b/a is unity. It
is expected that the b/a ratio will increase as the mobility of the radical probe
decreases. Figure 30.12 shows the ESR spectra of a cellulose acetate membrane
when radicals were brought into the membrane by immersing the membrane
into a radical solution. Figure 30.12 shows two overlapping spectra correspond-
ing to weakly restricted and highly restricted radical probes: one, which is
weakly restricted, is located inside the pores and the movement is restricted due
to the narrowness of the space in the pore; the other, whose motion is highly
restricted, is located in the polymer matrix and interacting with polymer
molecules. Even though the spectra are the overlap of signals from different
sources, the ratio of the peak heights, b/a, was considered as a measure of the
overall restriction on the motion of the radical probe. It was found that the b/a
-
20 G
H
Left: Fig. 30.11. Typical ESR spectra of TEMPO in an aqueous solution containing 0.1 wt.%
NaCI and 0.05 wt% TEMPO (from Ref. [21]).
Right: Fig. 30.12. ESR spectra of an asymmetric cellulose acetate membrane after immersing the
membrane in TEMO solution for 24 h and rinsing with water to remove the residual solution
(from Ref. [21]).
997
ratio increased from 1.3 to 1.9 as the temperature of heat treatment of the
cellulose acetate membrane was increased from 60 to 90 0C. The corresponding
change in the separation of sodium chloride was 0-95%. Interestingly, the b/a
ratio of 1.9 was also obtained for the dense homogeneous membrane in which the
radical probe was incorporated from the casting solution.
Note that the above equation includes friction factors, ai and i, which are
related to the ratio of the size of an ion and the size of the pore. After simplifying
the above equation by applying Boltzmann's distribution equation and the
Debye-Hiickel equation, a mono-dimensional equation that includes only the
axial co-ordinate in the pore was established. The general form is a complicated
matrix soluble for any number of ions by supplying numerical values specified
below:
- membrane; membrane porosity, pore size, pore length, charge density,
dissociation constant of the charge
- electrolytes; ionic valence, ionic diffusivity, ionic radius
- operating conditions; ionic concentrations in feed, pH of feed, operating
pressure
998
The model was tested by the experimentally obtained separation data for NaC1,
NaNO8, Na2SO4 , CdCl2 and CdSO 4 since the treatment of wastewater was her
primary concern. Calculation was also made for a mixture of NaCl and CdC 2.
999
In a membrane bioreactor, called ZenoGem®, the immersed membrane and
a bioreactor are integrated, thus combining clarification and filtration of an
activated sludge process in a single step. The main advantage of this system is
that no suspended solids will enter the effluent since particulate matters are
completely filtered by the immersed membrane. Since the settling of the sludge
can be minimized, a very high biomass concentration (1.0-1.5%) and high sludge
retention time (SRT 20-50 days) are allowed. This means that the membrane
bioreactor can be 2-5 times smaller than the conventional bioreactor. The
membrane bioreactor is simple in construction and operation.
Reversed micelles are molecular assemblies that are formed when a compound
with an amphipathic structure is added to an organic solvent such as hydrocar-
bon. A micelle can contain an aqueous phase, the size of which is as small as a few
nanometers. Amino acids, peptides and proteins that are usually immiscible in
organic solvents can be extracted into the micelles in the organic phase. Reverse
micellar extraction uses this unique property of micelles [25].
Protein extraction by reversed micelles is schematically shown in Fig. 30.14.
In the first step, proteins in aqueous phase are extracted to the interior of
micelles, when the aqueous phase is brought into contact with an organic phase
in which the micelles are retained. This is called the forward extraction step.
After separating the aqueous and organic phases, the organic phase is brought
into contact with another aqueous phase into which proteins are extracted from
the organic phase. This is the backward extraction step. Forward and backward
extraction are enabled by the proper adjustment of pH or salt concentration of
the aqueous phases. Because of the simplicity of the process, reverse micellar
extraction is expected to become a bio-separation technology by which a large
amount of bio-products can be processed.
Forward Extraction Backward Extraction
Aqueous Phase i l
Reversed Micella
Reversed Micellar MiA
Mixing R 0 eancPhase Aqueous Phase IOrganie Phase
Organic Phase Extraction
Salt oeR
Aq hese
Pha A Aqueous Phase
Hdrophobic
Hydrophyic
1000
Ichikawa reported his results of reverse micellar extraction using di-2-
ethylhexyl sodium sulfosuccinate, called AOT, as a compound with an amphi-
pathic structure and isooctane as an organic solvent. Extractions of cytochrome
c, ribonuclease A and lysozyme were attempted.
Figure 30.15 shows how many percent of cytochrome c was extracted into the
organic phase from the aqueous phase in the forward extraction step. It depends
on the concentration of AOT in the organic phase and a certain concentration of
AOT was necessary to achieve 100% extraction of cytochrome c. This concentra-
tion was called the minimal AOT concentration.
Forward extraction experiments were carried out while maintaining the
AOT concentration at its minimal AOT value. As the concentration of cyto-
chrome c in the aqueous phase increased, cytochrome c concentration in the
organic phase increased and so did water concentration in the organic phase. In
Fig. 30.16 the concentration of water and cytochrome c, both in the organic
phase, are correlated. Interestingly, those concentrations were proportional to
each other and the slope was equal to 3500. This means that 3500 water
molecules were necessary to extract one protein molecule in the micelle. This
number was called hydrophobic surroundings (H.S.). H.S. was found to be
linearly correlated to the polarity ratio of the protein molecule.
Figure 30.17 shows the circular dichroism spectra of cytochrome c. The spec-
tra in the aqueous phase before the forward extraction step and after the Back-
ward Extraction step are the same. However, the spectra from cytochrome c in
the organic phase after the forward extraction has shifted remarkably from the
spectra from cytochrome c in aqueous phase. This means that the configuration
of cytochrome c molecule in the micelle is quite different from that in the aque-
ous phase. The reason is probably that cytochrome c molecule is not floating in
the water pool inside the micelle but it is interacting strongly with AOT mole-
cules that constitute the micelle wall. It was further confirmed that the activity
of cytochrome c molecule was almost completely retained after transferring the
cytochrome c molecule from one aqueous phase to another aqueous phase by the
forward and backward extraction steps.
The applications of this technique for the extraction of antibiotics and DNA
are being investigated [26,27]. In order to make reverse micellar extraction use-
ful in practice, a reverse micellar system with high bio-compatibility has to be
developed. Improvement in selectivity in extraction of a targeted molecule by
introducing bio-affinity in the micelle is also necessary.
30.7 CONCLUSIONS
1001
rI,. . .....
n
...
100
o
oh 80 ,K -
x : 60
60 E
a) Minimal AOT T
tm -40 Concentration
iL 20
... .. I
10 100 0 0.2 04
AOT [mM] cytochrome c [mM]
Left: Fig. 30.15. Degree of protein extraction versus AOT concentration (from Ref. [25]).
5
Cytochrome c concentration, 7.8 x 10 moll1; KCI concentration, 0.1 mol/l; pH 7-8.
Right: Fig. 30.16. Water concentration versus protein concentration at minimal AOT (from Ref.
[25]).
E
x
1002
Attempts have been made, some of them very successfully, to develop hybrid
systems where membrane separation is combined with other separation
processes. We will hear more of the success stories of hybrid system in the future.
Extraction with reversed miscelles has potential in biotechnology applica-
tions. Attention should also be paid to recent development of stereo-specific
membranes [28]. Even though it is still in its infancy, its potential in the
separation of optical isomers is very high.
REFERENCES
1003
21 K.C. Khulbe, T. Matsuura, G. Lamarche, A.-M. Lamarche, C. Choi and S.H. Noh,
Polymer, 42 (2001), 6479-6484.
22 S. Chevalier, Mod6lisation math6matique des m6canismes de separation en nano-
filtration, Ph.D thesis, University of Bordeaux, 1999.
23 R. Rangarajan, M.A. Majid, T. Matsuura and S. Sourirajan, Ind. Eng. Chem. Proc.
Des. Dev., 24 (1985) 977.
24 D. Mourato, Desalinationand Water Reuse, 9/4 (2000) 27.
25 S. Ichikawa, Maku (Membrane), 22 (1997) 331.
26 N.W. Fadnavis, B. Satyavathi, and A.A. Deshpande, Biotechnol. Progr., 13 (1997)
503-505.
27 M. Goto, T. Ono, A. Horiuchi and S. Furusaki, J. Chem. Eng. Japan, 32 (1998)
123-125.
28 M. Yoshikawa, in: Molecularly and Ionic Recognition with ImprintedPolymers, ACS
Symposium Series, 703. ACS, Washington D.C., 1998, Chap. 12, pp. 170-187.
1004
Chapter 31
31.1 INTRODUCTION
During the past several decades, there has been extensive work done to monitor
and quantify the harmful effect of various pollutants to the environment. At the
same time, considerable effort has been dedicated to develop alternative
chemicals and processes which might not be as damaging to the environment.
However, very little emphasis has been placed on designing improved materials
that could selectively remove a wide range of impurities and permit for recovery
and reuse. In particular, new fibrous systems based on coated glass fibers appear
to offer many advantages over the currently available technology of granular
activated carbon (GAC) and ion exchange beads (IEB). These advantages include
the design flexibility to be fabricated in the form of felts or fabrics, which display
greatly improved contact efficiency leading to faster rates of reaction/ regenera-
tion. In addition, the wear resistance and dimensional stability are greatly
improved through the reinforcement of a glass fiber substrate.
In this chapter, the technological background of both activated carbons and
ion exchange materials is described. This is followed by a more detailed consider-
ation of recent work on the activated carbon fibers (ACF) and ion exchange
fibers (IEF).
Activated carbons have been used for many years for both air and water
purification and are described extensively in the literature [1-6]. The primary
raw material for activated carbons can be any organic material with an accept-
able char yield (i.e., coal, wood, coconut shells, or certain polymers). Activation
occurs when the carbon layers in the char are etched away through an oxidation
reaction resulting in the formation of a porous carbon network with high surface
1006
Fig. 31.1. SEM micrographs of (a) Purolite cationic exchange bead displaying effects of osmotic
shock and (b) ion exchange fibers.
Undoubtedly the first major innovation in this field occurred with the discovery
that ACF could be made by coating glass fibers (either woven or non-woven) with
a phenolic resin. The coating is then carbonized and activated from 600-800°C in
a single step process to yield the adsorbent material (Fig. 31.2). Given the low
costs of the constituents and the relatively simple processing, the potential exists
for producing a new adsorbent material that is cost competitive to GAC and with
Fig. 31.2 (a) Broken end of ACF showing the glass core surrounded by carbon coating. (b) Broken
end of an etched ACF showing a hollow carbon layer.
1007
greatly improved wear resistance over current ACF. In addition, the contact
efficiency is significantly improved compared to current ACF and almost an
order of magnitude better than GAC. Faster adsorption/desorption rates (due to
a decrease in the diffusion path length) maximizes the carbon usage resulting in
reduced critical bed depths and pressure drops. Better wear resistance means
lower attrition losses during handling and the potential for greatly increased
utility as protective garments. This ACF can be manufactured in a wide variety
of product forms (textiles, felts, papers), which allows for vast design flexibility
thus meeting the demands of a variety of applications.
Typical activated carbons have limited selectivity since their adsorption char-
acteristics are controlled to a considerable extent by the acidic nature of their sur-
face. In addition, the pore size distributions tend to be very broad with little
consideration as to what would be the most advantageous size range for a particu-
lar contaminant. Thus recent research describing the relationship between pore
size/surface chemistry and adsorption capacity becomes important. This led to
specific techniques for tailoring these variables through careful control of the acti-
vation conditions. The time/temperature of activation dictates the pore size while
the atmosphere can greatly alter the pore surface chemistry. Thus a wide range of
tailored materials with micropores ranging from 5-30 A and chemical structures
including nitrided, sulfonated, unsaturated, or a more highly oxidized surface
could be made to enhance adsorption of specific contaminants.
1008
- ,-- -,Samples were washed with 0.5N of HCIand DI. water.
-,
- -,Samples were only washed with D.I. water.
1000 60
...
--- 50
< 600 - -
Fig. 31.3. Effect of activation temperature on the surface area of phenolic resin-based ACF
produced by ZnC2, activation.
starts at about 300°C and increases rapidly up to 350, after which it gradually
rises to a broad maximum of -852 m2 /g of coating. At temperatures above 550°C
the surface area begins to decline due to significant widening of the pores.
The chemically activated ACF exhibit notable advantages over current acti-
vated carbon materials. For example, a wider range of fiber substrate materials
may be used offering greater versatility. This includes low melting point glass
microfibers, as found in current HEPA filters, allowing for potential adsorption/
filtration in a single step (of considerable importance in reducing the size of gas
masks, for example). Use of polymeric fiber substrates would result in even
better wear resistance over the glass substrate, further extending their utility
for protective clothing. This new approach makes it far easier and less expensive
to manufacture due to the low temperature activation and higher yields. By
utilizing different starting polymers, a wide variety of chemically modified
surfaces can be created which are capable of adsorbing/chelating many different
contaminants from both air and water.
1009
Fig. 31.4. STM images of ACF (a) surface and (b) cross-section.
fiber surface was composed of dispersed mesopores while the bulk of fiber
contained a homogeneous arrangement of micropores that were ellipsoidal in
shape (Fig. 31.4). This research also elucidated the mechanism of pore formation
showing that the first step of heating and charring the phenolic fiber creates a
microporosity of -600 m2 /g just from the evolution of volatiles. In the second
step, etchants such as steam, CO2 , air or ammonia act to produce a mesoporous
region to a depth of about 10 nm. The microporous structure in the core remains
relatively uniform in diameter but increases in size with increasing time of
activation. This type of process is in contrast to GAC where the activation must
initiate from the outside surface thus resulting in a very broad pore size
distribution.
In addition, experiments were carried out on a series of ACF with a range of
surface areas to examine the effect of pore size on adsorption of gases [11]. The
adsorption isotherms of n-butane for this series of ACF showed a counter-
intuitive relationship such that adsorption capacities at low concentrations
increased as surface areas decreased. This was explained by assuming that the
lower surface area material had a smaller pore size; this hypothesis was later
confirmed through STM. The smaller pores led to an increase in the overlapping
potentials from neighboring pore walls thus binding the molecules more tightly
at low concentrations. This resulted in the condensation of butane at much
lower concentrations, allowing for immediate use of its total pore volume. How-
ever, as the concentration was raised, the butane would eventually condense in
the larger pores. Since the higher surface area ACF also had a larger total pore
volume, they would eventually win out over the smaller pore materials resulting
in a crossover regime. Recently, this work was extended to examine the effect of
boiling point on the crossover phenomenon [12]. When the adsorption tempera-
ture is much greater than the boiling point of the adsorbate, the thermal energy
possessed by these molecules is sufficient to overcome the forces leading to lique-
faction. Therefore the molecule would prefer to remain in the gas phase. This
means that for gases with low boiling points such as methane, only the high-
1010
energy sites will be occupied because pore size is the critical variable. Since
methane does not condense in the pores, no crossover was observed (the small
pore ACF adsorbed better over the entire concentration range). It was shown
that as the boiling point of the adsorbate was increased, the crossover regime
shifted to higher concentrations. For low boiling point gases or at lower concen-
trations, the smaller pore ACF adsorbs to the highest capacity. But for higher
boiling point gases and at higher concentrations, pore volume becomes the con-
trolling feature. Therefore, this becomes a competition between the pore size
and the pore volume of the ACFs to determine which will be the most important
factor.
This type of information is critical in tailoring an activated carbon for
increased adsorption capacity. Thus, in most cases, it is the low surface area
(small micropore) material that is the most advantageous since one is typically
concerned about removing very low concentrations of contaminants. This is
counter-intuitive to the notion that higher surface area materials will always
perform better. As discussed previously, GAC is at an extreme disadvantage due
to its large pore size distribution whereas in ACF the distribution is much
narrower. In GAC, the same type of micropores only exist in the very deep
portions of the structure while the entrances on the surface are extremely wide
by comparison. Therefore, the pore structure of GAC has a weak attraction at
the surface and it takes a much longer time for molecules to diffuse and
penetrate to the point where they can be captured and held. Thus, the process of
filling the pores in GAC is more time consuming and much less efficient than
that of ACF.
1011
1
0.8
ACF-Untreated
--- ACF-Oxidized !
0.6 i ........... .... _
0
U
U
0.4
I..........
0.2 J............. .. . ..
., , ,,}1 i X
0
0 200 400 600 800 1000 1200 1400 1600
Time (min/g)
Fig. 31.5. Normalized breakthrough curves for ammonia (Co = 970 ppmv) on ACF.
-n 10
-e
0
0 ' ' ' ' ' ' ' ' ' ' ' ' ' '
0 1000 2000 3000 40 00 5000 6000
Concentration (ppmv)
Fig. 31.6. Adsorption isotherms of HCI gas on ammonia treated ACF.
1012
increases in adsorption capacity over currently used systems have been demon-
strated for SO2 , acetone, and trichloroethlyene [16-18]. Through these exam-
ples, one can note the effect that altering the pore surface chemistry has on
adsorption of various contaminants. This type of information, combined with
the knowledge generated on the effect of pore size, is invaluable in designing
highly efficient systems for removal of trace contaminants.
E
0
C
c
0 5 10 15 20 25 30 35 40
Volume (L)
1013
initial breakthrough, the GAC effluent concentration steadily climbed while the
ACF stayed level (close to zero) until a sharp breakthrough curve occurred. This
was indicative of fast adsorption kinetics and a very short mass transfer zone
giving the ACF an added advantage under dynamic conditions. This was in con-
trast to the GAC where the much longer diffusion path lengths and buried
micropores resulted in premature breakthrough. These same benefits were also
obtained with the chemical warfare agents where 3-fold adsorption superiority
was measured. In addition, the ACF could be easily regenerated multiple times
at low temperatures to yield 100% working capacity. Regeneration of GAC would
require much higher temperatures thus altering the pore structure and reducing
the adsorption capacity.
The concept of developing ion exchange materials in the form of fibers was first
reported by Economy et al. 30 years ago [20]. Fibrous ion exchange materials
have several advantages over the conventional ion exchange units. These advan-
tages include the ability to be fabricated in the form of felts or fabrics where the
contact efficiency is greatly improved. A fibrous form eliminates the need for
currently used retainers such as canisters, screens, etc. Recently, two new
concepts have been developed to yield fiber systems with a range of capabilities.
The first involves the concept of designing ion exchange materials by coating
glass fiber substrates with a polymeric ion exchange resin [21]. In the second
system, modifications of the aforementioned ACF where the pores have been
chemically treated to enhance their use for ion exchange as well as their ability
to chelate and remove specific contaminants from water [22].
1014
HC:CH 2
S Coatig
*CiR5
4h Fiberglass Substrate
with Resin Coating
t 2SO
V
Diivinylbenzene Oligomerization Reaction H H H
-c-c-c-c-C-C-
HS3'
-C-C--C-C--C-
H2 H2 | H2
SO3 H SO3-H
Sulfonation Reaction
Styrenic Sulfonic Acid Resin on
Glass
Fig. 31.8. Synthesis of polymeric ion exchange fibers.
Ion Exchange
Beads
Suspension
Poymerization
(organic
solvents
usedformacroporoity)
Ion Exchange Fibers
Ion Exchange
Fibers Distallation
forSolvent
Removal
toAllow
forPorosity
Oligomer Coating &Curing Stage
ofBeads
Preswelling Using Solvents
Organic
Functionalization Stage toAllow
Core
Functionallization
I Stepped
Dilution
Rinse
I~~~~~~~~~~~~~~~~~~
Beads
Shipped
&Stored
with~50%
Mdsture
Content
IPreswelling
ofBeads
Overnight
Prior
toEnd-use
Fig. 31.9. Comparison of fiber and bead preparation demonstrating that far fewer steps are
required with IEF.
ately in contact with the resin was rapidly depleted of ions by proton exchange
from the resin. Further exchange was delayed until more ions from the bulk
solution could diffuse into the surface film. Diffusion was easier with increasing
ionic concentration in bulk solution, making the exchange faster. This supports
the view that at low concentrations, the rate was controlled by "film" diffusion
1015
Low Temperature 25C)
5 HighTemperature (95 C}
4-
0
4 8 12 16 19,5 2 4 5.5
4
10 T ,
e~~~~~~~~~~~~~
, .
P~~~~~~~~~~
100
E--!
F i i
I I I . . - --
1u
0 0.5 1 1.5 2 2.5 3
Total Volume 0.1 M NaOH Added, (ml)
Fig. 31.11. Batch rates of exchange for two saline concentrations illustrating an order of
magnitude faster reaction rates for the IEF.
1016
90 -
80-
A Cationic Beads (5% DVB)
O Cationic Fiber (8% DVB) O
70-
60
E
50- A A A A A A A A A
0
(- 40-
- 30
Conditions:
n, 20- 100 ml/min flow rate
o pH = 4.65 at 27.9 C
Lowest residual (-1.2g resin each)
10- level <1 ppb
0
O- r 0 omm 0
. I I I I I I I I I I
0 500 1000 1500 2000 2500 3000 3500
31.3.2.3 Carbon-BasedIEFs
Carbon-based IEF provide a different mechanism of accessing the ionic species
via a microporous network in comparison to the swollen networks of the beads or
fiber coatings. This allows the capability to utilize both pore surface chemistry
(oxidation and/or sulfonation) and molecular sieving properties to selectively
1017
35-
Conditions:
100ml/min flow rate A
30-
pH 3.1 at 28.5C
(-2.0 g resin each) A
25
A
A v
A *
A Anionic Bead Filter
A O Anionic Fiber Filter
A
< 15-
a 10-
a' 00 0
5-
o lowest residual
0 00
O0 level <0.1ppm
0- 0 o oo o o o o o O
, i * . i . . I . I . i
20 40 60 80 100 120 140 160
Cumulative Effluent Volume/g resin, (ml/g)
Fig. 31.13. Dynamic mode breakthrough of Arsenate (V) on anionic fiber versus beads (influent
cone = 104.7 ppm).
remove ions. Utilization of ACF greatly increases the surface area available for
incorporation of ion exchange groups. In addition, these materials exhibit
excellent chemical and thermal resistance. This is required for many industrial
applications including the removal of cations from corrosive media, hot solutions
or melts.
Oxidation of carbons to prepare weak acid cationic (WAC) exchangers has
been extensively studied [16,23,24]. The ability to functionalize carbons with
strong acid cationic (SAC) groups would lead to systems with ion exchange
capability over the entire pH range. The preparation of high surface area
carbons typically involves treating precursors to high temperature (700-1100°C)
in activating environments (e.g. CO2steam) [25]. It has been reported that
degradation of phenolic resins at temperatures above 700°C was associated with
the loss of aromatic hydrogen [26]. The presence of aromatic hydrogen is
essential for electrophilic substitution reactions such as sulfonation. Due to the
poor reactivity of carbons activated at temperatures above 700°C, reports of
sulfonation in the literature are scarce. Instances where it has been
accomplished include coals and carbon blacks, albeit with low ion exchange
capacities [27,28]. Thus recent synthetic routes utilizing chemical activation at
much lower temperatures (allowing for retention of the aromatic hydrogen)
were examined to prepare SAC exchange fibers derived from phenolic resins.
Principal variables investigated to maximize the degree of sulfonation included
the carbonization and sulfonation temperatures.
After sulfonation, the increase in oxygen was significantly higher than
expected as calculated from addition of only sulfonic acid groups. Thus, it was
concluded that oxidation of the microporous carbon accompanied the high
temperature sulfonation of these systems; this result is consistent with reports
1018
.50
WX'
'0'
3
cat to
1A
Fig. 31.14. Strong acid ion exchange capacities of ACF series as a function of carbonization and
sulfonation temperatures.
1019
8 a
i'
Fig. 31.15. Weak acid ion exchange capacities of ACF series as a function of carbonization and
sulfonation temperatures.
31.4 SUMMARY
In this chapter it has been shown that preparing activated carbons or ion
exchange resins in the form of fibers has greatly improved their contact
efficiency and ease of regeneration. Because of these improvements, it is now
possible to remove a wide range of trace contaminants (both organic and
inorganic) to below one ppb. This represents a several fold improvement over
commercially available GAC and IEB. In the case of ACF, additional improve-
ments in efficiency are possible through further tailoring the pore dimensions
and surface chemistry. It is noteworthy that use of a glass fiber substrate for
preparing the ACF and IEF has greatly simplified the synthesis as well as
improving the wear resistance.
New directions for preparing activated organic fibers at relatively modest
temperatures of 200-400°C are indicated along with the preparation of an ACF
containing sulfonic acid groups. It is anticipated that both of these systems will
open up new and important pathways for removal of trace contaminants from
the environment.
1020
REFERENCES
1021
Chapter 32
32.1 INTRODUCTION
1024
developed) formats. In cases where recent extensive reviews are available for
particular types of polymeric extraction materials, the main attention is focused
on the most recent developments not covered by existing reviews. Molecularly
imprinted polymers are now well-established extraction materials for SPE, and
two very recent reviews [47,48] have thoroughly covered the existing literature.
However, in the SPME area, MIPs have been introduced only very recently. For
this reason, emphasis in this chapter has been placed on molecularly imprinted
materials as extraction media for SPME. Extensive reviews have also been
published on chemically modified polystyrene-divinylbenzene materials for SPE
[1,49,60]. Taking this into consideration, only a brief discussion is provided here
of these topics, mainly covering the general principles and latest developments
not covered by those recent reviews.
Special attention is focused on sol-gel polymers that advantageously combine
organic and inorganic components in their structures. A general scheme of
sol-gel chemistry is presented that allows for the in situ creation of these poly-
mers in a simple way under mild thermal conditions (often at room temperature)
using appropriately designed sol solutions. The sol-gel polymeric coatings and
monolithic beds prepared using such sol solutions get chemically bonded to the
surface of the substrate and, therefore, possess significantly higher thermal and
solvent stabilities-attributes that are important for effective extraction and
subsequent release of the analytes from the coatings for further analysis. The
organic-inorganic hybrid nature of these polymers provides advanced material
properties allowing for the extraction of both polar and nonpolar analytes pres-
ent in the matrix. A wide variety of functional groups can be incorporated in
these polymeric materials by properly designing the sol solution composition.
The preparation and extraction performance of sol-gel polymeric coatings and
monolithic beds with chemically incorporated polymers, dendrimers, macro-
cycles, and other organic ligands are presented. Future directions in the develop-
ment of sol-gel polymers for analytical extraction are pointed out. To our
knowledge, sol-gel polymers have not yet been covered by any of the existing
reviews on polymeric extraction materials.
1025
solutions include (a) cloud-point extraction (CPE) [63,64], (b) micellar extrac-
tion (ME) [65,66], (c) extraction based on aqueous biphasic systems (ABS) [67],
and (d) extraction using thermoseparating polymer systems (TPS) [68]. Rogers
and coworkers [62] have published an extensive review covering the existing
literature on all four of these polymer solution mediated extraction techniques
up to and including 1999. Here we will focus on new developments (not covered
by the above-mentioned review) with emphasis on future trends in this area.
1026
factors. In the case of temperature-sensitive polymers, the stimulus factor is
temperature. Temperature-responsive polymers respond to the temperature
change in the environment by a corresponding change in their molecular
structure and properties including phase transition, solubility, hydrophobicity,
hydrophilicity, etc. [72]. Examples of other types of stimuli-responsive polymers
are pH-responsive polymers, photo-responsive polymers, electric field-respons-
ive polymers, and chemical species-responsive polymers. In the latter cases, the
external stimuli are pH, light, electric field, and chemical species.
Temperature-responsive polymers are soluble in water. However, above a
certain temperature (a characteristic value for each TRP), called lower critical
solution temperature(LCST), a temperature-responsive polymer becomes insol-
uble in water due to temperature-induced phase transition. This physical change
is reversible, and the polymer goes back into solution when its temperature is
lowered below its LCST. Thanks to this property, TRPs have been extensively
used in diverse areas of science and technology: as an effective medium for con-
trolled drug delivery [73-77], as controlled permeability membranes [78,79], as
bioconjugates [80,81], as nontoxic media for cell culture [82,83], as chromato-
graphic stationary phases [72,84-89], and as environmentally benign extraction
media [90-99].
Poly(N-Isopropylacrylamide) (PNIPAAm) is the most widely used and best
studied among temperature-responsive polymers known to date. Scheme 32.1
illustrates synthesis of PNIPAAm through radical teleomerization of N-iso-
propylacrylamide using 3-mercaptopropionic acid as the chain transfer reagent.
It also shows the possibility of chemical attachment of PNIPAAm to silica
particles that can be used either as a chromatographic stationary phase or as a
solid-phase extraction media [72,80,100]. Cole et al. [101] described an alterna-
tive synthetic scheme via aqueous redox polymerization of N-isopropyl-
acrylamide.
