Considerations for conducting incubations to study the mechanisms of As release in reducing groundwater aquifers

Microbial Fe reduction is widely believed to be the primary mechanism of As release from aquifer sands in Bangladesh, but alternative explanations have been proposed. Long-term incubation studies using natural aquifer material are one way to address such divergent views. This study addresses two issues related to this approach: (1) the need for suitable abiotic controls and (2) the spatial variability of the composition of aquifer sands. Four sterilization techniques were examined using orange-colored Pleistocene sediment from Bangladesh and artificial groundwater over 8 months. Acetate (10 mM) was added to sacrificial vials before sterilization using either (1) 25 kGy of gamma irradiation, (2) three 1-h autoclave cycles, (3) a single addition of an antibiotic mixture at 1x or (4) 10x the typical dose, and (5) a 10 mM addition of azide. The effectiveness of sterilization was evaluated using two indicators of microbial Fe reduction, changes in diffuse spectral reflectance and leachable Fe(II)/Fe ratios, as well as changes in P-extractable As concentrations in the solid phase. A low dose of antibiotics was ineffective after 70 days, whereas autoclaving significantly altered groundwater composition. Gamma irradiation, a high dose of antibiotics, and azide were effective for the duration of the experiment. Using gamma irradiation as an abiotic control, shallow grey sediment and groundwater from 3 closely spaced locations along a gradient of dissolved As concentrations (60-130-210 μg/L) in Bangladesh were incubated for 8 months with and without organic C addition (0.9 and 0.6 mM of acetate and lactate). Unexpectedly, levels of dissolved As (64 ± 68, 92 ± 70, 217 ± 68 μg/L) and P-extractable As (0.7 ± 0.2, 2.1 ± 0.5 and 2.0 ± 0.3 mg/kg) at each location were highly variable over the duration of the experiment and prevented the detection of the relatively small levels of As release that were anticipated. Maintenance of an adsorptive equilibrium with the P-extractable As concentrations seems to govern dissolved As variability. The sediment variability is attributed to natural patchiness in the distribution of aquifer properties rather than a sampling artifact. Sub-sampling a single batch of groundwater and aquifer solids over time can alleviate this problem to some extent, but the issue of the representativeness of particular samples remains