Mercury Speciation in Highly Contaminated Soils from Chlor-Alkali
Plants
G.J.Zagury
Department of Civil, Geological and Mining Engineering, École Polytechnique de Montréal, Montreal,
Quebec, Canada
C.M.Neculita
Department of Civil, Geological and Mining Engineering, École Polytechnique de Montréal, Montreal,
Quebec, Canada
Chemical Engineering Department, École Polytechnique de Montréal, Montreal, Quebec, Canada
L.Deschênes
Chemical Engineering Department, École Polytechnique de Montréal, Montreal, Quebec, Canada
ABSTRACT: Mercury partitioning, speciation, and ecotoxicity were investigated in three highly
contaminated soils from chlor-alkali plants. Single extractions (for total, volatile, and methyl Hg) and a fourstep sequential extraction procedure (SEP) were used in order to assess Hg speciation and partitioning. Hg
was separated in fractions defined as water soluble (F1), exchangeable (F2), organic (F3) and residual (F4). In
order to evaluate the ecotoxicity of Hg-contaminated soils, mortality of earthworm (Eisenia andrei) was
assessed. Results revealed extremely (295-11500 mg Hg/kg) contaminated soils having an alkaline pH and
very low organic carbon content. In Soils 1 and 3, high percentages (88-98%) of the total Hg were present as
volatile Hg. The SEP indicated a relatively low mobility of Hg. The highest mobile fraction (F1+F2) was
found in Soil 1 (12 % of total Hg) whereas the lowest mobile fraction was found in Soil 3 (2.6% of total Hg).
Tissue-Hg concentrations of exposed earthworms were determined to establish bioconcentration factors
(BCFs). Relationships between soil extractable Hg, tissue-Hg concentrations and BCFs were established.
1 INTRODUCTION
Chlor-alkali plants (CAPs) which use metallic Hg
for the electrolytic production of chlorine and
caustic soda are potential sources of Hg pollution.
Extensive research on the role of CAPs in regional
Hg cycling has concluded that their contributions are
of particular concern to local and regional areas
(Maserti & Ferrara 1991, Rule & Iwashchenko
1998). The contamination of soils in the vicinity of a
CAP is due to emissions containing elemental Hg
(Hg0) which can react with metallic oxides and
organic matter (Schuster 1991, Stein et al. 1996). Hg
can be reemitted in the atmosphere especially during
periods of high temperature (Rule & Iwashchenko
1998) or can be oxidized under acidic conditions and
transformed into its bivalent form, Hg2+ (Hempel et
al. 1995). Divalent Hg is a precursor for the
formation of compounds with increased solubility
and mobility like Hg chloride (HgCl2) and Hg
hydroxide (Hg(OH)2) or increased bioavailability
such as methyl Hg (CH3Hg+) or ethyl Hg
((CH3)2Hg). Therefore, as for other metals, Hg
speciation is essential because it controls its
solubility, volatility, reactivity, bioavailability, and
finally, its toxicity (Stein et al. 1996). The toxicity of
Hg is well known for some species (HgCl2,
CH3HgCl) but is difficult to assess in real
contaminated sites because the differences in the
solubility of the Hg species affect the concentration
available to the exposure pathway. The
determination of speciation is an important step in
the evaluation of the toxicity potential of Hgcontaminated soils, and a strong need exists for
appropriate solid phase speciation schemes. To our
knowledge, mercury binding forms and speciation in
highly contaminated soils from CAPs has not yet
been reported. Furthermore, despite the widespread
knowledge that Hg binding forms and speciation are
key factors for conducting ecotoxicological hazard
assessments, few attempts have been made in order
to ascertain the relations between Hg fractionation
and speciation in highly contaminated soils and its
uptake by biota (Edwards et al. 1998).
The study focuses on Hg-contaminated soils toxicity
in relation to total Hg content, Hg partitioning, and
Hg uptake. A novel Hg-specific solid phase
speciation scheme was developed and tested on three
highly contaminated ambient soil samples obtained
from CAPs. It consists of four steps for estimating
the distinct fractions of Hg removed in specific
environments (neutral, slightly acidic with
exchangeable cations, strongly alkaline, and strongly
acidic). In order to evaluate the toxicity of Hgcontaminated soils, mortality of earthworms
(Eisenia andrei) was assessed. Tissues-Hg
concentrations of exposed earthworms were
determined to establish bioconcentration factors
(BCFs). Relationships have been established
between soil extractable Hg and tissues-Hg.
