Continuum approaches to solvation.
Deniz Gunceler (Physics, Cornell)

Electronic structure calculations of molecules and surfaces in a liquid can accelerate the development of many technologies ranging from solar energy harvesting to lithium batteries. The conventional approach to modeling solvation phenomena is ab-initio molecular dynamics, where forces obtained from quantum mechanical calculations are used together with a numerical integration scheme to do the thermodynamic sampling. While accurate, this approach demands enormous computational resources and time. A much more economical approach is the polarizable continuum model, where explicit solvent molecules are replaced with a continuous field. Polarizable continuum models (PCMs) have been applied to some solvated molecules; but they do not sufficiently capture solvation effects to describe highly polar systems like surfaces of ionic solids.

In this work, we present a nonlinear fluid functional within the framework of Joint Density Functional Theory. In this scheme, the fluid is treated as a continuous distribution of dipoles that responds to the solute, which is described starting from the exact free energy functional for point dipoles. We also show existing PCMs can be recovered as the linear limit of our functional. Our description is of similar computational cost to PCMs, and captures complex solvation effects like dielectric saturation without requiring new fit parameters. For both polar and nonpolar molecules, it achieves chemical accuracy in solvation energies. Our functional also makes it possible to investigate chemistry on the surface of lithium battery materials, which PCMs incorrectly predict to be unstable. Lastly, we outline a plan for predicting the optical and electron energy-loss spectrum of solvated ions, which requires us to approximate the stationary states of the combined electronic and fluid functional.