calendar Add meeting dates to your calendar.

 

Paper No. 2
Presentation Time: 1:45 PM

MOLECULAR MODELING OF CARBON DIOXIDE-WATER MIXTURES UNDER GEOLOGIC SEQUESTRATION CONDITIONS


NIELSEN, Laura C., Geochemistry, Lawrence Berkeley National Laboratory, 307 McCone Hall #4767, Berkeley, CA 94720-4767, BOURG, Ian C., Geochemistry, Lawrence Berkeley National Laboratory, 1 Cyclotron Road Mail Stop 90-1116, Berkeley, CA 94720 and SPOSITO, Garrison, Geochemistry, Lawrence Berkeley National Laboratory, 1 Cyclotron Road Mail Stop 90-1116, Berkeley, CA 94720-3114, lcnielsen@lbl.gov

The success of geologic carbon sequestration depends largely on the ability of the caprock to prevent the escape of CO2. Capillary breakthrough overpressure—the pressure above which CO2 may leak through the caprock—in turn depends upon the interfacial tension (IFT) between supercritical CO2 and brine. Molecular dynamics (MD) simulations are widely used to probe the interactions that lead to fluid-fluid properties such as IFT. Among the leading models of CO2-water interactions, three-site parametric approaches are favored for their accuracy and ease of implementation. We used MD simulations to compare three of these models [EPM2 (Harris and Yung, 1995), PPL (Panhuis et al., 1998), and DZ (Duan and Zhang, 2005, 2006)] for their ability to reproduce IFT and solubility data in CO2-SPC/E H2O mixtures over the range of P and T relevant to geologic sequestration. Long-range interactions in the models are characterized by Coulomb potentials between partial charges localized on three sites in CO2 and H2O, while short-range interactions are characterized by Lennard-Jones (LJ) potentials with two parameters, potential well depth (ε) and hard-sphere radius (σ). Each model differs significantly in one or more of these parameters as applied to the CO2-H2O interaction, enabling us to examine the effects of varying ε, σ, or q on IFT and solubility. None of these models perfectly reproduces experimental IFT or solubility data. For example, the PPL model under-predicts IFT and over-predicts CO2 solubility as functions of P, while the DZ and EPM2 models track IFT reasonably well but under-predict CO2 solubility. The best agreement overall was obtained with the EPM2 model, which may be improved to reproduce solubility more accurately by increasing its CO2 ε-parameter.
Meeting Home page GSA Home Page