Understanding of microscopic structure and ionic dynamics in battery components is indispensable to the development of viable all-solid-state batteries. Amorphous materials have attracted growing attention for their unique combination of desirable properties such as ionic conductivity, stability and processability. However, compared to bulk crystalline materials, theoretical understanding of these highly complex disordered systems is hindered by the scale limitations of DFT calculations. To improve atomic-level understanding and guide the rational design of amorphous phases and interfaces, we present a framework based on machine-learned interatomic potentials trained on the fly to explore structures and transport mechanisms in such systems. Applying this framework to the amorphous solid electrolyte Li3PS4 and a compatible amorphous coating, Li3B11O18, we explicitly simulate both the individual glass phases and their heterogeneous interface at the nanosecond timescale. Our results help explain the favorable rate performance of coated materials observed in experiment. Our study demonstrates the promising role of active learning in extending the application of ab-initio modeling to more complex and realistic systems such as amorphous materials and solid-state interfaces.

Atomistic study of amorphous interfaces in all-solid-state batteries using active learning

Fall 2022 abstracts, posters, and presentations will launch in mid-September.

ACS Fall 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.