Estimation of Se(VI)/Se(IV) ratio in water by the ratio recorded in barite



[1] It is possible that the distribution behaviors of redox-sensitive elements such as selenium (Se) between authigenic minerals and water can provide information on the oxidation state of the element in the coexistent water during mineral deposition. Considering the similar chemical characteristics of Se(IV) and Se(VI) oxyanions, we propose that the Se(VI)/Se(IV) ratio in a particular precipitate, such as barite, may enable us to estimate the Se(VI)/Se(IV) ratio in the coexistent water. The coprecipitation experiments of Se with barite coupled with the determination of the oxidation state of Se both in the aqueous phase and in barite were conducted by Se K-edge X-ray absorption near-edge structure and high-performance liquid chromatography connected to inductively coupled plasma-mass spectrometry, respectively, to investigate the influence of the oxidation state of Se on its immobilization into barite at pH 4.0 and 8.0. The results showed that the Se(VI)/Se(IV) ratio in barite was correlated with the Se(VI)/Se(IV) ratio in water, which in turn can provide physicochemical and biogeochemical information related to the Se(VI)/Se(IV) ratio in water in the environment where barite precipitated.

1. Introduction

[2] Generally speaking, oxyanions in water show conservative behavior compared with polyvalent cations that can be strongly affected by their adsorption on particulate matter in water. Among various elements that form oxyanions, such as sulfur, chromium, arsenic (As), selenium (Se), molybdenum, and rhenium, Se is unique in terms of having two oxidation states that form oxyanions in natural water. At equilibrium, the concentration ratios of various redox species such as Se(VI) and Se(IV) can be affected by the redox potential (Eh, if standardized to hydrogen electrode) in the environment. It is often the case, however, that the oxidation states of inorganic ions in seawater, selenate or selenite, are kinetically controlled, where thermodynamically unstable species can be present [Cutter and Bruland, 1984; Cutter, 1992; Rue et al., 1997]. The presence of such unstable species can be caused by biological effects, transport of the species by lateral advection and sinking particles, and so on. Thus, the Se(VI)/Se(IV) ratio in water can be influenced by various factors including redox condition, biological activities, and mixing of these species of different origins.

[3] Selenium is dissolved in water as selenate (SeO42−) or selenite/hydroselenite (SeO32−/HSeO3) ion under oxic and suboxic conditions, respectively, but can be immobilized under reducing condition by its reduction to native selenium (Se(0)) because of its low solubility [Sharmasarkar et al., 1998; Bujdos et al., 2005; Harada and Takahashi, 2009]. Selenate and selenite are highly soluble ions, but they can be incorporated into the solid phase by adsorption or coprecipitation. The distribution behavior of Se between the solid phase and water is variable depending on its oxidation state and chemical form [Fernandez-Martýnez and Charlet, 2009]. Thus, the Se(VI)/Se(IV) ratio recorded in the solid phase by coprecipitation can be used to estimate the Se(VI)/Se(IV) ratio in water.

[4] Barite (BaSO4), which is stable under a wide range of pressure, temperature, Eh, and pH, can incorporate various trace elements that may record geochemical information in water and eventually reflect depositional condition [Griffith and Paytan, 2012] (Figure 1). Several elements (for example, cations: Sr2+, Ca2+, Ra2+, Pb2+, and Nd3+; anions: SeO42−, CrO42−, and MnO42−) are incorporated into barite by substituting Ba2+ or SO42−, respectively [Hein et al., 2000; Zhu, 2004]. It has been reported that Se is strongly bound mainly in the sulfate site of the crystal lattice of barite [Andara et al., 2005]. If both selenite and selenate can be incorporated into barite, Se may preserve information about the Se(VI)/Se(IV) ratio in water in the depositional environment.

Figure 1.

Eh-pH diagrams for (a) Ba-S-H2O and (b) Se-H2O at 25°C and 1.0 bar calculated based on Visual MINTEQ. The total dissolved concentrations of Ba, Se, and S used for the calculations were 270 μg L−1, 140 μg L−1, and 2600 μg L−1, respectively, which approximately correspond to their concentrations in our experimental systems.