One important factor responsible for the wide use and popularity of poly(N-
isopropylacrylamide) as a temperature-responsive polymer is its moderate value
for the lower critical solution temperature (-32°C) which, being just above the
room temperature, is very convenient for experimentation [97,102]. Below
LCST, the polymer chains are hydrated to an expanded, water-soluble form.
Above LCST, the polymer chains dehydrate and shrink to a water-insoluble form
that separates out of the solution as a gummy precipitate that can be easily iso-
lated.
The temperature-responsive phase-transition property of these polymers
provides a powerful analytical tool for the preconcentration of trace analytes in
aqueous media. When a temperature-responsive polymer is solubilized in an
aqueous medium (below LCST) containing trace of organic contaminants, the
polymer chains are dispersed throughout the entire volume of the solution, and
can bind with individual molecules of the trace contaminant through hydropho-
bic and/or other forms of molecular level interactions. If, after an appropriate
contact period, the system is warmed above LCST, the temperature-responsive
1027
H2C=CH HooCCH 2 CH 2 S-CH 2 CH -H
C =O AIBN C=O
NH + HOOCCH 2CH 2 SH - NH
CH DMF CH
/
H3C CH3 H3C CH
0O
N-OH O
_________ IjN- 2 sSCHS
-CCH CH-CH
DCC 0 NH
CH
/
H3C \CH3
polymer chains precipitate out together with the contaminant molecules bound
to them. This translates into an effective capture of the contaminant molecules
(that were originally dispersed throughout the entire volume of the aqueous
sample) and their transfer to the precipitated polymeric matrix. Since the
volume of the precipitate is exceedingly small compared with the total volume of
the solution, the temperature-responsive precipitation results in a remarkable
enrichment of the target trace analytes.
The hydration/dehydration phenomena associated with the temperature-
responsive polymers also provide a mechanism to fine tune the selectivity of
these polymers based on the temperature to be used. At temperatures below
LCST, these polymers are hydrophilic while above LCST they are hydrophobic
[72]. This opens up new possibilities for liquid-phase separations in an environ-
mentally benign way by using aqueous mobile phase (no organic modifier) and
TRP-bonded stationary phases, and fine tuning the hydrophobicity of the sta-
tionary phase by adjusting the temperature. Essentially, this gives the possibil-
ity to replace mobile phase composition programming (gradient elution) with
temperature-programmed elution.
Temperature-responsive polymers have been used for the extraction of
hydrophobic and amphiphilic analytes. Figure 32.1 illustrates the preconcentra-
tion effect obtained for hydrophobic pesticides in ground water through poly-
mer-mediated extraction using PNIPAAm as the TRP.
1028
(A) (B)
8
c
X a.
0
0
IL
10min
ii
d 2
I , .
F ff 'O
i-I I II
Ii I - II -. .- 0 10 20 30 0 10 20 30
Lni ch I Time (min) Time (min)
Left: Fig. 32.1. Chromatograms of some pesticides (0.1 gM) in groundwater. Mobile phase, 70%
(v/v) acetonitrile; flow rate, 1 ml/min; detection wavelength, 254 nm. (A) without concentration,
(B) with polymer-mediated extraction. The polymer phase formed from a 10-ml sample solution
containing 0.1% (w/v) PNIPAAm was dissolved by 80 l of acetonitrile. A 20-a1l portion of the
resulting solution was injected. (Reproduced from Ref. [97], with permission of American
Chemical Society).
Right: Fig. 32.2. Chromatograms of river water: (A) without preconcentration, (B) with precon-
centration via polymer-mediated extraction. Conditions: 250 x 4.6 mm i.d. Intersil ODS column;
30% (v/v) aqueous acetonitrile mobile phase; 1.0 ml min' flow rate; UV detection at 254 nm.
(Reproduced from Ref. [98], with permission of Elsevier.)
1029
tridge or a precolumn, (b) sorbent disk, (c) monolithic bed (a single piece of
porous solid within a tubular confinement), and (d) flow-through membranes.
The ultimate aim is to obtain an enriched sample of the target analyte(s) and to
free it from any interferences that might be present in the original liquid sample.
To this end, the solid-phase should be carefully selected so that it provides high
selectivity for the target analyte(s). The solid-phase selection should be based on
physicochemical and structural characteristics of the analytes to facilitate vari-
ous types of molecular level interactions with the analytes present in the sur-
rounding environment. In order for SPE to be effective, such interactions
between the target analyte(s) and the solid sorbent should dominate over analo-
gous interactions of the target molecules with the original liquid matrix.
Once the solid phase has been selected, a number of steps needs to be
performed in sequence to obtain a clean, preconcentrated sample of the target
analyte(s) free of interferences. These are: (a) activation (conditioning)- pre-
treatment of the solid phase with an appropriate solvent to create a favorable
environment on the solid surface for intimate interaction between the target
analyte(s) and the solid sorbent; (b) sample application-passing the liquid
sample through the solid bed to provide sorbent/analyte interaction to achieve
retention of the target analyte(s); (c) washing-passing an appropriate liquid
through the solid bed to selectively remove the interferences (but not the target
analyte(s)); and (d) elution-recovery of the extracted and purified target
analyte(s) by passing through the sorbent bed a small volume of a solvent that
has higher affinity for the target analyte(s) than the affinity of sorbent toward
the analyte(s). The practical aspects of performing SPE on a cartridge are
illustrated in Fig. 32.3.
Solid-phase extraction is a more environmentally benign extraction tech-
nique than liquid-liquid extraction (LLE). SPE drastically reduces the consump-
tion of hazardous organic solvents. It is simple and easy to operate. Thanks to its
simplicity, reliability, and applicability to a diverse range of samples, SPE has
undergone explosive growth over the last few decades, and has now become the
most popular sample preconcentration technique [46].
It is evident from the above discussion that the heart of SPE is the solid
sorbent used for extraction. Future advances in SPE are strongly dependent on
the development of new types of sorbents with advanced sorption characteris-
tics. Novel synthetic polymeric sorbents have great promise for future develop-
ments in SPE. Before addressing this central topic of the chapter, a brief
discussion highlights the advantages and drawbacks of conventional solid-phase
extraction media with emphasis on how polymeric materials can overcome the
limitations of conventional extraction media and expand the scope of SPE.
1030
(A) Specimen
reservoir
Medical-grade
polypropylene
Polyethylene
Sorbent bed _ ? fritted disk (20 p)
Fritted disk
(20 )
Luer lip
..f~,
Elution
Communications). application
(PS-DVB) [1]. The following discussion on traditional SPE sorbents will be brief
since the focus of this review is on new polymeric sorbents.
1031
of silica-based stationary phases and extraction media) are characterized by low
retention for polar analytes. In solid-phase extraction applications, this results
in low breakthrough volumes that make such materials often unsuitable for the
preconcentration of polar compounds. Because of these limitations, much
research effort is directed towards developing alternative materials that possess
a wider range of pH stability. These include the development of polymeric
materials [1], zirconia- [117-122] and titania-based materials [123-127], and
polymer-encapsulated silica [128-132]. Recently, Kirkland and coworkers [133]
reported the possibility of creating silica materials with a wide pH stability range
(up to pH 11) via sol-gel technology.
1032
need to be quantitatively recovered. Because of this strong sorbate/sorbent
interaction, it is difficult to desorb such analytes from activated carbon materi-
als resulting in poor recovery. Besides, activated carbon materials are character-
ized by complicated surface and micropore structures containing various polar
groups [163] that are difficult to reproduce, resulting in significant batch-to-
batch variations in their properties. The adsorptive properties of activated car-
bons are greatly affected by the atmospheric humidity. The adsorptive capacity
of activated carbon strongly depends on its micropore structure. The huge
surface area (-1000 m2 /g) of activated carbon may promote certain catalytic
reactions. In such a case, the composition of desorbed analytes may not be repre-
sentative of those originally adsorbed.
(b) Graphitized carbon black (GCB): The term "graphitized carbon black"
refers to the forms of carbon prepared by thermal treatment of ordinary carbon
black to 2700-3000°C in an inert gas environment, usually in a graphitizing
furnace [164,165]. This high-temperature treatment leads to the breakdown of
various functional groups of the original carbon black surface, and results in a
material with a homogeneous surface that is practically free from unsaturated
bonds, ions, free radicals, and lone electron pairs. Graphitized carbons are also
free of micropores, and possess a typical surface area of 100 m2 /g [142]. Solute
adsorption takes place on the external surface of GCB predominantly due to
dispersive interaction between the solute molecules and the nonspecific adsorp-
tion sites on the GCB surface containing hexagonal array of carbon atoms. GCB
may partially regain surface heterogeneity during cooling of the carbonaceous
surface after graphitization [140] and the small amounts of polar adsorption
sites present on the surface are capable of offering specific interactions with
polar solutes. From a crystallographic point of view, three distinct types of
graphitized carbon structures can be differentiated: (a) Bernal perfect three-
dimensional hexagonal structure of graphite [166]; (b) Lipson-Stokes
three-dimensional rhombohedral structure of graphite [167]; and (c) Warren
two-dimensional turbostratic structure of graphite [168].
The use of graphitized carbon blacks and carbon-based extraction media in
solid-phase extraction have recently been reviewed [104,105]. Recent SPE appli-
cations of GCB include the preconcentration of phenols [140,142,144,169],
pesticides and herbicides [142,145,169-176], estrogens [177], alkylbenzene-
sulfonates and metabolites in fish [178], and surfactants in water [142]. Non-
polar nature of graphitized carbon and a wide range of pH stability makes this
material attractive for use as an effective sorbent for extraction and precon-
centration of analytes under extreme pH conditions not suitable for silica-based
extraction media. In this respect, GCB enjoys the advantages inherent in poly-
meric extraction materials. However, GCB has some significant shortcomings
that include excessive solute retention and poor mechanical stability. Attempts
have been made to improve mechanical strength of graphitized carbon
materials. To this end, a new type of sorbent, porous graphitic carbon (PGC),
was developed by immobilizing graphitized carbon on a silica structure.
1033
(c) Porous graphitic carbon (PGC): Graphitized carbon black is practically
nonporous, and has a homogeneous hydrophobic surface. Activated carbon, on
the other hand, is porous but suffers from surface heterogeneity. Therefore,
carbon-based porous materials with homogeneous hydrophobic surfaces present
considerable theoretical and practical interests [148-152,155]. PGCs are such
materials; like GCB, they are characterized by a wide pH stability range. They
are prepared by impregnating a porous template material (e.g., silica) with
organic polymerizing mixture (e.g., phenol-formaldehyde mixture, phenol-
hexamine mixture, etc.) [148,153,154] and the polymerization reaction is
conducted within the pores of the template material. After the polymerization
reaction, the in situ created polymer is converted to glassy carbon by heating to
-1000°C in an inert environment. At this point, the silica template is removed
by treating with an alkali to produce porous graphitic carbon which is further
heated to 2000-28000°C in an inert environment. This thermal treatment
removes the micropores, improves surface homogeneity, and produces some
degree of graphitization determined by the temperature at which the thermal
treatment is carried out. The particle and pore size characteristics of the
resulting PGC is determined by the chosen template material, while the surface
characteristics of the PGC is determined by the final thermal treatment and any
other chemical treatment that may follow.
A study was undertaken by Puig and Barcelo [144] to compare the perform-
ances of PGC and polymeric sorbents in SPE for the preconcentration of
phenolic compounds in environmental waters. The results showed that except
for one compound, 4-chloro-2-aminophenol, PGC provided inferior performance
both in terms of breakthrough volumes and detection limits.
(d) Carbonmolecularsieve (CMS): Carbon molecular sieves are carbon-based
microporous sorbents with very uniform pores [146,147]. They are characterized
by high surface area and strong retention of organic molecules. They are
prepared by controlled pyrolysis of suitable organic materials (e.g., polyvinyl
chloride, petroleum pitch, etc) at temperatures usually exceeding 400°C. They
possess disordered cavity-aperture structure resulting from the cross-linking of
crystallites. Important factors that determine the pore structure, pore distribu-
tion, and properties of CMS include (a) identity of the pyrolyzed polymer, (b)
purity of the polymer, (c) the used method of carbonization, and (d) presence of
impurities. Carbon molecular sieves are commercially available (e.g., Carbosieve
and Carboxen types from Supelco).
1034
PS-DVB-based sorbents have gained wide popularity among various polymers
that have been evaluated as extraction media. This is explained by the fact that
PS-DVB-based sorbents are characterized by a number of advantageous proper-
ties compared with silica-based surface-bonded extraction media, originally
developed for use as stationary phases in HPLC. Although both the silica-based
reversed-phase extraction materials and PS-DVB-based sorbents have hydro-
phobic surfaces, thanks to the presence of aromatic benzene moieties on the
surface, PS-DVB-based extraction media allow for selective extraction through
n-n: interactions with the analyte molecules.
By far the most outstanding advantage of PS-DVB-based polymeric sorbents
is their stability over a wide range of pH. This makes them suitable for extrac-
tion even under extreme pH conditions that are not compatible silica-based
sorbents with surface-bonded organic ligands. A second advantage of the PS-
DVB-based extraction materials is their ability to extract analytes from aqueous
media without requiring prolonged preconditioning with a polar organic solvent.
Silica-based surface-bonded extraction media often require such a precondition-
ing step because of the poor wettability of the bonded hydrophobic ligands with
water.
Further improvement in the wettability of PS-DVB-based sorbents with
water, and hence their ability to extract analytes from aqueous media, was
achieved through derivatization of PS-DVB with various polar groups including
acetyl [182-184], sulfonate [185], hydroxymethyl [182,183], benzoyl [186],
o-carboxybenzoyl [187], 2,4-dicarboxybenzoyl [188], 2-carboxy-3/4-nitrobenzoyl
[188], and tetrakis(p-carboxyphenyl)porphyrin [189,190]. The chemical basis of
all these derivatization reactions is presented in Scheme 38.2 [49]. PS-DVB has
already become a widely used SPE material. Various aspects of synthesis and use
of PS-DVB and functionalized PS-DVB have been covered in recent reviews
[1,49,188b]. Details on the development of PS-DVB-based extraction materials
and their use in analytical extractions can be found in these and other publica-
tions [182-190].
1035
AOH
N., __
% -T
CH3COCI H
A10 3
Cx/C' 0
\Cb
c0t
CO(t
IN(CHAJ
0-
1036
(a) fsH
+-OH
+2 I -
Cor 0 H
synthesise
adduct
0:)- I polymerise
(b)
n mamix
in excess
Fig. 32.4. Schematic representation of the molecular imprinting process for: (a) covalent
imprinting of phenyl -D-mannoside using vinylphenylboronic acid; and (b) non-covalent
imprinting of Boc-phenylalanine using methacrylic acid. (Reproduced from Ref. [191] by
permission of Elsevier).
1037
A B C
Naproxen
Nap roxen
Ibuprofcn Ibuprofen
Naproxen
6
I
5A tob1
r 20 o 5 lo i 2 -o 6 i- i15 20
Time (min) Time (min) Time (min)
Fig. 32.5. HPLC Chromatograms of a standard Ibuprofen sample (A), control plasma sample (B),
and control plasma sample spiked with Ibuprofen (C) using column switching techniques.
Conditions: 20 gl injection volume; precolumn, RAM-MIP-1 (10 x 4 mm i.d.); analytical column,
Cosmosil 5C-18-MS (150 x 4.6 mm i.d.); eluent for pretreatment, 20 mM phosphoric acid-
acetonitrile (78:22 (v/v), pH 2.24) at 1.0 ml min/min for 5 min; mobile phase, 20 mM sodium
phosphate buffer-acetonitrile (75:25 (v/v), pH 7.34) at 1.0 ml/min; detection, UV absorbance at
223 nm; Ibuprofen concentration at 5.0 )g/ml in (A) and (C). (Reproduced from Ref. [215] by
permission of American Chemical Society).
1038
to develop MIPs with restricted-access material (RAM) characteristics as will be
discussed in the next section.
1039
charge repulsion mechanisms. In GFF ISRP, for example, negatively charged
glycine groups are present at the pore entrance and can take part in the exclu-
sion of negatively charged proteins.
A semipermeable surface (SPS) RAM particle normally consists of a core
region of functionalized silica and a covalently bonded hydrophilic layer of
poly(oxyethylene) on the external surface. Practically, any derivatized silica
support (or even non-functionalized, bare silica) can be used to prepare this type
of RAM. The external coating of poly(oxyethylene) serves as a hydrophilic
barrier to proteins and other biopolymers. This hydrophilic polymeric barrier,
effectively prevents macromolecules from entering into the porous structure of
the RAM resulting in their exclusion. Small molecules, however, can diffuse
through the hydrophilic barrier and interact with the core material resulting in
their retention. Compared with ISRP RAMs, SPS restricted-access materials
possess greater analytical flexibility since they can be used for the extraction of
both polar and nonpolar analytes. This advantage of SPS RAMs stems from the
flexibility in using cores with a wide variety of functional groups-polar or
nonpolar. It should be mentioned that only nonpolar functional groups are used
in ISRP RAMs allowing for the extraction of only nonpolar analytes.
Shielded hydrophobicphase (SHP) RAM consists of silica particles contain-
ing a bonded hydrophilic polyethylene oxide) network with imbedded hydropho-
bic phenyl moieties. In this type of RAM the network covers the entire surface of
the silica support-both inside and outside the pores. Small molecules can
diffuse through the network and interact with the imbedded phenyl groups
resulting in their retention. Hydrated macromolecules (e.g., proteins), on the
other hand, are excluded because of the hydrophilic shield provided by the
poly(ethylene oxide) network.
Mixed functionalphase (MFP) RAM consists of silica particles functionalized
with a hydrophilic diol group and a hydrophobic group that can be chosen from a
wide selection of hydrophobic moieties including phenyl [231,232], butyl [232],
octyl [232], and cyclodextrin [221,233].
A general drawback of RAMs is the limited flexibility in manipulating the
retention characteristics. This is explained by the fact that to avoid protein
precipitation during extraction, the mobile phase organic modifier content needs
to be kept at a low value (< 25% for acetonitrile, < 20% for 2-propanol, and < 10%
for THF) [223]. Since silica-based RAMs are predominantly used in current
analytical practice, a second inherent disadvantage of using RAMs is the narrow
pH range over which these materials can be operated.
Development of organic polymer-based RAMs could effectively overcome
this difficulty. However, to date, there are very few reports on the development
and use of totally organic polymer-based restricted-access materials, although
polymers have found applications in silica-based RAM technology as a semi-
permeable hydrophilic barrier 218,2191 and as a hydrophilic shielding network
[220]. Haginaka et al. [215,234-235] have recently described a procedure to
prepare molecularly imprinted sorbents with restricted-access properties
1040
(S)-naproxen
V-65 4-vinylpyridine
dibutyl phthalate toluene EDMA
swelling _ swelling swelling
seed particle
hydrophi
GMMA
GCMA
KS2 0,
polymerization hydrophilization
Scheme 32.3. Synthetic scheme of surface-modified MIP for (S)-naproxen (Reproduced from Ref.
[235], with permission of Elsevier).
1041
Fig. 32.6. Scanning electron micrographs of the MIP (A, C) and RAM-MIP (B, D) for
(S)-naproxen. Magnifications: 2000x (A, B) and 10000x (C, D). (Reproduced from Ref. [215] by
permission of American Chemical Society).
1042
34
29
- 24
E
19
14
-1
0 5 10 15 20 25 30 35
Time (min)
Fig. 32.7. HPLC chromatograms of urine samples obtained after (a) C-18 SPE (upper plot) and
(b) immunoextraction (lower plot) of hydroxyphenanthrenes (OH-PHEs) and hydroxypyrenes
(OH-PYRs). Conditions: 10-ml volumes ofurine sample from the same subject, final volume 3 ml,
100 lp injection volume. Peaks: (1) 2-/3-OH-PHEs, (2) 9-OH-PHE, (3) 1-OH-PHE, (4)
4-OH-PHE, and (5) 1-OH-PYR. (Reproduced from Ref. [252] by permission of American
Chemical Society).
1043
TABLE 32.1
Extraction recyclability through four extraction/stripping cycles with 33 mg of sol-gel in 5.0 ml of
SrCl 2 solutions. Extraction was conducted were conducted at 23°C. (Reproduced from Ref. [255],
with permission of American Chemical Society).
Extraction cycle [Sr2 + ] (ppm) Capacity Kd %Sr2+ removed
Initial Final
1 25.0 2.50±0.18 3.9x10- 2 1.4x10 3 90.0+0.2
2 32.25 5.2±0.4 4.7x10 2 7.8x10 2 83.8+0.4
3 24.7 0.29±0.02 2.8x10 -2 1.3x10 4 98.8+0.2
4 27.6 1.44+0.11 4.2x10 2 2.8x10 3 94.8-0.1
1044
Fig. 32.8. Schematic representation of transport through mem- Donor Membrane Acceptor
branes. (Reproduced from Ref. [263] by permission ofElsevier). phase phase
side of the membrane that receives the transferred mass is called the acceptor
phase or permeate phase. Figure 32.8 illustrates the selective mass transfer
through a membrane [263].
Membrane-based extraction and separation processes are based on differen-
tial transport rates of various chemical species through the interface. In mem-
brane-based systems, the mass transport process can be divided into three
categories [263,264]: (a) passive, (b) facilitated, and (c) active. Passive transport
is commonly used in applications dealing with analytical extractions or separa-
tions. In this mode, the membrane serves merely as a physical barrier. Here, the
driving force for the transport process is the electrochemical potential gradient
of various molecular or ionic species between the donor and acceptor phases
located on the two sides of the membrane. In the facilitated mode of transport, in
addition to the electrochemical potential gradient, a carrier is used in the mem-
brane to enhance the permeability during the transport process. Active trans-
port is typical of living cellular membrane where the transport takes place
against the electrochemical potential gradient, and is caused by a chemical reac-
tion taking place inside the membrane.
Based on the nature of the used membrane and the trapping media (if any),
membrane extraction techniques can be differentiated into: (a) supported liquid
membrane extraction (SLME) [40,42], (b) microporous membrane liquid-liquid
extraction (MMLLE) [42,265], (c) membrane extraction with sorbent interface
(MESI) [257,266-271], and (d) polymeric membrane extraction (PME)
[272-278]. Here the discussion will be confined primarily to the use of polymeric
materials in membranes extraction.
Luo et al. [273] have studied the effect of polymer swelling on the mass trans-
fer performance in membrane extraction. Fu et al. have reported the use of
potentiometry to study extraction thermodynamics of polyanions into plasticized
polymer membrane doped with lipophilic ion exchangers [275]. Melcher and
coworkers [276-279] developed and studied membrane extraction systems with
silicone rubber membranes and various combinations of aqueous/organic liquids.
Recently, Hauser and Popp [261,280] described a silicone rubber hollow
fiber-based dynamic membrane extraction device (Fig. 32.9) to enrich and
analyze volatile organic compounds from water (or air). Being battery-operated,
portable, and easy-to-interface with a portable GC, the device appears to be
1045
Peristaltic
Ambient air Water sample - pump l
1 Silicone hollow
fibre
I
Activated
charcoal -+
Tosample inlet
of AirmoBTX
Fig. 32.9. Schematic of a flow cell for dynamic extraction with hollow fibers. (Reproduced from
Ref. [280] by permission of Elsevier).
perfectly suitable for field operation. Volatile organic compounds in water are
allowed to diffuse through the silicone hollow fiber, and are enriched in the
integrated activated charcoal-based sorption tube. Typical enrichment time is 10
min. The enriched analytes are then thermally desorbed and analyzed by GC. A
typical example of BTEX analysis in water using dynamic membrane extraction
using a silicone hollow tube membrane is presented in Fig. 32.10. Low /zg/l
extraction sensitivity was achieved. Li and Zhang [281] developed poly(vinyl-
chloride)-based membranes to extract zinc ions from aqueous media. Mayer et
al. [108] compared the performance of octadecylsiloxane-bonded silica- and
porous polymer particle-loaded membranes in solid-phase extraction.
Recently, Jonsson and Mathiasson published extensive reviews on supported
liquid membrane extraction and related techniques [44,282]. Cordero et al. [263]
reviewed the interfacing of membrane extraction techniques with GC, HPLC,
and CE. Van de Merbel published an elaborate review on on-line coupled mem-
brane-based extraction techniques [2] covering principles, instrumentation,
Benzene Toluene
60000
50000
Ethylbenzene
40000
m-Xylene
30000
o-Xylene
20000
10000
I,
'Y-PL------
1, , m
0t fWY-lea
_- PLW s, i XI ]AI ,1
Fig. 32.10. Gas chromatogram obtained after extraction of 5 .og/l BTEX in water with a 0.3 m
silicone hollow fiber 0.7 x 0.9 mm. Temperature, 20°C; extraction time, 10 min. (Reproduced
from Ref. [280] by permission of Elsevier).
1046
interfacing (to GC, LC, and CE), and application aspects of dialysis, electro-
dialysis, and filtration techniques. Bauer [259] reviewed application of mem-
brane extraction in conjunction with mass spectrometry (membrane
introduction mass spectrometry, MIMS).
One limitation of liquid membrane-based extraction systems is the short
membrane lifetime. This can be overcome by using a polymeric membrane
instead of liquid membranes. Silicone rubber-based membranes can be a good
choice since they possess extended lifetimes. Another significant advantage of
polymeric membranes is that they are practically insoluble in aqueous or many
organic media, and allow for the use of extraction systems with various combina-
tions of aqueous and organic liquids. On the other hand, the solute diffusion
through polymeric membranes is slower than that in the liquid membranes.
This results in slower extraction by polymeric membranes. Another disadvan-
tage of using polymeric membranes is that the composition of the membrane
remains fixed during the extraction process, which translates into reduced
flexibility in fine-tuning the extraction selectivity. Since the silicone rubber-
based membranes are hydrophobic, their use as the extraction medium is
primarily limited to the extraction of nonpolar analytes, and may not be suitable
for the extraction of polar compounds. To overcome this difficulty, hydrophobic
polymeric membranes have been surface-modified with a thin layer of some a
hydrophilic polymer (e.g., polyacrylate) [283,284].
Polymeric membrane extraction is still an emerging area; an important
trend is the flow-through solid-phase extraction using polymeric membranes
(membrane SPE). This is consistent with the general trend of miniaturization in
separation and extraction technologies. Materialization of membrane extraction
in a flow-though mode should make this extraction technology much faster
compared with the traditional passive or facilitated modes commonly employed
in liquid membrane extraction techniques (SLME or MMLLE).
In a recent report, Fritz and Masso [285] described a miniaturized flow-
thorough membrane extraction device using a miniaturized (1.2 mm in height
and 0.7 mm in diameter) sulfonated polystyrene-divinylbenzene resin disk
housed inside the removable needle of a 50 Al Hamilton gas-tight syringe as
illustrated in Fig. 32.11. The membrane rested on a stainless steel screen. All
these parts were assembled in a Teflon ferrule-a standard part of the used
Hamilton syringe. Aqueous samples containing eight substituted benzene
compounds (at 10 ppb level concentration) were automatically passed through
the membrane using a single-syringe infusion pump. It required only 5 ul
volume of acetonitrile to elute the extracted compounds. Analyte recovery was
higher than 90% with an average relative standard deviation of 4.6%.
Another significant advancement in the area of membrane SPE is the recent
development of polymeric membranes with a surface-grafted molecularly im-
printed layer [272,287,288]. Wang et al. [287] prepared membranes with surf-
ace-grafted MIPs. For this, they used a special photoreactive polymer, poly-
(acrylonitrile-co-(diethylamino)dithiocarbamoylmethylstyrene). Using theo-
1047
<_ Miniaturized
membrane
<_ Teflon
ferrule
It
1048
50
MP
40-
330-
permission of Elsevier).
-5
structure. Herbicide concentrations, 10 M
in water. (Reproduced from Ref. [272] by o LL~~E
terbumeton atrazine desmertyn metribuzine
the nonspecific interaction between the membrane and the analytes (thereby
maximize the effect of template-specific interaction), at the same time retaining
the high permeability of the membranes without blocking the membrane pores.
They photopolymerized a methanolic solution containing a functional monomer
(2-acrylamido-2-methyl-l-propane sulfonic acid (AMPS), methacrylic acid
(MAA), or acrylic acid (AA)) and a cross-linker (N,N-methylene-bis-acrylamide)
(MBAA) using UV radiation and benzophenone as the photoinitiator. A rela-
tively hydrophobic triazine herbicide, terbumeton, was selected as the template
molecule. In contrast to their previous work on the development of a surface-
grafted MIP synthesis procedure using aqueous solutions [288], the focus here
was on the synthesis of MIP grafted surface layer using polymerizing solutions
prepared in an organic solvent. Such a method should greatly increase the
number of templates that can be used for the synthesis.
