Analytica Chimica Acta 721 (2012) 22–27

Contents lists available at SciVerse ScienceDirect

Analytica Chimica Acta

journal homepage: www.elsevier.com/locate/aca

Speciation of mercury in fish samples by flow injection catalytic cold vapour

atomic absorption spectrometry
Yanlin Zhang, Samuel B. Adeloju ∗
NanoScience and Sensor Technology Research Group, School of Applied Sciences and Engineering, Monash University, Churchill, Victoria 3842, Australia

a r t i c l e i n f o a b s t r a c t

Article history: A rapid flow injection catalytic cold vapour atomic absorption spectrometric (FI-CCV-AAS) method is
Received 5 September 2011 described for speciation and determination of mercury in biological samples. Varying concentrations
Received in revised form 17 January 2012 of NaBH4 were employed for mercury vapour generation from inorganic and mixture of inorganic and
Accepted 19 January 2012
organic (total) Hg. The presence of Fe3+ , Cu2+ and thiourea had catalytic effect on mercury vapour gener-
Available online 31 January 2012
ation from methylmercury (MeHg) and, when together, Cu2+ and thiourea had synergistic catalytic effect
on the vapour generation. Of the two metal ions, Fe3+ gave the best sensitivity enhancement, achieving
Keywords:
the same sensitivity for MeHg and inorganic Hg2+ . Due to similarity of resulting sensitivity, Hg2+ was
Inorganic mercury
Methylmercury
used successfully as a primary standard for quantification of inorganic and total Hg. The catalysis was
Flow injection homogeneous in nature, and it was assumed that the breaking of the C Hg bond was facilitated by the
Cold vapour atomic absorption delocalization of the 5d electron pairs in Hg atom. The extraction of MeHg and inorganic mercury (In-
spectrometry Hg) in fish samples were achieved quantitatively with hydrochloric acid in the presence of thiourea and
Catalyst determined by FI-CCV-AAS. The application of the method to the quantification of mercury species in
Fish a fish liver reference material DOLT-4 gave 91.5% and 102.3% recoveries for total and methyl mercury,
respectively. The use of flow injection enabled rapid analysis with a sample throughput of 180 h−1 .
© 2012 Elsevier B.V. All rights reserved.

1. Introduction The majority of instrumental analytical methods developed

Mercury species present in the environment differ greatly in
their bio-physico-chemical properties, particularly in their toxi-
city, solubility, and rate of bioaccumulation by organisms [1–3].
For example, organomercurials are generally more toxic than
inorganic mercury compounds [4,5]. Yet organic mercury species,
such as monomethyl mercury (MMeHg or MeHg), dimethyl mer-
cury (DMeHg), ethyl mercury (EtHg), and phenyl mercury (PhHg)
are often detected in environmental samples, such as sediment,
water and soil [6–8]. Of these, MeHg is the only compound that
is bioaccummulated and biomagnified in the food chain and is
also by far the most toxic mercury compound which represents
a major health risk [9]. For this reason, the determination of total
Hg (T-Hg) concentrations in samples does not offer adequate
information on potential risk of mercury exposure, particularly
with respect to the presence of organic mercury. A knowledge
of mercury species present is therefore more relevant in gaining
better understanding of the impact of mercury pollution, and
information on MeHg content in environmental and biological
samples has special significance [10].
∗ Corresponding author. Tel.: +61 3 9902 6450; fax: +61 3 9902 6738.
E-mail address: Sam.Adeloju@monash.edu (S.B. Adeloju).
for the determination of MeHg to date are based on chromato-
graphic separation, such as GC [11–27], HPLC [28–37], and IC [38],
hyphenated with different detection techniques including ECD
[23,25], UV [30,32], FAAS [33], CVAAS [11,22,26,29,37,38], CVAFS
[12,13,16–18,21,24,36], MS [27], ICP-AES [20,32], and ICP-MS
[14,15,21,31,34,35]. In particular, GC hyphenated with spectro-
scopic detection is one of the most popular techniques. An obvious
advantage of chromatographic methods is the ability to distinguish
between different mercury species in samples. However, a common
disadvantage in these methods is that they involve complex and
tedious pre-treatment procedures for separation of organic mer-
cury (MeHg in biological samples) from the sample matrix by differ-
ent methods, such as distillation, alkaline digestion-solvent extrac-
tion and acidic digestion-solvent extraction [19,23]. The separation
of Hg species from sample matrix is one of the most critical steps
and, for biota and sediments, almost certainly the most critical.
The extraction of Hg compounds from biological samples can
be accelerated with the aid of microwave [39–41] or ultrasound
[23,36], but the whole analytical procedures when employed with
chromatographic methods are still tedious and time-consuming.
In addition, hyphenated analytical methods, such as GC-AAS, GC-
AFS, HPLC-AAS/AFS are not commercially or readily available, but
often assembled for specific use only in laboratories. Also, the for-
mation of extraneous MeHg during the analytical procedure is a
problem often encountered. In particular, when distillation method

