Understanding phase behaviour in the geological storage of carbon dioxide

Access to files

•   FULL-TEXT.PDF (pdf)

Abstract

Noble gas partitioning between supercritical CO2-H2O phases can be used to monitor Carbon Capture and Storage (CCS) sites and their natural analogues. However, in order for viable application, noble gas partitioning within these environments must be well constrained. Present estimates of partition coefficient for these systems are taken from the low pressure pure noble gas-water experiments of Crovetto et al. and Smith (Crovetto et al., 1982; Smith, 1985). The effect a supercritical CO2 phase may have on noble gas partitioning is assumed negligible, although this has not been empirically verified. In this work this assumption of noble gas behaviour within a supercritical CO2-H2O binary phase system is evaluated using a combined approach of experiment and simulation.Using a specially commissioned high pressure system at the British Geological Survey paired CO2 and H2O samples were collected from noble gas-enriched systems at pressures and temperatures ranging between 90 – 140 bar and 323.15 – 373.15 K. These were analysed for their noble gas content using a quadrupole mass spectrometer system developed specifically for this project. By comparing the relative concentrations of noble gases in each phase partition coefficients were defined for the experimental conditions. These were compared to their low pressure analogues. At higher CO2 densities all noble gases expressed a significant deviation from predicted partition coefficients. At the highest density (656 kg/m3) helium values decreased by -54% (i.e. reduced solubility within CO2) while argon, krypton and xenon values increased by 76%, 106% and 291% respectively. These deviations are due to supercritical CO2 acting as a polar solvent, the solvation power of which increases as a function of density. Polarisation is induced in each noble gas within this solvent based on their respective polarisabilities. Hence xenon, krypton and argon become more easily solvated as a function of CO2 density while solvating helium becomes harder. These deviation trends are well described using a second order polynomial. This fit defines a deviation coefficient which can be used to adapt low pressure partition coefficients to allow accurate predictions of partitioning within highly dense CO2 phases.Concurrently a Gibbs Ensemble Monte Carlo (GEMC) molecular model was iteratively developed to reproduce noble gas behaviour within these experimental systems. By optimising noble gas-water interactions a pure noble gas-water system was constructed for each noble gas at low pressure which replicated published partition coefficients. These optimised interactions were subsequently applied to low pressure CO2-H2O systems where partition coefficients were derived by calculating excess chemical potentials of noble gases in each phase. Again a good agreement was observed with published values. When the model was applied to the experimental conditions however, a poor agreement with the experimental values was observed. Instead simulated values replicated the low pressure Crovetto et al. and Smith datasets (Crovetto et al., 1982; Smith, 1985). This was due to no CO2-noble gas polarisation terms being included in the current iteration of the model. By including this within the model in the future a full reconciliation between the datasets is expected.

Keyword(s)

Gibbs Ensemble Monte Carlo; Henry's Law; carbon dioxide; noble gas; supercritical

Bibliographic metadata

Institutional metadata

Record metadata