American Journal of Analytical Chemistry, 2013, 4, 623-632
Published Online November 2013 (http://www.scirp.org/journal/ajac)
Open Access AJAC
The Challenge and Its Solution When Determining
Biogeochemically Reactive Inorganic Mercury (RHg):
Getting the Analytical Method Right
Lian Liang1*, Milena Horvat2, John Alvarez3, Lyman Young3, Jože Kotnik2, Lisa Zhang1
1Cebam Analytical, Inc., Bothell, USA
2Department of Environmental Sciences, J. Stefan Institute, Ljubljana, Slovenia
3Chevron Energy Technology Company, Richmond, USA
Received September 8, 2013; revised October 15, 2013; accepted October 28, 2013
Copyright © 2013 Lian Liang et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Biogeochemially reactive inorganic mercury (RHg) is an important fraction of Hg. Researchers have attempted to
measure RHg when characterizing Hg-impacted sites, conducting research and development of remediation practices, or
evaluating remediation efficiency. In these uses, RHg will be the best choice for analysis in ways that total methyl, and
other species of Hg cannot duplicate. The fraction has been inadequately measured using the Sn2+ reduction method and
operationally defined as “Sn2+ reducible Hg2+”, but the resulting data did not reflect well the nature of the fraction and
caused researchers to lose interest, thus limiting the use of RHg in past years. In this work, the problems of using the
Sn2+ reduction method were discovered to be generating irreproducible and negatively biased results. Negative bias
from 20% to 99% was found in different types of waters. To obtain reliable results, an ethylation-based GC-CVAFS
method was used to determine RHg. The performance of the method was evaluated by comparing it to the Sn2+ reduc-
tion method. Biogeochemically meaningful results have been obtained in the application of the method to determine
RHg in mercury mine-impacted waters from the Idrijca River in Slovenia.
Keywords: Mercury Speciation; Reactive Hg2+; Method Comparison; Ethylation-GC-CVAFS; Sn2+ Reduction
Methyl mercury (MeHg) is an important species due to
its persistence, bioaccumulation and biomagnification,
and because of its toxicity to man and ecosystems [1-5].
In past decades, MeHg and the methylation/demethyla-
tion process have drawn significant attention from scien-
tists worldwide. Because of its high solubility, mobility,
bioavailability, and especially methylation potential, bio-
geochemically reactive inorganic mercury (RHg) has
been recognized to be an important fraction of Hg for the
transformation, reaction, and methylation of Hg in bio-
geochemical cycling [6-9]. For characterization and as-
sessment of Hg impacted sites, research and development
of remediation practices, and evaluation of remediation
efficiency, RHg may be the best choice for analysis in
ways that total methyl, and other species of Hg cannot
duplicate. However, unlike MeHg, RHg measurement
has not gained widespread acceptance to match its im-
portance, even after over two decades of practice .
What is the challenge, and what is the solution?
In the past, in addition to “reactive Hg”, RHg was also
defined or described as “easily reducible Hg”, “Sn2+ re-
ducible Hg”, “acid-labile Hg”, “ionic Hg”, “labile inor-
ganic Hg”, and “bio-available-inorganic Hg” [10,11].
The names related to “Sn2+ reducible” are misleading
because RHg can be analyzed by other methods such as
the ethylation-based GC-CVAFS method that was the
technique used in this work [12-14]. The mechanism of
the method is to identify and quantify the fraction of
Hg2+ using the ethylation reaction, resulting in ethylat-
able Hg2+, a proxy for the fraction that is available for
abiotic and biotic methylation in an aqueous ecosystem.
Those terms related to acid leaching  are also mis-
leading because with acid leaching the measured fraction
is no longer the naturally-occurring RHg. In this work,
RHg is defined as “Hg2+ that readily enters into chemical
L. LIANG ET AL.
reactions” , and measured analytically in natural wa-
ters untreated with any chemicals including acidification
Careful selection of an appropriate analytical method
is the key to ensuring meaningful results. In the past, the
classic Sn2+ reduction method  has been virtually the
only method used for determination of RHg [10,11], so
RHg was called “Sn2+ reducible Hg”。 There is no doubt
that the classic Sn2+ reduction method is a reliable
method for determination of Hg2+ in chemically pre-
treated media such as acid/alkaline digestates or BrCl/
KMnO4 oxidates. However, problems were encountered
trying to measure RHg in waters that were not chemi-
cally treated prior to Sn2+ reduction. To gain insight into
these problems, this work focused on investigating the
effects of analytical conditions on results from the Sn2+
reduction method in comparison with those obtained us-
ing an ethylation-based GC-CVAFS method. Since the
methodology of the ethylation-based GC-CVAFS me-
thod has been detailed in previous publications [13,14],
and successfully used to determine Hg2+ in alkaline di-
gestates of biota and blood samples [17,18], the advan-
tage of using this new approach for RHg was demon-
strated by the discovery of bias in the Sn2+ reduction
method for the determination of RHg in environmental
waters, and examining whether the data generated was
more indicative of the RHg, compared to using the Sn2+
RHg-related studies conducted in past years focused
on fresh waters with low levels of Hg , or waters
already treated for removal of Hg . Since RHg is an
important and useful fraction to characterize Hg-im-
pacted sites and evaluate the results of remediation, the
present work includes analyses of both fresh waters with
Hg concentration < 12 ng/L (the lowest Nationwide Cri-
terion, 40 CFR 131.36) and Hg-impacted waters includ-
ing industrial waste waters. Moreover, as RHg is the
fraction directly linked to methylation and bioavailability,
RHg was previously measured mostly in filtered samples,
and defined as dissolved RHg when estimating bioavail-
able Hg [11,15]. However, methylation can take place
with RHg in both filtered waters and particles, so in this
work both filtered and unfiltered waters were studied.
