Reactive transport models for mineral CO2 storage in basaltic rocks

CO2 mineral storage in basalts may provide a long lasting, thermodynamically stable and environmentally benign solution to reduce anthropogenic CO2 in the atmosphere. We present here development of reactive transport models of this process with focus on the CarbFix experiment at Hellisheidi geothermal power plant in Iceland. There, up to 2.2 tons/year of purified CO2 of volcanic origin will be dissolved in water and injected at intermediate depths (400-800 m) into relatively fresh basaltic lava. Plans call for a full-scale injection if the experiment is successful. Reactive transport modeling is an important factor in the CarbFix project, providing tools to predict and optimize long-term management of the injection site as well as to quantify the amount of CO2 that has the potential of being mineralized. TOUGHREACT and iTOUGH2 are used to develop reactive fluid flow models that simulate hydrology and mineral alteration associated with injecting dissolved CO2 into basalts. The mineral reactions database in TOUGHREACT has been revised and extended, providing an internally consistent database suitable for mineral reactions of interest for this study. A multiple interacting continua (MINC) dissolution model was developed to simulate the long and short-term dissolution of basaltic glass taking into account dissolution kinetics, leached layer formation and diffusion-limited dissolution rates. Our main focus has been on developing a three dimensional field model of the injection site at Hellisheidi. Hydrological parameters of the model were calibrated using iTOUGH2 to simulate tracer tests that have been ongoing since 2007. Modeling results indicate groundwater velocity in the reservoir to be significantly lower than expected. The slow groundwater velocity may necessitate increasing groundwater flow by producing downstream wells at low rates after CO2 injection has started. The three dimensional numerical model has proven to be a valuable tool in simulating different injection and pumping schemes. Reactive chemistry was coupled to the model and TOUGHREACT used for reactive transport simulations, which are ongoing. Preliminary results confirm dissolution of primary basaltic minerals as well as carbonate precipitation. Secondary mineral abundance is highly dependent on temperature, pCO2 and flow rate. Optimally, simulations with the CarbFix field model should determine which injection scenario will maximize mineralization of injected CO2 as well as to show the depth and temperature range best suited for the mineralization.