2 MATERIALS AND METHODS
2.1 Sampling and preservation
The experiments were performed with three Hgcontaminated soil samples obtained from CAPs in
The Netherlands, Belgium and France. Upon
reception in the laboratory, all samples were
homogenized and stored in HDPE containers at 4°C
in order to limit Hg volatilization (Sakamoto et al.
1995). All analyses were performed in triplicate (at
least) with wet soils, after removal of particles larger
than 2 mm. The results were corrected for moisture
contents.
2.2 Soil characterization
The pH was measured in deionized water according
to Method D4972-95a (ASTM 1995). Moisture
content was determined at 45°C (Sakamoto et al.
1995) and cation exchange capacity (CEC) was
determined using the sodium acetate method
(Chapman 1965). Volatile solids were determined at
550°C (Karam 1993). Total carbon and total sulphur
were measured by combustion with an induction
furnace and total inorganic carbon was determined
following a phosphoric acid treatment (Ministère de
l’Environnement et de la Faune du Québec 1996).
Organic carbon was calculated by the difference
between total carbon and total inorganic carbon. The
samples were classified as gravel (> 2mm), sand
(2mm-75µm), and silt and clay (< 75µm) according
to Method D2487-83 (ASTM 1985).
2.3 Total Hg, Volatile Hg, and Methyl Hg
The total concentration of Hg in soils (n=5) was
determined following an acid digestion (Akagi &
Nishimura 1991). One gram of soil was placed in a
volumetric flask (100-ml) and 14 ml of 1 HNO3:5
H2SO4:1 HClO4 were added. The digestion was
performed at 250oC for 1h. The cold digest was
filtered (0.45 µm), and diluted with deionized water
to 100 ml. Total Hg was measured by cold vapor
atomic absorption spectrometry (CVAAS).
Volatile Hg was analyzed using a separate
extraction by the loss of Hg vapors after the soil (5g)
was heated (n=5) at 180°C for 2 days and then
digested using the same procedure as for total Hg.
Volatile Hg was calculated as the difference between
total Hg in the sample before and after heating.
Methyl Hg (CH3Hg+) analysis (n=4) was
performed by gas chromatography separation and
cold vapor atomic fluorescence spectroscopy
detection (GC/CVAFS) after
ethylation (Liang et al. 1994).
aqueous
phase
2.4 Sequential Extraction Procedure
Following a literature review covering the
determination of Hg speciation in solid samples, a
novel SEP was developed, based partially on the
work of Di Giulio & Ryan (1987), Lechler et al.
(1997) and Wallschläger et al. (1998). The proposed
SEP provides differentiation of Hg compounds into
the following four fractions: F1: water-soluble; F2:
exchangeable; F3: bound to organic matter, and F4:
residual Hg.
The sequential extractions were performed using
samples of soil (2 g) mixed with 20 ml of solvent in
a 50-ml PP centrifuge tube and shaken for 2 hrs at
20 ± 2°C using a wrist-action shaker. Between each
of the successive extractions and rinses, the
supernatant was obtained by centrifuging at 12,000 g
for 15-25 minutes at 10oC. Rinsing steps consisted
of washing the leached residues twice with
deionized water (20 and 10 ml) for 15 min. The
rinses were then added to the solvent extract from
the same sample and the combined supernatant was
analyzed for Hg by CVAAS. The solid residue was
used in the next extraction step. The water-soluble
fraction extractions were carried out using deionized
water. The exchangeable fraction was extracted
under slightly acidic conditions with 1M CaCl2
(pH=5). The organic fraction was separated by
successive extractions using 0.2 M NaOH and
CH3COOH 4 % (v/v). Residual Hg was extracted
by adding the same reagents as for total Hg to the
soil pellet directly in the original 50-ml centrifuge
tube. The sample was then transferred into a 100-ml
volumetric flask and the digestion was performed
using the same procedure as described for total Hg.
Statistical treatment of data was performed using
STATISTICA software (5.1 version, Statsoft 1997).
