Arsenic Speciation in Sulfidic Waters: Reconciling Contradictory

Nov 29, 2010 - Interplay of S and As in Mekong Delta sediments during redox oscillations. Van T.H. Phan , Fabrizio Bardelli , Pierre Le Pape , Raoul-M...
0 downloads 0 Views 1MB Size
Anal. Chem. 2010, 82, 10228–10235

Arsenic Speciation in Sulfidic Waters: Reconciling Contradictory Spectroscopic and Chromatographic Evidence Britta Planer-Friedrich,*,† Elke Suess,† Andreas C. Scheinost,‡ and Dirk Wallschla¨ger§ Environmental Geochemistry, Bayreuth University, Universita¨tsstrasse 30, Bayreuth 95440, Germany, and The Rossendorf Beamline (BM20), European Synchrotron Radiation Lab, 6 RUE JULES HOROWITZ, BP 220, Grenoble 38043, France, and Environmental and Resource Sciences Program, Trent University, 1600 West Bank Drive, Peterborough, Ontario K9J 7B8, Canada In recent years, analytical methods have been developed that have demonstrated that soluble arsenic-sulfur species constitute a major fraction of dissolved arsenic in sulfidic waters. However, an intense debate is going on about the exact chemical nature of these compounds, since X-ray absorption spectroscopy (XAS) data generated at higher (mmol/L) concentrations suggest the presence of (oxy)thioarsenites in such waters, while ion chromatographic (IC) and mass spectroscopic data at lower (µmol/L to nmol/L) concentrations indicate the presence of (oxy)thioarsenates. In this contribution, we connect and explain these two apparently different types of results. We show by XAS that thioarsenites are the primary reaction products of arsenite and sulfide in geochemical model experiments in the complete absence of oxygen. However, thioarsenites are extremely unstable toward oxidation, and convert rapidly into thioarsenates when exposed to atmospheric oxygen, e.g., while waiting for analysis on the chromatographic autosampler. This problem can only be eliminated when the entire chromatographic process is conducted inside a glovebox. We also show that thioarsenites are unstable toward sample dilution, which is commonly employed prior to chromatographic analysis when ultrasensitive detectors like ICP-MS are used. This instability has two main reasons: if pH changes during dilution, then equilibria between individual arsenic-sulfur species rearrange rapidly due to their different stability regions within the pH range, and if pH is kept constant during dilution, then this changes the ratio between OHand SH- in solution, which in turn shifts the underlying speciation equilibria. This problem is avoided by analyzing samples undiluted. Our studies show that thioarsenites appear as thioarsenates in IC analyses if oxygen is not excluded completely, and as arsenite if samples are diluted in alkaline anoxic medium. This also points out that thioarsenites are necessary intermediates in the formation of thioarsenates. * Corresponding author. † Bayreuth University. ‡ European Synchrotron Radiation Lab. § Trent University.

