Expansion of the Conceptual Model for the Accumulation of Selenium in Lentic Food Chains to Include Redox-Controlled Generation and Diffusion of Selenite and Dissolved Organo-Selenium Compounds.

The controls governing the availability of reduced selenium (Se) species, namely selenite (Se[IV]) and dissolved organo-Se (DOSe), to primary producers at the sediment-water interface in depositional environments (i.e., lentic systems) were assessed through consideration of theoretical principles and field data. Selenite is generated in suboxic sediment porewater via the microbially mediated reduction of selenate (Se[IV]) and/or reductive dissolution of Se-bearing iron oxides. Field data for lentic environments demonstrate that the production of DOSe in sediment porewaters can also be redox- and depth-dependent. In this manner, the remobilization depths of Se(IV) and DOSe in depositional environments are dependent on the vertical redox gradient (dEh/dz), where deeper depths of remobilization are observed in less reducing sedimentary environments (lower dEh/dz). In turn, remobilization depth has a direct bearing on the concentration of dissolved Se(IV) and DOSe that may be realized at the sediment-water interface because the depth of reaction governs the diffusive path length, concentration gradient, and rate of diffusional transport toward the sediment-water interface. The principles that link sediment redox gradients, depth of remobilization, diffusive transport processes, and concentration of reduced Se species at the sediment-water interface have a direct bearing on the potential for Se uptake by primary producers in lentic food chains (e.g., phytoplankton, biofilms, bacteria). Overall, these processes complement the current conceptual "benthic detrital food chain" model that describes the accumulation of Se in lentic systems. Environ Toxicol Chem 2022;41:2859-2869. © 2022 SETAC.

The identification of selenopolythionates in aqueous solutions by electrospray ionization-Fourier transform-ion cyclotron resonance-mass spectrometry.

Selenosulfate (SeSO32-) has been shown to occur in certain industrial process waters, and selenopolythionates (SenSxO62-) can be suspected to form from SeSO32- via oxidative or addition reactions. We report here the first observation of selenopolythionates in waters by mass spectrometry. The high mass accuracy and ultra-high resolution of Fourier transform-ion cyclotron resonance-mass Spectrometry with electrospray ionization (ESI-FT-ICR-MS) were used to analyze the isotope patterns of selenium (Se), sulfur (S), and oxygen (O) satellites, in order to provide unequivocal determination of the molecular sum formula of three different selenopolythionates. An aged aqueous solution of SeSO32- was shown to contain the sodium adducts of selenotrithionate (NaSeS2O6-), diselenotetrathionate (NaSe2S2O6-), and triselenopentathionate (NaSe3S2O6-). The identity of these ions was confirmed by accurate mass determination (Δ m/z < 3 ppm error) and by isotopic intensity ratio analysis of the [MIS+2] satellites. Furthermore, Collision Induced Dissociation (CID) was applied to selenotrithionate to distinguish between isomers, and the fragmentation mass spectrum reveals that the Se atom in NaSeS2O6- is located in the middle of the chalcogen chain. Ion chromatographic analysis of the analyzed selenosulfate solutions indicates that selenopolythionates are not suitable for determination by common separations employed for Se speciation analysis, which emphasizes the value of ESI-FT-ICR-MS for complete qualitative characterization of trace element speciation in solution.

Determination of selenocyanate, selenate, and selenite in mining wastewater by GC-MS using sequential derivatization and extraction.