[5] The purpose of this study is to know whether it is possible to estimate the Se(VI)/Se(IV) ratio in water based on the Se(VI)/Se(IV) ratio in barite. No study has been conducted on the utilization of the oxidation states of trace elements in precipitates to estimate the redox state of the element in the coexisting water. We already have conducted related studies based on a similar idea for As and Se incorporation into calcite [Yokoyama and Takahashi, 2011; Yokoyama et al., 2012]. However, we found that arsenate and selenite were selectively incorporated into calcite even when only arsenite or selenate was initially added into the solution, respectively. The results suggested that As and Se in calcite are not appropriate to estimate the As(V)/As(III) and Se(VI)/Se(IV) ratios in water. Therefore, the distribution behavior of Se between barite and water needs to be examined with speciation analyses of Se both in solid phase and water to establish it as a method for estimating the Se(VI)/Se(IV) ratio in water.

[6] In this study, artificial seawater (ASW) was used to estimate the solid-water distribution of Se in seawater. X-ray absorption near-edge structure (XANES) was used to determine the oxidation state of Se in the solid phase directly, whereas high-performance liquid chromatography (HPLC) connected to inductively coupled plasma mass spectrometry (ICP-MS) was used to determine the oxidation state of Se in water. The apparent distribution coefficient (Kd) of Se between barite and water was also measured by ICP-MS to clarify the difference in the immobilization mechanisms between Se(IV) and Se(VI) at pH 4.0 and 8.0.

2. Materials and Methods

[7] Laboratory experiments reproducing the immobilization of Se by barite were conducted at various Se(VI)/Se(IV) ratios in the initial solution prepared by mixing Se(IV) and Se(VI) solutions. Each Se(IV) and Se(VI) stock solution was prepared from analytical grade reagents of NaHSeO3 and Na2SeO4 (Wako, Japan), respectively. Aqueous speciation and mineral saturation states in the solutions were calculated based on the Visual MINTEQ program using a database from NIST [2003] [Gustafsson, 2009].

2.1. Coprecipitation Experiment Procedure

[8] In the present study, coprecipitation of Se with barite was conducted through the spontaneous precipitation of barite. Barite was precipitated from a mixture of (i) ASW containing 27 mM sulfate and (ii) 2.0 mM BaCl2•2H2O solution [Blount, 1974]. The ASW consisted of 440 mM NaCl, 50 mM MgCl2•2H2O, 27 mM Na2SO4, 9.6 mM CaCl2• 2H2O, and 2.2 mM NaHCO3. Right before the addition of BaCl2•2H2O solution, Se(IV) and/or Se(VI) were added to ASW. This solution was continuously adjusted to a specific pH (pH 4.0 or 8.0) by adding a small amount of HCl or NaOH solution during the experiments. The two pH conditions were used to assume (i) general hydrothermal pH condition (pH = 3.0 to 4.0) and (ii) seawater condition (pH = 8.0) [Von Damm, 1990]. The mixed solution was stirred under ambient condition with a magnetic stirrer at 25°C for 168 h (= 7 days). The saturation indices of barite (SIbarite) in the initial solutions were fixed at 4.2 by adjusting the Ba2+ concentration to keep the same initial precipitation rate of barite for all experiments in this study (Table 1). Under this condition, barite rapidly precipitated because of its high supersaturated condition. SI is defined as SI = log(IAP/Ksp), where IAP is the ion activity product and Ksp is the solubility product of a mineral. The initial solution (Se concentration: 0.10 mM) was unsaturated with respect to the solid phases of barium selenite and barium selenate to avoid the formation of these Se precipitates in the system (Table 1).

Table 1. Experimental Conditions of Coprecipitation Experiments
 Se in input solutionpHSe after the coprecipitationsSe in Barite 
Run no.Concentration (mg/L) inline image inline imageSIbariteaSIBaSeO3bSIBaSeO4bInput SolutionAfter ReactionConcentration (mg/L) inline image inline image inline image inline image inline image inline imageConcentration (mg/kg)Kdc(L/kg)
  1. a

    Each saturation index (SI) was calculated by Visual MINTEQ.

