Solubility and Changes of Mercury Binding Forms in Contaminated

Environ. Sci. Technol. 1998, 32, 2755-2762
Solubility and Changes of Mercury
Binding Forms in Contaminated
Soils after Immobilization Treatment
HARALD BIESTER* AND HOLGER ZIMMER
Institute of Environmental Geochemistry, INF 236,
69120 Heidelberg, Germany
Mobility at different pH and binding forms of mercury (Hg)
have been investigated in three Hg-contaminated soils
after immobilization treatment with alkali-polysulfide (APS)
and trimercapto-s-triazine trisodium salt solution (TMT).
Changes of solid-phase Hg binding forms after immobilization
were determined by Hg pyrolysis. Hg concentrations in
the water extracts of all samples increased after treatments
due to the formation of soluble mercury sulfides (APS
treatment), and the mobilization of humic acid bound Hg
at the high pH of the reagents. In contrast, Hg concentrations
decreased sharply at low pH due to decomposition of
soluble mercury sulfides and precipitation of humic acidbound Hg. Inorganic Hg compounds such as Hg0 or HgCl2
are effectively transformed to mercury sulfides by APS
treatment, whereas TMT could transform HgCl2 but not Hg0.
Both reagents were found to affect humic acid bound
Hg by way of increasing Hg desorption temperatures, although
APS was found not to desorb Hg completely from humic
acids and TMT-Hg complexes are actually incorporated
into humic acids.
Introduction
Contamination of soils with mercury (Hg) is a serious problem
due to the high toxicity of the metal and its compounds.
Methods for remediation of Hg-contaminated sites include
thermal treatment, leaching procedures, electrolysis (1-3),
or a combination of these methods if additional contaminants
are present. These methods are usually costly and timeconsuming. Therefore, in situ or on site immobilization of
Hg in contaminated soils by chemical treatment has been
considered as a stabilization treatment to reduce environmental risks (4) and as a chemical pretreatment of contaminated soils to maintain threshold values for landfill
deposition. In the case of large contaminated areas or where
the soil cannot be excavated, in situ immobilization of soluble
Hg compounds is considered an attractive alternative to stop
or prevent groundwater contamination. Due to the high
affinity of Hg compounds for sulfur, the use of sulfurcontaining reagents such as alkali-polysulfides or organic
compounds such as trimercapto-s-triazine salts have been
offered by remediation companies for the immobilization of
Hg in soil materials. Most of the available data concerning
the reaction of Hg with immobilization reagents are based
on the removal of Hg from wastewater by the formation of
insoluble Hg compounds (5). In general, the aim of using
sulfur-containing reagents is to reduce the mobility of soluble
or volatile Hg compounds in soils by forming stable Hg
* Corresponding author tel: +49-6221-544819; fax: +49-6221545228; e-mail: [email protected].
S0013-936X(97)00937-1 CCC: $15.00
Published on Web 08/11/1998
 1998 American Chemical Society
compounds such as metacinnabar (HgS) or insoluble organic
Hg-S compounds. It is unclear how and to what extent Hg
compounds in soils are transformed during such immobilization treatment. Although the magnitude of Hg
mobility reduction can be determined by leaching tests or
by the quantification of degassing volatile Hg compounds,
there is no published data regarding the quality and longterm stability of Hg compounds formed during immobilization. It is also unclear how the immobilizing reagents react
with different Hg compounds in soils contaminated with
metallic Hg (Hg0), HgCl2, or humic acid (HA)-bound Hg. Here,
we report that Hg binding forms in three Hg-contaminated
soils after treatment with two different immobilizing reagents.
Before and after treatment, solid-phase Hg binding forms
were determined by a pyrolysis technique similar to that
used in earlier studies to distinguish Hg binding forms in
soils by thermal Hg desorption (6-10). The results of the
solid-phase measurements were compared to those of
standard Hg-S compounds obtained by the reaction of the
reagents with HgCl2. It is known from numerous studies
that coupling of Hg to humic substances is the predominant
Hg binding form in most natural soils (11-13). Therefore,
one objective of our study was to determine the changes of
humic acid bound Hg binding characteristics during the
immobilization process. We extracted HA from soils before
and after treatment and compared changes in Hg content
and Hg desorption temperatures. Additionally, we investigated aqueous-phase Hg mobility in the untreated and
treated soils through leaching tests at different pH values.
Materials and Methods
Soil Samples. Soil samples were collected from three Hgcontaminated sites having different soil types and Hg
pollution histories. Sample CAP was taken from the top soil
layer of a former chlor alkali plant (Bitterfeld, Germany),
where Hg0 was directly spilled into the soil. The dark brown
soil consists mostly of sand and humic materials together
with small amounts of silty and clayey components (9.8%).