To further examine whether the ethylation method can
generate meaningful RHg results to characterize Hg-
impacted waters, the method was applied to measure-
ment of RHg in Hg mine-impacted waters from the Idri-
jca River in Slovenia, and results were compared with
those for total mercury (THg) and methyl mercury
(MeHg) to assess their biogeochemical significance.
2.1. Sample Collection and Process
Water samples were collected in glass bottles with Tef-
lon-lined caps. Special attention was paid to avoiding
contamination according to the procedures described in
EPA Method 1669 . For dissolved fractions, the
samples were filtered in the field immediately after sam-
ple collection using 0.45 um disposable cellulose nitrate
vacuum filter units. Unpreserved samples were packed in
coolers at 0˚C - 4˚C and shipped overnight to the labora-
tory in Bothell, WA, USA. In most cases, the lab ana-
lyzed the samples within 48 hours after receipt. Sample
bottles were shaken vigorously prior to sub-sampling.
2.2. Instrumentation, Standards, and Reagents
For standards and reagents, only those specifically pre-
pared for this work are addressed here, while others not
mentioned are detailed in EPA method 1631  or 1630
A cold vapor atomic fluorescence spectrometry (CVAFS)
analytical system utilizing a BR III Hg analyzer (Brooks
Rand, Seattle, USA) was employed for determination of
Sn2+ reduced Hg0 [20,22]. A lab-built GC/CVAFS system
with a BR III Hg analyzer was used for separation and
detection of ethylation derivatives of MeHg or RHg
[14,21]. An AB15 pH meter (Accumet) was used for pH
A 1000 ug/mL Hg2+ stock standard was prepared by dis-
solving 0.1353 g HgCl2 (Sigma-Aldrich) into 100 mL of
5% HNO3. A working standard of 10 ng/mL Hg2+ in
0.2% HNO3 was prepared from the stock solution by di-
A 0.5 M citrate buffer was prepared by dissolving 270 g
of reagent grade citric acid (C6H8O7·H2O) and 370 g of
reagent grade sodium citrate (C6H5Na3O7·2H2O) into
DDW to make a 2.5-L solution.
2.3. Analytical Procedures
In what follows, RHg determined by the ethylation me-
thod is designated “EtRHg”, while that determined by
Sn2+ reduction is designated “SnRHg”.
2.3.1. Determination of EtRHg
Appropriate aliquots up to 40 mL of unpreserved sample
were placed into bubblers, depending on the RHg con-
centration, and distilled, deionized water (DDW) was
added to bring the total volume to 50 mL. 1 mL of 0.5 M
citrate buffer was added, followed by 0.05 mL of freshly
thawed 1% NaBEt4 solution . The bubblers were
immediately capped and shaken. The bubblers sat at
Open Access AJAC
L. LIANG ET AL. 625
room temperature for 20 min as the ethylation reaction
proceeded. Then the samples were purged with N2 at 200
mL/min for 20 min. The ethylation derivatives purged
from the bubblers were collected onto Tenax traps, and
finally analyzed by GC/CVAFS [14,21]. The 10 ng/mL
Hg2+ working standard was used for calibration. The
method detection limit (MDL) is 0.05 ng/L.
2.3.2. Determination of SnRHg
Appropriate aliquots up to 100 mL of unpreserved sam-
ple were placed into bubblers, depending on the RHg
concentration. DDW was added to bring the total volume
to 100 mL. 0.2 mL of 25% SnCl2 (w/v) in 20% HCl was
added. The samples were purged with N2 at 300 mL/min
for 20 min to remove Sn2+ reduced Hg0. The purged Hg0
was collected on gold traps and then analyzed by CVAFS
[20,22]. The MDL is 0.2 ng/L.
2.3.3. Determination of THg
Samples were oxidized with BrCl overnight, and excess
BrCl was reduced with NH2OH solution prior to analysis
[20,22]. Appropriate sample aliquots depending on THg
concentration were placed into bubblers, and DDW was
added to adjust the volume to 100 mL. Samples were
then purged and Hg0 was quantified as described above
for the SnRHg procedure. The MDL is 0.2 ng/L.
2.3.4. Determination of MeHg
Samples were acidified to 4% HNO3, then allowed to sit
overnight or longer at room temperature prior to sample
preparation and analysis using EPA 1630 . The MDL
is 0.02 ng/L.
2.4. Experiments Examining the Effects of
Analytical Conditions on the Sn2+
2.4.1. The Effect of pH
Fourteen 500-mL bottles were prepared, each containing
400 mL DDW and 0.8 mL 25% SnCl2 (w/v, in 20% HCl).
Diluted NaOH and HCl solutions were used to adjust the
pH of these samples within a range of 2 to 13. From each
bottle, 3 aliquots of 100 mL each were transferred into 3
pre-cleaned bubblers. Two of the bubblers were spiked
with 280 pg of Hg2+ standard in 0.04 mL 0.2% HNO3
solution, while the third bubbler served as a blank. Sam-
ples in the bubblers were analyzed for Sn2+ reducible Hg
by purge/trap/CVAFS. The pH of each sample was
measured after purging.