  2. b

    BaSeO3 and BaSeO4: Typical barium selenite and barium selenate minerals, respectively.

  3. c

    Kd: Apparent distribution coefficient. Kd = [Se]B/[Se]W ([Se]B: Se concentration in barite; [Se]W: Se concentration in water).

ba-18.601004.21−1.89 × 1034.0 × 102
ba-28.820804.21−1.93− × 1033.8 × 102
ba-38.840604.21−2.11−0.838.08.08.206040010002.4 × 1032.9 × 102
ba-48.760404.21−2.28−0.658.07.98.204060010002.1 × 1032.6 × 102
ba-58.770304.21−2.41−0.588.08.08.203070010001.8 × 1032.2 × 102
ba-68.680204.21−2.59−0.538.08.08.302080088.111.91.1 × 1031.3 × 102
ba-78.790104.21−2.89−0.478.08.08.301090081.218.81.3 × 1031.6 × 102
ba-88.710004.21 −0.438.08.08.400100016.883.28.6 × 1021.0 × 102
ba-1′8.701004.21−5.32 × 1029.5 × 101
ba-2′8.620804.21−5.41− × 1028.8 × 101
ba-3′8.640604.21−5.54−0.834.03.98.40604015.152.332.57.3 × 1028.7 × 101
ba-4′8.360404.21−5.72−0.654.04.18.2040607. × 1024.4 × 101
ba-5′8.580204.21−6.02−0.534.03.98.302080041.758.27.2 × 1028.7 × 101
ba-6′8.410004.21 −0.434.03.98.200100013.186.96.6 × 1028.1 × 101

[9] The precipitates of barite and aqueous phase were separated by filtration with a 0.20 μm membrane filter (mixed cellulose ester, Advantec, Tokyo, Japan) and rinsed thrice with Milli-Q water. The X-ray diffraction (XRD) patterns of the precipitates were measured using a powder X-ray diffractometer (MultiFlex, Rigaku Co., Tokyo, Japan), where mineral phase identification was carried out by matching the XRD patterns to the standards of the International Center for Diffraction Data (ICDD) file. Total Se concentrations in the solution before and after filtration were analyzed by ICP-MS (7700cs, Agilent, Tokyo, Japan) or inductively coupled plasma atomic emission spectrometry (ICP-AES; SII, SEIKO, Chiba, Japan) after diluting them in 2 wt.% HNO3 solution. The oxidation states of Se in the precipitates and in the remaining solution after filtration were determined by XANES and HPLC-ICP-MS (7500cs, Agilent, Tokyo, Japan), respectively. A portion of the precipitates was immediately packed into an airtight polyethylene bag as slurry, and stored at 4°C until XANES measurement. The drying process was avoided because it can alter the Se(VI)/Se(IV) ratio in barite.

2.2. Analysis Methods

2.2.1. HPLC-ICP-MS

[10] The Se(VI)/Se(IV) ratio in the water sample was determined by HPLC-ICP-MS. The Pu-2089 Plus pump and a Co-2065 Plus (JASCO) oven were used in the experiment. An anion exchange column (Hamilton, PRP-X100; column length: 25 cm) was used at 40°C. The mobile phase was the aqueous solution containing 10 mM EDTA (Wako, Japan), 2.0 mM potassium hydrogen phthalate (Wako, Japan), and 3.0 wt.% methyl alcohol (Wako, Japan). To remove any Se remaining in the inlet and syringe, about 100 μL of Milli-Q water was injected into the syringe and also into the HPLC system after measurement to avoid contamination in subsequent analyses.

