Oxygen Evolution Reaction: Mechanisms & Catalysis

• Oxygen evolution reaction (OER) is the four-electron, four-proton oxidation of water/hydroxide forming triplet oxygen via multistep PCET.
• It is the kinetic bottleneck in water splitting, metal–air batteries, and solar fuel devices, driving the need for optimized catalyst design.
• OER research integrates experimental metrics, theoretical analysis, and data-driven models to achieve low overpotentials and high durability.

The oxygen evolution reaction (OER) is the four-electron, four-proton oxidation of water or hydroxide to molecular oxygen. It represents the principal kinetic bottleneck in electrolytic water splitting, recharge of metal–air batteries, and numerous solar fuel devices due to inherently sluggish multistep PCET (proton-coupled electron transfer) pathways and the requirement to form a spin-triplet O₂ product from closed-shell reactants. OER catalysis research thus focuses on the design, mechanistic elucidation, and optimization of earth-abundant transition-metal oxides, alloys, and emerging nonmetallic materials under both acidic and alkaline regimes. This article surveys OER mechanisms, activity descriptors, experimental and theoretical quantification, catalyst platforms, and contemporary design guidelines.

1. Fundamental Mechanisms and Electronic Structure

The OER proceeds via consecutive PCET steps involving adsorbed oxygen intermediates. In alkaline media, the canonical sequence on a generic surface site ($^*$) is:

1. $\mathrm{OH}^-$ adsorption: $$\ast + \mathrm{OH}^- \rightarrow \ast\mathrm{OH} + e^-$$
2. $\ast\mathrm{OH}$ deprotonation: $$\ast\mathrm{OH} + \mathrm{OH}^- \rightarrow \ast\mathrm{O} + \mathrm{H}_2\mathrm{O} + e^-$$
3. O–O bond formation: $$\ast\mathrm{O} + \mathrm{OH}^- \rightarrow \ast\mathrm{OOH} + e^-$$
4. O₂ release: $$\ast\mathrm{OOH} + \mathrm{OH}^- \rightarrow \ast + \mathrm{O}_2 + \mathrm{H}_2\mathrm{O} + e^-$$

Analogous steps operate under acidic conditions, with H₂O as reactant and H⁺ transfer (Mamun et al., 18 Mar 2026, Cheng et al., 2020, Antipin et al., 2020).

A spin-state transition is inherent: reactant $\mathrm{OH}^-$ (or H₂O) is a singlet, while O₂ is a triplet. Mechanistically, at least three electrons removed during OER must share the same spin to conserve spin angular momentum; deviation incurs an exchange penalty and increases overpotential (Li et al., 2020). This “three-same-spin” constraint is critical for the design of spin-polarized or half-metallic catalysts to facilitate efficient triplet O₂ evolution.

Two mechanistic families exist:

• Adsorbate Evolution Mechanism (AEM): O–O bond forms between surface-bound *O or *OOH moieties. The *O → *OOH step or subsequent *OOH → * + O₂ step is often rate-limiting, with overpotential determined by the largest free-energy difference $\Delta G_i$ among PCET intermediates (Antipin et al., 2020, Li et al., 2020, Zhao et al., 7 Mar 2026).
• Lattice Oxygen Mechanism (LOM): Lattice O directly participates in O–O coupling, often favored in highly covalent, oxygen-hole-rich oxides; this pathway can lower the highest kinetic barrier and overpotential (Li et al., 2020).

2. Benchmark Experimental Metrics and Kinetic Parameters

Quantitative assessment of OER catalysts employs several key metrics:

Metric	Definition/Descriptor	Typical Value for Leading Catalysts
Overpotential at 10 mA·cm⁻², $\eta_{10}$	$\mathrm{OH}^-$0	$\mathrm{OH}^-$1215–400 mV (NiFe-LDH, IrO₂, OsO₂)
Tafel slope, $\mathrm{OH}^-$2	$\mathrm{OH}^-$3 [mV/dec]	30–80 mV/dec
Turnover frequency, TOF	Rate per active site or atom	0.02–0.2 s⁻¹
Mass/specific activity	j/mass or j/ECSA	$\mathrm{OH}^-$40.4 A/g at 1.65 V (best perovskites)
Stability	Hours to weeks at 10–100 mA·cm⁻²; drift $\mathrm{OH}^-$510 mV	$\mathrm{OH}^-$610–120 h at 10 mA·cm⁻² (IrO₂, OsO₂, Fe–Ni LDH)

