Journal of Occupational and Environmental Hygiene, 10: 674–684

ISSN: 1545-9624 print / 1545-9632 online

DOI: 10.1080/15459624.2013.831982

Short Duration Needle Trap Sampling with Gas

Chromatography Analysis to Determine Nearly
Instantaneous Concentrations of Selected Organic Vapor
Contaminants
Simon J. Strating,1 Theodore J. Juarez,1 Michael E. Stevens Jr.,1
Duvel W. White,1 and Philip A. Smith1,2
1
Department of Preventive Medicine and Biometrics, Uniformed Services University of the Health
Sciences, Bethesda, Maryland
2
Health Response Team, Occupational Safety and Health Administration, U.S. Department of Labor, Salt
Lake Technical Center, Sandy, Utah

Needle trap device samplers were used for rapid (60 s)
quantitative sampling of short-term exposure limit (STEL) and
peak exposure standard concentrations using a manually oper-
ated pump to collect small volume (10 mL) gas phase samples
containing methylene chloride, benzene, toluene, and tetra-
chloroethylene vapors. Solventless introduction of chemical
samples for gas chromatography analysis with flame ionization
detection yielded linear results (R2 > 0.99) for vapor standard
mixtures of the four target analytes ranging from 10% to 200%
of their respective nominal STEL or peak exposure standard
concentrations. Needle trap samplers showed ≥86% recovery
(as GC-FID peak area responses) following 14-day storage
at room temperature compared to the same samplers analyzed
immediately, with better recovery values observed with shorter
storage (≥95% at room temperature for seven days, except
for methylene chloride) or with storage at 4◦ C. Calibration
for quantitation of concentrations of benzene, toluene, and
tetrachloroethylene was shown to be possible with the use of
an internal standard to account for injector discrimination
between the solventless NTD approach and injections of tar-
get analytes in carbon disulfide. Due to the simple sampling
method (no field calibration and battery-free pumping) and
the avoidance of solvent dilution, a needle trap sampling
approach could simplify sample collection and analysis to
chromatographically determine nearly instantaneous (1 min)
exposure concentrations.
Keywords needle trap device, gas chromatography, flame ioniza-
tion detector, air sampling, ceiling, short term exposure
limit
Address correspondence to: Philip A. Smith, U.S. Department of
Labor, OSHA, Health Response Team, Salt Lake Technical Center,
8660 S. Sandy Parkway, Sandy, Utah 84070; e-mail: smith.philip.a@
dol.gov
INTRODUCTION
S everal types of exposure standards for air contaminants
have been promulgated by the U.S. Department of La-
bor, Occupational Safety and Health Administration (OSHA)
besides the frequently used eight-hour time-weighted average
(TWA) permissible exposure limit (PEL) values. These include
the short-term exposure limit (STEL), ceiling, and combina-
tion of ceiling/peak exposure standards that exist for a number
of air contaminants. As described in 29 CFR 1910.1000, “an
employee’s exposure to any substance in Table Z-1, the expo-
sure limit of which is preceded by a ‘C,’ shall at no time exceed
the exposure limit given for that substance. If instantaneous
monitoring is not feasible, then the ceiling shall be assessed as
a 15-minute time weighted average exposure which shall not
be exceeded at any time during the working day.”(1)
For a number of stressors listed in Table Z-2 of the OSHA
air contaminants standard, ceiling exposure standards are given
that should not be exceeded except in some cases for a specified
duration and not to exceed a higher stated contaminant concen-
tration (i.e., the “peak concentration”). For airborne stressors
with peak type exposure standards, the peak concentration
should never be exceeded, while exposure greater than a STEL
standard averaged over 15 min is also unacceptable.
Commonly Used Methods to Measure STEL
and Ceiling/Peak Exposures
Current methods for evaluating both STEL and ceiling/peak
exposure compliance typically rely on sorbent tube sampling.
This is followed by solvent desorption and analysis by gas

674 Journal of Occupational and Environmental Hygiene December 2013

TABLE I. Analytes Studied
Allowable STEL Allowable Peak Vapor Hazard Mass of analyte
Analyte Exposure (ppm) Exposure (ppm) Ratio (VHR) A injected (ng)B
Benzene 5C 50D 125,263 12H
Methylene Chloride 125E —– 22,895 326 I
Tetrachloroethylene —– 300F 245 101 J
Toluene —– 500G 187 94 K
AVapor hazard ratio (VHR) is defined per Popendorf (6) as VHR = Pvapor × 106/TLV × 760; relevant 8-hour OSHA permissible exposure limit standards used in
place of TLVR , Pvapor = vapor pressure, mm Hg at 25 ◦ C
BCalculated by considering sampling duration, volumetric sampling rate, and sample dilution from solvent desorption following current methods for STEL and

peak exposure determination using charcole tube sampling

C15-min STEL, 29 CFR 1910.1028(7)
DCeiling standard is 25 ppm, 50 ppm acceptable maximum peak above the acceptable ceiling concentration for an 8-h shift, 29 CFR 1910.1000(1) Table Z-2.

Applies to industry segments exempt from 1 ppm 8-h TWA PEL value found in 29 CFR 1910.1028(7)
E15-min STEL, 29 CFR 1910.1052(8)
FCeiling standard is 200 ppm, 300 ppm acceptable maximum peak above the acceptable ceiling concentration for an 8-h shift (5 min in any 3 h), 29 CFR

1910.1000(1) Table Z-2

GCeiling standard is 300 ppm, 500 ppm acceptable maximum peak above the acceptable ceiling concentration for an 8-h shift (10 min), 29 CFR 1910.1000(1)