KYA soil was sampled from soils of a wood pressure treatment
site in Bad Krozingen, SW Germany. The site was contaminated by HgCl2 washed off from treated timber. This loess
soil consists of high amounts of carbonates (27%), clayey
and silty components (61.8%), and comparatively small
amounts of organic matter. Sample REC was taken as a mixed
sample from a 2.0-2.9-m section of soil core taken from a
former Hg recycling site in Frankfurt a.M., Griesheim, where
large amounts of metallic Hg were spilled into the top soil.
This soil consists of well-sorted medium grained sands. Clayey
components account for only 6.3% and the content of organic
carbon is low.
Immobilization. Immobilization tests were carried out
using either a sodium polysulfide solution (APS) (AC 2000/
aqua control) containing 5.29 mol/L sulfur or an aqueous
solution (15%) of trimercapto-s-triazine trisodium salt
(Na3C3N3S3) (TMT 15, Degussa) containing 1.85 mol/L sulfur.
For each test, 200 g of sample and 600 mL of demineralized
water were placed into a 1-L centrifuge bottle. After ultrasonic
dispersion (KLN System 582) of the samples, the reagents
were added at a molar concentration double the content of
Fe, Mn, Cu, Pb, Zn, Cd, Sn, As, and Hg in the sample. The
reagent concentrations were calculated to produce equimolar
amounts of sulfur and metals. The samples were shaken
end to end for 2 h and centrifuged (1 h/4300 rpm). The
supernatants were decanted and filtered through a 0.45-µm
nitrate cellulose filter. The solid residues were thoroughly
homogenized and frozen (-18 °C) until analysis. The pH of
VOL. 32, NO. 18, 1998 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2755
TABLE 1. Concentrations of Metals with Affinity to Sulfur (Fe,
Mn, Cu, Pb, Zn, Cd, Sn, Hg, As), Total Hg Content, and
Content of Organic Carbon (Corg) in Hg-Contaminated Soil
Samples (CAP, KYA, REC)
sample
total Hg
(mg/kg)
∑ Fe, Mn, Cu, Pb, Zn,
Cd, Sn, Hg, As (mol/kg)
Corg (%)
CAP
KYA
REC
1717 ( 331
161.3 ( 15.1
23.05 ( 2.44
0.22591
0.29166
0.31779
1.4 ( 0.08
0.7 ( 0.05
0.1 ( 0.008
the solutions was determined directly after centrifugation
by means of a glass electrode. Hg concentrations of the
solutions were determined by cold vapor atomic absorption
spectrometry (CV-AAS) after digesting 5 mL of the solution
in 20 mL of aqua regia (2 h/160 °C) and reducing Hg2+ with
stannous chloride using an automated Hg analysis system
(TSP mercury monitor 3200).
Standard Substances. Standard substances for mercury
sulfur compounds were derived from the reaction of Hg2+
with APS or TMT using a 2-fold excess of the reagents to a
0.01 M HgCl2 solution. The solutions were shaken for 2 h,
centrifuged, and decanted. The precipitates were washed
four times, freeze-dried, and stored frozen until analysis.
Polysulfides are generally described to react with mercury
according to
Hg + Sn2- f HgS + Sn-12-
(n ) 3-6)
(1)
The reaction of Hg with TMT leads to the formation of
insoluble organic complexes according to
S–
2+
Hg
+
S–
N
S–
N
S Hg S
S–
N
2
N
N
S–
N
N
S–
N
N
S–
(2)
However, it is assumed that the Hg binding is intermolecular
as well as intramolecular with the two S ligands of the TMT
molecule (14). Moreover, it is assumed that TMT-Hg
molecules can exist as monomers or polymers (15).
Solid-phase standards for HgCl2 and cinnabar were
obtained by mixing 0.001 M HgCl2 (Merck) and 0.001 M HgS
(red cinnabar, Merck) with 20 g of quartz powder for dilution.
Carboniferous schists bearing visible droplets of metallic Hg
were used as a standard of unbound metallic Hg. Hg0
incubated iron oxyhydroxides were prepared by incubating
dry iron oxyhydroxides in a sealed container for 14 d/40 °C
in a Hg0-saturated atmosphere. Humic acids were extracted
from untreated and treated samples CAP and KYA using 0.1
M NaOH according to the standard procedure of Calderoni
and Schnitzer (16). Due to the low amount of organic carbon
in REC (Table 1), we could not extract sufficient amounts of
solid HA from this soil to determine the Hg content of the
HA-bound Hg fraction or to analyze Hg desorption characteristics. The HA fraction was precipitated by acidifying
the extracts to pH <2 using hydrochloric acid, washing four
times, freeze-drying, and storing frozen until analysis.
