Influence of Reducing Conditions on Metallic Elements Released from

Sep 20, 2008 - Corresponding author phone: (+33) 555-457-485; fax: (+33) ... were obtained (A, 220−345 mV; B, 280−365 mV; C, 260−360 mV; and D, ...
0 downloads 0 Views 778KB Size
Environ. Sci. Technol. 2008, 42, 7615–7621

Influence of Reducing Conditions on Metallic Elements Released from Various Contaminated Soil Samples ´ NILLA, PRISCILLA PAREUIL, SONIA PE NURSEN OZKAN, FRANC ¸ OIS BORDAS,* AND JEAN-CLAUDE BOLLINGER Université de Limoges, Groupement de Recherche Eau-Sol-Environnement, Faculté des Sciences & Techniques, 123 avenue Albert Thomas, 87 060 Limoges, France

Received April 5, 2008. Revised manuscript received July 7, 2008. Accepted July 24, 2008.

The redox conditions of soil may have significant consequences for the mobility of metallic elements (ME), but unlike pH, very few studies have investigated this parameter. A procedure was established to study the solubilization of ME from soil samples in various reducing conditions using a batch method and sodium ascorbate solutions. The change in redox potential from +410 to +10 mV was studied from four contaminated soil samples (designated A-D) of different origins and compositions. The results showed that ME mobilization greatly increased with decreasing redox potential within a limited and very sensitive range. Depending on the soil sample studied, various sensitive ranges of potentials were obtained (A, 220-345 mV; B, 280-365 mV; C, 260-360 mV; and D, 240-380 mV), and the induced percentages of ME mobilization varied (i.e., maximal values for Zn: A, 45%; B, 59%; C, 53%; and D, 58%). The results could be explained by the combined effect of potential and pH decrease on ME-carrying phases; in particular, Fe and Mn (oxy)hydroxides.

Introduction Redox potential in soils can be modified by natural phenomena (flooding, etc.) or anthropic activities (irrigations, organic amendments, etc.). Like pH, this parameter is a main factor governing metallic element (ME) mobilization in soils through the dissolution of the ME-carrying phases (1, 2). Several experimental studies have focused on the influence of reducing conditions on ME mobility, either through batch experiments (3, 4), within soil columns (5, 6), or using undisturbed blocks of topsoil (2). The main methods used to create reducing conditions in a soil sample consisted of aerobic/anaerobic incubation (by fluxing either O2 or N2), water-flooding incubation, or addition of reducing agents. The first two approaches are mainly based on microorganism activity, which is hard to control. Moreover, they generally need a long contact time and do not allow many different reducing conditions to be obtained. The use of reducing reactants allows fast and easy experiments with a large range of reducing conditions. Davranche et al. (4, 7) studied the effect of reducing conditions on ME mobility in soils or synthetic pure solid phases using hydroxylamine hydrochloride or sodium ascorbate. More recently, Chatain et al. (8) evaluated the ability of sodium ascorbate or sodium * Corresponding author phone: (+33) 555-457-485; fax: (+33) 555457-203; e-mail: [email protected] 10.1021/es800953d CCC: $40.75

Published on Web 09/20/2008

 2008 American Chemical Society

borohydride to produce different oxido-reducing conditions in a mining soil. They emphasized the importance of the nature of the reducing agent and the pH range on the results. Larsen et al. (9), through parallel dissolution studies in HCl and ascorbic acid at pH 3, highlighted the importance of reductive dissolution in iron solubilization from a sandy aquifer. Depending on the technique employed, contradictory data have been reported, such as a simultaneous increase in Cd, Pb, Zn, and Fe when the redox potential E decreased using gas bubbling in a contaminated soil suspension (3) or a simultaneous increase in Pb, Cd, Fe, and Mn mobility under reducing conditions imposed by a chemical agent (4, 7). However, in any case, the solubilization of Fe and Mn (oxy)hydroxides seems the major mechanism controlling the release of ME under reducing conditions (3, 4, 10, 11) and is favored by acid conditions (12). In this context, our purpose was to propose a chemical experimental method creating a wide range of redox conditions to study the effect of reducing conditions on the release of ME. This methodology was applied to four contaminated soil samples presenting various characteristics. To understand the mechanisms involved, the dissolution of the main carrier phases of metals (iron and manganese (oxy)hydroxides) was monitored through the solubilization of Fe and Mn.