0003-2670/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.aca.2012.01.038
 Y. Zhang, S.B. Adeloju / Analytica Chimica Acta 721 (2012) 22–27
is applied to sediments, it may result in a significant bias in mea-
surements [42]. Methylation artefacts were also observed during
hot alkaline digestion and supercritical fluid extraction [42].
Due to these various disadvantages, non-chromatographic
methods have attracted significant interest in recent decades
and numerous methods have been reported [14,43–64]. Typi-
cally, discrimination between In-Hg and MeHg was realized by
the direct reduction of In-Hg with a weak reductant and reduc-
tion of total mercury (T-Hg) after digestion or in-line or off-line
oxidation, and MeHg concentration can be calculated by the
difference between In-Hg and T-Hg concentrations [43–49]. Selec-
tive extraction of MeHg from biological samples has also been
reported [50–54]. For example, MeHg can be selectively extracted
by using 2 M HCl without extracting In-Hg [50,51]. In recent years,
in-line separation method has been reported based on the dif-
ferent retention properties of In-Hg and MeHg [14,55–58]. An
additional benefit of this approach is the ease of analytes pre-
concentration which enabled achievement of improved sensitivity
[58].
Without digestion or oxidation, direct determination of T-Hg in
acid extracts by vapour generation-spectroscopic techniques has
proven to be difficult because the vapour generation efficiencies of
In-Hg and MeHg are different [43]. However, sequential determi-
nation of In-Hg and T-Hg in the same acid extracts by CVAAS has
been reported [50,60–62]. These were based on the discrimination
between In-Hg and MeHg by using different NaBH4 concentrations,
but the application of this approach is seldom considered.
It was reported that in the presence of Cu2+ , alkaline SnCl2
solution gave the same sensitivity for MeHg and Hg2+ , enabling
the determination of In-Hg and T-Hg in the same sample solution
without digestion or oxidation of MeHg [63]. However, this was
disregarded because of the formation of precipitate in the sample
solution and associated contamination problems [61]. The catalytic
effect of Fe3+ was also employed to achieve improved sensitivity
with NaBH4 as the reductant for MeHg [65,66], but no detailed opti-
mization of catalyst concentration or explanation of the catalytic
mechanism was reported.
In this study, we investigate the catalytic effect of Cu2+ , TU and
3+
Fe on the vapour generation of MeHg with NaBH4 as the reducing
agent and an attempt is made to explain the possible mechanism
involved. The possibility of achieving rapid ultrasound extraction of
In-Hg and MeHg with HCl in the presence of thiourea (TU) will also
be investigated. Furthermore, the application of the method to the
determination of In-Hg and MeHg in fish samples by FI-CCV-AAS
will be considered.
2. Experimental
2.1. Reagents and apparatus
All chemicals were of analytical reagent grade unless stated
otherwise. 12.5 mg of methylemercury chloride (Merck, Darm-
stat, Germany) was dissolved in 20 mL of 5 M HCl and diluted to
100 mL with 2 M HCl to give a MeHg concentration of 125 mg L−1
(100 mg Hg L−1 ). This stock solution was stored in the fridge
at 4 ◦ C. A diluted solution (1 mg L−1 ) was prepared fortnightly
and working standard solutions were prepared daily. Hg2+ stan-
dard solution of 1000 mg L−1 was purchased from BDH Chemicals
(Australia) Pty Ltd. (Kilsyth, Victoria, Australia) and adequate dilu-
tion was made prior to use. Milli-Q water was used throughout the
study.
A 1% NaBH4 solution was prepared by dissolving 1 g of NaBH4
in 20 mL of 0.1% NaOH solution and diluted to 100 mL with Milli-Q
water. Dilution of this solution for the vapour generation of Hg2+
or T-Hg was made with 0.1% NaOH solution. A 1% Fe3+ solution was
23
Fig. 1. FI-CCV-AAS system used for Hg determination. T: time delay relay; V: pinch
valve; P: peristaltic pump; GLS: gas–liquid separator; AC: absorption cell; FM: flow
meter; J: joining point; a and b: tubular reactors; c: gas delivery tube; S: sample; R:
reductant (NaBH4 solution); Ar: argon gas; C: carrier stream; W: waste.
prepared by dissolving 2.9 g of FeCl3 in 100 mL of 0.5 M HCl. A 1%
antifoaming solution was prepared by dissolving 0.25 g of antifoam
SE-15 (Sigma–Aldrich, St. Louis, USA) in Milli-Q water. A 1% sodium
tetraethylborate (NaBEt4 ) solution and a 1% sodium tetraphenylb-
orate (NaBPh4 ) solution were prepared by dissolving 0.25 g of solid
NaBEt4 (Sigma–Aldrich Chemie GmbH, Steinheim, Germany) and
0.25 g of NaBPh4 (Cros Organics, New Jersey, USA), respectively,
in 25 mL of Milli-Q water. Methylethyl mercury (MeEtHg) and
methylphenylmercury (MePhHg) were prepared by aqueous phase
ethylation and phenylation of MeHgCl at pH 4.5 with NaBEt4 and
NaBPh4 , respectively. Dimethyl mercury (DMeHg) and dibutyl mer-
cury (DBuHg) were synthesized through the Grignard reagents [67]
and their structure was confirmed with a GC-MS system which
comprised of a Hewlett Packard 6890 GC and a Hewlett Packard
5970 mass selective detector.
A Perkin-Elmer 3030 atomic absorption spectrometer (Perkin-
Elmer Pty Ltd., Rowville, Victoria, Australia) was used with Hg
hollow cathode lamp as the irradiation source for the measure-
ment of Hg. The instrument was operated in background correction
mode with a deuterium lamp as the continuum irradiation source.
The detection was carried out at 253.7 nm with a bandpass width
of 0.7 nm. An absorption cell was assembled in our laboratory, as
described previously [68]. The transient signals were recorded with
a Perkin-Elmer 56 chart recorder.
The sample extraction was carried out in a Soniclean 160 T ultra-
sound bath (Extech Equipment Pty. Ltd., Wantirna South, Victoria,
Australia). To perform the vapour generation, a flow injection (FI)
system was assembled in our laboratory as described previously
[69], except that in the present case we have included a time delay
relay and a cone-shaped absorption cell, as illustrated in Fig. 1.
Tygon pump tubing of 1.6 mm i.d. and 0.32 mm i.d. (Elkay Prod-
ucts, Inc. Worcester, Massachusetts, USA) were used to deliver the
sample and reductant solutions, respectively. Flexible pump tub-
ing and PTFE tubing of 1 mm in diameter (Cole-Parmer Instrument
Company, Verson Hills, Illinois, USA) were used for the connec-
tion and liquid delivery. The tubular reactors a and b were 1 mm
i.d. and 10 cm long. The pump was operated at 48 rpm, achieving
flow rates of 9.6 and 3.2 mL min−1 for the sample and reduc-
tant, respectively. A U tube was used as the gas–liquid separator
(GLS).
The operational procedure of the FI-CCV-AAS system composed
of two steps. In the first step, sample and reductant solutions were
loaded, while in the second step, the sample and reductant bands
were injected into their channels and merged at downstream of
the pump and vapour generation reaction occurred in the tubular
reactor a. The gas–liquid mixture was carried into the GLS where
the gas phase was separated and swept into the absorption cell,
whereas the liquid phase was discarded.
24 Y. Zhang, S.B. Adeloju / Analytica Chimica Acta 721 (2012) 22–27

