Rates and multiple sulfur isotope fractionations associated with the oxidation of sulfide by oxygen in aqueous solution

Abstract

Sulfide oxidation is a major component of the global sulfur cycle that requires consideration in isotope-based models of aquatic and sedimentary systems, but the isotope fractionations based on analyses of all three isotope ratios of sulfur (33S/32S, 34S/32S, 36S/32S) have yet to be documented for abiological sulfide oxidation processes. We present experimental determinations of the reaction rates and sulfur isotope fractionations associated with the oxidation of aqueous sulfide (principally HS−) by molecular oxygen (‘autoxidation’) in high pH (≈9.8), low ionic strength carbonate/bicarbonate buffered solutions as a function of temperature (5–45 °C) and trace metal catalysis (ferrous iron, added [Fe2+] ∼50–150 nM). Rates and isotope fractionations are quantified via the analysis of sulfide as a function of reaction progress over relatively low extents of reaction (ca. 33–47%). The oxidation of sulfide at pH = 9.8 and 25 °C without any catalyst added is associated with a computed second order rate constant (k) of lnk = 3.49 ± 0.38 (k in M−1 hr−1; 2 s.d., quadruple experiments) and major sulfur isotope discrimination of 34εP-R = −5.85 ± 0.43‰ (2 s.d., duplicate experiments) that are both consistent with previous studies, and a corresponding minor sulfur isotope fractionation relationship of 33/34θ = 0.509 ± 0.004 that translates to Δ33SP-R = 0.033 ± 0.018‰ (2 s.d.). The dependence of 34εP-R on reaction rate due to either temperature or ferrous iron catalysis over the ranges we have studied is small (<∼1‰ in 34εP-R) and similar values for 33/34θ and Δ33SP-R are obtained for all conditions studied (e.g., mean of all 7 experiments: 33/34θ = 0.5082 ± 0.0031 and Δ33SP-R = 0.037 ± 0.014‰; 2 s.d.). These results indicate that the process of sulfide autoxidation has a mass dependence that is resolvable from the expectations of typical equilibrium isotope exchange. Values for 36/34θ and Δ36SP-R may also exhibit deviations from typical equilibrium isotope exchange but are not resolved under all conditions studied. The shift in values of 33/34θ and Δ33S (and potentially 36/34θ and Δ36S) is consistent with the hypotheses that kinetic isotope effects can be associated with different mass laws than equilibrium processes or with reversibility occurring in the initial parts of the reaction network leading to oxidation products, but both hypotheses will likely require further investigation. We provide an example of how our experimentally calibrated ‘signature’ for sulfide autoxidation may be identified in natural data using the previously published δ34S and Δ33S values of dissolved sulfide in proximity to the oxic-sulfidic interface in the water column of the Cariaco Basin. The observation that the autoxidation of aqueous sulfide in high pH media is associated with a non-zero Δ33SP-R will influence how chemical oxidation processes are treated in environmental and global scale models of the sulfur cycle based on multiple sulfur isotopes.