Ground-state search and modification effects of lanthanide substitution in LiNiO2: a first-principles study

Abstract

LiNiO₂ (LNO) is a highly promising cathode material for lithium-ion batteries, but its performance is consistently limited by various stability issues. Lanthanide elements (Ln) possess all the necessary characteristics to serve as excellent dopants for modifying LNO. Experimental studies have demonstrated that Ln doping can effectively enhance the performance of high-nickel materials. However, the convergence issues in computational studies of systems containing Ln remain a significant challenge in the field, leading to a scarcity of computational research on LNO+Ln systems. In practical calculations, LNO+Ln models exhibit poor convergence and unstable convergence energies. We attribute this to the strong Coulomb interactions of the 4f electrons in Ln ions, which significantly affect the system’s energy, combined with their diverse electronic configurations that tend to produce multiple metastable states, resulting in a complex energy landscape. In our tests, we found a correlation between the specific values of the 4f electron magnetic moments of Ln ions and the convergence energy. The setting of the magnetic moment convergence parameters directly influences the model’s convergence quality and the energy of the converged state. Based on this, we developed a ground-state search method using the 4f electron magnetic moment values as a feature in the cutoff energy convergence plot. This method enables rapid and accurate calculations of LNO+Ln systems, significantly reducing computational resource consumption. Finally, we obtained the crystal and electronic structures of the ground state for the LNO+Ln(La-Gd) systems and calculated the Li/Ni disordering formation energy and oxygen vacancy formation energy. We discussed the results and analyzed the underlying mechanisms, revealing that the LNO+Ce model exhibits the feature of the most stable structure, the highest Li/Ni disordering formation energy, and the highest oxygen vacancy formation energy, making it a highly promising doping modification scheme for LNO. These findings are fully consistent with experimental conclusions on Ce-doped high-nickel materials. Our computational approach makes it possible to conduct purely computational studies on Ln-doped layered material systems, paving the way for further in-depth research in multiple directions. This work provides a reference for experimental studies on LNO+Ln systems, offers a solution to the computational challenges of lanthanide-doped systems, and holds significant importance for advancing the application of lanthanide elements in layered cathode materials.