Table Z-2
H15-min sample period (50 mL/min) at the STEL standard concentration, 1.0 mL solvent extraction and injection of 1.0 μL solvent extract for GC-FID analysis(2)
I15-min sample period (50 mL/min) at the STEL standard concentration, 1.0 mL solvent extraction and injection of 1.0 μL solvent extract for GC-FID analysis(3)
J1-min sample period (50 mL/min) at the peak standard concentration, 1.0 mL solvent extraction and injection of 1.0 μL solvent extract for GC-FID analysis(4)
K1-min sample period (50 mL/min) at the peak standard concentration, 1.0 mL solvent extraction and injection of 1.0 μL solvent extract for GC-FID analysis(5)

chromatography (GC). Benzene, methylene chloride, tetra-
chloroethylene, and toluene samples for STEL or ceiling/peak
exposure compliance are usually collected on coconut shell
charcoal (CSC) tubes to determine short-duration interval con-
centrations following analysis by GC with flame ionization
detection (GC-FID).(2–5) Solvent desorption requires the col-
lection of sufficient analyte to allow quantitative GC-FID
analysis. The respective masses collected over a 15-min pe-
riod at the STEL or over a 1 min period for peak exposure
measurement as applicable for each of the four analytes de-
scribed above are shown in Table I. Vapor hazard ratio values
calculated per Popendorf (6) are also provided, along with the
relevant OSHA exposure standards for these analytes.(1,7,8)
Possible Methods to Measure STEL
and Ceiling/Peak Exposures Without Solvent
Dilution
The pace of sampling method research using small sorbent
volumes has increased since the early 1990s when solid phase
microextraction (SPME) was first described.(9) “Microextrac-
tion is defined as an extraction technique where the volume
of the extracting phase is very small in relation to the volume
of the sample, and extraction of analytes is not exhaustive.
In many cases only a small fraction of the initial analyte is
extracted for analysis.”(10) The SPME technique is usually
employed for passive sampling, where analytes in the system
being sampled diffuse into a thin layer of polymer material
coated over a small fused silica fiber. Eventually, equilibrium
may be attained between the coating material and the system
being sampled where no further net analyte uptake occurs.
With SPME, the overall analyte concentration in the system
being sampled is not appreciably changed due to the small
scale of extraction.
Qualitative screening is an important practical use for SPME
due to the disadvantages associated with quantitative SPME
sampling: the need for gas phase calibration curves created
from sampling target analyte vapors under the same envi-
ronmental conditions as for field samples(11) and the poten-
tial for competitive displacement of relatively low affinity
volatile organic compounds from the sorbent surface when
adsorptive fiber coatings are used.(12) However, Augusto et al.
showed, for brief sampling periods, that SPME can be used
as a rapid quantitative field sampler without the need to create
vapor standards if a constant flow of air is passed across a
SPME fiber with an adsorptive coating, (13) and this makes it
a candidate of interest for STEL and ceiling/peak exposure
sampling.
Sampling with a needle trap device (NTD) sampler is not
a microextraction technique, but may be compared to a typi-
cal sorbent sampling approach—although on a much smaller
scale. If a few milligrams of sorbent material are immobilized
within a thin needle, analytes may be exhaustively extracted
from a finite (relatively small) volume of air passed through
the needle. The use of very small quantities of sorbent packed
in a needle allows introduction of an NTD sampler into a
standard GC injection port for thermal desorption of trapped
analytes, eliminating the need for solvent desorption. The
history and theory of NTD sampler development and exam-
ple applications have been reviewed recently by Lord and
Pawliszyn.(14)
Pawliszyn’s research group has demonstrated that NTD
samplers may be used for passive quantitative air sampling,(15)

