Comparative Modeling of Ions and Solvent Properties in Ca-Na Montmorillonite by Atomistic Simulations and Fluid Density Functional Theory

Abstract Molecular dynamics (MD) simulations provide an accurate description of the mineral–fluid interface from the perspective of the atomistic level taking into account all atom interactions. This simulation approach is computationally expensive if applied to large molecular systems. Classical Fluid Density Functional Theory (f-DFT) delivers structural and thermodynamic information at comparatively small computational costs. Numerous applications of f-DFT for electrolytes neglect an explicit consideration of solvent. In this work, an unrestricted three-component model (3CM) of f-DFT was applied, which incorporates Lennard-Jones (LJ) attractions for the description of the short-range interactions of fluid–fluid and fluid–wall rather than the hard sphere repulsions, named DFT/LJ-3CM. The DFT/LJ-3CM model considers ions as charged LJ particles and treats solvent molecules as neutral LJ particles. To validate the performance of the DFT/LJ-3CM, the f-DFT calculations were compared with atomistic simulations for montmorillonite (Mnt) with various hydrated states in electrolyte solutions. This benchmarking was used to assess critically the advantages and limitations of the f-DFT model. The calibrated DFT/LJ-3CM model for Na and Ca Mnt was applied to calculate cation selectivity for the ion exchange equilibrium with effective ion radius and swelling behavior of Mnt. The predictions of the DFT/LJ-3CM model were found to be in good agreement with the atomistic simulations and experimental data under a wide range of conditions. At the same time, the DFT calculations were 3–4 orders of magnitude faster than conventional MD simulations. Thus, the DFT/LJ-3CM model can be a computationally effective alternative to atomistic simulation in providing structural and thermodynamic properties of fluid–clay mineral interfaces. The DFT/LJ-3CM model provides a robust approach, which can be used for upscaling in reactive transport simulators and modeling ion migration taking place under more complex thermo-chemo-hydro-mechanical conditions.