10228

Analytical Chemistry, Vol. 82, No. 24, December 15, 2010

The formation of arsenic-sulfur species from sulfide and arsenite was investigated already extensively in solubility studies at the end of the 19th century.1-4 Mono-, di-, tri-, and tetrathioarsenates (AsO4-xSx3-) were proposed to form; the existence of thioarsenites (AsO3-xSx3-), which would be expected on the basis of geochemical expectations, was excluded. Modern analytical techniques, introduced for the measurement of arsenic-sulfur speciation in the late 1990s, yielded contradictory evidence on the formation of thioarsenites versus thioarsenates. Ion chromatography showed that thioarsenates formed both from arsenite-sulfide solutions5,6 as well as during acidification of tetrathioarsenate.7,8 Some confusion persists in the literature because one of the early studies9 on chromatographic separation of arsenic species from arsenite-sulfide model solutions reported them as thioarsenites. As the solutions were prepared under anaerobic conditions, the authors assigned the observed species with S:As ratios of 1, 2, 3, and 4 based on geochemical considerations to mono-, di-, tri-, and a somewhat dubious tetrathioarsenite (As(SH)4H0) species. Later studies disproved this identification as thioarsenites. Comparison with retention times of synthesized thioarsenates5,6 as well as the characterization of fractions collected from the ion exchange chromatography with known S:As ratios by electrospray mass spectrometry5 confirmed the species to be thioarsenates. To date, there is no conclusive evidence for thioarsenites in any chromatographic separation. In contrast to chromatographic results, spectroscopic evidence by X-ray absorption spectroscopy (XAS), including analysis of X-ray absorption near-edge structure (XANES) and the extended X-ray absorption fine structure (EXAFS), confirmed the species formed in arsenite-sulfide model solutions as thioarsenites based on their arsenic oxidation state of +3 and characteristic As(III)-S Brauner, B.; Tomı´cek, F. Fresenius J. Anal. Chem. 1888, 27, 508–513. McCay, L. W. Fresenius J. Anal. Chem. 1888, 27, 632–634. McCay, L. W. Z. Anorg. Chem. 1901, 29, 36–50. McCay, L. W.; Foster, W. Z. Anorg. Chem. 1904, 41, 452–473. Wallschla¨ger, D.; Stadey, C. J. Anal. Chem. 2007, 79, 3873–3880. Stauder, S.; Raue, B.; Sacher, F. Environ. Sci. Technol. 2005, 39, 5933– 5939. (7) Schwedt, G.; Rieckhoff, M. J. Chromatogr., A 1996, 736, 341–350. (8) Planer-Friedrich, B.; Wallschla¨ger, D. Environ. Sci. Technol. 2009, 43, 5007–5013. (9) Wilkin, R.; Wallschla¨ger, D.; Ford, R. Geochem. T. 2003, 4, 1–7. (1) (2) (3) (4) (5) (6)

10.1021/ac1024717  2010 American Chemical Society Published on Web 11/29/2010

bond distances of 2.23-2.24 Å.10 XAS spectra of thioarsenate standards have been provided recently; using these it was shown that thioarsenates can be distinguished from thioarsenites and that upon acidification of tetrathioarsenate, trithioarsenite formed,11 not trithioarsenate as found by chromatographic methods.7,8 This is in accordance with postulations in a 1970 paper12 that thioarsenates are only stable as long as AsOH-groups can form; once the formation of AsSH-groups is forced, the ions become unstable so quickly that hydrolysis to trithioarsenate is only a minor reaction. Polymeric trithioarsenites (AsS3)nn- were suggested to form instead. A recent modeling study provides further support for the coexistence of thioarsenites and thioarsenates.13 Thioarsenites thus appear to form either in solutions containing sulfide and arsenite or during decomposition of thioarsenates, but current chromatographic methods fail to detect them. Coelution,10 oxidation,10 and hydrolysis10,13,14 have been suggested as potential reasons. Coelution of thioarsenites and thioarsenates can be excluded on the basis of both the different chromatographic behavior of As(III) versus As(V) species13 as well as the characterization of ion chromatography fractions by electrospray mass spectrometry5 which showed only thioarsenates and no traces of thioarsenites. With the preparation of samples under anaerobic conditions, immediately followed by analysis with oxygen-free eluents, oxidation has been considered unlikely to quantitatively convert thioarsenites to thioarsenates.6 Hydrolysis could lead to conversion of thioarsenites to arsenite, considering that chromatographic elution of thioarsenates requires highly alkaline eluents and excess OH- may destabilize thioarsenites. In this study, we conducted comparative analyses of arsenitesulfide solutions with S:As ratios from 0.1 to 10 by XAS, and (without further dilution!) by IC-ICP-MS. We investigated the influence of eliminating oxygen during ion exchange chromatography and determined oxidation kinetics in arsenite-sulfide solutions. The effects of dilution were investigated using solutions from 10 mM to 0.1 mM arsenite at a S:As ratio of 10. For 0.1 mM arsenite, speciation changes were additionally monitored from pH 2 to 13 immediately after preparation and after one week of storage under anaerobic conditions. The experiments yielded new insight into thioarsenite stability and their importance for thioarsenate formation. METHODS All solutions were handled under anaerobic conditions in a glovebox with an atmosphere of 95% nitrogen and 5% hydrogen. Stock solutions were prepared of 20 mM arsenite (NaAsO2, Fluka) and 200 mM sulfide (Na2S · 9H2O, Sigma-Aldrich), dissolved in nitrogen-purged, deoxygenated ultrapure water without pH adjustment. From these, arsenite-sulfide model solutions with increasing S:As ratios of 0.1, 1, 2, 4, and 10 at a total arsenite concentration of 10 mM were obtained. A dilution series was (10) Beak, D. G.; Wilkin, R. T.; Ford, R. G.; Kelly, S. D. Environ. Sci. Technol. 2008, 42, 1643–1650. (11) Suess, E.; Scheinost, A. C.; Bostick, B. C.; Merkel, B. J.; Wallschla¨ger, D.; Planer-Friedrich, B. Anal. Chem. 2009, 81, 8318–8326. (12) Thilo, E.; Hertzog, K.; Winkler, A. Z. Anorg. Allg. Chem. 1970, 373, 111– 121. (13) Helz, G. R.; Tossell, J. Geochim. Cosmochim. Acta 2008, 72, 4457–4468. (14) Bostick, B. C.; Fendorf, S.; Brown, G. E. Mineral. Mag. 2005, 69, 781– 795.

