Understanding interfacial evolution in solid-state batteries using machine learning force fields*

The interfacial instabilities in solid-state batteries (SSBs) significantly affect the cell performance. In this talk, we will demonstrate direct observation of interfacial evolution at the Li₇La₃Zr₂O₁₂ (LLZO) solid-electrolyte/LiCoO₂ cathode interface from large-scale molecular dynamic simulations enabled by validated machine-learning force fields (MLFFs). Our results unravel the relationship between the surface chemistries of LLZO and LCO and propensities for interfacial degradation. We will further address the impact of element doping (as often found in LLZO) on Li-ion transport in single grain LLZO and its grain boundaries (GBs). It is observed that dopants tend to segregate at the LLZO GBs and form clusters that can lead to prenucleation of dopant-rich secondary phases at the GBs. This phenomenon is presumed to be general regardless of the type of GBs. Our findings imply that interlayer design may be needed to alleviate the intrinsic interfacial degradation for enhanced performance.