Evaluation of Surface and Transport Limitations to the Rate of Calcite Dissolution Using Pore Scale Modeling of a Capillary Tube Experiment at pCO2 4 bar

Mineral trapping is generally considered to account for most of the long-term trapping of CO2 in the subsurface. Prediction of mineral trapping at the reservoir scale requires knowledge of continuum-scale mineral dissolution and precipitation rates. However, processes that take place at the pore scale (e.g., transport limitation to reactive surfaces) affect rates applicable at the continuum scale. To explore the pore scale processes that result in the discrepancy between rates measured in laboratory experiments and those calibrated from continuum-scale models, we have developed a high-resolution pore scale model of a capillary tube experiment. The capillary tube (L=0.7-cm, D=500-μm) is packed with crushed calcite (Iceland spar) and the resulting 3D pore structure is imaged by X-ray computed microtomography (XCMT) at Berkeley Lab's Advanced Light Source at a 0.899-μm resolution. A solution in equilibrium with a partial pressure of CO2 of 4 bars is injected at a rate of 5 microliter/min and the effluent concentrations of calcium are measured to ensure steady state conditions are achieved. A simulation domain is constructed from the XCMT image using implicit functions to represent the mineral surface locally on a grid. The pore-scale reactive transport model is comprised of high performance simulation tools and algorithms for incompressible Navier-Stokes flow, advective-diffusive transport and multicomponent geochemical reactions. Simulations are performed using 6,144 processors on NERSC's Cray XE6 Hopper to achieve a grid resolution of 2.32 μm. Equivalent continuum scale simulations are also performed to evaluate the effect of pore scale processes. Comparison of results is performed based on flux-averaged effluent calcium concentrations, which are used as indicator of effective rates in the capillary tube. Results from both pore- and continuum-scale simulations overestimate the calcium effluent concentrations, suggesting that the TST rate expression parameters (fitted to experiments by Pokrovsky et al. 2005, Chem. Geol.) may not be appropriate for the Iceland spar used in the capillary experiment. Pore scale simulation effluent concentrations are lower than those obtained in the continuum-scale model. This difference is caused by mass transport limitations to reactive surfaces located in small pore spaces, where fluid velocities are significantly slower. Different reactive transport regimes are observed at the same slice of the capillary: far-from-equilibrium conditions in large pore spaces where dissolution is mostly surface-limited and near-equilibrium conditions in almost stagnant pore spaces where dissolution is transport-limited. Where different reactive transport regimes exist, the assumption of well-mixed conditions at each grid point of the continuum-scale model is not met.