Predicting Electrolyte Conductivity Directly from Molecular-Level Interactions

Ionic conductivity in liquid electrolytes depends on molecular interactions dictating the relative populations and behaviors of stoichiometric ion solvation clusters. However, the connections from molecular interactions to bulk ionic conductivity are not well-established, limiting the fast in silico evaluation of liquid electrolytes before experimental synthesis. To illustrate a bottom-up approach to predicting ionic conductivity, we outline a method using a chemical physics formalism with parameters computed by classical molecular dynamics (MD) simulations. The method is demonstrated on two liquid electrolyte chemistries with salts of differing electrolyte strengths. Using the proposed approach without empirical fitting, we achieve quantitative and qualitative prediction agreements with respect to conductivity measurements for strong and weak electrolytes, respectively. This approach provides the basis for closing the structure-based design computational loop to aid emerging high-throughput electrolyte discovery frameworks.