Statistical, Bottom-Up Model for Chemical Diffusion Based on Atomic Vacancy Sublattice Configurations in Layered Lithium Nickel Oxide Cathode Materials

Abstract

To investigate the influence of the local environment on Li-ion diffusivity in layered lithium nickel oxide (Li_xNiO₂) cathodes, a bottom-up, multiscale-modeling approach is applied, utilizing density functional theory (DFT) with corrected Coulomb and van der Waals interactions to describe the energy-structure relationship of Li_xNiO₂ (x = 0 – 1) in good agreement with previous experiments. The UNiversal CLuster Expansion (UNCLE) is employed to construct high-probability Li–vacancy configurations and the Nudged Elastic Band (NEB) method to compute energy barriers for representative Li diffusion mechanisms. By fitting a cluster expansion model to these barriers, diffusion barriers are determined for all possible Li–vacancy configurations within a nearest-neighbor approximation. Based on this description, Li-concentration-dependent diffusion coefficients are predicted for the entire Li-concentration range. For the Li_xNiO₂ crystal lattice, the computed Li chemical diffusivities well lie within experimental ranges, namely 10 – 10 cm² s⁻¹, at room temperature with activation energies around 37.9 kJ mol⁻¹. The maximum diffusivity of 4.23 × 10 cm² s⁻¹ is identified at x = 0.63. The new analytical, self-consistent approach here relies on configurational samplings of individual atomistic mechanisms and can be applied to investigate diffusion properties in further dilute and concentrated alloy systems more efficiently than common numerical procedures.