The prepared MIP-grafted membranes showed high specificity for the
template molecule (terbumeton), as is evident from Fig. 32.12. It should be noted
that flow-through membrane SPE is very attractive from an analytical point of
view, and new developments can be expected in this area in the near future.
1049
(A) 2
(B .- I
(B) 1 2--'
Left: Fig. 32.13. Modes of SPME operation: (a) direct extraction, (b) headspace SPME.
(Reproduced from J. Pawliszyn, B. Pawliszyn and M. Pawliszyn, The Chem. Educator,2 (1997) 1,
with permission of Springer).
Right: Fig. 32.14. Schematic of a sol-gel fiber (A) and a sol-gel open tubular microextraction
capillary (B). 1, sol-gel coating; 2, fused silica fiber (A) or capillary (B).
1050
silica capillary. A small segment of a wall-coated GC capillary column is often
used for this purpose. Figure 32.14 illustrates the construction of an SPME fiber
(Fig. 32.14A) and a wall-coated capillary for in-tube SPME (Fig. 32.14B).
Both in conventional and in-tube operation modes of SPME, the polymeric
coating is exposed to the environment containing the target analytes, and time is
allowed for an equilibrium to establish. Typically, it takes less than an hour to
reach the extraction equilibrium. The extraction process is aided either by
mechanically agitating the sample with a stir bar (as in conventional SPME) or
by flowing the sample over the coated surface (as in in-tube SPME or capillary
microextraction). After this, the fiber or the capillary is withdrawn from the
sample, and the extracted analytes are rapidly desorbed from the coating for
introduction into an analytical instrument-most commonly a GC, HPLC, SFC
or a CE system. The analyte desorption can be carried out either thermally by
introducing the fiber (with the extracted analytes on its coating) into a heated
GC injection port using an SPME syringe specially designed for SPME-GC anal-
ysis, or by rinsing the coated capillary with an appropriate solvent as is per-
formed in in-tube SPME/HPLC analysis.
SPME has gained wide popularity because of its simplicity, speed, solventless
extraction capability, portability, and suitability for automation. A fascinating
attribute of SPME techniques is that they simultaneously accomplish a number
of important analytical operations in a very simple way. These include sampling,
sample preparation, and analyte preconcentration. They also provide a robust
mechanism for the introduction of the extracted analytes into an analytical
instrument for quantitation. In addition, the fact that the extracted amount of
the analyte in SPME is practically independent of the sample volume [291]
makes the techniques especially valuable for direct field sampling and analysis.
In SPME techniques, a vital analytical role is played by the polymeric coating
that serves as the extraction medium, thus making the extraction process
environmentally benign by eliminating the need for hazardous organic solvents
that are used in large volumes by conventional extraction techniques.
Conventional sorbents for SPME include two types of extraction media
[314]: (a) homogeneous polymers and (b) composite extraction media consisting
of a partially cross-linked polymer embedded with porous particles of a second
component. New polymers for SPME coatings include: (1) Nafion, (2) Poly-
pyrrole, (3) molecularly imprinted polymers, and (4) sol-gel hybrid organic-inor-
ganic polymers (sol-gel PDMS, sol-gel PEG, sol-gel-crown ether, sol-gel
dendrimer, etc.). From the polarity point of view, conventional SPME coatings
can be classified into four categories: (1) nonpolar (e.g., PDMS), (2) moderately
polar (e.g., PDMS/DVB composite), (3) polar (e.g., polyacrylate), and (4) highly
polar (e.g., Nafion).
Since PDMS is nonpolar in nature, it can effectively extract nonpolar
analytes from a wide range of matrices. However, PDMS coatings are not
efficient for the extraction of polar analytes. More polar SPME coatings are
needed for the extraction of polar analytes. Polyacrylate (PA)-based coatings
1051
[317-322] are more suitable for this type of applications. Polyethylene glycol
(PEG)-based coatings (e.g., Carbowax 20M and Carbowax/divinylbenze) are also
used [323-325]. It should be mentioned that relatively thick coatings are to be
used in SPME to provide reasonable sample capacity and extraction sensitivity.
1052
linked polymeric component. Unlike the preparation of homogeneous polymeric
coatings, such a process is difficult to automate and is usually done manually.
Because of the presence of two sorbents in the coating, these coatings provide
improved selectivity. In addition, the fiber selectivity can be fine tuned for the
analytical problem at hand by proper selection of the two coating components. It
should be mentioned that such composite coatings possess lower mechanical
stability than the homogeneous polymer coatings. The following commercial
SPME coatings belong to this category: (1) polydimethysiloxane/divinylbenzene
(PDMS/DVB), (2) polydimethylsiloxane/Carboxen (PDMS/Carboxen), (c) Carbo-
wax/divinylbenzene (CW/DVB), and (d) Carbowax/templated resin (CW/TPR).
In written notation of these composite coatings, the non-particulate component
is usually put first, followed by a "/" sign that separates it from the particulate
component that follows. Small pieces (-2 cm) of these externally coated fibers
are then installed on individual fiber assembly to be used for extraction.
1053
100000
80000
9 60000
40000 a)
b)
1054
To this end, Pawliszyn and coworkers [304-307,327] recently developed
poly(pyrrole)-based coatings for SPME. Their main emphasis was to develop
polypyrrole (PPY) and poly(N-phenylpyrrole) (PPPY) coatings for reliable
hyphenation of in-tube SPME with HPLC for automated on-line extraction and
HPLC analysis. Outstanding solvent stability of these polymers made them an
appropriate choice for this job, since in such an interface the desorption of the
extracted analytes is normally carried out using organic-rich solvents. Another
important attribute of polypyrrole-based coatings is their outstanding pH stabil-
ity. For example, PPY coatings provided stable performance within a wide pH
range (1.5-10) [307]. PPY coatings possess a porous structure with an extraction
mechanism primarily governed by surface adsorption of the analytes. Table 32.2
compares the SPME performance of PPY coatings with that of commercially
available capillaries with the following coatings: (a) polar silica (host), (b)
Omegawax (Omeg), (c) SPB-1, and (d) SPB-5. The data clearly demonstrate the
superiority of the PPY coating over its commercial counterparts in the extrac-
tion of PAHs. The authors explained this enhanced extraction capability of PPY
TABLE 32.2
Comparison of the extraction efficiencies for PAHs on different capillary coatings (adapted from
Ref. [307], with permission of American Chemical Society)
PAH Compound Detector Amount of analyte extracted (ng)c or extraction yield
responseb (%)d
F
Host Omeg SPB-1 SPB-5 PPY
Naphthalene 0.058 0.9 1.7 1.9 3.0 5.9
Acenaphthylene 0.046 0.6 1.8 2.5 3.7 7.6
Acenaphthene 0.027 0.7 1.8 4.1 4.6 6.7
Fluorene 0.213 1.2 2.1 3.8 6.3 8.1
Phenanthrene 0.050 0.8 2.5 4.4 5.4 11.1
Anthracene 0.024 0.8 2.6 4.9 7.3 11.8
Fluoranthene 0.090 1.0 3.2 5.8 10.3 17.5
Pyrene 0.115 1.1 3.1 6.1 12.0 18.0
Benza[a]anthracene 0.082 1.5 4.8 6.1 9.9 20.7
Chrysene 0.052 1.4 4.4 5.2 9.2 18.7
Benzo[b]fluoranthene 0.072 1.9 6.3 6.4 8.6 23.3
Benzo[k]fluoranthene 0.107 1.5 5.8 5.7 8.4 15.7
Benzo[a]pyrene 0.079 1.8 6.0 6.0 8.8 18.7
Dibenz(a.h)anthracene 0.098 1.4 4.9 2.9 6.1 10.7
Benzo(ghi)perylene 0.099 1.6 5.2 3.1 6.7 10.8
Indeno(1,2,3-cd)pyrene 0.088 19 7.3 6.3 8.8 16.0
aCompound concentrations and experimental conditions (see notes b, c, and d).
bF (detector response factor) was obtained by injecting 10 l of 1,g/ml solution (for each PAH 10
ng was injected).
CA 1-ml sample (100 ng/ml for each PAH) was analyzed by in-tube SPME, and the amount of
analyte extracted (nA) was calculated by equation: nA = FA = (m/Ad)A; the extraction yield (%)=
nA x 100/M where is the total amount of analyte in the 1 ml solution (see Ref. [307] for details).
dSelectivity factor was calculated only for 15 cycles on the basis of nA relative to n (naphthalene).
1055
44.0
PPPY Film
19.5
14.5
9.5
coatings based on effective T-Ti and hydrophobic interactions between the PPY
coatings and PAH solutes. Enhanced extraction by PPY-coated capillaries was
also observed for polar compounds, such as phenol and its derivatives. However,
PPY-coated capillaries showed reduced extraction efficiency for monocyclic aro-
matic hydrocarbons compared with commercial capillaries. PPY-coated capillar-
ies also showed enhanced extraction efficiency for organoarsenic compounds and
for anionic species [305]. This may be explained based on the multifunctional
nature of the polymer and the positive charge that it carries.
Another significant advantage of PPY coatings was found to be their very lit-
tle affinity for nonpolar solvents like hexane. This fact, combined with high sol-
vent stability of PPY coatings, opens new possibility of extracting polar
impurities in nonpolar organic solvents. Figure 32.16 shows a chromatogram,
illustrating SPME (followed by GC analysis) of lower aliphatic alcohols from
hexane using PPY coating. It should be pointed out that PPY coatings are held
on the substrate by physical forces. Their high solvent stability is due to poor sol-
ubility of these polymers in a wide range of common organic solvents.
1056
and solvent stabilities. In most part, this is due to the lack of chemical bonding
between the stationary phase coating and the inner walls of the capillary to
which these coatings are applied. Sol-gel coating technology [332-336] easily
overcomes these problems by providing surface-bonded organic-inorganic
hybrid stationary phase coatings. The sol-gel coating technology is applicable for
creating stationary phase coatings both on the outer surface of conventional
SPME fibers and on the inner walls of the capillary used for in-tube SPME (cap-
illary microextraction). In addition, sol-gel technology can be used for the cre-
ation of both thin and thick coatings. Finally, sol-gel technique is very much
suitable for in situ creation of porous monolithic beds [337-352] within a fused
silica capillary using a properly designed sol solution that solidifies and turns
into a single piece of porous structure chemically bonded to the inner walls of the
capillary. Monolithic beds are analogs to particle-packed sorption beds. How-
ever, because of the facts that the porous sol-gel monolithic bed is created with-
out packing sorbent particles, and that the bed is directly bonded to the inner
walls of the capillary, sol-gel monolithic beds are characterized by enhanced
mechanical stability.
1057
TABLE 32.3
Chemical ingredients of a sol solution designed for the creation of a sol-gel PDMS coating.
(Reproduced from Ref. [335], with permission of the Royal Society of Chemistry).
Sol-gel solution Chemical name Function
ingredient
Sol-gel precursor Methyltrimethoxy- Serves as a source for the inorganic component
silane of the sol-gel organic-inorganic hybrid coating
and delivers it through hydrolytic polycondens-
ation reactions. Also provides active silanol
groups on the sol-gel coating to facilitate its
chemical bonding to the fiber/capillary surface.
Organic polymer 1) Hydroxy-terminated Provides organic component for the sol-gel
with sol-gel-active polydimethylsiloxane organic-inorganic hybrid coating through
functional groups polycondensation reactions with the hydrolyzed
products of the precursor.
Deactivation Poly(methylhydro- Provides desired level of high-temperature
reagent siloxane) derivatization of silanol groups on the sol-gel
coating and thus allows for the control of the
organic-inorganic hybrid coating polarity. The
derivatization reaction takes place during
thermal conditioning of the fiber after
completion of the sol-gel coating procedure.
Sol-gel catalyst Trifluoroacetic acid Plays a dual role int he sol-gel process: (a) as an
(TFA) acid-catalyst for the sol-gel process; (b) as a
source of water for the hydrolysis of the pre-
cursor (the commercial TFA used in this work
contained approx. 5% of water). It thus provides
a mechanism for the sol-gel process to be con-
ducted in organic solvent-rich environment with
controlled amount of water.
imately 1 cm end segment of a fused silica fiber using a cigarette lighter. The
bare end of the fiber was then carefully cleaned by using a piece of Kimwipe
tissue paper soaked with methanol, making sure that no soot particles remained
on the surface. The exposed end of the fiber was further treated with a 0.1 M
NaOH solution for 30 min to enhance the concentration of surface silanol groups
that would facilitate chemical bonding of the sol-gel coating to the exposed
surface. The treated end of the fiber was thoroughly washed with deionized
water. At this point, the fiber was ready for coating. To coat, the treated end of
the fiber was vertically dipped into the previously prepared sol solution placed in
a vial. The fiber was allowed to stay in this position for -20 min. Following this,
the fiber was withdrawn from the solution and subjected to a stepwise (or
programmed) thermal conditioning in a stream of helium. For a sol-gel PDMS-
coated fiber, the final conditioning temperature was as high as 360°C. If
necessary, the entire process of dipping, withdrawing, and thermal conditioning
of the fiber should be repeated to obtain the desired coating thickness. This
procedure was used and adapted by Wang et al. [353] to prepare sol-gel polyeth-
1058
Fig. 32.17. Scanning electron micrographs of a sol-gel PDMS-coated SPME fiber: (A) cross-
sectional view, 342x magnification, and (B) surface view, 3420x magnification. (Reproduced
from Ref. [333] by permission of American Chemical Society).
ylene glycol (PEG) coated SPME fibers, and by Zeng et al. [354-356] to prepare
SPME fibers with sol-gel crown ether coatings.
Scanning electron micrographs in Fig. 32.17 illustrate the cross-sectional
view (Fig. 32.17A) and surface view (Fig. 32.17B) of a sol-gel PDMS-coated
SPME fiber. They reveal that the sol-gel PDMS coating created on the SPME
fiber is remarkably uniform (Fig. 32.17A) and has a porous structure (Fig.
32.17B) that provides enhanced surface area, and facilitates fast mass transfer
and improved sample capacity.
Application of sol-gel coatings to the innersurface of the capillaryfor in-tube
SPME. A general procedure for the creation of sol-gel stationary phase coating
on the inner walls of a fused silica capillary was first described by Malik and
coworkers [332]. Briefly, to prepare a sol-gel PDMS-coated capillary, a hydro-
thermally treated fused silica tubing (e.g., 10 m x 250 gm i.d.) was filled with the
sol-gel solution using a home-made filling/purging device [334] under 100 psi
helium pressure. The solution was allowed to stay inside the capillary for 10-30
min, after which the solution was expelled from the capillary under the same
helium pressure. Following this, the capillary was dried by purging with helium
for 30 min at room temperature. The capillary was then thermally conditioned
under a continuous flow of helium: from 40°C to 350°C at 1°C min', holding the
column at the final temperature for 5 h. Finally, the capillary was rinsed with
methylene chloride/methanol (50:50 v/v) mixture and dried under helium purge.
At this point, the capillary was ready for use in capillary microextraction. The
sol-gel coated capillary was cut into small segments (3.5 cm and longer) for use in
capillary microextraction. Following the same procedure, longer sol-gel coated
1059
Fig. 32.18. Scanning electron micrograph of a
sol-gel PDMS-coated tubing for capillary
microextraction. Magnification 10000x.
(Reproduced from Ref. [308] by permission of
American Chemical Society).
1060
I. Hydrolysis of tilhesol-gel precursor: (1 alkyl or alk.xy groups, and R' = alkyl or hydroxy functionalities)
R R'
I Ca(alyst I
C1130-Si-OCII 3 + 20
0 HO-Si-O
O-Si- + C 3OII
OC113 OIl
i I
0 CH3 CH3 H3
_O it O-+iiOf o--+ li-OS --- oi-O-rSi-OH
? ? ClH3 CHj CH
3
UI I
-OH 0° 0° ICH
3 ICH
3 CH
3
-O. + HO- SiO-ito-i-O
. ,i_-oiii-
Column Wall ? ? CH 3 CH 3 CH 3
Scheme 32.4. Chemical reactions involved in sol-gel coating with hydroxy-terminated PDMS
stationary phase. (Reproduced from Ref. [332] with permission of American Chemical Society).
1061
s
3 4
I
TABLE 32.4
Effects of desorption temperature on the peak area repeatability data for PAHs and dimethyl-
phenols in SPME-GC using sol-gel PDMS-coated fibers conditioned at different temperatures.
Conditions: extraction time for each sample (room temperature), 30 min; split-splitless injection
(splitless desorption for 5 min), injection temperature is the same as the fiber conditioning
temperature; 10 m x 250 pm i.d. sol-gel PDMS capillary GC column; column temperature
programmed from 40°C (hold for 5 min) to 230°C at 6°C/min; detector: FID, 350°C. (Adapted
from Ref. [333] with permission of American Chemical Society).
Compound %RSD (n = 3)
Fiber conditioning temperature
250°C 300°C 310°C 320°C
(A) Polycyclic aromatic hydrocarbons
Naphthalene 1.67 1.04 1.01 0.98
Acenaphthylene 3.86 3.72 2.65 2.01
Fluorene 9.25 3.33 3.20 3.00
Phenanthrene 5.93 0.65 0.52 0.50
Anthracene 3.52 1.08 1.05 1.01
Fluoranthene 2.61 1.02 0.80 0.64
(B) Dimethylphenol isomers
2,5-Dimethylphenol 7.14 3.74 0.84 0.58
3,4-Dimethylphenol 4,72 3.11 2.84 1.79
2,6-Dimethylphenol 5.91 4.84 1.03 1.00
2,3-Dimethylphenol 8.73 4.78 2.65 1.23
1062
TABLE 32.5
The extraction precision of the sol-gel PEG coated fibers conditioned at 280°C. The evaluation
was made over an extended period of operation consisting of 150 runs. M.P. stands for mean peak
area. (Reproduced from Ref. [353] with permission of Elsevier).
Compound Extraction period
10th 50th 100th 150th
M.P. RSD M.P. RSD M.P. RSD M.P. RSD
( (%) (%) (%)
Benzene 14,355 6.2 14,210 5.3 14,307 5.7 14,345 5.8
Toluene 18,301 5.9 18,219 5.7 18,269 6.1 18,247 6.0
Ethylbenzene 25,220 4.7 25 341 5.1 25,274 5.3 25,297
p-Xylene 26,691 3.9 26,536 4.4 26,598 3.7 26,674 4.1
o-Xylene 29,732 4.4 29,775 4.1 29,826 4.7 29,793 3.9
1063
300
250
200
x
2 150
,-A
1t00 *-B
'-C
50 -D
-E
0
Fig. 32.20. Extraction time profile for 10 ng/ml BTEX (constant stirring at room temperature).
(A) Benzene; (B) toluene; (C) ethylbenzene; (D)p-xylene; (E) o-xylene. GC conditions: 40°C for 2
min, then programmed at 7°C/min to 110°C, and then 20°C/min to 130°C for 1 min; injection
temperature, 280°C; FID, 300°C; desorption time 20 s. (Reproduced from Ref. [353] by
permission of Elsevier).
1064
TABLE 32.6
Names and chemical structures of sol-gel reagents used to prepared a sol-gel ODS monolithic bed
in a fused silica capillary. (Reproduced from Ref. [337] with permission of American Chemical
Society).
Reagent function and reagent name Reagent structure
C13] OCH3
1065
Fig. 32.22. Scanning electron micrographs of a sol-gel ODS monolithic capillary: (A)
cross-sectional view, magnification 1800x; (B) longitudinal view of the wall region showing
anchorage of the monolithic structure to the capillary inner surface, magnification 1000x.
(Reproduced from Ref. [337] by permission of American Chemical Society).
anchorage to the capillary inner walls are presented in Figs. 32.22A and B,
respectively.
Scheme 32.5 presents the chemical reactions involved in the preparation of
sol-gel ODS monolithic bed that can be used either in capillary electrochroma-
tography (CEC), capillary LC, or in capillary microextraction.
Advantages of sol-gel monoliths as extraction media. Like sol-gel coatings,
sol-gel monolithic beds are also chemically bonded to the silica substrates to
which they are applied (e.g., to the inner surface of a fused silica capillary) (Fig.
32.22B). This chemical bonding gives mechanical stability to the extraction bed.
Unlike particle-packed extraction beds (as in SPE cartridges or HPLC columns),
in monolithic beds there are no free-to-move particles in the bed. The bed rather
represents a single piece of porous solid anchored to the capillary inner walls
through chemical bonding.
Sol-gel monolithic beds are characterized by enhanced permeability. Sol-gel
reaction conditions can be easily optimized to obtain monolithic beds with an
order of magnitude (or more) higher permeability than the conventional parti-
cle-packed beds. The presence of flow-though channels makes the monolithic
bed an effective medium for fast extraction. Like sol-gel polymeric coatings,
sol-gel monolithic beds are also characterized by enhanced thermal and solvent
stability thanks to their direct chemical bonding to the capillary inner walls. As a
result, capillary microextraction with monolithic beds can be easily interfaced,
either with a liquid-phase or a gas-phase separation technique.
Sol-gel monolithic capillary contains a significantly higher amount of the
sol-gel extraction medium compared with that on an SPME fiber or an open
tubular microextraction capillary. This translates into higher sample capacity,
and consequently, higher sensitivity of sol-gel monolithic microextraction com-
pared with conventional or in-tube SPME (open tubular capillary microextrac-
tion). Figure 32.23 presents two gas chromatograms illustrating extraction
1066
Complete Hydrolysis of N-Octadecyldimethyl[3-(trihydroxysilyl)propyl]ammonium Chloride
CH HCH o
CH(CH2 ) . t )Si-oCH
(C11 2 ?cH3 + 3 H20 c C(CH
o .
3 2)ir I (CH) 3 -Si-SI-OH + 3 CH 3OH
CH 3 OCH 3 CH3 OH
MCCH- OH H H3 ,
CH3 OH Ca, 0 0
OH O l13
Scheme 32.5. Chemical reactions involved in the preparation of sol-gel ODS stationary phase.
(Adapted from Ref. [336] with permission of American Chemical Society).
1067
1nnn
(B)
900 900 Open Tubular
800 800 -
700 700 -
o 600 600 -
M. 500 500 -
E
400 400 -
300 300 -
200 200 - o
O
100 100- : x
LL I
0 0-
0 2I I
I 10 20 30 min
Fig. 32.23. Comparison of the extraction sensitivity in capillary microextraction using: (A) a
sol-gel ODS monolithic capillary (50 Am i.d.) and (B) a sol-gel ODS-coated fused silica capillary
(250 Am i.d., -1 gm coating thickness). (From Ref. [368]).
architecture, dendrimer-based materials also have great potential for being used
in analytical chemistry (e.g., as effective extraction media, as selective stationary
phases or mobile phase additives for various chromatographic and electro-
migration separation techniques). However, only on rare occasions have these
analytical potentials of dendrimers been utilized [70,375-380]. Recently,
Newkome et al. [381] described a method for the synthesis of phenyl-terminated
dendrimers with a trimethylsilyl derivatized sol-gel-active root (Fig. 32.24).
Using sol-gel coating technology [232], the authors chemically bonded these den-
dritic moieties to the fused silica capillary inner surface in the form of a station-
ary phase coating. These sol-gel dendrimer-coated capillaries were further used
as GC columns that showed unique selectivity. In an ongoing project, these
authors have made use of these sol-gel dendrimer-coated capillaries to perform
microextraction [382].
Figure 32.25 shows a gas chromatogram of polycyclic aromatic hydrocarbons
that were extracted from an aqueous matrix using a sol-gel dendrimer-coated
capillary. In principle, a wide variety of dendrimers can be prepared and easily
bonded to silica particles for use as solid-phase extraction media or HPLC sta-
tionary phase. It should also be possible to prepared sol-gel monolithic beds with
dendritic ligands. Such monolithic beds can also be used as extraction media or
1068
- f)
(EtO) 3SiCH 2CH 2CH 2NH
32.7 CONCLUSION
1069
Fig. 32.25. CME-GC analysis of PAHs using a 4
I
sol-gel dendrimer-coated fused silica extraction
capillary. Extraction conditions: gravity-fed
extraction, 30 min; post extraction purge with
helium, 2 min. GC conditions: sol-gel PDMS
capillary column (10 m x 0.25 mm i.d.); desorp-
tion temperature, from 30 to 300°C; column
temperature programming, from 30°C (5 min) at
'IAs'
15°C/ min; detector, FID (300°C). (From Ref. 5 10 15 20 25 30
[382]). Minutes
1070
REFERENCES
1071
37 M. Ma and F.F. Cantwell, Anal. Chemn., 71 (1999) 388.
38 H. Liu and P.K. Dasgupta, Trends Anal. Chem., 15 (1996) 468.
39 A.L. Theis, A.J. Waldack, S.M. Hansen and M.A. Jeannot, Anal. Chem., 73 (2001)
5651.
40 G.A. Audunsson, Anal. Chem., 58 (1986) 2714.
41 R.A. Bartsch and J.D. Way, Chemical Separation with Liquid Membrane. ACS,
Washington, D.C., 1996.
42 J.A. Jonsson and L. Mathiasson, Trends Anal. Chem., 18 (1999) 318.
43 J.A. Jonsson and L. Mathiasson, Trends Anal. Chem., 18 (1999) 325.
44 J.A. Jonsson and L. Mathiasson, Adv. Chromatogr., 41 (2001) 53.
45 D. Kou, A.S. Juan and S. Mitra, Anal. Chem., 73 (2001) 5462.
46 M.C. Hennion, J. Chromatogr. A, 856 (1999) 3.
47 F. Lanza and B. Sellergren, Adv. Chromatogr.(2001) 41, 53.
48 N. Masque, R.M. Marce and F. Borrull, Trends .Anal. Chem., 20 (2001) 477.
49 M.E. Leon-Gonzales and L.V. Perez-Arribas, J. Chromatogr. A, 902 (2000) 3.
50 J. Pawliszyn, Solid-phase Microextraction. Theory and Practice. Wiley, New York,
(1997.
51 J. Pawliszyn (Ed.), Applications of Solid-Phase Microextraction. Royal Society of
Chemistry, Cambridge, UK, 1999.
52 M. de Fatima Alpendurada, J. Chromatogr.A, 889 (2000) 3.
53 K.G. Furton, J. Wang, Y.L. Hsu, J. Walton and J.R. Almirall, J. Chromatogr.Sci., 38
(2000) 297.
54 N.H. Snow, J. Chromatogr.A, 885 (2000) 445.
55 G. Theodoridis, E.H. Koster and G.J. de Jong, J. Chromatogr. B, 745 (2000) 49.
56 G.A. Mills and V. Walker, J. Chromatogr.A, 902 (2000) 267.
57 A.D. Harmon, Food. Sci. Technol., 79 (1997) 81.
58 Y. Ikada, Biomaterials, 15 (1994) 725.
59 R.P. Belardi and J. Pawliszyn, Water Pollut. Res. J. Canada, 24 (1989) 179.
60 M.R. Buchmeiser, Chem. Rev., 100 (2000) 1565.
61 T. Sekine and Y. Hasegawa, Solvent Extraction Chemistry. Fundamentals and
Applications, Marcel Dekker, New York, (1997.
62 J.G. Huddleston, H.D. Willauer, S.T. Griffin and R.D. Rogers, Ind. Eng. Chem. Res.,
38 (1999) 2523.
63 H. Ishii, M. Junichiro and H. Watanabe, Bunseki Kagaku, 26 (1977) 252.
64 W.L. Hinze and E. Pramauro, Crit. Rev. Anal. Chem., 24 (1993) 133.
65 H. Tani, T. Kamidate and H. Watanabe, J. Chromatogr. A, 780 (1997) 229.
66 J.F. Scamehorn and J.H. Harwell, Surfactants in Chemical/Process Engineering:
Surfactant Science Series. Marcel Dekker, New York, 1988.
67 J.G. Huddleston, A. Veide, K. Kohler, J. Flanagan, S.O. Enfors and A. Lyddiatt,
Trends Biotechnol., 9 (1991) 381.
68 H.O. Johansson, G. Karlstrom, B. Mattiasson and F. Tjerneld, Bioseparation, 5
(1995) 269.
69 E.L. Goetheer, M.W. Baars, L.J. van den Broeke, E.W. Meijer and J.T. Keurentjes,
Ind. Eng. Chem. Res., 39 (2000) 4634.
70 M.W. Baars, P.E. Froehling and E.W. Meijer, Chem. Commun., 1997 (20) (1997)
1959.
71 K.R. Powell, T.M. McCleskey, W. Tumas and J.M. DeSimone, Ind. Eng. Chem. Res.,
40 (2001) 1301.
