Search for a command to run...
Understanding the growth and dissolution behavior of calcium carbonate (CaCO 3 ) polymorphs is fundamental for studying biomineralization, environmental geochemistry, and reactive transport processes in porous systems. Among numerous carbonate phases, amorphous calcium carbonate (ACC) often plays the role of a metastable precursor that dissolves, releasing ions needed to form more stable crystalline phases such as calcite and vaterite. A complex dissolution–precipitation process of CaCO 3 polymorphs was observed in dedicated counter-diffusion experiments through silica-gel. To describe the reactive transport phenomena observed in the experimental system, a pore-scale three-dimensional reactive transport model was developed. This approach couples the lattice Boltzmann method, which simulates solute diffusion, with the growth and dissolution kinetics of calcite, vaterite, and ACC. To facilitate high-resolution simulations, a machine learning technique was applied to accelerate geochemical speciation calculations. The reactive transport model provides spatially and temporally resolved saturation indices at the micrometer scale, from which local growth and dissolution rates of the CaCO 3 polymorphs are computed. All key spatial and temporal features observed by in-situ X-ray imaging within the targeted experimental window are well reproduced, providing mechanistic insights into the chemical evolution of the system. In particular, the model reveals how local undersaturation triggers ACC dissolution, releasing Ca 2+ and CO 3 2- ions that sustain and spatially promotes the further growth of the pre-existing calcite and vaterite crystals identified by imaging within the targeted time window. More generally, it shows the tight coupling between mineral stability, ion availability, and growth kinetics, enabling a mechanistic interpretation of the coupled ACC dissolution and calcite/vaterite growth under diffusion-limited, chemically heterogeneous conditions. This approach advances the understanding of CaCO 3 transformations by combining in-situ imaging with detailed mechanistic modelling, and it lays the groundwork for future integration into real-time, simulation-assisted experimental workflows.