Journal of Occupational and Environmental Hygiene December 2013 675

even over periods as long as 95 h for benzene, toluene, ethyl-
benzene, and xylenes (BTEX).(16) A diffusive sampling rate
of up to 16 μL air/min was described for NTD samplers used
this way to passively sample BTEX analytes.(16) In the case
of benzene at the STEL concentration of 5 ppm, a 1 mL
air sample actively drawn through an NTD sampler packed
with suitable sorbent would result in trapping of 16 ng of
benzene (assuming no breakthrough), easily detectable by
thermal desorption followed by GC-FID analysis. Passively
sampling at the rate described by Gong et al.(16) for 15 min
would result in trapping < 1 ng of benzene. The mass of
benzene that would be trapped with only a 1 mL active NTD
sample is similar to the mass of benzene injected for GC-FID
analysis using current 15 min CSC tube sampling and solvent
desorption methods for STEL exposure assessment (Table
I).(2) Thus, for short duration quantitative sampling of the
kind needed to determine STEL compliance for benzene (with
the lowest short interval exposure standard values among the
analytes listed in Table I), active NTD sampling will potentially
trap sufficient analyte for analysis, even with a much smaller
air sampling volume compared to current methods.
A volumetric field sample collected over a very brief dura-
tion is relatively easy to obtain rapidly using an NTD sampler
and a manually operated pump, and a breathing zone sampling
approach capable of sampling over a relevant period (15 min
in the case of the benzene and methylene chloride STEL
standards) could be developed if NTD-based sampling and
analysis methods show promise for collection of short duration
occupational exposure information. Laboratory analysis of
NTD samples by thermal desorption in the heated inlet of a
calibrated gas chromatograph would provide for analysis with
no sample preparation while also avoiding the use of solvents.
This latter point eliminates the expense and environmental
risks related to collection and disposal of contaminated solvent,
while also eliminating solvent background that may obscure
early eluting analytes.
The purpose of this study was to examine the usefulness
of active NTD sampling for GC-FID analysis to determine
airborne contaminant concentrations over short durations. The
analytes listed in Table I were selected for study as these are
amenable to trapping by a simple single-bed sorbent system
for analysis on a non-polar liquid film capillary GC
column.
MATERIALS AND METHODS
Chemicals
Methylene chloride (99.9%), benzene (99%), toluene
(99.8%), and ethylbenzene (99%) were purchased from Fisher
Scientific (Pittsburgh, Pa.). Tetrachloroethylene (99%) and
carbon disulfide (CS2 , 99.9%) were purchased from Acros
Organics (Geel, Belgium).
Needle Trap Device and Pump
Pumps and NTD samplers were purchased from Kitagawa
America (Pompton Lakes, N.J.—a distributor for Kitagawa,
676 Journal of Occupational and Environmental Hygiene
FIGURE 1. Kitagawa AP-20N Gas Sample Pump shown with
Shinwa NeedlEx NTD sampler. The pump handle (1) can be pulled
and locked to provide 10, 50, or 100 mL sample volumes. The NTD
sampler (2 ) restricts the flow rate to approximately 10 mL per min.
A red “pop-out” indicator (3 ) releases for visual confirmation that
sampling is complete. The overall length with needle trap sampler
attached is 34 cm. For flow-assisted backflushed desorption the
NTD sampler (2) is detached from the pump and is connected
at its Luer fitting to a flexible gas line with carrier gas flow
diverted through the NTD sampler in the heated injector of the
GC instrument used.
Japan, and Shinwa, Japan. The Kitagawa AP-20N hand pump
is designed to sample 10, 50, or 100 mL volumes of air through
a Shinwa NeedlEx NTD sampler at a rate of approximately
10 mL per min. For the proposed measurement of air con-
taminant concentrations relevant to STEL or peak exposure
standards, the AP-20N pump was operated using a 10-mL draw
volume through an organic vapor type NeedlEx NTD sampler.
The sorbent provided in the NeedlEx sampler is a copolymer
of methacrylic acid and ethylene glycol dimethacrylate (6 μL
volume).(17) Figure 1 shows the Kitagawa pump with NeedlEx
NTD sampler and Figure 2 details the arrangement of the Leur
fitting connection for pump and backflush desorption flow,
sorbent bed, and sample inlet for this NTD sampler.
Quantitative Standards
Vapor Standards for Determination of Breakthrough,
NTD Sampler Variability, Optimal Inlet Liner, Analysis
Linearity, and Sample Stability
Tedlar bags with a single polypropylene fitting (5 L volume,
SKC, Eighty Four, Pa.) were prepared with a mixture of the
FIGURE 2. Photo of Shinwa NeedlEx NTD sampler (Figure 2a)
and diagram of the NTD sampler components (Figure 2b, not to
scale): (1) Luer-Lok fitting for connection to pump or desorption gas
flow, (2 ) narrow needle shaft measuring 0.7 mm O.D. by 85 mm
length, (3 ) sorbent bed containing 6 μL of copolymer, and (4)
side-hole for sample entry and analyte desorption when the NTD
sampler is inserted into a heated GC injector.
December 2013
TABLE II. Quantitative Vapor Standards
Nominal Percent of STEL or Peak Air Contaminant
ConcentrationA
10 50 100 150
Benzene (ppm) 0.5 2.5 5 7.5
Methylene Chloride 12.5 62.5 125 187.5
(ppm)
Tetrachloroethylene 30 150 300 450
(ppm)
Toluene (ppm) 50 250 500 750 1000
Stock solution 2 μL 9 μL 19 μL 28 μL 38 μL
volume injected
into 5.0 L bagB
AAs applicable to each analyte
BStock solution mixture ratio by volume was 5 parts benzene; 87 parts
methylene chloride; 333 parts tetrachloroethylene; and 576 parts toluene.
four target analytes having nominal concentrations expressed
as a percentage of their respective STEL or peak exposure
standard concentrations as shown in Table II. Sweet et al.
previously demonstrated that tetrachloroethylene was found
to be stable in Tedlar bags in concentrations between 5 and
100 ppm for at least 12 h.(18) One day prior to each experiment,
bags to be used were flushed with dry nitrogen gas three times,
and were then filled with 5.0 L of dry nitrogen using a large
gastight syringe (1 L volume, Hamilton, Reno, Nev.). Bags that
were to contain non-humid vapor standards were not handled
further, while bags for humid standards each received 85 μL of
deionized water, with the expectation that a portion of the water
would adhere to the inside wall of the bag. The bags were then
left to equilibrate overnight at 23◦ C yielding 0% and 40–50%
relative humidity (0 mg/L and 9–10 mg/L absolute humidity,
respectively), as measured by length of stain indicator tubes
for water vapor (1–40 mg/L, Draäger, Luebeck, Germany).
One hour prior to sampling events, analyte mixture was
injected into each bag from a stock solution containing the four
target analytes. For multi-hour experiments where more than
one bag would be sampled, the addition of stock solution was
staggered in order to maintain between 1 and 2 h of contact
time between vapor standards and bag walls. No noticeable
decreases in analyte concentrations were observed using this
approach.
Each bag concentration was validated with triplicate CSC
tube samples analyzed offsite by an American Industrial Hy-
giene Association (AIHA)-accredited laboratory using OSHA
methods that included correction for extraction efficiencies.(2–5)
Unless otherwise stated, experiments utilized 50% nominal
relative humidity with 100% nominal STEL and peak vapor
standard concentrations, and 10 mL of air was sampled from
each bag through an NTD sampler.
Gas Chromatography Analysis
An Agilent 6890N GC-FID instrument was used in this
study, employing a resistively heated low thermal mass (LTM)
Journal of Occupational and Environmental Hygiene
DB-5 column assembly (Agilent Technologies, Wilmington,
Del.). The column had an internal diameter of 0.25 mm, with a
length of 30 m and 0.25 μm df . The GC injector was operated at
200◦ C, and used a 4-mm gooseneck liner and a 20:1 split ratio
200 for most analyses (exceptions described below). The column
was operated at a constant pressure of 21.9 psi providing an
10 average linear velocity of 38 cm/s at 50◦ C column temperature.
250 At the start of each run for the initial experiments, the LTM
column module was held at 50◦ C for 1.0 min followed by