2.4.2. The Effect of SnCl2 Concentration
A 2.0 L glass volumetric flask was half-filled with DDW,
and 2 mL of concentrated HNO3 and 5.6 ng of Hg2+ stan-
dard were added. The flask was then filled to the neck
with DDW and shaken. The Hg2+ concentration in this
sample was 2.8 ng/L. Aliquots of 100 mL were placed
into bubblers, and then various volumes (0.05, 0.10, 0.20,
0.40, 0.60, 0.80, and 1.00 mL) of 25% SnCl2 (w/v, in
20% HCl) were added, corresponding to SnCl2 concen-
trations of 0.013, 0.025, 0.050, 0.100, 0.150, 0.200, and
0.250% (w/v). These samples were then analyzed for
Hg2+ by purge/trap/CVAFS. Analyses were performed in
duplicate for each SnCl2 concentration.
2.4.3. The Effect of Purge Gas Rate and Purge Time
Municipal waste water was collected in a 4-L glass bottle.
Observable particles in the water were removed by filtra-
tion using thin paper napkins. Filtration blanks showed
no significant Hg concentration. Concentrations of
EtRHg and THg in the sample were determined at the
optimal conditions of the methods, and found to be 13.54
± 0.62 ng/L (n = 6) and 34.7 ± 1.4 ng/L (n = 6), respec-
tively. The sample was then analyzed for THg and
SnRHg at various purge gas flow rates and purge times.
The analysis was performed in duplicate under each ana-
lytical condition, and the mean value of duplicate results
was calculated. Recoveries of THg at various conditions
were calculated relative to the result (34.7 ng/L) deter-
mined at optimal conditions, while SnRHg recoveries
were calculated relative to the optimal EtRHg result
2.4.4. The Effect of Purge Time on Recovery of
SnRHg in River Waters with Low Level Hg
Twelve river water samples were collected, and each was
analyzed for SnRHg using two different purge times (20
min and 40 min). These samples were visually clear (to-
tal suspended solids (TSS) < 4 mg/L), and THg concen-
trations ranged from 1 to 8 ng/L. Unfiltered and unpre-
served waters were analyzed. EtRHg concentrations in
these waters were also determined, and the results were
used as benchmarks for calculating the recoveries of
SnRHg for different purge times. All samples were ana-
lyzed in duplicate for both EtRHg and SnRHg, while
matrix spike (MS) measurements were performed on
each sample only for EtRHg analysis, to examine matrix
2.4.5. The Effect of Purge Time on Recoveries of
SnRHg in Hg Impacted Waters
Twelve industrial waste water samples collected from a
waste treatment site were used for the experiment. These
samples looked clear (TSS < 5 mg/L), and THg concen-
trations ranged from 5 to 100 ug/L. Unpreserved and
unfiltered samples were analyzed for SnRHg using two
different purge times (20 min and 48 hours). These sam-
ples were also analyzed for EtRHg, and the EtRHg con-
centrations were used for calculating SnRHg recoveries.
All samples were analyzed in duplicate for both EtRHg
Open Access AJAC
L. LIANG ET AL.
and SnRHg, while MS measurements were again per-
formed only for EtRHg analyses.
2.4.6. The Behavior of Purging SnRHg from a Typical
A Hg-impacted industrial waste water sample was used
for the experiment. The sample looked dirty and con-
tained some suspended solids. The sample was shaken
vigorously prior to sub-sampling, then analyzed for THg,
MeHg, and EtRHg; concentrations were 6.16 ug/L, 2.96
ng/L, and 2.97 ug/L, respectively. 3 mL of well-mixed,
unpreserved, and unfiltered water was added to 50 mL
DDW in a bubbler for reaction with 0.2 mL of SnCl2
(25%, w/v, in 20% HCl). The 20 min purge/trap/mea-
surement procedure was then repeated over and over, one
time after another. After 24 cycles, with Hg0 still coming
out of the sample and no idea how many more cycles
would be needed, the bubbler was purged for 4 hours for
the 25th cycle. Finally, the 26th cycle of 20 min showed
no detectable Hg0, suggesting that RHg and Hg0 had been
2.4.7. The Stability of RHg in Unpreserved Waters
Six unpreserved river water samples with various con-
centrations of RHg were collected in 1-L glass bottles
and shipped to the lab overnight, on ice. The samples
looked clear (TSS < 7 mg/L), suggesting the sample ma-
trices might not be complex. The samples were allowed
to warm to 20˚C after receipt, and then analyzed for
EtRHg. The analyses were carried out multiple times for
each sample, at 4, 24, 48, and 72 hours after sample re-
ceipt. All analyses were performed in duplicate and the
mean value of duplicate results was calculated. Samples
were well-mixed prior to sub-sampling.
3. Results and Discussion
3.1. Effect of Analytical Conditions on the
Results for Hg2+ Using the Sn2+ Reduction
The effects of pH and SnCl2concentration on the results
of the Sn2+ reduction method are shown in Figure 1. It is
worth noting here that the analyte is Hg2+ rather than
RHg because this experiment is designed to examine
whether pH or the SnCl2 concentration effect results
during Sn2+ reduction. Here, the reaction matrix was
DDW rather than natural water, and the Hg2+ was not
naturally occurring RHg. For the effect of pH, each point
is the mean of the results from two duplicate samples.
The relative percent difference (RPD) between duplicate
samples was <5% for 14 samples at various pH levels.
The average recovery of Hg2+was 100.8% ± 2.7% (n =
14), indicating the results are independent of pH. For the
effect of the SnCl2 concentration, average recovery was
Figure 1. The effect of pH and SnCl2 concentration on re-
sults of Hg2+.
