Ecotoxicology and Environmental Safety 112 (2015) 201–211

Contents lists available at ScienceDirect

Ecotoxicology and Environmental Safety journal homepage: www.elsevier.com/locate/ecoenv

The redox processes in Hg-contaminated soils from Descoberto (Minas Gerais, Brazil): Implications for the mercury cycle Cláudia C. Windmöller a,n, Walter A. Durão Júniorb, Aline de Oliveira a, Cláudia M. do Valle c a

Departamento de Química, ICEX, UFMG, Av. Antônio Carlos, 6627, 31270-901 Belo Horizonte, MG, Brazil IFMG—Campus Formiga, Rua Padre Alberico, 44, São Luiz, Formiga, MG, Brazil c IFAM—Campus Manaus Centro, Av. Sete de setembro, 1975, 69020-120 Manaus, AM, Brazil b

art ic l e i nf o

a b s t r a c t

Article history: Received 18 March 2014 Received in revised form 6 November 2014 Accepted 11 November 2014 Available online 24 November 2014

Investigations of the redox process and chemical speciation of Hg(II) lead to a better understanding of biogeochemical processes controlling the transformation of Hg(II) into toxic and bioaccumulative monomethyl mercury, mainly in areas contaminated with Hg(0). This study investigates the speciation and redox processes of Hg in soil samples from a small area contaminated with Hg(0) as a result of gold mining activities in the rural municipality of Descoberto (Minas Gerais, Brazil). Soil samples were prepared by adding Hg(0) and HgCl2 separately to dry soil, and the Hg redox process was monitored using thermodesorption coupled to atomic absorption spectrometry. A portion of the Hg(0) added was volatilized (up to 37.472.0%) or oxidized (from 367 7% to 88 716%). A correlation with Mn suggests that this oxidation is favored, but many other factors must be evaluated, such as the presence of microorganisms and the types of organic matter present. The interaction of Hg with the matrix is suggested to involve Hg (II)-complexes formed with inorganic and organic sulfur ligands and/or nonspecific adsorption onto oxides of Fe, Al and/or Mn. The kinetics of the oxidation reaction was approximated for two first-order reactions; the faster reaction was attributed to the oxidation of Hg(0)/Hg(I), and the slower reaction corresponded to Hg(I)/Hg(II). The second stage was 43–139 times slower than the first. The samples spiked with Hg(II) showed low volatilization and a shifting of the signal of Hg(II) to lower temperatures. These results show that the extent, rate and type of redox process can be adverse in soils. Descoberto can serve as an example for areas contaminated with Hg(0). & 2014 Elsevier Inc. All rights reserved.

Keywords: Mercury cycle Speciation Contaminated soils Redox process

1. Introduction Hg contamination of soil and/or sediment occurs globally. The most toxic species are short-chain organomercury compounds, as they can cross cell membranes more easily. Hg is bioaccumulated and biomagnified, and it can be transported through the atmosphere over long distances. Therefore, the metal can be found in regions where no observed natural or anthropogenic sources exist (Andrade et al., 2012), and the metal cycle in the environment is a subject of study by many research groups. Hg compounds are distributed in the environment as volatile species, for example, Hg(0) and dimethylmercury, reactive or soluble species in water, such as Hg(II), HgX2, HgX3
and HgX42
(where X¼ Brˉ, Clˉ or OHˉ), and less reactive species such as HgS and Hg(CN)2 (Guedes, 2009). Compounds of organic and inorganic Hg are generated in various industrial activities, mainly the n

Corresponding author. Fax: þ 55 31 3499 5700. E-mail address: [email protected] (C.C. Windmöller).

http://dx.doi.org/10.1016/j.ecoenv.2014.11.009 0147-6513/& 2014 Elsevier Inc. All rights reserved.

pharmaceutical, paper, electrochemical and pesticide industries and power plants (Mishra et al., 2005; Wang et al., 2012). In aqueous systems contaminated with Hg, a portion of the Hg may volatilize into the atmosphere. By way of rain, Hg may return to aqueous and terrestrial systems. Hg is easily transported in the atmosphere mainly in the form of Hg(0) and, to a lesser extent, as dimethylmercury. In the air or after deposition, Hg(0) can be converted into Hg(II) and organic forms and can bioaccumulate in various biotic and abiotic resources, such as fish, bryophytes, birds, soil, sediments and water (Palmieri et al., 2006). Physico-chemical soil parameters such as the pH, cation exchange capacity (CEC), particle size, organic matter content (OM) and types of clay minerals present play important roles in the variation of the oxidation state and the type of interaction with Hg soil components and, therefore, their mobility and availability (Rezende, 2009). The oxidation of metallic Hg, for example, can form soluble species (Hg (II)) and therefore make Hg more mobile in the environment because elemental Hg is barely soluble. In the case of the contamination of soil, sediment or other matrices, the choice of the

202

C.C. Windmöller et al. / Ecotoxicology and Environmental Safety 112 (2015) 201–211

samples using thermo-desorption coupled atomic absorption spectrometry (TDAAS) and determining the total Hg (THg) using hydride-generation atomic absorption spectrometry (HGAAS).

type of intervention for decontamination also depends heavily on the species of the metal present (Raposo et al., 2003). In addition to the well-known disaster in Minamata, Japan, in 1956, other cases of contamination have been reported, such as soil contamination in cinnabar extraction areas in Idrija, Slovenia (Biester et al., 2000; Gosar and Zibret, 2011; Gosar and Tersic, 2012; Tersic and Gosar, 2012; Tersic et al., 2011a, 2011b, 2014); San Joaquim, Mexico (Martinez-Trinidad et al., 2013); Almaden, Spain (Ruiz-Diez et al., 2012); Xunyang (Qiu et al., 2012) and Guizhou (Li et al., 2012), China; mining activities and/or gold mining for example, in the soils of Bolivia (Teran-Mita et al., 2013), Paracatu (Mulholland et al., 2012; Rezende et al., 2011), the Iron Quadrangle (Windmöller et al., 2007) and Descoberto (Durão Jr. et al., 2009) in Brazil, and industrial waste for example, the sludge from the petrochemical complex of Bandar Imam, Iran (Nik et al., 2013) and the industrial-waste landfills of Vale dos Sinos, Southern Brazil (Augustin and Viero, 2012). In 2003, the presence of high concentrations of Hg in a small area located in the rural municipality of Descoberto in Minas Gerais/Brazil was discovered (CDTN/FEAM, 2005). The contaminated area consists of fairly steep terrain, and it is right next to the Rico stream. This stream is a tributary of the Ribeirão do Grama River that belongs to the Paraíba do Sul River basin. The Fundação Estadual do Meio Ambiente (FEAM) made a study of Hg concentrations in stream and rivers waters and mapped the area in terms of its metal concentration using a network and in-depth sampling (CDTN/FEAM, 2006). Several scientific studies have been conducted in the area to collect information about metal speciation (Durão Jr. et al., 2009; Varejão et al., 2009). The work of Durão Jr. et al. (2009) stated that “a large part of the metal, which had contaminated the environment as Hg(0), oxidized and is principally bound to Fe, Mn and Al oxyhydroxides” and up to 30% of it can be bound OM present in the soil. Fractionation studies in the same work showed that the Hg is only slightly available to the solubilization conditions that normally occur in the environment. The oxidation state of the metal is essentially important in his mobility and even the possibility of methylation reactions, which are key steps in the bioaccumulation and biomagnification of the metal process. The objective of this work was to study the kinetics of oxidation, reduction and volatilization of Hg in the soils from the contaminated area of Descoberto (Minas Gerais/Brazil) by (1) spiking samples with various species of Hg and (2) monitoring the

