Probing surface degradation pathways of charged nickel-oxide cathode materials using machine-learning methods

While nickel-based layered oxide cathodes offer promising energy and power densities in lithium-ion batteries, they suffer from instability when fully delithiated upon charge. Ex-situ studies often report a structural degradation of the charged cathode materials, but the precise mechanism is still poorly understood on the atomic scale. In this work, we combine high level DFT calculations with molecular dynamics using machine-learning interatomic potentials to study structural degradation of fully delithiated LiNiO2 surfaces at the top of charge. We find a previously unreported, stable reconstruction of the (012) facet, with more facile oxygen loss compared to the pristine surfaces, and an oxygen vacancy formation energy closely corresponding to experimental decomposition temperatures of charged cathodes. Furthermore, we use molecular dynamics simulations to sample Ni ion migration into alkali-layer sites that is a kinetically plausible initiation step for surface degradation towards thermodynamically stable products.