100.4 ± 2.23% (n = 7) with RSD = 2.2%, indicating that
the SnCl2 concentration in the range tested does not af-
fect the results for Hg2+. The results shown in Figure 1
clearly indicate that the “chemically-related factors” of
pH and SnCl2 concentration do not affect the Hg2+ results.
In contrast, it was found that the “physical factors” of
purge gas flow rate and purge time critically affect the
results (Figure 2).
For the effect of gas flow rate shown in Figure 2, for
THg, the recovery increased with increasing gas flow
rate before reaching 100% recovery at flow rates be-
tween 260 to 370 mL/min. 370 mL/min is the highest
useable flow rate without breakthrough using our gold
traps (Cebam Analytical, Inc.), then recoveries decreased
at higher flow rates; for SnRHg, the recovery also in-
creased with increasing gas flow rate, but the highest
recovery was found to be only around 40%, and lower
values were recorded with further increases in flow rate,
again due to breakthrough. This indicated that for SnRHg
analysis with a purge time of 20 min at the highest flow
rate allowed, only 40% of the SnRHg in this sample was
recovered; i.e., the result was negatively biased by 60%.
Because there is no range of flow rates at which a plateau
in the recovery data is established, it would be difficult to
obtain reproducible results for replicate analyses. This
Open Access AJAC
L. LIANG ET AL. 627
Figure 2. Effect of purge gas flow rate (purge time 20 min),
and purge time on Hg recovery (purge flow rate 350 mL/
explains why previous results using the Sn2+ reduction
method were found to be irreproducible and imprecise.
For the effect of purge time (Figure 2), 100% recov-
ery of SnRHg can be obtained only when the sample is
purged for long enough and insufficient purge time leads
to lower results for SnRHg. This raises the question of
why 20 min was enough to purge the Sn2+-reduced Hg0
for analysis of THg (reaching 100% recovery), but a
longer time was required for analysis of SnRHg using the
same technique. The only difference in the analytical
procedures between THg and SnRHg is that for THg,
Sn2+ reduction was carried out in samples which had
been oxidized (or digested) with oxidants such as BrCl
, KMnO4, H2SO4, HNO3 (used by many EPA THg
methods), or others, while for SnRHg, the samples were
unpreserved and untreated chemically prior to Sn2+ re-
duction. It was also found that purging Sn2+-reduced Hg0
from DDW required as much purge time as the oxi-
dized/digested medium. Thus, recovering Sn2+-reduced
Hg0 from an unpreserved/undigested medium required a
longer purge time than from an oxidized/digested me-
dium or from DDW. The reason for this will be discussed
3.2. The Effect of Purge Time on Results of
SnRHg Analysis of River and Industrial
To confirm the effect of purge time on the results of
SnRHg analysis, more natural river waters with low Hg
concentrations and Hg-impacted industrial waste waters
with high Hg concentrations were analyzed by the Sn2+
reduction method. Figure 3 shows the effect of purge
time on SnRHg results for 12 river water samples. For
these relatively clean waters, a 20-min purge recovered
about 80% of the RHg, i.e. a negative bias of about 20%.
For filtered natural waters, similar patterns were ob-
served. However, for the highly contaminated industrial
waste waters with complex matrices (Figure 4), for a
purge time of 20 min recovered only very small fractions
(0.7% to 1.9%) of the RHg, but after purging for 48
hours, 100% recoveries were reached. This reveals how
severely biased previous results may have been when the
Sn2+ reduction method was used for Hg-impacted waters.
100% recovery can only be reached when the sample is
purged for a long enough time. How can we tell if the
RHg is recovered completely by this method? By purg-
ing a sample for a cycle time such as 20 min, the Hg col-
lected on the trap may be measured, and then a new trap
is used for the next purge cycle. This is continued until
the Hg collected on a trap is not detectable. Then the Hg
loaded on all the traps is summed to yield the result for
RHg in the sample.
Figure 3. Effect of purge time on recoveries of SnRHg in 12
river water samples.
Open Access AJAC
L. LIANG ET AL.
Figure 4. The effect of purge time on the recoveries of
SnRHg in 12 Hg impacted industrial waster waters.
3.3. Purging Sn2+-Reduced Hg0 from Industrial
Waste Waters Having High Hg
Concentrations and Complex Matrices
To get clearer insight into the purging behavior of Sn2+-
reduced Hg0 another experiment was conducted using an
Hg-impacted industrial waste water sample (Figure 5).
The sample was purged for 26 cycles as described above
and the concentration of RHg in the sample was calcu-
lated by summing the mass of RHg recovered over all
cycles; the result was found to be 2.99 ug/L.
Here, it is worth emphasizing that the concentration of
RHg by Sn2+ reduction was found to be equivalent to the
EtRHg concentration in the sample, i.e., SnRHg ≈ EtRHg.
This confirmed the finding in the low level Hg samples
(Figure 3), and the Hg-impacted samples (Figure 4), as
well as other samples investigated when samples were
purged for long enough: The two independently-defined
quantities based on two different chemical reactions,
ethylation and Sn2+ reduction, were found to represent
the same fractions, RHg.
Further consideration of the results in Figure 5 show:
1) the sum of RHg purged from the 1st to the 24th cycle
was 6707 pg, the average was 279 pg per cycle, and the
RSD was 7.7%, indicating that Sn2+-reduced Hg0 was
purged out steadily over time; and 2) the mass of RHg
purged in the 25th cycle was 2270 pg, and, according to
the average rate of the first 24 cycles, 3 hours would
have been enough for the 25th cycle. Again, the relation-
ship SnRHg ≈ EtRHg can be established only when the
sample is purged long enough to ensure the Sn2+-reduced
Hg0 is purged out completely.