2. Experimental 2.1. Study area, soil sampling and sample preparation Fig. 1 shows the map of the studied area that was selected because of contamination by Hg. The darkest part of the plot indicates Hg concentrations higher than 2.5 mg kg
1 in samples collected at different depths (CDTN/FEAM, 2005, 2006). A rainwater containment barrier that drains the contaminated area was constructed, and boxes and sedimentation tanks of solid material were also installed, as shown in the figure. The collection points selected were those likely to have low THg concentrations because the intention was to spike the samples with Hg. These collection points were surface samples (up to 30 cm) at D4, D6 and B7 shown in Fig. 1. Although the number of points was small, we can consider these to be representative of the area. The area is not large and is comprised of one type of soil with punctual variations in composition. Using a manual, stainless-steel auger sampler, the soil samples (D4, D6 and B7) were collected, packed into polyethylene bags and refrigerated at 4 °C. In the laboratory, the soil samples were dried at room temperature and sieved to particle sizes o2.0 mm. All analyses were conducted on this soil fraction. 2.2. Analytical procedures for determining the OM and Al, Fe and Mn content The soil pH was measured using a glass electrode in suspensions of the soil samples in distilled water in the proportion of 1:2.5 (Embrapa, 1999). The percentage (%) by weight of OM was determined by the Walkley–Black method according to Embrapa (1999). The method consists of the oxidation of an OM solution of potassium dichromate in the presence of sulfuric acid. The excess dichromate solution is then titrated with ferrous ammonium sulfate. The contents of Al, Fe and Mn were determined by aqua-regia extraction (3:1; HCl:HNO3). The mass of the samples was approximately 1 g, and the volume of aqua-regia used was 10 mL in a

D6 D4 B7

Fig. 1. Localization of the city of Descoberto, State of Minas Gerais, Brazil and a map of the contaminated area according to Hg contents displaying the sampling points (Durão Jr. et al., 2009).

C.C. Windmöller et al. / Ecotoxicology and Environmental Safety 112 (2015) 201–211

50-mL vial. The samples were then agitated for (18 74 h) and then diluted to 40 mL. The samples were then centrifuged, and the supernatant solution was analyzed by flame atomic absorption spectrometry (FAAS) on a Perkin-Elmer model AAnalyst 200 (Shelton, USA). Certified reference materials (CRMs) NIST SRM 2711—Montana Soil and NIST SRM 1944—New York/New Jersey Waterway Sediment were analyzed to evaluate the accuracy of the methods. Except for the determination of pH and OM, all analyses were performed in triplicate. 2.3. Total Hg analysis (THg) The THg was determined by extraction at room temperature in aqua-regia (3:1; HCl:HNO3). A volume of 10 mL of aqua-regia was added to a 50-mL vial with a maximum of 2.0 g of sample. The sample was then agitated for (18 74 h), diluted to 40 mL, and centrifuged for 20 min at 3000 rpm. The supernatant was analyzed by hydride generation (model FIAS 100, Shelton, USA) coupled to an atomic absorption spectrometer (HGAAS) from Perkin-Elmer (model AAnalyst 200, Shelton, USA). To evaluate the accuracy of the method, CRMs NIST SRM 2711 and NIST SRM 1944 were analyzed, and all measurements were performed in triplicate.

203

2.5. Mercury thermo-desorption analysis (TDAAS) Mercury thermo-desorption curves were determined using an in-house apparatus in accordance with Windmöller et al. (1996). This apparatus consisted of an electronically controlled heating unit (25 cm in length and internal diameter 2 cm) and a Hg detection unit. To detect Hg, a quartz tube, in which the thermally released Hg is purged, is placed in the optical system of an atomic absorption spectrometer (GBC 932-AA). Mercury is detected at 253.7 nm. The analysis was performed at a heating rate of 33 °C min
1 from room temperature up to 550 °C with a nitrogen gas flow of 200 mL min
1. Interferences, mainly from pyrolytic products of organic matter, were compensated by continuous deuterium background correction. The sample weight was  200 mg (spiked samples). The results are depicted as Hg thermo-desorption curves which show the release of Hg versus temperature. These curves were compared to standard thermodesorption curves for mercury compounds from previous works obtained with the identical equipment and operational conditions (Valle et al., 2005, 2006).

3. Results and discussion 3.1. Sample Physico-chemical characterization

2.4. Spiking of soil samples with Hg species As described in Section 2.1, samples with low THg contents (o 0.2–0.37 mg kg
1) were selected so that the Hg already present was much less than the amount of Hg added. Therefore, the already present Hg does not interfere in the interpretation of the test results. Portions of the samples (approximately 100 g of air dried soils) were spiked separately with Hg(0) and HgCl2 at a final concentration of E30 mg kg
1 (THg) by dry dilution. The incubation cells were dark glass bottles with Teflon lined lids. After spiking and homogenization, the samples were subdivided into two equal parts. One part was maintained at room temperature (RT) to simulate natural conditions, and the other part was maintained at low temperature (LT), 4 °C in sealed containers in the refrigerator to prevent contact with light. The intention was to evaluate the influence of temperature, notably in the volatilization process. All samples were periodically monitored by TDAAS until the thermodesorption curves obtained did not show any change. The determination of THg was performed simultaneously to verify whether volatilization of the metal occurred in addition to the redox processes. The % of Hg volatilized was calculated as the 100% concentration of THg in the first analysis of the sample after spiking. The difference in the levels of THg between the first analysis and the analysis on other days of monitoring is the % of Hg volatilized. The analyses of speciation were performed in two or more replicates.

