We will present recent progress using machine learning to better analyze the role of local solvation environments around ionic solutes in aqueous solution. The first half of the talk will show an automated cluster-continuum solvation procedure that can be used to correct shortcomings in continuum solvation models for predicting absolute solvation energies of ions needed to model acidity constants and standard redox potentials. Using a Smooth Overlap of Atomic Positions (SOAP) kernel, we show that local solvation environments around ions become more alike with increasing cluster size, thus demonstrating a measure of structural convergence correlating with converged solvation energies. The second half of the talk will show our progress using machine learning methods to accelerate the dynamics simulations needed for accurate free energy barriers for reactions in condensed phase environments.<br/>

Applications of machine learning for studying local solvation environments

255th National Meeting (2018)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.