Machine learning modeling of solid electrolyte interphase growth at Li₆PS₅Cl/Li-metal contacts

The structure and growth of the solid electrolyte interphase (SEI) region between an electrolyte and an electrode is one of the most fundamental, yet less-well understood phenomena in solid-state batteries. The initial stages of this growth are currently inaccessible to experiments, and previous parameterized models have suggested mutually contradicting hypotheses about the physics and chemistry at play [1]. I will present a parameter-free atomistic simulation of the SEI growth for one of the currently promising solid electrolytes (Li₆PS₅Cl). We employ moment tensor potentials [2,3], trained on ab initio simulations. Whereas ab initio molecular dynamics (MD) can only tackle a few hundred atoms for 100 ps [4], our machine-learned MD simulation comprises over 30,000 atoms for 10 ns. Thanks to this expanded range of space and timescales, we unveil a growth mechanism in two steps: a rapid initial influx of Li rapidly reduces the electrolyte into an amorphous phase; this is followed by a kinetically slower crystalization of the reduced phase into a Li₂(S_xP_yCl_1-x-y) solid solution. I will present numerical data and clarify the physics underlying the qualitatively different functional forms of the two-phase growth.



[1] Single, F.; Latz, A. and Horstmann, B. Identifying the Mechanism of Continued Growth of the Solid-Electrolyte Interphase. ChemSusChem 11, 1950–1955 (2018).

[2] Shapeev, A. Moment tensor potentials: a class of systematically improvable interatomic potentials. Multiscale Modeling & Simulation, 14(3):1153–1173 (2016).

[3] Novikov, I.; Gubaev, K.; Podryabinkin, E. and Shapeev, A. Mach. Learn.: Sci. Technol. 2 025002 (2021).

[4] Golov, A. & Carrasco, J. Molecular-Level Insight into the Interfacial Reactivity and Ionic Conductivity of a Li-Argyrodite Li6PS5Cl Solid Electrolyte at Bare and Coated Li-Metal Anodes. ACS Appl. Mater. Interfaces 13, 43734–43745 (2021).