600 a 30◦ C/min ramp to 110◦ C, at which point all four analytes
were observed to have eluted from the column. For calibration
using standards in solvent the initial 1.0 min hold at 50◦ C was
followed by a 30◦ C/min ramp to 120◦ C to ensure elution of the
ethylbenzene internal standard. In both cases the final portion
of a run involved a 300◦ C/min ramp to a final temperature of
250◦ C with a 0.5 min hold time.
After collection of an NTD sample the device was dis-
connected from the hand-operated pump shown in Figure 1.
It was then attached at its Luer type fitting (see Figures 1
and 2) to a flexible small-diameter gas line connected to one
of two automated external valves that were operated by the
GC instrument control software. Carrier gas flow could be
diverted through the valve connected to the Luer fitting with
the second valve (normally supplying the GC injector) closed.
Upon introduction of an NTD sample into the heated injector,
carrier gas was diverted through the NTD sampler for 60 s.
This arrangement provided for backflushed desorption with a
total desorption gas volume of 21 mL (1.0 mL/min through
the GC column, and 20 mL/min split flow) by maintaining the
constant carrier gas pressure that provided injector carrier gas
flow through the NTD sampler during the 60 s period of carrier
gas diversion.
At the conclusion of this desorption time, the GC instrument
control software cycled both external valves so that carrier gas
flow to the NTD sampler was stopped and reopened to the
carrier gas supply line found on the injector for the remainder
of a GC run. The 1.0 min thermal desorption eliminated
carryover and negated the need for NTD sampler conditioning
between samples as confirmed by a lack of response from
blank analyses run at regular intervals with NTD samplers
desorbed using this approach.
Needle Trap Device Performance
Breakthrough
To study breakthrough, two NTD samplers were connected
in series, and a 10 mL sample was collected and analyzed for
each of the five STEL or peak exposure standard concentra-
tions listed in Table II. Increasingly larger volumes were sam-
pled in 10 mL increments from a 100% nominal STEL/peak
standard concentration bag. As the Kitagawa AP-20N pump
only samples with 10, 50, or 100 mL volumes, repeated 10 mL
draws through the front NTD sampler were taken without
analysis while the back NTD sampler was analyzed after each
10 mL draw. This allowed separate breakthrough analyses of
the back NTD sampler at cumulative volumes through the front
NTD sampler of 10, 20, 30, 40, and 50 mL.
December 2013 677
NTD Sampler Variability
Variability was tested using nine NTD samplers with trip-
licate analyses for each from a 100% nominal concentration
standard bag (Table II). Run-to-run variability on a single nee-
dle was calculated individually for each NTD sampler using
the triplicate relative standard deviation (RSD) values for GC-
FID peak areas. Needle-to-needle variability was calculated
from the overall RSD of the triplicate analyses for all nine
NTD samplers (n = 27 analyses).
Additionally, variability was compared between NTD sam-
plers conditioned 10 days prior to use and freshly conditioned
NTD samplers. Nine NTD samplers were conditioned and
capped for 10 days with room temperature storage (23◦ C). Af-
ter the 10-day storage, the nine NTD samplers were uncapped
and used to collect 10 mL samples from a 100% nominal
STEL/peak standard concentration bag. Analysis/conditioning
of each NTD sampler was immediately followed by a sec-
ond 10 mL sample collection from the same 100% nominal
STEL/peak standard concentration bag. The second analysis
provided data for freshly conditioned NTD samplers to com-
pare to the samples taken after storing a sampler for 10 days
following conditioning. Since the data were paired for each of
the nine NTD samplers, two-tailed paired t-tests were used to
compare accuracy. Precision was compared using the resulting
RSD values.
GC Liner and Sample Humidity Effect on
Chromatography
Using a single NTD sampler, four commercially available
GC injector liner designs were evaluated for quantitative run-
to-run variability, qualitative chromatography aspects, and hu-
midity effects of water vapor expansion in the heated injector.
The evaluated liners included 2 mm straight, 2 mm gooseneck,
4 mm straight, and 4 mm gooseneck types. After each liner was
installed, two sets of triplicate samples were collected from
100% nominal STEL/peak standard concentration bags—one
triplicate sample set from a non-humid vapor standard bag and
the other triplicate sample set from a humid vapor standard bag.
Quantitative run-to-run variability was evaluated using RSD
comparisons. Chromatography quality and humidity effects
were evaluated visually with overlaid chromatograms.
Linearity
The linearity for NTD sampling was evaluated for the five
vapor standard concentrations of each analyte listed in Table
II. Triplicate NTD samples were drawn at each concentration
relative to the STEL and peak exposure standards. This yielded
15 data points for each target analyte. Each target analyte set
was then plotted separately and evaluated for linear regression
fit using least squares R2 values.
Stability Time-Course
Two 14-day stability time-course studies were conducted
on NTD samples drawn from a 100% nominal STEL/peak
standard concentration bag. One set of NTD samples was
stored refrigerated (4◦ C) and the other set of samples was
678 Journal of Occupational and Environmental Hygiene
stored at room temperature (23◦ C). Nine NTD samplers were
used for each time-course, with all nine samplers analyzed
on day 0 followed immediately by a second sample drawn
for time-course storage. At day 1 following sampling: NTD
samplers 1, 2, and 3 were analyzed for comparison to the values
obtained from analysis of the same NTD samplers on day 0.
This was repeated for NTD samplers 4, 5, and 6 analyzed again
on the 7th day of storage, and for NTD samplers 7, 8, and 9
analyzed on the 14th day of storage.
Calibration Using Standards in Solvent
After examining breakthrough, NTD sampler variability,
optimal inlet liner, analysis linearity, and sample stability for
the four analytes listed in Table II, work was completed to
calibrate the observed GC-FID responses for NTD sampling
of benzene, tetrachloroethylene, and toluene. Several variables
affected the slope of peak areas plotted against the mass of
analyte trapped on an NTD sampler relative to injection of
standards dissolved in CS2 , including injector temperature and
the depth of NTD insertion in the heated injector.
Assuming that injector discrimination was the same for all
analytes and an internal standard (ethylbenzene), whether from
an NTD sampler or a solvent injection, quantitation of NTD
analyses was possible without creating gas phase standards.
To accomplish this, injections were completed with the target
analytes and internal standard in CS2 at nine concentrations
to exceed the range of expected analyte recovery masses, and
the results were used to determine a response ratio (R) for
each analyte relative to the internal standard. These solvent
injection calibration curves were also used to determine the
concentration of analytes within dry Tedlar bag samples by on-
site analysis of desorption solvent obtained from CSC sorbent
tube samples.(2,4,5) Gas from the same dry Tedlar bag samples
that also contained a known internal standard concentration
(1.73 μg of this compound collected with a 10 mL NTD sample
volume) was pulled through an NTD sampler, and the values
of R for each analyte were used with Equation 1 to determine
the amount of each analyte trapped on an NTD sampler.
AX AI S
=R× (1)
MX MI S
In Equation 1, AX is the integrated peak area for analyte X,
AIS is the integrated peak area for the internal standard, MIS is
the known mass of internal standard desorbed from the NTD
sampler, and MX is the unknown mass of analyte X desorbed
from the NTD sampler.
RESULTS
Breakthrough
No breakthrough of detectable analyte on the backup NTD
sampler was observed in any of the 10 mL samples collected
and analyzed from 10% to 200% nominal STEL/peak standard
concentration bags. In sample volumes greater than 10 mL
from the 100% nominal STEL/peak standard concentration
bag, breakthrough was observed for methylene chloride at
December 2013

