A Simple Electrostatic Model for the Hard-Sphere Solute Component of Nonpolar Solvation

TL;DR

This paper introduces a simple electrostatic shell capacitor model to estimate nonpolar solvation free energies, achieving high accuracy and physical interpretability, and combining well with electrostatic models for comprehensive solvation predictions.

Contribution

The paper presents a novel electrostatic shell capacitor model for nonpolar solvation energy estimation, with physically meaningful parameters and improved accuracy over existing models.

Findings

RMSE of 0.45 kcal/mol for cavity energies

Total nonpolar model RMSE of 0.55 kcal/mol

Solvation free energies RMS error of 1.35 kcal/mol with combined electrostatic model

Abstract

We propose a new model for estimating the free energy of forming a molecular cavity in a solvent, by assuming this energy is dominated by the electrostatic energy associated with creating the static (interface) potential inside the cavity. The new model approximates the cavity-formation energy as that of a shell capacitor: the inner, solute-shaped conductor is held at the static potential, and the outer conductor (at the first solvation shell) is held at zero potential. Compared to cavity energies computed using free-energy pertubation with explicit-solvent molecular dynamics, the new model exhibits surprising accuracy (Mobley test set, RMSE 0.45 kcal/mol). Combined with a modified continuum model for solute-solvent van der Waals interactions, the total nonpolar model has RMSE of 0.55 kcal/mol on this test set, which is remarkable because the two terms largely cancel. The overall nonpolar model has a small number of physically meaningful parameters and compares favorably to other published models of nonpolar solvation. Finally, when the proposed nonpolar model is combined with our solvation-layer interface condition (SLIC) continuum electrostatic model, which includes asymmetric solvation-shell response, we predict solvation free energies with an RMS error of 1.35 kcal/mol relative to experiment, comparable to the RMS error of explicit-solvent FEP (1.26 kcal/mol). Moreover, all parameters in our model have a clear physical meaning, and employing reasonable temperature dependencies yields remarkable correlation with solvation entropies.