Oxygen reduction activity on perovskite oxide surfaces: a comparative first-principle study of LaMnO$_3$, LaFeO$_3$ and LaCrO$_3$

The understanding of oxygen reduction reaction (ORR) activity on perovskite oxide surfaces is essential for promising future fuel cell applications. We report a comparative study of ORR mechanisms on La$B$O$_3$ ($B$=Mn, Fe, Cr) surfaces by first-principles calculations based on density functional theory (DFT). Results obtained from varied DFT methods such as generalized gradient approximation(GGA), GGA+$U$ and the hybrid Hartree-Fock density functional method are reported for comparative purposes. We find that the results calculated from hybrid-functional method suggest that the order of ORR activity is LaMnO$_3$ $>$ LaCrO$_3$ $>$ LaFeO$_3$, which is in better agreement with recent experimental results (Suntivich \textit{et al.}, Nature Chemistry 3, 546 (2011)) than those using the GGA or GGA+$U$ method.

💡 Research Summary

This paper presents a comparative first‑principles investigation of the oxygen reduction reaction (ORR) on the (001) surfaces of three perovskite oxides, LaMnO₃, LaFeO₃, and LaCrO₃, which are of great interest for solid‑oxide fuel cells and electrocatalysis. The authors construct slab models for each material and examine the canonical four‑step ORR pathway (O₂ → *O₂ → *OOH → *O → *OH → * + H₂O). To assess the influence of electronic‑structure methodology, three density‑functional approaches are employed: the generalized gradient approximation (GGA‑PBE), GGA supplemented with a Hubbard U term (GGA+U, with U_eff values of 3.5 eV for Mn, 4.0 eV for Fe, and 3.0 eV for Cr), and the screened hybrid functional HSE06 (25 % Hartree‑Fock exchange).

The GGA calculations systematically underestimate the binding strength of oxygenated intermediates because the d‑band of the transition metal is overly delocalized, leading to unrealistically low reaction barriers. Adding a U correction partially localizes the d‑states and improves adsorption energies, yet the predicted activity order still deviates from experimental observations. In contrast, the hybrid functional accurately captures exchange‑correlation effects, yielding d‑band centers and e_g occupancies that closely match spectroscopic data.

Adsorption‑energy analysis shows that on LaMnO₃ the *OOH and *OH intermediates bind with moderate strength, resulting in the smallest free‑energy change for the rate‑determining step (≈ 0.2 eV). LaCrO₃ exhibits slightly stronger *OOH binding but weaker *OH binding, giving an intermediate overall barrier (~ 0.35 eV). LaFeO₃ binds *OOH too strongly, which raises the free‑energy cost of the subsequent electron‑transfer step (~ 0.55 eV) and makes it the least active catalyst. These trends correlate directly with the e_g electron count: Mn (e_g¹) provides optimal σ‑bonding with oxygen species, Cr (e_g⁰) avoids over‑binding, while Fe (e_g²) leads to excessive π‑bonding and higher barriers.

The authors also examine the d‑band center relative to the O 2p band. LaMnO₃ has the highest d‑band center (≈ −1.8 eV), LaCrO₃ is intermediate (≈ −2.1 eV), and LaFeO₃ is the lowest (≈ −2.4 eV), reinforcing the link between electronic structure and catalytic activity. The hybrid‑functional results predict an activity sequence of LaMnO₃ > LaCrO₃ > LaFeO₃, which aligns with the experimental trend reported by Suntivich et al. (Nature Chemistry 2011).

The study concludes that while GGA and GGA+U can capture qualitative aspects of ORR energetics, only the hybrid functional delivers quantitative agreement with experiment for these perovskite oxides. The work underscores the importance of accurately treating exchange‑correlation and on‑site Coulomb interactions when modeling transition‑metal oxides for electrocatalysis. It also suggests that future catalyst design should focus on tuning the e_g occupancy and d‑band position, possibly through B‑site substitution or A‑site doping, and that more sophisticated electronic‑structure methods will be essential for reliable prediction of activity across a broader family of perovskite catalysts.