2.2. Extraction and determination of mercury species 120

3
100

Relative signal intensity

0.1–0.4 g of homogenized fish tissue or liver sample was added
into a 10 mL centrifuge tube, followed by the addition of 5 mL of 2
2 M HCl and 0.2 mL of 0.5 M TU. The tube was screw capped and 80

(ABS)
manually shaken for 10 s. Extraction was then performed by son-
60
icating for 5 min at 100% of the ultrasound bath energy setting. 1
The slurry was centrifuged at 1300 rpm for 10 min and the super-
40
natant was transferred into a 50 mL volumetric flask. 0.5 mL of 1%
SE-15 antifoam solution was added and diluted to volume with 20
Milli-Q water. Hg2+ in this solution was determined initially by
CVAAS with 0.0001% NaBH4 solution. Then, an aliquot of 1% Fe3+ 0
solution was added into the flask to give a Fe3+ concentration of 0 1 2 3 4 5
100 mg L−1 and total mercury was determined by CCV-AAS with Cu2+ (mg /L)
1% NaBH4 solution.
For the determination of Hg2+ , air was used as the carrier stream Fig. 2. Catalytic effect of Cu2+ and TU on Hg response. (1) MeHg (20 ␮g L−1 in Hg),
for both the sample and the reductant and 0.0001% NaBH4 solution TU = 0, (2) MeHg (20 ␮g L−1 in Hg), TU = 0.1 M, and (3) Hg2+ (20 ␮g L−1 ). (All the signal
was used as the reductant. The pump speed was 48 rpm, giving flow intensity results are relative values to that of Hg2+ as 100%.)

rates of 9.6 and 3.2 mL min−1 for the sample and reductant, respec-
tively. The sample loading and injection time intervals were fixed at
5 and 15 s, respectively. To determine T-Hg, 0.2 M HCl and H2 O were
used as the carrier streams for the sample and reductant, respec-
tively. The same pump speed and sample loading and injection time
as for Hg2+ were used. 100 ppm Fe3+ was present in the sample
solution and 1% NaBH4 was used as the reductant. Hg2+ standard
solutions were used for the calibration for both Hg2+ and T-Hg and
quantification was performed by the use of peak height. The con-
centration of MeHg was calculated from the difference between
T-Hg concentration and Hg2+ concentration.
3. Results and discussion
3.1. Optimization of vapour generation process
The three basic parameters that influenced sensitivity of Hg2+
and T-Hg were the sampling time, injection time and argon gas flow
rate. The use of a sampling time of 5 s gave optimum signal inten-
sity, and an injection time of 15 s was employed to avoid sample
carry-over. Under these conditions, the sensitivity of the FI-CCV-
AAS measurement changed with an increase in argon gas flow rate
and the optimum sensitivity was obtained with an Ar flow rate of
125 mL min−1 .
Acid solution is normally used in FI-CVAAS measurement of
Hg2+ as the carrier stream for sample solution, while the reduc-
tant (SnCl2 or NaBH4 solution) is usually introduced continuously.
In this study, it was found that the same sensitivity can be achieved
by using water as the carrier streams for both the sample and NaBH4
solutions. It was also found that, when air was used as the carrier
streams for the sample and NaBH4 solutions, 25% improvement in
sensitivity was obtained for Hg2+ compared with the use of water
as the carrier streams. Under such conditions, the baseline obtained
Cu2+ (5 mg L−1 ) alone improved the sensitivity by about 4 folds,
while TU alone improved the sensitivity by only about 1 fold.
0.1 M TU added into the NaBH4 solution and 2 mg L−1 of Cu2+
added into the sample solution increased the sensitivity by
almost 8 folds and the sensitivity of MeHg was 87% of that of
Hg2+ .
Further study demonstrated that Fe3+ was a better sensitivity
enhancing agent than Cu2+ and TU. As shown in Fig. 3, the sensi-
tivity of MeHg response increased with increasing Fe3+ and NaBH4
concentrations and the optimum sensitivity was obtained when
≥100 mg L−1 Fe3+ was present in the sample solution with 0.1%
NaBH4 as the reductant. In the presence of 100 mg L−1 Fe3+ , the
same sensitivity was obtained for both MeHg and Hg2+ . This implies
that Hg2+ can be used as the sole standard for quantification of both
Hg2+ and MeHg. However, 1% NaBH4 is needed for fish liver samples
possibly due to matrix interference.
3.3. Possible mechanism of catalytic effects
Although Fe3+ has strong catalytic effect on the reduction of
MeHg, no such effect was observed with Fe2+ under the same con-
ditions. Given that Fe3+ ions were reduced to Fe atoms through Fe2+
by NaBH4 and the resulting atoms formed clusters when their con-
centration was high enough, we can conclude that the observed
catalytic effect in the vapour generation was due to homogeneous
catalysis. In other words, the catalysis was achieved by Fe3+ rather
than Fe2+ or its atoms or atom clusters. A possible mechanism for
this catalysis may involve interaction of Cl orbital in MeHgCl with
120
4
Relative signal intensity