produced for a fixed S:As ratio of 10 with arsenite concentrations of 0.1, 0.5, 1, 5, and 10 mM. Final pH values of the model solutions were determined with an HACH pH meter HQ 40d and ranged between 11.2 and 12.5. Immediately after mixing, 100 µL of sample was pipetted into PE sample holders for XAS analysis which were covered with Kapton tape inside the glovebox. The sample holders were flashfrozen in liquid nitrogen outside the glovebox, stored in the freezer, and transported on dry ice from the preparation laboratory in Bayreuth to the Rossendorf Beamline (BM20) at the European Synchrotron Radiation Facility in Grenoble where they were analyzed in a closed-cycle He-cryostat. Details of XAS analysis and spectra interpretation can be found in the Supporting Information. Total storage time until analysis was 5-8 days. To determine the influence of oxygen on thioarsenic speciation, two samples (10 mM arsenite; S:As ratio of 10) were prepared inside the glovebox, and then left exposed to ambient air for 1 and 24 h, respectively, before packing and flash-freezing. To determine the long-term stability of an arsenite-sulfide solution under anaerobic conditions, one sample was left at room temperature for 3 days in the glovebox before flash-freezing. For standard IC-ICP-MS analysis, 500 µL of sample was transferred to IC vials without further dilution. The vials were closed with filter caps which were so far assumed to provide sufficient protection against oxidation and left inside the glovebox for no more than 30 min. Samples were then transferred to the IC autosampler outside the glovebox where analysis with nitrogenpurged eluents was started within less than 1 min. Details on ICICP-MS analysis can be found in the Supporting Information. To determine the influence of pH and time on the speciation in arsenite-sulfide model solutions a subset of samples was prepared containing 0.1 mM arsenite and 1 mM sulfide at nominal pH values of 3, 5, 6, 7, 9, 11, 12, and 13 inside the glovebox. The samples were analyzed immediately after mixing and after one week in the glovebox with IC-ICP-MS outside the glovebox. The actual pH of each solution was measured immediately after taking the subsamples for analysis. To monitor the oxidation kinetics of thioarsenate formation, solutions containing 0.1 mM arsenite and 1 mM sulfide were prepared inside the glovebox, and then left outside either in open vials for 0.5, 1, 1.5, 3, 6, 24, and 96 h or in vials closed with filter caps for 1, 3, 6, 20, and 96 h. The samples were then analyzed with an HPLC gradient pump inside the glovebox to eliminate further species transformations due to oxidation during chromatographic separation. The stability of the original 0.1 mM arsenite-1 mM sulfide solution inside the glovebox was determined after 3 and 24 h. RESULTS AND DISCUSSION Predominance of Trithioarsenite in XAS and Dithioarsenate in IC-ICP-MS. The arsenic speciation determined by XAS and IC-ICP-MS for the five solutions with S:As ratios increasing from 0.1 to 1, 2, 4, and 10 only matched in sulfide-deficient conditions when arsenite was determined as predominant species. For solutions with excess sulfide, XAS showed a predominance of trithioarsenite, IC-ICP-MS a predominance of dithioarsenate (Figure 1). Analytical Chemistry, Vol. 82, No. 24, December 15, 2010

