Unraveling Li-Ion Solvation Dynamics in Ether-Based Electrolytes via High-Throughput Molecular Dynamics Simulations and Machine Learning

Abstract

Lithium-ion batteries (LIBs) are now widely used as energy storage systems in portable electronic devices and electric vehicles. The Li-ion solvation dynamics in electrolytes significantly influence the overall performance of LIBs. In this work, we performed high-throughput molecular dynamics smilations on thousands of ether-based electrolytes screened from a vast molecular database containing over 110 million organic molecules. Machine learning models combined with Shapley additive explanation analysis reveal that the presence of bulky alkyl groups reduce the solvent dipole moment, while enhancing the coordination ability of lithium ions with anions. Additionally, the introduction of ether-type groups (-O-) in the rings of cyclic ether molecules negatively affects the solvent dipole moment but positively influences the Li-anion coordination number. In contrast, ether-type groups (-O-) introduced in linear chains correlate positively with solvent dipole moment while negatively with the Li-anion coordination number. This work provides insights into the relation between molecular structure and Li-ion solvation dynamics in electrolyte solutions.