Table 1 shows the comparison of the certified values of Al, Fe, Mn and THg for the CRMs and the values obtained by digestion of the samples in aqua-regia. The Student-t test shows that the results agree to a confidence level of 95%. Table 1 also shows the results of the chemical characterization of the studied samples (D4, D6 and B7). The limits of detection (LOD) obtained for Fe, Mn, and Al were 0.063% m m
1, 2.8 mg kg
1 and 0.08% m m
1, respectively. The limits of quantitation (LOQ) obtained for Fe, Mn, Al and Hg were 0.190% m m
1, 8.5 mg kg
1, 0.24% m m
1 and 0.20 mg kg
1, respectively. These values were calculated for a solid-sample mass of approximately 1 g. Because the samples are from the same area, certain measured parameters are similar. The contents of THg ranged from o0.20 to 0.37 mg kg
1, this range is low because earlier studies in this area found up to 160 mg kg
1 (Durão Jr. et al., 2009). The pH values were slightly acidic, and the content of OM, an important parameter in the geochemistry of Hg, varied little (1.62–2.71% m m
1). The values found for Al and Fe were also similar, ranging from 0.807 to 1.2 and 1.9 to 2.1% m m
1, respectively. The only parameter that displayed a notably larger range of variation than the others was the amount of Mn, which ranged from 23.7 to 127 mg kg
1, indicating specific variations in the mineralogical composition of the site with respect to minerals containing this metal.

Table 1 Physico-chemical characterization of the samples and CRMs. Parameter


1

Al (% m m ) Fe (% m m
1) Mn (mg kg
1) THg (mg kg
1) OM (% m m
1) pH (H2O) a

Mean 7 S.D., n¼3.

Determined valuea

Certified value NIST 1944

NIST 2711

NIST 1944

NIST 2711

D4

D6

B7

5.3370.49 3.53 7 0.16 505 7 25 3.4 7 0.5 – –

6.53 7 0.09 2.89 7 0.06 638 7 28 6.25 7 0.19 – –

4.3570.36 3.42 7 0.32 5187 41 3.417 0.01 – –

5.87 7 0.50 2.727 0.15 563 7 30 6.90 7 0.31 – –

0.8077 0.006 1.9 7 0.1 23.7 70.4 o 0.2 1.6 5.6

1.187 0.03 2.1 70.1 1277 6 o 0.2 2.7 5.8

1.2 70.1 2.1 7 0.1 607 3 0.37 70.04 1.7 5.6

204

C.C. Windmöller et al. / Ecotoxicology and Environmental Safety 112 (2015) 201–211

studies (Valle et al., 2005, 2006) using the same equipment and operational conditions of this work. Hg(0) was released in the range between room temperature and 200 °C with a maximum at 150 °C; Hg(I) was released soon after with a maximum at approximately 200 °C, yielding a sharper signal than Hg(II) compounds (temperatures 4 200 °C). According to Valle et al. (2005), “even though there is a large temperature range in which peaks overlap, the peak of Hg bound to humic acid Fig. 2B pears at a higher temperature, from 250 to 430 °C, and Hg bound to mineral phases appears at lower temperatures… TDAAS does not always show the separation of mineral and organically bound Hg in soil clearly” Other studies (Biester and Scholz, 1997; Biester et al., 2000, 2002; Sladek et al., 2002; Gosar et al., 2006; Tersic et al., 2011a, 2011b; Pavel et al., 2014; Coufalik et al., 2014 and Bavec et al., 2014) have also shown that the thermodesorption temperatures of Hg standards are different, indicating that this technique provides data on speciation. Temperature ranges of the release of Hg standards varied because of differences in system operational conditions, mainly the gas flow (200 mL min
1 in this work and 300 mL min
1 in other works). The temperature heating rate was similar in the cited works (30 or 33 °C min
1). However, all studies agree that Hg(0) is released at a low temperature, followed by Hg (II) binding. Although many samples from this study area examined in previous studies have high levels of THg (Durão Jr. et al., 2009), samples with lower metal contents ( o0.20–0.37 mg kg
1) were selected for this study. Thermograms with approximately 2 g of the samples showed small concentrations of Hg, between 200 and 300 °C, corresponding to Hg(II). The strategy used was to ensure that the signal of Hg(II) present in the samples would not interfere in the signals to be monitored after spiking (corresponding to approximately 30 mg kg
1 of THg).

According to the reports of CDTN/FEAM (2005 and 2006), studies using X-ray diffraction showed that the soils in this area are composed of mixtures of quartz and kaolinite (430%), kaolinite, gibbsite and goethite (o 30%), and gibbsite and ilmenite (o10%). The minority phase (o3%) consists of magnetite, microcline, muscovite, hematite and monazite. Some of these minerals contain Si, Al and Fe in their structures. The values measured by X-ray diffraction are reported in percentages. The concentrations found for Mn are on the order of magnitude of mg kg
1, therefore Mn would not be detected by X-ray diffraction. No particle size analysis of these samples was performed because they have been extensively studied in a previous work by Durão Jr. et al. (2009). These soils have a % of silt plus clay typically ranging from 2.6 to 5.3%, i.e., they have high sand contents. A discussion of the speciation, distribution and transport of Hg and the influence of the physico-chemical parameters on the behavior of natural Hg in these samples was presented in Durão Jr. et al. (2009). A notable conclusion in that study was that the thermodesorption analysis revealed the presence of mainly Hg(II). Data are available for soils from another mining area (the state of Mato Grosso), in which Hg(0), Hg(I) and Hg(II) were found (Windmöller et al., 1996). These results indicate that Hg(0) (the source of metal contamination) undergoes oxidation. That study also showed the presence of oxidized metal in soils contaminated by waste from the chlor-alkali industry in Germany. 3.2. Study of the conversion processes between Hg species Calculations of percentage volatilization were performed using the GHAAS THg determination data which were always above the LOQ (0.2 mg L
1); the results showed an average relative standard deviation (RSD) of 5.2%. In the calculations of the % oxidation and reduction, measurements were obtained from replicates of the signal areas of the thermograms, and the average of the RSD was 14.6%. The thermodesorption results of Hg standard samples are shown in Fig. 2A and B. These results were obtained in previous

1.2 1.0 0.8 0.6 0.4 0.2 0.0

3.2.1. Spiking of soil samples with Hg(0) Fig. 3 shows the most representative thermograms obtained during the monitoring of samples (D4, D6 and B7) doped with Hg (0) and maintained at two temperature conditions.

HgO

0.8

0.4

1.5

HgCl2

0.2 0.0 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 1.2 1.0 0.8 0.6 0.4 0.2 0.0

1.0

Hg2Cl2

Hg release

Hg release

0.6

Humic Acid Al-Goethite Goethite Gibbsite Kaolinite

0.5

Hg(0) 0.0

100

200

300

400 o

Temperature C

500

100

200

300

400 0

Temperature C

Fig. 2. Thermograms of standard samples of mercury compounds from Valle et al. (2006) (A) and Valle et al. (2005) (B).

