SE154: Theoretical insights into hydrogenation and proton transport in ABO3 perovskites
Wednesday, November 4, 2026 at 1:40 PM to 2:10 PM · 30 min. (America/Chicago)
Room 107A (First Floor)
Oral Presentation
Information
The hydrogenation of ABO3 perovskites presents a promising avenue for enabling proton-coupled electron transfer in energy technologies, yet the fundamental thermodynamic and kinetic bottlenecks remain poorly understood. Here, we combine quantum-level simulations with machine learning (ML) to systematically investigate hydrogenation across 118 pristine ABO3 perovskites, spanning cubic, tetragonal, orthorhombic, and hexagonal phases. Density functional theory (DFT) calculations on 18 representative systems reveal that thermodynamic feasibility is strongly governed by B-site chemistry: redox-inert cations such as Zr and Ti yield high reaction energies (ΔG = 2–4 eV), while redox-active cations such as Fe, Co, Mn, and Ce favor hydrogenation (ΔG 3 where ab initio molecular dynamics indicated highly localized proton motion. In contrast, hexagonal SrMnO3 displayed balanced barriers and long-range diffusion with mean square displacements ~60 Å2 at 773 K, underscoring the role of phase symmetry in enabling conductivity.To accelerate discovery, we developed ML models incorporating both spatial (bond dissociation energies Eb(A–O), Eb(B–O)) and electronic (p- and d-band centers, band gap Eg, reduction potential Vr(B)) descriptors. Regression analyses identified Eb(B–O) and Eg as the dominant features, reflecting electron transfer between B-site d orbitals and hydrogen s orbitals during incorporation. Our final ML model achieved excellent accuracy (R2 = 0.96, MAE = 0.22 eV), enabling large-scale screening of 118 perovskites. Two trends emerged: (i) orthorhombic phases dominate pristine compositions and display strong hydrogenation driving forces but suffer from proton-trapping kinetics, necessitating phase engineering; (ii) cubic and hexagonal perovskites, while thermodynamically less favorable (1–4 eV), exhibit balanced kinetic barriers that promote conduction. This ML model led to the identification of two additional promising candidates, hexagonal BaPdO3 and cubic BaSnO3, which were subsequently validated by DFT calculations.
Author/Institution List
Zhen Jiang (Presenter, Middle Tennessee State University College of Basic and Applied Sciences); Chris Wolverton (Co-Author, Northwestern University)