72 H. Kanazawa and Y. Matsushima, Adv. Chromatogr., 41 (2001) 311.
73 T. Okano, Y.H. Bae and S. Kim, in: J. Kost (Ed.), Pulsed and Self-Regulated Drug
Delivery. CRC Press, Boca Raton, FL, 1990, p. 17.
74 N. Ogata, Polym. Prepr., 37 (1996) 113.
1072
75 C. Ganorkar, F. Liu, M. Baudys, et al. Polym. Prepr., 38 (1997) 560.
76 N.Y. Rapoport, W.G. Pitr and J. Pruitr, Polym. Prepr., 41 (2000) 716.
77 J.E. Chung, M. Yamato, M. Yokoyama, et al. Polym. Prepr., 39 (2000) 178.
78 N. Kubota, T. Matsubara and Y. Eguchi, J. Appl. Polym. Sci., 70 (1998) 1027.
79 T. Peng and Y.L. Cheng, J. Appl. Polym. Sci., 70 (1998) 2133.
80 Y.G. Takei, T. Aoki, K. Sanui, N. Ogata, T. Okano and Y. Sakurai, Bioconjugate
Chem., 4 (1993) 341.
81 Y.G. Takei, T. Aoki, K. Sanui, et al. Bioconjugate Chem., 5 (1994) 577.
82 T. Okano, N. Yamada, H. Sakai and Y. Sakurai, J. Biomed. Mater. Res., 27 (1993)
1243.
83 T. Takezawa, Y. Mori, T. Yonaha and K. Yoshizato, Exp. Cell Res., 208 (1993) 430.
84 K. Hosoya, K. Kimata, T. Araki, N. Tanaka and J.M. Frechet, Anal. Chem., 67 (1995)
1907.
85 H. Kanazawa, Y. Kashiwase and K. Yamamoto, Anal. Chem., 69 (1997) 823.
86 H. Kanazawa, Y. Matsushima and T. Okano, Trends Anal. Chem., 17 (1998) 435.
87 H. Lakhhirai, T. Okano, N. Nurdin, et al. Biochim. Biophys. Acta, 1379 (1998) 303.
88 H. Kanazawa, T. Sunamoto, Y. Matsushima, A. Kikuchi and T. Okano, Anal. Chem.,
72 (2000) 5961.
89 J. Kobayashi, A. Kikuchi, K. Sakai and T. Okano, Anal. Chem., 73 (2001) 2027.
90 C. Matsubara, S. Izumi, K. Takamura, S. Yoshioka and Y. Mori, Analyst, 118 (1993)
553.
91 C. Matsubara, N. Kikuchi and K. Takamura, Bunseki Kagaku, 44 (1995) 311.
92 M. Matsukata, T. Aoki, K. Sanui, N. Ogata, A. Kikuchi, Y. Sakurai and T. Okano,
Bioconjugate Chem., 7 (1996) 96.
93 T. Saitoh, T. Ohyama, T. Sakurai, T. Kaise and C. Matsubara,Anal. Sci., 13 (1997) 1.
94 T. Saitoh, T. Sakurai, T. Kaise and C. Matsubara, Anal. Sci., 13 (1997) 181.
95 T. Saitoh, T. Ohyama, K. Takamura, T. Sakurai, T. Kaise and C. Matsubara,
Talanta, 46 (1998) 541.
96 T. Saitoh, M. Haga, T. Sakurai, T. Kaise and C. Matsubara, Anal. Sci., 14 (1998) 929.
97 T. Saitoh, Y. Yoshida, T. Matsudo, S. Fujiwara, D. Akira, K. Iwaki, Y. Suzuki and C.
Matsubara, Anal. Chem., 71 (1999) 4506.
98 T. Saitoh, Y. Yoshida and C. Matsubara, J. Chromatogr.A, 891 (2000) 69.
99 T. Ohyama, K. Arai, T. Nakagawa, C. Matsubara and K. Takamura, Bunseki
Kagaku, 46 (1997) 59.
100 G.C. Chen and A.S. Hoffman, J. Biomater. Sci. Polym. Edn., 5 (1994) 371.
101 C. Cole, S.M. Schreiner, J.H. Priest, N. Monji and A.S. Hoffman, ACS Symposium
Series 350. ACS, Washington, DC, 1987, p. 245.
102 M. Heskins, J.E. Guillet and E. James, J. Macromol. Sci. Chem. A, 2 (1968) 1441.
103 M.C. Hennion and V. Pichon, Environ. Sci. Technol., 28 (1994) 576A.
104 M.C. Hennion, J. Chromatogr.A, 885 (2000) 73.
105 E. Matisova and S. Skrabakova, J. Chromatogr.A, 707 (1995) 145.
106 R.K. Iler, The Chemistry of Silica. Wiley, New York, 1979.
107 M.L. Lee, F.J. Yang and K.D. Bartle, Opem Tubular Column Gas Chromatograhy.
Wiley, Ner York, 1984.
108 M.L. Mayer, S.K. Poole and C.F. Poole, J. Chromatogr.A, 697 (1995) 89.
109 K.K. Unger, Packings and stationary phases in Chromatographic Techniques.
Marcel Dekker, NY, 1990.
110 S.O. Akapo and C.F. Simpson, J. Chromatogr.Sci., 28 (1990) 186.
111 M. Ashraf-Khorassani, L.T. Taylor and R.A. Henry, Anal. Chem., 60 (1988) 1529.
112 K.D. Altria, N.W. Smith and C.H. Turnbull, Chromatographia,46, (1997) 664.
113 M. Pursch and L.C. Sander, J. Chromatogr.A, 887 (2000) 313.
114 V.G. Berezkin, A. Malik and V.S. Gavrichev, J. High Resolut. Chromatogr./
1073
Chromatogr. Commun., 6 (1983) 388.
115 A. Malik, V.G. Berezkin and V.S. Gavrichev, Chromatographia,22 (1986) 117.
116 T. Takeuchi, M. Watanabe, T. Hamanaka, H. Haraguchi and D. Ishii, J. High
Resolut. Chromatogr., 13, (1990) 606.
117 J. Zhao and P.W. Carr, Anal. Chem., 72 (2000) 4413.
118 B. Yan, J. Zhao, J.S. Brown, J. Blackwell and P.W. Carr, Anal. Chem., 72 (2000) 1253.
119 J. Nawrocki, M.P. Rigney, A. McCormick and P.W. Carr, J. Chromatogr., 657 (1993)
229.
120 Y. Hu and P.W. Carr, Anal. Chem., 70 (1998) 1934.
121 J. Yu and Z. El Rassi, J. Chromatogr., 631 (1993) 91.
122 J. Yu and Z. El Rassi, J. High. Resolut. Chromatogr., 17 (1994) 705.
123 Z.T. Jiang and Y.M. Zuo, Anal. Chem., 73 (2001) 686.
124 M. Kawahara, H. Nakamura and T. Nakajima, J. Chromatogr., 515 (1990) 149.
125 A. Kurganov, U. Trudinger, T. Isaeva and K. Unger, Chromatographia,42 (1996)
217.
126 K. Tani and Y. Suzuki, J. Liq. Chromatogr. &. Rel. Technol., 19 (1996) 3037.
127 J. Winkler and S. Marme, J. Chromatogr. A, 888 (2000) 51.
128 C.H. Cars, L.M. van Elven, A.M. van Herk and A.L. German, Br. Polym. J., 21 (1989)
133.
129 H. Engelhardt and W.M. Cunat, Chromatographia,40 (1995) 657.
130 H. Figge, A. Deege, J. Kohler and G. Schomburg, J. Chromatogr., 351 (1986) 393.
131 Y. Shen, X. Shao, K. O'Neill, J.S. Bradshaw and M.L. Lee, J. Chromatogr. A, 866
(2000) 1.
132 G. Zhang and B. Xu, J. Chromatogr.A, 912 (2001) 335.
133 J.J. Kirkland, J.W. Henderson, J.J. Destefano, M.A. Vanstraten and H.A. Claessens,
J. Chromatogr.A, 762 (1997) 97.
134 A.V. Kiselev and J.P. Yashin, Gas Adsorption Chromatography.Plenum Publishing,
New York, 1969.
135 I. Halasz and C. Horvath, Anal. Chem., 34 (1969) 409.
136 J.H. Knox and P. Ross, Advances in Chromatography, Volume 37. Marcel Dekker,
New York, 1997.
137 F.D. Shire and B. McEnaney, Energeia, 2 (1991) 1.
138 J. Shurrock, M. Bauer, M. Holmes and A. Rachwal, Disinfection. By. Products. in.
Drinking. Water. Current.Issues. The Royal Society of Chemistry, Cambridge, UK,
(1999.
139 A. Di Corcia and R. Samperi, Anal. Chem., 62 (1990) 1490.
140 A. Di Corcia, S. Marchese and R. Samperi, J. Chromatogr., 642 (1993) 163.
141 B. Altenbach and W. Giger, Anal. Chem., 67 (1995) 2325.
142 C. Crescenzi, A. Di Corcia, G. Passariello, M.I. Turnes Carou and R. Samperi, J.
Chromatogr.A, 733 (1996) 41.
143 F. Mangani and R. Cenciarini, Chromatographia,41 (1995) 678.
144 D. Puig and D. Barcelo, J. Chromatogr. A, 733 (1996) 371.
145 J. Slobodnik, O. Oztezkizan, H. Lingeman and U.A. Brinkman, J. Chromatogr. A,
750 (1996) 227.
146 I.A. Dolova, A.V. Kiselev and Y.I. Yashin, Zh. Fiz. Khim., 48 (1974) 1285.
147 S.P. Nandi and P.L. Walker, Sep. Sci., 11 (1976) 441.
148 M.T. Gilbert, J.H. Knox and B. Kaur, Chromatographia,16 (1982) 138.
149 J.H. Knox, K.K. Unger and H. Mueller, J. Liq. Chromatogr., 6 (1983) 1.
150 K.K. Unger, Anal. Chem., 55 (1983) 361A.
151 E. Skuchtanova, L. Feltl and E. Smolkova-Keulemansova, J. Chromatogr., 292
(1984) 233.
152 J.H. Knox, B. Kaur and G.R. Millward, J. Chromatogr., 352 (1986) 3.
1074
153 0. Chiantore, I. Novak and D. Berek, Anal. Chem., 60 (1988) 638.
154 B.J. Bassler and R.A. Hartwick, J. Chromatogr. Sci., 27 (1989) 162.
155 D. Berek and I. Novak, Chromatographia,30 (1990) 582.
156 V.G. Berezkin and Y.S. Drugov, Gas Chromatography in Air Pollution Analysis.
Elsevier, Amsterdam, (1991.
157 Annual Book of ASTM Standards. Philadelphia, USA, 1991.
158 E. Atlas and S. Schauffler, Environ. Sci. Technol., 25 (1991) 61.
159 E. Atlas, S. Schauffler, J.T. Merril, C.J. Hahn, B. Ridley, J. Walega, J. Greenberg, L.
Heidt and P. Zimmerman, J. Geophys. Res., 97 (1992) 331.
160 K. Grob, J. Chromatogr., 321 (1985) 45.
161 B.V. Burger and Z. Munro, J. Chromatogr., 370 (1986) 449.
162 B.V. Burger and Z. Munro, J. Chromatogr., 394 (1987) 443.
163 C. Pierce, R.N. Smith, J.W. Wiley and H. Cordes, J.Am. Chem. Soc., 75 (1951) 4551.
164 W. Engewald, J. Porschmann and T. Welsh, Chromatographia,30 (1990) 537.
165 A. Di Corcia and A. Liberti, Adv. Chromatogr., 14 (1976) 305.
166 J.D. Bernal, Proc. Royal Soc. London, Ser. A, 100 (1924) 749.
167 H. Lipson and A.R. Stokes, Proc. Royal Soc. London, Se. A, 181 (1942) 93.
168 B.E. Warren, Phys. Rev., 59 (1941) 693.
169 N. Masque, R.M. Marce and F. Borrull, J. Chromatogr.A, 793 (1998) 257.
170 A. Lagana, G. D'Ascenzo, G. Fago and A. Marino, Chromatographia,46 (1997) 256.
171 R. Curini, A. Gentili, S. Marchese, et al. J. Chromatogr.A, 874 (2000) 187.
172 T. Potter, L. Martin, S. Belflower, et al. J. Agric. Food Chem., 48 (2000) 4103.
173 H. Sabnik and R. Jeannot, J. Chromatogr.A, 879 (2000) 73.
174 E. Majzik-Solymos, E. Visi, G. Karoly, et al. J. Chromatogr. Sci., 39 (2001) 325.
175 F.J. Schenk and D.J. Donoghue, J. Agric. Food Chem., 48 (2001) 6412.
176 K.N. Norman and S.H. Panton, J. Chromatogr.A, 907 (2001) 247.
177 C. Bartoni, R. Curini and G. D'Ascenzo, Environ. Sci. Technol., 34 (2000) 5059.
178 J. Tolls, M. Heller and D. Sijm, Anal. Chem., 71 (1999) 5242.
179 Q.C. Wang, W.K. Hosoya, F. Svec and J.M.J. Frechet, Anal. Chem. (1992) 64, 1232.
180 E. Baltussen, F. David, P. Sandra, H. Janssen and C. Cramers, J. High Resol.
Chromatogr., 21 (1998) 645.
181 B. Tienpont, F. David, C. Bicchi and P. Sandra, J. Micro Sep., 12 (2000) 577.
182 J.J. Sun and J.S. Fritz, J. Chromatogr., 522 (1990) 95.
183 J.J. Sun and J.S. Fritz, J. Chromatogr., 590 (1992) 197.
184 N. Masque, M. Galia, R.M. Marce and F. Borrull, J. Chromatogr.A, 771 (1997) 55.
185 P.J. Dumont and J.S. Fritz, J. Chromatogr.A, 691 (1995) 123.
186 N. Masque, M. Galia, R.M. Marce and F. Borrull, Analyst, 122, (1997) 425.
187 N. Masque, M. Galia, R.M. Marce and F. Borrull, J. Chromatogr. A, 803 (1998) 147.
188 (a) N. Masque, M. Galia, R.M. Marce and F. Borrull, Chromatographia,50 (1999) 21.,
(b) J. High Resolut. Chromatogr., 22 (1999) 547
189 A.D. Alder, F.R. Longo and J.D. Finerelli, J. Org. Chem. (1967) 32, 476.
190 D.G. Kim, M.W. Jung, I.R. Paeng, J.S. Rhee and K.J. Paeng, Microchem. J., 63 (1999)
134.
191 A. Mayes and K. Mosbach, Trends Anal. Chem., 16 (1997) 321.
192 L. Schewitz, L.I. Anderson and S. Nilsson, J. Chromatogr. A, 817 (1998) 5.
193 P.T. Vallano and V.T. Remcho, J. Chromatogr.A, 887 (2000) 125.
194 E. Reid, H.M. Hill and I.D. Wilson, Drug Development Assay Approaches, Including
Molecular Imprinting and Biomarkers. Methodol. Surv. Bioanal. Drugs 25. Royal
Society of Chemistry, Cambridge, UK, 1998.
195 S. Pleasance and R.A. Biddlecombe, DrugDevelopment Assay Approaches, Including
Molecular Imprinting and Biomarkers. Methodol. Surv. Bioanal. Drugs 25. Royal
Society of Chemistry, Cambridge, UK, (1998.
1075
196 R.A. Bartsch and M. Maeda, Molecular and Ionic Recognition with Imprinted
Polymers. ACS Symposium Series 703. ACS, Washington, D.C., (1998.
197 D. Kriz and K. Mosbach, Anal. Chim. Acta, 300 (1997) 71.
198 D. Kriz, O. Ramstrom and K. Mosbach, Anal. Chem., 69 (1997) 345A.
199 C. Berggren, S. Bayoudh, D. Sherrington and K. Ensing, J. Chromatogr. A, 889
(2000) 105.
200 B. Bjarnason, L. Chimuka and 0. Ramstrom, Anal. Chem., 91 (1999) 2152.
201 C. Crescenzi, S. Bayoudh, P.A. Cormack, T. Klein and K. Ensing, Anal. Chem., 73
(2001) 2171.
202 R. Koeber, C. Fleischer, F. Lanza, K.S. Boos, B. Sellergren and D. Bercello, Anal.
Chem. (2001) 73, 2437.
203 N. Masque, R.M. Marce, F. Borrull, P.A. Cormack and D.C. Sherington, Anal. Chem.,
72 (2000) 4122.
204 M.T. Muldoon and L.H. Stanker, Anal. Chem., 69 (1997) 803.
205 W.M. Mullett and E.P. Lai, Anal. Chem., 70 (1998) 3636.
206 A. Zander, P. Findlay, T. Renner and B. Sellergren, Anal. Chem., 70 (1998) 3304.
207 E.H. Koster, C. Crescenzi, W. den Hoedt, K. Ensing and G.J. de Jong, Anal. Chem., 73
(2001) 3140.
208 W.M. Mullett, P. Martin and J. Pawliszyn, Anal. Chem., 73 (2001) 2383.
209 C. Baggiani, G. Giraudi, C. Giovannoli, F. Trotta and A. Vanni, J. Chromatogr.A, 883
(2000) 119.
210 Y. Chen, M. Kele, S.B. Sajonz and G. Guiochon, Anal. Chem., 71 (1999) 928.
211 S. Vydyasankar, M. Ru and F.H. Arnold, J. Chromatogr.A, 775 (1997) 51.
212 K. Yoshizako, K. Hosoya, Y. Iwakoshi, K. Kimata and N. Tanaka, Anal. Chem., 70
(1998) 386.
213 B. Sellergren, Trends Anal. Chem., 18 (1999) 164.
214 E. Schoenzetter, V. Pichon, D. Thiebaut, A. Fernandez-Alba and M.C. Hennion, J.
Micro Sep., 12 (2000) 316.
215 J. Haginaka and H. Sanbe, Anal. Chem., 72 (2000) 5206.
216 D.J. Anderson, Anal. Chem., 65 (1993) 434R.
217 I.H. Hagestam and T.C. Pinkerton, Anal. Chem., 57 (1985) 1757.
218 C.P. Desilets, M.A. Rounds and F.E. Regnier, J. Chromatogr., 544 (1991) 25.
219 L.J. Glunz, J.A. Perry, B. Invergo, H. Wagner, T.J. Szczerba, J.D. Reteike and P.W.
Gunz, J. Liq. Chromatogr., 15 (1992) 1361.
220 D.J. Gisch, B.T. Hunter and B. Feibush, J. Chromatogr., 433 (1988) 264.
221 J. Haginaka and J. Wakai, Anal. Chem., 62 (1990) 997.
222 J. Haginaka and J. Wakai, Chromatographia,29 (1990) 223.
223 T.C. Pinkerton, T.D. Miller, S.E. Cook, J.A. Perry, J.D. Rateike and T.J. Szczerba,
Biochromatography, 1 (1986) 96.
224 T.C. Pinkerton, Am. Lab., 20 (1888) 70.
225 T. Nakagawa, A. Shibukawa, N. Shimono, T. Kawashima, H. Tanaka and J.
Haginaka, J. Chromatogr., 420 (1987) 297.
226 J.A. Perry, B. Invergo, H. Wagner and T.J. Szczerba, J. Liq. Chromatogr., 15 (1992)
3343.
227 J. Haginaka, H. Yasuda and Y. Kimura, Anal. Chem., 61 (1989) 2445.
228 J. Haginaka, J. Wakai, N. Yasuda, H. Yasuda and Y. Kimura, J. Chromatogr., 515
(1990) 59.
229 S.E. Cook and T.C. Pinkerton, J. Chromatogr., 368 (1986) 233.
230 J.A. Perry, J. Liq. Chromatogr., 13 (1990) 1047.
231 J. Haginaka and J. Wakai, Anal. Sci., 8 (1992) 141.
232 J. Haginaka and J. Wakai, J. Chromatogr., 596 (1992) 151.
233 J. Haginaka and J. Wakai, Anal. Sci., 8 (1992) 137.
1076
234 J. Haginaka, H. Sanbe and H. Takehira, J. Chromatogr.A, 857 (1999) 117.
235 J. Haginaka, H. Takehira, K. Hosoya and N. Tanaka, J. Chromatogr. A, 849 (1999)
331.
236 J. Haginaka and H. Sanbe, J. Chromatogr. A, 913 (2001) 141.
237 K.E. Gustavson, W.M. DeVita, A. Revis and J.M. Harkin, J. Chromatogr. A, 883
(2000) 143.
238 C. Schafer and D. Lubda, J. Chromatogr.A, 909 (2001) 73.
239 K.S. Boos and C.H. Grimm, Trends Anal. Chem., 18 (1999) 175.
240 K. Racaityte, E.S. Lutz, K.K. Unger, D. Lubda and K.S. Boos, J. Chromatogr.A, 890
(2000) 135.
241 R.A. van der Hoeven, A.J. Hofte, M. Frenay, H. Ikrth, U.R. Tjaden, J. van der Greef,
A. Rudolphi, K.S. Boos, G. Marko Varga and L.E. Edholm, J. Chromatogr.A, 762
(1997) 193.
242 C. Brinker and G. Scherer, Sol-Gel Science. Academic pPess, Boston, MA, 1990.
243 S. Acosta, P. Arnal, R.J. Corriu and D. LeClercq, in: A.K. Cheetham (Ed.), Better
Ceramics Through Chemistry VI (Mat. Res. Soc. Symp. Proc.). Materials Research
43
Society, Pittsburgh, 1994 p.
244 M. Andrianainarivelo, R.J. Corriu, D. Leclerq, P.H. Mutin and A. Vioux, J. Sol Gel
Sci. Technol., 8 (1997) 89.
245 G. Guerrero, P.H. Mutin and A. Vioux, Chem. Mater., 12 (2000) 1268.
246 J.N. Hay and H.M. Raval, Chem. Mater., 13 (2001) 3396.
247 P. Innocenzi, A. Sassi, G. Brusatin, M. Guglielmi, D. Favretto, R. Bertani, A. Venzo
and F. Bobanneau, Chem. Mater., 13 (2001) 3635.
248 J. Zuhlke, D. Knopp and R. Niessner, Fresen. J. Anal. Chem., 352 (1995) 654.
249 M. Cichna, P. Markl, D. Knopp and R. Niessner, Chem. Mater., 9 (1997) 2640.
250 B. Spitzer, M. Cichna, P. Markl, G. Sontag, D. Knopp and R. Niessner, J.
Chromatogr. A, 880 (2000) 113.
251 M. Cichna, P. Markl, D. Knopp and R. Niessner, J. Chromatogr.A, 919 (2001) 51.
252 M. Schedl, G. Wilharm, S. Achatz, A. Ketrup, R. Niessner and D. Knopp, Anal.
Chem., 73 (2001) 5669.
253 J. Seneviratne and J.A. Cox, Talanta, 52 (2000) 801.
254 J.D. Badjic and N.M. Kostic, J. Phys. Chem., 104 (2000) 11081.
255 T.L. Yost, B.C. Fagan, L.R. Allain, C.E. Barnes, S. Dai, M. Sepaniak and Z. Xue, Anal.
Chem., 72 (2000) 5516.
256 R.D. Blanchard and J.K. Hardy, Anal. Chem., 56 (1984) 1621.
257 M.J. Yang and J. Pawliszyn, Anal. Chem., 65 (1993) 2538.
258 N.C. van de Merbel, J.J. Hegeman and U.A.Th. Brinkman, J. Chromatogr.A, 634
(1993) 1.
259 S.J. Bauer and R.G. Cooks, Am. Lab., 25 (1993) 36.
260 G. Matz, G. Kibelka, J. Dahl and F. Lenneman, J. Chromatogr.A, 830 (1999) 365.
261 B. Hauser, P. Popp and A. Paschke, Int. J. Environ. Anal. Chem., 74 (1999) 107.
262 H. Startman, in: P.M. Bungay, H.K. Lonsdale and M.N. de Pinho (Eds.), Synthetic
Membranes-Science, Engineering,and Applications. Reidel, Dordecht, 1986.
263 B.M. Cordero, J.L. Pavon, C.G. Pinto, M.E. Laespada, R.C. Martinez and E.R.
Gonzalo, J. Chromatogr.A, 902 (2000) 195.
264 P. Meares, in: P.M. Bungay, H.K. Lonsdale and M.N. de Pinho (Eds.), Synthetic
Membranes-Science, Engineering, and Applications. Reidel, Dordecht, 1986.
265 Y. Shen, J.A. Jonsson and L. Mathiasson, Anal. Chem., 70 (1998) 946.
266 K.F. Pratt and J. Pawliszyn, Anal. Chem., 64 (1992) 2107.
267 M.J. Yang, S. Harms, Y.Z. Lu and J. Pawliszyn, Anal. Chem., 66 (1994) 1339.
268 Y.Z. Luo, M. Adams and J. Pawliszyn, Analyst, 122 (1997) 1461.
269 Y.Z. Luo, Anal. Chem., 72 (2000) 1058.
1077
270 A. Segal, T. Gorecki, M. Phillippe, J. Lips and J. Pawliszyn, J. Chromatogr. A, 873
(2000)
271 M.J. Yang and J. Pawliszyn, LC-GC, 14 (1996) 364.
272 T.A. Sergeyeva, H. Matuschewski, S.A. Piletsky, J. Bendig, U. Schedler and M.
Ulbricht, J. Chromatogr.A, 907 (2000) 89.
273 G.S. Luo, F. Xia and Y.Y. Dai, J. Appl. Polym. Sci., 73 (1999) 1555.
274 L.W. Schmidt, J.J. Sun, J.S. Fritz, D.F. Hagen, C.G. Markell and E.E. Wisted, J.
Chromatogr., 641 (1993) 57.
275 B. Fu, E. Bakker, V.C. Yang and M.E. Meyerhoff, Macromolecules, 28 (1995) 5834.
276 R.G. Melcher, Anal. Chim. Acta, 214 (1988) 299.
277 R.G. Melcher and S.A. Bouyoucos, Process Control Qual., 1 (1990) 63.
278 P.L. Morabito and R.G. Melcher, Process Control Qual., 3 (1992) 35.
279 R.G. Melcher and P.L. Marbito, Anal. Chem., 62 (1990) 2183.
280 B. Hauser and P. Popp, J. Chromatogr.A, 909 (2001) 3.
281 F. Li and X. Zhang, Am. Lab., 33 (2001) 36.
282 J.A. Jonsson anmd L, Mathiasson, J. Chromatogr.A, 902 (2000) 205.
283 M. Ulbricht, K. Richau and H. Kamusewitz, Colloids Surfaces A, 138 (1998) 353.
284 M.J. Steuck, US Patent 4618533, 1986.
285 J.S. Fritz and J.J. Masso, J. Chromatogr.A, 907 (2001) 89.
286 J.S. Fritz and J.J. Masso, J. Chromatogr.A, 909 (2001) 79.
287 H.Y. Wang, T. Kobayashi and N. Fuji, J. Chem. Technol. Biotechnol., 70 (1997) 355.
288 S.A. Piletsky, H. Matuschewski, U. Schedler, A. Wilpert, E.V. Piletskaya, T.A. Thiele
and M. Ulbricht, Macromolecules, 33 (2000) 3092.
289 H. Kataoka, S. Narimatsu, H.L. Lord and J. Pawliszyn, Chromatography (Japanese
Review), 20 (1999) 237.
290 A. Penalver, E. Pocurull, F. Borrull and R.M. Marce, Trends Anal. Chem., 18 (1999)
557.
291 Z. Zhang, M.J. Yang and J. Pawliszyn, Anal. Chem., 66 (1994) 844A.
292 R. Eisert and J. Pawliszyn, Anal. Chem., 69 (1997) 3140.
293 Y. Gou, R. Eisert and J. Pawliszyn, J. Chromatogr.A, 873 (2000) 137.
294 Y. Gou, C. Tragas, H. Lord and J. Pawliszyn, J. Micro Sep., 12 (2000) 125.
295 Y. Guo and J. Pawliszyn, Anal. Chem., 72 (2000) 2774.
296 H. Hartmann, J. Burhenne and M. Spiteller, Fresen. Enuiron. Bull., 7 (1998) 96.
297 H. Kataoka, H.L. Lord and J. Pawliszyn, J. Chromatogr.B, 731 (1999) 353.
298 H. Kataoka, H.L. Lord and J. Pawliszyn, J. Anal. Toxicol., 24 (2000) 257.
299 H. Kataoka, H.L. Lord and J. Pawliszyn, J. Chromatogr.A, 880 (2000) 35.
300 H. Kataoka, H.L. Lord, S. Yamamoto, S. Narimatsu and J. Pawliszyn, J. Micro Sep.,
12 (2000) 493.