10229

Figure 1. Comparison of (a) the arsenic-sulfur species distribution derived from EXAFS spectra by ITT (Table SI-2), (b) As-S and As-O coordination obtained from EXAFS shell fit (Table SI-1), and (c) that determined by IC-ICP-MS analysis (Table SI-3). For the ITT calculation of species concentrations, the concentrations of AsS3 and AsIII references were fixed at 100%, while the concentrations of the mixtures were calculated without constraints; i.e., they were not normalized to 100%.

Figure SI-1 shows the XANES spectra as well as EXAFS chi and Fourier transformations for the model solutions. The whiteline position of the XANES spectra for the sulfide-deficient sample exactly matched that for sodium arsenite (11 868.5 eV, Figure SI1, Table SI-1). With an increasing excess of sulfide the white-line positions shifted toward lower energies and at a 10-fold sulfide excess the white-line position matched that for the AsIII-S species trithioarsenite (11 867.0 eV,11 Figure SI-1, Table SI-1). Overall, the XANES spectra suggested a mixture of species, probably between the two end-members, arsenite and trithioarsenite, with arsenic in its +3 redox state. However, the white-line position of the pentavalent tetrathioarsenate (11 869.3 eV) was also nearby; hence, it cannot be fully excluded that it also contributes to the observed mixed XANES white-line position signals (Figure SI-1, Table SI-1). To determine the number of dominant species in the model solutions, the EXAFS data were further evaluated with iterative 10230

Analytical Chemistry, Vol. 82, No. 24, December 15, 2010

factor analysis (ITFA). A minimum Malinowski indicator value (IND-factor) was derived which showed that the obtained EXAFS spectra could be reproduced by two components. A linear combination of these two components also yields a good fit of both the chi spectra and Fourier transforms, further confirming that only two structural components are required to reproduce all spectra (Figure SI-1). The principal component analysis (PCA) factor loadings after varimax rotation showed that component 1 dominated the sulfide-deficient sample (98.7%), while component 2 increased with increasing sulfide concentration to a maximum of 97.8% in the sample with a 4-fold sulfide excess and to 95.4% with 10-fold sulfide excess (Table SI-2). These samples were thus assumed to contain almost exclusively the end-member species for the S:As ratio series. The structure of each end-member was resolved by EXAFS shell fitting (Table SI-1). Component 1 showed a single As-O shell and no indication for an As-S coordination. The XANES edge position was indicative of an AsIII-O bond, and the bond distance of 1.78 Å and the coordination number of 3.1 were in line with a typical AsIII-O coordination such as in sodium arsenite (CN ) 3.4, R ) 1.79 Å, Table SI-1). AsV-O bond distances in (thio)arsenates are typically shorter (1.68-1.70 Å,11,15 Table SI-1). Component 2 showed an As-S coordination with bond lengths characteristic for thioarsenites (2.21-2.28 Å10,14,16). The specific bond lengths of 2.24 and 2.23 Å in the mixes with S:As ratios of 4 and 10, the CNs of 3.2 and 2.8, and the XANES white-line at 11 867.0-11 867.5 eV were all typical of trithioarsenite.11,17,18 To verify the identity of the end-members, reference spectra for arsenite and trithioarsenite were added to the ITFA calculations (Figure SI-1, Figure 1a). The first model using the trithioarsenite spectra from our latest study11 failed by not reaching a clear minimum of the IND-factor. This failure might have been due to the influence of an As-As coordination and a beginning AsIII-S precipitation in our sample which could cause slight differences in the molecular structure. In order to have a pure reference spectrum for trithioarsenite, its chi-function was modeled as described in the Methods section. The improved model confirmed the predominance of arsenite in the model solution with a S:As ratio of 0.1 and of trithioarsenite in the model solutions with a S:As ratio of 4 and 10. Varimax rotation and ITT (iterative target test) indicate small portions of a component with an As-O coordination in the latter two solutions, which increases from 3% in the solution with a S:As ratio of 4 to 12% in the solution with a S:As ratio of 10 (Table SI-2). However, this small As-O coordination could not be verified in the EXAFS spectra. For the solution with a S:As ratio of 10, the whole experiment and ITFA were repeated and reanalyzed with the same result (Table SI-1). If the observed increased importance of an As-O coordination at a S:As ratio >4 is no analytical artifact (and IC-ICP-MS results seem to confirm that, see below), it indicates that, at sulfide ratios exceeding the S:As stoichiometry of 3 required for trithioarsenite formation, further reactions occur besides formation of trithioarsenite from arsenite which either (15) Paktunc, D.; Foster, A.; Laflamme, G. Environ. Sci. Technol. 2003, 37, 2067– 2074. (16) Helz, G.; Tossell, G.; Charnock, J.; Pattrick, R.; Vaughan, D.; Garner, D. Geochim. Cosmochim. Acta 1995, 59, 4591–4604. (17) Palazzi, M. Acta Crystallogr., Sect. B 1976, 32, 3175–3177. (18) Seung, D. Y.; Gravereau, P.; Trut, L.; Levasseur, A. Acta Crystallogr., Sect. C 1998, 54, 900–902.

