Polaron-Ready Partition DFT for Metal–Oxide Interfaces without Hybrids or U

Decompose an interface into fragments (oxide slab region, metal cluster, and vacancy/defect region) within Partition-DFT (PDFT), and approximate the partition energy via MFOA (Shi et al., 2023). Introduce a localization-sensitive correction in the partition term that penalizes spurious delocalization, enabling symmetry-preserving polaron formation at Ti sites. This approach applies PDFT with MFOA, previously shown to suppress LDA/GGA errors for strong correlation in model hydrogen chains, to heterogeneous interfaces, sidestepping empirical U/hybrid choices. It constrains localization at the fragment level via the partition potential rather than global parameter tuning. The method promises polaron energies and localization patterns at near-GGA cost but with accuracy comparable to hybrids, critical for large supercells and defect statistics, while avoiding spin-symmetry breaking. The impact is a transferable, low-cost route to treat charge localization at catalytic interfaces, enabling predictive screening of co-catalysts, vacancy engineering, and light-assisted charge separation without empiricism.

References:

1. Strong electron correlation from partition density functional theory.. Yi Shi, Yuming Shi, A. Wasserman (2023). Journal of Chemical Physics.