100
for the Hg measurement was more stable. However, for the deter-
mination of T-Hg, the use of 0.2 M HCl and water as the carrier 80
(ANS)

streams for the sample and reductant, respectively, was necessary.

60
3
3.2. Catalytic effect of TU and metal ions on MeHg reduction 40

Sensitivity enhancement of MeHg response by Cu2+ and TU 20 2

was observed with the use of NaBH4 as the reductant. How- 1
0
ever, the sensitivity enhancement caused by TU alone was
0 100 200 300 400 500
much less than with Cu2+ . Evidently, synergistic enhancement
was observed when both Cu2+ and TU were present together Fe3+ (mg/L)
in the sample solution. Optimum sensitivity enhancement was
Fig. 3. Influence of Fe3+ and NaBH4 concentrations on MeHg response. NaBH4 con-
obtained when TU was added into the NaBH4 solution, while centration (%): (1) 0.0001, (2) 0.001, (3) 0.01, and (4) 0.1. MeHg (in Hg) = 12 ␮g L−1 .
Cu2+ was present in the sample solution. Fig. 2 shows that (All the signal intensity results are relative values to that of Hg2+ as 100%.)
 Y. Zhang, S.B. Adeloju / Analytica Chimica Acta 721 (2012) 22–27
Table 1
Comparative results of the reduction of dialkyl mercury in absence and presence of
Fe(III) ions.
Hg species Absorbance
Fe3+ = 0 mg L−1 Fe3+ = 100 mg L−1
2+
Hg 0.059 0.059
MeHgCl 0.0142 0.059
DMeHg 0.0143 0.060
DBuHg 0.0092 0.042
MeEtHg 0.0102 0.050
MePhHg 0.0086 0.048
*0.1% NaBH4 was used for vapour generation.
the vacant hybridized d2 sp3 orbital of Fe3+ , i.e. donation of the 3p
electron pair of Cl atom to the vacant d2 sp3 orbital of Fe3+ . This delo-
calization of the 3p electrons of Cl atom may weaken the Me Hg
bond in the molecule and facilitate the breakdown of the bond and
the formation of volatile Hg atoms. An alternate mechanism is the
interaction between the 5d orbital of the Hg atom in MeHgCl and
the d2 sp3 hybrid orbital of Fe3+ , i.e. the delocalization of the 5d
electron pair in Hg atom and their donation to the vacant d2 sp3
orbital of Fe3+ . This delocalization can also result in the weakening
of the Me Hg bond in the MeHgCl molecule and facilitate the bond
breaking.
To obtain more evidence, different organic mercury species
were used for comparative study. Organomercurials studied
include DMeHg, MeEtHg, MePhHg and DBuHg. These organic mer-
cury species were analysed by CVAAS under different reduction
conditions. As shown in Table 1, reduction of all the dialkyl mer-
curials was catalysed by Fe3+ as evident by the higher sensitivity
obtained in the presence of Fe3+ . The reduction of DMeHg in the
presence of Fe3+ was much easier than for MeHg, achieving 90% of
its maximum sensitivity with the use of 0.0001% NaBH4 , while only
45% of the maximum sensitivity was achieved for MeHg under the
same conditions.
To obtain more evidence on whether any Lewis acid may act as
a potential catalyst for the vapour generation of MeHg, a group of
metal ions, including Ti3+ , Cr3+ , Mn2+ , Fe3+ , Fe2+ , Co2+ , Co3+ , Ni2+ ,
Cu2+ , Zn2+ , Cd2+ , Pb2+ , Au3+ , Pt4+ , Pd2+ and Rh3+ were investigated
for the vapour generation of organic mercury species, but only Fe3+
and Cu2+ exhibited catalytic effect. Two factors may be responsi-
ble for the result. First of all, a vacant orbital in the catalyst ion is
essential for accepting the electron pair from the Hg atom in organic
mercury molecule. Obviously, free metal ions can offer more vacant
orbitals than coordinated ions and this is determined by the coor-
dination strength of the metal ions with the ligands in the solution.
In this study, the major ligands present in the sample solution were
H2 O and Cl− . Most of the above metal ions are significantly coor-
dinated with H2 O molecules or Cl− ions. H2 O coordinate strongly
with Mn2+ , Fe2+ , Co2+ , Ni2+ and Zn2+ , while the coordination of Ti3+ ,
Cr3+ , Fe3+ , Co3+ and Cu2+ by H2 O is relatively weaker [70]. Cd2+ ,
Zn2+ , Au3+ , Pt4+ , Pd2+ and Rh3+ are strongly coordinated by Cl− [71].
This means that only Ti3+ , Cr3+ , Fe3+ , Co3+ and Cu2+ have consid-
erable proportion of vacant p orbital in the solutions used in this
study. Secondly, the catalytic effect may also be determined by the
strength of interaction between the substrate and the metal ion.
Fe3+ and Cu2+ may have strong interaction with Hg, while other
metal ions do not have considerable interaction with Hg and, hence,
their catalytic effect is negligible. Our attempt to investigate the
possible catalytic effect of Co3+ was not successful because it is
