Deep Potential Model for Analyzing Enhancement of Lithium Dynamics at Ionic Liquid and Perovskite (BaTiO₃) Interface

Recently, ab-initio molecular dynamics (AIMD) based DeePMD (DPMD) potential has not only improved computational accuracy and speed but has also overcome the limitations of traditional force-field-based methods. In this study, we will discuss the technical details and considerations for utilizing this deep learning-based force field, particularly in multi-component systems such as ionic liquids. We will compare the results of structural & dynamical properties calculated from traditional force fields, scaled charge force fields, polarizable force fields, and DPMD force fields.<br/>Finally, we will apply the DPMD force field to an ionic liquid and perovskite interface system, which is virtually impossible to simulate from using traditional force fields. Experiments have shown that the hybrid solid electrolyte (HSE), consisting of nanoscale perovskite particles mixed with an ionic liquid (IL), exhibits excellent flame retardancy, thermal stability, and improved ionic conductivity compared to pure IL electrolyte. We will analyze the structural and dynamical differences induced by the perovskite interface in comparison to bulk ionic liquid using DPMD force field simulations.