You are on page 1of 11

Anal Bioanal Chem (2007) 388:837–847

DOI 10.1007/s00216-007-1194-2

ORIGINAL PAPER

Determination of total mercury and methylmercury


in biological samples by photochemical vapor generation
Mariana A. Vieira & Anderson S. Ribeiro &
Adilson J. Curtius & Ralph E. Sturgeon

Received: 16 November 2006 / Revised: 6 February 2007 / Accepted: 8 February 2007 / Published online: 27 February 2007
# Springer-Verlag 2007

Abstract Cold vapor atomic absorption spectrometry (CV- cases. Limits of detection of 6 ng/g for total mercury using
AAS) based on photochemical reduction by exposure to formic acid, 8 ng/g for total mercury and 10 ng/g for
UV radiation is described for the determination of methyl- methylmercury using TMAH were obtained. The proposed
mercury and total mercury in biological samples. Two methodology is sensitive, simple and inexpensive, and
approaches were investigated: (a) tissues were digested in promotes “green” chemistry. The potential for application
either formic acid or tetramethylammonium hydroxide to other sample types and analytes is evident.
(TMAH), and total mercury was determined following
reduction of both species by exposure of the solution to UV Keywords Mercury . Methylmercury . Speciation .
irradiation; (b) tissues were solubilized in TMAH, diluted Photochemical reduction . Vapor generation . Formic acid .
to a final concentration of 0.125% m/v TMAH by addition Tetramethylammonium hydroxide
of 10% v/v acetic acid and CH3Hg+ was selectively
quantitated, or the initial digests were diluted to 0.125% m/v
TMAH by addition of deionized water, adjusted to pH 0.3 Introduction
by addition of HCl and CH3Hg+ was selectively quantitated.
For each case, the optimum conditions for photochemical Mercury (Hg) is one of the most hazardous of environmental
vapor generation (photo-CVG) were investigated. The pollutants, having significant routes of ingestion to humans
photochemical reduction efficiency was estimated to be through consumption of marine foodstuffs or via industrial,
∼95% by comparing the response with traditional SnCl2 agricultural and laboratory exposure. Its toxicity depends on
chemical reduction. The method was validated by analysis of the chemical form, which predominantly includes elemental
several biological Certified Reference Materials, DORM-1, mercury (Hg0), inorganic mercury (Hg2+) and methyl-
DORM-2, DOLT-2 and DOLT-3, using calibration against mercury (CH3Hg+). The accumulation of mercury in fish
aqueous solutions of Hg2+; results showed good agreement tissues is usually as methylmercury, which is more toxic than
with the certified values for total and methylmercury in all inorganic mercury [1].
Interest in, and the need for, determination of Hg in
samples such as water, soil, sediments and biological
materials has led to significant progress in the development
of analytical measurement techniques. The most popular
M. A. Vieira : A. S. Ribeiro : A. J. Curtius approach remains cold vapor generation, wherein the mercury
Departamento de Química,
Universidade Federal de Santa Catarina,
species are typically mineralized to Hg2+ through acidic
Florianópolis, SC 88040 900, Brazil attack and then reduced to the element by reaction with
SnCl2 or NaBH4 [2, 3]. A variety of detection techniques is
R. E. Sturgeon (*) utilized, most frequently including long-path quartz tube
Institute for National Measurement Standards,
National Research Council Canada,
atomic absorption spectrometry [4, 5], atomic fluorescence
Ottawa, ON K1A 0R9, Canada spectrometry [6, 7] and in situ collection of the cold vapor
e-mail: ralph.sturgeon@nrc-cnrc.gc.ca coupled to an electrothermal atomizer for AAS [8, 9].
838 Anal Bioanal Chem (2007) 388:837–847