2.2.2. XANES Measurement and Data Analysis

[11] XANES technique is a suitable method to determine the Se(VI)/Se(IV) ratio in barite, because it is a nondestructive analytical method. The K-edge XANES spectra of Se were obtained in the beamline BL01B1 of SPring-8 (Hyogo, Japan) and in the beamline BL-12C of Photon Factory, KEK (Tsukuba, Japan) with an Si(111) double-crystal monochromator and two mirrors. The monochromator was calibrated using a reference sample, NaHSeO3, which was prepared as a pellet after dilution with boron nitride powder and the energy value of 12.6545 keV was assigned to the peak maximum in the XANES region for the compound. The XANES spectra of the reference compounds were collected in transmission mode, whereas those of the experimental samples were obtained in fluorescence mode using a 19 element germanium semiconductor detector at 90° to the incident beam. In the fluorescence mode, the samples were positioned at 45° with respect to the incident beam. Multiple scans were obtained to check the radiation damage that can change the valence ratio of Se in the samples, where no appreciable change was found in every spectrum obtained here. The measurements were carried out at room temperature under ambient condition. The XANES data were analyzed using a REX2000 (Rigaku Co.) for fitting the sample spectra by linear combination of the reference materials. The Se(VI)/Se(IV) ratio in the solid sample was determined by least-squares fitting using the reference spectra of BaSeO3 (Wako Pure Chem., Ltd., Japan) and BaSeO4 (Alfa Aesar, UK), which were received as analytical grade reagents.

3. Results and Discussion

3.1. The Oxidation State of Se in Barite

3.1.1. pH 8.0 System

[12] X-ray diffraction results for the samples formed at pH 8.0 indicated that the precipitates consisted of only barite. The formations of barium selenite and barium selenate were not likely, because they were under saturation in terms of solubility products. Barium sulfate mineral can be exclusively present as barite in nature without any other polymorphs based on the survey of Inorganic Crystal Structure Database (ICSD). Thus, the transformation of barite to other minerals associated with aging or pH condition is unlikely. This result suggested that if Se was strongly bound mainly in the sulfate site in the crystal of barite, the Se(VI)/Se(IV) ratio in barite can be preserved during aging. The intensities and widths of the peaks in the XRD were similar to those of the reference samples of barite, showing that the incorporation of Se did not alter the main mineralogy of barite because of the low Se concentration added to the system.

[13] The oxidation state of Se in the precipitates was determined by XANES based on the fitting of the spectra of samples by the linear combination of the spectra of BaSeO3 and BaSeO4, where most of the samples could be fitted by these two reference materials (Figure 2; Table 1). The Se(VI)/Se(IV) ratios obtained from this method showed that (i) most Se in the precipitates were coprecipitated as Se(IV) for the samples when inline image (= Se(VI)/ΣSe ratio in water determined by HPLC-ICP-MS) was lower than 0.7, but (ii) inline image (= Se(VI)/ΣSe ratio in barite determined by XANES) increased when inline image was over 0.7 (Figure 3). These findings indicated that inline image was primarily correlated with inline image in water. When inline image ( inline image: Se(IV)/ΣSe ratio in water determined by HPLC-ICP-MS), Se was incorporated only as Se(IV) in the precipitates. When inline image, on the other hand, a part of Se was incorporated as Se(IV), but more than 70% of Se was mainly incorporated as Se(VI). All these results were reproducible at least under our experimental conditions based on our repeated coprecipitation experiments (N = 3) with XANES and HPLC-ICP-MS analyses.

Figure 2.

Normalized Se K-edge XANES spectra of the reference materials (BaSeO3, BaSeO4, and native Se) and Se in barite at (a) pH 8.0 and (b) pH 4.0 as a function of inline image (= Se(VI)/ΣSe).

Figure 3.

Relationship between inline image and inline image. The Se(VI)/Se(IV) ratios in barite and in solution were measured by XANES and HPLC-ICP-MS, respectively. The Se(VI)/{Se(IV) + Se(VI)} ratio was also shown for pH 4.0 system.