FIGURE 3. Overlaid GC-FID chromatograms resulting from
analysis of a second NTD sampler placed in series to capture
analyte breaking through a front NTD sampler used to sample
a humid 100% nominal STEL/peak standard concentration bag:
125 ppm methylene chloride (peak 1); 5 ppm benzene (peak 2,
peak not present until 50 mL sample volume -not shown here);
500 ppm toluene (peak 3 ); and 300 ppm tetrachloroethylene
(peak 4, peak not present until 50 mL sample volume - not
shown here). All chromatograms are magnified 10x for peak 3 .
The lightly dotted 10 mL sample volume chromatogram shows
no FID response for any of the analytes. The solid 20 mL
sample volume chromatogram shows a slight FID response for
methylene chloride, indicating the beginning of breakthrough.
The dashed 30 mL sample volume chromatogram shows more
pronounced breakthrough of methylene chloride and slight toluene
breakthrough. The dotted 40 mL sample volume chromatogram
shows increased breakthrough of methylene chloride and toluene.
20 mL sample volume, then for toluene at 30 mL sample
volume, and finally for all four analytes at 50 mL sample
volume. Although toluene was sampled at its 500 ppm peak
concentration and methylene chloride was sampled at its lower
125 ppm STEL, it was not surprising to observe methylene
chloride as the first analyte to break through as other estab-
lished sampling methods acknowledge difficulties in capturing
this analyte. The documentation for OSHA Method 59 calls for
collection of methylene chloride through three CSC sections
(350 mg in each section) in series to avoid breakthrough
loss. The documentation for this method also states that high
(80%) humidity increases methylene chloride breakthrough in
charcoal tubes.(3) Figure 3 chromatograms show breakthrough
study results for sample volumes ranging from 10 to 40 mL.
NTD Sampler Variability
In preliminary studies, the use of a single NeedlEx NTD
sampler produced highly repeatable results, whereas sampling
conducted with a different NeedlEx NTD for each subsequent
sample collection and analysis was found to produce results
that were slightly less reproducible. Run-to-run variability (as
Journal of Occupational and Environmental Hygiene
RSD values of GC-FID peak areas) on a single needle ranged
from 1 to 5%, while needle-to-needle variability was 5% for
three of the analytes and 10% for methylene chloride. Figure 4
shows a graphical representation for these results with run-
to-run variability of each needle represented by error bars
showing the RSD, and needle-to-needle variability represented
by the y-axis showing differences for each NTD sampler
from the overall mean. The bar chart is useful for quick
visual evaluation of individual NTD sampler performance in
comparison to the others but the RSD values mentioned above
provide the best overall needle-to-needle statistical evaluation.
For example, comparing methylene chloride results from
triplicate samples on NTD sampler #2 to triplicate samples on
NTD sampler #8 yields the largest difference for needle-to-
needle comparison (+16% to −11% compared to the overall
mean) but the overall methylene chloride needle-to-needle
RSD was 10%. Since run-to-run variability on a single nee-
dle was lower than needle-to-needle variability, it may be
prudent to pull a very small volume (e.g., 1 mL) from a
stable, concentrated vapor of an analyte that would serve as
an internal standard prior to sampling. Field samples were
not analyzed in the scope of this preliminary study, but in
future studies post-analysis correction of some type would
be warranted due in part to the observed needle-to-needle
variability.
Addressing the question related to post-conditioning stor-
age prior to sample collection, results for experiments to detect
changes in 10-day stored vs. fresh NTD samplers are shown
in Figure 5. The t-tests failed to show a statistical difference
between the means of GC-FID peak areas for 10-day stored and
fresh NTD samplers at an alpha value of 0.05. Additionally,
the variabilities were similar for both conditioning approaches
(4–6% RSD for three analytes and 9–11% RSD for methylene
chloride) as shown by the error bars in Figure 5.
GC Inlet Liner and Sample Humidity Effects on
Chromatography
During the early stages of the work described in this article,
a 2-mm straight liner was used in the GC-FID instrument
because the volume of a 0.75-mm liner designed for SPME
was thought to be too small to allow for analyte thermal
expansion from the relatively high capacity NTD sampler,
and a 4-mm liner designed for liquid injection was thought to
be too large for rapid analyte transfer to the column during
thermal desorption. NeedlEx NTD sampler instructions call
for a 10 s delay between insertion of the NTD sampler into
the heated GC injector and pushing an inert gas through the
sampler to achieve flow assisted thermal desorption.(19)
However, the four target analytes all exhibited peak dou-
bling with this approach, which was thought to be caused
by an initial thermal expansion peak followed by a second
flow-assisted peak. To eliminate the extra peak, the carrier gas
bypass valve described previously was used to allow simulta-
neous NTD sampler insertion and gas flow across the heated
NTD sorbent bed. Chromatography of the four target analytes
December 2013 679
 FIGURE 4. Results from the use of 9 separate NTD samplers with triplicate samples for each drawn from a humid 100% STEL/peak standard
concentration bag: methylene chloride (white bars); benzene (light gray bars); toluene (gray bars); and tetrachloroethylene (dark gray bars).
The X-axis shows the 4 separate analytes on each of the 9 individual NTD samplers. The Y-axis represents the needle-to-needle GC-FID peak
area variability by showing the difference of each triplicate NTD sampler average for each analyte from the overall average for all 27 analysis
events. The error bars show run-to-run GC-FID peak area variability (triplicate RSD) for a single NTD sampler.

