Résumé
A field-pilot bioreactor exploiting microbial iron (Fe) oxidation and subsequent arsenic (As) and Fe co-precipitation was monitored during 6 months for the passive treatment of As-rich acid mine drainage (AMD). It was implemented at the Carnoulès mining site (southern France) where AMD contained 790–1315 mg L−1 Fe(II) and 84–152 mg L−1 As, mainly as As(III) (78–83%). The bioreactor consisted in five shallow trays of 1.5 m2 in series, continuously fed with AMD by natural flow. We monitored the flow rate and the water physico-chemistry including redox Fe and As speciation. Hydraulic retention time (HRT) was calculated and the precipitates formed inside the bioreactor were characterized (mineralogy, Fe and As content, As redox state). Since As(III) oxidation improves As retention onto Fe minerals, bacteria with the capacity to oxidize As(III) were quantified through their marker gene aioA. Arsenic removal yields in the pilot ranged between 3% and 97% (average rate (1.8 ± 0.8) ✕ 10−8 mol L−1 s−1), and were positively correlated to HRT and inlet water dissolved oxygen concentration. Fe removal yields did not exceed 11% (average rate (7 ± 5) ✕ 10−8 mol L−1 s−1). In the first 32 days the precipitate contained tooeleite, a rare arsenite ferric sulfate mineral. Then, it evolved toward an amorphous ferric arsenate phase. The As/Fe molar ratio and As(V) to total As proportion increased from 0.29 to 0.86 and from ∼20% to 99%, respectively. The number of bacterial aioA gene copies increased ten-fold during the first 48 days and stabilized thereafter. These results and the monitoring of arsenic speciation in the inlet and the outlet water, provide evidences that As(III) oxidized in the pilot. The biotreatment system we designed proved to be suitable for high As DMA. The formation of sludge highly enriched into As(V) rather than As(III) is advantageous in the perspective of long term storage.