Sequential Reconstruction Unlocks Durable Oxygen Redox for High-Rate Green Hydrogen

Green hydrogen has long been haunted by a stubborn chemical bottleneck: the oxygen evolution reaction, the slow half of water electrolysis that steals voltage, wastes electricity, and degrades catalysts. A team led by researchers at the National University of Singapore now reports a way out of this trap, and the numbers they present are striking. Writing in Nature Sustainability, the group describes a ‘sequential reconstruction activation’ strategy in which a catalyst builds its own optimal active phase, step by step, under the very conditions that would normally destroy it. The resulting electrode drives an anion-exchange-membrane water electrolyzer at 4.1 amperes per square centimeter at just 1.8 volts in 1 M potassium hydroxide at 80 degrees Celsius, and survives nearly 3,500 hours of accelerated ageing at 60 degrees Celsius. In a field where catalysts often fade within days at industrial current densities, that combination of rate and endurance is the kind of result that redraws design rules.

The core insight concerns a mechanism that electrochemists have courted for a decade: lattice-oxygen redox. In conventional metal oxide and oxyhydroxide catalysts, oxygen atoms in the crystal framework sit passively as spectators; every bond made and broken during water oxidation happens at adsorbed intermediates such as *OH and *O. That is safe but slow, because the adsorption energies of these intermediates are locked together by rigid scaling relationships. Make *OH binding stronger to speed up one step, and *O binding strengthens too, penalizing another. The scaling relations impose a theoretical floor on the overpotential, a tax every conventional catalyst must pay. Lattice-oxygen redox breaks that tax. When framework oxygen atoms themselves participate in the redox chemistry, oxygen-oxygen bonds can form through direct coupling of lattice oxygen, sidestepping the adsorbate scaling conundrum entirely.

The catch is stability, and it has been brutal. Oxidizing lattice oxygen means ripping electrons from the very orbitals that hold the crystal together. Once activated indiscriminately, that oxidation delocalizes across the lattice, holes accumulate on framework oxygen, metal-oxygen bonds soften and break, and the material dissolves or amorphizes. Many catalysts that unlock lattice oxygen in the laboratory die within hours under the punishing polarization of a real electrolyzer. The literature is littered with tantalizing activity reports that collapse under accelerated ageing. The Singapore-led team, whose senior authors include Junmin Xue, Haoyin Zhong, Bin Tian and Xiaopeng Wang, asked a different question: what if lattice oxygen could be activated not all at once, but in a spatially ordered, electronically programmed sequence?

Their answer begins with an unlikely precursor: a composite of chromium(III) oxide and nickel diselenide, Cr2O3/NiSe2. When the electrode is held under anodic polarization in alkaline electrolyte, the reconstruction unfolds in two chemically distinct stages, and the order matters enormously. In the first stage, selenium leaches out of the NiSe2 phase. Losing the chalcogen opens up the nickel 3d band manifold, specifically broadening the antibonding eg band, the orbital set that overlaps with oxygen 2p states. A broader eg band strengthens nickel-oxygen covalency, and stronger covalency does two things at once: it accelerates the deprotonation of *OH intermediates, a known rate-limiting step on nickel oxyhydroxide surfaces, and it begins to tilt the electronic structure toward oxygen participation.

Then comes the second, decisive stage. Chromium begins to leach from the oxide component, and crucially, the chromium departure selectively depletes oxygen atoms that were coordinated to nickel across the composite interface. What remains is a modified nickel oxyhydroxide, SRA-NiOOH, in which oxygen non-bonding states, oxygen 2p orbitals stripped of their bonding partners, are generated not randomly across the lattice but in spatially confined pockets. This is the heart of the strategy. Confined non-bonding states act as local redox reservoirs for lattice oxygen, close enough to the surface to join the catalytic cycle, yet electronically quarantined so that oxidation holes cannot wander freely through the framework. Destructive redox delocalization, the killer mechanism of lattice-oxygen catalysts, is suppressed by construction. The oxygen-redox activity and the lattice stability, usually two ends of a seesaw, are decoupled.