OER activity is commonly benchmarked at $\mathrm{OH}^-$7 in 0.1–1 M KOH or acidic media (typically 1 M H₂SO₄). State-of-the-art catalysts achieve $\mathrm{OH}^-$8 values of $\mathrm{OH}^-$9215–350 mV (NiFe-LDH, Ni42 steel, OsO₂), with Tafel slopes as low as 31–71 mV/dec (Schäfer et al., 2017, Zhao et al., 7 Mar 2026, Gong et al., 2014, Cheng et al., 2020, Fujioka et al., 13 Aug 2025). Long-term operation and Faradaic efficiency close to unity are critical for industrial viability.

3. Materials Platforms and Structure–Activity Relationships

Transition Metal Oxides, Spinels, and Perovskites

• NiFe Oxides and Oxyhydroxides: Exhibit $$\ast + \mathrm{OH}^- \rightarrow \ast\mathrm{OH} + e^-$$0 as low as 230–260 mV (exfoliated LDH, amorphous NiFeOx) with Tafel slopes 29–40 mV/dec, attributed to interplay between Fe (active site) and Ni (conductivity/structural promotion). Extended studies confirm that Fe–O motifs, not Ni–O, are the true OER active centers; Ni modulates Fe oxidation and enables low-barrier O–O coupling (Gong et al., 2014, Hao et al., 2020, Mirabella et al., 2021).
• Perovskite Oxides (ABO₃): OER activity correlates with both geometric and electronic structure. Geometric descriptors such as the Goldschmidt tolerance factor $$\ast + \mathrm{OH}^- \rightarrow \ast\mathrm{OH} + e^-$$1 and the octahedral factor $$\ast + \mathrm{OH}^- \rightarrow \ast\mathrm{OH} + e^-$$2 successfully predict activity trends. Symbolic regression reveals $$\ast + \mathrm{OH}^- \rightarrow \ast\mathrm{OH} + e^-$$3 (in V) fits 18 known and four newly discovered perovskites, four of which (e.g., Cs₀.₄La₀.₆Mn₀.₂₅Co₀.₇₅O₃) outperform BSCF (Weng et al., 2019). Structurally, large A-site cations and small, oxidized B-cations with eg occupancy ≈1 and oxygen vacancy content δ ≈ 0.20 are required for high activity (Cheng et al., 2020, Antipin et al., 2020).
• Bulk OsO₂ Single Crystals: OsO₂ achieves $$\ast + \mathrm{OH}^- \rightarrow \ast\mathrm{OH} + e^-$$4 of 390 mV, $$\ast + \mathrm{OH}^- \rightarrow \ast\mathrm{OH} + e^-$$5 = 43 mV/dec, and $$\ast + \mathrm{OH}^- \rightarrow \ast\mathrm{OH} + e^-$$6120 h stability in 1 M KOH, with the (110) facet offering the lowest theoretical overpotential (0.31 V) due to moderate *OH adsorption, minimal charge transfer, and kinetic stability against dissolution. Crystal integrity (low defect, low grain boundary) is a key stability descriptor (Zhao et al., 7 Mar 2026).

Alloys, Graphene Hybrids, and Non-Precious Metals

• Zn–Al Alloys: Zn₀.₉Al₀.₁ outperforms pure Zn (lowest $$\ast + \mathrm{OH}^- \rightarrow \ast\mathrm{OH} + e^-$$7 = 0.240 V @ 12 mA·cm⁻² in 1 M KOH, $$\ast + \mathrm{OH}^- \rightarrow \ast\mathrm{OH} + e^-$$8 = 53.5 mV/dec) due to optimal eutectic microstructure, maximal exchange current, and low charge transfer resistance (Mamun et al., 18 Mar 2026).
• Co₃O₄/Graphene Hybrids: Co₃O₄/N–rmGO hybrids achieve $$\ast + \mathrm{OH}^- \rightarrow \ast\mathrm{OH} + e^-$$9 ≈ 0.31 V and $\ast\mathrm{OH}$0 mV/dec, rivaling IrO₂/RuO₂ benchmarks. Synergy arises from coupled Co–O–C (and Co–N–C) bonding, high nanocrystal density, and conductive graphene backbones (Liang et al., 2011).

