Column experiment for assessing microbial behavior around radioactive waste repositories, including migration of potentially radionuclide-accumulating bacteria

• A microbial index to assess radioactive waste repository performance is proposed.
• Distribution coefficient for bacteria on solids was measured in a sand column.
• Sensitivity analysis revealed lower distribution coefficient would be critical.

To assess microbial behavior at anticipated repositories of nitrate-containing radioactive waste such as TRU waste, we set up an anoxic single horizontal column filled with Pleistocene sand with indigenous microorganisms as model samples. The column was supplied with artificial groundwater containing nitrate and acetate for 9 weeks (Run 1) or nitrate-amended groundwater from the same Pleistocene stratum for 6 weeks (Run 2). Bacterial communities, including culturable denitrifiers, were established in the sand bed, resulting in acridine orange direct counts per pore water of 3 × 108 cell mL−1 in Run 1 and 5 × 107 cell mL−1 in Run 2 and nitrate-reducing activity per pore water of roughly 13 mg L−1 d−1 in Run 1 and 1–4 mg L−1 d−1 in Run 2. Eh and hydraulic conductivity declined in Run 1, indicating microbial activity capable of retarding radionuclide transport. However, the ratio of bacterial cell concentration found in the effluent water (free-living bacteria) to the total bacterial concentration in sand (Rmobile) exceeded 2%. This finding is relevant to the increase in radionuclide transport associated with free-living cells. As a tool for quantifying this influence, we introduced an index, Kd,att (distribution coefficient for microbes on sand particles), and calculated this value from the Rmobile value. By sensitivity analysis using a numerical simulation model (MINT), we then demonstrated that higher Kd,att values would suppress the detrimental effects of the free-living bacteria. Quantification of microbial influences can be made more realistic by obtaining Kd,att values in a column experiment and incorporating this index into radionuclide transport models.