improved so that the peak doubling was only observed at the

apex of peaks. In order to eliminate the double apex, the GC
injector split ratio was increased from 1:10 to 1:20 and column
flow was decreased from 2.0 mL per min to 1.0 mL per min
in order to maintain a nearly identical flow rate through the
NTD sampler during the period of carrier gas flow through
the heated NTD sampler. This change lessened the apex peak
doubling, leaving only a fronting peak shoulder as shown in
Figure 6a.
However, analysis of humid samples resulted in a slight
increase in the fronting shoulder (e.g., the toluene peak in
Figure 6a). The use of a 4-mm straight liner eliminated fronting
peak shoulders but GC peak widths increased, apparently due
to slower analyte transfer caused by the larger volume of this
liner. Figure 6b shows that the 2-mm gooseneck liner improved
the chromatography peak sharpness and reduced the fronting
peak shoulder associated with inefficient analyte transfer that
was observed with the 2-mm straight liner. Figure 6c shows
that the 4-mm gooseneck liner completely eliminated fronting
peak shoulders due to slow analyte transfer from the NTD FIGURE 5. Effect of conditioning NTD samplers immediately
sorbent, with minimal loss of peak sharpness. The RSD for prior to use, and use following ten-day post-conditioning storage.
replicate samples of the four target analytes decreased as well GC-FID peak areas for NTD samplers used immediately following
conditioning were compared with peak areas for the same
from 4–7% RSD for the 2 mm liners to 1–2% RSD using the
samplers used following ten days of post-conditioning storage.
4 mm gooseneck liner.

680 Journal of Occupational and Environmental Hygiene December 2013

 FIGURE 6. GC injector liner and humidity effects. Overlaid GC-FID chromatograms resulting from 10 mL NTD samples drawn from both
humid and non-humid 100% STEL/peak standard concentration bags: methylene chloride (peak 1); benzene (peak 2 ); toluene (peak 3 ); and
tetrachloroethylene (peak 4 ). A 6x magnified view for peaks 1 and 2 is provided for each set of overlaid chromatograms. Figure 6a shows a
slight humidity effect (larger fronting peak shoulders) in a chromatogram comparison of humid and dry samples introduced into a 2-mm straight
liner (RSD values ranging from 3.5 to 5.9%). Figure 6b shows reduced fronting peak shoulder size and humidity effect in a comparison of humid
and dry samples introduced into a 2-mm gooseneck liner (RSD values ranging from 5.4 to 6.7%). Figure 6c shows elimination of fronting peak
shoulders and humidity effect in a comparison of humid and dry samples introduced into a 4-mm gooseneck liner (RSD values ranging from 0.3
to 2.2%). The GC peak fronting observed for toluene and tetrachloroethylene are related to the relatively large concentrations of these analytes
(e.g., the toluene peak exposure standard concentration is two orders of magnitude higher than the benzene STEL concentration).

The fronting observed for the toluene and tetrachloroethy-
lene peaks indicates that the column capacity was exceeded
for these analytes, and this is due to the relatively high con-
centrations needed to match the respective allowable peak
concentration values for these two analytes. The use of an NTD
sampler to measure the wide range of concentrations studied
(5 to 500 ppm, Table II) in a workplace would either require
predetermined analyte-specific sample volumes, or different
injector split ratios and/or a GC column with thicker stationary
phase film to improve the chromatography. This is important
to avoid co-elution of potential interferents, and these details
should be optimized during robust method development.
for all four target analytes, with each of the triplicate analysis
RSD values below 4%. Results from laboratory analysis of
CSC tube samples from Tedlar bags showed excellent linearity
when plotted against expected analyte concentrations (R2 <
0.99). These samples for the Tedlar bag analyte concentration
estimates ranged from 18.0 to 25.5% lower than expected
values for unknown reasons. It is unlikely that this deviation
from theoretical in the measured concentration was due to
incomplete analyte desorption during laboratory analysis, as
the analytical results reported consider the known CS2 /CSC
desorption efficiencies which are updated on a regular basis
by the laboratory.

Linearity Stability Time-Course

The use of a single NTD sampler with the five vapor stan- No statistically significant difference was observed between
dard bag concentrations yielded R2 values greater than 0.99 samples stored at 4◦ C and those stored at 23◦ C over a period