extremely unstable in aqueous solution in the absence of stronger
ligands.
The synergistic catalytic effect of Cu2+ and TU could be
attributed to the interaction between the S atom in TU molecule
and the Hg atom in the organomercuric compound which increased
25
the electron density of the Hg atom and, hence, enhanced the
delocalization of the 5d electrons of Hg besides the possible Hg–Cu
interaction. This assumption is in accordance with the fact that the
addition of Cu2+ and TU into the acidic sample solution had weaker
catalytic effect than that achieved by the addition of Cu2+ into the
sample solution and TU into the NaBH4 solution, respectively. In
the latter case, coordination between TU and Hg is stronger due to
the improved availability of the lone pair of the sulphur atom in
TU in alkaline medium.
The sensitivity enhancement by TU alone is assumed to be due to
the accelerated decomposition of MeHg by TU because such effect
was observed in our study when a MeHg solution was left overnight
in the presence of TU.
Munaf et al. [72] reported that the same vapour generation effi-
ciency was achieved for Hg2+ and MeHg by using SnCl2 in the
presence of Cu2+ or Cd2+ as the catalyst. However, it has been
shown that precipitation occurs in alkaline medium [61]. Our study
showed that in acidic fish liver extract, no absorption response was
obtained even for Hg2+ , possibly due to the masking effect of thiol
groups in the liver samples.
Serafimovski et al. [61] compared different results in the litera-
ture and presented their own results in the vapour generation and
determination of mercury species. They pointed out that the results
reported in the literature on the vapour generation of MeHg are
rather controversial. For example, one study [73] showed that sim-
ilar sensitivity was achieved for both In-Hg and MeHg during their
simultaneous determination even with atomization at room tem-
perature. However, a different study reported that equal calibration
slopes were achieved for both species only if elevated temperatures
(800–900 ◦ C) are employed for the atomization [74]. Interestingly,
there are also results showing that even with elevated atomiza-
tion temperatures, it is impossible to achieve the same sensitivity
for both In-Hg and MeHg [75]. Evidence has shown that MeHgH is
formed using relatively high NaBH4 concentrations or SnCl2 in the
presence of Cd2+ or Cu2+ and this species is stable enough during
the measurement [75,76]. However, as different hydride genera-
tion systems, chemical conditions and atomizers were used, it was
difficult to identify the important factors affecting the efficiency
atomization of the two mercury species. It was suggested that
kinetic factors during generation or release of MeHgH should be
taken into account in optimization studies as well as the efficiency
of the atomization of MeHgH using different atomizers [61]. Differ-
ent atomization modes including electrically heated quartz tube
(QT) at different temperatures and ICP-AES for the determination
of MeHg and In-Hg in different reaction media were compared. It
was shown that the NaBH4 concentration did not influence the effi-
ciency of MeHgH generation significantly. However, much lower
sensitivity was observed for MeHg compared with In-Hg in QT-AAS,
independent of the reductant concentration and reaction media.
Even at 900 ◦ C, isoformation of the behaviour of both Hg species was
not achieved and around 20–30% lower sensitivity was observed
for MeHg in both acidic and alkaline reaction media. In contrast,
very good leveling of the sensitivities of both In-Hg and MeHg was
achieved by ICP-AES detection [61]. This study suggested that ICP
ensures highly efficient atomization of MeHgH. It was also assumed
that batch system is appropriate for the determination of MeHg
since it ensures better mixing of sample and reductant solutions,
relatively long reaction time and more efficient release of MeHgH
[61].
These previous results show that MeHgH could be an interme-
diate in the vapour generation of MeHg. Therefore, the addition
of Fe3+ ions may facilitate the atomization of MeHgH in liquid
phase at the room temperature detection employed in this study.
In this case, delocalization of the 5d electrons of Hg atom is more
likely to occur because all the other atoms, except Hg in the
MeHgH molecule, have no valence electron pair for donation.
26 Y. Zhang, S.B. Adeloju / Analytica Chimica Acta 721 (2012) 22–27

Table 2
Analytical performance of the FI-CCV-AAS method for In-Hg and MeHg.

