High-Throughput DFT Screening of ABO_3-d Perovskites for Hydrogen Storage

Hydrogen is a promising optimal energy source due to its clean and versatile nature, enabling the decarbonization of various fields and contributing to a sustainable and low-carbon future. To utilize the hydrogen energy in many applications ranging from energy science to electronics, we aim to discover the host materials for excellent hydrogen capacity and conduction. Recently, all charge states of the H carrier: H⁺ (proton), H⁰ (atom), and H^- (hydride ion) have been experimentally reported in specific ABO_3-d perovskites,^[1-2] which indicates the prospect of perovskite compounds in hydrogen energy and information science. However, it remains unclear that how the crystal, electronic structure, and chemical property of each element in ABO_3-d perovskites affect their performance in hydrogen incorporation and transport. To this end, we performed a comprehensive design of bulk channels in near-equilibrium perovskites by high-throughput density of functional theory (DFT) calculations. The perovskites were diversely screened with notable cations for A-sites (Ca, Sr, Ba) and B-sites (Co, Fe, Mn, Ti, and Zr) in a collection of phases (cubic, tetragonal, orthorhombic, and hexagonal). Our results showcased that the perovskites with active B-site cations (Co, Fe and Mn) can only uptake protons (H⁺, rather than hydride H^-) in all configurations of stoichiometric and non-stoichiometric ABO_3-dconfigurations with minor effects from crystal phases. In comparison, the perovskites with d⁰ cations (Ti and Zr) on B sites are inert to uptake protons (H⁺) in their ground-state phases of pristine structures, however, the resulting hydride (H^-) will be easily observed in their non-stoichiometric configurations. Moreover, we also provided machine learning (ML) predictors for the solubility of O vacancies in the ABO_3-dPerovskites, which compiles the electronic effective mass as correlation term into ML predictors of single-site O vacancy reported in previous studies.^[3-4]Finally, we also investigated the Grotthuss mechanism of H transfer in our ABO_3-dperovskites to uncover the competitive effects of hopping and rotation motions towards the H conduction. Overall, we highlighted the importance of many metrics of the perovskites toward hydrogen uptake and conductivity from HT-DFT screening of our ABO_3-dcandidates. The stable cubic-phase CaMnO₃ and BaTiO_3-dconfigurations are the most promising parent models for H⁺ and H^-host materials, respectively. Our ML-based solubility predictors for O deficient and kinetic understanding of H diffusion provide broader avenues to achieve precision-guided discovery of the next-generation materials for hydrogen storage.<br/><br/>1. Vera, C.Y.R., et. al., Ding, D. A mini-review on proton conduction of BaZrO₃-based perovskite electrolytes. <i>J. Phys. Energy</i> 3, 13 (2021).<br/>2. Liu, X., et. al., Haugsrud, R. Highly correlated hydride ion tracer diffusion in SrTiO_3-xH_x oxyhydrides. <i>J. Am. Chem. Soc.</i> 141, 4653-4659 (2019).<br/>3. Wexler, R. B., et. al., Carter, E. A. Factors governing oxygen vacancy formation in oxide perovskites. <i>J. Am. Chem. Soc.</i> 143, 13212-13227 (2021).<br/>4. Baldassarri, B., et. al., Wolverton, C. M. Oxygen vacancy formation energy in metal oxides: high-throughput computational studies and machine-learning predictions. <i>Chem. Mater. </i>35, 10619-10634 (2023).