Materials and Methods All chemicals were of analytical grade. The different solutions were prepared in high-purity deionized water (HPW) (Milli-Q system: resistivity 18.2 MΩ · cm, TOC e 10 µg · L-1). All glassware and containers were previously decontaminated in 10% (v/v) nitric acid for at least 24 h. Each experiment and each analysis were carried out in triplicate; the results are given as mean value ( standard deviation. Origin and Chemical Characterization of Soil Samples. Four samples of French contaminated soils, designated A-D, from different geographical origins were studied (Table S1 of the Supporting Information). They were pretreated according to the ISO 11464 standard (air-drying and sieving at 2 mm after deagglomeration). The pHH2O and pHKCl (1 mol · L-1) were determined according to ISO 10390. The cation exchange capacity (CEC) was determined by the cobaltihexammine method according to the French standard NF 31-130. Organic carbon (OC) content was determined by sulfochromic oxidation according to the ISO 14135 standard. Inorganic carbon (IC) content was evaluated by difference between the total carbon obtained by C/S analyzer (ELTRA CS-200) and OC content. Total ME Content in Soil Samples. ME content was determined by microwave-assisted digestion (Anton Paar Multiwave 3000). Each air-dried soil sample (0.5 g) was digested with 3 mL of 35% HCl and 9 mL of 69% HNO3 in PTFE vessels. A power ramp was carried out for 5 min to reach 1400 W and held for 25 min. Digests were filtered through 0.2 µm Sartorius cellulose acetate filters into 50 mL volumetric flasks completed with HPW. The correctness of this approach was monitored by comparison with HF digestion (unpublished data). Mineralogical Fractionation of Iron. Percentages of amorphous, organic, and crystalline iron were distinguished according to Schwertmann (13) and McKeague (14). Amorphous and organic iron contents were determined by ammonium oxalate extraction: 0.05 g of air-dried soil sample and 20 mL of ammonium oxalate (0.2 mol · L-1 acidified at VOL. 42, NO. 20, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7615

TABLE 1. Chemical Characteristics, Total Metallic Elements’ Contents, Iron Fractionation and Cr(VI) Content in the Soil Samples soil samples A

Chemical Characteristics 6.4 ( 0.2/5.2 ( 0.2 7.9 ( 0.2/7.4 ( 0.2 +389 ( 1 +360 ( 3 11.8 ( 0.4 18.4 ( 0.2 23.5 ( 0.9 29.2 ( 1.0 6.7 ( 1.0 54.9 ( 1.7

pHH2O and pHKCl Eha (H2O) (mV) CEC (meq · 100 g-1) organic carbon (g · kg-1) inorganic carbon (g · kg-1)

Major Metallic Elements (mg · kg-1) 51 903 ( 5640 29 458 ( 2683 2664 ( 197 206 218 ( 7591 total 39 026 ( 2272 29 625 ( 1373 crystalline (%) 66.3 ( 5.2 62.0 ( 3.7 organic (%) 10.2 ( 1.0 4.9 ( 0.3 amorphous (%) 23.5 ( 2.1 33.2 ( 2.7 982 ( 54 556 ( 7

Al Ca Fe

Mn Cu Cr

total Cr(VI) (%)

Ni Pb Zn a

Minor Metallic Elements (mg · kg-1) 32.4 ( 3.2 67.4 ( 0.2 203.1 ( 29.6 37.7 ( 1.9 13.8 ( 1.5 ND 348.8 ( 62.6 18.2 ( 0.6 26.4 ( 2.5 36.3 ( 1.2 167.6 ( 8.9 190.3 ( 12.4

C

D

6.7 ( 0.2/5.9 ( 0.2 +361 ( 2 21.4 ( 0.3 48.3 ( 1.9 6.4 ( 2.6

5.7 ( 0.2/4.4 ( 0.2 +502 ( 5 10 ( 1 57 ( 9 8.2 ( 3.2

29 262 ( 446 8972 ( 231 36 145 ( 895 45.8 ( 1.4 15.5 ( 1.6 38.8 ( 1.0 377 ( 14

28 726 ( 733 3100 ( 300 18 620 ( 650 82.3 ( 0.7 8.7 ( 1.8 9.1 ( 1.1 553 ( 61

37.3 ( 1.1 52.6 ( 1.4 ND 18.2 ( 0.6 36.3 ( 1.2 225.6 ( 8.1

223 ( 13 14.8 ( 0.6 20.3 ( 6.8 4.3 ( 0.3 89 ( 3 81.3 ( 0.6

Eh expressed in mV was measured after system stabilization ND: Not Detected.