Elemental Analysis. Metal concentrations were determined by flame AAS (Perkin-Elmer 4100) after digestion of
2 g of samples in 20 mL of aqua regia (160 °C/2 h). Hg in
HA was analyzed by CV-AAS (Dr. Seitner, Hg Monitor 254A)
using stannous chloride mercury reduction after digesting
10-30 mg of the extracted HA in 5 mL of aqua regia. The
total carbon content of the samples was determined by
photometric detection of CO2 after combustion of the
homogenized sample (0.5 g) in a high-frequency induction
furnace (CS-225 LECO). The content of inorganic carbon
was calculated from the amount of carbonates in the samples
2756
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 32, NO. 18, 1998
determined by means of a “carbonate bomb” (17). The
content of organic carbon was obtained by subtracting the
inorganic carbon content from the total carbon content.
Hg Pyrolysis. Hg pyrolysis is based on thermal decomposition or desorption of Hg compounds from solids at
different temperatures and continuous determination of
released Hg. The solid soil samples are heated in a pyrolysis
unit connected to a measuring cell that is placed inside the
detection unit of an atomic absorption spectrometer (PerkinElmer AAS 3030). A detailed description of the apparatus is
given elsewhere (9). Measurements were carried out at a
heating rate of 0.5 °C/s and a N2 gas flow of 300 mL/min. The
detection limit of the system is 40 ng at a maximum sample
weight of about 200 mg for a single peak curve. Sample
weights range between 1 and 50 mg (wet weight). Weights
of extracted HA were always below 1 mg (dry weight) due to
their high Hg content (approximately 1.2-2.7%). Replicates
of the same sample can vary up to 10 °C in Hg release onset
temperatures. Peak heights can vary substantially if wet
sample material is used (10).
Leaching Tests. Leaching tests were performed by
shaking 20 g of the untreated and treated samples for 24 h
end-to-end in 1-L polyethylene centrifuge bottles with 200
mL of leach solution. Solution pH was adjusted by using
glycine/HCl buffer solution (pH 3), acetic acid/sodium acetate
buffer solution (pH 4.6), and demineralized water (pH 5.7).
All solutions were added to the fresh sample material. After
centrifugation (1 h/4300 rpm) all solutions were filtered
through 0.45-µm nitrate cellulose filters, and the pH was
determined. We distinguished easily reducible soluble Hg
from Hg bound to soluble organic complexes such as fulvic
acids (total soluble Hg minus easily reducible Hg) according
to Meili et al. (18). The easily reducible soluble Hg fraction
was determined in the undigested extracts directly after
stannous chloride mercury reduction. Analysis of total Hg
in the extracts was carried out by CV-AAS after digesting 5
mL of the solution in aqua regia.
Results and Discussion
Hg desorption curves of the untreated samples show that in
all samples the main Hg release occurs in the temperature
range between 120 and 300 °C (Figure 1) similar to the results
of other studies (6-8). Hg released in this temperature range
was generally assigned to the desorption of Hg from
nonspecific soil matrix components. It could be seen from
the Hg release curves of the standard substances (Figure 2)
that Hg sorbed to mineral soil components is released at
lower temperature than Hg that is bound to humic acids by
covalent bonding (e.g., R-S-Hg). Accordingly, the lower
Hg release temperature of REC as compared to CAP and KYA
suggest a predominant bonding of Hg to mineral surfaces.
However, at high Hg concentrations after saturation of humic
acid binding sites, Hg can also be bound by adsorption,
showing Hg release temperatures similar to Hg bound to
mineral surfaces. Therefore, Hg bound to different components in the same soil cannot readily be distinguished by
pyrolysis measurements alone. Comparing the Hg release
curves of the soils with those of standard Hg sulfides, it is
also indicated that in none of the soils does Hg occur as red
or black cinnabar. Concentrations of metals with affinity to
sulfur (Fe, Mn, Cu, Pb, Zn, Cd, Sn, As), Hg, and organic carbon
concentrations in the soils are given in Table 1.
In untreated CAP, we observed a strong Hg enrichment
in the extracted humic acids (Figure 6). The amount of HA
extracted from this soil was the highest of the three soils,
indicating that humic acid-bound Hg is the predominant Hg
binding form in this soil. The desorption curve of this sample
indicates the occurrence of free metallic Hg by an additional
peak starting below 100 °C (Figure 1) according to the Hg0
standard (Figure 2). Earlier investigations (10) reported about
FIGURE 1. Hg release curves of the untreated samples CAP, KYA, and REC.
