Manganese (Mn) biomineralization has promising applications for remediation of soils and groundwater contaminated by mine drainage. This microbially mediated process involves the enzymatic oxidation of aqueous Mn(II) to solid-phase Mn(III, IV), effectively immobilizing Mn and producing reactive minerals capable of sorbing co-occurring toxicant metals (e.g., lead, cadmium, nickel). While Mn biomineralization been characterized in simple batch systems and column bioreactors, reaction behavior in complex natural environments is poorly understood. The physical heterogeneity of porous media, such as soils and sediments, promotes nonuniformity of fluid flow and formation of reactant gradients. We hypothesize that these heterogeneities decrease both the rate and extent of Mn biomineralization relative to well-mixed batch conditions. We integrate novel “soil-on-a-chip” microfluidic reactors with brightfield microscopy to directly visualize and quantify Mn biomineralization by the model Mn-oxidizing bacterium Pseudomonas putida GB-1 in a simulated soil pore structure. Understanding the pore-scale hydrodynamic and biogeochemical controls on the growth of Mn-oxidizing bacteria, the rate of Mn biomineralization, and the spatial distribution of Mn oxides will allow us to enhance contaminant removal efficiency when upscaling.