pH 3 with HNO3) were mixed in the dark for 2 h at 20 °C. Organic iron content was determined by sodium pyrophosphate extraction: 0.04 g of another air-dried soil sample and 20 mL of 0.1 mol · L-1 sodium pyrophosphate were mixed for 16 h at 20 °C. Crystalline iron content was determined by difference from total iron. After each extraction, the solutions were filtered through 0.2 µm Sartorius cellulose acetate filters into 50 mL volumetric flasks completed with HPW. Fractionation of ME by Sequential Extractions. The accelerated BCR sequential extraction procedure proposed by the European Bureau Communautaire de Référence (BCR), and developed by Pérez-Cid et al. (15) using focused ultrasound (Bandelin HD 70) was carried out. Extracts were separated from solid residues by centrifugation at 1500g for 15 min. Since the four fractions obtained were operationally defined, they were specified here as F1 (exchangeable and soluble in acidic medium), F2 (reducible), F3 (oxidizable), and F4 (residual). The complete procedure for each set of analyses using the same reagents was carried out in the absence of a sample. No significant contamination due to glassware or reagents was observed. Chromium Speciation. Cr(VI) content was determined according to the MA. 200-CrHex 1.0 Canadian method (16). One gram of air-dried soil sample and 40 mL of basic extracting solution (0.5 mol · L-1 NaOH and 0.28 mol · L-1 Na2CO3) were mixed for 1 h at 90-95 °C. After Cr(VI) extraction, the pH was adjusted to 7-8 with HNO3, and the Cr(VI) content was measured by the colorimetric reaction using diphenylcarbazide (DPC). The absorbance of the Cr(VI)-DPC complex was measured at 540 nm (Varian Cary 50 Probe Spectrometer). The Cr(III) content was determined by difference from total Cr. ME Mobilization According to Redox Potential. Experiments were performed with a solid/liquid ratio of 10 g · L-1 in HDPE amber bottles (1 L). The various reducing agents and their concentration ranges are listed in Table S2 of the Supporting Information. The initial pH of the reducing solutions was adjusted with HNO3 or NaOH to the soil pHH2O values but was free to change during the experiments; this was taken into account in the discussion. A reference batch experiment was carried out with HPW under the same experimental conditions. The soil samples were maintained 7616

B

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 20, 2008

in suspension by continuous stirring on an orbital shaking table (IKA-Labortechnik K550 Digital) at 250 rpm and room temperature (22 ( 2 °C). Previously, the contact time corresponding to the stabilization of pH, E, and mobilization of ME was determined from a kinetic study. For these tests, three batch experiments conducted under the same experimental conditions were carried out for each soil sample with HPW and 0.02 and 1 mol · L-1 of each reducing agent. Aliquots of the suspensions were taken at various periods; pH and E were monitored every 24 or 48 h (Crison GLP22 pH meter: combined electrode Crison 52 21 for pH and indicating platinum electrode, Ag/ AgCl/KCl 3 mol · L-1 reference electrode Crison 52 62 for potential measurement). The measured E values were transformed into Eh values (vs she) (the redox potential relative to the standard hydrogen electrode) by addition of the potential for the reference electrode to better compare our results with literature. Analytical Methods. ME concentrations were determined using either flame atomic absorption spectrometry (Varian SpectrAA 220 equipped with a deuterium background correction) or graphite furnace atomic absorption spectrometry (Varian SpectrAA 880 Z equipped with a Zeeman background correction). Before analysis, each soil suspension was centrifuged at 25000g for 10 min. The interference due to sodium ascorbate during the AAS analysis was limited by a microwave-assisted digestion adapted from the EPA 3015 method: 5 mL of sample was digested with 3 mL of 30% H2O2 and 3 mL of 69% HNO3 at 165 °C for 10 min, followed by a ramp of 10 min to reach 170 °C (Anton Paar Multiwave 3000); the digests were filtered through 0.2 µm Sartorius cellulose acetate filters into 20 mL volumetric flasks completed with HPW. For the higher concentrations of sodium sulfite and sodium thiosulfate, internal standards were used because of matrix interferences.

Results and Discussion Chemical Characteristics of Soil Samples. Chemical characteristics and total ME contents are presented in Table 1 with iron fractionation and Cr(VI) content. (i) For sample A, the OC content was low, and the pHH2O was slightly acid. Its Al content was twice that of the other

FIGURE 1. Metallic elements’ distribution in soil samples. Black, solid area, F1: exchangeable fraction. Unmarked area, F2: reducible fraction. Cross-hatched area, F3: oxidizable fraction. Gray, solid area, F4: residual fraction.