FIGURE 2. Hg release curves of standard Hg compounds: (A) Hg(0), (B) Hg(0) incubated iron oxyhydrates, (C) HgCl2, (D) extracted humic
acids (CAP), (E) APS + HgCl2, (F) synthetic red cinnabar, (G) TMT + HgCl2.
44% for Hg0 (RSD 66%) and 56% (RSD 7.5%) for matrix-bound
Hg in this soil.
After the APS treatment, no more free Hg0 could be
detected in this sample. All Hg was released between 100
and 400 °C (Figure 3). The thin but high peaks between 100
and 200 °C were only observed in samples where visible
droplets of metallic Hg occurred. We suggest that these peaks
represent Hg released from Hg0 droplets which were coated
by mercury sulfide during immobilization, thus preventing
Hg0 from further reaction with the reagent. Additional Hg
desorption at 250-400 °C indicates the occurrence of Hg
compounds that are even more stable than metacinnabar.
Humic acids extracted from CAP before the treatment
both show an overlapped double peak between 150 and about
300 °C, possibly indicating two kinds of Hg binding sites
(Figure 4). The physical and chemical properties of the Hg
binding in humic materials are not well understood, but it
is generally believed that Hg in humic substances is
predominately bound to reduced sulfur groups (12, 19, 20).
However, Hg in humic substances is also reported to be
bound through different processes such as inner- or outersphere complexation, adsorption, or ionic exchange (21),
which could not be distinguished by pyrolysis measurements
alone.
After the APS treatment, the Hg content in HA extracted
from CAP has been decreased to 7684 mg/kg, indicating that
about 40% of the Hg has been desorbed from the HA and
transformed to mercury sulfides. The Hg release curve shows
that most of the humic acid-bound Hg of the APS-treated
sample (Figure 4) is released between 250 and 350 °C, which
VOL. 32, NO. 18, 1998 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2757
FIGURE 3. Hg release curves of CAP before and after 2-fold excess APS and TMT treatment as compared to the APS + HgCl2 and TMT
+ HgCl2 standards.
FIGURE 4. Hg release curves and Hg content of humic acids extracted from CAP before and after the 2-fold excess treatment with APS
and TMT.
is even higher than the Hg release temperature of metacinnabar (Figure 2). The second peak (250-400 °C) of the entire
APS-treated sample (Figure 3) is compatible with HA-bound
Hg, which has reacted with the APS by forming stable
organosulfide complexes. The reasons for the increase of
the Hg-HA bonding stability are unknown. There might be
some cross-linking of the Hg-S group with other humic acid
sulfur groups formed during the immobilization treatment.
Only small amounts of Hg were released below 250 °C, which
is in the same temperature range found for the HA of the
untreated sample (Figure 4), indicating that some of the HAbound Hg was not transformed during the treatment.
After the TMT treatment, the Hg desorption curve of the
sample CAP still shows high amounts of Hg0, indicating that
2758
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 32, NO. 18, 1998
metallic Hg has not been converted during the TMT treatment
(Figure 3). Moreover, no shift of the Hg release temperature
toward that of the TMT-Hg standard substance could be
observed. Hg concentrations in the extracted HA were found
to show almost the same values before and after the TMT
treatment. The Hg release curve of the HA shows an increase
of the Hg release temperature of about 50 °C as compared
to the HA of the untreated sample (Figure 4). We conclude
that TMT, in contrast to APS, does not desorb Hg from HA
preferentially and precipitates as the TMT-Hg complex but
is effectively sorbed or incorporated into the HA, resulting
in greater thermal stability of the HA-bound Hg.
The Hg release curve of untreated KYA shows only a single
peak between 150 and 300 °C (Figure 1), indicating the
FIGURE 5. Hg release curves of KYA before and after 2-fold excess APS and TMT treatment as compared to the APS + HgCl2 and TMT
+ HgCl2 standards.
predominance of matrix-bound Hg forms. The comparatively high concentration of easily reducible Hg (0.25%)
determined in the water extracts (Figure 8) also supports the
assumption that some HgCl2 occurs in this soil. As the curve
does not match that of the HgCl2 standard (Figure 2), we
assume that HgCl2 in this sample does not occur as a free
salt but is predominately sorbed to matrix components.
Previous studies have shown that HgCl2 in this sample is
mostly incorporated into the calcareous matrix of the soil
(9).