An alternative to chemical reduction is the use of with tetramethylammonium hydroxide (TMAH) has been
photoreduction, based on the exposure of a sample to a studied, providing very simple sample preparation proce-
UV field. Early work on the photoassisted generation of dures [32, 33]. Torres et al. [34] developed a CV-AAS
volatile species focused on the production of mercury in the technique for the speciation of mercury following digestion
aquatic environment. Thus, Akagi et al. [10–12] noted the of biotissues with TMAH. CVG using NaBH4 permitted
photochemical alkylation of mercury induced by the action detection of total mercury with a heated quartz tube
of sunlight in the presence of several low molecular weight atomizer (QTA) and inorganic mercury with the QTA at
organic acids (acetic and propionic). These early findings room temperature. Formic acid is also an efficient tissue
are today supported by additional observations on the solubilizer [35] and has been successfully used for the
release of mercury from intertidal sediments [13] as well as determination of mercury by CV-AAS [36].
from the aerobic surface waters of lakes [14] in direct Based on the efficiency with which biological tissues can
response to sunlight and the presence of dissolved organic be solubilized with either formic acid or TMAH, coupled
matter of relatively low molecular weight (<5 kDa). with the known capabilities of photo-CVG in formic acid
The action of UV light on dissolved organic and inorganic media for the reduction of mercury, this study was
compounds results in the formation of many intermediate undertaken to evaluate such a combined procedure for the
products [15]. UV irradiation of aqueous systems can speciation of mercury in tissue samples. Although TMAH
produce OH• and O• radicals, and in the presence of organic has never been examined as a suitable medium for photo-
compounds additional radicals such as CO• and R• arise that CVG, this reagent was found to be effective for the direct
can promote photoreduction of other species [15, 16]. Guo et determination of not only total mercury, but both inorganic
al. [17] were the first to report on the analytical use of photo- and methylmercury by performing individual measurements
CVG using UV radiation for vapor generation of Se(IV) and under optimized conditions.
they identified the products of the reaction arising in a
formic acid solution as both SeCO and SeH2. Subsequent
studies by these authors of Se [18], As [19, 20] and Ni [21] Experimental
as well as a host of other elements [22, 23] in solutions
containing low molecular weight organic acids (formic, Instrumentation
acetic or propionic) suggested that radical-induced UV
alkylation reactions were leading to the synthesis of volatile All measurements were conducted using an AAnalyst 100
analyte species. More recently, these observations have been atomic absorption spectrometer (Perkin Elmer, Norwalk, CT,
confirmed in other laboratories [24–26]. USA) equipped with deuterium arc background correction
Recent reports have shown that photocatalytic reduction and a mercury hollow cathode lamp source. The spectrometer
of mercury can be achieved using UV irradiation in the was operated under the following conditions: wavelength,
presence of TiO2 as a catalyst and either formic acid [27] or 253.7 nm; spectral band pass, 0.7 nm; lamp current, 6 mA. A
methanol (as a hole scavenger) [28]. Zheng et al. [29] QTA, mounted in the spectrometer’s flame compartment, was
undertook photo-CVG in the presence of formic acid to maintained at room temperature and steady-state absorbance
investigate the speciation of mercury in water by atomic measurements made in continuous graphics mode. A flow-
fluorescence spectrometry. They reported that, when irradi- through photoreactor was constructed by coiling either quartz
ating sample solutions containing formic acid, both Hg2+ (2.0 mm i.d. ×3 mm o.d. ×200 cm long) or Teflon tubing
and CH3Hg+ are reduced to Hg0, thereby permitting the (2.5 mm i.d. ×3.5 mm o.d. ×160 cm long) around a low-
quantitation of total mercury. On the other hand, only Hg2+ pressure UV mercury vapor lamp (254 nm, 15 W, Cole
is reduced (with a comparative efficiency of 10%) to Hg0 Parmer, Vernon Hills, IL, USA). The resulting internal
using visible light. Li et al. [30] proposed a simple method volumes of the coils were comparable at 6 and 8 mL,
of determining mercury in wine or liquor samples by AFS respectively. Samples were pumped through the reactor with
wherein Hg2+ was reduced to Hg0 by UV irradiation as a the aid of a Minipuls 2 peristaltic pump (Gilson, Middleton,
result of the presence of ethanol in the sample matrix. A WI, USA). A 56 mL/min flow of Ar purge gas was
limit of detection of 70 pg/mL was achieved. Bendl et al. introduced into a gas–liquid separator where it merged with
[31] reported a method for the determination of Hg in fish the effluent from the UV photoreactor, transporting the
tissues using photo-CVG of Hg0 in acetic acid solutions. volatile species directly to the QTA. Figure 1 schematically
The determination of total mercury and its speciation in illustrates the system used in this work. Samples were
biological samples requires careful consideration of sample weighed using an M2P microbalance (Sartorius, Göttingen,
pretreatment. A substantial number of different methods are Germany). A Branson 2200 ultrasonic bath (Branson
currently available for this [1]. Recently, the determination Ultrasonic, Danbury, CT, USA) was used to facilitate sample
of mercury in biological materials following their digestion solubilization.
Anal Bioanal Chem (2007) 388:837–847 839

