Improving the reliability of classical molecular dynamics simulations in battery electrolyte design
Xin He, Yujie Zhang, Haomiao Li, Min Zhou, Wei Wang, Ruxing Wang, Kai Jiang, Kangli Wang
Abstract
Explorations into new electrolytes have highlighted the critical impact of solvation structure on battery performance. Classical molecular dynamics (CMD) using semi-empirical force fields has become an essential tool for simulating solvation structures. However, mainstream force fields often lack accuracy in describing strong ion-solvent interactions, causing disparities between CMD simulations and experimental observations. Although some empirical methods have been employed in some of the studies to address this issue, their effectiveness has been limited. Our CMD research, supported by quantum chemical calculations and experimental data, reveals that the solvation structure is influenced not only by the charge model but also by the polarization description. Previous empirical approaches that focused solely on adjusting ion-solvent interaction strengths overlooked the importance of polarization effects. Building on this insight, we propose integrating the Drude polarization model into mainstream force fields and verify its feasibility in carbonate, ether, and nitrile electrolytes. Our experimental results demonstrate that this approach significantly enhances the accuracy of CMD-simulated solvation structures. This work is expected to provide a more reliable CMD method for electrolyte design, shielding researchers from the pitfalls of erroneous simulation outcomes.