Selenium speciation analysis is usually carried out using complex hyphenated analytical systems such as LC-ICP-MS. Here we present a novel selenium speciation approach based on a sequential derivatization and extraction combined with gas chromatography mass spectrometry for the simultaneous determination of selenite, selenate, and selenocyanate in aqueous mine wastewater samples. Selenocyanate was derivatized with triethyloxonium tetrafluoroborate to ethylselenocyanate, which was extracted into chloroform, following which the sample was split into two aliquots. One aliquot was acidified and 3,5-bis(trifluoromethyl)-o-phenylenediamine was used for the novel derivatization of selenite to 4,6-bis(trifluoromethyl)-2,1,3-benzoselenadiazole, for the determination of selenite. For the second aliquot, concentrated hydrochloric acid was added along with 4-nitro-o-phenylenediamine to simultaneously reduce selenate to selenite and derivatize the combined "selenite + selenate" fraction to 5-nitro-2,1,3-benzoselenadiazole. The benzoselenadiazoles were extracted with chloroform and all extracts were combined for GC-MS analysis. Low ng g-1 detection limits were reported for all three species. The method is unhindered by concentrations of chloride and sulphate up to 3%, as well as nitrate concentrations up to 3% for selenocyanate and selenite analysis, with minor losses in sensitivity for selenate up to 100 ppm nitrate, making the method particularly suitable for aqueous mine waste characterization. Quantitative trace selenium speciation was achieved using cost-effective materials and apparatus on a simple-to-operate benchtop instrument. The novel methodology was tested on gold mine wastewater samples; comparing to total selenium, a 63-149% recovery as the sum of species was observed. Additionally, this novel speciation approach was compared to LC-ICP-MS based selenium speciation and a reasonable agreement was found in the species distribution.

Gas chromatography-mass spectrometric characterization of SenSm compounds in synthetic melts and sediments.

A gas chromatography-mass spectrometry (GC-MS) method was used to characterize and identify mixed selenium-sulfur compounds with the formula SenSm in synthetic selenium-sulfur melts and selenium-impacted sediments. This method shows strong fragmentation of SenSm compounds in the electron ionization source, which makes conclusive identification of discrete compounds difficult. Despite these limitations, three SenSm compounds (SeS4, SeS5, and SeS6) were identified conclusively in synthetic samples, and several others were shown to be mixed SeS compounds, without being able to determine a molecular formula. Three selenium-impacted sediments from sites containing similar total selenium concentrations (11.0-18.5 ppm) were extracted with carbon disulfide (CS2); each of the three sediments contained at least 12 SenSm compounds, ten of which were common to all three sediments. Two of the compounds identified in the synthetic samples were also found in all three sediments, SeS4 and SeS5, and a third, SeS6 was found in one sediment sample. This is the first time that cyclic SenSm compounds have been confirmed to exist in the environment.

Determination of Se8 in sediments by gas chromatography-mass spectrometry.

A gas chromatography-mass spectrometry (GC-MS) method was developed to analyze cyclooctaselenium (Se8) in sediments after extraction with carbon disulfide (CS2). The method suffers from some analytical complications, most notably a poor peak shape for Se8, manifesting in a broad peak (2 min) with strong fronting, and a decrease in Se8 retention time with decreasing concentration. Detailed analysis of the mass spectral data suggests that (thermal) decomposition of Se8 on the GC column is responsible for both of these phenomena. Despite these limitations, GC-MS with selected ion monitoring (SIM) yielded a sufficiently low detection limit, 10 µg/kg (dw), to analyze Se8 in selenium-impacted sediments. The CS2 extraction appears to be selective enough to allow the determination of Se8 in sediments without further cleanup when GC-SIM-MS is used for analysis. One selenium-impacted sediment contained 7.1 mg/kg (dw) Se8, corresponding to 41% of its total selenium and as much as 50% of its total elemental selenium. This is the first time that Se8 has been positively identified in the environment. Preliminary results also indicated the presence of mixed Se-S rings in this sediment.

Rates and processes affecting As speciation and mobility in lake sediments during aging.

Sediments from an arsenic (As) contaminated groundwater vent site were used to investigate As(III) binding, transformation and redistribution in native and iron oxide amended lake sediments using aging spiked batch reactions and a sequential extraction procedure that maintains As(V) and As(III) speciation. In the native sediments, fractionation analysis revealed that 10% of the spiked As(III) remained intact after a 32-day aging experiment and was predominantly adsorbed to the strongly sorbed (NH4H2PO4 extractable) and amorphous Fe oxide bound (H3PO4 extractable) fractions. Kinetic modelling of the experimental results allowed identifying the dominant reaction path for depletion of dissolved As(III) to As(III) absorbed on to the solid phase, followed by oxidation in the solid phase. Arsenite was initially adsorbed primarily to the easily exchangeable fraction ((NH4)2SO4 extractable), then rapidly transformed into As(V) and redistributed to the strongly sorbed and amorphous Fe oxide bound fractions. Oxidation of As(III) in recalcitrant fractions was less efficient. The iron oxide amendments illustrated the controls that iron oxides can have on As(III) binding and transformation rates. In goethite amended samples As(III) oxidation was faster and primarily occurred in the strongly sorbed and amorphous Fe oxide bound fractions. In these samples, 19.3µg Mn was redistributed (compared to the native sediment) from the easily exchangeable and crystalline Fe oxide bound fractions to the strongly sorbed and amorphous Fe oxide bound fractions, indicating that goethite may act as a catalyst for Mn(II) oxidation, thereby producing sorbed Mn(III/IV), which then appears to be involved in rapidly oxidizing As(III).