Fig. 1 Schematic of photo-


chemical flow-through UV
reactor

Reagents and samples Procedure 2, for the Approximately 5 mL of a


determination of total solution of 25% m/v TMAH
Analytical-reagent-grade materials were used for all of the mercury using TMAH: in methanol were added to a
experiments. All solutions were prepared using 18 MΩ cm 100–300 mg subsample of
deionized, reverse osmosis water (DIW) obtained from a tissue or mussel. The mixture
mixed-bed ion exchange system (NanoPure, model D4744, was left to stand for at least
Barnstead/Thermoline, Dubuque, IA, USA). A standard stock 3 h at room temperature and
solution (1000 mg/L) of inorganic mercury was prepared by then diluted to 50 mL with
dissolution of mercury chloride (Gold Star, Alfa Chemicals, DIW for a final concentration
Ward Hill, MA, USA) in nitric acid. Similarly, a 1000 mg/L of 2.5% m/v TMAH.
stock solution of methylmercury (CH3HgCl) was prepared by
dissolving methylmercury chloride (Alpha Division, Danvers, Procedure 3, for the Samples were prepared as
MA, USA) in propan-2-ol. Formic and glacial acetic acids determination of described above for total
were sourced from Fisher Chemicals (Fair Lawn, NJ, USA); methylmercury using mercury but 1 g subsamples
tetramethylammonium hydroxide pentahydrate was from TMAH: were taken for analysis and
Alfa Aesar (Ward Hill, MA, USA) and antifoam “B” from after the 3 h incubation
Sigma Aldrich (Steinheim, Germany). Plastic and glass period an aliquot of 2.5 mL
containers were washed with tap water and with a diluted of the digest was diluted with
Extran solution followed by immersion in 10% v/v HNO3 for 10% v/v acetic acid to a final
at least 48 h and a thorough rinse with DIW prior to use. volume of 50 mL, yielding a
The following Certified Reference Materials from the final concentration of 0.125%
National Research Council Canada were used to assess the m/v TMAH in the solution.
accuracy of the method: DOLT-2 and DOLT-3 (dogfish liver),
DORM-1 and DORM-2 (dogfish muscle) and TORT-2 Procedure 4, for the Samples were prepared as
(lobster hepatopancreas). determination of described in Procedure 3, and
methylmercury using sufficient HCl was added to
TMAH and a final solution adjust the pH to 0.3 before
Sample preparation pH of 0.3: dilution with 10% v/v acetic
acid to a final concentration
Samples were prepared using four different procedures, as of 0.125% m/v TMAH. The
follows: resulting slurried sample was
Procedure 1, for the Approximately 250 mg subjected to analysis.
determination of total subsamples of tissue were Five drops of the antifoam “B” were added to all solutions
mercury using formic acid: weighed into 50 mL dark glass to eliminate foaming of the sample in the gas-liquid
flasks and 10 mL of formic separator. Corrections for moisture content were undertaken
acid added. The flasks were following the determination of dry mass in accordance with
placed in an ultrasonic bath at the recommendations of the CRM supplier.
50 °C for 6–8 h to complete
solubilization of the tissues, as
described by Scriver et al. [35]. Procedure
After cooling, the samples
were diluted with DIW to a Using the prepared samples, a sufficient volume of each was
final concentration of 10% v/v introduced into the photoreactor by pumping it through
formic acid. either the Teflon or quartz coil at flow rates of 5–12 mL/min
840 Anal Bioanal Chem (2007) 388:837–847

to produce steady-state absorbance signals. A number of acid was selected for subsequent experiments with this
parameters were examined (acid or TMAH concentration, reagent. Based on these observations, it is possible that the
type of acid, flow rate/residence time in the reactor, transport results presented by Bendl et al. [31] may correspond not to
gas flow rate) in order to optimize response using each of the measurement of total mercury in the fish tissues they
three sample preparation procedures and acquire information investigated but only to methylmercury. This would likely
on total and speciated mercury. A blank response was not result in a significant bias in their results, as the major
measured before every analytical run using solutions con- fraction of the mercury present is methylmercury.
taining only the low molecular weight organic acids or
TMAH. Effect of irradiation time

The residence time of the sample in the flow-through reactor


Results and discussion is an important parameter because the dose of UV radiation
received determines the extent of radical formation and the
Formic acid system efficiency of reduction of mercury. Fixing the length and
volume of the coil, the effect of irradiation time on response
Effect of reagent concentration was investigated by changing the sample flow rate using a
solution of 10% v/v formic acid spiked with 10 ng/mL Hg2+
Low molecular weight organic acids, such as formic and and CH3Hg+. Increasing the pump speed (i.e., decreasing
acetic acid, are typically employed for photochemical the irradiation time) produced a linearly proportional
generation due to their ability to generate volatile species increase in signal intensity for both species. No significant
[15]. The effect of formic acid concentration on the difference in performance between the Teflon and the
absorbance by Hg0 derived from photo-CVG of 10 ng/mL quartz coil was evident, nor between the responses for Hg2+
solutions of Hg2+ and CH3Hg+ is shown in Fig. 2a and b. and CH3Hg+ for residence times varying from 38 to 216 s
Both quartz and Teflon coils were investigated using an (2.5–14 mL/min). An irradiation time of 45 s (sample flow
irradiation time of 45 s (12 mL/min sample flow rate). The rate of 12 mL/min) was selected for subsequent experi-
concentration of formic acid ranged from 0.5 to 50% v/v. In ments using solutions of formic acid. Note that these
all cases, no signal was detected in the absence of formic conclusions were valid for processing a simple calibration
acid. At concentrations as low as 0.5% v/v formic acid standard; it will be shown later that only the longer
the signals appear for both species and above 10% v/v the residence times can be recommended for the processing of
responses are similar from both species irrespective of the digested fish tissue because of the increased opacity of the
type of coil used. It was reported previously [18] that less test solution.
than 5% of the incident UV radiation is transmitted through
the thin Teflon tube wall. Despite this, the use of a 15 W UV Effect of sample pH
lamp is more than sufficient to ensure efficient reduction of
inorganic mercury and methylmercury in simple solutions, as The possibility of undertaking speciation measurements was
evidenced by the fact that the CH3Hg+ is reduced to the investigated by examining the effect of sample pH on response
same extent as the Hg2+ despite the additional need for prior following the addition of NH4OH. The initial pH of 10% v/v
destruction of the carbon–mercury bond. It is also clear, formic acid is 2. Increasing the pH did not improve the
however, that it is not possible to differentiate Hg2+ from separation in response from the two species, and at a pH
CH3Hg+ because both species yield equivalent signals. A higher than 6 the response decreased for both species.
concentration of 10% v/v formic acid was adopted for Adding concentrations of TiO2 in the range 1–4 mg/L as a
subsequent experiments with this reagent. Figure 2c and d photocatalyst did not remediate this situation. When TiO2
summarize results obtained from the use of acetic acid was added to the standard solutions, precipitation of the
during UV irradiation under otherwise similar conditions. In reagent could be observed in the coils.
this case, a concentration of acetic acid less than 5% v/v is
sufficient to achieve maximum response from CH3Hg+ with Tetramethylammonium hydroxide system
no signal from Hg2+; more than 20% v/v acetic acid is
required to promote reduction of Hg2+ from a standard Effect of concentration of TMAH As noted earlier, TMAH
solution. Without a clear understanding of the mechanism of has been successfully used as a solubilizer for biological
the reaction it is not possible to comment further on this tissues and was thus investigated for its efficacy as a
empirical observation. In any case, the use of acetic acid for reagent to promote photoassisted reduction of the mercury
photo-CVG reduction of mercury permits direct speciation species. Figure 3 shows the effect of concentration of
of Hg2+ and CH3Hg+. A concentration of 10% v/v acetic TMAH on the responses from 10 ng/mL of Hg2+ and
Anal Bioanal Chem (2007) 388:837–847 841

