The study, published in Discover Electrochemistry by Prince Gollapalli, Santu Biswas, Leeor Meir, Nofar Refaeli and Maytal Caspary Toroker, tackles a question that has lingered because of an experimental catch. The catalysis does not happen on the stable nickel hydroxide phase that is easy to characterize in the lab. Instead, water oxidation occurs on a metastable nickel oxyhydroxide (β-NiOOH) surface that forms only under operating potentials of roughly 1.3 to 1.5 volts versus the reversible hydrogen electrode, when the material sheds protons and nickel transitions from the +2 to the +3 oxidation state. Capturing that fleeting phase at the atomic scale requires demanding in situ experiments, so most studies have settled for characterizing the stable precursor. Density functional theory (DFT) offers a way around the problem, letting researchers build atomistic models of the active phase and watch the reaction unfold electron by electron.

To make the model as realistic as possible, the team anchored it in experiment. Transmission electron microscopy studies of nickel hydroxide consistently reveal hexagonal, plate-like nanoparticles, so the researchers carved a hexagonal nanocrystal out of the unit cell of staggered β-NiOOH, a structural variant predicted to be stable in earlier first-principles work. The cluster measures roughly one nanometer across in width and length, closely matching the approximately 2.5-nanometer particles synthesized and imaged in a combined experimental and theoretical study the team used as a benchmark. They then built two versions of the particle: a three-layer stack about 1.35 nanometers tall, and a single layer just 0.38 nanometers thick. Crucially, the top layer is identical in both, so any difference in catalytic performance could be attributed purely to what lies beneath.

On that shared top surface, the researchers identified three chemically distinct places where the reaction could begin. At the edge of the particle, an undercoordinated nickel atom bonded to only five oxygens offers a natural docking point for water, with no vacancy needed. In the interior, the surface alternates between oxygen-terminated and hydroxyl-terminated triangular sites, and the team created vacancies at each, removing an oxygen atom in one case and an oxygen plus a hydrogen atom in the other, to open adsorption sites. This choice of locations matters because the conventional wisdom in nanocatalysis holds that edges are the hot spots of reactivity. The simulations were set to test whether that wisdom survives on a particle this small.

The answer, at least for thermodynamics, is a resounding no. Running the full four-step OER mechanism, in which water adsorbs and then loses four proton-electron pairs through *OH2, *OH, *O and *OOH intermediates, the team computed Gibbs free energies for every step on every site of both particle sizes. The overpotential, the extra voltage beyond thermodynamic requirement needed to make every step proceed downhill, dropped consistently when the particle was thinned from three layers to one. On the hydroxyl-terminated site, the overpotential fell from 0.78 volts to 0.57 volts; on the oxygen-terminated site, from 0.98 to 0.75 volts; and on the edge site, from 1.62 to 1.11 volts. The single-layer hydroxyl-terminated configuration, at 0.57 volts, matched or beat values reported in the literature for extended surfaces of the same material.

Two mechanisms appear to explain the advantage of going thin. First, shrinking the particle raises the catalytically active surface area per unit volume, a straightforward geometric effect. Second, and more subtly, a single layer has no neighboring layer above it for protons to migrate into or to hydrogen-bond with. Previous work had shown that hydrogen desorption is easier by about half an electron-volt in monolayers precisely because interlayer hydrogen bonding is absent, and that adding protons to β-NiOOH is otherwise spontaneous. Without a second layer to hold protons back, deprotonation steps become easier, and the tallest energy barrier on the reaction coordinate shrinks. The team confirmed this with an implicit solvation model, which raised the best-case overpotential only modestly, by 0.06 to 0.1 volts, leaving the single-layer advantage intact.

The identity of the potential determining step, the slowest rung on the energetic ladder, also depended on where the reaction happened. At the edge site, where the reacting intermediate touches only a single nickel atom, the hardest step was the second deprotonation, converting *OH to *O. At both interior sites, where three nickel atoms surround the adsorbate, the bottleneck shifted to the third deprotonation, from *O to *OOH, and the overpotential dropped by roughly a factor of two. In every configuration, the *O intermediate turned out to be the pivotal species, appearing either as the product or the reactant of the rate-limiting step, underscoring its central role in limiting oxygen evolution on this catalyst.

