Data-driven discovery and intelligent design of artificial hybrid interphase layer for stabilizing lithium-metal anode

Lithium metal is a promising anode material for high-energy-density batteries, but its application is hindered by safety concerns arising from dendrite growth. In this work, we propose a high-throughput workflow that combines quantum-mechanical simulations with machine learning to accurately predict self-assembled monolayers (SAMs) that can assemble an artificial inorganic-organic hybrid interphase layer on the Li-metal anode to enhance cycling stability and mitigate dendrite growth. The workflow comprises automatic data collection, first-principles simulations, and screening of candidate molecules using machine learning. We screened out 128 molecules from the PubChem database and identified the eight best candidates with low Li diffusion barriers and high mechanical stability. A structure-property relationship was established between the Li diffusion barrier and the structural characteristics of head, middle, and tail groups in the SAMs using simple quantum mechanical (QM) dipole and electrostatic potential descriptors. These results open new avenues for designing highly stable Li-metal anodes for practical use in Li-metal batteries.