Hydrological and biogeochemical controls governing the speciation and accumulation of selenium in a wetland influenced by mine drainage.

Controls governing the speciation and accumulation of Se in a 3.7-ha marsh influenced by mine drainage were assessed through examination of water balance, water quality, sediment, and plant tissue components. Over the 8-mo study period (April through November, 2009), mean monthly flows ranged from 1600 to 2300 m3 d-1 (hydraulic retention time of 1-3 d). Total Se concentrations in the marsh outflow were lower than the inflow by 0.4 to 6.2 µg L-1 (mean difference = 3.3 µg L-1 ), illustrating Se removal. The Se accumulation pathways are illustrated by elevated concentrations of Se in sediments (3-35 mg kg-1 dry wt) as well as in below-ground (2-41 mg kg-1 dry wt; mean = 10 mg kg-1 dry wt) and above-ground (0.8-6.3 mg kg-1 dry wt; mean = 2 mg kg-1 dry wt) emergent plant tissues. Redox stratification in the shallow water column had a marked effect on Se speciation and behavior, illustrating bottom water removal of dissolved selenate in suboxic horizons and increased mobility of dissolved organo-Se. Mass balance data yielded inflow and outflow loading rates for Se of 27 and 23 g d-1 , respectively (net accumulation rate of 4 g d-1 or 0.11 mg m2 d-1 ). The rate of accumulation as calculated from the mass balance agrees with independently measured rates of Se accumulation in sediments for the site (3.6-8.1 g d-1 or 0.10-0.22 mg m-2 d-1 ). Environ Toxicol Chem 2018;37:1824-1838. © 2018 SETAC.

The role of phosphorus in the metabolism of arsenate by a freshwater green alga, Chlorella vulgaris.

A freshwater microalga, Chlorella vulgaris, was grown in the presence of varying phosphate concentrations (<10-500µg/L P) and environmentally realistic concentrations of arsenate (As(V)) (5-50µg/L As). Arsenic speciation in the culture medium and total cellular arsenic were measured using AEC-ICP-MS and ICP-DRC-MS, respectively, to determine arsenic biotransformation and uptake in the various phosphorus scenarios. At high phosphate concentration in the culture medium, >100µg/L P, the uptake and biotransformation of As(V) was minimal and dimethylarsonate (DMAs(V)) was the dominant metabolite excreted by C. vulgaris, albeit at relatively low concentrations. At common environmental P concentrations, 0-50µg/L P, the uptake and biotransformation of As(V) increased. At these higher As-uptake levels, arsenite (As(III)) was the predominant metabolite excreted from the cell. The concentrations of As(III) in these low P conditions were much higher than the concentrations of methylated arsenicals observed at the various P concentrations studied. The switchover threshold between the (small) methylation and (large) reduction of As(V) occurred around a cellular As concentration of 1fg/cell. The observed nearly quantitative conversion of As(V) to As(III) under low phosphate conditions indicates the importance of As(V) bio-reduction at common freshwater P concentrations. These findings on the influence of phosphorus on arsenic uptake, accumulation and excretion are discussed in relation to previously published research. The impact that the two scenarios of As(V) metabolism, As(III) excretion at high As(V)-uptake and methylarsenical excretion at low As(V)-uptake, have on freshwater arsenic speciation is discussed.