samples; Fe and Mn content were also higher. Fe was mainly in the crystalline form, which is less sensitive to dissolution than the amorphous and organic ones (18). Sample A was contaminated mainly by Ni, Zn, and Cr, which was essentially in the Cr(III) form. (ii) Sample B was an alkaline soil, Ca content was up to 100 times higher than in the others. Moreover, the IC content was about 10 times higher than in the other samples. Its basic pH is due to the presence of calcium carbonate. The presence of calcite was confirmed by XRD analysis (Figure S1). The percentage of organic iron was very low. Sample B was contaminated mainly by Zn and Cu. (iii) Sample C was a neutral soil. Its OC content and CEC values were the greatest in the present series; nevertheless, these values could be generally considered as low. The mineralogical fractionation of iron differs from the other soil samples: the crystalline fraction was the smallest, whereas the amorphous one was the highest. This was confirmed by the XRD analysis (Figure S1). Therefore, the dissolved iron proportion would be probably greater in this sample under reducing conditions. Sample C was contaminated mainly by Zn and Cr. (iv) Sample D was slightly acidic, like sample A. The Fe content was the lowest, and its fractionation underlined that it was mainly in crystalline form. Amorphous and organic forms represented only 20%. Sample D was contaminated mainly by Cu, Pb and Zn. ME Fractionation in Soil Samples. The percentages of ME in the various fractions were calculated from the total content obtained by digestion (Figure 1). Consequently, the sum of the four fractions usually varied between 80 and 120% (18). For all of the soil samples studied, Fe was mainly found in F4, whereas Mn was distributed within the four fractions, with predominance in F2 and F4. Concerning the main metallic contaminants,

(i) For sample A, Ni and Cr were found mainly in F4. Zn fractionation was more heterogeneous; it was found principally in F4 but also in the reductive fraction. (ii) For sample B, Cu was found mainly in F4, and also in F3. Zn was found in the four fractions. (iii) For sample C, Cr was found mainly in F4. Zn was distributed within the four fractions, in a way similar to sample B. (iv) For sample D, 5% of Cu, 10% of Zn, and 25% of Pb were found in F2. These three metallic contaminants were also present in the other three fractions. ME fractionation was dependent on both the ME studied and the soil sample concerned. Iron was found mainly in F4, but this was not coherent with its mineralogical fractionation. For example, the residual fraction contained all of the Fe, including the crystalline Fe, which represented only 45% (hematite may be reported for all the samples except for C; see Figure S1). This underlined the lack of selectivity of hydroxylamine hydrochloride, which did not allow the iron crystalline oxides to be extracted in the reductive fraction (17, 19). The four samples differed in their composition (i.e., Fe and Mn contents, IC content, etc.), the nature of contaminants, and their distribution in the soil sample, so a difference in behavior under various redox conditions could be expected. Imposition of a Redox Potential Gradient in the Soil Sample by Adding a Reducing Agent. Although the thermodynamic significance of redox potential measurements has been questioned, their evolution is a useful indicator of the changes in the redox status of any given soil sample. To develop an experimental method that allowed the imposition of a wide range of reducing conditions in soil suspension, eight well-known reducing agents (4) were tested at different concentrations with soil sample D. The results of batch experiments showed different effects of the reducing VOL. 42, NO. 20, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7617

FIGURE 2. Mobilization percentages of Fe, Mn, Ni and Zn versus time for the soil sample A (black square, HPW; gray circle, 0.02 mol · L-1 ascorbate; gray triangle, 1 mol · L-1 ascorbate). agents on pH and Eh for the soil sample studied (Table S2). A broader Eh range was obtained with sodium ascorbate, which led to various redox potentials ranges: from “oxidized” (+430 mV) to “reduced” (+26 mV), according to the scale of Trolard et al. (21). In the presence of sodium ascorbate, the pH of the soil sample D suspension was only slightly modified: variations of less than (0.5 pH unit were observed, depending on the concentration. Many studies have emphasized that both pH and E influence ME release from a soil (3, 4, 7-9, 22, 23). Obtaining a relatively stable pH during the experiments is, thus, crucial to assess the effect on ME solubilization of reducing conditions only. According to these results, only sodium ascorbate was later used to impose reducing conditions with the different soil samples. ME Mobilization According to the Contact Time. The mobilization of ME as a function of time showed two kinds of behavior (see Figure 2 for sample A). For Fe, Mn, and Ni, the percentages of solubilized ME increased and were stabilized after 250 h, whereas for Zn, the solubilized percentage reached a maximum before decreasing. Fe and Mn mobilization depended on the ascorbate concentration, although Zn mobilization seemed to be less dependent on reagent concentration. Indeed, Zn was partly released in HPW (up to 5.8%), which is consistent with its fractionation (5.6% of Zn in F1). ME mobilization therefore depends on ME fractionation in the soil sample. Similar trends were observed for all soil samples. The Mn mobilization rate was faster than for Fe. With 1 mol · L-1 sodium ascorbate, Mn was mobilized to about 60% after 30 h, whereas Fe mobilization reached the same percentage after 175 h (Figure 3), in general accordance with the literature data (19, 20). The effect of sodium ascorbate on pH and Eh depended on both the sample buffer effect and the reagent concentration. During kinetics experiments, the pHs of samples B and D showed smaller variations than those of samples A and C. For each sample, the differences between pH in HPW and in 1 mol · L-1 sodium ascorbate for the maximum contact 7618