Parameter Optimum achievement

Sample consumption per assay (mL) 0.8

Sample throughput (h−1 ) 180
Linear regression fit of Hg2+ y = 0.0060x − 0.0012 R2 = 0.9981
Linear regression fit of MeHg (in Hg) y = 0.0059x − 0.0004 R2 = 0.9996
Detection limit of Hg2+ (3, ␮g L−1 ) 0.2
Detection limit of MeHg (in Hg, ␮g L−1 ) 0.2

Table 3
Influence of NaBH4 concentration on the recovery of T-Hg, Hg2+ and MeHg in fish
liver (DOLT-4) sample.

Form of Hg Found RSD (%, Certified Recovery

(␮g g−1 , n = 6) n = 6) valued (␮g g−1 ) (%)
Fig. 4. Dependence of MeHg extraction efficiency on HCl concentration.
With 1% NaBH4
T-Hg 2.36 ± 0.06 2.9 2.58 ± 0.22 91.5 ± 2.3
Nevertheless, another possibility that could not be ignored is that Hg2+ 1.00 ± 0.09 13.1 1.25a 80.0 ± 7.2
the interaction of metal with NaBH4 could produce borane-metal MeHg 1.36 ± 0.10 10.2 1.33 ± 0.12 102.3 ± 7.5
With 0.5% NaBH4
compounds and/or lead to the formation of intermediates metal- T-Hg 2.22 ± 0.06 3.7 2.58 ± 0.22 86.0 ± 2.3
borane-hydride, or hydride-metal with different reactivity which Hg2+ 1.00 ± 0.09 13.1 1.25a 80.0 ± 7.2
may act as catalysts in the hydride generation reactions [77,78]. MeHg 1.22 ± 0.11 13.1 1.33 ± 0.12 91.7 ± 8.3
With 0.1% NaBH4
T-Hg 1.92 ± 0.07 4.6 2.58 ± 0.22 74.4 ± 2.7
3.4. Extraction of MeHg and Hg2+ Hg2+ 1.00 ± 0.09 13.1 1.25a 80.0 ± 7.2
MeHg 0.92 ± 0.07 9.7 1.33 ± 0.12 69.2 ± 5.3
The extraction efficiency of MeHg and Hg2+ in DOLT-4 was inves- a
Difference between T-Hg and MeHg given by the supplier.
tigated by using 5 mL of different HCl concentrations. As shown in
Fig. 4, the extraction efficiency of MeHg increased with increas-
ing HCl concentration and close to 100% of the certified value of 3.6. Analytical performance and application
MeHg in the reference material was extracted with 1 M HCl or
higher. To ensure achievement of efficient extraction even when Table 2 gives the performance data of the FI-CCV-AAS methods
larger sample size was used, 2 M HCl was employed for real sam- for Hg and MeHg measurement. The data show that high sensitivity
ples in this study. However, the extraction of Hg2+ was much less and precision were achieved. A high sample throughput of 180 h−1
efficient under the same conditions. With increasing HCl concen- was achieved with a detection limit of 0.2 ␮g L−1 for both Hg and
tration, extraction efficiency of Hg2+ improved, but quantitative MeHg (as Hg).
extraction was difficult even when 6 M HCl was used. By adding Hg2+ and T-Hg concentrations in fish liver reference material
thiourea into the HCl, extraction of Hg2+ was significantly improved DOLT-4 were determined by using the extraction and determi-
as shown in Fig. 5. It was observed that 100% of the certified value nation procedures described above and MeHg concentration was
of Hg2+ in the reference material was extracted in the presence of calculated from the difference between T-Hg and In-Hg. The results
0.02 M TU in 5 mL of 2 M HCl with sonication for 5 min. in Table 3 demonstrate that the method gave satisfactory accuracy
and precision for T-Hg and MeHg. The results also demonstrate that
3.5. Inorganic interference the T-Hg and MeHg concentrations found increased with increasing
NaBH4 concentration and, thus, demonstrate the need for adequate
The addition of 10 mg L−1 of Cr6+ , Mn2+ , Co2+ , Ni2+ , Cu2+ , Zn2+ , reductant concentration. The relatively low recovery obtained for
Pb2+ and Cd2+ , and 400 mg L−1 Fe3+ into the standard solution had Hg2+ resulted possibly from organic interference in the samples.
no interference (signal variation less than 5%). These concentra- Nevertheless, a comparison of the concentrations found for differ-
tions are much higher than expected in fish samples and should, ent mercury species with those of the certified values shows a very
therefore, cause no problem. good agreement.

4. Conclusions

The vapour generation of Hg2+ and a mixture of Hg2+ and MeHg

Fig. 5. Dependence of Hg2+ extraction efficiency on TU concentration. HCl concen-
tration was 2 M.
was successfully performed by using 0.0001% NaBH4 and 0.1%
NaBH4 plus 100 ppm Fe3+ as catalyst, respectively. In standard solu-
tion measurement, with the use of 0.1% NaBH4 as the reductant and
100 ppm Fe3+ as the catalyst, MeHg gave the same vapour genera-
tion efficiency as Hg2+ and, thus, eliminated the need for separation
or oxidation. TU and Cu2+ had catalytic effect on the vapour gen-
eration individually, and synergistic catalytic effect was observed
when they were both present. The catalysis was suggested to be
homogeneous and the delocalization of the 5d electron pair of the
Hg atom in the MeHgCl molecule to permit coordination with Fe3+
as the electron acceptor is possibly responsible for the catalysis.
SnCl2 cannot reduce In-Hg or MeHg in the fish liver extract pos-
sibly due to the interference by thiol groups in the biomolecules.
The method was successfully applied to the quantitative analysis of
 Y. Zhang, S.B. Adeloju / Analytica Chimica Acta 721 (2012) 22–27 27

