Article pubs.acs.org/est
Fractionation of Selenium during Selenate Reduction by Granular Zerovalent Iron Heather K. Shrimpton, David W. Blowes,* and Carol J. Ptacek Department of Earth and Environmental Sciences, University of Waterloo, 200 University Avenue West, Waterloo, Ontario N2L 3G1, Canada S Supporting Information *
ABSTRACT: Batch experiments were conducted using granular zerovalent iron (G-ZVI) with either ultrapure water or CaCO3 saturated simulated groundwater to assess the extent of Se isotope fractionation in solution under the anaerobic conditions characteristic of many aquifers. G-ZVI is a common remediation material in permeable reactive barriers (PRB) to treat Se-contaminated groundwater, and stable isotopes are a potential tool for assessing removal mechanisms. The solution composition, speciation of Se, and Se isotope ratios were determined during both sets of experiments. Dissolved Se concentrations decreased from 10 to 9 when Ca is present.7 In the absence of Ca, the maximum Se(IV) adsorption occurs within a pH range from 3 to 5.50 The adsorption of Se(VI) on Fe minerals is chemically similar to the behavior of sulfate.51 Adsorption of Se(VI) on iron oxides is most extensive at pH < 7.5.51 The removal rate of Se(IV) has also been found to be pH dependent;45 selenite removal in a ultrapure water system was found by others to be more effective at lower pH, with no sorption occurring at pH 8.45 Dissolved Fe concentrations remained below the quantifiable limit of 0.7 mg L−1 for all samples in the CaCO3 experiment. After about 21 h, Fe concentrations were above the instrument detection limit of 0.2 mg L−1. During the ultrapure water experiment, there was an average of 700 ± 160 mg of Fe in solution per kg of G-ZVI used. The ultrapure water control sample, filtered after only brief contact with the G-ZVI, had an Fe concentration higher than any sample at 5.54 mg L−1, and the ultrapure water control sample collected at the end of the experiment had a higher Fe concentration of 12.99 mg L−1 (see the Supporting Information). The presence of Se in the system resulted in a decrease in mass of Fe in solution, suggesting reactions between the two species are occurring. One of the 15 h replicates had an unusually high Fe concentration and a correspondingly low Se concentration compared to samples of a similar duration. A lower Se concentration is not unexpected, as removal in non-Ca anoxic systems is highly dependent on aqueous Fe(II) concentrations.52 The increase in dissolved Fe concentrations following the addition of G-ZVI to water has been observed in previous studies and was not unexpected.12,53 Although the aqueous speciation of the Fe was not measured, it is reasonable to assume that at the pH of the experiments dissolved Fe is predominately Fe(II), with some Fe(III) present on the G-ZVI surface, allowing for the formation of Fe2(SeO3)3.54 Fe(II) can act as an electron donor to further reduce Se.52 Although the presence of O2 at atmospheric levels can retard the reduction of Se(IV) by G-ZVI, low levels of O2 can be beneficial due to the increased release of Fe(II),55 although the general effectiveness is highly dependent on other influences.45,48 Several factors, including high ionic strength and basic conditions, all result in diminished rates of Se(VI) reduction.56 The competition for electrons and the formation of iron oxides
Figure 1. Variation in pH and concentrations of Fe, Se, and Ca in solution. Blue circles represent the experiment including CaCO3; red triangles are data from the experiment with ultrapure water. Dashed lines are a first-order rate removal model fit to the Se concentration data.
decreased at a rate similar to the decreases in Ca concentration in this experiment. The decreases in Ca concentration and alkalinity of the solution are likely due to the precipitation of CaCO3. The ultrapure water batch experiment contained no detectable Ca in solution, and alkalinity remained below quantifiable limits. The IC Se speciation results showed that the stock solution contained negligible Se(IV), and none of the filtered, unacidified samples contained any quantifiable Se(IV). Dissolved Se concentrations declined from 10.33 to 1.54 mg L−1 in samples taken over the 72 h duration of the CaCO3 experiment and from 10.5 mg L−1 to below detection in the ultrapure water experiment (Figure 1). The concentration of Se in G-ZVI-free Se(VI) input controls showed little change over the duration of both experiments (see the Supporting Information). Overall, replicate measurements differed by less than 0.9%; the average difference between duplicate bottles was