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 20, 2008

time were the following for A and D, ∆pH ≈ 1; C, ∆pH ≈ 1.5; and B, ∆pH ≈ 2. In the presence of 1 mol · L-1 sodium ascorbate, the pH seemed to be imposed by the reagent (pH ≈ 6). Significant variations were observed with 0.02 mol · L-1 sodium ascorbate, probably due to its interaction with soil sample components and the release of protons during ascorbate oxidation. Moreover, protons bound to solid phases could be exchanged with Na+ brought by the ascorbate solution, as with pHKCl measurement. The evolution of Eh according to time was similar whatever the soil sample. In HPW and with 1 mol · L-1 sodium ascorbate, Eh (stabilized before 50 h) was imposed either by the soil sample in HPW or by the excess of ascorbate (1 mol · L-1). Stabilization of Eh was the result of the balance between O2 from air and the soil sample interaction with the ascorbate solution. With regard to 0.02 mol · L-1 sodium ascorbate, the evolution of Eh was similar to that of pH, probably due to ascorbate consumption by soil sample components. According to the results obtained from the monitoring of pH, Eh, and ME mobilization versus time, the contact time chosen was about 350 h for each soil sample. This contact time is consistent with the slow kinetics of this reagent (4, 24). ME Mobilization According to Redox Potential. The percentages of ME mobilized are represented according to Eh in order to observe the evolution for each soil sample at the maximum contact time studied (Figure 3). For all ME considered and soil samples studied, ME mobilization occurred in three steps: (i) in an oxidizing medium, ME mobilization was slightly influenced by Eh variations; (ii) in a moderately reducing medium, a decrease in Eh induced a great increase in the ME percentage mobilized; and (iii) in a reducing medium, the percentage mobilized was stabilized. Each soil sample has its own buffer capacity respective to variations in redox conditions (1) and reacts differently to ascorbate addition. Hence, ME were mobilized at various sensitive ranges of Eh (A, 220 to 355 mV; B, 280 to 365 mV; C, 260 to 360 mV; and D, 240 to 380 mV; see Figure 3). Sample D, which was the most oxidizing soil sample and presented

FIGURE 3. Mobilization of major and minor ME for various soil samples (contact time, 350 h). The sensitive ranges (crosshatching) were evaluated from the derivative curves ∆(ME mobilized)/∆Eh ) f (Eh). the lowest pHH2O value (Table 1), was the most sensitive to low Eh variations. The percentages of mobilization depended on both the ME studied and soil sample concerned. For example, 57% of Fe was mobilized in sample D, whereas up to 98% was mobilized in sample C. These results could be explained by Fe fractionation in soil samples: crystalline iron represented 83% in D but only 46% in sample C, in which Fe is mainly in the amorphous form. Crystalline iron is the more difficult to solubilize when Eh and pH change. Whatever the soil sample considered, between 70 and 80% of Mn was mobilized when Eh reached 220 mV. Mn (oxy)hydroxides could be dissolved under weaker reducing conditions than Fe (oxy)hydroxides (19). Mn mobilization seems to be less dependent on the soil sample nature than Fe. As concerns the main metallic contaminants, the percentages mobilized were smaller as compared to Fe and Mn. The mobilized amounts were consistent with ME fractionation (Figure 1): Zn for samples B and C, Cu, Zn, and Pb for sample D were the most sensitive to Eh variations; Cu, Cr, Ni, and Zn in samples A, B, and C were less influenced by Eh variations. But from a quantitative point of view, the