Fig. 2 Effect of concentration 0.6 0.5


of formic (a and b) and acetic a b
acid (c and d) on CV-AAS Quartz coil Teflon coil
response from 10 ng/mL solu-
0.5
tions of: (squares) Hg2+and 0.4
(circles) CH3Hg+. a, c: quartz
coil used; b, d: Teflon coil used
0.4
0.3

0.3

0.2
0.2

0.1
0.1
Absorbance

0.0 0.0
0 10 20 30 40 50 0 10 20 30 40 50

% (v/v) HCOOH
0.5 0.5
c Quartz coil d Teflon coil

0.4 0.4

0.3 0.3

0.2 0.2

0.1 0.1

0.0 0.0

0 10 20 30 40 50 0 10 20 30 40 50

% (v/v) CH3COOH

CH3Hg+ in standard solutions using the quartz (Fig. 3a) and CH3Hg+ was prepared in a similar way, but no response
Teflon coil (Fig. 3b) flow-through UV reactor and a could be detected at concentrations of 0.5, 1.0 and 2.5 ng/mL
residence time of 45 s. Substantial signals for CH3Hg+ due to inadequate detection power. In an attempt to
occur in both systems when the concentration of TMAH is improve the sensitivities for these low concentration
as low as 0.025% m/v. Signals gradually increase for Hg+2 standards and to determine CH3Hg+ in the samples, acetic
with increasing concentration of TMAH, and above 2.5% acid was added, in order to achieve an effect evident from
m/v both species are efficiently reduced. This concentration the earlier results shown in Fig. 2. The use of acetic acid
was adopted for further experiments. presumably provides for an increase in methyl radicals in
The responses from the two mercury species are clearly the irradiated solutions, thereby enhancing the rate of
separated according to the data in Fig. 3; when the reduction of CH3Hg+ in TMAH. The effect of acetic acid
concentration of TMAH is 0.125% m/v, CH3Hg+ is prefer- concentration on the response from 1 and 10 ng/mL Hg2+
entially reduced. Samples were thus initially prepared using and CH3Hg+ in 0.125% m/v TMAH standard solutions
2.5% m/v TMAH and then diluted to a final concentration of using an irradiation time of 108 s is shown in Fig. 4. For
0.125% m/v TMAH using DIW. A calibration curve for the 1 ng/mL standard solution of Hg2+ or CH3Hg+, it is not
842 Anal Bioanal Chem (2007) 388:837–847

Fig. 3 Effect of TMAH con- 0.5


centration on response from
10 ng/mL (squares) Hg2+or
0.4
(circles) CH3Hg+ using (a) a
quartz and (b) a Teflon coil
flow-through UV reactor 0.3

0.2

0.1

Absorbance
0.0 a
0 2 4 6 8 10
0.5

0.4

0.3

0.2

0.1

0.0 b
0 2 4 6 8 10
% (m/v) TMAH

possible to discern separation of the species due to the poor suggesting that it is due to the intervention of the chloride
signal-to-noise ratio (Fig. 4a). However, with the 10 ng/mL anion in suppressing the reduction of inorganic mercury.
solutions it is clear that such separation can be achieved Experiments were undertaken to discern the effect of added
(Fig. 4b). Significantly higher absorbance arises from chloride (as NaCl) on the responses from both Hg2+ and
CH3Hg+ at concentrations of acetic acid above 2.5% v/v CH3Hg+. Signal intensity was reduced 50% at 0.5% m/v
and thus 10% v/v was selected for subsequent study. added NaCl, completely suppressed for Hg2+ at 1% m/v
Mineral acids such as HCl or H2SO4 were also used to NaCl, and 90% suppressed for CH3Hg+ at 2.5% m/v NaCl,
adjust the sample pH in order to evaluate the ability to confirming this supposition.
separate the mercury species. Figure 5 shows the effect of Based on this information, the possibility of investigat-
pH on signal intensity for solutions adjusted with HCl. ing speciation in real samples was tested by adjusting the
Samples of DORM-1 were prepared with a final concen- pH to 0.3 with HCl. Samples of DORM-1 were solubilized
tration of 0.125% m/v TMAH in 10% v/v acetic acid in TMAH, diluted to a final concentration of 0.125%
spiked with 10 ng/mL CH3Hg+ and Hg2+, and the final pH TMAH and the solution adjusted to pH 0.3 by the addition
was adjusted to fall in the range 0–1.3 using HCl. As noted of HCl. Both the supernatant and the residue arising from
by Bendl et al. [31], the addition of HCl caused precipitation subsequent centrifugation were examined separately. The
in sample digests, which was removed by centrifugation residue was diluted with 10% v/v acetic acid. By subjecting
before measurements. An initial pH of 2 (without the each phase to photo-CVG it was found that signals from the
addition of HCl) provides for responses from both species supernatant were low compared to those from the residue
in calibration solutions and in those of digested tissues. By and did not provide satisfactory results for CH3Hg+.
decreasing the pH, an effective separation of the response Mercury species are thus retained on the precipitate and
from each species occurs due to selective suppression of the can subsequently be reduced by the UV. As such, analysis
reduction of Hg2+ and, at pH 0.3, it appears that it is possible of DORM-1 samples that were not subjected to a prior
to undertake speciation measurements not only in simple centrifugation provided results for CH3Hg+ which were
standards but also in solutions of digested tissue samples. It concordant with the certified values.
is worth noting that, although the use of H2SO4 to adjust the
pH did not give rise to a precipitate, it was not possible to Effect of visible light on photoreduction of Hg Zheng et al.
separate the signals from the two species in this medium, [29] reported on the ability of visible light to photoreduce
Anal Bioanal Chem (2007) 388:837–847 843