Metal-Free Systems and Earth-Abundant Architectures

• Boron Clusters (B₁₂ Frameworks): Fully hole-doped AlB₁₄ achieves $\ast\mathrm{OH}$1 mV, $\ast\mathrm{OH}$2 mV/dec, and current densities an order of magnitude higher than Co₃O₄. The unique mechanism features molecular water—not OH⁻—adsorption at high-field B₁₂ surfaces. Outward-pointing p-orbitals and large local electric fields enable efficient H₂O activation and robust durability without redox-active metals (Fujioka et al., 13 Aug 2025).
• Phosphorized Steels: 3D nanoporous S235-P and anodized Ni42 steel develop high-surface-area Fe phosphide cores with in situ conversion to active Fe/Mn or Ni–Fe oxyhydroxides during OER. They exhibit $\ast\mathrm{OH}$3–326 mV, 68–72 mV/dec Tafel, and $\ast\mathrm{OH}$4400\,000 s durability in 1–7 M KOH or near-neutral pH (Schäfer et al., 2017, Han et al., 2018).
• Fe-doped CuO: Optimally doped Cu₀.₉₈₄Fe₀.₀₁₅₆O₁₊δ achieves $\ast\mathrm{OH}$5 mV, $\ast\mathrm{OH}$6 mV/dec, surpassing most Cu-based and even IrO₂ standards due to conductivity, vacancy, and surface area enhancements (Baral et al., 2024).

4. Descriptor-Based and Multivariate Optimization

No single electronic or structural parameter unambiguously predicts OER performance across platforms. Multivariate approaches show high-performance perovskites require:

• Electronic conductivity σ ≳ 10 S·cm⁻¹
• Oxygen vacancy concentration δ ≳ 0.20
• Flat-band potential $\ast\mathrm{OH}$7 V vs RHE

Simultaneous satisfaction of these conditions yields high activities; violation of any yields outliers (e.g., BSCF with suppressed $\ast\mathrm{OH}$8 at high $\ast\mathrm{OH}$9) (Cheng et al., 2020).

Data-driven modeling (symbolic regression) identifies the $$\ast\mathrm{OH} + \mathrm{OH}^- \rightarrow \ast\mathrm{O} + \mathrm{H}_2\mathrm{O} + e^-$$0 descriptor as directly predictive for perovskites, and supervised optimization in self-driving laboratories accelerates mixed-metal oxide discovery by one to two orders of magnitude over manual methods (Weng et al., 2019, Fatehi et al., 2023).

Electronic-structure-derived metrics ($$\ast\mathrm{OH} + \mathrm{OH}^- \rightarrow \ast\mathrm{O} + \mathrm{H}_2\mathrm{O} + e^-$$1 occupancy, band gap, DOS at $$\ast\mathrm{OH} + \mathrm{OH}^- \rightarrow \ast\mathrm{O} + \mathrm{H}_2\mathrm{O} + e^-$$2), geometric factors ($$\ast\mathrm{OH} + \mathrm{OH}^- \rightarrow \ast\mathrm{O} + \mathrm{H}_2\mathrm{O} + e^-$$3, $$\ast\mathrm{OH} + \mathrm{OH}^- \rightarrow \ast\mathrm{O} + \mathrm{H}_2\mathrm{O} + e^-$$4), and surface energetics ($$\ast\mathrm{OH} + \mathrm{OH}^- \rightarrow \ast\mathrm{O} + \mathrm{H}_2\mathrm{O} + e^-$$5, $$\ast\mathrm{OH} + \mathrm{OH}^- \rightarrow \ast\mathrm{O} + \mathrm{H}_2\mathrm{O} + e^-$$6) all enter volcano plots, but deviations ("descriptor breaks") are resolved only in multi-dimensional spaces or with explicit inclusion of interfacial phenomena and surface reconstructions (Cheng et al., 2020, Antipin et al., 2020).

5. Mechanistic Probing by Advanced Spectroscopy, Imaging, and Modeling

Spectroscopy: Operando XANES (O K-edge, Ir L₃-edge) tracks deprotonation, oxidation, and formation of electron-deficient O species in IrO₂, correlating spectral shifts to particular surface reconstructions and potential-driven changes (Nattino et al., 2019). Raman and EXAFS elucidate dynamic formation of resting Fe=O in NiFe-LDH or phase transitions in bulk α-Li₂IrO₃ during OER (Pandya et al., 2022, Hao et al., 2020).