maximum amount of ME mobilized when the redox potential decreases was greater than the ME content found in F2. These results could be explained by the combined effect of Eh and pH on ME mobilization and Fe and Mn (oxy)hydroxides dissolution. The acidification, led by the increase in sodium ascorbate concentration, favored ME mobilization but also Fe and Mn (oxy)hydroxides dissolution (12), which played a major role in the ME mobilization under reducing conditions. However, the apparent release of ME from the oxidizable fraction (F3) by a reducing agent can also be explained as a displacement of Cu, Ni, Zn, or Cr from organic sites by the high concentrations of Fe and Mn generated during reduction. Sometimes, the percentage of ME mobilized with ascorbate was greater than the sum of F1 + F2 + F3. Part of the ME found in F4 could have been dissolved. This result could be explained by the dissolution of Fe and Mn (oxy)hydroxides, not solubilized during the determination of F2 with hydroxylamine hydrochloride as reported by Neaman et al. (19), but which could occur under higher reducing conditions in the presence of ascorbate, releasing the associated ME. Study of Metallic Elements Solubilization Mechanisms. To check that ME solubilization resulted from decreasing redox potential rather than complexation with ascorbate under our experimental conditions, speciation calculations were performed using the MINEQL+ program (25) (detailed results not shown), and additional batch experiments with a nonreducing but complexing agent (sodium acetate) were carried out with soil sample D. Considering the calculated speciation of ME in solution under our pH conditions, complexes with ascorbate represented a significant part for all total ME, from 10-20% for Fe and Zn to 70-80% for Cu and Pb. For every metallic element (M), the MHAsc form was predominant from pH 4 to 6. This result shows that ME released in solution were prevented in part from readsorption on residual solid phases by complexation in solution with ascorbate. Since calculations raised possible interactions between ascorbate and ME, it was crucial to ascertain that the main factor affecting ME mobility was the redox status of the soil sample suspension. The effect of complexation on ME mobilization in the presence of ascorbate was checked by comparison with a nonreducing ligand, the sodium acetate, whose complexation constants are at least equal to or higher than those with ascorbate and which is not effective at solubilizing metal oxides. At the same Eh (ca. +430 mV) and pH (5.7) values, ascorbate induces a slight increase in ME solubilization relative to acetate, whose results were close to those of the reference experiment in HPW (Figure S3). However, at the same concentration (0.07 mol · L-1) and pH (5.7), the ME solubility induced by ascorbate (strong reducing medium, Eh ) +97 mV) was up to 10 orders of magnitude greater (Figure S3) than with acetate (strong oxidizing medium, Eh ) +420 mV). This result suggests that the high concentration of soluble ME measured with ascorbate resulted from the effect of decreasing redox potential rather than a complexation mechanism. The dissolution of natural or synthetic iron and manganese oxides induced by the presence of a reducing ligand has been widely studied (4, 9, 11, 26). However, scarce research has been published on the use of ascorbate to study the effect of reducing conditions on ME mobility in soils. Among these studies, only Larsen et al. (9) have tried to differentiate the proton-assisted dissolution from the reductive dissolution, but none has discussed the possibility of a solubilization of ME promoted by complexation with ascorbate. Eh greatly decreased as sodium ascorbate concentration increased (i.e., see Figure S2 for soil sample D). The result is similar to a titration curve of the soil redox capacity. For each of the four VOL. 42, NO. 20, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7619

soils studied, two buffering steps toward redox potential changes appeared before the effect of reducing conditions on the soil was at maximum. The shape of the curve in Figure S2 may indicate that different mineral forms of Mn and Fe are solubilized when Eh varies and are involved in the buffering capacity of the soil (22, 23). As expressed by Barcelona and Holm (1), each soil has its own redox capacity, which represents a considerable resistance to changes in redox conditions. Each soil will, thus, react differently after addition of the same concentration of reducing agent, as already observed from two contaminated soil samples (industrial and agricultural) (4). The simultaneous release of various ME as observed in our study, seemed to indicate that their solubilization was related to the reductive dissolution of the carrier phases formed by Fe and Mn oxides. In another study of the release of trace ME from a contaminated soil by suspension experiments, the solubilized concentrations of Pb, Cd, and Zn were positively correlated with Fe concentrations (3); the authors concluded that trace ME were released from iron oxide surfaces as the solid dissolves. On the contrary, from column experiments, an inverse relationship between Pb and Fe concentrations in solution was observed (6) and suggested the importance of the reductive dissolution of Fe oxides in controlling ME solubility and mobility in soils. Quantin et al. (26) compared the chemical extraction and the bacterial dissolution of natural metal-rich soils; they concluded that during reducing conditions induced by temporary water-logging, the sources of released Cr and Ni are iron but mainly manganese oxides. The Concept of a Redox Mobilization Edge. Very few researchers have tried to represent the solubilization of ME according to different E values. Gotoh and Patrick studied the distribution of different forms of manganese (22) and iron (23) in a waterlogged soil under various redox potential and pH conditions. They showed the existence of a critical redox potential range for their reduction and their induced dissolution: that is (in mV), +200 e Eh e +400 for Mn at 6 e pH e 8 and +100 e Eh e +300 at 6 e pH e 7 or Eh ≈ -100 mV at pH 8. Atta et al. (27) revealed, from rice soil, that redox potential and pH affect water-soluble and exchangeable iron and manganese, and threshold values were about -150 mV at pH 8 and 7 for Fe and +150 mV for Mn and +200 mV at pH 6 for Fe. Our results are consistent with those from the literature, whose values were obtained from soil samples of different origins and composition subject to reduction through “natural causes”, such as aerobic/anaerobic incubation or water-flooding (see also refs 20 and 28). By analogy with the effect of pH on ME solubilization (29), we suggest the introduction of the concept of a “redox mobilizationedge”, describing the influence of the decrease in the redox potential on ME solubilization from a given soil. Whatever the soil sample studied or the methodology used, the ME amount released was slightly influenced by Eh variations. Redox potential, as well as pH, is a key factor to be taken into consideration for risk assessment studies in the case of metal-contaminated sites. In our case, the maximum amounts of the ME solubilized when the potential decreases were greater than the ME content found in fraction F2. This result shows that the sequential extraction protocols such as BCR (15) are not suitable for the evaluation of their potential mobility under these conditions. The need for a standardized method for assessing the leachability and the mobility of metallic contaminants over reducing conditions is pointed out herein. Then an easy chemical approach, using a series of sodium ascorbate solutions, could be used to simulate a gradient of reducing conditions and to test potential mobility of ME; for example, during land use changes. 7620