[14] These analyses suggested that the reduction of Se(VI) occurred at inline image possibly because the stability of Ba-selenite is larger than that of Ba-selenate, which promotes Se(VI) reduction. To confirm this result, the inline image was also measured during the formation of barite by HPLC-ICP-MS. The results showed that inline image did not change before and after coprecipitation, suggesting that the initial inline image was preserved during the experimental period. The redox change induced by the complexation reaction was also confirmed by Yokoyama et al. [2012], showing that As was immobilized in calcite as As(V) after the oxidation of As(III) to As(V) because of the larger stability of Ca-arsenate than that of Ca-arsenite.

[15] When inline image was above 0.7, inline image shows a positive correlation with inline image at pH 8.0. At inline image below 0.7, Se(VI) was not found in barite within the detection limit of XANES analysis (= about 5% as shown for the arsenate-arsenite system in Takahashi et al., 2003). Thus, although the reduction of Se(VI) to Se(IV) inhibits precise determination of inline image by inline image, we can estimate whether more than 70% of dissolved Se is Se(VI) or not based on the presence or absence of Se(VI) in barite determined by XANES analysis.

3.1.2. pH 4.0 System

[16] Coprecipitation experiments at various inline image were also conducted at pH 4.0 to simulate Se distribution to barite in hydrothermal systems. XRD results showed that the precipitates consisted only of barite without other minerals above the detection limit (>5%). The initial inline image was preserved again at pH 4.0 during aging as revealed by HPLC-ICP-MS analysis, indicating that the direct reduction of Se(VI) to Se(IV) and/or Se(IV) to Se(0) cannot be observed in water.

[17] As for Se species in the solid phase, however, Se(IV), Se(VI), and a smaller amount of Se(0) were detected at pH 4.0 by Se K-edge XANES (Figure 2b). Unfortunately, it was impossible to unequivocally verify the Se(0) precipitates by XRD, because the Se(0) concentration was lower than the detection limit of XRD (>5%). The oxidation state of Se in barite showed that Se(VI) was incorporated into barite to a larger degree than that at pH 8.0 as shown by XANES analysis (Figure 3). Results showed that (i) inline image was strongly correlated with inline image, (ii) Se(0) was also detected under high inline image condition in water at pH 4.0. In particular, Se was incorporated as multiple Se species of 72% Se(IV), 23% Se(0), and 5% Se(VI) in the precipitates at inline image. On the other hand, at inline image, 13% Se(IV) and 87% Se(VI) were detected in the precipitates without any contribution of Se(0). The inline image represented a positive correlation against inline image (Figure 3), as we also found in the pH 8.0 system. Selenium(0) was detected as well in barite at a high inline image in this experiment. This result suggested that Se(0) was formed more readily in the presence of barite, because inline image was stably conserved without the formation of Se(0) in a barite-free system for the experimental period of this study. To exclude the effect of Se(0) formation, the ratio of Se(VI)/{Se(IV)+Se(VI)} was also plotted against inline image (note that Se(0) is not present as dissolved species in water), where a good correlation was still observed between them. Thus, it is suggested that the Se(VI)/Se(IV) ratio in barite can reflect the ratio in water at pH 4.0 even in the presence of Se(0) in the solid phase.

[18] As shown above, reduction of Se(VI) to Se(IV) or Se(IV) to Se(0) was observed during the incorporation of Se into barite, but the reductant in the reactions is not clear at present. The effect of X-ray beams on the redox change of Se was carefully examined by repeated scans of XANES spectra, which shows that no radiation effect was observed in our measurement. Given the low Se concentration, we believe that the reduction of Se(IV) to Se(0) can occur by any reductants in the system. One possibility can be water itself by the reaction of H2O → 2H+ + 1/2O2 + 2 e. Selenium(0) was not found at pH 8.0, but only found at pH 4.0, where the stability field of Se(0) in the Eh-pH diagram is larger (Figure 1b). In addition, Se(0) was found only in the presence of barite, possibly because the precipitation of Se(0) can be promoted in the presence of the solid surface.

