Inhomogeneous Fluid Approach to Solvation Thermodynamics. 2. Applications to Simple Fluids

In the previous paper expressions for the partial molar energy and entropy at infinite dilution have been derived based on the inhomogeneous forms of the energy equation and the correlation expansion for the entropy. These expressions are here applied to a series of solutes of varying size in dense hard-sphere and Lennard-Jones solvents, some of which serve as reference systems for comparison with water. Numerical results are obtained under the assumption that the inhomogeneous solvent-solvent pair correlation function in the mixture is equal to the bulk solvent radial distribution function (Kirkwood superposition approximation). The correlation functions required are obtained by both integral equation theory (Percus-Yevick approximation) and Monte Carlo simulations. The thermodynamic results are compared with equation of state, integral equation, and free energy simulation results for the same systems. For hard-sphere systems the excess entropies are in good agreement with equation-of-state results but in many Lennard-Jones systems the calculated partial molar energies and entropies are lower than the expected values. This is attributable to overestimation of the structure of the bulk triplet correlation function by the superposition approximation. The decomposition of the chemical potential shows that similar solvation free energies can have entirely different physical origins. Specifically, in solvents of high cohesive energy density the chemical potential is dominated by the breakup of solvent-solvent interactions locally around the solute. In solvents of low cohesive energy density it is dominated by the pressure-volume term. Increase in solvent-solvent interaction strength leads to increase in the chemical potential of the solute due to the higher solvent reorganization energy, which is insufficiently compensated by an increase in solvent reorganization entropy.