2.5 Eisenia andrei mortality, tissue-Hg
concentrations and BCFs
Eisenia andrei survival tests were carried out using a
modification of the standard procedure of Quebec’s
Ministry of the Environment (Centre d’Expertise en
Analyse Environnementale du Québec 2003). Sieved
soil (< 2mm) samples (200g) were transferred in
plastic pots and 10 organisms were added to each
recipient. The bioassays (n=3) were conducted at
20±2°C under the dark, for 21 days, at a soil
moisture content of 95% of the water holding
capacity. The number of survivors was recorded
weekly. After 7, 14, and 21d, randomly selected
earthworms (n=3) were removed from soil and
prepared for Hg-tissue analysis. The earthworms
were washed with cold tap water and placed into
glass finger bowls lined with filter paper (changed
every 12 h over a 48-h period). The earthworms
were then rinsed in deionized water, killed by deep
freezing (-20°C), thawed at room temperature, and
dried at 45°C for 48 hours to constant weight. The
same digestion procedure as the one described for
total Hg analysis was performed for Hg-tissue
analysis
of
earthworms.
Uncontaminated
experimental soil was used for the negative controls.
3 RESULTS AND DISCUSSION
3.1 Soil Physicochemical Characteristics
Soil characterization showed that all samples had an
alkaline pH with values of 7.90 ± 0.05 (Soil 1),
7.92 ± 0.04 (Soil 2) and 9.11 ± 0.03 (Soil 3). This
pH range is typical of contaminated soils originating
from CAPs. Organic carbon content was below
detection or very low in all soils (below detection
(Soil 3), 0.24% (Soil 1), and 1.82% (Soil 2)).
Particle-size distribution analysis indicated sandy
soils with the majority of particles having a diameter
higher than 75 µm (91.1 %, 90.8 %, and 86.1 % for
soils 1, 2, and 3 respectively). Soils 1 and 2 were
classified as sandy soils and Soil 3 as a gravelly sand
with fines. The CECs were 12.1 meq/ 100g (Soil 1),
10.6 meq/ 100g (Soil 2) and 5.9 meq/ 100g (Soil 3).
These values are typical of coarse-grained inorganic
soils.
3.1.1 Total Hg, Volatile Hg and Methyl Hg
All soils were highly contaminated (Table 1).
Compared to other published studies dealing with
chlor-alkali contaminated soils, the three soil
samples in our study were much more contaminated
(80.5-104 mg Hg/kg (Maserti & Ferrara 1991) and
0.47-4.2 mg Hg/kg (Biester et al. 2002)). The
highest percentages of volatile Hg were found in
Soils 1 and 3, with 88 and 98 % of total Hg
respectively. Volatile Hg was much lower in Soil 2
(14.4%). Such levels are unusual, but are not odd
(Biester & Nehrke 1997) since Hg0 is used as a
cathode in caustic soda productions.
Methyl Hg concentrations were low (from 0.24 to
19.3 µg/kg) compared to published soil values (from
670 to 10 000 µg/kg) used in screening, cleanup or
monitoring of contaminated sites (American
Petroleum Institute 2000). The higher organic
carbon content of Soil 2 (1.82 %) did not entail a
high proportion of methyl Hg (0.008 % of total Hg).
3.1.2 Sequential Extraction Procedure
The precision of the developed SEP was very
satisfactory as reflected by the small standard
deviations obtained using 3-4 replicates (Table 1).
Recoveries (the sum of extracted Hg fractions
divided by the independently determined total Hg
concentration) ranged from 74 ± 1 to 130 ± 6 %.
SEP results indicate that the sum of fractions F1,
F2 and F3 represented low percentages of total Hg
(below 14 %, 4 %, and 3 % for soils 1, 2 and 3,
respectively). The most important contribution to the
sum was the second extraction (F2) with results
varying between 2.1 % (Soil 3) and 11.1 % (Soil 1).
In contrast, fraction F3 (Hg associated to organic
matter) was very low (0.3 to 2.0 %) and consistent
with the low organic carbon content of the three
study soils.
Although F1 only accounts for at most 0.7 % of
total Hg (Soil 1), this fraction is very important from
an environmental risk point of view due to its easy
availability in environmental weathering conditions
(Wallschläger et al. 1998, Bloom et al. 2003).