Characterization of colloidal arsenic at two abandoned gold mine sites in Nova Scotia, Canada, using asymmetric flow-field flow fractionation-inductively coupled plasma mass spectrometry.

Asymmetric flow-field flow fractionation-inductively-coupled plasma-mass spectrometry was used to determine whether colloidal arsenic (As) exists in soil pore water and soil extract samples at two arsenic-contaminated abandoned gold mines (Montague and Goldenville, Nova Scotia). Colloidal arsenic was found in 12 out of the 80 collected samples (=15%), and was primarily associated with iron (Fe) in the encountered colloids. The molar Fe/As ratios indicate that the colloids in some samples appeared to be discrete iron-arsenic minerals, whereas in other samples, they were more consistent with As-rich iron (oxy)hydroxides. Up to three discrete size fractions of colloidal As were encountered in the samples, with mean colloid diameters between 6 and 14nm. The pore water samples only contained one size fraction of As-bearing colloids (around 6nm diameter), while larger As-bearing colloids were only encountered in soil extracts.

Validation of an updated fractionation and indirect speciation procedure for inorganic arsenic in oxic and suboxic soils and sediments.

A sequential extraction procedure (SEP) for the speciation analysis of As(III) and As(V) in oxic and suboxic soils and sediments was validated using a natural lake sediment and three certified reference materials, as well as spike recoveries of As(III) and As(V). Many of the extraction steps have been previously validated making the procedure useful for comparisons to similar previous SEP studies. The novel aspect of this research is the validation for the SEP to maintain As(III) and As(V) species. The proposed five step extraction procedure includes the extraction agents (NH4)2SO4, NH4H2PO4, H3PO4 + NH2OH·HCl, oxalate + ascorbic acid (heated), and HNO3 + HCl + HF, targeting operationally defined easily exchangeable, strongly sorbed, amorphous Fe oxide bound, crystalline Fe oxide bound, and residual As fractions, respectively. The third extraction step, H3PO4 + NH2OH·HCl, has not been previously validated for fraction selectivity. We present evidence for this extraction step to target As complexed with amorphous Fe oxides when used in the SEP proposed here. All solutions were analyzed on ICP-MS. The greatest concentrations of As were extracted from the amorphous Fe oxide fraction and the dominant species was As(V). Lake sediment materials were found to have higher As(III) concentrations than the soil materials. Because different soils/sediments have different chemical characteristics, maintenance of As species during extractions must be validated for specific soil/sediment types using spiking experiments.

Production and Release of Selenomethionine and Related Organic Selenium Species by Microorganisms in Natural and Industrial Waters.

Laboratory algal cultures exposed to selenate were shown to produce and release selenomethionine, selenomethionine oxide, and several other organic selenium metabolites. Released discrete organic selenium species accounted for 1.6-13.1% of the selenium remaining in the media after culture death, with 1.3-6.1% of the added selenate recovered as organic metabolites. Analysis of water from an industrially impacted river collected immediately after the death of massive annual algal blooms showed that no selenomethionine or selenomethionine oxide was present. However, other discrete organic selenium species, including a cyclic oxidation product of selenomethionine, were observed, indicating the previous presence of selenomethionine. Industrial biological treatment systems designed for remediation of selenium-contaminated waters were shown to increase both the concentration of organic selenium species in the effluent, relative to influent water, and the fraction of organic selenium to up to 8.7% of the total selenium in the effluent, from less than 1.1% in the influent. Production and emission of selenomethionine, selenomethionine oxide, and other discrete organic selenium species were observed. These findings are discussed in the context of potentially increased selenium bioavailability caused by microbial activity in aquatic environments and biological treatment systems, despite overall reductions in total selenium concentration.

Multi-dimensional water quality assessment of an urban drinking water source elucidated by high resolution underwater towed vehicle mapping.

Spatial surveys of Ramsey Lake, Sudbury, Ontario water quality were conducted using an innovative underwater towed vehicle (UTV) equipped with a multi-parameter probe providing real-time water quality data. The UTV revealed underwater vent sites through high resolution monitoring of different spatial chemical characteristics using common sensors (turbidity, chloride, dissolved oxygen, and oxidation/reduction sensors) that would not be feasible with traditional water sampling methods. Multi-parameter probe vent site identification is supported by elevated alkalinity and silica concentrations at these sites. The identified groundwater vent sites appear to be controlled by bedrock fractures that transport water from different sources with different contaminants of concern. Elevated contaminants, such as, arsenic and nickel and/or nutrient concentrations are evident at the vent sites, illustrating the potential of these sources to degrade water quality.