The experimental evidence for this picture is layered across multiple structural and spectroscopic probes. X-ray absorption spectroscopy traces the changes in nickel coordination and oxidation state as reconstruction proceeds, while synchrotron-based measurements carried out at the Singapore Synchrotron Light Source document the emergence of the modified oxyhydroxide phase. Density functional theory calculations, performed with the ASTAR Computational Resource Centre and the National Supercomputing Centre in Singapore, complement the experiments, showing how the broadened eg* band and the confined oxygen non-bonding states reshape the energy landscape of the oxygen evolution mechanism. The team further demonstrates that the approach generalizes: the sequential-reconstruction activation paradigm is not a one-off trick of one composition but a design framework that can be exported to related systems, a claim their fourth figure addresses directly.

The device-level performance is where the work will command the most attention. Anion-exchange-membrane water electrolyzers are the commercial sweet spot for many of these nickel-based catalysts, because AEM technology can use earth-abundant transition metals instead of the iridium that proton-exchange-membrane electrolyzers demand. But AEM electrolyzers still need to run at high current density, above 2 amperes per square centimeter, to reach the cost targets that would make green hydrogen competitive with fossil-derived hydrogen. At 4.1 A cm-2 at 1.8 V in 1 M KOH at 80 degrees Celsius, the SRA-NiOOH anode sits squarely in industrially relevant territory, and the nearly 3,500-hour stability test under accelerated ageing at 60 degrees Celsius suggests the catalyst does not merely sprint; it runs a marathon. Sustained oxygen-redox operation at high current density, the authors argue, is precisely what the sequential electronic reconstruction buys: a durable oxygen-redox window that stays open under demanding polarization.

The conceptual contribution may outlast the specific numbers. For years, the field has treated catalyst reconstruction as something to minimize, a necessary evil on the path to stable operation, or at best a one-shot activation. This work reframes reconstruction as something to choreograph. By choosing a precursor whose constituents leach in a controlled order, selenium first, chromium second, the researchers wrote the electronic structure of the final catalyst as a timed sequence rather than a static composition. Each leaching event performed a specific electronic function, band broadening here, oxygen non-bonding state generation there, and the final NiOOH inherited the benefits of both without the instability that an uncontrolled surface transformation would bring. It is a programming philosophy applied to surfaces, and it echoes a broader shift in electrocatalysis toward operando-informed precatalyst design, in which the working electrode is understood as a dynamic, self-assembling entity rather than a fixed material.

There are, of course, caveats and open questions. The reported metrics come from laboratory-scale AEM cells and accelerated ageing protocols, and the road from such benchmarks to kilowatt-scale stacks involves membrane durability, flow uniformity, and cost questions that no single anode catalyst can answer alone. The precise atomic-scale geography of the confined non-bonding states, and how far the confinement holds as reconstruction continues over thousands of hours, will reward further study. But the bar the paper clears is a meaningful one: it shows that the activity-stability trade-off, long treated as a law of nature for lattice-oxygen catalysis, can be suspended by electronic design. If sequential reconstruction proves as general as the authors suggest, the oxygen evolution bottleneck that has constrained green hydrogen may finally have an engineering-grade exit, built not from a miracle material but from a catalyst taught to rebuild itself in the right order.

Subject of Research: Sequential electronic reconstruction of a Cr2O3/NiSe2 precursor to enable durable lattice-oxygen redox in anion-exchange-membrane water electrolysis

Article Title: Sequential electronic reconstruction enables durable oxygen redox in water electrolysis

Article References: Yu, J., Zhong, H., Zhang, Q., Zhang, X., Ye, Y., Diao, C., Yu, Z. G., Xi, S., Tian, B., Wang, X., & Xue, J. (2026). Sequential electronic reconstruction enables durable oxygen redox in water electrolysis. Nature Sustainability. https://doi.org/10.1038/s41893-026-01940-6

Image Credits: AI Generated

DOI: 10.1038/s41893-026-01940-6

Keywords: water electrolysis, oxygen evolution reaction, lattice oxygen redox, nickel oxyhydroxide, green hydrogen, anion-exchange membrane, electrocatalysis, catalyst reconstruction, eg band, catalyst stability, nickel selenide, Nature Sustainability