Because nickel can access multiple oxidation states, the team could use magnetism as a fingerprint of electronic change. Assigning magnetic moments of roughly 1.7, 1.2 and 0.2 Bohr magnetons to nickel in the +2, +3 and +4 states, they found that at the edge site only one nickel atom oxidizes from +3 to +4 during the reaction, while at the oxygen-terminated site none do, and at the hydroxyl-terminated site two of the three supporting nickel atoms make the jump to +4. Robustness checks with different values of the DFT+U correction parameter and with the hybrid HSE06 functional confirmed the assignments. The hydroxyl-terminated site, with its more oxidized nickel atoms, sits closest to the highly oxidized state the material is expected to adopt during real oxygen evolution, marking it as the genuinely active reaction center.

The final piece was dynamics rather than thermodynamics. Each reaction step ejects a proton-electron pair, leaving behind a hole in the valence band that must travel to the active site to drive oxidation. The team propagated quantum wave packets through the electrostatic potential landscapes of each intermediate, using a split-operator algorithm to solve the time-dependent Schrödinger equation on a femtosecond timescale, and measured the cumulative hole flux crossing a barrier near the adsorbate. The results tracked the thermodynamics beautifully: the edge site showed the lowest hole flux and the highest electrostatic barriers, while the hydroxyl-terminated site showed the highest flux, with a cumulative probability of 0.83 for its potential determining reactant compared with just 0.37 at the edge. More nickel atoms sharing the load and a more oxidized surface mean more holes available to do chemistry.

Taken together, the study rewrites a piece of nanocatalyst intuition. On β-NiOOH nanoparticles, the bulk-like hydroxyl-terminated site outperforms the edge, contradicting the general rule that edges are more active, and the thinnest possible particle delivers the lowest overpotential. For engineers designing anode materials for green hydrogen electrolyzers, the prescription is concrete: maximize surface-to-volume ratio by exfoliating the catalyst toward single layers, and engineer hydroxyl-terminated, vacancy-rich surfaces where multiple nickel atoms can participate and change oxidation state. The work also suggests magnetism itself could serve as a control handle, since spin-polarized band structures mean the number of available states at the Fermi level differs for spin-up and spin-down carriers. As the hydrogen economy scales up, such atomistic design rules, extracted from simulations that can finally see the metastable phase where the real chemistry happens, may prove as valuable as any new material discovery.

Subject of Research: Computational study of how nanoparticle size and active site location govern the thermodynamics and hole transport dynamics of β-NiOOH catalysts for the oxygen evolution reaction

Article Title: Effect of nanoparticle size and active site position on the thermodynamics and charge dynamics of NiOOH catalyst for oxygen evolution reaction

Article References: Gollapalli, P., Biswas, S., Meir, L., Refaeli, N., & Caspary Toroker, M. (2026). Effect of nanoparticle size and active site position on the thermodynamics and charge dynamics of NiOOH catalyst for oxygen evolution reaction. Discover Electrochemistry, 3(1), Article 50. https://doi.org/10.1007/s44373-026-00135-0

Image Credits: AI Generated

DOI: 10.1007/s44373-026-00135-0

Keywords: oxygen evolution reaction, NiOOH, water splitting, green hydrogen, density functional theory, electrocatalysis, nanoparticles, overpotential, hole transport, oxidation state, nickel oxyhydroxide, DFT+U

The team approached the problem with a combination of in situ surface-enhanced Raman spectroscopy and density functional theory calculations. Surface-enhanced Raman spectroscopy, or SERS, amplifies the ordinarily faint Raman scattering from molecules and structures close to a nanostructured metal surface, making it one of the few techniques capable of probing an electrode while it is actually operating in electrolyte. This operational capability is critical because catalyst surfaces are dynamic: they restructure, protonate, and oxidize in response to the applied potential, and a structure observed ex situ, after the current has been switched off and the electrode removed from solution, may bear little resemblance to the one doing the catalysis in the moment.