500

C.C. Windmöller et al. / Ecotoxicology and Environmental Safety 112 (2015) 201–211

2

D4(RT)+Hg(0)

1

1

10 d

10 d

1 0

0 2

44 d

1 0 2

90 d

1

Hg release

2

44 d

1 0 2

90 d

1 0 2

0 2

140 d

1

180 d

1 0 2

0 2

200 d

1

200 d

1

0

0 2

D6(LT)+Hg(0) 1h

1

D6(RT)+Hg(0)

1

1h

0 1

0 1

1d 0 2

6d

1 0 1

1d Hg release

Hg release

1d

0 2

1

Hg release

D4(LT)+Hg(0)

1d

0 2

205

0 2

6d

1 0 1

90 d

90 d 0 2

0 1

110 d

110 d

1

0

0 1

1

B7(RT)+Hg(0)

B7(LT)+Hg(0)

1h 0

0

1

1

1h 1d

1d 0

0

1

0 1

Hg release

1

7d

7d 0

1

41 d 0 0.5

41 d 0 1

83 d 0.0 0.5

83 d 0 1

99 d 0.0

99 d 0

100

200

300

400

500

Temperature (oC)

100

200

300

400

500

o

Temperature ( C)

Fig. 3. Thermograms of samples spiked with Hg(0) for the conditions of LT and RT.

In general, with the addition of Hg(0) to the soil samples, Hg (0) oxidizes to Hg(II), forming generally broad peaks with a range of desorption between 250 °C and 400 °C and a maximum

absorbance at a temperature of E300 °C. This oxidation is independent of the temperature conditions during the storage of samples. In several samples (indicated by arrows in Fig. 3), a third

206

C.C. Windmöller et al. / Ecotoxicology and Environmental Safety 112 (2015) 201–211

peak (or a shoulder indicating a third peak) is noted between a peak at temperatures lower than 100 and a peak at temperatures higher than 250 °C. This additional peak or shoulder appears in some thermograms during the monitoring and disappears at the end to form a peak that begins above 250 °C (samples D4RT, D4LT, D6RT, and D6LT). In the case of the samples B7RT and B7LT the peak of Hg(0) shifts to slightly higher temperatures, indicating a possible mixture of Hg (0) and Hg(I). Previous studies attribute the peak at approximately 220 °C to Hg (I) (Valle et al., 2006). This previous study investigated Hg oxidation and reduction in soil samples from the Amazon. This peak description is reasonable because the thermograms of doped samples often display the appearance and disappearance of this signal, as in Valle et al. (2006) and this work. This signal does not consistently appear always because of the instability of the specie. This indicates that the Hg(0) added is first oxidized to Hg(I) and subsequently to Hg(II) (above 250 °C). Other authors have reported (near 200 °C) the release of “Hg in forest soil” (Bavec et al., 2014) and noncinnabar Hg forms (Hg sorbed to organic or mineral soil or sediment components) (Biester et al., 2000). However, these studies employed a carrier gas flow 50% greater than that used in this work. This difference complicates a direct comparison of the results. The release of Hg bound to humic substances obtained using the current conditions is shown in Fig. 2B, with a maximum absorbance occurring at approximately 300 °C. Differences between the temperature conditions can be evident only with a quantitative interpretation of the data, for which the area calculations of a greater number of thermograms shown in Fig. 3 were used. Comparing the three samples, sample D6 appears to have experienced a faster oxidation of the added Hg, and sample B7 shows the presence of Hg(0) even when the monitoring showed “stabilization” of the added Hg. These results are related to the redox reactions of Hg in soils that can be induced by light, notably by UV light. In addition, soil conditions such as the pH, the amount and type of OM and the concentration of Fe and Al may affect the redox kinetics of Hg (Wallschläger et al., 1998; Magarelli and Fostier, 2005). Other metals that can present various oxidation states, such as Mn, also interfere in the redox conditions of the environment, as well as microorganisms. Several minerals containing Fe were found in all of the soil samples from Descoberto/MG (Durão Jr. et al., 2009) that were analyzed by X-ray diffraction. Mn, analyzed by FAAS, was found in smaller concentrations (Table 1). However, the quantities of Mn were large compared to the natural THg values of these samples and on the order of the Hg added to the samples. Valle et al. (2006) also observed the oxidation of the Hg(0) added to Amazon samples and suggested that Fe3 þ is the species responsible for the oxidation of Hg(0). Furthermore, they stated that the OM can be the main source of the negative charge in the soil that contribute to a higher affinity for metallic cations, favoring the oxidation of Hg(0) because of the complexation and stabilization of the oxidized metal. The OM of soil displays groups such as acetic acid (-COOH), phenols (–C6H5OH), hydroxides (–OH), sulfhydryls (–SH) and amines (–NH2). At pH 45, the organic constituents contribute positively to the CEC of the soils, providing sites for the adsorption of Hg (Fiorentino et al., 2011; Licht et al., 2007). In general, the OM is responsible for up to 90% of the CEC (Melamed and Villas Bôas, 2002; Linhares, 2009) of soils. The average pH in the samples studied in this work was 5.7, indicating that the adsorption of cations is favored. Strong Hg(II)–sulfur (Hg–SR) bonds in natural organic matter have been reported (Nagy, et al., 2011), and reactions between Hg (II) and inorganic sulfides are also of note (Skyllberg and Drott,

