Microbially induced calcite precipitation (MICP) is a promising reinforcement technique that has been applied in soil improvement, CO 2 storage, and enhanced oil recovery. The spatial uniformity of CaCO 3 precipitation largely determines the bonding performance. However, in field applications, environmental and operational conditions often lead to limited reactant concentrations, which shorten the transport distance of reactants and make it difficult for them to distribute evenly within the porous medium. This results in non-uniform precipitation and weak bonding, which limits the engineering performance and restricts the large-scale application of MICP. To achieve uniform bonding under concentration-limited conditions, it is necessary to study the mass transport and precipitation evolution properties and to explore ways to improve precipitation uniformity, such as optimizing the reactant concentration ratio. In this study, we developed a GPU-accelerated pore-scale MICP model that couples fluid flow, reactive mass transport, and precipitation. Using three representative pore structures (homogeneous, heterogeneous, and natural soil), we systematically analyzed how the bacteria-to-Ca 2+ ratio (BCR) affects mass transport processes and the evolution of precipitation patterns under concentration-limited conditions. The results show that when the Ca 2+ concentration is maintained at or above 0.5 M , and the BCR is properly controlled (e.g., kept below 2.0), the spatial uniformity of precipitation can be improved. Moreover, under limited reactant conditions, an injection rate of 4 × 10 −4 m/s provides a balance between total precipitation and uniformity. Compared with staged injection, parallel injection performs better because it avoids early clogging and extends the reaction duration. This study identifies the main mechanisms governing precipitation distribution under concentration-limited conditions in MICP, and provides a numerical tool for optimizing injection strategies to improve precipitation uniformity.
Chu et al. (2026) studied this question.