mercury species in a fish liver reference material DOLT-4, achiev- [39] L.H. Reyes, G.M.M. Rahman, H.M.S. Kingston, Anal. Chim. Acta 631 (2009) 121.
ing recoveries of 91.5% and 102.3% for total and methyl mercury, [40] M. Abuín, A.M. Carro, R.A. Lorenzo, J. Chromatogr. A 889 (2000) 185.
[41] S. Chen, A.S. Chou, D. Hwang, J. Chromatogr. A 1024 (2004) 209.
respectively. [42] M. Leermakers, W. Baeyens, P. Quevauviller, M. Horvat, Trends Anal. Chem. 24
(2005) 383.
References [43] C.E. Oda, J.D. Ingle Jr., Anal. Chem. 53 (1981) 2305.
[44] E.V. Alonso, M.T.S. Cordero, A.G. de Torres, P.C. Rudner, J.M.C. Pavón, Talanta 77
(2008) 53.
[1] S.J. Balogh, E.B. Swain, Y.H. Nollet, Environ. Pollut. 154 (2008) 3. [45] S. Río Segade, C. Bendicho, Ecotoxicol. Environ. Saf. B 42 (1999) 245.
[2] M. Gochfeld, Exotoxicol. Environ. Saf. 56 (2003) 174. [46] C. Fernandez, A.C.L. Conceicao, R. Rial-Otero, C. Vaz, J.L. Capelo, Anal. Chem. 78
[3] D. Karunasagar, M.V.B. Krishna, Y. Anjaneyulu, J. Arunachalam, Environ. Pollut. (2006) 2494.
143 (2006) 153. [47] M.V.B. Krishna, M. Ranjit, D. Karunasagar, J. Arunachalam, Talanta 67 (2005)
[4] L. Friberg, J. Vostal, Mercury in the Environment, CRC Press, Cleveland, OH, USA, 70.
1972. [48] P. Cava-Montesinos, A. Dominguez-Vidal, M.L. Cervera, A. Pastor, M. de la
[5] S. Havarinasab, P. Hultman, Autoimmun. Rev. 4 (2005) 270. Guardia, J. Anal. At. Spectrom. 19 (2004) 1386.
[6] Y. Cai, R. Jaffé, R. Jones, Environ. Sci. Technol. 31 (1997) 302. [49] J.L. Capelo, I. Lavilla, C. Bendicho, Anal. Chem. 72 (2000) 4979.
[7] H. Hintelmann, M. Hempel, R.D. Wilken, Environ. Sci. Technol. 29 (1995) 1845. [50] S. Río Segade, C. Bendicho, J. Anal. At. Spectrom. 14 (1999) 263.
[8] L. Ebdon, M.E. Foulkes, S.L. Roux, R. Muñoz-Olivas, Analyst 127 (2002) 1108. [51] S. Fragueiro, I. Lavilla, C. Bendicho, J. Anal. At. Spectrom. 19 (2004) 250.
[9] V. Celo, D.R.S. Lean, S.L. Scott, Sci. Total Environ. 368 (2006) 126. [52] K. Matsunaga, S. Takahashi, Anal. Chim. Acta 87 (1976) 487.
[10] J.B. Rivers, J.E. Pearson, C.D. Shultz, Bull. Environ. Contam. Toxicol. 8 (1972) 257. [53] J.J. Bisogni, A. Lawrence, Environ. Sci. Technol. 8 (1974) 850.
[11] U. Harms, Mikrochim. Acta 109 (1992) 131. [54] D. Beauchemin, K.W.M. Siu, S.S. Berman, Anal. Chem. 60 (1988) 2587.
[12] J.L. Gómez-Ariza, F. Lorenzo, T. García-Barrera, Chemosphere 61 (2005) 1401. [55] M.V.B. Krishna, D. Karunasagar, S.V. Rao, J. Arunachalam, Talanta 68 (2005)
[13] Z. Jókai, L. Abrankó, P. Fodor, J. Agric. Food Chem. 53 (2005) 5499. 329.
[14] L. Yang, Z. Mester, R.E. Sturgeon, J. Anal. At. Spectrom. 18 (2003) 431. [56] G.A. Zachariadis, A.N. Themidis, E.I. Daftsis, J.A. Stratis, J. Anal. At. Spectrom. 20
[15] J. Qvarnström, L. Lambertsson, S. Havarinasab, P. Hultman, W. Frech, Anal. (2005) 63.
Chem. 75 (2003) 4120. [57] W. Jian, C.W. Mcleod, Talanta 39 (1992) 1537.
[16] M. Logar, M. Horvat, I. Falnoga, V. Stibilj, Fresenius J. Anal. Chem. 366 (2000) [58] H. Wu, Y. Jin, W. Han, Q. Miao, S. Bi, Spectrochim. Acta B 61 (2006) 831.
453. [59] S. Río Segade, J.F. Tyson, J. Anal. At. Spectrom. 18 (2003) 268.
[17] Y. Cai, S. Monsalud, K.G. Furton, Chromatographia 52 (2000) 82. [60] S. Río Segade, J.F. Tyson, Spectrochim. Acta B 58 (2003) 797.