Under alkaline water oxidation, nickel-based anodes accumulate positive charge as the potential is raised. The conventional narrative held that the octahedral layers of nickel hydroxide, Ni(OH)2, are oxidized first to the trivalent nickel oxyhydroxide NiOOH, and that this phase hosts the highest-valence intermediates responsible for extracting electrons and protons from water. Raman spectra recorded under operating conditions appeared to support this picture: characteristic bands emerged at elevated potentials and were assigned to superoxide-like oxygen species or to vibrational fingerprints of the NiOOH lattice itself. The new work demonstrates that this assignment rested on an overlooked spectroscopic subtlety, the contribution of overtone and combination modes.

In Raman spectroscopy, the dominant, or Stokes, signal arises from photons that lose energy equal to a single vibrational quantum of the material, producing bands at fundamental vibrational frequencies. But weaker features also appear at roughly twice a fundamental frequency, from two-phonon overtone scattering, and at sums or differences of two different fundamentals, from combination modes. These second-order processes are intrinsically less intense than fundamentals, yet under surface-enhancement conditions the signal from a single atomic layer can be amplified dramatically, and overtone and combination features can become prominent enough to be mistaken for first-order bands from entirely different species. That, according to the study, is precisely what happened on nickel anodes.

By comparing measured spectra with computed vibrational spectra for candidate surface structures, the researchers found that the bands previously attributed to superoxide species, or to NiOOH, could instead be explained as overtones and combination bands of a nickel dioxide surface termination. In other words, the spectral signature that seemed to confirm the textbook NiOOH picture was a harmonic echo of the true structure: an NiO2-terminated surface in which nickel reaches a formal oxidation state above that of NiOOH. When the overtones were properly accounted for, the calculations converged on NiO2 as the structure consistent with the spectra at the potentials where water oxidation actually proceeds.

The implications for electrocatalysis are substantial. Nickel anodes are central to alkaline water electrolysis, a technology positioned as a cornerstone of green hydrogen production, and they also appear as key components in nickel-iron and nickel oxyhydroxide composite catalysts. Mechanistic models, from the oxygen evolution reaction pathway to the role of iron impurities in boosting activity, have been built on the premise that NiOOH is the operative phase. If the working surface is instead NiO2-terminated, then the elementary steps of water activation, the identity of the rate-limiting intermediate, and the way iron or other dopants modify activity may all need to be re-examined through a different structural lens.

The finding also carries a broader methodological lesson for the field of operando catalysis research. Spectroscopic identification of catalyst phases under reaction conditions is the gold standard, but spectra are interpretations, not photographs. The danger highlighted here, that second-order scattering can masquerade as fundamental modes of a different compound, is not unique to nickel. Other transition-metal oxides and hydroxides with strong lattice vibrations could harbor similar spectral ambiguities, particularly where surface enhancement or resonance effects amplify weak overtone features. The study suggests that confident phase assignment under operating conditions should routinely include comparison against computed spectra for candidate structures, and that candidate lists should include high-valence terminations that past convention may have excluded prematurely.

The synergy between experiment and theory was essential to the result. Density functional theory allows researchers to predict the vibrational spectrum of a hypothetical surface structure with atomic precision, including not only the fundamental modes but also the intensities and positions of overtones and combination bands. When the computed spectra of NiO2-terminated surfaces were laid against the in situ SERS data, the mismatch that plagued NiOOH-based assignments resolved. This iterative loop, in which measured bands constrain the candidate structures and computed spectra test each candidate, offers a template for revisiting other contested catalyst surfaces, from cobalt and iron oxides to mixed-metal systems used in oxygen evolution.

There is also a practical dimension to the discovery. Understanding the true surface structure is a prerequisite for rational catalyst design. Electrolyzer developers seeking to lower the overpotential of the oxygen evolution reaction, one of the major efficiency losses in hydrogen production, rely on structural knowledge to tune composition, morphology, and doping strategies. A surface terminated in NiO2 changes the picture of where and how the oxygen-oxygen bond forms, which in turn informs which compositional modifications are likely to stabilize the most active configurations. The revised assignment may thus redirect experimental efforts that were optimized against an imperfect structural target.