Moreover, the small percentages should be treated
with caution because they can represent high Hg
concentrations. In Soil 3, for example, the soluble
fraction is about 0.5 % of total Hg, but it still
represents more than 50 mg Hg/kg which exceeds
the ecological soil criteria of Finland and Sweden
(De Vries & Bakker 1998) by a factor of 250.
In all soils, the largest Hg proportion was found
within the residual fraction (70.8 –116 % of total
Hg). In naturally occurring conditions, this is the
least available form of Hg, depending on the matrix
under study and the source of contamination
(Wallschläger et al. 1998, Bloom et al. 2003). But in
our study, this result is misleading due to the
presence of very high concentrations of volatile Hg
in Soils 1 and 3. Furthermore, the relatively lower
recovery percentages for the SEP conducted on Soil
3 (74 %) suggest the possible loss of volatile Hg
species during the manipulations (Wallschläger et al.
1998).
Globally, Soil 1 showed the highest percentages
of mobile (water soluble and exchangeable) Hg
(12% of total Hg). Hence, mobile Hg in this soil
could easily migrate to deeper soil layers and
disperse via pore water. Soil 2 had the lowest
concentration of total Hg and appeared to be less
hazardous in terms of volatility and mobility.
Nevertheless, the Hg found in this soil remains
potentially phytoavailable. Soil 3 was extremely
contaminated (11 500 ± 457 mg/kg) with 98 % of
total Hg in the volatile form. The proportion of
easily leachable Hg (F1+F2) was the lowest for this
soil (2.6% of total Hg), suggesting relatively lower
potential Hg mobility via pore water. However, as a
consequence of the higher total Hg concentration in
this soil, the mobile fractions actually represent very
high Hg concentrations (300 mg/kg). Furthermore,
even though the concentration of methyl Hg is
relatively low (19.3 ± 1.47 µg/kg), this species is
still potentially toxic due to its bioconcentration
potential.
3.1.3 Mortality of E. Andrei, tissue-Hg
concentrations and BCFs
No mortality occurred during the 21-day incubation
in Soils 1 and 2. Fischer & Koszorus (1992) also
found no lethal effects on earthworms (Eisenia
fetida) during 56-day exposure in soils artificially
contaminated with HgCl2 (500 mg HgCl2/kg).
Results also revealed 0 % mortality in control soils.
At the highest tested concentration (Soil 3), the
mortality was 5% (7d), 85% (14d), and 100% (21d).
Distinct earthworm behavior, judged against the
common pattern observed in control soil was noted
in this extremely contaminated soil, where divided
worms were often observed.
Analysis of tissue-Hg concentrations showed
higher values in soils with higher concentrations of
mobile Hg (F1+F2). Tissue-Hg concentrations also
increased with exposure period. At 7, 14 and 21d,
the highest Hg concentrations were found in
earthworms placed in Soil 1 (480, 979 and 1410 mg
Hg/kg) whereas in Soil 2 lower values were
measured (94.7, 139 and 178 mg Hg/kg). The high
mortality of earthworms in Soil 3 at 14 and 21 days,
only allowed the 7-day tissue-Hg analysis which
revealed 870 mg Hg/kg.
BCFs showed a trend similar to tissue-Hg
concentrations: higher BCFs were measured in soils
with higher concentrations of mobile Hg (F1+F2),
and BCFs increased with exposure period (7-day to
21-day). BCFs varied from 0.32 to 0.60 (Soil 2) to
0.8 to 2.5 (Soil 1). In Soil 3, a very low value of
0.08 was observed after 7 days of exposure.
Comparison of soil extractions results with the
concentrations of Hg in earthworms produced good
correlations. The mortality was correlated with total
Hg, as well as with the soluble fraction. The BCFs
were also positively correlated with total Hg (at 14
and 21d) and with the water soluble fraction (at 14
and 21d).
having an alkaline pH and a very low organic carbon
content. Methyl Hg concentrations were relatively
low in all soils but very high concentrations of
volatile Hg were found.