301 H. Kataoka, S. Narimatsu, H. Lord and J. Pawliszyn, Anal. Chem., 71 (1999) 4237.
302 H. Kataoka and J. Pawliszyn, Chromatographia,50 (1999) 532.
303 M.E. McComb, R.D. Oleschuk, E. Giller and H.D. Gesser, Talanta, 44 (1997) 2137.
304 J. Wu, H.L. Lord, J. Pawliszyn and H. Kataoka, J. Micro. Sep., 12 (2000) 255.
305 J. Wu, Z. Mester and J. Pawliszyn, Anal. Chim. Acta, 424 (2000) 211.
306 J. Wu and J. Pawliszyn, J. Chromatogr. A, 909 (2001) 37.
307 J. Wu and J. Pawliszyn, Anal. Chem., 73 (2001) 55.
308 S. Bigham, J. Medlar, A. Kabir, C. Shende, A. Alli and A. Malik, Anal. Chem., 74
(2002) 752.
309 D. Djozan and Y. Assadi, Chromatographia,(1997) 45 (suppl), 183.
310 H.B. Wan, H. Chi, M.K. Wong and C.Y. Mok, Anal. Chim. Acta (1994) 298, 219.
311 Y. Liu, M.L. Lee, K.J. Hageman, Y. Yang and S.B. Hawthorne, Anal. Chem., 69 (1997)
5001.
312 Y. Liu, Y. Shen and M.L. Lee, Anal. Chem., 69 (1997) 190.
1078
313 J. Pawliszyn, J. Chromatogr.Sci., 38 (2000) 270.
314 V. Mani, Applications of Solid-phase Microextraction. Royal Society of Chemistry,
Cambridge, UK, 1999.
315 D.W. Potter and J. Pawliszyn, Environ. Sci. Technol., 28 (1994) 298.
316 K.D. Buchholz and J. Pawlyszyn, Anal. Chem., 66 (1994) 160.
317 W. Vaes, C. Hamwijk, E.U. Ramos, H.J. Verhaar and J.L. Hermens, Anal. Chem., 68
(1996) 4458.
318 A. Podgornik, M. Barut, J. Jancar and A. Strancar, J. Chromatogr.A, 848 (1999) 51.
319 R. Swart, S. Brouwer, J.C. Kraak and H. Poppe, J. Chromatogr.A, 732 (1996) 201.
320 R. Swart, J.C. Kraak and H. Poppe, J. Chromatogr.A., 689 (1995) 177.
321 R. Swart, J.C. Kraak and H. Poppe, Chromatographia,40 (1995) 587.
322 Z.X. Tan and V.T. Remcho, Anal. Chem., 69 (1997) 581.
323 B.J. Hall and J.S. Brodbelt, J. Chromatogr.A., 777 (1997) 275.
324 A. Penalver, E. Pocurull, F. Borrull and R.M. Marce, J. Chromatogr.A, 922 (2001)
377.
325 Y.C. Wu and S.D. Huang, Anal. Chem., 71 (1999) 310.
326 L.I. Andersson, Anal. Chem., 68 (1996) 111.
327 J. Wu, X. Yu, H. Lord and J. Pawliszyn, Analyst, 125 (2000) 391.
328 B.W. Wright, P.A. Peaden, M.L. Lee and T.J. Stark, J. Chromatogr., 248 (1982) 17.
329 L.G. Blomberg, J. Microcol. Sep., 2 (1990) 62.
330 K. Janak, M. Horka and M. Krejci, J. Microcol. Sep., 3 (1991) 115.
331 V.G. Berezkin, V.E. Shiryaeva and T.P. Popova, Zh. Analit. Khim., 47 (1992) 825.
332 D.X. Wang, S.L. Chong and A. Malik, Anal. Chem., 69 (1997) 4566.
333 S.L. Chong, D. Wang, J.D. Hayes, B.W. Wilhite and A. Malik, Anal. Chem., 69 (1997)
3889.
334 J.D. Hayes and A. Malik, J. Chromatogr.B, 695 (1997) 3.
335 A. Malik and S.L. Chong, in: J. Pawliszyn (Ed.), Applications of Solid-phase
Microextraction. Royal Society of Chemistry, Cambridge, UK, 1999, pp. 73-91.
336 J.D. Hayes and A. Malik, Anal. Chem., 73 (2001) 987.
337 J.D. Hayes and A. Malik, Anal. Chem., 72 (2000) 4090.
338 Z. Chen and T. Hobo, Anal. Chem., 73 (2001) 3348.
339 D.C. Collins, Q. Tang, N. Wu and M.L. Lee, J. Micro Sep., 12 (2000) 442.
340 M.T. Dulay, J.P. Quirino, B.D. Bennett, M. Kato and R.N. Zare, Anal. Chem., 73
(2001) 3921.
341 L. Roed, E. Lundanes and T. Greibrokk, J. Chromatogr.A, 890 (2000) 347.
342 L. Roed, E. Lundanes and T. Greibrokk, J. Micro Sep., 12 (2000) 561.
343 Q. Tang, N. Wu and M.L. Lee, J. Microcol. Sep., 11 (1999) 550.
344 Q. Tang, N. Wu and M.L. Lee, J. Micro. Sep., 12 (2000) 6.
345 Q. Tang, B. Xin and M.L. Lee, J. Chromatogr.A, 837 (1999) 35.
346 N. Ishizuka, H. Minakuchi, K. Nakanishi, N. Soga, K. Hosoya and N. Tanaka, J.
High Resol. Chromatogr., 21 (1998) 477.
347 H. Minakuchi, K. Nakanishi, N. Soga, N. Ishizuka and N. Tanaka, J. Chromatogr.A,
762 (1997) 135.
348 H. Minakuchi, K. Nakanishi, N. Soga, N. Ishizuka and N. Tanaka, J. Chromatogr.A,
797 (1998) 121.
349 N. Ishizuka, H. Minakuchi, K. Nakanishi, N. Soga, H. Nagayama, K. Hosoya and N.
Tanaka, Anal. Chem., 72 (2000) 1275.
350 N. Ishizuka, H. Minakuchi, K. Nakanishi, N. Soga and N. Tanaka, J. Chromatogr.A,
797 (1998) 133.
351 H. Minakuchi, K. Nakanishi, N. Soga, N. Ishizuka and N. Tanaka, Anal. Chem., 68
(1996) 3498.
352 K. Nakanishi, H. Minakuchi, N. Soga and N. Tanaka, J. Sol Gel Sci. Technol., 8
1079
(1997) 547.
353 Z.Y. Wang, C.H. Xiao, C.Y. Wu, C. Wu and H. Han, J. Chromatogr.A, 893 (2000) 157.
354 Z. Zeng, W.L. Qiu and H. Xing, et al. Anal. Sci., 16 (2000) 851.
355 Z. Zeng, W. Qiu and Z. Huang, Anal. Chem., 73 (2001) 2429.
356 Z. Zeng, W. Qiu, M. Yang, X. Wei, Z. Huang and F. Li, J. Chromatogr.A, 934 (2001) 51.
357 SUPELCO Catalog 2000: ChromatographyProducts for Analysis and Purification.
Aldrich-Sigma Co., USA, 2000, p. 259.
358 K. Jinno, M. Taniguchi and M. Hayashida, J. Pharm.Biomed. Anal., 17 (1998) 1081.
359 D.A. Volmer and J.P. Hui, Rapid. Commun. Mass. Spectrom., 12 (1998) 123.
360 K. Jinno, M. Taniguchi, H. Sawada and M. Hayashida, Applications of Solid-Phase
Microextraction. Royal Society of Chemistry, Cambridge, UK, 1999, pp. 527-539.
361 M. Moder and P. Popp, in: J. Pawliszyn (Ed.), Applications of Solid-Phase Micro-
extraction, Royal Society of Chemistry, Cambridge, UK, 1999, pp. 311-326.
362 Y. Hirata and J. Pawliszyn, J. Microcol. Sep., 6 (1994) 443.
363 D. Figeys, A. Ducret, J.R. Yates and R. Aebersold, Nat. Biotechnol., 14 (1996) 1579.
364 D. Figeys, Y. Zhang and R. Aebersold, Electrophoresis, 19 (1998) 2338.
365 R. Carabias-Martinez, E. Rodriguez-Gonzalo, J. Dominguez-Alvarez and J.
Hernandez-Mendez, J. Chromatogr.A, 869 (2000) 451.
366 S.M. Fields, Anal. Chem., 68 (1996) 2709.
367 G. Chirica and V.T. Remcho, Electrophoresis,20 (1999) 50.
368 S. Bigham, J. Medlar, K. Hamlet, A. Kabir and A. Malik. Manuscript in preparation.
369 G.R. Newkome, C.N. Moorefield and F. Vogtle, Dendrimersand Dendrons: Concepts,
Syntheses, Applications. Wiley-VCH, Weinheim, Germany, 2001.
370 F. Zeng and S.CI Zimmerman, Chem. Rev., 97 (1997) 1681.
371 R.S. Roy and K. Das, Chem. Commun. 2000 (7) (2000) 519.
372 D. Astruc, J.C. Blais, E, Cloutet, L. Djakovitch, S. Rigaut, J. Ruiz, V. Sartor and C.
Valerio, Top. Curr. Chem., 210 (2000) 229.
373 M.T. Reetz, Heterocycl. Chem., 35 (1998) 1065.
374 A. Hierlemann, J.K. Campbell, L.A. Baker, R.M. Crooks and A.J. Ricco, J. Am. Chem.
Soc., 120 (1998) 5323.
375 A.I. Cooper, J.D. Londono, G. Wignall, et al. Nature (London), 389 (1997) 368.
376 S.A. Kuzdzal, C.A. Monnig, G.R. Newkome and C.N. Moorefield, J. Chem. Soc.,
Chem. Commun, (1994) 2139.
377 S.A. Kuzdzal, C.A. Monnig, GR. Newkome and C.N. Moorefield, J. Am. Chem. Soc.,
119 (1997) 2255.
378 P.C. Muijselaar, C. Cramers, J.M. Jansen, E. van de Barbander-van den Berg, J.
High Resolut. Chromatogr., 18 (1995) 121.
379 M. Castagnola, L. Cassiano, A. Lupi, I. Messana, M. Patamia, R. Rabino, D.V.
Rossetti and B. Giardina, J. Chromatogr.A, 694 (1995) 463.
380 C. Stathakis, E.A. Arriaga and N.J. Dovichi, J. Chromatogr.A, 817 (1998) 233.
381 G.R. Newkome, K.S. Yoo, A. Kabir and A. Malik, TetrahedronLett., 42 (2001) 7537.
382 A. Kabir, C. Hamlet, S.Y. Yoo, G.R. Newkome and A. Malik, Manuscript in
preparation.
1080
Chapter 33
There is, nowadays, an increasing effort to monitor the environment for com-
pounds that may pose a risk to human and ecosystem health. This requires ana-
lytical methods that can be applied to different types of matrix (waters, soils,
sediments, foodstuff, biological fluids, etc.). Despite advances in the develop-
ment of highly sensitive analytical instrumentation for the final determination
of analytes, a sample pretreatment is usually necessary in order to extract, con-
centrate and isolate the analytes of interest in complex matrices. Solid-phase
extraction is, nowadays, widely accepted as a powerful technique for sample
preparation. Recent years have witnessed important developments in SPE for-
mats and sorbents with improved n-alkylsilicas, new highly cross-linked copoly-
mers and carbonaceous sorbents. These sorbents now have the capability for
extracting organic compounds over a wide range of polarities for screening pur-
poses [1]. However, analyte retention is based on hydrophobic interactions and
their selectivity is often too low for trace analysis in complex samples. Co-
extraction of other analytes and matrix interferences generally occurs, and this
can even become a major problem when analytes of interest are at the trace level
and interferences are present at high concentrations. Additional clean-up proce-
dures are then required, but then the sample pretreatment involves several
steps, with both an increase in the risk of loss and contamination and a decrease
of the reliability of the results.
Selectivity can be greatly enhanced by using materials involving antigen-
antibody interactions, thus providing selective extraction methods based on
molecular recognition. The antibodies are covalently bonded onto an appropri-
ate sorbent to form a so-called immunosorbent (IS), to be packed in a SPE car-
tridge or precolumn. Thanks to the high affinity and high selectivity of the
antigen-antibody interactions, they allow a high degree of molecular selectivity
and have been shown to be a unique tool in the sample preparation area in recent
years [2-6]. Extraction, concentration and clean up from complex liquid samples
ComprehensiveAnalytical Chemisty XXXVII
J. Pawliszyn (Ed.)
0 2002 Elsevier Science B.V. All rights reserved 1081
are achieved in the same step from large volumes when required. Their applica-
tion to extracts from solid matrixes is solvent-free and simpler than any other
clean-up procedure.
First, ISs have been described in the biological field because of the availabil-
ity of antibodies, which can be very selective for large molecules and easily
obtained. Many examples have been described for the immunoextraction of anti-
bodies, hormones, peptides, enzymes, recombinant proteins, viruses and sub-
cellular components. Obtaining selective antibodies for small size molecules was
more difficult and the development of immunochemical methods in the SPE field
targeting low molecular weight analytes is rather recent. The binding of analyte
to antibody is the result of a good spatial complementarity and is a function of
the sum of intermolecular interactions. Therefore, an antibody can also bind one
or more analytes with a structure similar to the analyte that has induced the
immune response, and this is the so-called cross-reactivity of antibodies. It is
usually considered a negative feature for immunoassays, but it was exploited in
extraction, so that ISs have been developed for single analytes, single analytes
and metabolites or a class of structurally related analytes. During the last
decade, commercial IS have been introduced for the clean up of samples for the
analysis of toxins (aflatoxins, ochratoxins, fumonisins) and veterinary drugs
(clenbuterol) in food. ISs for many contaminants or class of contaminants have
been described in the literature [4].Several class-selective immunosorbents
have been optimized for trapping groups of pesticides and priority industrial
organic pollutants [6]. Immunosorbents are robust and validation studies using
certified reference materials have demonstrated their reliability.
The objective of this chapter is to describe the potential of immunoextrac-
tion. When synthesizing an immunosorbent (IS) and developing an immuno-
extraction procedure, several parameters have to be carefully controlled and
optimized for obtaining a robust and selective SPE method. These parameters
first include the nature and the amount of the antibodies and the sorbent
selected for their immobilization. The importance of the capacity and its rela-
tionship with the analyte recovery and breakthrough volume are highlighted.
Emphasis is also given to the optimization of the extraction sequence, especially
to the desorption step.
33.2.1 Antibodies
The overall process for the preparation of an immunosorbent first involves the
production of antibodies. Because compounds of low molecular weight (<1000
Dalton) are unable to evoke an immune response, they have to be modified by
binding to a larger carrier molecule, usually a protein (e.g., bovin serum
albumin, P-lactoglobuline) to form an immunoconjugate that can be used for the
immunization. It is often necessary to introduce a functional group into the
1082
selected molecule in order to make possible this coupling. The design of the
so-called hapten is the most crucial step in the development of an immunochemi-
cal technique for small molecules because it determines the features of the
resulting antibodies. The optimum hapten for a selected target analyte has to be
a near-perfect mimic of this molecule in structure, geometry, electronic and
H-bonding capabilities and hydrophobic properties [7]. Moreover, the strategy
will be different whether a single compound is targeted or a whole class of
structurally related compounds. For the class-selective assay, the handle will be
best located at a position that leaves the common sites exposed to the immune
system. For small molecules, the use of a spacer arm in the linker is required in
order to favor recognition by the immune system. The length and the chemical
structure of the linker should be selected in order to maximize the exposure of
the target molecules and the spacer itself should not be selectively recognized.
The length of the spacer from three to six atoms has been shown to be optimal. A
good immune response seems to be obtained with 10-30 haptenic groups per 100
kDa of carrier protein [8].
New technologies for developing specific antibodies (Abs) are an active area,
using sometimes very sophisticated methods, but one has to realize that to date,
commercial Abs use mainly polyclonal or monoclonal antibodies. In both cases,
the different steps, required to produce Abs, are similar up to the immunization
of the host animal with the immunogen. After a period of a few weeks, the serum
is collected and the G-type immunoglobulins (IgG) fraction is isolated an puri-
fied. This fraction contains polyclonal antibodies (pAbs) made of an heteroge-
neous mixture of antibodies that can bind the antigen with a variety of
strengths, because they are directed against the various antigenic determinants
(epitopes) on the antigen. When a small molecule is targeted, one can expect the
number of epitopes to be low. It is difficult to measure the concentration of
specific antibodies in the whole IgG fraction. It is commonly found that it
contains about 15% of active antibodies specific of the target molecule. If
polyclonal antibodies have proven to be powerful in immunoassay techniques,
they have some limitations because polyclonal antiserum varies from one animal
to another and the supply of Abs ends when the animal dies. This can be
overcome by the use of the hybridoma technique based on the fusion of spleen
lymphocytes (which produce Abs) after isolation from immunized mice with
myeloma cells. The advantage of this technology is the capability of an unlimited
production of monoclonal Abs (mAbs) with constant affinities. However, in
contrast to pAbs production, this technique is difficult, laborious, time-
consuming and expensive.
1083
has been widely used for the coating of microtiter plates in immunoassays, is
based on hydrophobic interactions and cannot be used when the desorption of
the trapped compounds requires an organic modifier (see immunoextraction
procedure). The affinity adsorption, using protein A or G immobilized on a
stationary phase, allows a good orientation of the antibodies, but this process
also suffers from the risk of lingand leakage and steric problems due to the size of
the immobilized proteins. Covalent binding appears to be the best process of
immobilization to obtain a good stability between the sorbent and the antibod-
ies. In this case, some specific properties are required for the sorbents since they
should be porous and possess large-size pores, contain appropriate functional
groups for the bonding and have a high capacity. They should also be hydrophilic
in order to avoid additional non-specific interactions. Most of the studies report
the use of silica and agarose gels for the covalent immobilization of antibodies.
Another method-the sol-gel method-consists in encapsulating antibodies
in the pores of a hydrophilic glass matrix. This method was applied to the synthe-
sis of immunosorbents for PAHs [9,10]. Many authors have also attempted to
use hydrophilic organic polymers because they allow an oriented bonding of the
antibodies and they have higher capacities. However, hydrophilic polymers all
contain 7r-bonds that give rise to non-specific hydrophobic interactions and
decrease dramatically the selectivity of the ISs, which is their primary interest in
sample preparation methods.
<
J -< Antibodies
o 4 Analyte-antigen
011 1o: Interferences
Fig. 33.1. Off-line procedure used for the immunosample pretreatment on immunosorbents. (a)
Percolation of the sample; (b) washing to eliminate the nonretained compounds; (c) elution of
compounds retained by the immobilized antibodies.
1084
percolation, a washing step using a few mL of pure water or water containing a
small amount of organic solvent in order to eliminate compounds that do not
react with antibodies by specific interactions, and, finally, the desorption step
with a few ml of an organic solvent-water mixture. Regeneration is possible
using a PBS solution and storage at 4°C.
The capabilities of ISs for trapping the antigen and related analytes depends on
the cross-reactivity of the antibodies. The easiest way to measure the potential of
an IS for trapping a group of related analytes is to measure the recoveries (ratio
between the extracted amount and the percolated amount) that are obtained
when samples which are spiked with all the analytes are percolated. Column A of
Table 33.1 shows this potential for ISs obtained for phenylureas when bonding
polyclonal anti-isoproturon and anti-chlortoluron antibodies [11-14]. It was
difficult to predict that the anti-isoproturon IS should be class-specific because
its structure is slightly different from those of other phenylureas (see Fig. 33.2).
This was easier for anti-chlortoluron antibodies and the IS show good recoveries
for almost all the phenylureas. Figure 33.3 shows the chromatograms
corresponding to the on-line analysis of 50 ml of drinking water (Fig. 33.3a) and
surface-water (Fig. 33.3b) samples spiked with 0.5 rg/1l of a mixture of phenyl-
ureas using a precolumn packed with the anti-chlortoluron IS [12,15]. The
spiking mixture contained 13 phenylureas and only fenuron and fluometuron
TABLE 33.1
Potential for class selective immunoextraction sorbents. A: Comparison of two IS prepared using
anti-isoproturon and anti-chlortoluron polyclonal antibodies; B: comparison of two IS prepared
with monoclonal and polyclonal anti-isoproturon antibodies. Average recoveries obtained for
nine phenylureas after the percolation of water, spiked with 1 g/l of each analyte (from Refs.
[11-14])
A B
Polyclonal Polyclonal Monoclonal Polyclonal
anti-isoproturon anti-chlortoluron anti-isoproturon anti-isoproturon
(V = 50 ml) (V = 50 ml) (V = 25 ml) (V = 25 ml)
Fenuron - - 0 <10
Metoxuron 21 (3) 80 (5) 19 (5) 61 (4)
Monuron 98 (4) 78 (8) 28 (5) 101 (5)
Chlortoluron 95 (4) 95 (5) 100 (2) 102 (8)
Isoproturon 99 (3) 90 (8) 99 (1) 99 (3)
Difenoxuron 37 (4) 17 (9) 6 (6) 67 (6)
Linuron 61 (6) 85 (8) 83 (2) 96 (5)
Chlorbromuron 95 (5) 102 (5) 100 (9) 101 (5)
Diflubenzuron 101 (3) 76 (9) 59 (7) 99 (6)
1085
H3 C' \' No OH3
,%~cH H3C -- N /CH 3
H3C - C--
CH 3 Cl 0 CH
isoproturon chlortoluron
H
H3C\ N5H
CI~ N N /CH3
HC > -N <
O H 01 Cr>I-N\OMe
di-demethyl-isoproturon linuron
H
N H H3
CI - N N<CH
CH3
O CF 3 CH 3
monuron
fluometuron
H
C0 /\ N N<OH
Br N
H
OMe
Cl 0 CH3
CI O~N
\CH 3
diuron chlorbromuron
CH,O / \ N CH3
H
CiN CH 3
CI O CH3
methoxuron C CneburonH,
neburon
Fig. 33.2. Structure of commonly used phenylureas.
were not recovered. The chromatograms for these two different water matrices
are similar, since no interfering analytes are showing up in the River Seine
sample, thus pointing out the selectivity of the immunoextraction.
The values reported in Table 33.1 indicate recoveries of 80-100% for several
analytes and lower recoveries for other analytes. This is explained by the differ-
ent affinity constants that exist between the structurally related analytes and
the polyclonal antibodies bonded onto silica. The affinity order can be made clas-
sifying analytes with increasing recoveries, but even for analytes having total
recovery, the affinity can be different.
1086
(b)
ISs have been developed for single analytes and their metabolites. The list
includes drugs such as clenbuterol, chloramphenicol, diethylstilbestrol, oest-
radiol-17-3, thiobendazole, tetracyclins or aflatoxins [16-26], algal toxins [27-
29], pesticides and their transformation products [11-13,15,17,30-41]. Class-
selective IS have been described for the extraction of triazines, phenylureas,
polyaromatic hydrocarbons, BTEX (benzene, toluene, ethylbenzene and xylene)
isomers, benzidine and azodyes [9-15,36,37,42-50].
1087
comparison of individual recoveries obtained with mAbs and pAbs requires the
knowledge of the amount of bonded antibodies in each case and will be discussed
later. The fact that mAbs show almost the same cross-reactivity as pAbs can be
easily explained by the small size of the molecules. When the molecules of
interest are very small, the polyclonal antibody mixture cannot contain a large
number of specific antibodies for the different parts of the small molecule as it
can contain when the molecule is very large and has several epitopes. Then, the
probability is high for having similar properties between monoclonal and poly-
clonal antibodies.
1088
S
2
TI
Concentration
(pg/ll)
Fig. 33.4. Binding-capacity curve obtained with an anti-atrazine IS by measuring the amount of
atrazine adsorbed onto the IS as a function of the analyte concentration in the percolated
samples (from Ref. [61).
The objective of a recent study was to optimize the production of ISs using
well defined monoclonal antibodies at different bonding densities [14]. First, the
capacities obtained onto silica and Sepharose have been compared. When anti-
isoproturon mAbs were bonded to silica (glutardialdehyde-activated silica with
30 nm pore size) at a density of 1 mg per 100 mg silica, the capacity of 100 mg of
the bonded support was 1.35 ± 0.1 Ag. When the same antibodies were bonded to
Sepharose 4B at a density of 1 mg per 0.5 ml gel, the capacity of the 0.5 ml of
bonded support was 1.34 ± 0.1/ g. These similar capacities for the two ISs show
that the accessibility of the antibodies is equivalent and high. Up to the bonding
density of 1 mg of antibodies per 100 mg silica, the relation between capacity and
amount of antibodies was linear. Above this value, the capacity was lower than
expected, certainly because of steric hindrance between antibodies.
When an IS targets several structurally related analytes, the definition of the
capacity is more complex. The competition process between the compounds in a
group during the percolation can be illustrated by the measurements of capaci-
ties using a mixture of these structural analogs. An example of measurements of
capacity obtained after the percolation of a mixture of compounds on an anti-
atrazine IS is illustrated in Fig. 33.5. The results show that a higher capacity is
obtained for the analyte-antigen, i.e. atrazine, when it is alone in the sample
than when the percolated samples were spiked with atrazine and several tri-
azines. The capacities measured for related compounds depend on the affinity of
the antibodies for the compounds. In this example, a stronger affinity for the
propazine than for atrazine induces a larger apparent capacity for propazine
than for the analyte-antigen. Then, the capacity measured with the analyte-
antigen cannot be transposed to the other compounds within the group and
depends strongly on the competition process, i.e. on the number of compounds
and their respective concentrations in the sample.
1089
-a
2
I
a.
Fig. 33.5. Calibration curves of triazines in anti-atrazine IS. Percolation of 25 ml of water spiked
at concentration from 0.1 to 3 g/l for each compound (from Ref. [12]).
Quantitative analysis can be made only if recoveries are constant over the
whole linear calibration range and if the calibration curves do not depend on this
competition process. In other words, recoveries and calibration curves should be
independent of whether the analyte-antigen or any other related analytes is on
its own or in a mixture of related compounds. Therefore, the calibration range
should be defined by experiments based on the percolation of samples simulta-
neously spiked with a mixture of compounds. In this case, the more restricted
linear range will be obtained for the compounds for which the affinity of the anti-
bodies is the lowest (terbutylazine in the example of Fig. 33.5). The upper limit
of this linear range defines the concentration range of the method for all the
compounds of the group for quantitative analyses of real samples. Increasing the
amount of antibodies immobilized on the sorbent can extend this range of con-
centration. However, the amount of sorbent has to be adapted in order to keep a
bonding density that ensures a good accessibility of the analytes for the
antibodies.
One important parameter in the SPE sequence is the sample volume which can
be percolated without loss in recovery. Two parameters can affect the recovery
which are the IS capacity and/or the affinity of antibodies towards compounds.
Insufficient retention induces a low breakthrough volume and an incomplete
recovery when overcome. In general, immunoextraction is used to extract
compounds at trace levels, so, the greater the breakthrough volume and the
capacity are, the better the limit of detection is. As seen in the section above,
1090
capacities are rather low compared to that of other reversed-phase sorbents for
which overloading of the capacity is unlikely to occur. So, it is essential to verify
if experimental conditions do not induce overloading of the IS capacity or of the
breakthrough volume, because then calibrations are no longer linear.
The high affinity for the analyte-antigen is seen by the high sample volume
that can be handled without breakthrough. One litre of water containing 100 ng
of isoproturon was percolated with 100% recovery through an IS obtained with 1
g of silica and 200 Al of crude anti-isoproturon serum [51]. High breakthrough
volumes were also reported using anti-phenylurea or anti-triazine ISs when
water was spiked with the analyte-antigen alone [13,42] allowing detection
limits in water in the range 5-50 ng/l.
Comparison of recoveries provided by different ISs requires knowledge of the
nature of the bonded antibodies, which can be made by the capacity measure-
ments. Columns B of Table 33.1 gives extraction recoveries obtained after the
percolation of 25 ml of water spiked with nine phenylureas on anti-isoproturon
ISs based on monoclonal and polyclonal antibodies [11]. For these experiments,
the amount of the respective antibodies was adapted in order to have the same
capacity (4200 ng). This capacity was larger than the sum of the amount of
compounds in the sample (25 ng of each compound) in order to avoid a decrease
of the recoveries caused by the overloading of the IS. As expected, the recoveries
of the analyte-antigen, i.e. isoproturon, were close to 100% for both -ISs. They
were also similar for fenuron, chlortoluron, linuron and chlorbromuron. For the
other compounds, the recoveries obtained with monoclonal antibodies were
lower than with the polyclonal ones. Nevertheless, the cross-reactivity of the
monoclonal antibodies remains wide ensuring their application to the selective
extraction of class of related compounds with the advantage to provide more
reproducible ISs.
