Assessment of the risks and environmental impacts of carbon geosequestration requires knowledge about the wetting behavior of mineral surfaces in the presence of CO2 and the pore fluids. In this context, the interfacial tension (IFT) between CO2 and the aqueous fluid and the contact angle, θ, with the pore mineral surfaces are the two key parameters that control the capillary pressure in the pores of the candidate host rock. Knowledge of these two parameters and their dependence on the local conditions of pressure, temperature, and salinity is essential for the correct prediction of structural and residual trapping. We have performed classical molecular dynamics simulations to predict the CO2–water IFT and the CO2–water–calcite contact angle. The IFT results are consistent with previous simulations, where simple point charge water models have been shown to underestimate the water surface tension, thus affecting the simulated IFT values. When combined with the EPM2 CO2 model, the SPC/Fw water model indeed underestimates the IFT in the low-pressure region at all temperatures studied. On the other hand, at high pressure and low temperature, the IFT is overestimated by ∼5 mN/m. Literature data regarding the CO2/water/calcite contact angle on calcite are contradictory. Using our new set of force field parameters, we performed NVT simulations at 323 K and 20 MPa to calculate the contact angle of a water droplet on the calcite {10.4} surface in a CO2 atmosphere. We performed simulations for both spherical and cylindrical droplet configurations for different initial radii to study the size dependence of the water contact angle on calcite in the presence of CO2. Our results suggest that the contact angle of a cylindrical droplet, is independent of droplet size, for droplets with a radius of 50 Å or more. On the contrary, spherical droplets make a contact angle that is strongly influenced by their size. At the largest size explored in this study, both spherical and cylindrical droplets converge to the same contact angle, 38°, indicating that calcite is strongly wetted by water.

Density functional theory (DFT) with semiempirical dispersion corrections (DFT-D2) has been used to calculate the binding energy of a CO2 molecule on the calcite {10.4} surface for different positions and orientations. This generated potential energy landscape was then used to parametrize a classical force field. From this, we used metadynamics (MTD) to derive free energy profiles at 300 and 350 K for CO2 binding to calcite, CO2 binding with Ca2+, and pairing of two CO2 molecules, all for aqueous conditions. We subsequently performed classical molecular dynamics (MD) simulations of CO2 and water on the {10.4} surface at pressures and temperatures relevant for CO2 geological storage. Density profiles show characteristic structured water layering at the calcite surface and two distinct phases of water and CO2. We have also calculated the densities of the CO2-rich and water-rich phases and thereby determined the mutual solubilities. For all the pressures and temperatures in the studied range, CO2 was unable to penetrate the ordered water layers and adsorb directly on the solid surface. This is further confirmed by the free energy profiles showing that in the presence of water there is neither direct adsorption to the {10.4} surface nor contact binding of CO2 with Ca2+. Rather, we saw a weak affinity for the surface of the ordered water layers. At 5 MPa and 323 K, we observed the nucleation of a CO2 droplet located above two structured water layers over the solid. It could not penetrate the structured water but remained bound to the second water layer for the first 10 ns of the simulation before eventually detaching and diffusing away.

Ion incorporation or removal from a solid at the interface with solution is a fundamental part of crystal growth. Despite this, there have been few quantitative determinations of the thermodynamics for such processes from atomistic molecular dynamics due to the associated technical challenges. In this study, we compute the free energies for ion removal from kink sites at the interface between NaCl and water as an illustrative example. To examine the influence of the free energy technique used, we compare methods that follow an explicit pathway for dissolution with those that focus on the thermodynamics of the initial and final states using metadynamics and free energy perturbation, respectively. While the initial results of the two approaches are found to be completely different, it is demonstrated that the thermodynamics can be reconciled with appropriate corrections for the standard states, thus illustrating the need for caution in interpreting raw free energy curves for ion binding as widely found in the literature. In addition, a new efficient approach is introduced to correct for the system size dependence of kink site energies both due to the periodic interaction of charges in an inhomogeneous dielectric system and due to the dipolar interactions between pairs of kinks along a row. Ultimately, it is shown that with suitable care, both classes of free energy techniques are capable of producing kink site stabilities that are consistent with the solubility of the underlying bulk solid. However, the precise values for individual kink sites exhibit a small systematic offset, which can be ascribed to the contribution of the interfacial potential to the pathway-based results. For the case of NaCl, the free energies of the kink sites relative to a 1 M aqueous solution for Na+ and Cl– are found to be surprisingly different and of opposite sign, despite the ions having very similar hydration free energies.

In this study, the adaption of the recently published molecular GFN-FF for periodic boundary conditions (pGFN-FF) is described through the use of neighbor lists combined with appropriate charge sums to handle any dimensionality from 1D polymers to 2D surfaces and 3D solids. Numerical integration over the Brillouin zone for the calculation of π bond orders of periodic fragments is also included. Aside from adapting the GFN-FF method to handle periodicity, improvements to the method are proposed in regard to the calculation of topological charges through the inclusion of a screened Coulomb term that leads to more physical charges and avoids a number of pathological cases. Short-range damping of three-body dispersion is also included to avoid collapse of some structures. Analytic second derivatives are also formulated with respect to both Cartesian and strain variables, including prescreening of terms to accelerate the dispersion/ coordination number contribution to the Hessian. The modified pGFN-FF scheme is then applied to a wide range of different materials in order to examine how well this universal model performs.

Knowledge about the interaction between fluids and solids and the interfacial tension (IFT) that results is important for predicting behavior and properties in industrial systems and in nature, such as in rock formations before, during, and after CO injection for long-term storage. Many authors have studied the effect of the environmental variables on the IFT in the CO-HO system. However, experimental measurements above CO supercritical conditions are scarce and sometimes contradictory. Molecular modeling is a valuable tool for complementing experimental IFT determination, and it can help us interpret results and gain insight under conditions where experiments are difficult or impossible. Here, we report predictions for CO-water interfacial tension performed using density functional theory (DFT) combined with the COSMO-RS implicit solvent model. We predicted the IFT dependence as a function of pressure (0-50 MPa), temperature (273-383 K), and salinity (0-5 M NaCl). The results agree well with literature data, within the estimated uncertainty for experiments and for molecular dynamics (MD) simulations, suggesting that the model can be used as a fast alternative to time-consuming computational approaches for predicting the CO-water IFT over a range of pressures, temperatures, and salinities.