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 20, 2008

Acknowledgments This work was financially supported by the Contrat de plan Etat-Région Limousin and by the Conseil Régional du Limousin. S.P. and P.P. thank the ANRT for funding. The authors thank Dr. E. Joussein (GRESE, Universite´ de Limoges) for characterization of soils; the three anonymous reviewers; and the Associate Editor (Prof. L. Sigg) whose constructive critical comments helped to improve the content of this paper.

Supporting Information Available Origin and characterization of the studied soil samples (Table S1 and Figure S1), ranges of pH and Eh obtained with different concentrations of various reducing agents (Table S2) and variation of Eh according to ascorbate concentration in solution for soil sample D (Figure S2), comparison between metal solubilization following the addition of sodium ascorbate or sodium acetate solutions (Figure S3). This material is available free of charge via the Internet at http:// pubs.acs.org

Literature Cited (1) Barcelona, M.; Holm, T. R. Oxidation-reduction capacities of aquifer solids. Environ. Sci. Technol. 1991, 25, 1565–1572. (2) Charlatchka, R.; Cambier, P. Influence of reductive conditions on solubility of trace metals in contaminated soils. Water, Air, Soil Pollut. 2000, 118, 143–167. (3) Chuan, M. C.; Shu, G. Y.; Liu, J. C. Solubility of heavy metals in a contaminated soil: effects of redox potential and pH. Water, Air, Soil Pollut. 1996, 90, 543–556. (4) Davranche, M.; Bollinger, J. C. Heavy-metals desorption from synthetised and natural iron and manganese oxyhydroxides: effect of reductive conditions. J. Colloid Interface Sci. 2000, 227, 531–539. (5) Schlieker, M.; Schu ¨ring, J.; Hencke, J.; Schulz, H. D. The influence of redox processes on trace element mobility in a sandy aquiferan experimental approach. J. Geochem. Explor. 2001, 73, 167– 179. (6) Ma, L. Q.; Dong, Y. Effects of incubation on solubility and mobility of trace metals in two contaminated soils. Environ. Pollut. 2004, 130, 301–307. (7) Davranche, M.; Bollinger, J. C.; Bril, H. Effect of reductive conditions on metal mobility from wasteland solids: an example from the Mortagne-du-Nord site (France). Appl. Geochem. 2003, 18, 383–394. (8) Chatain, V.; Sanchez, F.; Bayard, R.; Moszkowicz, P.; Gourdon, P. Effect of experimentally induced reducing conditions on the mobility of arsenic from a mining soil. J. Hazard. Mater. 2005, B122, 119–128. (9) Larsen, O.; Postma, D.; Jakobsen, R. The reactivity of iron oxides towards reductive dissolution with ascorbic acid in a shallow sandy aquifer (Rømø, Denmark). Geochim. Cosmochim. Acta 2006, 70, 4827–4835. (10) Wang, G.; Staunton, S. Evolution of water-extractable copper in soil with time as a function of organic matter amendments and aeration. Eur. J. Soil Sci. 2006, 57, 372–380. (11) Suter, D.; Banwart, S.; Stumm, W. Dissolution of hydrous iron(III) oxides by reductive mechanisms. Langmuir 1991, 7, 809–813. (12) Deng, Y. Effect of pH on the reductive dissolution rates of iron (III) hydroxide by ascorbate. Langmuir 1997, 13, 1835–1839. (13) Schwertmann, U. Differenzierung der Eisenoxide des Bodens durch Extraktion mit Ammoniumoxalat-Lo¨sung. Z. Pflanzenerna¨hr. Du ¨ ng. Bodenk. 1964, 105, 194–202. (14) McKeague, J. A. An evaluation of 0.1 M pyrophosphate and pyrophosphate dithionite in comparison to oxalate as extractants of the accumulation products in podzols and some other soils. Can. J. Soil Sci. 1967, 47, 95–99. (15) Perez-Cid, B.; Lavilla, I.; Bendicho, C. Speeding up of a threestage sequential extraction method for metal partitioning using focused ultrasound. Anal. Chim. Acta 1998, 360, 35–41. (16) Centre d’Expertise en Analyse Environnementale du Que´bec. De´termination du chrome hexavalent: me´thode colorime´trique MA 200-CrHex 1.0, re´v. 4; Ministe`re du De´veloppement Durable, de l’Environnement et des Parcs du Que´bec, 2006; p 12. (17) La Force, M. J.; Fendorf, S. Solid-phase iron characterization during common selective sequential extractions. Soil Sci. Soc. Amer. J. 2000, 64, 1608–1615.

