Solubility of carbon dioxide in aqueous fluids and mineral suspensions at 294 K and subcritical pressures

An experimental investigation has been carried on the solubility of CO2 in water and 1 M NaCl between 0.3 and 4 MPa, in order to test the validity of the results given by various modelling codes. In addition to experiments with pure fluids, the effect of a range of likely reservoir minerals on CO2–water interactions, including K-feldspar, kaolinite, calcite, Ca-montmorillonite and Na-montmorillonite were also investigated. In addition to measurements of CO2 solubility, the pH of the CO2-saturated suspensions was also measured directly at pressures of up to 1 MPa. The results demonstrate that predictions of CO2 solubility made with PHREEQC and Geochemist’s Workbench agree to within 20% with the experimental value, provided corrections are first made off-line for the fugacity coefficient of CO2, while predictions from standalone models are slightly more accurate. In the presence of mineral suspensions, PHREEQC and Geochemist’s Workbench give good results for calcite and kaolinite but underestimate the pH of montmorillonite-bearing assemblages while slightly overestimating the pH of K-feldspar suspensions. These results are significant because they indicate that CO2-charged fluids reacted with clays may be less acidic than indicated by the models, which will impact predictions of the potential for dissolution of reservoir and cap rock minerals, as well as the potential for leaching of toxic metals.

► CO2 solubility at subcritical P may be lower than predicted by modelling codes.
► Standalone models predict CO2 solubility better than fluid–rock interaction codes.
► Modelled CO2 solubilities depend strongly on the fugacity coefficient used.
► CO2-charged fluids reacted with clays may be less acidic than models suggest.
► CO2-charged fluids may thus mobilise less metal from grain surfaces than thought.