2010). Bouffard and Amyot (2009) also reported that the OM can intensify the process of Hg(0) oxidation because of its complexation with Hg(II). No recognizable pattern was noted in the thermograms that differentiated the probable interactions. However, an interaction with the organic and mineral phases is inferred, notably the organic and inorganic sulfur ligands and/or nonspecific adsorption onto oxides of Fe, Al and/or Mn. To confirm this inference, systematic studies using X-ray absorption spectroscopy are required. Although several parameters can be suggested as possible oxidants of Hg(0), identifying the species affecting this process in soils is complex (Windmöller et al., 2007). Sample B7 likely has two peaks at the end of the monitoring because it has a lower content of OM (Table 1); that is, oxidation is less favored in this case. The peak at the lower temperature is likely because of Hg(I). However, Hg(II) adsorption is also possible in some mineralogical phase, such as the Hg species that appeared in the temperature range between 180 and 240 °C. Valle et al. (2006) spiked uncontaminated Amazonian soils with Hg(0) in conditions comparable to this study and found similar results for a sandy sample. The % Hg volatilized was evaluated by monitoring the THg. The % oxidation of Hg(0) added was calculated using the ratio of the area of the Hg signal and the sample mass submitted for analysis by TDAAS. It was calculated considering the Hg remaining after volatilization. Whereas the Hg(0) begins to be released from ambient temperature to a maximum of 200 °C (Windmöller et al., 1996), any Hg signal from 200 °C to 550 °C (the temperature range limit of the thermodesorption equipment) was considered to represent oxidized Hg(0). The samples were monitored until the thermograms were unchanged; this time ranged from 99 days (sample B7) to 200 days (sample D4). Table 2 shows the results of the % Hg(0) oxidized and volatilized at the end of the monitoring periods. The data show that the extent of oxidation was generally high, reaching almost 100% for samples D4 and D6 for the two storage conditions. Sample B7 shows a lower extent of 87% at room temperature and 54% at low temperatures. Regarding volatilization, this sample, which showed less oxidation, showed the highest % volatilization. Therefore, when Hg is oxidized, it interacts with the matrix and is fixed; when Hg is present as Hg(0), it is more easily volatilized. According to Choi and Holsen (2009), most of the volatile Hg in soils is Hg(0). Gustin and collaborators (2002) confirmed that sunlight is one of the most important factors in Hg emissions. Although the containers used for storing the samples at room temperature were dark, some passage of light should be allowed. The samples that were refrigerated were more protected from contact with light, thus reducing the volatilization of Hg(0). Beyond solar radiation, other parameters, such as moisture and temperature, influence the process of volatilization of Hg in contaminated soils. Studies by Almeida and Souza (2007) demonstrated that the meteorological Table 2 Percentage of Hg(0) oxidized and volatilized at the end of the monitoring period. Amostras

Time (day)

Hg(0) (%) Oxidized

D4 D6 B7

200 110 99

LT ¼ low temperature. RT ¼room temperature.

Volatilized

LT

RT

LT

RT

80 713 88 716 36 77

76 714 80 713 54 710

16.9 7 0.9 10.4 7 0.5 33.3 7 2.0

22.17 1.0 18.6 7 1.0 37.4 7 2.0

C.C. Windmöller et al. / Ecotoxicology and Environmental Safety 112 (2015) 201–211

parameters showed a strong correlation with the flow of gaseous Hg in contaminated soils. Bahlmann et al. (2006) also demonstrated that the emission of Hg is strongly related to the exposure of the sample to solar radiation and temperature. Experiments by Choi and Holsen (2009) showed that, despite the evident relationship between the volatilization of Hg and solar radiation, soil temperature is the most important parameter for the volatilization of Hg. In general, the volatilization ranged from 10.7% to 37.4% and was lower in samples stored at lower temperatures. The differences between the volatilization at low and room temperatures for the same sample were between 4.0% and 8.1% for the studied samples. Considering the difference of the temperatures ranged from 15 to 20 °C during the monitoring period, contrasting values of volatilization between the two conditions were not observed.

207

Graphs of the kinetics of oxidation were constructed with the % of oxidized Hg(0) and time. We used the equation for first-order kinetics, similar to a previous study (Valle et al., 2006). The consumption of a generic reagent “A” is represented by the following equation:

d[A]/dt = − k[A]

(1)

and its solution can be written as the following:

ln([A]/[A]0 ) = kt.

(2)

The % of Hg(0) oxidized was converted into the natural logarithm (ln) and related to time (Fig. 3). These graphs show that the mechanism of Hg oxidation occurs in two steps, which have been adjusted for two first-order reactions that yield two linear equations corresponding to steps I and II.

5.0

2.5

D4+Hg(0)

D4+Hg(II)

4.0 3.5 3.0

2.0

2.5 2.0

Hg(II) peak area decrease

Hg(II) peak area increase

4.5

1.5 1.0

1.5 0

50

100

150

200

0

20

40

60

80

100

5.0

2.5

D6+Hg(II) 2.0

4.5 1.5

1.0

4.0 0.5

Hg(II) peak area decrease

Hg(II) peak area increase

D6+Hg(0)

0.0

3.5 0

20

40

60

80

100

120

0

20

40

60

80

100

5.0

2.2

B7+Hg(II)

B7+Hg(0)

2.0 4.0

3.5

1.8

3.0

1.6 2.5

2.0

1.4

1.5 0

20

40

60

Time (days)

80

100

0

20

40

60

80

100

Time (days)

Fig. 4. Plots of the kinetics of oxidation of the added Hg(0) (a) and of reduction of the added HgCl2 (b) in the soil samples.

Hg(II) peak area decrease

Hg(II) peak area increase

4.5

208

C.C. Windmöller et al. / Ecotoxicology and Environmental Safety 112 (2015) 201–211

Table 3 Half-lives (t1/2) obtained from the kinetics graphs for the oxidation and reduction of Hg in soil samples spiked with Hg(0) and HgCl2, stored under the conditions LT and RT. Sample

Step

Hg(0) oxidized

Hg(II) reduced

LT

RT

LT

RT

I II

5 231

2 270

14 693

12 693

D6

I II

2 139

1 139

6 161

5 116

B7

I II

3 128

2 107

7 100

ND 100

t1/2 (day) D4

RT ¼room temperature. LT ¼ low temperature. ND¼ not determined.

The value of k was obtained by the angular coefficient of the linear fit. The parameter that is obtained for chemical reactions is the first-orderhalf-life (t1/2), corresponding to the time interval in which the % Hg(0) oxidized becomes half of its initial value. The calculation is given by the following equation:

t1/2 = (/k), being k = (y2 −y1)/x2−x1).

(3)

As the thermograms (Fig. 3) showed, only Hg(II) was present at the end of the monitoring period (except for in sample B7). Step I, the faster step, corresponds to the conversion of Hg(0)/Hg(I), and step II, which is slower, corresponds to the conversion of Hg(I)/Hg (II), as observed by Valle et al. (2006). The information obtained from the graphs in Fig. 4 is summarized in the half-life (t1/2) values for the two steps of the oxidation processes (Table 3). The first observation displays that the t1/2 of the first step (2–5 days) is smaller than that of the second (107–270 days). The D4 sample showed a second step that is slower than the identical step for the other samples. Because this sample had the lowest Mn content, we hypothesize that Mn is one of the major species responsible for the Hg-oxidation process. Small differences between the two storage conditions were not clear because of the random errors of the entire process. These results show that differences in the processes of oxidation of Hg can be considered important for the biogeochemical cycle of metal. Information that can better explain these processes are therefore relevant. 3.2.2. Spiking of soil samples with HgCl2 The thermodesorption curves of samples spiked with HgCl2 under both temperature conditions are shown in Fig. 5. In general, a small shift of the signal of Hg(II) to lower temperatures is observed compared to those on the first-day thermograms. This shift is characterized by the formation of broad peaks with a range of desorption between 180 °C and 450 °C and a maximum absorbance at a temperature of E300 °C, independent of the storage conditions of the samples. The thermograms show that, in the days following the start of the experiment, the Hg(II) desorption temperature range varied little, indicating that a possible reduction process or a new interaction of the metal with the matrix occurred to a small extent. For example, for sample D4 under the RT condition, a small shoulder characterizing the reduction of Hg(II) (the sign of Hg below 200 °C) was observed after 44 days of monitoring. Valle and coworkers, in 2006, attributed the changes in the Hg in soil samples from Amazon spiked with HgCl2, to the reduction to Hg(I) and

