Arsenic speciation in natural sulfidic geothermal waters

Abstract

The speciation of arsenic in natural sulfidic geothermal waters was studied using chemical analyses and thermodynamic aqueous speciation calculations. Samples were collected in three geothermal systems in Iceland, having contrasting H2S concentrations in the reservoir (high vs. low). The sampled waters contained 7–116ppb As and <0.01–77.6ppm H2S with pH of 8.56–9.60. The analytical setup used for the determination of arsenic species (Ion Chromatography-Hydride Generation Atomic Fluorescence Spectrometry, IC-HG-AFS) was field-deployed and the samples analyzed within ∼5min of sampling in order to prevent changes upon storage, which were shown to be considerable regardless of the sample storage method used. Nine aqueous arsenic species were detected, among others arsenite (HnAsIIIO3n-3), thioarsenite (HnAsIIIS3n-3), arsenate (HnAsVO4n-3), monothioarsenate (HnAsVSO3n-3), dithioarsenate (HnAsVS2O2n-3), trithioarsenate (HnAsVS3On-3) and tetrathioarsenate (HnAsVS4n-3). The results of the measured aqueous arsenic speciation in the natural geothermal waters and comparison with thermodynamic calculations reveal that the predominant factors determining the species distribution are sulfide concentration and pH. In alkaline waters with low sulfide concentrations the predominant species are AsIII oxyanions. This can be seen in samples from a liquid-only well, tapping water that is H2S-poor and free of oxygen. At intermediate sulfide concentration AsIII and AsV thio species become important and predominate at high sulfide concentration, as seen in two-phase well waters, which have high H2S concentrations in the reservoir. Upon oxidation, for instance due to mixing of the reservoir fluid with oxygenated water upon ascent to the surface, AsV oxyanions form, as well as AsV thio complexes if the sulfide concentration is intermediate to high. This oxidation process can be seen in samples from hot springs in the Geysir geothermal area. While the thermodynamic modeling allows for a first-order estimation of the dominant species, discrepancies between the model results and the field data highlight the fact that for such dynamic chemical systems the exact speciation cannot be calculated, thus on-site and preferentially in-situ analysis is of crucial importance.