Activity Coefficients and Solubility of CaCl₂ from Molecular Simulations

Abstract

We obtain the activity coefficients and lower bounds to the solubility of CaCl2 in aqueous solutions at temperatures between 298.15 and 473.15 K using molecular simulations with three previously developed nonpolarizable force fields. We find that a scaled-charge force field gives incorrect activity coefficients at low concentration and has a different absolute chemical potential than experiments, but still predicts an accurate solubility for the calcium chloride dihydrate. The two full-charge models have chemical potentials and activity coefficients closer to experiments, but there is considerable variation between them, with the chemical potentials differing by over 100 kJ·mol−1. The slow dynamics of the full-charge models at high concentrations are unrealistic and require advanced sampling methods to obtain the activity coefficients. We find that development of polarizable models is likely necessary to accurately represent both thermodynamic and transport properties of divalent electrolyte solutions.

Individual Ion Activity Coefficients in Aqueous Electrolytes from Explicit-Water Molecular Dynamics Simulations

“…This is attributed to the limitations associated with classical nonpolarizable force fields, as also observed in prior studies. 72,77,79,81 Having established that our approach, based on calculations for individual ions, gives overall activity coefficients (MIACs) in agreement with literature values, we now turn our attention to the activity coefficients of the individual ions (IIACs). Figure 5 shows our results and compares them to experimental data and prior simulations for other models.…”

“…Simulation results for all four salts compare reasonably well with the experimental data of Robinson and Stokes at low concentrations before deviating at higher concentrations. This is attributed to the limitations associated with classical nonpolarizable force fields, as also observed in prior studies. ,,, …”

“…Because of the strong electrostatic interactions, insertion of single ions is carried out slowly. LJ and Coulombic interactions are expressed as functions of a coupling parameter (λ for LJ and ϕ for Coulombic), as in previous studies by Panagiotopoulos and co-workers. ,,, LJ interactions are turned on first with 21 windows with λ equally spaced in the range of [0,1]. While turning on LJ interactions, a soft-core “shifted” potential is applied to avoid discontinuity.…”

Computing the Liquidus of Binary Monatomic Salt Mixtures with Direct Simulation and Alchemical Free Energy Methods

“…Solid–liquid equilibrium calculations in systems with two or more components have largely focused on computing the solubilities of higher-melting components in a solvent − rather than computing liquidus/solidus curves in mixtures where the pure components have similar melting temperatures. − Both the direct coexistence and thermodynamic free-energy-based methods are more challenging for multicomponent systems. For direct coexistence, multicomponent systems appear to exacerbate the problems presented by system size effects and slow equilibration.…”

Prediction of Alkanolamine pKa Values by Combined Molecular Dynamics Free Energy Simulations and ab Initio Calculations

“…The standard state chemical potential μ i † (T, P) for a solute and for the solvent may then be expressed in terms of their reference state properties (infinite dilution in the case of solutes, and pure solvent in the case of the solvent) by the same formal expression involving molecular simulation quantities as recently derived for the solute case, 26,27 first implemented for solutes by Wilkins 28 and subsequently by Young et al 29 μ μ ρ μ ρ…”