prevent complete reaction of arsenite with sulfide, or form new arsenite. A potential mechanism could be the formation of polysulfides, which can act as elemental sulfur donors. Complexation of elemental sulfur with trithioarsenite will lead to formation of tetrathioarsenate which is not stable at pH 11-12 and will partially transform to trithioarsenate and arsenite.8 However, ITT calculations did not reveal any indications for the presence of trior tetrathioarsenate. The XAS species distribution as derived by ITT is shown in Figure 1a. Fitted oxygen coordination numbers CNAs-O of 2.8 (S: As ) 1) and 1.0 (S:As ) 2) and fitted sulfur coordination numbers CNAs-S of 0.9 (S:As ) 1) and 2.4 (S:As ) 2) show a decrease in the As-O coordination with increasing sulfur presence and an increasing importance of As-S bonds (Table SI-1, Figure SI-1, Figure 1b). The mix with a S:As ratio of 1 could indicate the presence of monothioarsenite with matching bond lengths for AsIII-S and AsIII-O bonds, while the mix with a S:As ratio of 2 could indicate a dithioarsenite, despite even lower bond distances compared to typical AsIII-O bonds. However, for both species we had no reference spectra, and it is questionable if they could be discriminated in mixtures, since their As-S and As-O bond lengths vary too little. With one exception (S:As ) 1), the species sum up to 100 ± 5%, which is the typical error of this kind of analysis, suggesting that first the two references are real components of the mixtures and second no further components are present in significant amounts. The larger deviation of sample S:As ) 1 may indicate the presence of a small amount (e10%) of an unidentified species. Our observations regarding shifts of the XANES white-line positions from orpiment to arsenite, the predominance of arsenite at low S:As ratios, and the formation of thioarsenites with increasing sulfide concentrations generally agree with those of others.10,14 In contrast to Bostick et al.,14 neither Beak et al.10 nor we found evidence for polymeric arsenic-sulfur species. The reason for this discrepancy is unclear as arsenite concentrations were only slightly lower in these latter studies (0.7-3.4 mM in ref 10 and 10 mM in our study, compared to 17-50 mM14). While Bostick and co-workers interpreted their spectra as evidence for the occurrence of solely thioarsenites, their reported As-S bond lengths of 2.12 Å could also indicate the presence of thioarsenates, especially at their S:As ratio of 5. It is thus questionable whether their samples really give unequivocal proof for thioarsenites or have already undergone some oxidation. In contrast to Beak et al.10 who had to use a specially designed nitrogen-purged flowcell for XAS analysis to decrease the likelihood of sample redox transformations, we found that thioarsenites were preserved by sample flash-freezing and analysis in a closed-cycle He cryostat. Summarizing, we believe that the predominance of arsenite at S:As ratios of 0.1 and the predominance of trithioarsenite at S:As ratios >2 as determined by XAS depict the “true” species distribution in our arsenite-sulfide model solutions. Analyzing the identical solutions without further dilution by IC-ICP-MS, we find comparably a predominance of arsenite at S:As ratios of 0.1 but, in contrast to the trithioarsenite predominance in XAS, the formation of dithioarsenate as predominant species besides monothioarsenate at S:As ratios of 1 and trithioarsenate at S:As