This comparison was carrying out using ISs characterized by the same
capacity and by percolating the same volume of sample. The modification of
these two parameters can modify the recoveries. Table 33.2 shows the recoveries
obtained after the percolation of 25 and 50 ml of water samples spiked with a
single phenylurea on 100 mg and 350 mg of an anti-isoproturon IS (bonding
density of 1 mg of monoclonal Abs for 100 mg of silica). Results obtained with
100 mg of IS show that a constant recovery about 100% is obtained for the
analyte-antigen when increasing the sample volume due to the strong affinity of
the antibodies for this compound. For other compounds, the increase of the
percolated volume causes a decrease of the recoveries due to the lower affinity of
the antibodies for these compounds. The breakthrough of these analytes can be a
problem for real application when low levels of concentration have to be
detected. High enrichment factors necessitate the percolation of high volumes of
sample with high recoveries. The increase of the amount of IS causes an increase
of the capacity of the IS. Then, the breakthrough of analytes occurs for larger
volumes of sample. This is demonstrated by the results in Table 33.2 obtained
with 350 mg of IS, which contains 3.5 times more antibodies than 100 mg of IS
1091
TABLE 33.2
Average recoveries (%) and standard deviation (n = 3), obtained after the percolation of 10, 25, or
50 ml LC grade water spiked with phenyureas through 100 or 350 mg of IS bonded with 1 mg and
3.5 mg of monoclonal anti-isoproturon antibodies, respectively
Amount of IS: 100 mg 350 mg
Sample volume: 25 ml 50 ml 25 ml 50 ml
Isoproturon (ISO) 84 (2) 94 (10) 95 (8) 100 (1)
Demethyl-ISO 72 (4) 52 (15) 100 (1) 102 (3)
Didemethy-ISO 66 (21) 24 (8) 99 (2) 100 (2)
Chlorbromuron 32 (3) 14 (7) 102 (3) 59 (8)
Diuron 24 (13) 15 (6) 95 (8) 64 (9)
Fluometuron 11 (0) 8 (2) 51 (13) 25 (5)
Chlortoluron 19 (2) 11 (5) 97 (20) 49 (3)
Linuron 5 (1) 9 (4) 51 (5) 35 (7)
Neburon 15 (0) 13 (7) 60 (9) 27 (4)
Monuron 8 (3) 2 (2) 28 (5) 17 (6)
(with the same bonding density). The increase of the capacity allows the simulta-
neous extraction of a larger number of compounds with higher recoveries.
For desorption or elution from the IS, the analyte-antibody complex formed dur-
ing the percolation of the sample has to be disrupted. As the biochemical interac-
tion energies are high, desorption will occur only by notably modifying the
experimental conditions. Different approaches are possible: displacer agents,
chaotropic agents, pH variations or water-organic modifier mixtures. Solutions
of chaotropic ions disrupt the water structure around the antibody. In fact, vari-
ous aqueous solutions that have been successfully applied for many years for the
desorption of proteins were shown to be unable to desorb small molecules. Prob-
ably, the desorption of proteins is mainly based on the structure change of the
bound protein and not only on the antibody structure change. Elution with low
pH solutions have been described (acetic acid 2%, formic acid 0.1%, citric acid 0.1
M, 0.2 M glycine-HCl), but one inconvenience is that large volumes are required
for complete desorption [30,40]. Another solution is to reconcentrate this vol-
ume using classical SPE sorbent before analysis by LC or GC.
A water-organic mixture is often required to elute small molecules from the
IS. Common organic solvents are acetonitrile [10,12,13,15,20,31,33,34,42-47],
ethanol [35,52] or methanol [11-13,15,36,37,48]. The desorption volume is then
optimized to a few ml or less, thus leading to a concentrated extract. Pure
organic solvents such as methanol are often recommended in the procedure
supplied with commercialized ISs. The use of a mixture of water with organic
1092
solvent currently used in reversed-phase LC also induces an easy on-line coup-
ling of immunoextraction with LC (see below). The choice of the elution condi-
tions are important for reusability of the IS.
The effect of the flow-rate on the binding of analytes on the IS has been studied
[25,50,55,56]. When the application of the sample on the IS is performed by a
pump as in on-line systems, this parameter can easily be controlled. An increase
in recovery from 25 to 95% was observed as the flow-rate decreased from 2.0 to
0.2 ml/min when 1 ml of 50 g/l toluene in water was percolated through a 2 cm
x 1 mm i.d. precolumn packed with a silica-based anti-toluene IS [55]. Using a
1093
1.5 cm x 4.6 mm i.d. precolumn packed with a silica-based IS containing
anti-carbendazim antibodies, a similar effect of the flow-rate was observed
between 0.2 and 10 ml/min but the recovery decrease was from 75 to 60% [47].
Other studies did not notice any loss when increasing the sample flow rate up to
2-5 ml/min using 4.6-mm i.d. precolumn[31]. The effect of the flow-rate is
certainly more important with 1-mm i.d. precolumns due to the increase of the
linear velocity when reducing the internal diameter of the column.
33.5 APPLICATIONS
The main advantage of using IS for the handling of liquid samples is that
extraction and clean-up are performed using a single cartridge which, although
the price is higher, shouldd be compared to the use of two cartridges and the
laborious and sometimes difficult use of Florisil sorbents for clean-up. Liquid
samples include serum, plasma, urine, milk and water. Urine is often diluted in
an equal volume of PBS.
Many off-line techniques are in fact used as clean-up extracts from various
matrices. After elution and further evaporation, GC or LC are applied for the
1094
final determination of analytes. The selectivity is the most important feature of
ISs and has been demonstrated for a variety of complex matrixes such as soil,
sediments, sludges, biological and plant tissues or food. The analysis of phenyl-
urea and triazine herbicides in several food samples was shown to be highly
simplified compared to conventional methods [36,37]. PAHs could be deter-
mined in waste sludges and sediments [45]. The method was validated using
certified reference sludges and sediments and the clean-up provided by the
anti-fluorene IS was shown to be better than that provided by conventional silica
clean-up [43].
1095
(a)
(b)
9 10 20 T(min)
Fig. 33.6. On-line analysis of 2.5 ml of textile effluent diluted with 47.5 ml of PBS using a
precolum packed with (a) the non-selective PRP-1 and (b) the anti-benzidine IS. The 50-ml
sample was spiked with 1 gfl of each analyte: (1) benzidine, (2) 3,3'-dichlorobenzidine, (3)
4-aminobenzene (from Ref. [44]).
too high amounts of acetonitrile for a long time, the connecting valve was
switched when the percentage of acetonitrile was 50%. At each cycle, the IS was
washed with 5 ml of a solution containing 70% methanol and then regenerated
with 6 ml of PBS. The same precolumn could be re-used for more than 50 cycles
including a half of dirty samples. Thanks to this high degree of extraction
selectivity, the sample volume could be reduced. When combining the selectivity
of the IS with the high sensitivity achieved by LC-atmospheric pressure chemical
ionization (APCI)-MS in positive operation mode, detection and confirmation
could be obtained at the low ng/l from preconcentration volume as low as 20 ml
[33,34]. The method was validated through an interlaboratory study with
Aquacheck certified samples.
The on-line coupling of IS to LC is particularly appropriate for the trace
analysis of polar and/or volatile analytes because no evaporation of samples
occur. When sample matrices are very complex such as some industrial effluents,
the detection of polar analytes is hindered by many polar interferences. Owing to
their carcinogenic nature, benzidine and dichlorobenzidine are included in the
US EPA lists and in the European restricted list of "very toxic substances for the
environment". These analytes are still produced in large amounts as
intermediary compounds in the manufacture of dyes and pigments. An
anti-benzidine IS was designed which was able to trap aminobenzene and
related azodyes [26]. Figure 33.6 shows the chromatogram obtained for the
on-line analyses of 2.5 ml of a textile effluent diluted with 47.5 ml of PBS
solution, using a polylmeric precolumn (a) or using the IS precolumn (b). The
1096
selectivity provided by the IS is very obvious for such an industrial waste which
contains many polar interferences co-extracted by the polymer and co-eluted at
the beginning of the chromatogram. Another example deals with the 16 priority
PAHs of the US EPA list which contains some 2-3 rings PAHs that are both
volatile and hydrophobic [46]. Due to their hydrophobicity, the addition of an
organic solvent in the sample before percolation was necessary to avoid
adsorption on containers or connection tubes. Using fluorescence detection, it
was possible to quantify PAHs at 20 ng/l from 20 ml of surface water (regulatory
level for some PAHs in Europe for surface water taken for drinking water
plants).
In order to recover a whole class of compounds, two antibodies have some-
times been mixed in the cartridge bed. Mixed-bed sorbents using antibodies
against chlortoluron and isoproturon have been obtained for recovering the
group of phenylureas [15]. A mixed-bed IS was also tailored for trapping
ampicillin and cloxacillin [62].
Immunoextraction was also coupled to GC although the interfacing between
the immunoextraction and the GC part is not as straightforward as it is with LC.
After trace-enrichment of analytes on a 10 x 3 mm immunoprecolumn packed
with monoclonal anti-atrazine antibodies immobilized onto cellulose, they were
desorbed and re-concentrated on a reversed-phase precolumn packed with a
copolymer by means of an acidic buffer [42]. After clean-up and drying with
nitrogen, desorption and transfer to the GC was done with ethylacetate via an
on-column interface. The selectivity of the system was such that non-selective
flame ionization detection (FID) could be used to detect several triazines in river,
waste water and orange juice. The detection limits for 10 ml samples were 15-25
ng/l for FID and about 1.5 ng/l for nitrogen-phosphorus detection.
The potential for rapid sample handling was demonstrated with the direct
on-line coupling of IS with microLC-UV DAD. The direct on-line set-up was
performed using a 10 x 1 mm precolumn prepacked with a silica-based anti-
phenylureas IS and a 250 x 1 mm analytical column. Quantification limits of 10
ng/l have been obtained from sample volumes of 5 ml with UV-DAD detection
[61]. Figure 33.7 illustrates the high selectivity of the immunoextraction as
compared with the use of the same precolumn packed with C18 silica. Concen-
trations were around 50 ng/l for both chlortoluron and linuron, 150 ng/l for
isoproturon and 200 ng/l for diuron. Since there is no co-extraction of humic
materials, the analysis can be more rapid because usually the first eluted peak is
delayed in order to be after the humic acids peaks. The analysis of diuron used in
antifouling paints could be done in sea water in less than 10 min. This was
possible because the 4 0-um silica particles used for preparing the IS which
allowed a rapid sample flow-rate and because the sample volume is low.
1097
mAbs
30
20
10
0 10 mi 20 30
33.6 CONCLUSION
1098
trapping occurs on the basis of a structure recognition. Moreover, a clear
baseline enhances the overall sensitivity of the analysis and allows the handling
of a lower sample in comparison with non-selective extraction procedures. The
on-line combination of this selective preconcentration with a separation using a
very short micro-analytical column and MS detection is a good solution for rapid
analysis.
REFERENCES
1 M.-C. Hennion, J. Chromatogr. A, 856 (1999) 3.
2 E. Hogendoorn and P. Van Zoonen, J. Chromatogr.A, 892 (2000) 435.
3 N. Delaunay-Bertoncini, V. Pichon and M.-C. Hennion LC-GC Europe, 14/3 (2001)
162.
4 N. Delaunay, V. Pichon and M. C. Hennion, J. Chromatogr.B, 745 (2000) 15.
5 D.S. Hage and N.A. Nelson, Anal. Chem., 71 (2001) 199A.
6 V. Pichon, M. Bouzige, C. Miege and M.-C. Hennion, Trends Anal. Chem., 18 (3)
(1999) 219.
7 M.P. Marco, S.G. Gee and B.D. Hammock, Trends Anal. Chem., 14 (7) (1995) 415.
8 M. Worterg, K. Camman, K. Strupat and F. Hillenkamp, Fresenius' J. Anal. Chem.,
348 (1994) 240.
9 M. Cichna, P. Markl, D. D. Knopp, R. Niessner, Chem. Mater., 9 (1997) 2640.
10 M. Cichna, D.D. Knopp and R. Niessner, Anal. Chim. Acta, 339 (1997) 241.
11 V. Pichon, L. Chen, M.-C. Hennion, R. Daniel, A. Martel, F. Le Goffic, J. Abian and D.
Barcelo, Anal. Chem., 67 (1995) 2451.
12 V. Pichon, L. Chen, N. Durand, F. Le Goffic and M.-C. Hennion, J. Chromatogr. A,
725 (1996) 107.
13 V. Pichon, H. Rogniaux, N. Fisher-Durand, S. Ben Rejeb, F. Le Goffic and M.-C.
Hennion, Chromatographia,45 (1997) 289.
14 N. Delaunay-Bertoncini, V. Pichon and M.-C. Hennion, Chromatographia,53 (2001)
S-224.
15 V. Pichon, M. Bouzige and M.-C. Hennion, Anal. Chim. Acta, 376 (1998) 21.
16 J.F. Lawrence and C. Menard, J. Chromatogr.B, 696 (1997) 291.
17 J. Cai and J. Henion, J. Chromatogr.B, 691 (1997) 357.
18 H. Hooijerink, R. Schilt, E.O. van Bennekom and F.A. Huf, J. Chromatogr. B, 660
(1994) 303.
19 Th. Gude, A. Preiss and K. Rubach, J. Chromatogr. B, 673 (1995) 197.
20 R. Bagnati, M.G. Castelli, L. Airoldi, M. Paleologo-Oriundi, A. Ubaldi and R. Fanelli,
J. Chromatogr., 527 (1990) 267.
21 A.L. Savage, S.H. Sarijo and J. Baird, Anal. Chim. Acta, 375 (1998) 1.
22 E. Horne, T. Coyle, M. O'Keefe and D.L. Brandon, Analyst, 124 (1999) 87.
23 A. Kussak, B. Andersson and K. Andersson, J. Chromatogr.A, 708 (1995) 55.
24 A. Kussak, B. Andersson and K. Andersson, J. Chromatogr. B, 672 (1995) 253.
25 A. Farjam, G.J. De Jong, R.W. Frei, U.A.Th. Brinkman, W. Haasnoot, A.R.M.
Hamers, R. Schilt and F.A. Huf, J. Chromatogr., 452 (1988) 419.
26 W. Haasnoot, R. Schilt, A.R.M. Hamer, F.A. Huf, Ai. Farjam, R.W. Frei and U.A.Th.
Brinkman, J. Chromatogr., 489 (1989) 157.
27 C. Rivasseau and M.-C. Hennion, Anal. Chim. Acta, 394 (1999) 243.
28 L. Ten-Hage, N. Delaunay, V. Pichon, A. Coute, S. Puiseux-Dao and J. Turquet,
Toxicon, 38 (2000) 1043
29 N. Delaunay, V. Pichon, J.-P. Le Caer and M.-C. Hennion, Anal. Chim. Acta, 407
(2000) 173.
1099
30 G.S. Rule, V. Mordehal and J. Henion, Anal. Chem., 66 (1994) 230.
31 V. Pichon, L. Chen and M.-C. Hennion, Anal. Chim. Acta, 311 (1995) 429.
32 A. Martin-Esteban, P. Fernandez, D. Stevenson and C. Camara, Analyst, 122 (1997)
1113.
33 I. Ferrer, M.-C. Hennion and D. Barcelo, Anal. Chem., 69 (1997) 4508.
34 I. Ferrer, V. Pichon, M.-C. Hennion and D. Barcelo, J. Chromatogr.A, 777 (1997) 91.
35 S.J. Shahtaheri, M.F. Katmeh, P. Kwasowski and D. Stevenson, J. Chromatogr. A,
697 (1995) 131.
36 J.F. Lawrence, C. Menard, M.-C. Hennion, V. Pichon, F. Le Goffic and N. Durand, J.
Chromatogr.A, 732 (1996) 277.
37 J.F. Lawrence, C. Menard, M.-C. Hennion, V. Pichon, F. Le Goffic and N. Durand, J.
Chromatogr.A, 752 (1996) 147.
38 V. Pichon, E. Aulard-Macler, H. Oubihi, P. Sassiat, M.-C. Hennion and M. Caude,
Chromatographia,46, 9/10 (1997) 529.
39 J. Dalluge, T. Hankemeier, R.J.J. Vreuls and U.A.Th. Brinkman, J. Chromatogr. A,
830 (1998) 267.
40 J. Cai and J.D. Henion, Anal. Chem., 68 (1996) 72.
41 D.H. Thomas, M. Beck-Westermeyer and D.S. Hage, Anal. Chem., 66 (1994) 3823.
42 M. Bouzige, V. Pichon and M.-C. Hennion, J. Chromatogr.A, 823 (1998) 197.
43 S. Perez, I. Ferrer, M.-C. Hennion and D. Barcelo, Anal. Chem., 70 (1998) 4996.
44 M. Bouzige, P. Legeay, V. Pichon and M.-C. Hennion, J. Chromatogr.A, 846 (1999)
317.
45 C. Miege, M. Bouzige, S. Nicol, J. Dugay, V. Pichon and M.-C. Hennion, J. Chroma-
togr. A, 859 (1999) 29.
46 M. Bouzige, V. Pichon and M.-C. Hennion, Environ. Sci. Technol., 33 (1999) 1916.
47 S. Perez and D. Barcelo, Analyst, 125 (2000) 1273.
48 R.K. Betsen-Farmen, I.V. Botnen, H. Noto, J. Jacob and S. Ovrebo, Int. Arch. Occup.
Environ. Health, 72 (1999) 161.
49 S.D. Thomas and Q.X. Li, Environ. Sci. Technol., 34 (2000) 2649.
50 S. Ouyang, Y. Xu and Y.H. Chen, Anal. Chem., 70 (1998) 931.
51 S.J. Shahtaheri, P. Kwasowski and D. Stevenson, Chromatographia,47, (1998) 453.
52 A. Martin-Esteban, P. Kwasowski and D. Stevenson, Chromatographia,45 (1997)
364.
53 T.M. Phillips and J.M. Krum, J. Chromatogr.B, 715 (1998) 55.
54 N. Delaunay-Bertoncini, V. Pichon and M.-C. Hennion, submitted.
55 D.H. Thomas, V. Lopez-Avila, L.D. Betowski and J. Van Emon, J. Chromatogr.A,
724 (1996) 207.
56 J.G. Rollag, M. Beck-Westermeyer and D.S. Hage, Anal. Chem., 68 (1996) 3631.
57 A. Farjam, A.E. Brugman, A. Soldaat, P. Timmerman, H. Lingeman, G.J. De Jong,
R.W. Frei and U.A.Th. Brinkman, Chromatographia,31 (1991) 469.
58 A. Farjam, R. de Vries, H. Lingeman, R.W. Frei and U.A.Th. Brinkman, Int. J.
Environ. Anal. Chem., 44 (1991) 175.
59 A. Farjam, C. van der Merbel, H. Lingeman, R.W. Frei and U.A.Th. Brinkman, Int. J.
Enuiron.Anal. Chem., 45 (1991) 73.
60 A. Farjam, N.C. van der Merbel, A.A. Nieman, H. Lingeman and U.A.Th. Brinkman,
J. Chromatogr., 589 (1992) 141.
61 E. Schoenzetter, V. Pichon, D. Thi6baut, A. Fernandez-Alba and M.-C. Hennion, J.
Microcolumn Sep., 12/5 (2000) 316.
62 R. Dietrich, E. Usleber and E. Martlbauer, Analyst, 123 (1998) 2749.
63 J. Henion, E. Brewer and G. Rule, Anal. Chem., 70 (1998) 650A.
1100
Abbreviations
AA acrylic acid
ABS aqueous biphasic system
ACN acetonitrile
AFM atomic force microscopy
AMANDA ammonia measurement by annular denuder sampling with
on-line analysis
AMPS 2-acrylamido-2-methyl-1- propane sulfonic acid
AOT di-2-ethylhexyl sodium sulfosuccinate
APHA American Public Health Association
ASE accelerated solvent extraction
BC black carbon
BOSS Brigham Young University Organic Sampling System
BSA bovine serum albumin
BTEX benzene, toluene, ethylbenzene, xylenes
CBA ethylcarboxylic acid
CBG Coomassie Blue G
CE capillary electrophoresis
CEC capillary electrochromatography
CEPPWD circular end parallel plate wet denuder
CH cyclohexyl
CHD cyclohexanedione
CIF charcoal impregnated filter
CL chemiluminescence
CME capillary microextraction
CMS carbon molecular sieves
CPE cloud-point extraction
CVD chemical vapor deposition
CW Carbowax
DAD diode array detector
DBT dibenzothiophene
DCB 1,4-dichlorobenzene
DCCC droplet countercurrent chromatography
DI direct immersion
DM dialysis membrane
DME dynamic membrane extraction
DMG dimethylglyoxime
1101
DNPH 2,4-dinitrophenylhydrazine
DS diffusion scrubber
DS- dodecylsulfate anion
DTA diethylenetriamine
DVB divinylbenzene
EC elemental carbon
ECD electron capture detector
ECN Energieonderzoek Centrum Nederland (Netherlands Energy
Research Institute) (Ch. 5)
ED electrochemical detector
EDMA ethylene glycol dimethacrylate
EDTA ethylenediaminetetraacetic acid
ELISA enzyme-linked immunosorbent assay
EPA Environmental Protection Agency (of the U.S.A.)