The behavior shown in Figure 5 was similarly found
in purging naturally occurring Hg0 from various Hg-im-
pacted waste waters, and also found by Horvat  in
measurements of Hg0 evasion in Hg-impacted river wa-
ters. The analysis of Hg0 requires only the purge/trap step
prior to CVAFS detection; there were no chemical reac-
tions/treatments involved. This may suggest that the long
purge time required was for the isolation of Hg0 from the
sample medium, regardless of any chemical processes.
Based on the nature of Hg0, it may be adsorbed to or-
Figure 5. Behavior of purge 8977 pg of Hg0 reduced by Sn2+
from 3 mL Hg contaminated water, the 25th purge cycle is 4
hours, other 20 min.
ganic/inorganic substances in the medium, and this bind-
ing makes purging the Hg0 difficult. Conversely, the
purging of Hg0 from DDW requires less time compared
to unpreserved/undigested water because Hg0 occurs free
of binding in DDW. An additional or alternative expla-
nation for this observation is that the binding between
RHg and natural organic material (NOM) makes RHg
resistant to reduction by Sn2+ [24-26].
Moreover, as mentioned above, isolating Sn2+-reduced
Hg0 from DDW required the same purge times as from
oxidized/digested medium. Thus, Hg0 is also present free
of binding in oxidized/digested medium, suggesting that
this apparent binding mechanism can be destroyed by
oxidation/digestion, setting Hg0 free in the medium. This
also explains the difference in recoveries between THg
and RHg in Figures 4 and 5, where THg can be fully
recovered easily by a 20 min purge, but SnRHg cannot.
In addition to the negative bias described above, there
is potentially a positive bias when waters contain natu-
rally occurring Hg0 or dimethyl Hg. It is clear that these
species may also be purged, collected, and measured
together with Sn2+-reduced Hg0. This positive bias will
be discussed below.
3.4. Matrix Interference in the Ethylation
Matrix interference with the ethylation reaction is a po-
tential problem in the determination of MeHg in various
matrices [17,18,21], and the same applies to RHg. This is
why distillation or other techniques were developed for
the isolation of MeHg from matrices prior to ethylation
[27,28]. However, existing techniques for the isolation of
MeHg cannot be applied for the isolation of RHg because
the nature of this fraction will change during processing.
Fortunately, matrix interference of the ethylation reaction
was found to be easily eliminated by simply diluting the
samples. The ratio of RHg to MeHg generally ranges
from 10 to 10n depending on the site, and the more im-
pacted the site is, the greater the value of n. This allows
Open Access AJAC
L. LIANG ET AL. 629
diluted samples to be analyzed without matrix interfer-
ence [29,30]. The lowest RHg concentrations were found
to be >0.5 ng/L, which is about 10 times the MDL, 0.05
ng/L. Matrix interference was determined by analyzing
samples spiked with Hg2+ standard. Spike recoveries
ranging within 75% to 125% were considered to show no
matrix interference. In the experiments on the effect of
purge time on the recovery of SnRHg, a MS sample was
analyzed for each sample for determination of EtRHg.
MS recoveries for 12 unfiltered river samples were found
to range from 83% to 107%, while recoveries of 91% to
112% were found for 12 industrial waste water samples.
The industrial waste waters were analyzed after dilution
by up to 1000 or more times.
3.5. Stability of RHg in Unpreserved Water
The results shown in Figure 6 indicate that RHg was
stable at 20˚C for at least 72 hours after sample receipt. If
the sample shipping time was 24 hours, the samples
would thus have been stable for 96 hours after collection.
The samples used here were relatively clear, with TSS <
7 mg/L, suggesting that the sample matrices were not
complex. Since the stability of RHg may depend strongly
on various biogeochemical factors, its relative stability in
these water samples for 96 hours does not necessarily
ensure its stability in other samples. To obtain reliable
results, water samples should be analyzed as soon as
possible after sample collection. Freezing of samples is
not recommended because adsorption of RHg on the
container wall could increase with decreasing tempera-
Generally, water samples for analysis of trace metals
are preserved with acids such as HNO3 and HCl to main-
tain the stability of the analytes for a longer time. In past
years, water samples analyzed for RHg were sometimes
acidified with HCl prior to analysis [31-34]. However, to
maintain the original state of the RHg, water samples
Figure 6. Observation of stability of RHg in unpreserved
surface water at 20˚C.
should not be preserved. Nevertheless, waters acidified to
4% HNO3 or HCl were also investigated in this work,
and results around 30% lower than those found in unpre-
served waters were observed. This result is similar to that
reported by Bloom . MeHg was also involved in
these experiments, but in contrast to RHg, higher results
for MeHg were found in acidified waters.
3.6. Application of the Ethylation Method to the
Determination of RHg in Hg-Impacted
Waters from the Idrijca River, Slovenia
The Idrijca River in Slovenia drains the area of the Idrija
Hg mine, now closed but formerly the world’s second
largest Hg mine. Five sampling locations were selected
along the Idrijca River (Figure 7), representative of dif-
ferent levels of Hg impact as determined by previous
studies [15,35-39]. The first location, above the Belca
inflow, is a pristine site with very low Hg contamination,
while all the downstream sampling sites are affected by
the Hg mine to some extent . The second site, in the
town of Idrija, is where the mine drainage enters the river.