Fig. 4 Effect of acetic acid


0.04
concentration on response from
standard solutions containing
0.125% m/v TMAH for (a) 0.03
1 ng/mL and (b) 10 ng/mL
(squares) Hg2+or (circles)
CH3Hg+using a quartz coil 0.02
flow-through UV reactor and an
irradiation time of 108 s
0.01

Absorbance
a
0.00
0 10 20 30 40 50
0.4

0.3

0.2

0.1

0.0 b
0 10 20 30 40 50
% (v/v) CH3COOH

Hg2+ in formic acid, leading to the conclusion that study to evaluate the effect of visible light on the
speciation was possible by alternately exposing a solution photoreduction of Hg2+ and thus the stability of calibration
to UV for the quantitation of total mercury, and visible light solutions and digested samples exposed to ambient labora-
for the selective measurement of inorganic mercury. A tory lighting in the presence of formic acid and TMAH was
Fig. 5 Effect of pH (using HCl 0.35
for adjustment) on response
from (a) a calibration standard
0.28
solution consisting of 0.125%
m/v TMAH and 10%
v/v acetic acid, and (b) a solu- 0.21
bilized sample of DORM-1
containing 0.125% m/v TMAH 0.14
and 10% v/v acetic acid. Both
samples were spiked with 0.07
10 ng/mL (squares) Hg2+and
Absorbance

(circles) CH3Hg+. A quartz coil 0.00 a


flow-through UV reactor was
used and the irradiation time 0.0 0.5 1.0 1.5 2.0
was 108 s
0.28

0.21

0.14

0.07

0.00 b
0.0 0.5 1.0 1.5 2.0
pH
844 Anal Bioanal Chem (2007) 388:837–847

1.2 flasks. A control solution was prepared and stored in the


a
dark for 2 h. A fresh standard solution was prepared prior to
1.0
the actual measurement sequence and used to normalize the
0.8 responses from all other samples following their processing
through the UV photoreactor.
0.6 In the formic acid system, exposure of the sample to
0.4 visible light did not result in any statistically significant loss
Normalized absorbance

of mercury species from the solution, as shown by the


0.2 results in Fig. 6a. This contrasts with the results of Zheng et
al. [29] for which a reduction of approximately 10% was
0.0 achieved and used for analytical purposes. For the TMAH
, ,
0/0 1/0 2/0 0/2 0 /0 system (Fig. 7b), exposure of standard solutions to visible
light resulted in significant losses of Hg2+, whereas losses
1.0 b
of CH3Hg+ were minor. Moreover, significant losses of
0.8
Hg2+ were also measurable for solutions kept in the dark
for 2 h, suggesting an intrinsic ability of TMAH to reduce
0.6 Hg2+ at room temperature in pure solutions of the standard.
This has significant implications for the analysis of biotissues
0.4 solubilized with TMAH—inorganic mercury appears unsta-
ble in calibration standards and methylmercury is thus
0.2 recommended for the preparation of such solutions.
An additional study was performed to evaluate the
0.0
, , stability of mercury in solutions of DOLT-2 tissue samples
0/0 1/0 2/0 0/2 0 /0
digested with formic acid or TMAH to which a 10 ng/mL
Exposure time, h spike of Hg2+ was added. The samples were purged with a
Fig. 6 Effect of visible and ambient laboratory fluorescent light on 65 mL/min flow of Ar for 4 min prior to processing them
photo-CVG of mercury using 10 ng/mL (light bars) Hg2+or (dark through the UV photoreactor for measurement of residual
bars) CH3Hg+. (0/0) freshly prepared solution; (1/0) 1 h exposure to
visible light; (2/0) 2 h exposure; (0/2) 2 h exposure to ambient mercury content. Figure 7a and b display results for Hg2+,
fluorescent laboratory light and (0’/0’) 2 h in the dark. a, 10% v/v from which it is evident that no losses have occurred.
formic acid; b, 2.5% m/v TMAH Chemical reduction of inorganic mercury (noted from the
results of Fig. 6a) either does not occur in such digested
undertaken. Standard solutions in 10% v/v formic acid or tissue samples because of the complexity of the matrix
2.5% m/v TMAH spiked with 10 ng/mL of Hg2+ or (possible inhibition of the formation of radicals or reducing
CH3Hg+ and contained in glass volumetric flasks were gases) or the presence of partially digested or solubilized
exposed for one and two hours to light from a 50 W high amino acids and proteins in the solution stabilizes any Hg0
intensity tungsten lamp (from a slide projector), and for two formed. This suggests that quantitation of Hg2+ be
hours to ambient laboratory (fluorescent) lighting. Care was undertaken in such digests using the method of additions.
taken to ensure there was no increase in the temperature of A third study was undertaken using standard solutions
the solutions during exposure by forced air cooling of the containing 50 ng/mL Hg2+ or CH3Hg+. In addition to 10% v/v