Hg(0). They observed the formation of three peaks in the monitoring better than in this case. Table 4 shows the % of reduced Hg(II) and the Hg volatilized in samples D4, D6 and B7 under both temperature conditions. The calculation of the % reduced Hg(II) was obtained using the identical method used to calculate the % of oxidized Hg(0). The area of the reduced Hg(II) was considered to be from room temperature (25 °C) to 200 °C. The % of reduction was calculated considering the Hg remaining in the matrix after volatilization. The % of reduced Hg(II) was small compared to that of oxidized Hg(0), indicating that oxidation is more favored than reduction in these matrices. A similar result was obtained in a previous study (Valle et al., 2006) that stated that these differences are because of the presence of oxidants in the matrix (such as Fe3 þ ) and to the presence of other components that easily stabilize Hg(II). The extent of reduction varied from 5 71% to 9 71%. However, sample D4 shows a slightly larger extent of reduction than the others. Volatilization also occurred to a lesser extent than in the case of the samples spiked with Hg(0). The Hg added in the form of Hg(II) is assumed to be reduced before being volatilized, which explains the greater volatilization in the first case, for which the reduced form was in a higher concentration and therefore more easily volatilized. The range of volatilization was from 2.370.1% to 7.370.4%, whereas in the case of spiking with Hg(0), the range of volatilization was from 10.470.5% to 37.472%. These values are significantly different. The differences between the storage conditions for samples D6 and B7 showed the expected results: higher volatilization at room temperature than at low temperatures. Differences in the volatilization were much smaller than in the case of oxidized Hg(0), likely because of the smaller random errors in all measurements. The graphs of the kinetics of the reduction of Hg(II) (Fig. 4) show that reducing the Hg(II) also occurs in two steps. Step I was attributed to the mechanism of Hg(II)/Hg(I) reduction, and step II is attributed to Hg(I)/Hg(0). The t1/2 values obtained for the first step of the reduction processes were higher than those obtained for oxidation, ranging from 5 to 14 days. This higher value demonstrates a greater resistance to reduction. Observing the t1/2 for the second step, this is more evident for sample D4, whose value is 693 days, i.e., twice the value obtained for sample D6. Sample D4 displayed slower kinetics for both oxidation and reduction, likely because the metal Mn plays an important role in these two processes. As stated earlier, sample D4 has the lowest Mn content; therefore, Mn could be important to both the oxidation and the reduction processes. Additionally, the various oxidation states complicate the identification of species affecting this process in soils. Solar radiation, the constituents of OM and microorganisms (Choi and Holsen, 2009) must also be considered. Studies by Valle et al. (2006) on the reduction of Hg(II) in soil samples from the region of Manaus (Amazon) did not observe a clear correlation between the OM content and the extent of reduction. These authors worked with soil samples with a higher % of OM than the samples used in this study. However, the t1/2 values obtained by reduction to Hg(II) were similar, ranging from 4 to 14 days in the first step and 301– 408 days in the second step. Unlike other metals, Hg can be volatilized from the soil/air interface in the form of Hg(0) or, to a lesser extent, as other volatile compounds. The % of volatilized Hg in Hg(II)-spiked samples was low when compared to that of the same samples spiked with Hg (0). A previous study by Scholtz and collaborators (2003) confirms that Hg(II) tends to be only minimally reduced and volatilized in soil samples throughout the year, even in differing climatic conditions. Kinetic studies of the conversion of Hg(0)/Hg(II) in the Descoberto samples showed that the mechanism of oxidation of the samples spiked with Hg(0) occurred faster than the reduction of

C.C. Windmöller et al. / Ecotoxicology and Environmental Safety 112 (2015) 201–211

Hg release

2.0 1.5 1.0 0.5 0.0

D4(LT)+Hg(II) 1d

2.0 1.5 1.0 0.5 0.0

8d

2.0 1.5 1.0 0.5 0.0

44 d

2.0 1.5 1.0 0.5 0.0

88 d 100

Hg release

2.5 2.0 1.5 1.0 0.5 0.0

200

300

400

D6(LT)+Hg(II)

1d

7d

2.5 2.0 1.5 1.0 0.5 0.0

45 d

2.5 2.0 1.5 1.0 0.5 0.0

94 d 100

Hg release

200

300

400

1d

2.5 2.0 1.5 1.0 0.5 0.0

8d

2.5 2.0 1.5 1.0 0.5 0.0

45 d

2.5 2.0 1.5 1.0 0.5 0.0

108 d 100

200

300

400

500

1d

8d

2.0 1.5 1.0 0.5 0.0

44 d

2.0 1.5 1.0 0.5 0.0

88 d 100

2.5 2.0 1.5 1.0 0.5 0.0

200

300

400

D6(RT)+Hg(II)

500

1d

2.5 2.0 1.5 1.0 0.5 0.0

7d

2.5 2.0 1.5 1.0 0.5 0.0

45 d

2.5 2.0 1.5 1.0 0.5 0.0

94 d 100

500

B7(LT)+Hg(II)

D4(RT)+Hg(II)

2.0 1.5 1.0 0.5 0.0

500

2.5 2.0 1.5 1.0 0.5 0.0

2.5 2.0 1.5 1.0 0.5 0.0

2.0 1.5 1.0 0.5 0.0

209

2.0 1.5 1.0 0.5 0.0

200

300

400

B7(RT)+Hg(II)

500

1d

2.0 1.5 1.0 0.5 0.0

8d

2.0 1.5 1.0 0.5 0.0

45 d

2.0 1.5 1.0 0.5 0.0

108 d 100

0

200

300

400

500

0

Temperature ( C)

Temperature ( C)

Fig. 5. Thermograms of samples spiked with HgCl2 under the conditions LT and RT.

Table 4 Percentage of Hg(II) reduced and Hg volatilized at the end of the monitoring period. Amostras

Time (day)

Hg(II) (%) Reduced

D4 D6 B7

100 100 94

LT ¼low temperature. RT ¼ rom temperature.