Humic acids extracted from untreated KYA also show high
Hg concentrations (Figure 6), but the amount of extractable
HA was much lower than for CAP. In HA of this sample, we
observed an intense formation of Hg0 after extraction of the
HA (Figure 6) that we attribute to the enrichment of HA in
the extract and the reduction of coextracted Hg2+ to Hg0.
Hg2+ reduction, frequently reported to occur in HA (22, 23),
was not observed in the HA extracted from CAP (Figure 4),
which might be caused by the absence of easily reducible Hg
compounds such as HgCl2 in CAP or differences in the HA
potential to reduce Hg2+.
After the APS treatment, KYA shows a bimodal Hg release
curve with the additional peak occurring between 250 and
about 350 °C, indicative of metacinnabar or other mercury
sulfides (Figure 5). According to the size of the two peaks,
most of the Hg was transformed to mercury sulfides. Only
a small amount of Hg remains unaltered, releasing Hg at the
same temperature as the peak of the untreated sample.
In contrast to the HA extracted from the untreated sample,
those extracted after the APS treatment release most of the
Hg in the temperature range between 150 and 250 °C and
only small amounts between 250 and 350 °C (Figure 6). More
than 90% of the HA-bound Hg was desorbed by the APS
treatment and predominately precipitated as metacinnabar
as indicated by Hg release curve of the APS-treated entire
sample (Figure 5). Moreover, no more free Hg0 could be
found in the HA sample after the APS treatment. We assume
that this is mostly due to the transformation of the HgCl2.
However, it is unknown whether the APS treatment changed
the Hg reducing properties of the HA.
After the TMT treatment, sample KYA showed a slight
increase of the Hg release temperature of about 20 °C. As
observed for sample CAP, the Hg release temperature of the
TMT-treated sample KYA is much lower than that of the
TMT Hg standard (Figure 5).
HA extracted from TMT-treated KYA released most Hg
within the same temperature range (200-300 °C) as HA of
the untreated sample but did not show formation of Hg0
(Figure 6). In contrast to HA extracted from CAP and KYA
after the APS treatment, HA extracted after the TMT treatment
shows the same Hg release temperatures, indicating the
formation of similar Hg compounds in both samples. The
high Hg content in HA of KYA indicates again that Hg was
not desorbed from the HA by TMT. There is even an increase
of the HA Hg content that we attribute to complexation of
chloride-mercury by TMT and final incorporation of the
TMT-Hg complex into HA. The different Hg concentrations
and Hg desorption characteristics of the HA extracted from
the two samples indicate that humic acid-bound Hg reacts
differently with the two reagents. We concluded that the Hg
desorption peaks of the HA extracted from the TMT-treated
samples indicate the release of Hg from reduced sulfur groups.
Accordingly, most of the HA bound Hg of the untreated
sample, KYA would be already bound to reduced sulfur groups
as the HA curve widely overlaps with that of the TMT-treated
HA. In contrast, the first peak of the Hg release curve of HA
extracted from CAP indicates that most of the Hg in this
sample is bound less strongly to other functional groups or
by adsorption processes.
The chemical properties of the HA-bound Hg compounds
formed during the treatment are mostly unknown, as Hg
desorption measurements provide only fingerprints and
corresponding standard Hg compounds are not yet found.
Accordingly, the long-term stability of this Hg compound
could hardly be estimated. Despite the stabilty increase of
HA-bound Hg through the treatment with both reagents, it
is still unknown to what extent this HA-bound Hg is still
available to the transformation processes such as Hg2+
reduction to Hg0 or Hg methylation, known to occur as
interactions between HA and Hg or induced by microbiological activity (23, 24). Therefore, despite the transformation
of mobile Hg compounds such as Hg0 or HgCl2, Hg in soils
immobilized with sulfur-containing reagents could be hardly
considered as an inert chemical compound.
VOL. 32, NO. 18, 1998 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2759
FIGURE 6. Hg release curves and Hg content of humic acids extracted from KYA before and after the 2-fold excess treatment with APS
and TMT.
FIGURE 7. Hg release curves of REC before and after 2-fold excess APS and TMT treatment as compared to the APS + HgCl2 and TMT
+ HgCl2 standards.
REC shows the lowest Hg release temperature of all
samples (Figure 1). Due to the lack of organic material in
this sample, we assumed that Hg is predominantly adsorbed
to mineral soil components such as sesquioxides, which
show similar Hg release characteristics between 120 and 220
°C (Figure 2). REC shows the most distinct changes of Hg
release temperatures of all samples after the APS treatment.
