On the Prediction of Gas Solubility in Brine Solutions for Applications of CO2 Capture and Sequestration

Abstract

Geological storage of CO2 in deep saline aquifers has become a well-accepted method for CO2 sequestration. CO2-solubility in these brine solutions is one of the most important factors in determining the amount of CO2 sequestered in these aquifers. Since the type of salt can significantly alter the CO2-solubility in brine solutions, the impact of water chemistry on CO2 solubility is investigated in this work and results are compared against experimental observations. The current work for predicting solubility of a gas in brine solution containing various salts is based on the extension of well-known Setschenow relation that has been commonly used for salts with monovalent ions. In this research, we extend the Setschenow approach by expressing the solubility in terms of ionic strengths and molar concentrations of each salt. The method also characterizes each component (e.g., gas, anions and cations) against the experimental measurements. A simple methodology, developed with a theoretical framework, is presented to predict the impact of different types of salts on solubility of CO2. This approach can be extended to any type of gases or other solutes (e.g. CH4, H2S etc.) in brine solutions. In particular, The gas solubility in brines is expressed in terms of molar components and ionic strength of each salt. The expression contains unique/characteristic parameters for each component (gas, anions and cations). These parameters for anions and cations of typical formation water (present in oil/gas reservoirs) and CO2/novel solvents are obtained from literature or using regression on experimental data.Results of CO2-solubility were compared with published data in literature, demonstrating that the methodology (presented in the work) can predict the effect of water-chemistry on solubility predictions.The proposed method was tested for a novel solvent (dimethyl ether) and comparison with experimental solubility data show an excellent match between the predictions and measurements.