Volatilized

LT

RT

LT

RT

87 1 57 1 77 1

97 1 67 1 77 1

4.2 7 0.2 3.5 7 0.2 2.3 7 0.1

7.3 7 0.4 5.4 7 0.3 3.3 7 0.2

the Hg(II)-spiked samples. The physico-chemical characteristics of the soil and the climatic conditions of the region of Descoberto favor the oxidation of Hg over reduction. This preference for oxidation is validated by the tendency of Hg identified in the contaminated area to be found predominantly in the form of Hg(II); however, specific differences depend on the soil composition. Additionally, microbial activity is notable in the redox activity of soils. The drying of the soil decreases, but does not eliminate, the microbiological activity. A systematic study would be necessary to evaluate the influence of microbial activity in these soils. 4. Conclusions In the study of the conversion of Hg species, the thermograms show that the oxidation process of the added Hg(0) in the sample

210

C.C. Windmöller et al. / Ecotoxicology and Environmental Safety 112 (2015) 201–211

usually occurs to a greater extent and at a faster rate than the reduction of Hg(II). The plots of the kinetics of the samples spiked with Hg(0) and Hg(II) showed that both the oxidation and reduction processes occur in two steps, and the first step in both studies was faster. The presence of Mn may play an important role in these processes, but more detailed studies are required to clarify this effect. The % of OM is an important factor in these processes, mainly because of the ability of OM to complex Hg(II). It was not possible the identification of which species are affecting this process in soils. Overall, the presence of OM favors Hg(0) oxidation and avoids the reduction of Hg(II). Studies of the extent of Hg(0) oxidation showed that oxidation is more favored at room temperatures than at low temperatures. The interaction of Hg with the matrix is suggested to occur in Hg(II)-complexes with inorganic and organic sulfur ligands and/or nonspecific adsorption onto oxides of Fe, Al and/or Mn The results obtained show large differences in the kinetics and thermodynamics of the processes of oxidation and reduction of Hg in soils. Additional information covering the conditions under which these processes occur is important to better understand the key process of Hg methylation and elucidate the geochemical cycle of this metal. In the case of Descoberto, which can serve as an example for other soils contaminated with Hg(0), the oxidation and volatilization of Hg(0) are important processes that occur. The oxidation tends to fix the metal in the fine grain size soil matrix and leach the metal into the aqueous environment, favoring the methylation of Hg. Studies investigating abiotic factors such as Mn mineralogical components and microbiological redox effects are of great value.

Acknowledgments The authors thank CNPq (150482/2012-4), FAPEMIG (CRA-APQ03861-09 and PPM-00664-11) and PRPq/UFMG (Pró-reitoria de Pesquisa) for their financial support.

References Almeida, M.G., Souza, C.M.M., 2007. Associação geoquímica de mercúrio com oxihidróxidos de alumínio, ferro e manganês em sedimentos superficiais da bacia inferior do Rio Paraíba do Sul, RJ, Brasil. Anais do XI Congresso Brasileiro de Geoquímica. Atibaia, Sociedade de Geoquímica, CD-ROM, 4 p. Andrade, R.P., Michel, R.F.M., Schaefer, C.E.G.R., Simas, F.N.B., Windmöller, C.C., 2012. Hg distribution and speciation in Antarctic soils of the Fildes and Ardley peninsulas, King George Island. Antarct. Sci. 24, 395–407. Augustin Jr., P.V., Viero, 2012. Environmental impact and geochemical behavior of soil contaminants from an industrial waste landfill in Southern Brazil. Environ. Earth Sci. 67, 1521–1530. Bahlmann, Ebinghaus, Ruck, 2006. Development and application of a laboratory flux measurement system (LFMS) for the investigation of the kinetics of mercury emissions from soils. J. Environ. Manag. 81, 114–125. Bavec, Biester, Gosar, 2014. Urban sediment contamination in a former Hg mining district, Idrija, Slovenia. Environ. Geochem. Health 36, 427–439. Bouffard, Amyot, 2009. Importance of elemental mercury in lake sediments. Chemosphere 74, 1098–1103. Biester, Gosar, Covelli, 2000. Mercury speciation in sediments affected by dumped mining residues in the drainage area of the Idrija mercury mine. Environ. Sci. Technol. 34, 3330–3336. Biester, Müller, Schöler, 2002. Binding and mobility of mercury in soils contaminated by emissions from chlor-alkali plants. Sci. Total Environ. 284, 191–203. Biester, Scholz, 1997. Determination of mercury binding forms in contaminated soils: mercury pyrolysis versus sequential extractions. Environ. Sci. Technol. 31, 233–239. CDTN/FEAM, 2005. Diagnóstico da Contaminação Ambiental em Descoberto, Minas Gerais, em decorrência do afloramento de mercúrio em dezembro de 2002. Relatório Técnico, Belo Horizonte, Minas Gerais, Brazil. CDTN/FEAM, 2006. Diagnóstico da Contaminação Ambiental em Descoberto, Minas Gerais, em decorrência do afloramento de mercúrio em dezembro de 2002. Relatório Técnico Final, Belo Horizonte, Minas Gerais, Brazil.