Figure 2. Chromatogram of 10 mM arsenite and 100 mM sulfide solution prepared inside the glovebox and analyzed with an IC outside the glovebox using nitrogen-purged eluents (green line) in comparison to an identical sample prepared and analyzed inside the glovebox (red line). Comparison to a 10 mM pure arsenite solution shows a retention time shift for arsenite (blue line).

ratios of 10 (Figure 1a, Table SI-3). We observe a discrepancy between the sum of the arsenic species and the initial arsenite concentration (“loss”) which increases with increasing sulfide concentrations from 6% (S:As of 0.1) to 26% (S:As of 10). The continuous decrease of arsenite in all solutions up to a S:As ratio of 4 and its reappearance in the solution with the highest sulfide excess (S:As 10) as observed in PCA and ITT calculations was also observed during IC-ICP-MS analysis (Figure 1c). As the S:As 10 solution was reanalyzed numerous times, always with the same result, we assume that this is no analytical artifact. An observed increase in tri- and tetrathioarsenate is in line with the postulations made for the ITT results above about the formation of tetrathioarsenate from trithioarsenite at excess sulfide concentrations. The chromatographic behavior of the arsenite peak is interesting. Instead of the original arsenite peak with a retention time of 185 s two peaks were observed: a larger one at 205 s and a smaller one at 160 s (Figure 2). Both with a decreasing S:As ratio at a fixed arsenite concentration of 10 mM and a decreasing total arsenite concentration at a fixed S:As ratio of 10, the two peaks reunited. One wide peak with two maxima was observed for 10 mM arsenite at a S:As ratio of 3 and for 1 mM arsenite at a S:As ratio of 10. Increasing the S:As ratio of a 1 mM arsenite solution to 50 led to further peak splitting. With a 1 mM NaOH eluent instead of the usually applied 100 mM NaOH, two clearly separated peaks were visible down to a concentration of 0.5 mM arsenite at a S:As ratio of 10. Unambiguous identification of the two peaks was impossible. Neither peak showed a simultaneous peak in the sulfur track; thus, there was no indication that either peak represented thioarsenites. Spiking samples with arsenite to determine which peak was the “true” arsenite peak led to recombination of both peaks with the original arsenite retention time. The observed peak splitting depends thus not only on a simple matrix effect due to the high concentrations used, but also on total arsenite and sulfide concentrations, as well as on the S:As and the SH-:OH- ratios. For the present interpretation, both peaks were integrated as arsenite peaks. When comparing the different speciation distributions obtained by XAS and IC-ICP-MS, one has to keep in mind that species with relative concentrations of OH-

thioarsenites

SH- ) OH-

initially arsenite, after some days thioarsenites arsenite (thioarsenites are unstable due to excess of OH-)

SH- < OH-

a

species analyzed by chromatography under anaerobic conditions

species analyzed by chromatography under aerobic conditions

arsenite (artifact: transformation of thioarsenites to arsenite due to excess of OH- by elution at pH 13) arsenite (initially true speciation, later artifact) arsenite

thioarsenates (artifact: oxidation of thioarsenites before alkaline transformation to arsenite; thioarsenates are stable at excess OH-) initially arsenite (true speciation), then thioarsenates (artifact) arsenite (true speciation, even though there is oxygen, there is no formation of thioarsenates without initial formation of thioarsenites)

Indicated by italics.