[19] Similarly, we think that the reduction of Se(VI) at inline image occurred due to the increase in stability of Se(IV) in the system through the formation of barium selenite, which is a much more stable complex than barium selenate. The redox change during the incorporation into such precipitates was also reported by (i) Yokoyama et al. [2012], who have shown the As(III) oxidation during its incorporation into calcite and (ii) Tsuno et al. [2003], who have shown the reduction of Yb(III) to Yb(II) during its incorporation into calcite. Thus, it is suggested that the immobilization of Se(IV) was promoted by the reduction of Se(VI) to Se(IV) via formation of barium selenite complexes. The electron donor in the solution was unclear, but we think that water can be a reductant because of the oxidation of water into O2 (H2O → 2 H+ + 1/2 O2 + 2 e) as suggested above.

3.1.3. Comparison of the Distribution Behavior of Se in Both pH Systems

[20] In addition to the examination of two different pH conditions, similar coprecipitation experiments were conducted at various systems as summarized in Table 2. All experiments were conducted at inline image or inline image at pH 8.0. At first, other chemical compositions of the solution was tested, including the use of ASW (results were already shown above), ASW without bicarbonate (carbonate), ASW added with phosphate, and just Milli-Q water, where the saturation index (SI) of barite and Se concentration added to the system were kept constant. Second, the SI of barite was also varied from 2.9 to 4.2 at constant Se concentration with various barium concentrations to evaluate the effect of the precipitation rate. Previous studies showed that the precipitation rate depends on the saturation state in solution, which has a significant influence on the distribution behavior of several ions between mineral and water [Lorens, 1981; Alexandratos et al., 2007].

Table 2. Additional Experiments With Various Initial Solution, Where pH was Fixed at 8.0
 SampleSe in Input Solution (mg/L)Se Species in WaterSI of Barite inline image (%) inline image (%)KSe(IV) (L/kg)aKSe(VI) (L/kg)b
  1. a

    KSe(IV) = Kd × ( inline image/ inline image), where inline image and inline image are the Se(IV)/ΣSe ratio in the barite and in the solution, respectively.

  2. b

    KSe(VI) = Kd × ( inline image/ inline image), where inline image and inline image are the Se(VI)/ΣSe ratio in the barite and in the solution, respectively.

Exp 1Milli-Q water10.5 4.2116.383.7 7.0 × 101
Exp 2Artificial Sea Water (ASW8.6 4.2112.088.0 8.7 × 101
Exp 3ASW without carbonate9.1 4.2112.187.9 1.3 × 102
Exp 4ASW with phosphate8.9 4.2113.886.2 8.2 × 101
Exp 5ASW with higher Se conc.93.6Se(VI)4.2110.189.9 2.7 × 102
Exp 6ASW with various SI of barite7.63.9119.280.8 2.4 × 102
Exp 7ASW with various SI of barite7.5 3.6122.777.3 1.5 × 102
Exp 8ASW with various SI of barite7.6 3.2118.481.6 5.6 × 102
Exp 9ASW with various SI of barite7.7 2.9125.974.1 1.0 × 103
Exp 10ASW with various SI of barite7.8 2.6120.479.6 1.5 × 103
Exp 11AWS8.6Se(IV4.2110003.6 × 102 
Exp 12ASW with various SI of barite9.03.2110001.2 × 103 

[21] The results showed that the apparent Kd value can be variable under different conditions: Kd increased (i) in the absence of carbonate, (ii) at a larger initial Se concentration, and (iii) at a lower SI. Especially, SI had strong influence on the partitioning of Se into barite, showing that log Kd linearly decreased with increasing SI (Figure 4). Based on the XANES analyses, however, the inline image was constant at 83% ± 5%, which showed that the inline image was not changed at various SI employed in this study. All these results suggested that the relationship between inline image and inline image was not affected even when the chemical composition, SI, and Se concentration were changed to a certain degree.

Figure 4.

Relationship between the saturation index (SI) and apparent distribution coefficient (Kd) at pH 8.0. SI = log(IAP/Ksp), where IAP and Ksp are the ion activity product and solubility product of barite, respectively.

