Accurate Performance Assessment of Lithium-Oxygen Batteries by Obtaining the Globally Stable Structures of Reaction Intermediates.

Single-atom catalysts hold great promise for enhancing redox efficiency in nonaqueous lithium-oxygen batteries, but inconsistent performance evaluations persist due to the vast configurational space of LixOy discharge products. Here, we address this challenge by integrating first-principles calculations with a differential evolution algorithm to perform exhaustive global structure searches. We screened over 1,500 adsorption geometries of LiO2, Li2O2, and (Li2O2)2 on transition-metal and nitrogen codoped graphene (TM-N4C, TM = Ru, Fe, Co, Pt, Ni) and identified the true ground-state configurations. These globally optimized structures exhibit significantly greater thermodynamic stability than the manually selected models commonly used in prior studies, revealing that reliance on a limited set of representative geometries can lead to erroneous performance predictions. Free-energy analyses further highlight Co-N4C as the optimal catalyst, with ultralow overpotentials for oxygen reduction (0.17 V) and evolution (0.10 V), and designate Li2O2 formation and (Li2O2)2 decomposition as the rate-limiting steps in discharge and charge, respectively. The exceptional activity of Co-N4C arises from its balanced adsorption energies, which promote uniform deposition of discharge products during cycling and facilitate their efficient desorption upon charging. By eliminating the biases of manual structure selection, this work establishes a rigorous framework for accurate electrocatalytic performance assessment in lithium-oxygen batteries.