[18] K.C. Bowles, S.C. Apte, Anal. Chim. Acta 419 (2000) 145. [61] I. Serafimovski, I. Karadjova, T. Stafilov, J. Cvetković, Microchem. J. 89 (2008)
[19] P. Quevauviller, M. Filippelli, M. Horvat, Trends Anal. Chem. 19 (2000) 157. 42.
[20] H.E.L. Palmieri, L.V. Leonel, Fresenius J. Anal. Chem. 366 (2000) 466. [62] S. Río Segade, J.F. Tyson, Talanta 71 (2007) 1696.
[21] H.E.L. Armstrong, W.T. Corns, P.B. Stockwell, G. O’Connor, L. Ebdon, E.H. Evans, [63] C.C.Y. Chan, R.S. Sadana, Anal. Chim. Acta 282 (1993) 109.
Anal. Chim. Acta 390 (1999) 245. [64] P. Rivaro, C. Ianni, F. Soggia, R. Frache, Microchim. Acta 158 (2007) 345.
[22] B. He, G. Jiang, Z. Ni, J. Anal. At. Spectrom. 13 (1998) 1141. [65] A.da C.P. Monteiro, L.S.N. de Andrade, R.C. de Campos, Fresenius J. Anal. Chem.
[23] A.M. Caricchia, G. Minervini, P. Soldati, S. Chiavarini, C. Ubaldi, R. Morabito, 371 (2001) 353.
Microchem. J. 55 (1997) 44. [66] A.da C.P. Monteiro, L.S.N. de Andrade, A.S. Luna, R.S. de Campos, Spectrochim.
[24] L. Liang, M. Horvat, E. Cernichiari, B. Gelein, S. Balogh, Talanta 43 (1996) 1883. Acta B 57 (2002) 2103.
[25] E. Bailey, A.G.F. Brooks, Mikrochim. Acta 109 (1992) 121. [67] C.S. Marvel, V.L. Gould, JACS 44 (1922) 153.
[26] M. Horvat, K. May, M. Stoeppler, A.R. Byrne, Appl. Organomet. Chem. 2 (1988) [68] Y. Zhang, S.B. Adeloju, Talanta 74 (2008) 951.
515. [69] Y. Zhang, S.B. Adeloju, Talanta 76 (2008) 724.
[27] R. Ito, M. Kawaguchi, N. Sakui, N. Okanouchi, K. Saito, Y. Seto, H. Nakazawa, [70] J.J. Lagowski, Modern Inorganic Chemistry, Marcel Dekker Inc., New York, USA,
Talanta 77 (2009) 1295. 1973, p. 595.
[28] C.F. Harrington, Trends Anal. Chem. 19 (2000) 167. [71] J.J. Lagowski, Modern Inorganic Chemistry, Marcel Dekker Inc., New York, USA,
[29] Y. Yin, J. Liu, B. He, J. Shi, G. Jiang, J. Chromatogr. A 1181 (2008) 77. 1973, p. 668.
[30] T. Hashempur, M.K. Rofouei, A.R. Khorrami, Microchem. J. 89 (2008) 131. [72] E.H. Munaf, H. Haraguchi, T. Takeuchi, Anal. Chim. Acta 243 (1991) 247.
[31] I. Cattani, S. Spalla, G.M. Beone, A.A.M. Del Re, R. Boccelli, M. Trevisan, Talanta [73] S. Kagaya, Y. Kuroda, Y. Serikawa, K. Hasegawa, Talanta 64 (2004) 554.
74 (2008) 1520. [74] D.P. Torres, M.A. Vieira, A.S. Ribeiro, A.J. Curtius, J. Anal. At. Spectrom. 20 (2005)
[32] L. Xia, B. Hu, Y. Wu, J. Chromatogr. A 1173 (2007) 44. 289.
[33] A.J. Percy, M. Korbas, G.N. George, J. Gailer, J. Chromatogr. A 1156 (2007) 331. [75] J. Gutiérrez, H. Travieso, M. Pubillones Aurora, Water Air Soil Pollut. 68 (1993)
[34] S.C. Hight, J. Cheng, Anal. Chim. Acta 567 (2006) 160. 315.
[35] R. Rai, W. Maher, F. Kirkowa, J. Anal. At. Spectrom. 17 (2002) 1560. [76] P.J. Craig, R.O. Jenkins, G.H. Stojak, Chemosphere 39 (1999) 1181.
[36] S. Río-Segade, C. Bendicho, Talanta 48 (1999) 477. [77] E. Pagliano, M. Onor, J. Meija, Z. Mester, R.E. Sturgeon, A. D’Ulivo, Spectrochim.
[37] E. Ramalhosa, S. Río Segade, E. Pereira, C. Vale, A. Duarte, Anal. Chim. Acta 448 Acta, Part B 66 (2011) 740.
(2001) 135. [78] A. D’Ulivo, J. Dědina, Z. Mester, R.E. Sturgeon, Q. Wang, B. Welz, Pure Appl.
[38] C. Sarzanini, G. Sacchero, M. Aceto, O. Abollino, E. Mentasti, Anal. Chim. Acta Chem. 83 (2011) 1283.
284 (1994) 661.