Identification of trace levels of selenomethionine and related organic selenium species in high-ionic-strength waters.

A new anion-exchange chromatographic separation method was used for the simultaneous speciation analysis of selenoamino acids and the more ubiquitous inorganic selenium oxyanions, selenite and selenate. For quantification, this separation was coupled to inductively coupled plasma-mass spectrometry to achieve an instrumental detection limit of 5 ng Se L(-1) for all species. This chromatographic method was also coupled to electrospray tandem mass spectrometry to observe the negative ion mode fragmentation of selenomethionine and one of its oxidation products. Low detection limits were achieved, which were similar to those obtained using inductively coupled plasma-mass spectrometry. An extensive preconcentration and cleanup procedure using cation-exchange solid-phase extraction was developed for the identification and quantification of trace levels of selenomethionine in environmental samples. Preconcentration factors of up to five were observed for selenomethionine, which in addition to the removal of high concentrations of sulphate and chloride from industrial process waters, allowed for an unambiguous analysis that would have been impossible otherwise. Following these methods, selenomethionine was identified at an original concentration of 3.2 ng Se L(-1) in samples of effluent collected at a coal-fired power plant's biological remediation site. It is the first time that this species has been identified in the environment, outside of a biological entity. Additionally, oxidation products of selenomethionine were identified in river water and laboratory algal culture samples. High-resolution mass spectrometry was employed to postulate the chemical structures of these species.

Production and release of selenocyanate by different green freshwater algae in environmental and laboratory samples.

In a previous study, selenocyanate was tentatively identified as a biotransformation product when green algae were exposed to environmentally relevant concentrations of selenate. In this follow-up study, we confirm conclusively the presence of selenocyanate in Chlorella vulgaris culture medium by electrospray mass spectrometry, based on selenium's known isotopic pattern. We also demonstrate that the observed phenomenon extends to other green algae (Chlorella kesslerii and Scenedesmus obliquus) and at least one species of blue-green algae (Synechococcus leopoliensis). Further laboratory experiments show that selenocyanate production by algae is enhanced by addition of nitrate, which appears to serve as a source of cyanide produced in the algae. Ultimately, this biotransformation process was confirmed in field experiments where trace amounts of selenocyanate (0.215 ± 0.010 ppb) were observed in a eutrophic, selenium-impacted river with massive algal blooms, which consisted of filamentous green algae (Cladophora genus) and blue-green algae (Anabaena genus). Selenocyanate abundance was low despite elevated selenium concentrations, apparently due to suppression of selenate uptake by sulfate, and insufficient nitrogen concentrations. Finally, trace levels of several other unidentified selenium-containing compounds were observed in these river water samples; preliminary suggestions for their identities include thioselenate and small organic Se species.

Identification and determination of selenosulfate and selenocyanate in flue gas desulfurization waters.

In this work, 13 selenium species in flue gas desulfurization (FGD) waters from coal-fired power plants were separated and quantified using anion-exchange chromatography coupled to inductively coupled plasma mass spectrometry. For the first time, we identified both selenosulfate (SeSO(3)(2-)) and selenocyanate (SeCN(-)) in such waters, using retention time matching and confirmation by electrospray mass spectrometry. Besides selenite and selenate, selenosulfate was the most frequently occurring selenium species. It occurred in most samples and constituted a major fraction (up to 63%) of the total selenium concentration in waters obtained from plants employing inhibited oxidation scrubbers. Selenocyanate occurred in about half of the tested samples, but was only a minor species (up to 6% of the total selenium concentration). Nine additional Se-containing compounds were found in FGD waters, but they remain unidentified at this point.

Biogeochemical mechanisms of selenium exchange between water and sediments in two contrasting lentic environments.

