Ruthenium atoms placed into precise crystallographic positions of a cobalt oxide lattice have produced an acidic oxygen evolution catalyst that runs for more than 240 hours at 10 mA cm⁻² with an overpotential of just 240 mV, according to a new study published in the journal Catal. The key was not simply adding ruthenium, but choosing which cobalt site it replaces: substitution at octahedral Co³⁺ sites in spinel Co₃O₄ dramatically outperformed substitution at tetrahedral Co²⁺ sites, which required a higher overpotential of 280 mV and survived only 42 hours under identical testing conditions.
The work addresses one of the central bottlenecks in proton exchange membrane water electrolysis (PEMWE). While PEMWE offers high current density, high-purity hydrogen, and compact design compared with alkaline systems, its anodic oxygen evolution reaction (OER) suffers large overpotentials that inflate cell voltages and operating costs. Under the harsh acidic conditions and high anodic potentials involved, only a handful of materials combine sufficient activity with long-term durability, and commercial systems currently lean heavily on iridium- and ruthenium-based noble metal catalysts.
Iridium oxides are exceptionally stable in acid because dissolving Ir species requires a high oxidation potential, but the scarcity and cost of iridium severely limit industrial scalability. Ruthenium-based catalysts are more active and cheaper, yet they oxidize into soluble RuO₄ species at anodic potentials above 1.4 V versus RHE, leading to rapid degradation. Non-noble alternatives such as Co₃O₄, MnO₂, TiO₂, Nb₂O₅, and CeO₂ have been explored, but cobalt oxide in particular suffers from limited intrinsic activity and susceptibility to degradation in acidic media.
Spinel Co₃O₄ offered the researchers a structural handle that few materials provide. The compound crystallizes in cubic Fd3m symmetry with the general formula AB₂O₄, where A-site Co²⁺ cations occupy tetrahedral positions and B-site Co³⁺ cations reside in octahedral sites. Octahedral Co³⁺ forms strong σ-σ interactions with oxygen, while tetrahedral Co²⁺ displays weaker π-π interactions. Because the two sites can be distinguished and selectively substituted, the team could probe the intrinsic role of each coordination environment directly, an approach previous substitution studies could not fully deliver since the substituting ion’s own activity confounded the results.
The researchers synthesized two families of catalysts: Oct-RuₓCo₃₋ₓO₄ with ruthenium replacing octahedral cobalt (x = 0.08, 0.13, 0.47, 0.59) and Tet-RuₓCo₃₋ₓO₄ with tetrahedral substitution (x = 0.07, 0.18, 0.42), with atomic ratios confirmed by ICP-OES. Density functional theory calculations of dissolution energies motivated the design: while cobalt dissolution energy was similar at roughly 4.50 eV in both environments, ruthenium at the octahedral site required 6.07 eV to dissolve versus only 3.22 eV at the tetrahedral site, indicating that octahedral ruthenium is far more stable.
Extensive characterization confirmed that the site-selective substitution had actually been achieved. XRD patterns matched the standard spinel Co₃O₄ phase with no impurity peaks besides a carbon signal, and EDX mapping showed homogeneous Ru, Co, and O distribution, ruling out surface RuO₂ clusters. Raman spectroscopy revealed distinct redshift patterns: octahedrally substituted samples showed stronger shifts in the A1g and Eg modes associated with octahedral units, while tetrahedrally substituted samples shifted more in the F2g modes tied to tetrahedral units. Ru K-edge EXAFS sealed the case, with a 2.47 Å second-shell peak in Oct-Ru₀.₁₃Co₂.₈₇O₄ matching the octahedral Co-Co distance of pristine Co₃O₄ (2.46 Å), and a 3.12 Å peak in the tetrahedral sample assigned to Ru_tet–Co_oct coordination.
