Benchmarking density functional theory in the correlated nickelate LiNiO₂

Understanding the electronic structure of correlated oxides remains a central challenge for theory, as standard density functional theory (DFT) approximations often fail to capture the subtle balance of charge transfer, magnetism, and structural distortions. We present systematic benchmarks of widely used functionals—LDA, PBE, PBE+U, SCAN, r²SCAN, PBE0, and SCAN0—against diffusion Monte Carlo (DMC) for Li_0.5NiO₂, LiNiO₂, and NiO₂. We introduce a robust atomic partitioning scheme for noisy DMC densities. DMC establishes a consistent physical picture: NiO₂ is nonmagnetic with undistorted octahedra, an apical distortion in LiNiO₂ leads to a Ni³⁺ state and 1 µB moment, and a disproportionated phase in Li_0.5NiO₂ results in alternating Ni³⁺- and Ni⁴⁺-like sites. Benchmarking shows r²SCAN performs the best across charge, spin, and radial density descriptors. SCAN also performs well, while hybrid functionals (PBE0, SCAN0) perform unexpectedly poorly, and PBE+U+V yields inconsistent trends between charge and spin densities. We identify r²SCAN as an efficient alternative to DMC for correlated nickelates and highlight the importance of accurate benchmarks to advance predictive modeling of charge-transfer driven phenomena in cathodes, catalysts, and correlated electron systems.