High-Voltage, High Capacity Aluminum-Rich Lithium Cathode Materials: A Bayesian Optimization and First-Principles Study

High voltage and low volumetric change are key objectives in the design of lithium battery cathodes. Increasing the voltage leads to a higher energy density of the batteries, which can satisfy the requirements of emerging electric vehicles and power grid markets. Reducing the volumetric change mitigates chemomechanical degradation. Layered nickel–cobalt cathode materials are currently dominant in the rechargeable battery market but have limited voltages and suffer from significant volumetric changes during cycling. In this work, we efficiently navigate the space of possible compositions for the formula LinNin–x–yCoxMyX2n (M = Mn, Al, Cr or Fe, X = O, S) by combining Bayesian Optimization and density functional theory. Our approach enables us to find combinations of metal elements in layered oxides that are potential candidates for future lithium battery cathodes. In particular, we report aluminum-rich cathode structures with voltages of ∼4.4 V, volume change lower than 1%, and that are thermodynamically stable against crystal decomposition. With theoretical capacities exceeding 350 mAh/g, the proposed materials suggest that aluminum can, which is usually used as a coating material and dopant in layered Co/Ni cathode materials, can be examined as the principal element in the cathode.