ePTFE expanded poly(tetrafluoroethylene)
ESR electron spin resonance
FIA flow injection analysis
FID flame ionization detector
FL fluorescence detector
FMA fluorescein mercuric acetate
FP filter pack
FPD flame photometric detector
FSA Fluorosililic acid
FTIR Fourier transform infra red
FV flash volatilization
GAMS gas and aerosol monitoring system
GC gas chromatography; also graphitic carbon (Chap. 6)
GC/MS gas chromatography/mass spectrometry
GC-FID gas chromatography-flame ionization detection
GCB graphitized carbon black
GDMA glycerol dimethacrylate
GFF glycine-L-phenylalanine-L-phenylalanine
GMMA glycerol monomethacrylate
HF: Hydrofluoric acid
HMDS hexamethyldisilazane
HMHP hydroxymethyl hydroperoxide
HMX 1,3,5- 7-tetrahydro -1,3,5, 7-tetranitrotetrazine
HPLC high performance liquid chromatography
HPLC/UV high performance liquid chromatography/ultraviolet detector
HRP horseradish peroxidase
HS headspace
H.S. hydrophobic surroundings
HV high voltage
i.d. inside diameter
1102
IAE immunoaffinity extraction
IC ion chromatography
IC-CD-UV-MS ion chromatography-suppressed conductivity-ultraviolet
detection-mass spectrometry
ICP-MS induction coupled plasma mass spectrometer
IMPROVE Interagency Monitoring of Protected Visual Environments
IR infra red
ISRP internal surface reversed-phase
LC liquid chromatography
LC/MS liquid chromatography/mass spectrometry
LCST lower critical solution temperature
LCW liquid core waveguide
LED light emitting diode
LLE liquid-liquid extraction
LOD limit of detection
LPM liters per minute
LPME liquid-phase microextraction
LSD lysergic acid diethylamide
LSE liquidsolid extraction
MAA methacrylic acid
MAE microwave assisted extraction
MALDI-MS matrix-assisted laser desorption ionization mass spectrometry
4-MAP 4-methylacetophenone
MASE microwave-assisted solvent extraction
MB + methylene blue cation
MB.DS methylene blue-dodecylsulfate ion pair (Ch
MBAA N,N-methylenebis(acrylamide)
MBTH 3-methyl-2-benzothiazoline hydrazone
MCME monolithic capillary microextraction
MDA methylenedioxy amphetamine
MDEA methylenedioxy ethylamphetamine
MDMA methylenedioxy methamphetamine
ME micellar extraction
MECC micellar electrokinetic capillary chromatography
MEKC micellar electrokinetic chromatography
MESI membrane extraction with a sorbent interface
MF microfiltration
MFP mixed-functional phase
MFV methanol-fueled vehicle
MHP methyl hydroperoxide
MI molecular imprinting
MIMS membrane introduction mass spectrometry
MIP molecularly imprinted polymer
MMAD mass median aerodynamic diameter
1103
MMLLE microporous membrane liquid-liquid extraction
MS mass spectrometry or mass spectrometer
MTBE methyl tert-butyl ether
MTMS methyltrimethoxysilane
MW molecular weight
NAAQS National Ambient Air Quality Standard
NCAR National Center for Atmospheric Research
NMDS nation membrane diffusion scrubber
NMR nuclear magnetic resonance
NPD nitrogen-phosphorus detector
o.d. outside diameter
OC organic carbon
ODS octadecylsilane
OTCME open tubular capillary microextraction
PA polyacrylate
PAH polynuclear aromatic hydrocarbons; also polycyclic aromatic
hydrocarbons
PAR 4-(2-pyridylazo resorcinol)
PBA phenylboronic acid
PCBs polychlorobiphenyls
PCE perchloroethylne
PCS particle collection system
PDMS poly(dimethylsiloxane)
PEEK polyether ether ketone
PEG poly(ethylene glycol)
PES polyethersulfone
PFA Teflon perfluoroalkoxy Teflon
PFE pressurized fluid extraction
PGC porous graphitic carbon
PHCs petroleum hydrocarbons
PIPAAm poly(N-isopropylacrylamide)
PMDS porous membrane diffusion scrubber
PME polymeric membrane extraction
PMHS poly(methylhydrosiloxane)
PMMA poly(methyl methacrylate)
PMT photomultiplier tube
PNIPAAm poly(N-isopropylacrylamide)
PO-CL peroxyoxalate chemiluminescence
ppb parts per billion
ppbv parts per billion by volume
PPD parallel plate denuder
ppm parts per million
ppmv parts per million by volume
PPO poly(2,6-dimethyl-1,4-phenylene) oxide
1104
ppq parts per quadrillion
ppt parts per trillion
pptv parts per trillion by volume
PPWD parallel plate wet denuder
PPY polypyrrole
(PPPY) poly(N-phenylpyrrole)
PS poly(styrene)
PS-DVB Poly(styrene-divinylbenzene)
psi pounds per square inch
PTFE poly(tetrafluoroethylene)
PVC polyvinyl chloride
PVDF polyvinylidene fluoride
RAM restricted access material
RAMS real-time ambient mass sampler
RDX 1,3,5-hexahydro-1,3,5-trinitrotriazine
RH relative humidity
RO reverse osmosis
ROMP ring opening metathesis polymerization
rpm revolutions per minute
RSD relative standard deviation
RWAD rotating wet annular denuder
SIN signal to noise ratio
SAX trimethyl aminopropyl
SC-CO2 supercritical carbon dioxide
SCE saturated Calomel electrode
SCX benzenesulfonic acid
SEM scanning electron micrograph also microscopy
SFC supercritical fluid chromatography
SFE supercritical fluid extraction
SID surface ionization detector
SJAC steam jet aerosol collector
SLM supported liquid membrane extraction (Chap. 23)
SLME supported liquid membrane extraction (Chap. 32)
SLPM standard liters per minute
SMM surface modifying macromolecules
SPE solid phase extraction
SPME solid phase microextraction
SPS semipermeable surface
SS stainless steel
STEM scanning transmission electron microscopy
TC total carbon
TCE trichloroethylene
TCP 2,4-trichlorophenol
TDLAS tunable diode laser absorption spectrometer
1105
TDM therapeutic drug monitoring
Teflon AF Teflon amorphous fluoropolymer
TEOM tapered element oscillating microbalance
TEOS tetraethoxysilane
THF tetrahydrofuran
TMB tetramethylbenzidine
TMOS tetramethoxysilane
TNT 2,4,6-trinitrotoluene
TOFMS time-of-flight mass spectrometry
TPH total petroleum hydrocarbon
TPR templated resin
TPS thermoseparating polymer system
TRP temperature-responsive polymer
TSP total suspended particulate matter
UF ultrafiltration
UV ultraviolet
UPW ultrapure water
V-65 2,2'-azobis(2,4-dimethylvaleronitrile)
VC vinyl chloride
VOCs volatile organic compounds
WAD wet annular denuder
WEDD wet effluent diffusion denuders
XPS X-ray photoelectron spectroscopy
1106
Index
1107
air pollution, 162 4-aminosulfonyl-7-fluoro-2,1,3-benzoxad
air sampling pumps, personal, 681 iazole, 625
air segmented liquid, 123 ammonia, 98, 103, 115, 120, 136, 140,
airborne particles, 163 141, 148, 153, 1011
alarm pheromones, 672 ammonium, 196, 198
alcohol oxidase, 124 ammonium acetate, 106
alcohols, 635, 650, 656, 753 ammonium iodide, 644
aldehydes, 9, 26 ammonium molybdate, 166
algal pheromones, 691 ammonium nitrate, 97
Aliquat-336, 508 ammonium sulfate, 97, 164
alkali metal ion adducts, 170 amperometry, 621
n-alkanes, 9 amphetamine, 650, 927
alkyl bromides, 610 analyte binding, 241, 242
alkyl chloroformates, 648 analyte collection, 592
alkyl hydroperoxide, 112 analyte distribution in natural systems,
alkylating anticancer drugs, 641 471
alkylation, 647 analyte enrichment, 757
alkylboronylation, 652 analyte hydrophobicity, 514
alkylchloroformates, 610 analyte trapping, 512
1-alkylthiol 2-alkyl substituted analyte-matrix interactions, 560
isoindoles (AAI), 612 analytical distillation procedures, 702
alkyltin, 941 analytical instrumentation,
allergens, 1, 24 interferences, 422
alpha track detectors, 17 analytical process steps, 253
alumina, 347, 738, 763 anandamide, 635
alveolar region, 205 anion exchangers, 739
Alzheimer's disease, 614 ANSYL hydrazones., 628
ambient aerosol, 205 antibodies, 741
- analysis of, 183 - affinity of, 1090
ambient air measurement, 200 - cross-reactivity, 741
ambient outdoor air, 204 - immobilization, 1083
American Public Health Association - mimics, 409
methods, 899 antidepressants, 928
American Society for Testing and antifouling compounds, 747
Materials (ASTM), 924 antigen-antibody interaction, 741
amidation, 610 antioxidants, 622
amines, 115, 120, 284, 612, 630, 638, 643 ants, 689
amino acids, 612, 614, 616, 622, 625, aphrodisiac pheromones, 672
638, 642, 653 aquatic colloids, 228
aminoacridines, 631 aqueous and organic liquids, 532
2-aminobenzoic acids, 630 aqueous environmental matrices,
2-aminonaphthalene trisulphone, 655 definition, 219
aminoglycan, 629 aqueous matrix, 468
aminoglycoside, 639 aqueous phase, 541
aminopropyl silica, 738 arable layers, 74
1108
arachadonyl ethanolamide, 635 azodyes, 741
arginine, 648
Aroclor, 903, 913 back extraction, 299
aromas, 699 bacteria, 24
aromatic dialdehydes, 612 balanced pressure sampling, 730
aromatic hydrocarbons, 752, 754 basic compounds, 724
aromatic sulfonates, 747 batch equilibrium, 733
aromatic sulfonic acids, 740 batch techniques, 267
aromatics, 294 bathometers, 36
arsenic speciation, 232 Beer's law behavior, 131
arson investigation, 921 benzaldehyde, 294
asbestos, 17 benzene, 23, 128, 293
ascending/descending drop, 314 benzo(a)pyrene, 13
ASHRAE, 22 benzodiazepines, 928
asthma, 161 benzoic acid, 294, 630
ASTM methods, 13 benzophenones, 928
asymmetric membranes, 105, 126 Berthelot reaction, 129
atmospheric aerosols, 163 BET method, 236
atmospheric chemistry, 162 binding energies, 736
atomic absorption, 16 bioaccumulation, 761
atomic force microscopy, 994 bioaerosols, 1, 24, 202, 203, 205
attractants, 670 bioanalyses, 244, 845
auto exhaust, 125 bioavailability, 241
autoanalyzer, 307 biodesulfurization of dibenzothiophene,
automated evaporative concentration 913
system, 728 biological activity of insect pheromones,
automated headspace, 290 669
automated in-tube solid-phase biological agents, 24
microextraction, 428 biological fluids, 284, 286, 293, 659
automated liquid-liquid extraction, 727 - volatiles in, 292
automated monitoring, 727 biological matrices, 293, 532, 548
automated processes, future trends, 865 biomedical analyses, 247, 524
automated sequential trace enrichment biotransformation, 647
of dialysates, 861 - pathways, 913
automated solid-phase extraction, 742 bisphenols, 635
automated SPE-GC coupling, 381 bisphosphonates, 618
automatic sampling systems, 52 bitumen, 913
automation, 377, 464, 729, 760 black carbon, 172
- applications, 844 blood, 926, 929-931
- for bioanalysis, 845 - alcohol, 279, 287, 289, 930
- for clinical chemistry, 861 - bone and tissue, 548
- strategies, 839 - gases, 531, 548
automobile exhaust, 162 Bodman's sampler, 38
autosamplers, 454 bonding density, 1088, 1090
- for GC, 430 boreholes, 42
1109
bound residues, 907 carbon-based extraction materials, 1032
boundary layer, 268, 270 carbon-based ion exchange fibers, 1017
bovine serum albumin, 204 carbon-based sorbents, 739
brain dialysate, 617 carbon content percentage, 737
breakthrough curves, 736 carbon dioxide (C02), 22
breakthrough volumes, 359, 724, 734, carbon measurement, 174
736, 743, 745, 1090 carbon molecular sieve, 1034
- determination, 360 carbon monoxide (CO), 22
- methods for estimating, 361 carbonaceous particles, 172
- models for prediction, 361 carbonization in a soot-forming flame,
breath, 792 170
bromide, 196 carbonyls, 610, 632, 639, 643
bromoacetonitrile, 652 - as analytes and reagents, 650
bronchoconstriction, 205 carboxene, 751
brown algae, 688 carboxylic acids, 420, 610, 634, 635, 640,
brown-field sites, 897 643, 653
bubble gas, 84 carboxylic groups, 739
bubbler/impinger, 98 cardiopulmonary and lung cancer
buffers, 798 related mortality, 161
bulk liquid membranes, 505 carrier-mediated transport, 507
bulk materials, 796 carryover, 300, 743
busulfan, 641 cartridges, 343, 346, 360, 362, 732, 733,
butanal, 294 738, 741, 746, 851
butyltin, 660 cation exchange capacity, 230
butyric acid, 121, 147 cation exchangers, 739, 1014
Celgard, 105, 123
C18 silica, 739 CEN methods, 733
cadaverine, 621 centrifugation, 163
caffeine, 930 ceramic honeycomb denuder, 164
Calberla stain, 26 certified reference material, 459, 940
calcium, 198 charcoal filter, 236, 680
calibration, 497, 751 charged particle, 176
- selection of method, 458 chemical activation, 1008
cannabinoids, 635, 932 chemical cytometry, 617, 636
capillary, 130 chemically bonded reversed phase silica,
capillary electrophoresis, 308, 967 737
capillary gas chromatography-mass chemically bonded sorbents, 349
spectrometry, 625 chemically modified resins, 739
capillary isoelectric focusing, 978 chemically nonspecific particle
capillary isotachophoresis is, 973 measurements, 161
capillary scale analyzers, 114 chemiluminescence, 113, 166, 635
capillary zone electrophoresis (CZE), chemoluminescence, 19-21
609, 611,613, 617, 621, 631, 638 chiral separations, 658
capsules, 797 chloracetaldehyde, 632
carbamate pesticides, 747 chloride, 198
1110
chlorinated solvents, 895 competition process, 1089, 1090
chloroacetaldehyde, 632 composite membranes, 990
chlorodifluoroacetic anhydride, 649 composite sample, 75
chloromethylanthracene, 634 compost, 219
chlorophenoxy acetic acid herbicides, 647 computer-aided strategies for SPE
cholesterol, 124 method development, 375
Chromasil, 119 concentration enrichment, 516
Chromosorb, 13 condensation, 162
chromotropic acid method, 11 condensation nuclei counters, 193
chrysene, 13 conditioning step, 373
chrysolite, 17 conductivity of an aqueous solution, 225
cigarette smoke, 102 constant heating time, 290
circular end parallel plate denuder, 101, contact SPME, 689
146 contamination, 97
citric acid, 147 continuous collections, 176, 680
class-selective trapping, 1087 continuous LLE, 726
clay, particle size distribution of, 233 continuous phase, 298
cleanup, 297, 723, 738, 742, 756, 758, controlled-release binding agents, 798
762, 763, 817, 878, 968 controlled toxicological studies, 162
clinical chemistry, automation for, 861 convection, 305
closed loop stripping, 691 convective-diffusive mass transfer, 298,
cloud chamber, 193 302
coagulation, 162 convolution model of extraction, 264
coated ion exchange fibers, 1014 cooking bags, 681
coatings, 797 cooling, 494
cocaine odor, 932 Coomassie Blue G, 204
co-current chromatography, 314 copper acetate, 126
coefficient of haze, 174 copper diethyldithiocarbamate, 126
co-elution, 762 corers, 81
co-extractives, 874 corona discharge, 178
coiled configuration, 100 coupled-column techniques, 378
coke oven gas, 118 Cour traps, 26
cold finger, 416 creams/ointments/gels, 795
cold trap, 285 p-cresol, 111
collection efficiencies, 105, 121 critical limits, 724
collection efficiency, 106, 129, 178 cross-reactivity, 1087
colony forming units, 27 cryogenic condensation, 730
colorimetric analysis, 20 cryogenic focusing, 685, 731
colorimetry, 19, 898 cryogenic oven, 285
colorings, 222, 798 cryogenic sampler, 82
column switching methods, 763 cryogenic trapping, 692, 730, 953
combined hot water extraction-SPME, CTC Analytics, 430
578 cumene hydroperoxide, 135
T M
CombiPal autosampler, 430 cuticular pheromones, 689
commercial devices, 429 cyanide, 118, 932
1111
cyanopropyl, 738 1,4-dichlorobenzene), 137
1-cyano-2-substituted-benz[f]isoindoles, dichlorofluoroacetyl, 649
615 dienophiles, 655
cyclohexanedione, 106 diethyl sulphate, 647
cyclohexyl, 738 diethyldithiocarbamate, 641
cyclone, 183 diethylhexyl phosphoric acid, 509
cyclophosphamide, 632 diffusion, 181, 759
cystine, 623 diffusion-based calibration, 269
Czapla-1 sampler, 41 diffusion coefficient, 98, 271, 412
diffusion constant, 534
dansyl chloride, 621 diffusion denuders, 98, 99, 102, 163,
dansyl derivatives, 621 164, 169, 205
dansyl hydrazine, 627, 632 diffusion scrubbers, 104
dansyl hydrazone, 632 diffusiophoresis, 144
decreased pulmonary function, 161 diffusive mass transfer, 302
defensive compounds, 677 dihydrorhodamine, 616
defensive secretions, 672 dimethamphetamine, 927
deisopropylatrazine, 737, 738 1,3-dimethyl-2-imidazolidinone, 293
demethylation, 618 dimethyl mercury, 659
dendrimers in solution-mediated dimethyl sulphate, 647
extraction, 1026 dimethylamine, 116
density, 221 DIN methods, 733
- of solids, 229 2,4-dinitrofluorobenzene, 653
denuder liquid considerations for IC 2,4-dinitrophenylhydrazine, 124, 639
coupling, 151 diol groups, 740
denuders, 3, 16 dioxane, 283
deproteinization, 800 direct extraction procedures, 712
derivatization, 10, 419, 686, 727, 752, 954 direct immersion SPME, 249, 810
- combined with extraction, 274 direct insertion membrane probe, 537,
- selection of reagent, 443 759
desorption, 1092, 1095 direct on-line coupling, 1095
- conditions, optimisation, 447 direct SPME, 750
desorption chemical direct trapping, 512
ionization-membrane inlet mass directional heterogeneity, 65
spectrometry, 541 directional variability, 66
detection, 943 discrete sample, 34
detergent, 1094 discs, 346, 360, 362, 733
determination of D and L-Leucine, 651 dispersed phase, 298
developmental biology, 617 displacement, 751
dextran ladder, 631 displacement chromatography, 734
dialysis, 483, 503, 761, 818, 825, 878, dissolution, initial rate of, 319
882, 978 dissolved organic carbon, 226
1,2-diamino-4,5-methylene-dioxy- dissolved organic matter, 249
benzene, 628 distribution coefficient, 224, 227, 749,
diazepam, 929 750, 751, 757, 759
1112
distribution constant electret detectors, 17
- calculation of, 454 electric field, 176
- prediction of, 438 electrochemical detection, 642
distribution ratio, 297, 533, 726 electrodialysis, 485
disulphide, 627 electron capture atmospheric pressure
dithiocarbamates, 295 chemical ionization, 653
dithiothreitol, 623 electron capture detector, 6, 13
DNA analysis, 920 electron energy loss spectroscopy, 170
DNPH, 10 electron microscopy, 168
DNPH impregnated cartridges, 91 electron spin resonance, 996
Donnan exclusion, 106 electronic interactions, 742
donor-controlled extraction, 511 electrophoresis, 114, 609
donor phase, 754, 757, 759 electropolished canisters, 87
dosimeter tubes, 5 electrospray ionization, 639, 646, 656
double-tube samplers, 69 electrostatic aerosol analyzer, 178
doughy materials, 67, 68 electrostatic and thermal precipitation,
dredged material, 219 163
dredges, 81 electrostatic attraction, 181
drinking water, 219 electrostatic collection, 176
drop dissolution in the aqueous phase, electrostatic precipitators, 176
318 elixirs, 795
drop-bottle system, 40 elution, 369, 735, 736, 745, 1092, 1093
droplet countercurrent emulsifying agents, 799
chromatography, 314 emulsion liquid membrane, 505
drops, 129, 314 emulsions, 300, 726, 727
drug binding studies, 472 enantiomer, 651
drug-receptor binding, 243 end-capping, 737
drugs, 926, 930 endocrine disruption, 14
- illicit samples, 792 endotoxins, 24
drying step, 375 enrichment factor, 299, 511, 723
dual channel ion chromatograph, 198 enrichment techniques, 950
dual wavelength absorbance detection, environmental analysis, 244, 294, 525
132 environmental hazards, 727
Dufours gland, 689 environmental samples, 532
Durham sampler, 26 environmental tobacco smoke, 1, 23,
dynamic (flowing) extractions, 592 201
dynamic headspace, 279, 287, 730 enzymatic reactions, 615
- analysis, 705 ePTFE, 113, 126
- sampling, 670 equilibration, 685
dynamic LPME, 322 equilibration time, 460, 683
dynamic systems, 685 equilibrium conditions, 395
dynamics of release, 684 equilibrium dialysis, 244
equilibrium partitioning, 247
eicosanoids, 654 equilibrium principles, 256
Einstein diffusion equation, 316 essential oil, 702
1113
esterification, 610 FeCl 3, 129, 130
estradiol, 649 fenfluramine, 650
estrogens, 650, 741, 747, 748 fermentation, 540, 545
estrone, 649 - volatiles, 686
ethanethiol, 644 fermenting fruit, 691
ethanol, 293, 647, 930, 1092 fermentors, 531
- in blood, 292, 293 ferrocene, 656
ethyl acetate, 287, 745 fiber SPME, 810
ethylamine, 116 fibre assemblies, 429
ethylation, 957 fibre coatings, 408, 441
ethylene oxide, 287 fibre conditioners, 431
European Pharmacopoeia, 293 fibre preparation, 414
evaporation, 162, 532, 534 field analysis, 722, 723, 727, 732, 746,
excimer, 621 753, 760
exhaustive electromigration injection, 134 field portable SPME-FastGC, 431
exhaustive extraction, 255, 257, 487, field TWA sampling using the modified
733, 748 SPME device, 407
expanded PTFE, 105 field-flow fractionation, 222, 234
experimental determination of retention film/drop sampler, 135
factors, 366 films, 129, 132
experimental parameters affecting filter-based automated particle
extraction efficiency, 415 collection system, 183
explosives, 895, 899, 905, 924 filter collection, 173
extraction, 255, 723, 944 filters, 98
- combined with derivatisation, 274 filtration, 98, 163, 180, 482, 722, 733,
- conditions for heterogeneous 848
samples, 458 fine particles, 98, 161
- conditions, optimization of, 265, 456 fingerprinting, 901
- convolution model, 264 flame ionization detector, 6, 9, 175
- efficiency, 224, 511 flame photometric detector, 13, 164
- kinetics, 750 flash vaporization-flame photometric
- mode, selection of, 444 detection (FV-FPD), 164
- of solids, 260 flat sheet membranes, 481
- parameters, 439, 724 flattened glass coil, 203
- phases, 751 flavor, 699
- rate law, 303 flavorings, 797
- syringe, 522 flophemesyl, 646
- time, 452, 453 Florisil, 738, 763
- with coated fibre, 397 flow injection analysis, 116, 307
- with in-tube SPME, 401 flow proportional composite sample, 34
exuvia, 677 flow rate, 515, 1093
flow-through equilibrium, 733
fabric denuders, 138 flow-through techniques, 266, 446, 683
faeces, 792 9-fluorenylmethylchloroformate, 617
falling drops, 130, 132 fluorescamine, 205
1114
fluorescein mercuric acetate, 119 galactose, 652
fluorescence, 12, 14, 20, 109, 119, 612, galactosemia, 652
621, 631, 635 gamma-hydroxybutyric acid, 932
- quantum yield, 616 Garret's sampler, 47
- tagging, 205 gas chromatography, 102, 321, 447, 643,
fluoresceneisothiocyanate, 619 728, 941
1-fluoro-2,4-dinitrophenyl-5-L-alanine, gas chromatography-photoionization
658 detection, 164
fluorogenic reactions between gas chromatography-mass spectrometry
nucleosides and aldehydes, 633 (GC-MS), 6, 7, 27, 92, 103
fluorogenic reagent, 616 gas chromatography-flame photometric
4-fluoro-7-nitro-2,1,3-benzoxadiazole, detection (GC-FPD), 21
620 gas diffusion detection device, 117
fluorophore, 619 gas diffusion flow injection analysis, 125
fluorophoric reagent, 616 gas mixtures, 464
FMA, 130 gas separation membranes, progress in,
fodder plants, 79 987
food analysis, 609, 869 gaseous matrices, 464
- and industrial analysis, 526 gaseous pollutants, 98
- and packaging industries, 294 gaseous samples, 532
forensic science, 919 gases, interference from, 199
forest soil, 75 gases, sampling from, 538
formaldehyde, 1, 9, 10, 14, 106, 124, gas-liquid equilibrator, 135
421, 651 gasoline, 899, 903, 922
formic acid, 121, 134, 147, 151, 646, 647 gas-phase reactions, 541
fossil fuel, 913 gas-tight syringe, 290
Fourier transform infrared Gatling gun design, 100
spectroscopy, 9, 168 gene expression, 617
fractional distillation, 682 gentamicin, 639
fragrance analysis, 699 gentrisone, 644
frass, 677 glass honeycomb denuders, 100
freely dissolved concentrations, 241 glass annular denuders, 101
freons, 284 glass denuder, 103
fritted glass filter, 181 glass fiber filter, 183
frontal chromatography, 361, 724, 734 glass parallel plate wet denuder, 176
Frossling's equation, 129 glucose, 124
full evaporative technique (FET), 289 glutamate, 614
fulvic acid, 762 glutaraldehyde, 9
fungal volatiles, 686 glutathione, 616, 641
fungi, 24, 26, 688 glycerol/phosphoric acid, 104
5-furoylquinoline-3-carboxaldehyde glycine, 1093
(FQ), 616 glycols, 137
fusible materials, 68 glycoproteins, 630
gold film, 130
GAC, 1006 Gore-Tex, 105, 113
1115
Gormley-Kennedy equation, 105 - removal, 1017
grab samplers, 81 helium ionization detector, 23
gradient elution, 122 helium-purge membrane inlet, 537
granular activated carbon (GAC), 1005 hematin, 111, 112
graphitized carbon black, 349, 352, 739, Henry's constant, 295, 729
744, 1033 Henry's law, 729
gravel, particle size distribution of, 233 heptafluorobutyl, 650
gravitational settling, 163, 181 herbicides, 619, 741, 906
grease, 67, 901 heterogeneity, 63, 65
green leaf volatiles, 677 hexavalent, 183
Griess-Saltzman reagent, 20, 128, 133 hexylchloroformate, 649
Grignard reaction, 958 high performance liquid
gross sample, 61 chromatography see HPLC
groundwater, 219 high-resolution transmission electron
guanine, 633 microscopy, 170
high specific surface area, 739
H202, 103, 124 Hildebrand solubility parameter, 584
H2S, 117, 130 Hirst-type samplers, 26
H2SO4, 102, 103, 129, 137, 164 HNO3, 137
hair, 789 hold-up volume, 365
halide generation, 958 hollow fiber, 481, 520, 759, 762, 975, 977
haloacetic acids, 647 homocysteine, 622, 623, 625
halobimanes, 624 homogeneous polymeric SPME coatings,
halocarbons, 727 1052
halogenated hydrocarbons, 294, 752, horizontal pipelines, 49
754 horseradish peroxidase, 110
Hantzsch synthesis, 12 hot water extraction, 587
hazardous air pollutants, 6 hot water reverse phase LC, 601
hazardous nuclear material, 208 house-flies, 687
HCHO, 103, 106, 107 HPLC, 10, 13, 27, 611, 613, 633
headspace, 245, 272, 320, 329, 444, 461, HPLC-FLU, 631
491,678, 689, 690, 704, 728, 759 HPLC-MIMS, 548
- configuration, 397 HPLC-UV, 14
- gas chromatography, 279, 295 human breath, 500, 932
- solid phase microextraction, 248, human colon adenocarcinoma, 617
250, 398, 580, 709, 750, 810, 922, human exposure to pollutants, 201
931 humic acid, 249, 762
headspace vial, 280, 282, 284, 285, 287, humic matter, 237, 721, 740, 753, 758,
289, 291, 295 761
- multiple sampling from, 285 humidifier, 164
health effects, 2 hybrid systems for water treatment, 999
health hazards, 727 hydrazines, 632, 638
heavy metals, 162, 237, 895, 908, 909 hydrazones, 632, 640, 650
- effect of pH and redox potential on, hydride generation, 956
225 hydrodistillation, 702
1116
hydrogen bonding, 736, 737 indoors inhalable particulate matter
hydrogen chloride, 151 concentrations, 201
hydrogen exchange rate, 170 indophenol blue, 103, 104, 115, 129
hydrogen fluoride, 151 induction coupled plasma-mass
hydrogen interactions, 742 spectrometers (ICP-MS), 209
hydrogen peroxide, 109, 111, 135 inductively coupled argon plasmas, 171
hydrophobic chemicals in soil, 250 inertial impaction, 180
hydrophobic drugs, 294 inertial methods, 163
hydrophobic filter, 192 infinite-dilution activity coefficients, 471
hydrophobic interactions, 736, 737, 738, infrared, 9
740 inhalable particulate matter
hydrophobic membrane, 757, 759 concentrations indoors, 201
hydrophobic polytetrafluoroethylene inhalants, 796
filter, 192 injectables, 796
hydrophobicity, 113 injection volume, 284, 285
5-hydroxy-2-aminovaleric acid (HAVA), inorganic amines, 545
648 inorganic analyses, 659
hydroxyapatite, 139 inorganic anions, 659
hydroxyethyl starch, 644 inorganic cations, 659
hydroxyls, 643 inorganic ions, 724, 740
hydroxymethyl hydroperoxide, 109 inorganic oxide adsorbents, 347
2-hydroxymethylpiperidine, 103 inorganic salts, 721, 753
insect cuticles, 684
ifophosphamide, 632 insect pheromones, collection and
ignitable liquid, 921 isolation of, 670
illicit drug samples, 792 insecticides, 906
Illicit pharmaceuticals, 793 insects, 24
immersed membrane, 999 insulin, 615
immobilized antibodies, 410 interception, 180
immobilized fluorophore, 123 interfacial adsorption, 305
immunoaffinity extraction, 814, 861, interferences from prolonged sample
878 collection, 168
immunoaffinity sorbents, 741 inter-laboratory studies, 463
immunoassay, 898 internal standards in headspace
immunoconjugate, 1082 quantitation, 289
immunogenicity, 617 internal surface, 236
immunological trapping, 513 interparticle velocity, 363
immunosorbents, 357, 744, 1088 intramolecular condensation, 640
impaction, 179, 193 in-tube SPME, 266, 401, 810
impactor, 203 in-tube SPME-HPLC, 426
indirect trapping, 513 in-vial desorption, 347
indoor air environment, 203 iodoacetamide, 642
- kerosene-fueled space heaters, 201 ion chromatography, 16, 18, 19, 120, 757
indoor HONO from open flame sources, - coupling, 151
137 ion exchange beads, 1006
1117
Ion exchange fibers to remove inorganic leachates, 219
contaminants, 1014 leaching techniques, 944
ion exchange groups, 740 lead, 660
ion exchange mechanisms, 737, 739 leafminer moths, 688
ion exchange membranes, 105, 125 lepidopteran pheromones, 688
ion-exchange sorbents, 353, 739 leucine, 620
ion exchange preconcentration, 176 light extinction, 172
ion pair sorbents, 739 limit of detection, 724, 743
ion suppression, 722 Limulus amebocyte lysate (LAL) assay,
ionic interactions, 736, 737 24, 25
ionic strength, 286, 740 linalool, 287
ionizable compounds, 752 linear alkylbenzene-sulfonates, 740
ion-pair extraction, 509, 726, 727, 740 linear dynamic range, 457, 460
isoamyl alcohol, 292 linear sample capacity, 298
isobutylchloroformate, 649 liquid chromatography/electrospray
isobutyric acid, 121 ionisation mass spectrometry, 409
isokinetic samplers, 36 liquid chromatography/mass
isopeptides, 638 spectrometry, 409LC-MS, 27
isoprene, 932 liquid core waveguide (LCW), 109
isopropanol, 282 liquid drop, 130
isotachophoresis, 969 liquid extraction system, 187