The curve of this sample was shifted completely toward the
temperature range of the APS standard substance, showing
that all of the Hg was transformed to metacinnabar (Figure
7).
TMT treatment of REC caused an increase of the Hg release
temperature of about 50 °C but did not reach the temperature
of the TMT-Hg standard (Figure 7), as found for CAP and
2760
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 32, NO. 18, 1998
KYA. The composition of the TMT-Hg compound formed
in this sample is unknown. Although, we could not extract
HA-bound Hg from REC, the TMT-treated sample shows the
same Hg release temperature as KYA where we observed an
enrichment of TMT-Hg in extracted HA. It seems that
TMT-Hg complexes bound to HA cannot be distinguished
from those bound to mineral surfaces. We assume that
pyrolysis measurements show in both cases the breakdown
of similar Hg-S bonding. The high thermal stability of the
TMT-Hg standard substance is attributed to the formation
of TMT-Hg polymers. From the lower Hg desorption
temperature of the soil-bound Hg following TMT treatment,
we concluded that polymerization of TMT-Hg molecules
does not occur in the same way in soils.
FIGURE 8. Percentage of total Hg in water extracts of the samples CAP, KYA, and REC at pH 3, at pH 4.6, and with demineralized water
before and after 2-fold excess APS and TMT treatment. Relative standard deviations (percentage from left to right): CAP, pH 3: 22.5, 18.2,
18, 18.8; pH 4.6: 12.5, 12.5, 38, 35.4; demineralized water: 15, 15, 40, 37.4; KYA, pH 3: 16.7, 18.8, 23.5, 29; pH 4.6: 21, 24.5, 28, 29.4. demineralized
water: 20, 13.6, 29.3, 27.2. REC, pH 3 13.3, 13.6, 37.9, 42.4; pH 4.6: 16.6, 15.8, 38.7, 26.7. demineralized water: 8, 20.5, 34.6, 28.8.
Despite the occurrence of Hg release peaks that indicate
the formation of metacinnabar in all samples after the APS
treatment, the chemical composition and the chemical
properties of the formed mercury sulfides are not well
understood. Moreover, we found that if the reagents are
used at higher concentrations (25-fold excess), mercury
sulfide peaks could not be detected in any of the samples.
It is known that soluble mercury sulfides (SMS) are formed
at excess sulfide concentrations (5, 25) according to
HgS + S2- f HgS22-
TABLE 2. pH Values of the Extracts after Leaching the
Untreated and Treated Samples at pH 3, at pH 4.6, and with
Demineralized Water
pH pH demineralized
3 4.6
water
CAP
KYA
REC
CAP-APS
CAP-TMT
3.9
6.7
3.0
4.2
4.6
4.6
6.1
4.5
4.7
4.7
7.8
7.8
5.6
10.1
10.8
pH pH demineralized
3 4.6
water
KYA-APS
KYA-TMT
REC-APS
REC-TMT
7.1
7.2
3.3
3.0
6.2
6.2
4.6
4.7
10.5
10.6
10.0
10.5
(3)
Hg concentrations in the reagent solutions after the 2-fold
excess APS treatment show that 46% (( 13%) of total Hg was
extracted from REC, 20.5% (( 3.7%) from KYA, and 6.8% ((
2.4%) from CAP. Assuming that most of the Hg was extracted
as SMS, the amount of mercury sulfide formed during the
treatment is actually higher than indicated by the pyrolysis
measurements of the solid phase. During the TMT treatment,
only 2.2% (( 0.38%) of total Hg were extracted from REC,
1.7% (( 0.36%) from KYA, and 3.9 (( 0.71%) from CAP as
TMT does not form soluble complexes with Hg.
Leaching Tests
Hg Concentrations in Water Extracts before Immobilization. Hg mobility in all untreated samples was found to be
greatest in the extracts of the pH 3 leachates (Figure 8),
attributed to increasing desorption or dissolution of Hg bound
to organic or inorganic soil components at low pH (26). The
general increase of the Hg concentrations in the undigested
pH 3 extracts also indicates that a large fraction of the
dissolved Hg at this pH exists as easily reducible Hg2+. The
differences between Hg concentrations found in the water
extract and those in the pH 3 extract depend on the buffer
capacity of the soils. The high carbonate content of KYA
buffers the solutions of pH 3 and pH 4.6 to values higher
than pH 6 (Table 2). The final pH of the pH 3 and pH 4.6
extracts of CAP and REC, however, is lower (Table 2), which
explains the higher Hg concentrations in these extracts
(Figure 8). The undigested water extracts of KYA show nearly
the same Hg concentrations as the digested extracts,
indicating that most of the soluble Hg of this sample occurs
as easily reducible Hg compounds such as HgCl2. All samples
show higher Hg concentrations in the soluble complexes of
the bound fraction than in the undigested extracts (digested
minus undigested). This indicates that the water-soluble
Hg in these samples is predominantly bound to soluble metal
complexes such as fulvic acids, including organomercurials,
which occur in the percolate of contaminated soils (27).