Fig. 7 Influence of visible light


1.0 a 1.0 b
Normalized absorbance

on the photo-CVG response


from a 10 ng/mL Hg2+spike to
digested DOLT-2 in (a) 10% v/v 0.8 0.8
formic acid and (b) 2.5% m/v
TMAH. Conditions: (0/0) 0.6 0.6
freshly prepared sample; (1/0)
1 h exposure to visible light; 0.4 0.4
(2/0) 2 h exposure; (0/2) 2 h
exposure to ambient fluorescent
0.2 0.2
laboratory light and (0’/0’) 2 h
in the dark. Each sample was
purged for 4 min at 65 mL/min 0.0 0.0
0/0 1/0 2/0 0/2
, ,
0 /0 , ,
Ar prior to measurement of 0/0 1/0 2/0 0/2 0 /0
residual mercury
Exposure time, h
Anal Bioanal Chem (2007) 388:837–847 845

Fig. 8 Effect of purge time 1.2


(65 mL/min Ar) on the photo- a 1.0 b

Normalized absorbance
CVG response from 50 ng/mL 1.0
(light bars) Hg2+or (dark bars) 0.8
CH3Hg+in: (a) 10% v/v formic 0.8
acid and (b) 2.5% m/v TMAH
0.6
(for Hg2+) or 0.125% m/v 0.6
TMAH plus 10% v/v acetic acid
(for CH3Hg+) 0.4
0.4

0.2 0.2

0.0 0.0
0 2 4 6 8 10 0 2 4 6 8 10
Time, min

formic acid, a medium optimized for speciation of methyl- summarized in Table 1. Relative standard deviations were
mercury consisting of 0.125% m/v TMAH and 10% v/v below 5% in most cases, except for TORT-2 (15% RSD),
acetic acid or 2.5% m/v TMAH for total Hg was also which most likely arose because the analyte concentration is
examined. A 65 mL/min flow of Ar was passed through the so low in this sample. Samples of DOLT-2, DOLT-3 and
solutions for varying periods of time while they were TORT-2 produced dark solutions when digested with formic
maintained in the dark. Results are presented in Fig. 8a acid, whereas that for DORM-1 was yellow. This may be the
and b. Inorganic mercury is unaffected by the flow of gas reason why poor results were obtained when a high sample
whereas losses of CH3Hg+ occur which increase with flow rate was used; penetration of UV into the solution is
increasing purge time in the formic acid solution. For the more efficient for clear and colorless samples than for the
TMAH/acetic acid system (Fig. 8b), both species suffer more opaque ones. When a Teflon coil was used in the UV
losses with increasing Ar purge time, consistent with the photoreactor, results were satisfactory only for DORM-1.
conclusions drawn from Fig. 6b, but more pronounced Clearly, the efficiency of the process is dependent on the
because of the mechanical stripping of the Hg0 from solution dose of UV radiation received which, in turn, is influenced
induced by the flow of Ar. by not only the residence time of the sample in the UV field,
but also the absorptivity of the sample solution and the
Analytical applications Using optimized experimental con- transmission efficiency of the tubing.
ditions, the accuracy of the proposed methodology was For the determination of total mercury following
evaluated by analyzing several Certified Reference Materials. digestion with TMAH (procedure 2), external calibration
Total mercury was determined in samples that had been curves were prepared by spiking Hg2+ into 2.5% m/v
solubilized in formic acid and diluted to a final concentration TMAH. Sample processing was initially undertaken using a
of 10% v/v. An external calibration curve was constructed quartz coil at a flow rate of 12 mL/min (residence time of
using the same formic acid medium spiked with Hg2+. 45 s). As for the formic acid system, the results were
Samples and standards were subjected to photo-CVG by discordant with the certified values for total mercury for
pumping the solutions through a quartz coil in the photo- DOLT-2, DOLT-3 and TORT-2 and, as such, the samples
reactor at a rate of 12 mL/min (45 s irradiation time). At this were re-run at a flow rate of 5 mL/min to increase exposure
flow rate the results were not concordant with the certified time to the UV field to 108 s. Under these conditions,
values. By reducing the flow rate to 5 mL/min (108 s results were satisfactory, as evidenced by data summarized
irradiation time), satisfactory results were achieved, as in Table 1. When the Teflon coil was used it was only

Table 1 Determination of total mercury in biological materials by CV-AAS following photo-CVG*

CRM Certified value, mg/kg Solubilized in 10% v/v formic acid Solubilized in 2.5% m/v TMAH