isotherm, 298 liquid filled fiber optics, 126
isothiocyanates, 619, 638 liquid matrices, 467
isotopic dilution, 458 liquid membrane, 486, 504
isotopic enrichment, 614 liquid phase microextraction, 301, 523
isotopic ratios in sugars, 652 liquid-liquid interface, 298
liquid-liquid extraction, 297, 701, 712,
kaolinite, 230, 236 725, 756, 757, 762, 801, 849, 878,
kerosene heaters produce acid aerosols, 913, 926, 968, 969, 1025
206 liquid-phase microextraction, 821
ketones, 26 liquids, sampling from, 541
kinase, 657 living systems, 472
K6nig reaction, 118 Los Angeles, fine particle mass in, 172
Kovats retention index, 685 low-specificity sorbents, 349
Kuderna-Danish evaporator, 14, 728 lumpy materials, 65
lymphocytes, 617
lag time, 333 lysimeters, 46
laminar discs, 344
large volume injector, 723, 745 macrocyclic ligands, 356
lasalocid, 640 magnetic stirring, 446
laser ablation, 919 Makrofol polycarbonate plastic, 17
laser desorption/fast gas malachite green, 130
chromatography, 391 malathion, 931
laser desorption/ionization, 169 MALDI (matrix assisted laser
laser induced fluorescence (LIF), 614 desorption ionization), 325, 615
1118
MALDI-TOF, 627, 631, 656 - with sorbent interface (MESI), 493,
maleic acid, 147 759
maleimides, 626 membrane inlet designs, 536
malondialdehyde, 650 membrane inlet mass spectrometry
malonic acid, 147 (MIMS), 531, 759
malonyldialdehyde, 648 membrane matrix, 117
manual headspace, 290 membrane probe, 531, 537
Marfey's reagent, 658 membrane separation, 479
marine pollution, 762 membrane surface characterization
marking pheromones, 671 - by atomic force microscopy, 994
mass median aerodynamic diameter - by electron spin resonance, 996
(MMAD), 102 - by measurement of pore sizes, 993
mass spectrometry, 13, 531 membrane techniques, 881
- for aerosol analysis, 169 membrane temperature, 535
mass transfer, 506 membrane transport, progress in, 998
matrix-assisted laser desorption membrane vesicles, 248
ionization see MALDI memory effects, 300
matrix effects, 306 Meperidine, 928
matrix interferences, 721 mercaptans, 118
matrix matching, 306 2-mercaptoethanesulphonic, 641
matrix modifier, 287, 294 mercaptoethanol, 623
matrix solid-phase dispersion, 875, 885 2-mercaptoethanol (2ME), 613
maximum rate extraction, 489 3-mercaptopropionic acid (MPA), 613
maze type impactor, 179 mercury, 659
MBTH, 11, 129 MESI, 494
- reaction with HCHO, 129 MESI-GC, 495, 500
McIntyre-Day scoop, 81 metabolic cytometry, 636
measurement of the fate of emitted H2 S metabolites, 741, 913
in an oil field, 207 metal complexation, 356
measuring cell, 537 metals, 908, 930
MECC, 615, 619 methacrylic acid, 121
medical air, 680 methadone, 928, 929
membrane-assisted solvent extraction, methamphetamine, 650, 927
523 methanator, 175
membrane-based denuders, 104 methane, 9
membrane-based extraction, new methanesulfonic acid, 151
polymeric materials, 1044 methanol, 113, 647, 930, 1092, 1094
membrane-based passive samplers, 54 methanol-fueled vehicle (MFV), 125
membrane-based sample preparation, method detection limit, determination
818 of, 462
membrane-based techniques, 754 method selection, 372
membrane-controlled extraction, 511 method validation, 463
membrane extraction, 275, 520, 819 methyl-4-aminomethyl benzoate, 631
- apparatus, 520 methylamine, 116, 620
- progress in, 1000 methylarsonic acids, 660
1119
methyl benzoate, 932 mixed-bed ion exchange column, 187
3-methyl-2-benzothiazoline hydrazone, mixed-bed sorbents, 1097
129 mixed-mode sorbents, 354, 740
N-methylcarbamates, 727 mixing, 284, 290
methyl ethyl ketone, 292 mixing chamber, 193, 203
methyl formate, 647 mobile phase, 735
N-methylglycamine, 620 modifiers, addition of, 597
methyl hydroperoxide, 109 molar absorbtivity, 616
methylenedioxymethamphetamine, 928 mold, 26, 204
methylmercury, 931, 941 molecular imprinted SPE, 742
Metricel, 105 molecular imprinting, 816
Mexican fruit flies, 686 molecular recognition, 741
micellar electrokinetic chromatography, molecular sieve membranes, 987
134 molecularly imprinted polymers, 357,
microcosms, 911 741, 744, 816, 878, 1035, 1053
microdialysis, 614, 878, 882 momentum transfer, 305
microdrop-LPME, 315, 321 monensin, 640
microenvironmental exposure monitor, monoamines, 621
15 monobromobimane, 623, 624
microextraction techniques, 880 monochlorobenzene, 746
micro-fibres, 473 monoclonal antibodies, 1083, 1087, 1091
microfilms of water, 46 monocytes, 617
microfiltration membranes, 992 monodisperse aerosols, 169, 193
microinjection-microspectroscop, 326 monofunctional silane, 737
micro-LC, 1097 monomers in polymers, 279, 284
micromachining, 642 monosaccharides, 644, 652
microplates, 854 monosegmented flow extraction, 312
microporous membrane liquid-liquid montmorillonite, 230, 236
extraction (MMLLE), 505, 518, 757, morpholinoethanesulfonic acid, 152
763, 821 mortality data, 161
micro-ring electrode, 130 Muencke-type absorbers, 14
microscale glassware, 678 multichannel membrane collector, 126
microsomes, 248 multichannel sampler, 38
micro-SPE, 389 multimodal size distribution, 162
microsyringe, 321 multimode extractions, 354
microtiter plates, 746 multiple concentric tubes, 101
microwave-assisted extraction, 566, 572, multiple headspace extraction, 285, 287
822, 883, 945 multiple PPD, 102
microwave ovens, 568
mildew, 26 Na 2 B40, 133
milk, 788 Nafion, 105, 140, 164
MIMS theory, 532 Nafion dryer, 731
mineral insulating oils, 294 Nafion membrane DS (NMDS), 106
mineral oil, 901 Nafion suppressor, 206
mixed functional phase, 1040 nanodrop-LPME, 315, 325
1120
nanofiltration, 998 non-porous membranes, 504, 755, 756
nanopore membranes, 113 non-target analysis, 722
naphthalene, 13 non-volatile organic compounds, 532
naphthalene sulfonates, 740, 748 non-wetting phase, 310
naphthalene-2,3-dicarboxaldehyde nuclear solid waste, 909
(NDA), 615 nucleation, 162
narasin, 640 nucleophiles, 622
National Ambient Air Quality Nurnik-1 sampler, 41
Standards (NAAQS), 18, 19, 22, 98 nylon filters, 168
National Institute for Occupational
Safety and Health (NIOSH), 4, 5, 03, 102
13, 23, 651 OASIS column, 631, 656
natural organic matter, 753, 762 octanol/water partition coefficient, 736,
needle vibration technique, 446 897
negligible depletion SPME, 245, 247 octanol/water distribution constant, 361
nephelometers, 16 odors, 699
Nernst's distribution law, 725 off-flavors, 699
neurodegenerative disorders, 614 off-line automation of SPE, 746
neutron activation analysis, 16 off-line cleanup, 969
neutrophils, 617 off-line equipment, 523
NH3, 102, 129 off-line immunoextraction, 1094
N-hydroxysuccinimidyl esters, 610 off-line preconcentration, 969
N-hydroxysuccinimidyl off-line sample clean-up, 968
fluorescein-O-acetate , 622 off-site analysis, 722
nicotine, 23, 102, 928 oil film thickness on the surface of
Niemisto corer, 81 water, 47
nitrate, 139, 166, 196, 198, 659 oligosaccharides, 630, 654, 655
nitric acid, 151, 168 on-column electrophoretic
nitrite, 139, 200, 166, 659 concentration, 975
nitroaromatics, 638 on-disc derivatization, 347
nitrogen blowdown, 728, 733 on-line concentration, 972
nitrogen dioxide, 1, 19, 104 on-line connection to chromatography,
nitrogen-phosphorus detector, 13 522
4-nitrophenol, 742 on-line coupling, 379, 969, 978, 1095
nitrous acid, 128, 151 on-line electrophoretic concentration, 972
NMR, 27 on-line LLE-GC, 727
NO 2, 23, 101, 102 on-line membrane cleanup, 977
non specific interactions, 1094 on-line preconcentration, 1095
non-dispersive infrared analyzer, 9, 22, on-line solid-phase extraction, 743
101 on-line vs. off-line Analysis, 843
non-exhaustive extraction, 256, 277, on-site analysis, 722
733, 748 on-site monitoring, 544, 747, 753, 760,
non-hygroscopic aerosol, 196, 197 897
non-porous hydrophobic membrane DS OPA, 613
devices, 125 organic aerosols, 201
1121
organic-carbon normalized partition - optical microscopy, 234
coefficient, 237 - radiation scattering, 234
organic chloramines , 545 - sedimentation, 234
organic compounds, 540 - sieving, 233
organic matrixes, 546 particle-embedded glass fiber discs, 344
organic pollutants, 560 particle-loaded membranes, 344
organic solvents, 259, 319, 753, particle-particle reactions, 169
1093-1095 particles
organic vapors, 137, 169 - prehumidification of, 164
organic volatiles in drinking water, 287 - thermal decomposition of, 164
organoarsenic, 941 particulate air pollution, 161
organochlorine pesticides, 12, 727 particulate matter, 1, 14, 23
organophosphate pesticides, 931 particulate sulfate, 208
organophosphorous pesticides, 746 partition coefficient, 228, 280, 281-286,
oscillometric conductometry, 114 289, 290, 294, 295
OSHA (Occupational Safety and Health partition law, 725
Administration), 4 partitioning, 725
outer particle surface area, 233 passive dosimetry, 53
overaspiration, 206 passive methods, 54
oxalic acid, 147, 151 passive samplers, 4, 7, 54, 346, 761
oxidase enzymes, 124 passive time weighted average
oxidation, 174, 648 sampling, 272
oximations, 654 passive water sampling, 53
ozone, 18, 124, 208, 646 pattern-recognition techniques, 294
- damage to the lung, 161 PDMS, 400, 408, 441, 442, 465, 494
ozonides, 646 peat samples, 549
Peltier cooled maze, 203
packing density, 363 pentaacetylaldonitrile, 652
palm weevils, 687 pentadecafluorooctanoyl, 650
Palmes tubes, 19 pentafluorobenzaldehyde, 651
parallel plate denuders, 101, 104, 141 pentafluorobenzyl bromide, 653
parallel processing, 743, 746 pentafluorobenzyl hydroxyamine, 12
Paraquat, 170 o-(2,3,4,5,6-pentafluorobenzyl)-
pararosaniline, 10, 21 hydroxylamine, 421
Parkinson's disease, 614 pentafluorobenzylation, 646, 650, 654
partial pressure, 729 pentafluorophenyldimethylsilyl, 646
partially concurrent solvent pentanedione, 109
evaporation, 745, 746 peptides, 615, 616, 622, 653
particle collection, 179 perfluorinated reagents, 649
- by vapor condensation, 192 perfluorotolyl, 650
particle deposition, 125, 140, 144 periodic heterogeneity, 65, 66
particle size distribution, 232 periodical analyses, 49
particle sizing permanent gases, 294
- centrifuging, 234 permeation, 117, 486, 489
- electron microscopy, 234 - denuders, 117
1122
- principles, 256 o-phthalic acid, 147
peroxidase substitute, 111 phosphatidylcholine, 634
peroxyacetyl nitrate, 101, 137 phosphines, 623
peroxyoxalate chemiluminescence photoacoustic spectroscopy, 9, 173
(PO-CL), 123 photoacoustic-IR, 22
perpendicular pipelines, 49 photoionization, 169
pertrimethysilylated, 643 photoionization detector, 9
perturbation, 336 photolytic activation, 636
pervaporation membranes, progress in, phthalates, 13
989 physicochemical applications, 470
pesticides, 12, 295, 609, 741, 742, 895, physicochemical constants, 276
906, 931 physicochemical properties of aqueous
Petri dish, 26 and solid environmental matrices,
petroleum, 895 219
petroleum hydrocarbons, 899, 909 phytoextraction, 909
pH, 416, 456, 724, 752 phytoremediation, 897, 909
- value of soil or sediment samples, picodrop-LPME, 315, 326
231 pipe sampler, 72
- value of water, 222 piped water, 52
- variations, 1092 planar diffusion scrubbers, 116
Phanaerocheatechrysosporium, 912 planar DS, 117
pharmaceuticals, 293 planar molecules, 739
pharmacokinetics, 642 plant volatiles, 691
phase ratio, 280, 285, 299, 725 plants, 688
phase separation, 307, 727 plasma, 209, 785, 929
phenanthrene, 13 plate number, 362
phencyclidine, 929 platelet activating factor, 650
phenolic compounds, 747 platinum-DDTC, 642
phenolic estrogens, 653 Plexiglas denuder, 153
phenolic substances, 134 plutonium, 209
phenols, 13, 653, 727, 737, 747 PM10, 98, 99
phenothiazines, 931 PM2.5, 98, 99, 162
phenoxy acid herbicides, 754 poisons, 926
phenyl silica, 738 polar compounds, 540
phenylboronic acids, 356 polarizable electrons, 739
phenylisothiocyanate, 638 pollen, 24, 25
phenylureas, 1091, 1095, 1097 polluted aqueous samples, 763
pheromones polluted wastewaters, 52
- biological contexts, 673 poly(dimethylsiloxane) (PDMS) coating,
- definition, 670 399
- gland extracts, 677 poly(styrene-divinylbenzene)
- methods for sampling and copolymers, 738
collection, 674 poly(styrene-divinylbenzene)-based
- types of, 670 materials for SPE, 1034
o-phthaldialdehyde, 115, 612 polyacrylate, 751
1123
polyamines, 621 porous hydrophobic membranes, 113
polyaromatic hydrocarbons, 227, 899, porous membrane diffusion scrubbers
1096 (PMDS), 113, 114, 118
polychlorinated biphenyls, 13, 100 porous membranes, 117, 754, 975, 977
polychlorobiphenyls, 903 porous polymer beads, 343
polyclonal antibodies, 1083, 1087 porous polymers, 349
polycyclic aromatic compounds, 13, 23, post-sampling aerosol analysis:, 167
92, 741, 747 potassium, 198
polydimethylsiloxane, 751, 752 Prandtl boundary layer, 440
polyether antibiotics, 640 precipitation, 219
polymer coatings, 752 precision, 460, 744
polymer sorbents, 738 precolumn, 743, 744
polymerase chain reaction, 27 precolumn dimensions, 379
polymeric composite coatings, 1052 preconcentration, 183, 297, 968, 969
polymeric extraction materials in SPE, - of volatile organic compounds, 45
1034 pre-equilibrium principles, 256
polymeric materials, 1024 preservatives, 799
polymeric membrane extraction, 505, pressure drop, 362
761 pressure-balanced headspace sampling,
polymeric resins, 738 291
polymeric surface coatings for SPME, pressurised fluid extraction, 573
1050 pressurized liquid extraction, 885, 886
polymers, 552 primary samples, 61
polynuclear aromatic hydrocarbons, - of wild plants, 79
100, 170 primer pheromones, 672
polypyrrole, 408, 409 probe for collecting samples from
polypyrrole-based coatings, 1054 pipelines, 49
polysaccharides, 627, 654 profile sampling, 72
polystyrene-divinylbenzene, 732, 744, proline, 648
751 n-propanol, 292, 293
polystyrene-divinylbenzene-N-vinylpyrr propionic acid, 121, 147
olidone, 739 proportionality constant, 685
polyurethane foam, 12, 14, 92 prostaglandins, 654
polyvinylidene fluoride, 113 protein precipitation, 626, 847
Porapak Q, 679, 681 protein removal, 800
pore blockage, 125 proteins, 615, 616, 656
pore size, 113 proteome analysis, 978
pore structure, 1009, 1011 Pseudomonas pseudoalcaligenes,913
pore tortuosity, 113 pulsed fluorescence analyzer, 164
Poreflon porous PTFE scrubbers, 123 purge-and-trap, 279, 280, 287, 294, 688,
Poreflon tubes, 114 690, 691, 730, 903
Poropak, 13 purge-and-membrane mass
porosity, 236 spectrometry, 550
porous graphitic carbon, 352, 736, 738, push-pull systems, 680
739, 1034 putrescine, 621
1124
1-pyrenyldiazomethane, 420 response time, 533
pyrolysis, 552 restricted access material, 740, 744, 860,
pyrrolidone group, 739 1039
pyruvic acid, 147 restricted access sorbents, 355
retaining column, 745
quadrupole MS, 169 retention, 365
quantitation, 287, 289, 487 retention factors, 366, 735, 738
- techniques, 286 - methods for estimating, 366
- problems with, 289 retention gap, 745, 763
quantitative analysis by LLE, 306 retention volume, 735
quantitative extraction, 418 retinitispigmentosa, 656
quartz fiber filters, 168, 175 reverse osmosis membranes, 985
quaternary ammonium herbicides, 747 reverse osmosis, progress in, 983
quaternary ammonium salts, 740 reversed phase sorbents, 735
reversed-phase liquid chromatography,
racemization, 652 137
radioactive waste, 909 reversed-phase membrane inlet mass
radioimmunoassay, 25 spectrometry, 546
radon, 1, 16 rhinoceros beetles, 687
rainwater, 48 rhodamine, 616, 636
Raman microspectroscopy, 18 Rhodococcus, 913
random heterogeneity, 65 rhodopsin, 656, 657
random variability, 69 ribosides, 633
rate determining step, 302 rinse solvent, 369
reactions, monitoring, 545, 547 robots, 746
reactive oxygen species (ROS), 616 root layer soils, 75
real-time emission spectrometry, 171 rotary evaporator, 728
real-time single particle mass rotating wet annular denuders, 140
spectrometry, 171 Rotorod sampler, 26
receiving phase, 330 rubber, 24
recognition pheromones, 671 rumen fluid, 549
recovery, 511, 723, 735, 1090
redox potential, 224 saccharides, 636, 654
- in soils and sediments, 231 - determination via derivatization of
reductive amination, 629, 630, 636 the carbonyl, 627
reference material, definition, 940 safe sampling volume, 369, 734
refluxing with a hydrophobic filter, 192 salinomycin, 640
regeneration, 1093 saliva, 549, 787
relative response factors, 685 salt
repeated sampling, 689 - adding to aqueous samples, 282,
representative sample, 67 284, 457, 753
residual solvents in pharmaceuticals, - concentrations, 416, 984
279, 293 - salting out, 284, 312, 724, 726
respirable particulates, 15, 23 sample cleanup, 297, 723, 738, 742, 756,
respiratory ailments, 161 758, 762, 763, 817, 878, 968
1125
sample degradation, 722 scintillation cells, 17
sample handling, 289 seashore waters, 46
sample holding time, 92 secondary chemical equilibria, 298
sample preparation, 722 secondary flow, 311
- for analysis, 55 sedimentary matter, 219
- on-line ASTED, 861 sediments, 81, 549
- on-line immunoaffinity extraction, segmented flow, 183, 301, 727
861 segmentor, 307, 727
- on-line RAM columns, 860 segregation, 64
- on-line turbulent flow, 859 selective volatilization, 174
sample processing considerations, 373 selectivity, 517, 758
sample stability, 746 semi-infinite spherical diffusion, 318
sample stacking, 972 semiochemicals, 670
sample volume, 372, 744, 751, 753 semipermeable membrane devices), 54,
- and extraction recovery, 1090 761
- optimization, 448 semipermeable surface, 1040
samplers with a shutter closure, 82 semi-preparative column
samplers for loose and lumpy material, chromatography, 631
69 semi-volatile compounds, 540
samplers for loose materials, 71 semi-volatile organic compounds, 87, 91,
samples 759
- of undisturbed structure, 73 sensitivity, 460, 1097
- number necessary, 62 separating funnel, 726
- without structure preservation, 73 separation/detection technique,
sampling selection, 447
- biological plant materials, 78 Sepharose, 1089, 1095
- for speciation analysis, 944 sequential composite sample, 34
- from horizontal barrel boilers, 51 sequential injection analysis, 115, 129
- from streams, 36 sequential injection extraction, 312
- ground cover, 79 sequestration, 897
- leaves, 80 serial processing, 746
- groundwater, 42 serial sample processing, 742
- water and vapour from closed Serpentine-II, 106
circuits, 49 serum, 785, 926
sampling plan, 62 serum protein binding, 248
sampling port diameter, 685 serum proteins, 242
sampling protocol, 62 sewage sludge, 219
sampling site selection, 77 sex pheromones, 671
sampling soil from surface layers, 76 SFE see supercritical fluid extraction
sampling stick, 74 shielded hydrophobic phase, 1040
sand, particle size distribution of, 233 sialic acids, 627, 655
scanning electron microscopy (SEM), sick building syndrome, 24, 202, 203
994 silanol groups, 737
scanning probe microscopy, 170 silica, 347, 656, 732, 738, 763
Schiff Base, 629 silica-based extraction materials, 1031
1126
silicone membrane DS, 126 sol-gel monolithic beds for extraction,
silicone rubber membrane, 118 1065
silt, particle size distribution of, 233 sol-gel polymeric extraction media, 1056
silylation, 643, 654 sol-gel polymers, 1025, 1041
simple composite sample, 34 solid adsorbents, 89
simple permeation, 506 solid analysis, 532
simultaneous distillation-extraction, solid environmental matrices,
702 definition, 219
single-cell analysis, 980 solid environmental samples, 560
single drop-LPME, 315, 318, 320 solid matrices, 469
single jet inertial impactor, 198 solid phase analytical derivatization,
single liquid drop techniques with two 618
phases, 301 solid phase extraction see SPE
single particle mass spectrometry, 170 solid phase microextraction see SPME
single-tube denuders, 99 solid sampling, 550
single-tube glass denuder, 104 solid versus liquid sorbents, 268
single-tube samplers, 69 solid-liquid extraction, 731
single-tube wet denuder, 138, 137, 181 solid-solution partitioning of metals,
size exclusion, 740 238
size exclusion chromatography, 234, solid-supported LLE, 850
761, 763 solids, extraction of, 260
Slocum's sampler, 39 solids injection syringe, 690
sludge, 469 solubilities, 725
small-volume injector liners, 685 solubility parameter, 583
SO2, 129, 151 solvation parameter model, 367
soda-lime trap, 187 solvent evaporation, 728
sodium, 198 solvent extraction, 677
sodium acetate, 644 solvent-free standards, 413
sodium borohydride, 623 solvent selection, 583, 726
sodium cyanoborohydride, 630, 631 solvent vapor exit, 745, 746
sodium dodecylsulphate, 654 solventless gas chromatographic
sodium perchlorate, 183 injection, 647
sodium tetraethylborate, 660 solventless sample preparation, 898, 899
soil, 72, 219, 284, 286 sonication, 901, 902, 906, 950
- soil analysis, 77, 895, 899 soot particles, 170
- (bio)remediation, 911 sorbent, 732
- hydrophobic chemicals in, 250 - with enhanced selectivity, 740
- profile, 72 sorption of organic compounds, 238
- samples, 532, 550 source layer, 319
- slurries, 577 Soxhlet, 14, 582, 682, 899, 901, 902
- surface pollution, 76 SPE, 27, 713, 245, 731, 805, 878, 968
- texture, 233 - cartridge, 266
- volatile organics in, 287 - historical development, 342
soil-water partition coefficients, 897 - -liquid chromatography, 379
solanesol, 23 - off-line/cartridges, 851
1127
- off-line/microplates, 854 stationary phase, 735
- on-line, 858 steady-state extraction, 490
- new polymeric materials in, 1029 steady-state flow, 532
- theory, 359 steam distillation, 677
- traditional materials, 1030 steam injection rates, 193
- working principle of, 1029 steric hindrance, 620
speciation, 335, 908 steroids, 930
- definition, 939 stiffening agents, 799
- monitors, 161 stinkbugs, 688
specific attenuation cross section, 173 stir bar sorptive extraction, 824, 880
specific interactions, 1094 stirring rate, 321
spectrophotometer, 21 stomach fluid, 549
spectroscopic methods for aerosol storage conditions, 94
analysis, 167 storage stability, 88
spermidine, 621 subcritical water extraction, 587
spermine, 621 submersible rectangular sampler, 47
spherical drop diffusion, 317 succinic acid, 147
spin bar sorptive extraction (SBSE), 878 sugarcane weevils, 687
SPME, 4, 7, 12, 98, 102, 223, 227, 389, sulfamic acid, 104, 130
683, 701, 807, 880, 906, 919, 951, sulfate, 98, 102, 161, 164, 196, 198, 206,
970 207
- applications, 464, 686 sulfite, 124
- devices, 749 sulfite oxidase, 124
- evolution of, 391 sulfonation, 739, 1018
- for measuring freely dissolved sulfonic acid groups, 739
concentrations, 245 sulfosalicylic acid, 626
- -gas chromatography interface, 422 sulfur, 641
- -high performance liquid sulfur dioxide, 21, 151, 164
chromatography interface, 425 sulfuric acid, 103, 162
- in biomedical analyses, 247 sulfurous acid, 151
- in-tube, 266 SUMMA canisters, 7
- kinetics, 227 Super Q, 681
- method development and supercritical fluid extraction (SFE), 701,
optimization of, 435 714, 803, 878, 886, 902, 949
- new polymeric materials, 1049, 1053 - of food samples, 888
- optimization and calibration, 436 - instrumentation, 562
- principles of, 395 - method development, 566
- time-weighted average sampling, 403 supercritical fluids, 561
standard addition, 287, 306, 458 superheated water extraction, 587
static (non-flowing) extractions, 595 supersaturation, 193
static drops, 129 supported liquid membrane (SLM), 301,
static headspace, 279, 280, 294, 704, 504, 506, 756, 820
729, 903 suppositories, 796
static in-tube SPME time weighted surface-active impurities, 226
average sampling, 403 surface-active substances, 305
1128
surface area, 236 2,3,5,6-tetrafluorophenol, 641
surface-bound macrocyclic ligands, 356 tetramethylammonium hydroxide, 648
surface layers of water, 46 tetramethylbenzidine, 127
surface modification by surface theoretical plates, 735
modifying macromolecules (SMM), thermal decomposition of particles, 164
990 thermal desorption, 3, 8, 12, 102, 103,
surface porosity, 113 683, 690, 730
surface tension, 226 thermal treatment, 174
surface water, 36, 219, 753 thermodenuders, 102, 103, 164
surfactants, 740, 747, 748 thermodynamic constants, 283, 295
suspended particles, 722, 733, 734, 761 thermodynamic studies, 284
suspended sediments, 83 thermodynamics, 257, 436, 752
suspending agents, 799 thermoelectrically cooled maze, 193
sweat, 790 thermophoresis, 140, 193
swelling, 534 thiamine, 112
syringe-based automated system, 290 thin layer chromatography, 27
system map, 368 thiochrome, 112
thioglycol methylate, 660
Taber trap, 26 thiols, 622, 624
tablets, 795 - fluorophoric tagging of, 624
tamoxifen, 636 throughfall waters, 48
tandem mass spectrometry (MS/MS), Ti(IV)-4-(2-pyridylazo)-resorcinol, 123
543 time weighted average sampling, 272
tapered element oscillating time-of-flight mass spectrometry, 169,
microbalance (TEOM), 16, 102 615
target analysis, 722, 747 time-weighted average, 3, 406
Tedlar® bags, 681 tin, 660
Teflon, 105, 204, 117, 126, 127, 169 tissues, 792
temperature, 447, 448, 456, 498, 574, toluene, 128
589, 724, 752 topicals, 795
- aqueous media, 221 tortuosity, 113
- dependence of SLM extraction, 512 total metal burden, 238
- solid samples, 228 total organic carbon, 226, 237
temperature-programmed desorption total suspended particulate matter, 99,
membrane inlet mass spectrometry, 172
540 toxicological analysis, 926
temperature-responsive polymers in trace analysis, 1096
solution mediated extraction, 1026 trace element speciation, 939
template, 741, 742 - instrumentation, 941
Tenax, 13, 679, 681,688, 730, 731 trace gases, 97, 100, 129, 154
termites, 689 trace metals, 208
tert-butanol, 292 - in aerosols, 171
tertiary amine, 618 - mobilization of, 223
test sample, 61 trail pheromones, 672
tetrabutylammonium, 647 transacylation, 650
1129
transducin, 656 - Method IP-2B, 24
trap-and-release MIMS, 535, 541 - Method IP-5A, 19, 20
trap-in-needle system, 433 - Method IP-6A, 10
trapping efficiency, 686 - Method IP-6B, 10
trapping on glass, 691 - Method IP-7, 14
trapping time, 498 - Method IP-8, 12
triazines, 740, 741, 742, 747 - Method TO-13A, 92
trichloroethylene, 128 - Method 5021, 294
1,1,1-trichloromethane, 283 U.S. Pharmacopeia Method 467, 293
trifluoroacetyl, 650 UV absorption, 10
trifunctional silane, 737
trimethyl orthoacetate, 647 vacuum manifolds, 733
trimethylamine, 116 valeric acid, 147
Trimethyliodosilane, 643 validation, 463
trimethyllead, 941 valve and loop-based headspace
trimethylsilylation, 610, 643, 645 autosampler, 290
trimethylsilylimidazole, 644 valve and loop-based systems, 290
trimethyltrifluorotoluylammonium, 648 van Veen grab, 81
tri-octyl phosphine oxide (TOPO), 510 vapor condensation, 192, 208
triolein, 761 vapor distillation, 702
triphenylphosphine, 634 variability, 63
tunable diode laser absorption vial size, 285
spectrometer (TDLAS), 107 vinylacetic acid, 121
turbidity, water, 222 viruses, 24
turbulent flow, 859 visibility degradation due to aerosol
two-stage inlets, 537 light absorption, 173
tylenol, 622 visibility reduction, 174
visible light, 638
ultrafiltration, 244, 800 vitamin B1, 112
ultrasonic extraction, 904 vitamin D3, 655
ultraviolet, 638 volatile acids, 284
unconventional poisons, 930 volatile compound, 744
unforced water discharge, 52 volatile losses, 728
unsaturated soil water, 45 volatile organic compounds, 1, 5, 23, 26,
unsupported liquid membrane 40, 45, 72, 87, 294, 531, 534, 542,
techniques, 301, 329 727, 754, 759, 899, 903
urea, 1093 volatile organics in soil, 287
urine, 293, 786, 926, 927, 930, 931 volatiles in biological fluids, 292
U.S. EPA Methods, 5, 6, 15, 21, 92, 287, volatiles in high-boiling oils, 294
624, 731, 899 volatilization, 169
- Method IP-1OA, 15 volatilization loss, 221
- Method IP-1OB, 16 Vortex technique, 446
- Method IP-1A, 7
- Method IP-1B, 6 washing solution, 734
- Method IP-2A, 23 wasps, 689
1130
wastewater, 219, 750, 753 - glass, 137
water, 541 wet effluent systems, 183
water analysis, applications of SPE for, wetted denuders, 135, 136
747 wetting film, 310, 322
water analysis, procedures, 734 wetting phase, 310
water content of solid matrices, 228 Whitman two-film model, 303
water extraction, 578 whole blood, 785
water purification, 1013 whole-body extraction, 669
water salinity, 221 whole cell suspensions, 980
water sample storage, 56 whole-column cryotrapping, 731
water trap, 731 wines, 651, 881
water-organic modifier, 1092
water-soil partition coefficients, 910 XAD-2, 13, 14, 92
wet aerosol collectors, 199 XAD-4, 23
wet analysis, 175, 176, 200 xenobiotics, 879
wet annular denuders, 140, 141 X-ray diffraction, 18
wet denuders, 104 X-ray emission, 16, 170
- coupled IC analysis of acid gases X-ray fluorescence, 168
and ammonia, 148 xylene, 746
- which design to choose, 147 o-xylene, 289
wet diffusion denuder, 205
wet effluent diffusion denuders zeolite membranes, 989
(WEDD), 136 zone broadening, 307
1131