Diffusion studies in UO2 with an improved tight-binding potential: SMTB-QB

The theoretical modeling of metal oxides represents a fundamental challenge in materials science due to their complex iono-covalent bonding, multiple valence states, and the crucial role of defects in determining their physical properties. Among these materials, uranium dioxide (UO₂) stands as both a technologically important compound and a model system for studying actinide oxides. We present here an improved tight-binding potential, SMTB-QB, which introduces variable charges ensuring local electroneutrality at the bond level. This key feature enables an accurate description of charged defects, the emergence of an electronic gap, and the ability to handle multiple valence states of the same cation (e.g U³⁺, U⁴⁺, U⁵⁺, etc.). The model extends the capabilities of previous potentials while maintaining computational efficiency applicable to various oxide systems. Its accuracy is validated through extensive comparisons with DFT calculations and experimental data properties of UO₂. Using molecular dynamics simulations, we investigate oxygen and uranium diffusion mechanisms across different temperature ranges and stoichiometries (UO₂±ₓ). Our results reveal three distinct diffusion regimes for oxygen, with migration energies of 0.47 ± 0.03 eV in UO₂₋ₓ and 0.79 ± 0.03 eV in UO₂₊ₓ below 2400 K, converging to an activation energy of 3.67 ± 0.14 eV above 2500 K regardless of stoichiometry. For uranium diffusion, we demonstrate a vacancy-mediated mechanism with a migration energy of 4.07 ± 0.76 eV in stoichiometric UO₂, which shows agreement with experimental data. The SMTB-QB potential thus provides a robust framework for studying nuclear fuel materials and potentially other complex oxide systems.