Determined**, mg/kg RSD, % Determined**, mg/kg RSD, %

DORM-1 0.798±0.074 0.75±0.01 1.5 0.78±0.02 2.7


DOLT-2 2.14±0.28 2.00±0.10 5.0 2.20±0.05 2.3
DOLT-3 3.37±0.14 3.27±0.03 0.9 3.38±0.06 1.8
TORT-2 0.27±0.06 0.27±0.04 14.8 0.30±0.02 6.7

*with quartz coil; no significant difference with Teflon coil


**mean and standard deviation (n=3), dry mass corrected.
846 Anal Bioanal Chem (2007) 388:837–847

possible to determine total Hg in samples of DORM-1 at Table 3 Figures of merit (in ng/mL)
this flow rate because of the poor penetration of the already Slope−1±std dev R2 LOD LOD (ng/g)
reduced intensity of UV through the opaque sample
solutions. It is possible that longer residence times for Hg2+
irradiation could alleviate this problem, but this was not (a) 0.032±0.004 0.9999 0.03 6
(b) 0.026±0.006 0.9999 0.04 8
investigated.
CH3Hg+
For the determination of CH3Hg+ in biological samples
(c) 0.018±0.002 0.9994 0.05 10
solubilized with TMAH (procedure 3), an external calibra- SnCl2
tion curve was prepared in a mixture of 0.125% m/v (d) 0.033±0.003 0.9999 0.02 –
TMAH and 10% v/v acetic acid by spiking with CH3Hg+.
With the exception of TORT-2, results summarized in Calibration curves prepared in: (a) 10% v/v formic acid; (b) 2.5% m/v
TMAH with Hg2+ standard; (c) 0.125% m/v TMAH+10% v/v acetic
Table 2 show good agreement with the certified values. For acid with CH3Hg+ standard and (d) following SnCl2 reduction (see
TORT-2, the concentration of CH3Hg+ is lower than the text for optimized conditions).
limit of quantitation. Relative standard deviations of
replicate measurements were better than 6.4% for measure-
ments of both total and organic mercury using a TMAH Calibration curves based on standard solutions of Hg2+ or
sample preparation. Although speciation analysis is difficult CH3Hg+ in the range 0.5–25 ng/mL were generated using
because of the substantial dilution of the sample to achieve formic acid and TMAH systems optimized for the
a concentration of 0.125% m/v TMAH, this impediment determination of mercury in samples of digested biological
could be readily overcome by using a more powerful tissue (i.e., 108 s residence time). Correlation coefficients
detection technique, such as AFS or ICP-MS. were higher than 0.999. Estimated limits of detection
For the determination of CH3Hg+ in biological samples (LOD), defined as three times the standard deviation of
which were solubilized with TMAH diluted with acetic acid ten replicate measurements of the blank divided by the
and adjusted to pH 0.3 with HCl (procedure 4), an external slope of the calibration curve, were calculated for each
calibration curve was prepared using the same medium as system and also for samples. The data in Table 3 illustrate
the samples by spiking with CH3Hg+. The results presented that the LODs obtained using photo-CVG based on UV
in Table 2 also show good agreement with the certified irradiation of formic acid or TMAH for the determination of
values for methylmercury, demonstrating yet another photo- total Hg are similar to those obtained using a conventional
CVG approach to the speciation of mercury. SnCl2 system to reduce Hg2+ to Hg0. Procedural LODs for
Results reported in Tables 1 and 2 for the determination the samples are generally adequate for the analysis of these
of both total mercury and methylmercury in tissue samples CRMs, and the use of more powerful instrumentation for
using photo-CVG were subjected to a t-test to ensure that detection should extend the lower range of concentrations
there was no statistical difference from the certified value of which can be addressed.
the measurand at the 95% level of confidence. It is of interest to determine the efficiency of generation,
which is defined by the convolution of the efficiency of
Figures of merit Figures of merit for the determinations of formation of the volatile species with those of its gas–liquid
total mercury and methylmercury in biological tissues using separation and transport to the detector. Generation effi-
a quartz coil flow-through UV reactor are shown in Table 3. ciency was estimated from a comparison of CV-AAS

Table 2 Determination of methylmercury in biological materials by CV-AAS following photo-CVG*

CRM Certified value, mg/kg A B

Determined**, mg/kg RSD, % Determined**, mg/kg RSD, %

DORM-1 0.731±0.060 0.78±0.05 6.4 0.76±0.02 2.6


DORM-2 4.47±0.32 4.25±0.07 1.6 4.19±0.07 1.7
DOLT-2 0.693±0.053 0.68±0.04 5.9 0.66±0.02 3.0
DOLT-3 1.7 1.63±0.10 6.1 1.41±0.08 5.7
TORT-2 0.152±0.013 <LOD ND ND ND

