Resolving the Non-Arrhenius Behavior of Oxygen Anion Conductivity in YSZ Using Off-Lattice and Coarse-Grained Kinetic Monte Carlo

Solid oxide fuel cells (SOFCs) are a key clean energy technology that converts chemical energy into electricity, but the operation at low temperatures (e.g., below 1000 K) remains extremely challenging because the oxygen anion (O–2) conductivity of yttria-stabilized zirconia (YSZ) electrolytes drops too rapidly with decreasing temperature. To provide deeper insights into the low-temperature transport mechanism, here, we developed a combined off-lattice and coarse-grained kinetic Monte Carlo (OC-KMC) algorithm to simulate the O–2 migration at 800–2000 K. Using the Y–Zr–O global neural network (G-NN) potential, we established a comprehensive reaction table from parallel off-lattice KMC (OL-KMC) at different temperatures, which contains 24,089 distinct minimum configurations and 621,557 elementary ion migration pathways. The OC-KMC overcomes the low-temperature conductivity prediction difficulty and predicts the ionic conductivity σ = 0.00071–0.38 S/cm at 800–1400 K, which aligns well with experimental measurements and resolves the physical origin of the abrupt drop in low-temperature conductivity. We show that (i) the rate-determining step of O–2 migration corresponds to the escape event from the global minimum and (ii) the non-Arrhenius conductivity arises from the coupled, temperature-dependent decreases in the apparent activation energy and the activation entropy. This work demonstrates the power of the OC-KMC approach in resolving complex kinetics in SOFC electrolyte materials, pointing to new directions for the design of electrolyte materials.