Microbial-induced desaturation and precipitation in stratified soils with fine sand and silt layers

A tank test was performed simulating two-dimensional planar flow conditions at a meter scale to evaluate the effectiveness of microbial-induced desaturation and precipitation (MIDP) in stratified soil conditions. The tank setup (116.5 cm tall, 122 cm wide, and 5.25 cm thick) was filled with two layers of fine sand (a target layer of 40 cm and nontarget layer of 21 cm above) that were confined by silt layers above (6 cm), between (9 cm) and below (9 cm) the sand layers. Multiple flushes of substrate solution, containing calcium, nitrate, and acetate, were injected into the lower sand layer to stimulate indigenous nitrate-reducing bacteria to produce biogenic gas, biominerals, and biomass. Embedded sensors were used to measure the changes in electrical conductivity, volumetric water content, and pore pressure in both the target and nontarget sand layers during and between treatment cycles. Time-lapse camera images were used to determine flow velocity distributions in the target layer and identify modes of gas migration. At the end of the test, hydraulic conductivity, calcium carbonate content, and soil–water characteristic curves (SWCCs) were measured on intact samples of the treated material. The results showed that most of the reaction products were formed in the targeted sand layer. During the first treatment cycle, the degree of saturation in the target sand layer decreased to 75% within 5–12 days, at which point it started to migrate upwards until it got trapped and formed a lens underneath the silt layer above. During the second and subsequent treatment cycles, seepage velocity increased due to the entrapment of biogenic gas, the reaction rate increased due to the accumulation of biomass, and the gas formed channels through the silt and migrated further upwards into and through the upper sand and silt layers by irregular venting events. After 4–5 cycles, an equilibrium condition was reached at which the degree of saturation fluctuated from 65 to 80% when gas was being produced and vented to 80–90% when substrates were depleted. The CaCO₃ content after 10 cycles over 12 weeks ranged from 1.6% close to the inlet to 0.5% close to the outlet, with an average of 0.68%. The formation of biomass and CaCO₃ had a relatively large impact on the saturated hydraulic conductivity but a very limited impact on the SWCC.