The biogeochemical mechanisms of Se exchange between water and sediments in two contrasting lentic environments were assessed through examination of Se speciation in the water column, porewater, and sediment. High-resolution (7 mm) vertical profiles of <0.45 µm Se species across the sediment-water interface demonstrate that the behavior of dissolved Se(VI), Se(IV), and organo-Se are closely linked to redox conditions as revealed by porewater profiles of redox-sensitive species (dissolved O2, NO3-, Fe, Mn, SO4(2-), and ΣH2S). At both sites Se(VI) is removed from solution in suboxic near-surface porewaters demonstrating that the sediments are serving as diffusive sinks for Se. X-ray absorption near edge spectroscopy (XANES) of sediments suggests that elemental Se and organo-Se represent the dominant sedimentary sinks for dissolved Se. Dissolved Se(IV) and organo-Se are released to porewaters in the near-surface sediments resulting in the diffusive transport of these species into the water column, where between-site differences in the depths of release can be linked to differences in redox zonation. The presence or absence of emergent vegetation is proposed to present a dominant control on sedimentary redox conditions as well as on the recycling and persistence of reduced Se species in bottom waters.

Stabilization of thioarsenates in iron-rich waters.

In recent years, thioarsenates have been shown to be important arsenic species in sulfidic, low-iron waters. Here, we show for the first time that thioarsenates also occur in iron-rich ground waters, and that all methods previously used to preserve arsenic speciation (acidification, flash-freezing, or EDTA addition) fail to preserve thioarsenates in such matrices. Laboratory studies were conducted to identify the best approach for stabilizing thioarsenates by combination and modification of the previously-applied methods. Since acidification was shown to induce conversions between thioarsenates and precipitation of arsenic-sulfide minerals, we first conducted a detailed study of thioarsenate preservation by flash-freezing. In pure water, thioarsenates were stable for 21d when the samples were flash-frozen and cryo-stored with a minimal and anoxic headspace. Increasing headspace volume and oxygen presence in the headspace were detrimental to thioarsenate stability during cryo-storage. Addition of NaOH (0.1M) or EtOH (1% V/V) counteracted these effects and stabilized thioarsenates during cryo-storage. Addition of Fe(II) to thioarsenate solutions caused immediate changes in arsenic speciation and a loss of total arsenic from solution during cryo-storage. Both effects were largely eliminated by addition of a neutral EDTA-solution, and thioarsenates were significantly stabilized during cryo-storage by this procedure. Neutralization of EDTA was required to prevent alteration of thioarsenate speciation through pH change. With the modified method (anoxic cryo-preservation by flash-freezing with minimal headspace after addition of neutralized EDTA-solution), the fractions of mono- and dithioarsenate, the two thioarsenates observed in the iron-rich ground waters, remained stable over a cryo-storage period of 11d. Further modifications are needed for the higher SH-substituted thioarsenates (tri- and tetrathioarsenate), which were not encountered in the studied iron-rich ground waters.

Release of reduced inorganic selenium species into waters by the green fresh water algae Chlorella vulgaris.

The common green fresh water algae Chlorella vulgaris was exposed to starting concentrations of 10 µg/L selenium in the form of selenate, selenite, or selenocyanate (SeCN(-)) for nine days in 10% Bold's basal medium. Uptake of selenate was more pronounced than that of selenite, and there was very little uptake of selenocyanate. Upon uptake of selenate, significant quantities of selenite and selenocyanate were produced by the algae and released back into the growth medium; no selenocyanate was released after selenite uptake. Release of the reduced metabolites after selenate exposure appeared to coincide with increasing esterase activity in solution, indicating that cell death (lysis) was the primary emission pathway. This is the first observation of biotic formation of selenocyanate and its release into waters from a nonindustrial source. The potential environmental implications of this laboratory observation are discussed with respect to the fate of selenium in impacted aquatic systems, the ecotoxicology of selenium bioaccumulation, and the interpretation of environmental selenium speciation data generated, using methods incapable of positively identifying reduced inorganic selenium species, such as selenocyanate.

Arsenic speciation in sulfidic waters: reconciling contradictory spectroscopic and chromatographic evidence.

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 OH(-) and 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.