Choi, Holsen, 2009. Gaseous mercury fluxes from the forest floor of the Adirondacks. Environ. Pollut. 157, 592–600. Coufalik, Zverina, Komárek, 2014. Determination of mercury species using thermal desorption analysis in AAS. Chem. Pap. 68, 427–434 (4). Durão Jr., W.A., Palmiere, Trindade, Branco, Carvalho Filho, Fleming, Silva, Windmöller, 2009. Speciation, distribution, and transport of mercury in contaminated soils from Descoberto, Minas Gerais, Brazil. J. Environ. Monit. 11, 1056–1063. EMBRAPA/Empresa Brasileira de Pesquisa Agropecuária, 1999. Sistema Brasileiro de Classificação de Solos, 2th ed. Rio de Janeiro, Brasil. Fiorentino, Enzweiler, Angélica, 2011. Geochemistry of mercury along a soil profile compared to other elements and to the parental rock: evidence of external input. Water Air Soil Pollut. 221, 63–75. Gosar, Sajn, Biester, 2006. Binding of mercury in soils and attic dust in the Idrija mercury mine area (Slovenia). Sci. Total Environ. 369, 150–162. Gosar, Zibret, 2011. Mercury contents in the vertical profiles through alluvial sediments as a reflection of mining in Idrija (Slovenia). J. Geochem. Explor. 110, 81–91. Gosar, Tersic, 2012. Environmental geochemistry studies in the area of Idrija mercury mine, Slovenia. Environ. Geochem. Health 34, 27–41. Guedes, 2009. Protocolo de avaliação do mercúrio potencialmente tóxico em águas naturais (Tese, Doutorado em Química). Instituto de Química da Universidade de Campina, Campinas/São Paulo. Gustin, Biester, Kim, 2002. Investigation of the light-enhanced emission of mercury from naturally enriched substrates. Atmos. Environ. 36, 3241–3254. Li, Feng, Qiu, Shang, Wang, 2012. Mercury pollution in Wuchuan mercury mining area, Guizhou, Southwestern China: the impacts from large scale and artisanal mercury mining. Environ. Int. 42, 59–66. Licht, Mello, Silva, 2007. Prospecção geoquímica. Depósitos minerais metálicos, não-metálicos, óleos e gás. CPRM, Rio de Janeiro 788 pp.. Linhares, 2009. Avaliação da vulnerabilidade de solos representativos de Minas Gerais à contaminação por metais pesados baseada em estudos de adsorção e processos de extração (Tese, Doutorado em Química). Departamento de Química da Universidade Federal de Minas Gerais, Belo Horizonte/MG. Magarelli, Fostier, 2005. Quantificação de fluxos de mercúrio gasoso na interface solo/atmosfera utilizando câmara de fluxo dinâmico: aplicação na bacia do rio Negro. Quim. Nova 28, 968–974. Martinez-Trinidad, Silva, Islas, Reyes, Munguia, Valdez, Martinez, 2013. Total mercury in terrestrial systems (air–soil–plant–water) at the mining region of San Joaquin, Queretaro, Mexico. Geofis. Int. 52, 43–58. Melamed, Villas Bôas, 2002. Mecanismos de interação físico-química e mobilidade do mercúrio em solos, sedimentos e rejeitos de garimpo de ouro. Ser. Tecnol. Ambient., 25; (Rio de Janeiro: CETEM/MCT). Mishra, Tripathi, Bhalke, Shukla, Puranik, 2005. Determination of methylmercury and mercury(II) in a marine ecosystem using solid-phase microextraction gas chromatography–mass spectrometry. Anal. Chim. Acta 551, 192–198. Mulholland, Boaventura, Araújo, 2012. Geological and anthropogenic influences on sediment metal composition in the upper Paracatu River Basin, Brazil. Environ. Earth Sci. 67, 1307–1317. Nagy, Manceau, Gasper, Ryan, Aiken, 2011. Metallothionein-like multinuclear clusters of mercury(II) and sulphur in peat. Environ. Sci. Technol. 45, 7298–7306. Nik, Shahbazi, Naserian, 2013. Electrokinetics remediation of mercury contaminations of Bandar Imam petrochemical complex of Iran Asian. J. Chem. A 25, 3942–3946. Palmieri, Nalini, Leonel, Windmöller, Santos, Brito, 2006. Quantification and speciation of mercury in soils from the Tripuí Ecological Station, Minas Gerais, Brazil. Sci. Total Environ. 368, 69–78. Qiu, Feng, Meng, Sommar, Gu, 2012. Environmental geochemistry of an active Hg mine in Xunyang, Shaanxi Province, China. Appl. Geochem. 27, 2280–2288. Raposo, Windmöller, Durão Jr., W.A., 2003. Mercury speciation in fluorescent lamps by thermal release analysis. Waste Manag. 23, 879–886. Rezende, 2009. Avaliação da distribuição e mobilidade de elementos traços em sedimentos da bacia hidrográfica do rio São Francisco (Dissertação de Mestrado). Departamento de Química da Universidade Federal de Minas Gerais, Belo Horizonte/MG. Rezende, Moura, Durão Jr., W.A., Nascentes, Windmöller, Costa, 2011. Arsenic and mercury mobility in Brazilian sediments from the São Francisco River Basin. J. Braz. Chem. Soc. 22 (5), 910–918. Ruiz-Diez, Quinones, Fajardo, Lopez, Higueras, Fernandez-Pascual, 2012. Mercuryresistant rhizobial bacteria isolated from nodules of leguminous plants growing in high Hg-contaminated soils. Appl. Microbiol. Biotechnol. 96, 543–554. Scholtz, Van Heysta, Schroederb, 2003. Modelling of mercury emissions from background soils. Sci. Total Environ. 304, 185–207. Skyllberg, Drott, 2010. Competition between disordered iron sulphide and natural organic matter associated thiols for Mercury(II)—an EXAFS study. Environ. Sci. Technol. 44, 1254–1259. Teran-Mita, Faz, Salvador, Arocena, Acosta, 2013. High altitude artisanal small-scale gold mines are hot spots for mercury in soils and plants. Environ. Pollut. 173, 103–109. Tersic, Gosar, Biester, 2011a. Environmental impact of ancient small-scale mercury ore processing at Psenk on soil (Idrija area, Slovenia). Appl. Geochem. 26, 1867–1876 (11). Tersic, Gosar, Biester, 2011b. Distribution and speciation of mercury in soil in the area of an ancient mercury ore roasting site, Frbejzene trate (Idrija area, Slovenia). J. Geochem. Explor. 110, 136–145.

C.C. Windmöller et al. / Ecotoxicology and Environmental Safety 112 (2015) 201–211

Tersic, Gosar, 2012. Comparison of elemental contents in earthworm cast and soil from a mercury-contaminated site (Idrija area, Slovenia). Sci. Total Environ. 430, 28–33. Tersic, Biester, Gosar, 2014. Leaching of mercury from soils at extremely contaminated historical roasting sites (Idrija area, Slovenia). Geoderma 226–227, 213–222. Valle, Santana, Augusti, Egreja Filho, Windmöller, 2005. Speciation and quantification of mercury in Oxisol, Ultisol, and Spodosol from Amazon (Manaus, Brazil). Chemosphere 58, 779–792. Valle, Santana, Windmöller, 2006. Mercury conversion processes in Amazon soils evaluated by thermodesorption analysis. Chemosphere 65, 1966–1975. Varejão, Bellato, Fontes, 2009. Mercury fractionation in stream sediments from the Quadrilátero Ferrífero gold mining region, Minas Gerais State, Brazil. Environ. Monit. Assess. 157, 125–135.

211

Wallschläger, Desai, Spengler, Windmöler, Wilken, 1998. How humic substances dominate mercury geochemistry in contaminated floodplain soils and sediments. J. Environ. Qual. 27, 1044–1054. Wang, Feng, Anderson, Xing, Shang, 2012. Remediation of mercury contaminated sites —a review. J. Hazard. Mater. 221–222, 1–18. Windmöller, Santos, Atahyde, Palmieri, 2007. Análise Qualitativa e Quantitativa de Mercúrio em Sedimentos de Áreas de Garimpo de Ouro do Quadrilátero Ferrífero (MG). Quim. Nova 30, 1088–1094. Windmöller, Wilken, Jardim, 1996. Mercury speciation in contaminated soils by thermal release analysis. Water Air Soil Pollut. 89, 399–416.