Hg Concentrations in Water Extracts after Immobilization. Hg concentrations in the water extracts generally
increase after treatment with both immobilization reagents.
Moreover, Hg concentrations found in the extracts were
generally lower after the TMT treatment than after the APS
treatment (Figure 8). Unlike the untreated samples, the
treated samples CAP and KYA show a distinct decrease of the
Hg concentration in the extracts with decreasing pH of the
solution. Additionally, Hg concentrations of the pH 3 extracts
of these samples were always lower after the treatment than
before. This tendency was not observed for REC, where the
highest Hg concentrations after both treatments were found
in the pH 3 extracts. In none of the extracts did we observe
any easily reducible Hg, indicating successful immobilization
VOL. 32, NO. 18, 1998 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2761
of this fraction. The extracts generally show higher pH values
after treatment with both reagents due to the high pH (pH
12) of the reagents (Table 2). The water extracts, which in
case of the samples CAP and KYA show the highest Hg
concentrations, also show the highest pH values. From the
high pH of the reagents and the water extracts, we assume
that an important factor contributing to the enhanced
solubility of Hg is the mobilization of Hg-bearing HA from
the soils.
For all samples the amount of Hg found in the APS
solutions is far greater than Hg concentrations in the TMT
solutions (Figure 8). Moreover, the amount of Hg extracted
during the APS treatment increases with decreasing content
of extractable HA, indicating that soluble SMS are more easily
formed from inorganic Hg species than from humic acidbound Hg. Accordingly, the lowest percentage of Hg has
been extracted from CAP and the highest from REC. In
contrast, Hg concentrations in the used TMT solution
decrease with decreasing content of extractable HA in the
samples in the order CAP > KYA > REC. These results
indicate that the formation of SMS is the predominant process
causing the increase in Hg mobility after APS treatment,
whereas TMT mainly increases the Hg mobility by mobilization of HA at high pH.
We conclude that the strong decrease of the Hg concentrations in the acid extracts is attributed to the precipitation
of HA and the decomposition of mercury polysulfides to HgS,
S0, and H2S in the APS-treated samples. The decomposition
of the polysulfides could be observed by the precipitation of
white sulfur (S0) and the smell of H2S during the acid leaching.
Conversely, the lower Hg concentrations in the extracts after
the acid leaching of the TMT-treated samples CAP and KYA
are caused by precipitation of the Hg-bearing HA mobilized
during the treatment.
Different from CAP and KYA, Hg concentrations in the
water extracts of treated REC were not higher than Hg
concentration in the pH 3 extracts (Figure 8). Moreover, Hg
concentrations in the extracts of the TMT- and APS-treated
REC do not show the same high differences as found for the
other samples. The comparatively low Hg concentrations in
the water extract of the APS-treated sample REC are attributed
to the fact that more than 46% of the total Hg in this sample
was previously extracted during the APS treatment.
The TMT-treated sample generally does not show large
differences in the Hg content of the extracts obtained by
leaching at different pH, whereas the APS-treated sample
shows increasing Hg concentrations in the extract after the
pH 3 leaching (Figure 8). As the absence of easily reducible
Hg in the extracts confirms that metacinnabar or TMT-Hg
were not dissolved during the acid leaching, we assume that
one reason for the high Hg concentrations in the pH 3 extracts
(Figure 8) is the low amount of gleyey and organic matrix
components in this soil where HgS or TMT-Hg complexes
could be adsorbed. Moreover, no coprecipitation of HgS or
TMT-Hg by precipitation of HA occurs due to the lack of HA
in this sample. Despite the fact that Hg release curves after
APS and TMT treatment indicate successful transformation
of all Hg in REC (Figure 7), which is assumed to be due to
the lack of organic materials, Hg compounds formed in this
sample seem to be more easily remobilized at low pH for the
same reason.
2762
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 32, NO. 18, 1998
Acknowledgments
This research was funded by the Environmental Ministry of
Baden Württemberg/Germany Projekt Wasser, Abfall, Boden
(PWAB) No. PD 95172.