Imaging: Compressive Raman imaging demonstrates the coexistence of bulk-mediated electrochemical–chemical (EC) and classical adsorbate evolution (AEM) pathways in α-Li₂IrO₃. At low overpotentials, both pathways operate (Tafel ~60 mV/dec); at high overpotential ($$\ast\mathrm{OH} + \mathrm{OH}^- \rightarrow \ast\mathrm{O} + \mathrm{H}_2\mathrm{O} + e^-$$7 mV), AEM dominates (Tafel ~120 mV/dec) (Pandya et al., 2022). AFM mapping resolves nanoscale O₂ bubble nucleation at step edges of Pt, correlating bubble blocking with current density losses (Bodappa et al., 15 Nov 2025).

Theoretical Analysis: DFT calculations provide thermodynamic and kinetic profiles for both conventional and oxygen-lattice-involving mechanisms. For perovskite oxynitrides, cation site, local strain, and N → O substitution are found to modulate $$\ast\mathrm{OH} + \mathrm{OH}^- \rightarrow \ast\mathrm{O} + \mathrm{H}_2\mathrm{O} + e^-$$8, $$\ast\mathrm{OH} + \mathrm{OH}^- \rightarrow \ast\mathrm{O} + \mathrm{H}_2\mathrm{O} + e^-$$9, and overpotential by up to 0.8 V (Ouhbi et al., 2019, Ouhbi et al., 2018).

6. Design Guidelines and Optimization Strategies

Rational OER catalyst optimization encompasses:

• Tuning Multidescriptors: For perovskites, target simultaneously σ ≳ 10 S·cm⁻¹, δ ≳ 0.20, $$\ast\mathrm{O} + \mathrm{OH}^- \rightarrow \ast\mathrm{OOH} + e^-$$0 V; employ strategic A-site/B-site doping, vacancy engineering, and post-synthetic surface treatments (e.g., carbon sonication) (Cheng et al., 2020, Antipin et al., 2020).
• Exploiting Electronic Structure: Choose B-site 3d cations and local coordination that yield eg occupancy $$\ast\mathrm{O} + \mathrm{OH}^- \rightarrow \ast\mathrm{OOH} + e^-$$1, optimal covalency, and spin polarization near the Fermi level for efficient spin-aligned electron extraction (Li et al., 2020).
• Maximizing Surface Area and Defects: Leverage nanoarchitecturing, oxygen vacancy modulation, higher surface area supports, and controlled heterostructures to maximize density and accessibility of active sites (Liang et al., 2011, Baral et al., 2024).
• Stabilizing Surface Phases: Preserving crystal integrity (low-grain boundary, minimal amorphization) is essential to prevent rapid dissolution, as in bulk OsO₂, while still enabling dynamic surface reconstruction to catalytically active oxyhydroxides (Zhao et al., 7 Mar 2026, Han et al., 2018).
• Data-Driven Discovery: Employ machine learning, symbolic regression, and closed-loop robotic laboratories to explore compositional and structural parameter spaces efficiently, accelerating optimization cycles by up to 100× (Weng et al., 2019, Fatehi et al., 2023).

7. Future Directions and Open Problems

• Multiobjective Optimization: Balancing activity, stability, cost, and corrosion resistance for industrially relevant current densities and ambient or acidic environments.
• Mechanism–Structure Mapping: Further dissecting the contributions of AEM, LOM, and EC mechanisms with site-specific and dynamic imaging/spectroscopy across complex oxide, alloy, and cluster systems.
• New Materials Paradigms: Metal-free frameworks (boron clusters), molecular water activation, and lattice engineerable perovskite/oxynitride hybrids to transcend the limitations of redox-active transition metal dependence (Fujioka et al., 13 Aug 2025).
• Industrial Integration: Direct translation of nanoporous steel-based OER architectures and scalable synthesis of durable perovskite catalysts to device-relevant electrolyzer and battery formats (Schäfer et al., 2017, Han et al., 2018).
• Theory-Guided Design: Extending volcano and multi-descriptor models to account for interface, strain, dynamic dissolution, and the role of the electrolyte in modulating electron transfer and intermediate stabilization.

The field continues to converge toward descriptors, mechanistic understanding, and platform materials that enable low-overpotential, high-durability OER at scale in earth-abundant systems, with emerging computational, spectroscopic, and automated optimization tools setting the stage for next-generation catalyst discovery.