The third location, above the town of Spodnja Idrija, is
situated approximately 200 m downstream of the outflow
of a municipal wastewater treatment plant. The fourth
site, Kozarska grapa, is approximately 15 km down-
stream from Spodnja Idrija in a rural area; and Bača pri
Modreju, the final sampling site, is about 1 km upstream
of the confluence of the Idrijca River with the Soča River
(Isonzo). The pH in the river is between 7.73 and 8.82,
and the solute composition is dominated by 3
and Mg2+ . Water temperatures range from about 6˚C
in the winter to 18˚C in summer.
Monitoring RHg in the Idrijca River was carried out to
demonstrate the advantages of the ethylation based
method. In the past, Sn2+ reduction has been the only
method for measuring RHg at the site. Since the river-
drains the area of the former Hg mine, the water contains
a significant amount of Hg0 . To eliminate the posi-
tive bias caused by naturally occurring Hg0, the samples
Figure 7. Sampling locations on the river Idrijca.
Open Access AJAC
L. LIANG ET AL.
were pre-purged prior to adding the Sn2+ reagent. How-
ever, purging for 10 to 20 min might not be enough to
completely remove naturally occurring Hg0 from highly
contaminated samples like these (THg up to 650 ng/L),
and a positive bias might result. In contrast to this posi-
tive bias, negative bias might be introduced in the sub-
sequent Sn2+ reduction of RHg due to an insufficient
purge of Sn2+-reduced Hg0. As a result, the data might be
highly variable and uncertain due to both positive and
negative bias effects. When the ethylation based method
was used, neither positive nor negative bias effects were
To examine whether the ethylation method can gener-
ate results meaningful for the characterization of Hg
speciation in this river, water samples from the five loca-
tions were collected in the summer of 2012. Filtration
was performed immediately after sample collection at all
sites. Both filtered and unfiltered samples were analyzed
for THg, MeHg, and EtRHg. Results for RHg determined
using the ethylation based method and for other Hg spe-
cies measured in the same samples are shown in Figure
8. Concentrations of the various species at the five sam-
pling locations were compared and found to vary in ways
consistent with each other in response to changes in dilu-
tion, mixing, rainfall, etc.
For all Hg species/fraction, the highest concentrations
were found at the location of the Idrija Hg mine, and
concentrations decreased downstream but in different
patterns for the different Hg species/fraction. The THg
concentration dropped quickly within a relatively short
distance (about 10 km), then gradually decreased, ap-
proaching the RHg concentration. This indicates that, in
addition to RHg, THg includes other Hg species such as
Hg0, HgS, Hg2Cl2, HgSe, etc. In the river, Hg0 evasion
and the adsorption/precipitation of HgS, Hg2Cl2, and
HgSe may take place, resulting in decreasing THg con-
Figure 8. Concentrations of Hg species/fractions measured
in the River Idrija in Slivenia.
centrations which gradually approach the RHg concen-
tration. In this particular case, Hg0 evasion enhanced by
the turbulent, torrential nature of the river is likely the
major cause of the dramatic decrease in the THg concen-
tration downstream from the Hg mine. Compared to THg,
RHg concentrations vary differently and in a pattern that
was not affected by Hg0 evasion. Concentrations of
MeHg decreased downstream of the mine, but in another,
different pattern compared to both THg and RHg.
This work has identified and characterized appropriate
and reliable analytical methods capable of producing
accurate, precise, and meaningful results for RHg. The
positive and negative biases that can arise in using the
Sn2+ reduction method for the determination of RHg have
been described, and the use of the alternative ethylation
GC-CVAFS method is recommended. Using the ethyla-
tion based method, it is confident that results generated
are the biogeochemically reactive fractions of Hg2+. As
any good measurements always can promote the research
growing up, it is believed that the discovery of the ana-
lytical problems using the Sn2+ reduction method for de-
termination of RHg will draw significant attention from
environmental researchers in the world. Once the prob-
lems outlined in this work are recognized and the ethyla-
tion based GC-CVAFS method is used, measurement of
RHg will become widespread because it will provide the
researchers with meaningful data. It is believed that a
breakthrough in the research of Hg biogeochemistry is a
The work was partially funded by the Slovenian Re-
search Agency (ARRS) through programme P1-0143 and
project J1-4288, and also supported by Chevron Energy
Technology Company under Contract CW831200. V.
We thank Fajon for sampling the water samples from the
 T. W. Clarkson and L. Magos, “The Toxicology of Mer-
cury and Its Chemical Compounds,” Critical Reviews in
Toxicology, Vol. 36, No. 8, 2006, pp. 609-662.
 T. Barkay and I. Wagner-Dobler, “Microbial Transforma-
tions of Mercury: Potentials, Challenges, and Achieve-
ments in Controlling Mercury Toxicity in the Environ-
ment,” In: A. I. Laskin, J. W. Bennett and G. M. Gadd,
Eds., Advances in Applied Microbiology, Vol. 57, El-
sevier Academic Press Inc., San Diego, 2005, pp. 1-52.
 M. Horvat, J. Snoj Tratnik and A. Miklavcic, “Mercury:
Biomarkers of Exposure and Human Biomonitoring,” In:
Open Access AJAC
L. LIANG ET AL. 631
L. E. Knudsen and D. F. Merlo, Eds., Biomarkers and
Human Biomonitoring, Issues in Toxicology, No. 9, RSC
Publishing, Cambridge, 2012, pp. 381-417.