Literature Cited
(1) Stepan, D. J.; Fraley, R. H.; Charlton, D. S. Remediation of
Mercury-Contaminated Soils: Development and Testing of
Technologies; Topical Reports of the Gas Research Institute;
GRI-94/0402; GSI: Palo Alto, 1995; 40 pp.
(2) Pedroso, A. C. S.; Gomes, L. E. R.; De Carvalho, J. M. R. Environ.
Technol. 1994, 15, 657-667.
(3) Charlton, D. S.; Harju, J. A.; Stepan, D. J.; Kühnel, V.; Schmit,
C. R.; Butler, R. D.; Henke, K. R.; Beaver, F. W.; Evans, J. M. In
Mercury Pollution-Integration and Synthesis; Watras, C. J.,
Huckabee, J. W., Eds.; Lewis Publishers: Chelsea, MI, 1994; pp
595-600.
(4) Thöming, J.; Sobral, L. G. S.; Santos, R. L. C.; Hempel, M.; Wilken,
R. D. Abstracts of the Fourth International Conference on Mercury
as a Global Pollutant, Hamburg, Germany; 1996; p 161.
(5) Findlay, D. M.; McLean, R. A. N. Environ. Sci. Technol. 1981, 15,
1388-1390.
(6) Azzaria, L. M.; Aftabi, A. Water, Air Soil Pollut. 1991, 56, 203217.
(7) Bombach, G.; Bombach, K.; Klemm, W. Fresenius J. Anal. Chem.
1994, 66, 18-20.
(8) Windmöller, C.; Wilken, R. D.; Jardim W. Water, Air Soil Pollut.
1996, 89, 399-416.
(9) Biester, H.; Scholz, C. Environ. Sci. Technol. 1997, 31, 233-239.
(10) Biester, H.; Nehrke, G. Fresenius J. Anal.Chem. 1997, 358, 44464452.
(11) Meili, M. Water, Air Soil Pollut. 1991, 56, 333-348.
(12) Schuster, E. Water, Air Soil Pollut. 1991, 56, 667-680.
(13) Stein, E. D.; Cohen, Y.; Winer, A. M. Crit. Rev. Environ. Sci.
Technol. 1996, 26 (1), 1-43.
(14) Nakamura, Y. Patent Ger. Offen. 2,240,733 (Cl. A 01n), March
8, 1973; Japan Appl. 7163740, August 20, 1971, 17 pp.
(15) Feher, F.; Hirschfeld, D.; Linke, K. H. Z. Naturforsch. 1962, 17b,
624.
(16) Calderoni, G.; Schnitzer, M. Geochim. Cosmochim. Acta 1984,
48, 2045-2051.
(17) Müller, G., Gastner, M. N. Jb. Miner. Mh. 1971, 10, 466-469.
(18) Meili, M.; Iverfeld, A° .; Håkanson, L. Water, Air Soil Pollut. 1991,
56, 439-453.
(19) Hintelmann, H.; Welbourn, P. M.; Evans, R. D. Environ. Sci.
Technol. 1997, 31, 489-495.
(20) Skyllberg, U. L.; Bloom, P. R.; Nater, E. A.; Xia, K.; Bleam, W. F.
Proceedings of Extended Abstracts from the Fourth International
Conference on the Biogeochemistry of Trace Elements, Berkeley,
CA, June 23-26, 1997; pp 285-286.
(21) Kerndorf, H.; Schnitzer, M. Geochim. Cosmochim. Acta 1980,
44, 1701-1708.
(22) Allard, B.; Arsenie, I. Water Air Soil Pollut. 1991, 56, 457-464.
(23) Weber, J. H.; Reisinger, K.; Stoeppler, M. Environ. Technol. Lett.
1985, 6, 203-208.
(24) Weber, J. H. Binding and Transport of Metals by Humic Materials.
In Humic Sustances and their Role in the Environment; Frimmel,
F. H., Christman, R. F., Eds.; John Wiley & Sons: New York,
1988; pp 165-178.
(25) Paquette, K. E.; Helz, G. Environ. Sci. Technol. 1997, 31, 21482153.
(26) Andersson, A. Mercury in Soils. In The Biogeochemistry of
Mercury in the Environment; Nriagu, J. O., Ed.; Elsevier/Holland
Biomedical Press: Amsterdam, 1979; pp 79-112.
(27) Hempel, M.; Wilken, R. D.; Miess, R.; Hertwich, J.; Beyer, K.
Water, Air Soil Pollut. 1995, 80, 1089-1098.
Received for review October 23, 1997. Revised manuscript
received June 9, 1998. Accepted June 25, 1998.
ES9709379