Predicting Surface Atomic Structures on Doped Perovskite Oxides Using Grand Canonical Monte Carlo: Model System of La0.6Sr0.4FeO3–δ
Surface properties of metal oxides are determined by their atomic structures, which can be exceedingly complex, including different surface terminations, chemistries and phase precipitates. First-principles studies have advanced our understanding of oxide surface reconstructions, but often face challenges due to the limited exploration of surface compositional and configurational spaces. To overcome this limitation, this study integrates bulk and surface thermodynamics consistently to effectively sample surface configurations. Our method is Grand Canonical Monte Carlo (GCMC) calculations, accelerated by a machine-learning interatomic potential. This approach is applied to a model perovskite oxide, La0.6Sr0.4FeO3–δ (LSF), a state-of-the-art oxygen electrode material for solid oxide cells. To make GCMC calculations computationally feasible, we first assessed the competing bulk phases in equilibrium with LSF. We used these phases as initial guesses at the surfaces of LSF. GCMC then refined these surface structures to identify equilibrium surface structures on LSF under different thermodynamic conditions. In calculating surface energies, the chemical potentials of the cations were evaluated as functions of the oxygen partial pressure, the bulk oxygen and cation vacancy concentrations, and the overall bulk composition. Surface segregations of SrO2, SrO, La2O3, Fe, and Ruddlesden–Popper (RP) phases were found at different oxygen chemical potentials. Surface phase diagrams of LSF (001) show that SrO-terminated RP phase segregation dominates under moderate oxygen environment, while oxidizing and reducing environments favor SrO2 formation and Fe exsolution, respectively. A decrease in temperature reduces the oxygen partial pressure window for RP segregation on the LSF (001) surface. The results provide useful insights into the surface thermodynamics of LSF (001), and the method can be leveraged to investigate surface atomic structures across a broader range of complex oxides.