[22] Compared with these parameters, pH markedly affected the incorporation of Se species into barite, possibly because of the different chemical forms of dissolved species of Se(IV) at different pH values, as will be discussed later. The effect of temperature, which has not been examined here, must be studied in the future especially to estimate the Se(VI)/Se(IV) ratio in hydrothermal water based on the ratio in barite. Similarly, the vital effect on the Se(VI)/Se(IV) ratio must be studied if barite precipitated in biogenic debris.

3.2. Distribution Coefficient of Se

[23] Variation of the apparent distribution coefficient (= Kd) of Se between barite and water were plotted against inline image for systems aged at pH 8.0 and 4.0 (Figure 5). The apparent distribution coefficient of Se between barite and water is defined as

display math

where [Se]B and [Se]W are the total Se concentration in the barite (mol/kg) and in water (mol/L), respectively. In this study, the amount of barite was calculated from the difference in concentration of barium between the input (initial) and final solutions by ICP-AES measurements [Yokoyama et al., 2012]. The calculated amount of barite formed in each solution was (1.10 ± 0.03) × 10−1 g in the experiment at SI = 4.2.

Figure 5.

The apparent distribution coefficient (Kd) at various inline image obtained by the experiments at pH 8.0 and pH 4.0.

[24] For the pH 8.0 system, the total Se concentration in barite and water were determined after 168 h, when the precipitation process of barite was completed. Consequently, a much larger amount of Se was incorporated into barite at inline image compared with that at inline image (Table 2). The apparent Kd values at pH 8.0 decreased from 4.0 × 102 L/kg to 1.0 × 102 L/kg with increasing inline image in water (Figure 5). Thus, Se(IV) was more preferentially incorporated into barite than Se(VI) under surface seawater condition at pH 8.0. As for the pH 4.0 system, on the other hand, the distribution of Se to barite showed a different trend from that at pH 8.0; the apparent Kd value remained nearly constant at various Se(VI)/Se(IV) ratios in water; Kd values were 9.5 × 101 L/kg and 8.1 × 101 L/kg at inline image and inline image, respectively (Figure 5).

[25] The apparent Kd of Se(IV) and Se(VI) defined as KSe(IV) and KSe(VI) were determined from the equation:

display math


display math

where inline image is Se(IV)/ΣSe ratio in barite determined by XANES. In addition, the apparent Kd value can be written as

display math

[26] This equation implies that Kd is correlated with inline image if KSe(IV) and KSe(VI) are constant at various inline image. The results showed a linear correlation between inline image and apparent Kd values at pH 8.0 (r2 = 0.95; Figure 5). Calculations showed that KSe(IV) and KSe(VI) at pH 8.0 were 4.0 × 102 L/kg and 8.5 × 101 L/kg, respectively. The KSe(IV)/KSe(VI) can be estimated as 4.7, indicating that Se(IV) was preferentially coprecipitated into barite compared with Se(VI) in the pH 8.0 system (Table 3). On the other hand, in the pH 4.0 system, KSe(IV) and KSe(VI) are calculated as 9.0 × 101 L/kg and 7.0 × 101 L/kg, respectively, indicating that the distribution of Se to barite was almost identical between Se(IV) and Se(VI) in the pH 4.0 system.

Table 3. Apparent Kd of Se(IV) and Se(VI) Between Water and Barite and Their Ratios at pH 4.0 and 8.0
 Kd (L/kg)
 pH 8.0pH 4.0
KSe(IV)4.0 × 1029.0 × 101
KSe(VI)8.5 × 1017.0 × 101

