Self‐Healing Mechanism of Lithium in Lithium Metal

Li is an ideal anode material for use in state‐of‐the‐art secondary batteries. However, Li‐dendrite growth is a safety concern and results in low coulombic efficiency, which significantly restricts the commercial application of Li secondary batteries. Unfortunately, the Li‐deposition (growth) mechanism is poorly understood on the atomic scale. Here, machine learning is used to construct a Li potential model with quantum‐mechanical computational accuracy. Molecular dynamics simulations in this study with this model reveal two self‐healing mechanisms in a large Li‐metal system, viz. surface self‐healing, and bulk self‐healing. It is concluded that self‐healing occurs rapidly in nanoscale; thus, minimizing the voids between the Li grains using several comprehensive methods can effectively facilitate the formation of dendrite‐free Li. Li is an ideal anode material for use in state‐of‐the‐art secondary batteries. A deep neural network is adopted to construct a Li potential with quantum‐mechanical computational accuracy. Using this potential, this study achieves a large‐scale Li–metal system simulation and reveals two self‐healing mechanisms, viz. Li surface self‐healing and Li bulk self‐healing.