The developed extraction procedure distinguished
water soluble (F1), exchangeable (F2), organic (F3),
and residual (F4) Hg. Sequential extractions
indicated that the majority of the Hg was associated
with the residual fraction which generally represents
the least available form of Hg. In Soils 1 and 3,
however, high percentages (88-98%) of the total Hg
were present as volatile Hg. Soil 1 showed the
highest percentages of mobile (F1+F2) Hg (about 12
% of total Hg) whereas mobile Hg percentages were
much lower in the two other soils. In all soils, Hg
associated with organic matter (F3) was very low
(0.3-2% of total Hg) in agreement with the low
organic carbon content of the soils. In Soil 3, the low
percentage of mobile Hg (2.6 %) is misleading
because it actually represents a very high Hg
concentration (300 mg/kg). The developed SEP was
very precise and efficient for all Hg-contaminated
soils tested, with satisfactory recovery percentages
ranging from ranged from 74 ± 1 to 130 ± 6 %.
Toxicity tests showed no lethality in Soils 1 and 2
but a high lethality was recorded in Soil 3. In Soils 1
and 2, the higher mobile Hg concentrations entailed
higher tissue-Hg concentrations as well as higher
BCFs. Soil 3, with the highest concentration of
mobile Hg showed the highest tissue-Hg
concentration at 7 days and the highest toxicity, but
showed limited bioconcentration after 7 days
suggesting a threshold in Hg uptake.
Relationships have been established between soil
extractable Hg and toxicity, as well as between soil
extractable Hg and BCFs. Toxic effects on
earthworms and BCFs were positively correlated
with total Hg and with soluble Hg. Globally, this
study indicates that both mobile Hg concentrations
and total Hg are important factors in the
environmental impact assessment of Hgcontaminated soils following the activity of chloralkali plants.
5 ACKNOWLEDGEMENTS
4 CONCLUSION
In order to assess Hg partitioning in highly
contaminated soils affected by the activity of CAPs,
a novel four-step sequential extraction procedure
was developed. Total, volatile, and methyl Hg
concentrations were also determined using separate
single extractions. The soil characterization revealed
extremely contaminated coarse-grained sandy soils
This research was supported by Solvay and Total
Fina Elf. The authors gratefully acknowledge the
assistance provided by Mr. Roger Jacquet and Ms.
Patricia De Bruycker of Solvay, and Dr. Frédéric
Périé
of
Total
Fina
Elf.
6 REFERENCES
Akagi H, Nishimura H (T. Suzuki et al.) 1991. Advances in
Mercury Toxicology. New York: Plenum Press.
American Petroleum Institute 2000. A critical review of
methods for developing ecological soil quality guidelines
and criteria. Washington, USA.
American Society for Testing and Materials 1985.
Classification of Soils for Engineering Purposes.
American Society for Testing and Materials 1995. Standard
test method for pH of soils.
Biester H, Nehrke G 1997. Quantification of mercury in soils
and sediments: acid digestion versus pyrolysis. Fresenius
Journal of Analytical Chemistry 358: 446-452.
Biester H, Müller G, Schöler HF 2002. Binding and mobility
of mercury in soils contaminated by emissions from chloralkali plants. The Science of the Total Environment 284:
191-203.
Bloom NS, Preus E, Katon J, Hiltner M 2003. Selective
extractions to assess the biogeochemically relevant
fractionation of inorganic mercury in sediments and soils.
Analytica Chimica Acta 479: 233-248.
Centre d’Expertise en Analyse Environnementale du Québec
2003. Détermination de la toxicité létale avec le ver de terre
(Eisenia andrei, Eisenia fetida). Ministère de
l’Environnement, Québec, Canada.
Chapman HD (Black CA) 1965. Methods of soils analysis. Part
2: Chemical and microbiological properties. Madison:
American Society of Agronomy Inc.
De Vries W, Bakker DJ 1998. Manual for calculating critical
loads of heavy metals for terrestrial ecosystem: Guidelines
for critical limits, calculation methods and input data. Den
Helder, The Netherlands.
Di Giulio RT, Ryan EA 1987. Mercury in soils, sediments, and
clams from a North Carolina peatland. Water, Air, and Soil
Pollution 33: 205-219.
Edwards SC, Macleod CL, Lester JN 1998. The bioavailability
of copper and mercury to the common nettle (Urtica Doica)
and the earthworm Eisenia fetida from contaminated
dredge spoil. Water, Air, and Soil Pollution 102:75-90.