 M. C. Newman, X. Y. Xu, A. Condon and L. Liang, “Flood-
plain Methylmercury Biomagnification Factor Higher
than That of the Contiguous River (South River, Virginia
USA),” Environmental Pollution, Vol. 159, No. 10, 2011,
 J. C. Wang, M. C. Newman, X. Y. Xu and L. Liang,
“Higher and More Variable Methylmercury Biomagnifi-
cation Factors for Floodplain Than the Contiguous River
(South River, Virginia USA),” Ecotoxicology and Envi-
ronmental Safety, 2013.
 H. Hsu-Kim, K. H. Kucharzyk, T. Zhang and M. A. De-
shusses, “Mechanisms Regulating Mercury Bioavailabil-
ity for Methylating Microorganisms in the Aquatic Envi-
ronment: A Critical Review,” Environmental Science &
Technology, Vol. 47, No. 6, 2013, pp. 2441-2456.
 J. K. Schaefer, S. S. Rocks, W. Zheng, L. Y. Liang, B. H.
Gu and F. M. M. Morel, “Active Transport, Substrate
Specificity, and Methylation of Hg(II) in Anaerobic Bac-
teria,” Proceedings of the National Academy of Sciences,
Vol. 108, No. 21, 2011, pp. 8714-8719.
 W. Zheng, L. Y. Liang and B. H. Gu, “Mercury Reduc-
tion and Oxidation by Reduced Natural Organic Matter in
Anoxic Environments,” Environmental Science & Tech-
nology, Vol. 46, No. 1, 2012, pp. 292-299.
 M. E. Hines, J. Faganeli, I. Adatto and M. Horvat, “Mi-
crobial Mercury Transformations in Marine, Estuarine
and Freshwater Sediment Downstream of the Idrija Mer-
cury Mine,” Applied Geochemistry, Vol. 21, No. 11, 2006,
 N. S. Bloom, “Influence of Analytical Conditions on the
Observed ‘Reactive Mercury’, Concentrations in Natural
Fresh Waters,” In: J. Huckabee and C. J. Watras, Eds.,
Mercury as a Global Pollutant, Lewis Publishers, Ann
 J. D. Dean and R. P. Mason, “Estimation of Mercury Bio-
accumulation Potential fromWastewater Treatment Plants
in Receiving Waters: Phase II,” Final Report, 2009, 05-
WEM-1COa, WERF, Co-published by IWA Publishing.
 S. Rapsomanikis, O. F. X. Donard and J. H. Weber,
“Speciation of Lead and Methyllead Ions in Water by
Chromatography/Atomic Absorption Spectrometry after
Ethylation with Sodium Tetraethylborate,” Analytical
Chemistry, Vol. 58, No. 1, 1986, pp. 35-38.
 N. S. Bloom, “Determination of Pictogram Levels of Me-
thylmercury by Aqueous Phase Ethylation, Followed by
Cryogenic Gas Chromatography with Cold Vapor Atomic
Fluorescence Detection,” Canadian Journal of Fisheries
and Aquatic Sciences, Vol. 46, No. 7, 1989, pp. 1131-
 L. Liang, M. Horvat and N. S. Bloom, “An Improved
Speciation Method for Mercury by GC/CVAFS after
Aqueous Phase Ethylation and Room Temperature Pre-
collection,” Talanta, Vol. 41, No. 3, 1994, pp. 371-379.
 S. Žižek, R. Milačič, R. Jaćimivić, M. J. toman and M.
Horvat, “Periphyton as a Bioindicator of Mercury Pollu-
tion in a Temperate Torrential River Ecosystem,” Che-
mosphere, Vol. 85, No. 5, 2011, pp. 883-891.
 W. R. Hatch and W. L. Ott, “Determination of Sub-Cold
Vapour Atomic Absorption Spectrophotometry,” Ana-
lytical Chemistry, Vol. 40, No. 14, 1968, pp. 2085-2087.
 L. Liang, Bloom and M. Horvat, “Simultaneous Deter-
mination of Mercury Speciation in Biological Materials
by GC/CVAFS after Ethylation and Room Temperature
Precollection,” Clinical Chemistry, Vol. 40, No. 4, 1994,
 L. Liang, C. Evens, S. Lazoff, J. S. Woods, E. Cernichiari,
M. Horvat, M. D. Martin and T. DeRouen, “Determina-
tion of Methyl Mercury in Whole Blood by Ethylation-
GC-CVAFS after Alkaline Digestion-Solvent Extrac-
tion,” Journal of Analytical Toxicology, Vol. 24, No. 5,
2000, pp. 328-332. http://dx.doi.org/10.1093/jat/24.5.328
 US EPA Method 1669, “Sampling Ambient Water for
Trace Metals at EPA Water Quality Criteria Levels,”
Washington, DC, 1996.
 US EPA 1631, “Mercury in Water by Oxidation, Purge
and Trap, and Cold Vapor Atomic Fluorescence Spec-
trometry,” Environmental Protection Agency, Washing-
ton, DC, 2002.
 US EPA 1630, “Methyl Mercury in Water by Distillation,
Aqueous Ethylation,” US Environmental Protection
Agency, Washington, DC, 2001.
