High-Throughput MD Simulations of Li⁺ Conductivity in the Discharge Products of Li-O₂ Batteries Using Machine Learning Potentials

Lithium-oxygen secondary batteries (LOBs) are a type of next-generation battery with the potential to achieve gravimetric energy densities surpassing those of current lithium-ion batteries. However, sufficient charge-discharge cycle performance has not been realized under the expected high energy density conditions due to the decomposition of the electrolyte caused by various reactive oxygen species generated during charge and discharge. To address these challenges, all-solid-state LOBs, which use solid electrolytes stable against reactive oxygen species, have attracted significant attention.¹ Generally, constructing a three-phase interface—composed of the electrolyte phase, porous cathode phase, and gas phase—is difficult when using solid electrolytes, significantly limiting the number of reaction sites. Additionally, the deposition of discharge products (Li₂O₂ and its analogues) leads to the disappearance of reaction sites, causing reactions on the cathode to stop prematurely.
In principle, if the discharge products possess Li⁺ conductivity, these challenges could be overcome, raising expectations for all-solid-state LOBs with higher current density and energy density. However, the relationship between the structure of discharge products and Li⁺ conductivity remains poorly understood. While ab initio molecular dynamics (AIMD) simulation is a powerful tool for predicting Li⁺ conductivity, the required computational resources are vast, making it challenging to predict the Li⁺ conductivity of various discharge products.
In this study, we conducted high-throughput MD simulations based on machine learning potentials with accuracy comparable to AIMD. Specifically, we systematically investigated the Li⁺ conduction properties of various kinds of discharge product structures, including those with Li⁺ defects and heteroatom doping.

1. A.Kondori, M.Esmaeilirad, A.M.Harzandi, R.Amine, M.T.Saray, L.Yu, T.Liu, J.Wen, N.Shan, H.Wang, A.T. Ngo, P. C. Redfern, C. S. Johnson, K. Amine, R. Shahbazian-Yassar, L. A. Curtiss, M. Asadi, A room temperature rechargeable Li₂O-based lithium-air battery enabled by a solid electrolyte. Science 379,499-505(2023)