Journal of Occupational and Environmental Hygiene December 2013 681


FIGURE 7. Storage stability of the four target analytes over
14 days at 4◦ C and 23◦ C yielded the following respective average
recoveries (as GC-FID peak area responses) at 14 days: 80.1
and 86.0% for methylene chloride (Figure 8a); 97.9 and 89.9% for
benzene (Figure 8b); 96.1 and 88.1% for toluene (Figure 8c); and
97.3 and 87.5% for tetrachloroethylene (Figure 8d). At 7 days,
recoveries were ≥95.0% for all analytes with the exception of
methylene chloride, which was recovered at 91.7% compared to
samples analyzed on day 0. Two-tailed paired t-tests (alpha =
0.05) failed to show a statistical difference between 4◦ C and 23◦ C
storage for the four analytes.
of 14 days. At room temperature, GC-FID peak areas for NTD
samplers were 92.9% or greater for the sampled target analytes
with storage for 7 days but fell to 86.0% or greater with storage
of 14 days. Results are shown graphically in Figure 7.
TABLE III. Results of Quantitative NTD Sampling with Internal Standard Correction for Calibration Using
Standards Dissolved in CS2 (see Figure 8)
Concentration Range
Benzene 5–100 ppm
Toluene 50–1100 ppm
Tetrachloroethylene 30–750 ppm
682 Journal of Occupational and Environmental Hygiene
FIGURE 8. Results of quantitative NTD sampling with internal
standard correction for calibration using standards dissolved in
CS2 (See Table III).
Calibration Using Liquid Standards
The ability to calibrate the GC-FID instrument response for
NTD samples using standards dissolved in CS2 and internal
standard correction is demonstrated by Figure 8 where NTD
results are plotted against CSC results for the same bags. With-
out internal standard correction, plots for the same analytes
produced regression lines with good linearity, but with slope
values far from unity. This was due to injector discrimination
between the two types of sample introduction approaches, and
a possible explanation is expansion of the CS2 solvent delivery
vehicle in the heated injector while this does not occur to the
same degree for the NTD samples. The slopes of the regression
lines for the three analytes with internal standard correction
shown in Figure 8 are close to 1, with high correlation between
the NTD and CSC results (see Table III).
DISCUSSION
T he primary advantage to the use of an NTD sampler for
collection of a nearly instantaneous sample is the ability
to potentially desorb all of the trapped analyte into a GC instru-
ment, while traditional methods that rely on solvent desorption
require longer sampling times that would not be considered to
R2 Slope
0.998 0.94
0.997 1.07
0.997 0.99
December 2013
be nearly instantaneous. For example, OSHA Method 111 for
toluene recommends a sampling time >10 min for diffusive
sampling to detect exposures above the 300 ppm ceiling value.
While the use of diffusive samplers for periods as short as 1 min
was shown to be possible for determination of the peak toluene
concentration (500 ppm) under carefully controlled conditions,
as this type of device begins to sample immediately upon
removal from its sealed container the potential for significant
sampling error exists for such short duration sampling periods
and the “shortest recommended sampling time for toluene
using diffusive samplers was determined to be 10 minutes.”(5)
The volumetric nature of NTD sampling, its speed, and
the elimination of solvent use for desorption deal with these
problems. While rapid volumetric sampling across sorbent
packed within a traditional thermal desorption tube would pro-
vide the same benefits, dedicated thermal desorption hardware
would be required for analysis. The relatively high flow from
the desorption of a large (slowly heated) tube with internal
diameter of for example, 4 mm, must be concentrated onto
a smaller secondary trap that may be rapidly heated, with
a much lower volumetric flow to closely match that of a
modern capillary GC column typically having a diameter
≤0.32 mm.(20)
The use of a hand pump or vacuum reservoir and a constant
restricted flow for NTD sampling could simplify short duration
air contaminant collection in the field by eliminating the need
for calibration of a battery operated sampling pump. In addition
to the potential use of NTD sampling for target analytes
with OSHA STEL or ceiling/peak exposure standards, the
approach may also be of interest to those wishing to compare
worker exposures to TLV values published by the American
R
Conference of Governmental Industrial Hygienists (ACGIH). R
Ceiling or STEL TLV values have been promulgated for more
R
than 150 chemical substances.(21)
Although rapid collection of ceiling or peak exposure sam-
ples is advantageous, a problem exists with the method de-
scribed here for measurement of STEL values, as a sample
averaging time of 15 min is needed for these measurements.
An approach to remedy this while avoiding breakthrough could
entail the use of a critical orifice to allow for 15 min sample
collection at a constant flow rate, with a total air sample volume
of 10 mL or less. Also, the ceiling/peak and STEL exposure
standards apply to personal breathing zone exposures, and
modifications to the sampling system would be needed to
allow for mounting of a combined NTD sampler and vacuum
reservoir for worker breathing zone sampling, with subsequent
activation of sampling at a time when a maximum exposure is
expected.
Another limitation of the methods described here is NTD
sampler cost relative to traditional sampling media and the
current lack of validated methods. However, reusability of
NTD samplers (estimated 25–30 uses)(19) along with solvent-
less laboratory sample desorption are likely to mitigate the
cost issue to some degree. Another limitation is related to
the side-hole sample entry/exit port (Figure 2) that tends to
damage a typical injector septum during insertion or removal,
Journal of Occupational and Environmental Hygiene
resulting in faster than normal septum degradation and occa-
sionally clogging of the NTD sampler entry/exit port. Addi-
tionally, the recommended injector temperature limit of 200◦ C
for thermal desorption of the copolymer sorbent bed could
limit the usefulness of the method for larger, less volatile
compounds.
However, in preliminary work we observed that tributyl
phosphate (a relatively much less volatile analyte compared
to those listed in Table I) was successfully desorbed at 200◦ C
without observable carryover from the copolymer sorbent bed
using 60 s NTD flow assisted injection. No evidence was seen
for any degradation of the sorbent in the NTD samplers used
(e.g., spurious GC peaks, loss of sampling capacity, increased
GC baseline) during this study with injections carried out at the
recommended temperature of 200◦ C. If needed, other sorbent
bed materials that may be heated to temperatures greater than
300◦ C are available, and further work to validate the approach
discussed here for very rapid sampling of STEL or ceiling/peak
concentrations should occur after considering the wide range
of sorbents that were not available in commercially available
NTD samplers at the time this work was completed.
The need for internal standard application to the NTD
sampler introduces additional problems, although this could
be solved by pulling a very small precisely metered volume
of a constant-concentration vapor standard through the NTD
sampler in the field prior to sampling. With knowledge of the