(18) Davidson, C. M.; Duncan, A. L.; Littlejohn, D.; Ure, A. M.; Garden, L. M. A critical evaluation of the three-stage BCR sequential extraction procedure to assess the potential mobility and toxicity of heavy metals in industrially-contaminated land. Anal. Chim. Acta 1998, 363, 45–55. (19) Neaman, A.; Waller, B.; Mouélé, F.; Trolard, F.; Bourrié, G. Improved methods for selective dissolution of manganese oxides from soils and rocks. Eur. J. Soil Sci. 2004, 55, 47–54. (20) Cornu, J. Y.; Denaix, L.; Schneider, A.; Pellerin, S. Temporal evolution of redox processes and free Cd dynamics in a metalcontaminated soil after rewetting. Chemosphere 2007, 70, 306– 314. (21) Trolard, F.; Bourrie´, G.; Soulier, A.; Maitre, V.; Genin, J. M. R.; Abdelmoula, M. Dynamique de l’oxydo-re´duction dans les zones humides. in Agriculture Intensive et Qualite´ des Eaux; Cheverry C., Ed.; INRA Editions: Paris, 1998; pp 185-208. (22) Gotoh, S.; Patrick, W. H. Transformation of manganese in a waterlogged soil as affected by redox potential and pH. Soil Sci. Soc. Am. Proc. 1972, 36, 738–742. (23) Gotoh, S.; Patrick, W. H. Transformation of iron in a waterlogged soil as influenced by redox potential and pH. Soil Sci. Soc. Am. Proc. 1974, 38, 66–71. (24) Liang, L.; Hofmann, A.; Gu, B. Ligand-induced dissolution and

(25)

(26)

(27)

(28)

(29)

release of ferrihydrite colloids. Geochim. Cosmochim. Acta 2000, 64, 2027–2037. Schecher, W. D.; McAvoy, D. C. MINEQL+, A Chemical Equilibrium Modelling System, Version 4.5 for Windows, User’s Manual, 1st ed; Environmental Research Software Publisher: Hallowell, 1998. Quantin, C.; Becquer, T.; Berthelin, J. Mn-oxide: a major source of easily mobilisable Co and Ni under reducing conditions in New Caledonia Ferralsols. C.R. Geosci. 2002, 334, 273–278. Atta, S. Kh.; Mohammed, S. A.; Van Cleemput, O.; Zayed, A. Transformations of iron and manganese under controlled Eh, Eh-pH conditions and addition of organic matter. Soil Technol. 1996, 9, 223–237. Miao, S.; DeLaune, R. D.; Jugsujinda, A. Influence of sediment redox conditions on release/solubility of metals and nutrients in a Louisiana Mississippi River deltaic plain freshwater lake. Sci. Total Environ. 2006, 371, 334–343. Schindler, P. W.; Fürst, B.; Dick, R.; Wolf, P. U. Ligand properties of surface silanol groups. Surface complex formation with Fe3+, Cu2+, Cd2+, and Pb2+. J. Colloid Interface Sci. 1976, 55, 469–475.

ES800953D

VOL. 42, NO. 20, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7621