CO2-brine-mineral Interfacial Reactions Coupled with Fluid Phase Flow

AbstractDue to their widespread occurrence and large capacities, deep geological saline formations are regarded as an important storage option for anthropogenic CO2. Injection of supercritical CO2 into such a formation will result in a multi-phase flow porous media system. Both the CO2 and brine phase compositions are influenced by multiphase flow and mass transport processes as well as by interfacial reactions (gas dissolution, water vaporization, mineral dissolution and precipitation). For a model based assessment of CO2 storage, most simulation codes apply an operator-splitting approach to solve the coupled problem, where multi-phase flow and geochemical reactions are handled by separate routines sequentially. This approach relies on two approximations: (I) the dissolution of CO2 in the brine, which is usually quantified by the multiphase flow routine by using an equation of state approach, is treated as instantaneous, and (II) the amount of CO2 consumed during geochemical reactions quantified by the reaction routine is small compared to the amount dissolved, as during geochemical reactions CO2 is not resupplied from the CO2 phase by dissolution.To investigate these two approximations, the multiphase flow and multi-component reactive transport simulator OpenGeoSys was extended and now allows to simulate mineral-brine as well as the brine-CO2 interface reactions either kinetically controlled or by using an equilibrium approach, and to account for the presence of a CO2 phase during brine-mineral reactions. The code is used here to investigate a simple gas-liquid-solid phase (CO2-H2O- CaCO3) system controlled by fast reaction rates. Batch reaction calculations are performed for the multiphase system at various temperature and pressure conditions for different initial CO2 saturations. Two methods of approximating the equilibrium state of the system by an operator splitting approach are compared. The first method determines the gas-liquid and solid-liquid equilibria in separate subsequent steps. At reservoir conditions relevant for storage of CO2 (323K, 100bar) and for high CO2 saturations the error in predicted CO2 concentrations in the liquid phase reaches up to -2%. This error can be reduced to less than -0.5% by the second method, where a conjoint gas-liquid-solid equilibrium is accounted for in the reaction calculations. Accordingly, the latter approach should preferably be employed in multiphase flow reactive transport modeling based on operator splitting techniques