Fischer E, Koszorus L 1992. Sublethal effects, accumulation
capacities and elimination rates of As, Hg and Se in the
manure worm, Eisenia foetida (Oligochaeta, Lumbricidae).
Pedobiologia 36:172-178.
Hempel M, Wilken RD, Geilhufe C, Richter-Politz I (van den
Brink WJ, Bosman R, Arendt) 1995. Contaminated Soils.
Netherlands: Kluwer Academic Publishers.
Karam (Carter MR) 1993. Soil sampling and methods of
analysis. Boca Raton: Lewis Publishers.
Lechler PJ, Miller JR, Hsu LC, Desilets MO 1997. Mercury
mobility at the Carson River Superfund Site, west-central
Nevada, USA: interpretation of mercury speciation data in
mill tailings, soils, and sediments. Journal of Geochemical
Exploration 58: 259-267.
Liang L, Horvat M, Bloom NS 1994. An improved speciation
method for mercury by GC/CVAFS after aqueous phase
ethylation and room temperature precollection. Talanta 41:
371-379.
Maserti BE, Ferrara R 1991. Mercury in plants, soil and
atmosphere near a chlor-alkali complex. Water, Air, and
Soil Pollution 56: 15-20.
Ministère de l’Environnment et de la Faune du Québec 1996.
Solides – Détermination du carbone inorganique total,
dosage par spectrophotométrie IR. Québec, QC, Canada.
Rule JH, Iwashchenko MS 1998. Mercury concentrations in
soils adjacent to a former chlor-alkali plant. Journal of
Environmental Quality 27: 31-37.
Sakamoto H, Tomiyasu T, Yonehara N 1995. The content and
chemical forms of mercury in sediments from Kagoshima
Bay, in comparison with Minamata Bay and Yatsushiro
Sea, southwestern Japan. Geochemistry Journal 29:97-105.
Schuster E 1991. The behavior of mercury in the soil with
special emphasis on complexation and adsorption
processes-a review of the literature. Water, Air, and Soil
Pollution 56: 667-680.
StatSoft, Inc. STATISTICA for Windows (Computer program
manual) 1997. Tulsa, OK: StatSoft, Inc.
Stein ED, Cohen Y, Winer AM 1996. Environmental
distribution and transformation of mercury compounds.
Critical Reviews in Environmental Science and Technology
26: 1-43.
Wallschläger D, Desai MVM, Spengler M, Wilken RD 1998.
Mercury speciation in floodplain soils and sediments along
a contaminated river transect. Journal of Environmental
Quality 27: 1034-1044.
Table 1. Total Hg and Hg partitioning in contaminated soils from CAPs using the developed four step sequential extraction procedure.
Soil
Total Hg
(mg/kg)
*
1*
568 ± 24.2
2*
295 ± 18.3
3*
1.15x104 ± 457
Fraction 1
Water soluble
(mg/kg)
(%)**
4.08 ± 0.9
0.7 ± 0.2
5
0.27 ± 0.0
0.1 ± 0.0
1
53.4 ± 1.5
0.5 ± 0.0
9
Fraction 2
Exchangeable
(mg/kg)
(%)
63.3 ± 2.1
11.1 ± 0.4
0
9.20 ± 0.3
3.1 ± 0.1
2
Fraction 3
Organic
(mg/kg)
(%)
246 ± 18.4
2.1 ± 0.2
Fraction 4
Residual
(mg/kg)
(%)
Sum of
fractions
(mg/kg)
Calculated as: (Hg extracted in concerned fraction/ total Hg) x 100
***
Calculated as: (sum of extracted Hg fractions / total Hg) x 100
(%)***
11.5 ± 2.37
2.0 ± 0.4
661 ± 32.1
116 ± 6
740 ± 36
130 ± 6
1.43 ± 0.06
0.5 ± 0.0
314 ± 7.30
106 ± 2
325 ± 7.5
110 ± 3
35.3 ± 4.17
0.3 ± 0.0
8.12 x 103 ± 125
70.8 ± 0.9
Mean values and S.D. are calculated from 3 different determinations for Soil 2 and 4 different determinations for Soil 1 and Soil 3.
**
Recovery
8.45 x
103 ± 108
74 ± 1