 L. Liang and N. S. Bloom, “Determination of Total Mer-
cury by Single-Stage Gold Amalgamation with Cold Va-
por Atomic Spectrometry,” JAAS, Vol. 8, 1993, pp. 591-
 M. Horvat, “Determination of Mercury and Its Com-
pounds in Water, Sediment, Soil and Biological Sam-
ples,” In: N. Pirrone and K. R. Mahaffey, Eds., Dynamics
of Mercury Pollution on Regional and Global Scales:
Atmospheric Processes and Human Exposures around the
World, Springer, New York, 2005, pp. 154-190.
 B. H. Gu, Y. R. Bian, C. L. Miller, W. M. Dong, X. Jiang
and L. Y. Liang, “Mercury Reduction and Complexation
by Natural Organic Matter in Anoxic Environments,”
Proceedings of the National Academy of Sciences, Vol.
108, No. 4, 2011, pp. 1479-1483.
 W. M. Dong, L. Y. Liang, S. Brooks, G. Southworth and
B. H. Gu, “Roles of Dissolved Organic Matter in the
Speciation of Mercury and Methylmercury in an Con-
taminated Ecosystem in Oak Ridge,” Tennessee Envi-
ronmental, Vol. 7, No. 1, 2010, pp. 94-102.
 C. L. Miller, G. Southworth, S. Brooks, L. Y. Liang and
B. H. Gu, “Kinetic Controls on the Complexation be-
Open Access AJAC
L. LIANG ET AL.
Open Access AJAC
tween Mercury and Dissolved Organic Matter in a Con-
taminated Environment,” Environmental Science & Tech-
nology, 2009, Vol. 43, No. 22, 2009, pp. 8548-8553.
 M. Horvat, N. S. Bloom and L. Liang, “A Comparison of
Distillation with Other Current Isolation Methods for De-
termination of Methyl Mercury Compounds in Low Level
Environmental Samples, Part 1: Sediment,” Analytica
Chimica Acta, Vol. 281, No. 1, 1993, pp. 135-152.
 M. Horvat, L. Liang and N. S. Bloom, “A Comparison of
Distillation with Other Current Isolation Methods for the
Determination of Methyl Mercury Compounds in Low
Level Environmental Samples, Part 2: Water,” Analytica
Chimica Acta, Vol. 282, No. 1, 1993, pp. 153-168.
 L. Liang, M. Berndt, T. K. Bavin and P. Pang, “Defini-
tion and Application of Alkylatable Mercury (AHg) for
Estimation of Bioavailable Mercury,” 9th ICMGP, Gui-
yang, June 7-12, 2009, pp. S19-52.
 L. Liang, D. Hwang, L. Young, E. A. Aultman and P.
Pang, “The Use of Alkylatable Mercury for Risk Assess-
ment and Remediation Evaluation of Hg Contaminated
Water/Soil/Sediment,” 10th ICMGP, Halifax, Nova Sco-
tia, July 24-29, 2011.
 K. Matsunaga, S. Konishi and M. Nishimura, “Possible
errors Caused Prior to Measurement of Mercury in Natu-
ral Waters with Special Reference to Seawater,” Envi-
ronmental Science & Technology, Vol. 13, No. 1, 1979,
pp. 63-70. http://dx.doi.org/10.1021/es60149a013
 J. A. Dalziel and P. A. Yeats, “Reactive Mercury in the
Central North Atlantic Ocean,” Marine Chemistry, Vol.
15, No. 4, 1985, pp. 357-365.
 N. S. Bloom and E. A. Crecelius, “Determination of Mer-
cury in Seawater at Subnanogram per Liter Levels,” Ma-
rine Chemistry, Vol. 14, No. 1, 1983, pp. 59-64.
 G. A. Gill and W. F. Fitzgerald, “Picomolar Mercury
Measurements in Seawater and Other Materials Using
Stannous Chloride Reduction and Two Stage Gold
Amalgamation with Gas Phase Detection,” Marine Che-
mistry, Vol. 20, No. 3, 1987, pp. 227-243.
 T. Kanduc, D. Kocman and N. Ogrinc, “Hydrogeochemi-
cal and Stable Isotope Characteristics of the River Idrijca
(Slovenia), the Boundary Watershed between the Adriatic
and Black Seas,” Marine Chemistry, Vol. 14, No. 3, 2008,
 M. E. Hines, M. Horvat, J. Faganeli, J. C. J. Bonzongo, T.
Barkay, E. B. Major, K. J. Scott, E. A. Bailey, J. J. War-
wick and W. B. Lyons, “Mercury Biogeochemistry in the
Idrija River, Slovenia, from above the Mine into the Gulf
of Trieste,” Environmental Research, Vol. 83, No. 2,
2000, pp. 129-139.
 M. Horvat, S. Covelli, J. Faganeli, M. Logar, V. Mandic,
R. Rajar, A. Sirca and D. Zagar, “Mercury in Conta-
minated Coastal Environments, a Case Study: The Gulf of
Trieste,” Science of The Total Environment, Vol. 237-238,
1999, pp. 43-56.
 M. Horvat, V. Jereb, V. Fajon, M. Logar, J. Kotnik, J. Fa-
ganeli, M. E. Hines and J.-C. Bonzongo, “Mercury Dis-
tribution in Water, Sediment and Soil in the Idrijca and
Soča River Systems,” Geochemistry: Exploration, Envi-
ronment, Analysis, Vol. 2, No. 3, 2002, pp. 287-296.
 D. Kocman, T. Kanduc, N. Ogrinc and M. Horvat, “Dis-
tribution and Partitioning of Mercury in a River Catch-
ment Impacted by Former Mercury Mining Activity,”
Biogeochemistry, Vol. 104, No. 1-3, 2011, pp. 183-201.