(A) solubilization of tissues with 2.5% m/v TMAH and diluted to 0.125% m/v with 10% v/v acetic acid
(B) solubilization with TMAH and diluted to 0.125% m/v with acetic acid and adjusted to pH 0.3 using HCl
*with quartz coil; no significant difference with Teflon coil
**mean and standard deviation (n=3), dry mass corrected
ND: not determined
Anal Bioanal Chem (2007) 388:837–847 847

response using a conventional SnCl2 chemical reduction 6. Cai Y (2000) Trends Anal Chem 19:62–66
method with the results obtained using photo-CVG. The 7. Gámiz-Gracia L, Luque de Castro MD (1999) J Anal Atom
Spectrom 14:1615–1617
reduction conditions were optimized for the tin chloride 8. Yan X, Ni Z, Guo Q (1993) Anal Chim Acta 272:105–114
system (4% v/v HCl; 0.5%, m/v SnCl2 in 3% v/v HCl; gas 9. Flores EMM, Welz B, Curtius AJ (2001) Spectrochim Acta B
flow rate of 50 mL/min and 60 cm length of reaction coil) 56:1605–1614
and the flow rates of reductant and sample set to 2.5 mL/min 10. Akagi H, Sakagami Y (1972) J Hyg Chem (Jpn) 18:358
11. Akagi H, Takabatake E (1973) Chemosphere 3:131–133
to match the flow rates in the photo-CVG system in order to 12. Akagi H, Fujita Y, Takabatake E (1976) Chem Lett 49:1–4
ensure delivery of identical fluxes of Hg2+ to the gas–liquid 13. Canario J, Vale C (2004) Environ Sci Technol 38:3901–3907
separator. From a comparison of the slopes of the 14. Siciliano SD, O’Driscoll NJ, Tordon R, Hill J, Beauchamp S,
calibration curves, the efficiency of photo-CVG with UV Lead DRS (2005) Environ Sci Technol 39:1071–1077
15. Golimowski J, Golimowska K (1996) Anal Chim Acta 325:
irradiation was determined to be approximately 95% of that 111–133
obtained by conventional chemical-CVG. 16. Capelo-Martínez JL, Xímenes-Embún P, Madrid Y, Cámara C
(2004) Trends Anal Chem 23:331–340
17. Guo X, Sturgeon RE, Mester Z, Gardner GJ (2003) Anal Chem
75:2092–2099
18. Guo X, Sturgeon RE, Mester Z, Gardner GJ (2003) Environ Sci
Conclusions Technol 37:5645–5650
19. Guo X, Sturgeon RE, Mester Z, Gardner GJ (2005) J Anal Atom
Photo-CVG with UV irradiation combined with CV-AAS Spectrom 20:702–708
20. McSheehy S, Guo X-M, Sturgeon RE, Mester Z (2005) J Anal
detection for analysis of samples solubilized with formic acid Atom Spectrom 20:709–716
or TMAH has been successfully used to determine total 21. Guo X, Sturgeon RE, Mester Z, Gardner GJ (2004) Appl
mercury and methylmercury in biological tissues. Both Organomet Chem 18:205–211
reagents provide stable solutions of the analyte and promote 22. Guo X, Sturgeon RE, Mester Z, Gardner GJ (2004) Anal Chem
76:2401–2405
reduction of Hg2+ and CH3Hg+ to Hg0 following exposure to 23. García M, Figueroa R, Lavilla I, Bendicho C (2006) J Anal Atom
the UV field. The methodology does not require extensive Spectrom 21:582–587
sample preparation, it eliminates the need to use (unstable 24. Wang Q, Liang J, Qiu J, Huang B (2004) J Anal Atom Spectrom
and expensive) reducing agents, and it provides limits of 19:715–716
25. Figueroa R, García M, Lavilla I, Bendicho C (2005) Spectrochim
detection that are comparable to conventional methods for Acta B 60:1556–1563
the determination of mercury by vapor generation. 26. Sun YC, Chang YC, Su CK (2006) Anal Chem 78:2640–2645
27. Wang X, Penkonen SO, Ray AK (2004) Electrochim Acta
49:1435–1444
Acknowledgement M. A. Vieira and A.J. Curtius are grateful to the 28. Khalil LB, Rophael MW, Mourad WE (2002) Appl Catal
Conselho Nacional de Pesquisas e Desenvolvimento Tecnológico B36:125–130
(CNPq) for a research scholarship. 29. Zheng C, Li Y, He Y, Ma Q, Hou X (2005) J Anal Atom Spectrom
20:746–750
30. Li Y, Zheng C, Ma Q, Wu L, Hu C, Hou X (2006) J Anal Atom
Spectrom 21:82–85
References 31. Bendl RF, Madden JT, Regan AL, Fitzgerald N (2006) Talanta
68:1366–1370
1. Leermakers M, Baeyens W, Quevauviller P, Horvat M (2005) 32. Tao G, Willie SN, Sturgeon RE (1998) Analyst 123:1215–1218
Trends Anal Chem 24:383–393 33. Nóbrega JA, Santos MC, De Sousa RA, Cadore S, Barnes RM,
2. Welz B, Sperling M (1999) Atomic absorption spectrometry. Tatro M (2006) Spectrochim Acta B 61:465–495
Wiley-VCH, Weinheim 34. Torres DP, Vieira MA, Ribeiro AS, Curtius AJ (2005) J Anal
3. Ribeiro AS, Vieira MA, Curtius AJ (2004) J Braz Chem Soc Atom Spectrom 20:289–294
15:825–831 35. Scriver C, Kan M, Willie S, Soo C, Birnboim H (2005) Anal
4. Sturman BT (1985) Appl Spectrosc 39:48–56 Bional Chem 381:1460–1466
5. Welz B, Schubert-Jacobs M, Schlemmer G (1988) Fresenius Z 36. Kan M, Willie SN, Scriver C, Sturgeon RE (2006) Talanta
Anal Chem 331:324–329 68:1259–1263

You might also like