What controls the development of heterogenous dissolution patterns in carbonate rocks?

Porosity and permeability changes are anticipated when carbonate rocks are percolated with and dissolved by acidic fluids. The ability to predict the location, extent, and impact of these changes could benefit acid-relevant operations in carbonate rocks, specifically CO₂ storage by improving our estimates of CO₂ flow and storage performance in the subsurface. In this work, we combine percolation experiments and numerical simulations to capture the chemical effects of CO₂-saturated water (weak acid) and HCl solution (strong acid) on cm-scale limestone cores. Numerical simulations are parameterized and validated against experimental data, including effluent solution chemistry, porosity distribution, and observed dissolution features in CT images of the reacted specimens. CT imaging data of intact cores are employed to construct porosity and permeability distribution maps over the core domain serving as input for reactive transport models of the experiments. The results indicate that the pore space heterogeneity controls the mineral dissolution from the onset of the acidic fluid injections, while the acid type becomes progressively important as the dissolution front further penetrates the rock. The compact dissolution pattern formed in the HCl-treated cores due to the complete dissociation of the strong acid could be numerically simulated using a generalized power-law porosity-permeability relationship with a power value of 3, applied at the numerical grid scale. However, the formation of the lengthwise wormhole in CO₂-treated cores due to partial dissociation of the weak acid and its buffering capacity could be only simulated using a large power value of 15 at the grid scale in the porosity-permeability relationship. This exponent increases to 27.6 for the bulk flow behavior of the limestone core containing the wormhole, illustrating large-scale dependence of acid-induced permeability evolutions in carbonate rocks. These findings highlight the need for developing robust upscaling approaches to account for the hydraulic behavior of reactive, intrinsically heterogeneous carbonate rocks in large-scale simulations.