Abstract

Due to its immense importance as an amorphous solid electrolyte in thin-film devices, lithium phosphorus oxynitride (LiPON) has garnered significant scientific attention. However, investigating Li⁺ transport within the LiPON framework, especially across a Li||LiPON interface, has proven challenging due to its amorphous nature and varying stoichiometry, necessitating large supercells and long time scales for computational models. Notably, machine-learned interatomic potentials (MLIPs) can combine the computational speed of classical force fields with the accuracy of density functional theory (DFT), making them the ideal tool for modeling such amorphous materials. Thus, in this work, we train and validate the neural equivariant interatomic potential (NequIP) framework on a comprehensive DFT-based data set consisting of 13,454 chemically relevant structures to describe LiPON. With optimized training (validation) energy and force mean absolute errors of 5.5 (6.1) meV/atom and 13.6 (13.2) meV/Å, respectively, we employ the trained potential to model Li transport in both bulk LiPON and across Li||LiPON interfaces. Amorphous LiPON structures generated by the optimized potential resemble those generated by ab initio molecular dynamics, with N being incorporated on nonbridging apical and bridging sites. Subsequent analysis of Li+ diffusivity in the bulk LiPON structures indicates broad agreement with prior computational and experimental literature. Further, we investigate the anisotropy in Li⁺ transport across the Li(110)||LiPON and Li(111)||LiPON interface, where we observe Li transport across the interface to be one order of magnitude slower than Li motion within the bulk Li and LiPON phases. Nevertheless, we note that this anisotropy of Li transport across the interface is minor, and we do not expect it to cause any significant impedance buildup. Finally, our work highlights the efficiency of MLIPs in enabling high-fidelity modeling of complex noncrystalline systems over large length and time scales.