3.3. Comparison of the Distribution Behaviors of Se in Both pH Systems

[27] The degree of adsorption of Se depends on pH, especially for Se(IV) [Sharmasarkar et al., 1998; Seby et al., 2001; Torres et al., 2010]. Selenium(IV) is a weak acid that can exist as H2SeO3, HSeO3, or SeO32−, depending on the solution pH (pKa1 = 2.70, pKa2 = 8.54). On the other hand, Se(VI) is dissolved only as SeO42− in the pH region of natural water due to the low pKa1 and pKa2 (pKa1 = −2.01, pKa2 = 1.80) [Torres et al., 2011; Tanaka et al., 2013]. Speciation calculations based on pKa showed that Se(IV) dissolved as HSeO3 in the pH range from 2.7 to 8.5 and as SeO32− at pH above 8.5 under our experimental conditions. Thus, Se(IV) is dissolved only as the monovalent anion HeSeO3 in the pH 4.0 system, whereas 60% SeO32− and 40% HSeO3 were present in the pH 8.0 system in this study (Figure 1b). A number of previous studies have shown that SeO32− forms more stable complexes with cations than SeO42−, whereas HSeO3 forms less stable complex than SeO32− because of the lower charge on HSeO3 [Abdullah et al., 1995; Smedley and Kinniburgh, 2002]. As a result, an obvious trend between Se speciation and the incorporation behavior of Se into barite was observed in this study. In the pH 8.0 system, Se(IV) was incorporated into barite to a larger degree than Se(VI). This result was possibly due to the larger stability of surface complex of SeO32− to barite surface compared with SeO42−, which was suggested from the larger stability of BaSeO3 complex than that of BaSeO4 based on the Linear Free Energy Relationship (LFER) [Stumm, 1992], describing the linear relationship between the logarithms of stabilities of the surface complex and corresponding complex in solution. In the pH 4.0 system, on the other hand, Se(IV) was incorporated to a lesser degree compared with the pH 8.0 system because of the relative decrease in the affinity of Se(IV) for barite as the Se(IV) species in water changes from SeO32− to HSeO3.

[28] These findings indicated that pH can have a strong influence on incorporation of Se into barite because of the different proton dissociation behaviors of Se(IV) and Se(VI) at various pH. This result indicates that pH information where barite precipitated is also necessary to estimate the Se(VI)/Se(IV) ratio in water based on the ratio in barite.

4. Implications and Conclusions

[29] The present work shows that the inorganic Se(VI)/Se(IV) ratio in water can be estimated by that of the Se(VI)/Se(IV) ratio in barite. To the best of our knowledge, this is a first study to estimate the redox species in water by using the valance ratio of a trace element in the precipitates. For instance, Yokoyama and Takahashi [2011], Yokoyama et al. [2012] demonstrated that, in the case of arsenic and selenium incorporation into calcite, calcite selectively incorporates arsenate rather than arsenite, and selenite rather than selenate. These phenomena are associated with the higher stability of calcium arsenate and calcium selenite complexes than those of arsenite and selenate, respectively. However, in the barite-selenium system, the affinities of selenite and selenate for barite are more or less similar, which enables us to use the Se(VI)/Se(IV) ratio in barite to estimate the redox species in water.

[30] If the Se(VI)/Se(IV) ratio in water can be estimated by the ratio in barite, what are the geochemical implications obtained by the estimated ratio? If the Se(VI)/Se(IV) ratio in water is under redox equilibrium, the present method enables us to estimate the redox condition in the environment where barite deposited, since the Se(VI)/Se(IV) ratio allows us to calculate redox potential if it is under equilibrium. In most cases, however, the Se(VI)/Se(IV) ratio in natural water such as seawater is known to be controlled by kinetic factors, where thermodynamically unstable species can be found such as Se(IV) and Se(VI) under oxic and reducing conditions, respectively [Cutter and Bruland, 1984; Cutter, 1992; Rue et al., 1997]. Thus, if we could recover barite in various environments, the Se(VI)/Se(IV) ratio in barite may provide information on these processes such as biological effect and transport of specific species of Se from other sources. Thus, it is expected that the Se(VI)/Se(IV) ratio in water estimated by the ratio in barite can have implications on biogeochemical processes of Se and transport of Se(VI) or Se(IV) in the environment where barite was deposited.


[31] This research was supported by a Grant-in-Aid for Scientific Research. This work was performed with the approval of KEK-PF (2012G111 and 2013B052) and SPring-8 (2012B1428, 2013A1177, and 2013B1212).