| dc.description.abstract | Selenium is well named after the moon, because statements about this element must often be qualified, making it feel ever changing; for instance, Se is a nutrient, but an excess of 400 µg per day is toxic for humans. High Se concentrations can cause reproductive decrease or complete failure in fish, aquatic birds, amphibians, and reptiles. These animals can also bioconcentrate Se, so high aqueous Se concentrations are not required to lead to toxic consequences. The problem of Se is not limited to wetlands; plants that uptake Se can cause selenosis to animals that forage on land, such as sheep and cattle. Remediating Se before it can reach a receptor is important in preventing the loss of the next generation of wildlife.
When treating Se in ground water, or measuring Se concentrations at contaminated sites, it can be difficult to determine where the Se is going unless extensive measurements are made, including solid sampling and speciation measurements. Non-traditional stable isotopes are an emerging tool in the remediation of groundwater contamination caused by anthropogenic activities. Taking Se stable isotope measurements in conjunction with information about removal mechanisms could produce a powerful predictive system to solve remediation problems.
A combination of theoretical calculations, laboratory experiments, and field measurements were used to evaluate Se stable isotopes as a remediation tool. Molecules for SeO4^2-, SeO3^2-, HSeO3^-, CaSeO4, and CaSeO3 were all modeled using Gaussian 09 (Frisch et al. 2009). The vibrational energies from these models were then used to calculate the equilibrium fractionation between each pair of molecules. These equilibrium fractionation factors were found to be within 0.21 ‰ of other values from the literature, where available. Calculations were of the same magnitude as laboratory studies: 13.4 ‰ for reduction of SeO4^2- to SeO3^2-, and 0.8 ‰ for SeO3^2- to HSeO3^-, which is similar to the range for adsorption of 0 – 1.24 ‰. These theoretical values can be used to establish baseline values when no data is available from laboratory experiments, as is the case for the formation of CaSeO4 and CaSeO3. Calcium plays an important role in the sequestration of Se at pH above 7 (Goldberg and Glaubig 1988), making it relevant to the Se isotope literature.
Three separate laboratory experiments were conducted. Abiotic reduction of Se(IV) by Na2S(aq) was investigated to determine whether isotopic fractionation can differentiate reduction by H2S(g) and direct respiration or enzymatic reduction due to sulfur reducing bacteria (SRB) in the environment. A solid precipitate formed rapidly and was collected for PXRD analysis. The precipitate was either yellow, orange, or red depending on the starting pH and the Se:S ratio in solution. All three precipitate colors had different powder X-ray diffraction (PXRD) ring patterns. The yellow precipitate had no pattern, and may have been amorphous S(0), the orange precipitates were most similar to selenium sulfide, and the red precipitate was most similar to a mixture of Se(0) and S(0). The fractionation factor when samples were filtered at 3-4 hours was 7.9 ‰. When the solid was left in contact with solution for a longer duration, the fractionation factor increased to 10.9 ‰.
An experiment on biotic reduction of Se(IV) by a natural SRB consortium, including a comparison of results to the work of others, provides contrasting data for the abiotic experiment. Batch vessels were loaded with a mixture of lucerne hay (alfalfa), silica sand, and a small amount of zero valent iron (ZVI) before being transferred to an anaerobic chamber. A solution containing MgSO4, Na-lactate, and SRB culture was then added to each vessel, and they were crimp sealed. Samples were taken along a time series. Scanning electron microscope (SEM) images show microbes colonizing the sand crevices and much of the organic matter, but no Se precipitates are obvious. Initial fractionation in the reduction of Se(IV) was positive (10‰ < ε ≤ 19 ‰), followed by a decrease in δ82Se in solution. Because a negative fractionation factor is unlikely, it is probable that multiple pools of reduced and organic Se species are entering solution, causing a final δ82Se of -13.2 ‰ after three days, with the lowest δ82Se of -16.7 ‰ seen at two days.
The final experiment used ZVI to reduce Se(VI) in a flow through cell system, while simultaneously collecting XANES data and isotope samples. A column with a transparent window packed with ZVI was placed in the hutch at a synchrotron (Sector 13, Advanced Photon Source (APS), Argonne, IL, USA), and a Na2SeO4, CaCO3 solution was pumped through. The Na2SeO4 concentration was increased at 8 hour increments, and removed from influent solution at the end of the experiment to observe the effect of rinsing the column with a CaCO3 only solution. The linear combination fit (LCF) shows progressively reduced species of Se accumulating on the solid over time, with more Se present near the input of the cell than the output. The Se(VI) component decreased rapidly in the rinse phase of the LCF, suggesting most Se in solution was Se(VI). The δ82Se could be fit with a straight line, yielding an isotopic discrimination of 9.6 ‰. The δ82Se of the rinse solution could be fit with a Rayleigh type curve, with a fractionation factor of 2.4 ‰. This fractionation factor is between adsorption of Se onto iron minerals, and reduction of Se by ZVI in the presence of CaCO3, as measured in an earlier batch experiment. Simultaneously obtaining isotope and solid phase data helps link removal mechanisms to fractionation factors. Isotope results from all three laboratory experiments suggest reductive processes are of the same magnitude as theoretical calculations, and Se reduction experiments conducted by others.
Samples for isotope and cation analysis were collected from along the length of an Se-bearing groundwater plume. Laboratory obtained Se isotope fractionation factors were then used to model the processes occurring in the plume. Geochemical and redox data were used to support the isotope modeling results. These modeling results, supported by redox data, allowed us to infer the processes occurring in the subsurface. The main processes within the plume include adsorption and dispersive dilution, as indicated by only small changes in δ82Se values (δ82Se = 1.5 – 3.4 ‰) and a large decrease in Se concentration (from 9770 to 774 µg L-1). Reduction is occurring within the source area (δ82Se = 1.3 – 3.8 ‰), and is the cause of the sharp increase in δ82Se (8.7 ‰) under the wetland complex. Very low δ82Se values behind the source area (-26 ‰) and at the distal end of the plume (-16 ‰) are likely due to the oxidation of low δ82Se from local shales. Low concentration (< 900 µg L-1), low δ82Se values (-0.2 – 0.9 ‰) at the plume’s edge are either the result of desorption/oxidation of previously reduced plume Se, or mixing of plume Se with background Se with a low δ82Se. The results from this field study demonstrate the potential use for Se stable isotope measurements in environmental samples. | en |