values of R for target analytes relative to the internal standard
obtained from the analysis of standards dissolved in solvent,
the calculation of target analyte air concentration could be
completed.
Lastly, a substantial limitation of this NTD sampling and
analysis method is the need for a carrier gas bypass valve
to be retrofitted to a GC instrument to avoid split GC peaks
observed for the light analytes studied here when the bypass
valve was not used. This is a minor modification compared to
the addition of an entire thermal desorption module that would
be necessary to use traditional thermal desorption tubes, and
the issue of split peaks would be a lesser problem (if at all)
for semivolatile analytes trapped on an NTD sampler. For
analytes of this type, nearly instantaneous thermal desorption
is less important compared to light solvent analytes as the less
volatile analytes tend to focus at the proximal GC column
end where the column interfaces with a heated injector, as
temperature programmed GC can employ a relatively cool
initial column temperature (e.g., 40◦ C).
Another approach that has shown promise in simplifying
NTD injection in a standard GC instrument injector is the use
of a side-hole located above the sorbent packing in a NTD
sampler. Sampling is through a hole in the tip of the NTD,
identical to the NeedlEx NTD samplers used in this work. This
is accomplished with vacuum applied to the side-hole above
the sorbent. For analysis, the tip of the NTD bottoms out in an
injector liner designed to seal the tip, and the pressurized and
heated injector gas in the liner then pushes down through the
side-hole, through the sorbent, and out the tip of the NTD for
backflushed thermal desorption of trapped analytes.(14)
December 2013 683
CONCLUSION
A t the time of this preliminary study, the NTD samplers
used were obtainable from a commercial source, but
NTD samplers with the wide range of well-known sorbents
used for thermal desorption tube sampling were not. Further
validation studies are needed for NTD sampling, and these
should examine carefully selected sorbents that are already
available for use in traditional thermal desorption tubes, -
possibly using a multi-bed design. Robust testing with op-
timized sorbent types should include dynamic standard gen-
eration with relative humidity levels ≥ 80%. This preliminary
study demonstrated the fundamental ability to vary rapidly
sample target chemical vapors using an NTD sampler for
analytes with both STEL and ceiling/peak type exposure stan-
dards. The GC-FID results obtained were highly repeatable
and linear from 10% to 200% of the nominal respective OSHA
STEL or peak exposure standard concentrations for benzene,
methylene chloride, toluene, and tetrachloroethylene. Analytes
remained stable after collection on the NTD sampler for at
least 7 days at room temperature. A short duration expo-
sure assessment approach where nearly instantaneous analyte
concentration information is desired could potentially benefit
from NTD sampling and its associated low-effort, solventless
sample preparation/introduction method without the purchase
and maintenance of dedicated thermal desorption equipment.
ACKNOWLEDGMENTS
T his work received financial support from the U.S. Army
Center for Environmental Health Research and material
support from the U.S. Department of Labor, OSHA Salt Lake
Technical Center. The views presented are those of the authors,
and do not represent those of their respective organizations.
This document is not a standard or regulation. It creates no
new legal obligations and alters no existing obligations created
by OSHA standards or the Occupational Safety and Health
Act. It may contain recommendations that are advisory in
nature, informational in content, and are intended to support
the provision of safe and healthful workplaces. This article is
not subject to US copyright law.
REFERENCES
1. “Air Contaminants,” Code of Federal Regulations Title 29, Part
1910.1000. 2013.
2. “Benzene: Method 1005: OSHA Sampling and Analytical Meth-
ods.” Available at http://www.osha.gov/dts/sltc/methods/validated/
1005/1005.html (accessed May 1, 2012).
684 Journal of Occupational and Environmental Hygiene
3. “Methylene Chloride: Method 59: OSHA Sampling and Analytical
Methods.” Available at http://www.osha.gov/dts/sltc/methods/organic/
org059/org059.html (accessed May 1, 2012).
4. “Tetrachloroethylene Trichloroethylene: Method 1001. OSHA Sam-
pling and Analytical Methods.” Available at http://www.osha.gov/dts/sltc/
methods/mdt/mdt1001/1001.html (accessed May 1, 2012).
5. “Toluene: Method 111: OSHA Sampling and Analytical
Methods.” Available at http://www.osha.gov/dts/sltc/methods/organic/
org111/org111.html (accessed May 1, 2012).
6. Popendorf, W.: Source control via substitution. In Industrial Hygiene
Control of Airborne Chemical Hazards. Boca Raton, Fla: CRC/Taylor &
Francis, 2006. pp. 185–214.
7. “Benzene,” Code of Federal Regulations Title 29, Part 1910.1028, 2013.
8. “Methylene Chloride,” Code of Federal Regulations Title 29, Part
1910.1052, 2013.
9. Arthur, C.L., and J. Pawliszyn: Solid-phase microextraction with
thermal-desorption using fused-silica optical fibers. Anal. Chem.
62:2145–2148 (1990).
10. Lord, H., and J. Pawliszyn: Microextraction of drugs. J. Chromatogr. A
902:17–63 (2000).
11. Bocchini, P., D. Dello Monaco, R. Pozzi, et al.: Solid-phase microex-
traction coupled to gas chromatography with flame ionization detection
for monitoring of organic solvents in working areas. Microchim. Acta
165:271–278 (2009).
12. Mocho, P., V. Larroque, and V. Desauziers: Modelling of toluene solid-
phase microextraction for indoor air sampling. Anal. Bioanal. Chem.
388:147–156 (2007).
13. Augusto, F., J. Koziel, and J. Pawliszyn: Design and validation of
portable SPME devices for rapid field air sampling and diffusion-based
calibration. Anal. Chem. 73:481–486 (2001).
14. Lord, H.L., W. Zhan, and J. Pawliszyn: Fundamentals and applications
of needle trap devices: A critical review. Anal. Chim. Acta 677:3–18
(2010).
15. Cai, J., G. Ouyang, Y. Gong, and J. Pawliszyn: Simultaneous sampling
and analysis for vapor mercury in ambient air using needle trap
coupled with gas chromatography-mass spectrometry. J. Chromatogr. A
1213:19–24 (2008).
16. Gong, Y., I.-Y. Eom, D.-W. Lou, D. Hein, and J. Pawliszyn:
Development and application of a needle trap device for time-
weighted average diffusive sampling. Anal. Chem. 80:7275–7282
(2008).
17. Saito, Y., I. Ueta, K. Kotera, et al.: In-needle extraction device designed
for gas chromatographic analysis of volatile organic compounds. J.
Chromatogr. A 1106:190–195 (2006).
18. Sweet, N.D., G.E. Burroughs, L. Ewers, and G. Talaska:
A field method for near real-time analysis of perchloroethy-
lene in end-exhaled breath. J. Occup. Environ. Hyg. 1:515–520
(2004).
19. “NeedlEx Instruction Manual.” Available at http://www.airmet.com.au/
admin/Products/DownloadProductFile.aspx?FileTitle=Manual&uniquen
ame=NeedlEx manual.pdf (accessed May 4, 2012).
20. Woolfenden, E.: Sorbent-based methods for volatile and semi-volatile
organic compounds in air. Part 1: Sorbent-based air monitoring options.
J. Chromatogr. A 1217:2674–2684 (2010).
21. American Conference of Governmental Industrial Hygienists
(ACGIHR ): 2012 Threshold Limit Values for Chemical Substances
and Physical Agents and Biological Exposure Limits. Cincinnati, Ohio:
ACGIH Worldwide, 2012.
December 2013