For a field that has invested decades in the NiOOH framework, the result is a reminder that foundational assumptions deserve periodic challenge, especially as instrumentation and computational methods grow more powerful. The active surface of a nickel anode during alkaline water oxidation, once thought settled, is now understood to be a nickel dioxide termination whose spectral fingerprint was hiding in plain sight, doubled in frequency and blended into features everyone thought they had already explained. As the community absorbs the finding, both the mechanistic models of oxygen evolution and the spectroscopic conventions used to build them are likely to change together.

The reinterpretation also speaks to the peculiar electrochemistry of nickel in alkaline media. Nickel hydroxide electrodes exhibit well-defined redox features in cyclic voltammetry, and the potential region associated with the Ni(II)/Ni(III) transition has long served as a convenient marker for when the oxyhydroxide phase should form. The new assignment implies that the charge stored beyond that conventional transition does not simply stop at trivalent nickel but continues to be accommodated by the surface lattice, consistent with a termination whose formal nickel oxidation state exceeds that of NiOOH. This helps rationalize observations that have puzzled researchers for years, including the strong potential dependence of Raman features in the high-potential regime and the apparent persistence of catalytic activity at potentials where NiOOH alone seemed an insufficient electron sink.

Surface-enhanced Raman spectroscopy itself has a history intertwined with nickel electrochemistry. Early applications to corroding and battery-related nickel surfaces established many of the band positions that later became embedded in the literature, and those assignments propagated through successive studies, sometimes with little independent verification. The present work illustrates how citation chains can entrench a misassignment: once a band position is labeled in an influential early paper, subsequent authors may adopt the label rather than rederive it, and the error compounds quietly across decades until a systematic re-examination, armed with modern computational power, forces a correction.

The overtone problem is particularly insidious because the frequencies of second-order features are not arbitrary. An overtone appears near twice a fundamental frequency, and a combination band near the sum of two fundamentals, so a misassigned overtone can land almost exactly where a plausible first-order mode of a different species is expected. On nickel anodes, lattice vibrations of the oxidized surface are energetic enough that their doubled frequencies fall in the same spectral window as oxygen-oxygen stretching modes of superoxide-like species, which is why the misreading was so persuasive. Only by computing the full second-order spectrum of candidate structures, rather than just their fundamentals, could the true origin of the bands be pinned down.

For experimentalists planning operando studies, the result argues for a more cautious use of single-technique identification. Complementary probes such as X-ray absorption spectroscopy, which is sensitive to oxidation state and coordination geometry rather than to vibrational selection rules, can provide independent constraints on surface structure and help disambiguate cases where Raman features admit multiple interpretations. Combining such measurements with isotopic labeling, for instance using H2-18O electrolyte to test whether a band shifts as expected for an oxygen-oxygen stretch, offers a practical route to catching overtone masquerades before they harden into consensus assignments.

Finally, the finding arrives at a moment when alkaline water electrolysis is scaling rapidly, and when even small improvements in anode understanding could translate into meaningful efficiency gains at industrial current densities. If the working surface of nickel anodes is a high-valence dioxide termination, then stability against dissolution, reconstruction, and degradation must be assessed for that phase, not for NiOOH, and accelerated stress tests may need reinterpretation as well. The correction thus ripples outward from spectroscopy into the engineering practice of a technology central to the hydrogen economy.

Subject of Research: Reassignment of Raman overtone bands indicating a NiO2-terminated, rather than NiOOH, active surface on nickel anodes during alkaline water electrolysis.

Article Title: Consequences of overtones in Raman spectra for assigning nickel anode surface structures as NiO2 during alkaline electrolysis

Article References: Leist, J., Neufischer, A., Jacob, T., & Engstfeld, A. K. (2026). Consequences of overtones in Raman spectra for assigning nickel anode surface structures as NiO2 during alkaline electrolysis. Nature Catalysis. https://doi.org/10.1038/s41929-026-01613-9

Image Credits: AI Generated

DOI: 10.1038/s41929-026-01613-9

Keywords: nickel anode, Raman spectroscopy, NiO2, NiOOH, water oxidation, alkaline electrolysis, oxygen evolution reaction, surface-enhanced Raman spectroscopy, density functional theory, electrocatalysis, green hydrogen, catalyst surface structure