In 0.5 M H₂SO₄ with roughly 2 mg cm⁻² catalyst loading, the activity differences were clear. The best octahedral sample, Oct-Ru₀.₅₉Co₂.₄₁O₄, reached 10 mA cm⁻² at only 192 mV overpotential, compared with 380 mV for undoped Co₃O₄. At similar ruthenium contents, octahedral substitution consistently beat tetrahedral substitution, a trend that persisted when currents were normalized by electrochemically active surface area. Oct-Ru₀.₁₃Co₂.₈₇O₄ showed a double-layer capacitance of 88.74 mF cm⁻² versus 42.91 mF cm⁻² for its tetrahedral counterpart and 27.82 mF cm⁻² for pristine Co₃O₄, indicating more exposed active sites.
Stability testing sharpened the contrast further. Oct-Ru₀.₁₃Co₂.₈₇O₄, containing only about 4 atomic percent ruthenium, operated for over 240 hours at 10 mA cm⁻² with a decay rate of just 0.6 mV h⁻¹, whereas Tet-Ru₀.₁₈Co₂.₈₂O₄ and higher-loadings tetrahedral samples surged in voltage within 50 and 20 hours, comparable to pure Co₃O₄. Post-test XRD and Raman on the tetrahedral sample revealed an emergent RuO₂ crystalline phase and loss of Co₃O₄ vibrational modes, suggesting dissolution or phase separation. In contrast, the octahedral catalyst retained its composition after 240 hours, with a post-test Ru:Co atomic ratio of 0.13:2.87 essentially unchanged from the fresh sample.
The mechanism behind this divergence emerged from in-situ Raman spectroscopy and computation. On the octahedral catalyst, Raman peaks at roughly 560 cm⁻¹ beyond 1.45 V versus RHE indicated CoOOH intermediates from cobalt oxidation, while a peak near 1080 cm⁻¹ above 1.6 V signaled the direct *O-*O coupling characteristic of the oxide path mechanism (OPM). Neither feature appeared on the tetrahedral electrode. DFT calculations showed the OPM energy barrier on Ru_oct–Co_oct dual sites drops to 0.46 eV, while the tetrahedral Ru–O–Co configuration raises the barrier to 0.82 eV because the ruthenium-cobalt spacing exceeds 3 Å, hindering orbital overlap. The shorter 2.79 Å Moct–Cooct distance in the octahedral sample, versus 2.89 Å in the tetrahedral model, is key to triggering the dual-site pathway.
Electronic analysis completed the picture. Bader charge calculations showed octahedral ruthenium substitution reduces the positive charge on adjacent octahedral cobalt centers (1.11 versus 1.33 in pristine Co₃O₄), mitigating the over-oxidation of Co³⁺ to unstable species that drives dissolution, whereas tetrahedral substitution raises the charge on neighboring cobalt sites. Charge density difference maps revealed directional electron transfer from ruthenium to cobalt through Ru-O-Co bridges in the octahedral case, but restrained transfer in the tetrahedral geometry. Translating to a device scale, a 2 × 2 cm² PEM water electrolyzer with the octahedral catalyst on the anode and Pt/C on the cathode required a cell voltage of 1.66 V for 250 mA cm⁻² and ran stably at 20 mA cm⁻² for over 30 hours with a 4.25 mV h⁻¹ decay rate, using a remarkably low noble metal loading of only 0.15 mg cm⁻². The authors present the strategy as a design paradigm for cost-effective, low-precious-metal catalysts for hydrogen production.
Subject of Research: Site-selective ruthenium substitution in spinel cobalt oxide for acidic oxygen evolution catalysis
Article Title: Site-selective Ru doping in spinel Co3O4 unlocks dual-site synergy for acidic water electrolysis
Article References: Wang, Y., Zhang, Y., Chen, S., Fan, Y., Peng, J., Shifa, T. A., Li, Y., Vomiero, A., Wang, F., & Sun, X. (2026). Site-selective Ru doping in spinel Co3O4 unlocks dual-site synergy for acidic water electrolysis. Catal, 2(1), Article 4. https://doi.org/10.1007/s44422-026-00018-w
Image Credits: AI Generated
DOI: 10.1007/s44422-026-00018-w
Keywords: water electrolysis, oxygen evolution reaction, spinel Co3O4, ruthenium doping, PEMWE, acidic OER, oxide path mechanism, dual-site catalysis, DFT calculations, electrocatalyst stability, hydrogen production, EXAFS