competitive dissociation as explained above and also at acidic conditions (pH 2.8, 5, 6) due to transformation before precipitation as amorphous As2S3 as reported previously.11 However, at pH 7 and 9.2, trithioarsenate predominates which proves that as the relative SH-:OH- ratio in sulfidic waters is increased by lowering the pH at a constant sulfide concentration, the first reaction (at neutral to moderately alkaline pH) is the formation of thioarsenites, which are then oxidized to thioarsenates upon analysis under aerobic conditions. With the solutions left under anaerobic conditions inside the glovebox for one week (Figure 4b), some speciation changes occurred: Dithioarsenate increased in solutions with pH 0.2 mg/L19) which are sufficient to spontaneously oxidize thioarsenites to thioarsenates. We are thus convinced that the thioarsenates we determined there reflect natural speciation, no sampling or analysis artifact. However, as we know now that thioarsenites must form as intermediate species for thioarsenates to occur, it should be possible to detect them in completely anoxic milieus such as at depth of sulfidic hot springs or in the anoxic bottom waters of Mono Lake. There is thus an urgent need for a suitable analytical method to determine thioarsenites at environmentally relevant concentra(21) Hollibaugh, J. T.; Carini, S.; Gu ¨ rleyu ¨ k, H.; Jellison, R.; Joye, S. B.; LeCleir, G.; Meile, C.; Vasquez, L.; Wallschla¨ger, D. Geochim. Cosmochim. Acta 2005, 69, 1925–1937. (22) Charlet, L.; Chakraborty, S.; Appelo, C. A. J.; Roman-Ross, G.; Nath, B.; Ansari, A. A.; Lanson, M.; Chatterjee, D.; Mallik, S. B. Appl. Geochem. 2007, 22, 1273–1292. (23) Scheinost, A. C.; Charlet, L. Environ. Sci. Technol. 2008, 42, 1984–1989. (24) Webb, S. M. Phys. Scr. 2005, 115, 1011–1014. (25) Ressler, T. J. Synchrotron Radiat. 1998, 5, 118–122. (26) Rossberg, A.; Reich, T.; Bernhard, G. Analyt. Bioanalyt. Chem. 2003, 376, 631–638. (27) Rossberg, A.; Scheinost, A. C. Analyt. Bioanalyt. Chem. 2005, 383, 56–66. (28) Palazzi, M. Acta Crystallogr., Sect. B 1982, 24, 1968–1938. (29) Emmerling, F.; Roehr, C. Z. Naturforschung, B: Anorg. Chem., Org. Chem. 1987, 42. (30) Kempa, P. B.; Wiebke, M.; Felsche, J. Acta Crystallogr., Sect. C 1983, 39. (31) Jaulmes, S.; Palazzi, M. Acta Crystallogr., Sect., B. 1976, 32, 2119–2122. (32) Dittmar, G.; Schaefer, H. Z. Naturforsch., B 1978, 33, 678–681. (33) Mullen, D. J. E.; Nowacki, W. Z. Kristallogr., Kristallgeom., Kristallphys., Kristallchem. 1972, 48–65.

tions. We believe that, due to the instability of thioarsenites under acidic as well as highly alkaline conditions and in the presence of even trace amounts of oxygen, methods based on ion exchange chromatography suffer serious limitations and optimization of such techniques for thioarsenite separation may not be possible. At present, the best available practice is to sample and analyze thioarsenic species under complete exclusion of oxygen and to report so-detected “arsenite” as the sum of arsenite and potentially present thioarsenites. ACKNOWLEDGMENT We would like to acknowledge generous funding from the German Research Foundation within the Emmy Noether program (Grant PL 302/3-1) as well as the financial support for a Ph.D. stipend to Elke Suess from the German National Academic Foundation. Technical support and helpful discussions by the team at the Rossendorf beamline at ESRF and assistance during sample preparation by Stephan Weiss from the Research Center DresdenRossendorf, Germany, are greatly acknowledged. SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review September 22, 2010. Accepted November 10, 2010. AC1024717

Analytical Chemistry, Vol. 82, No. 24, December 15, 2010

10235