Hydrogen has long been touted as the clean fuel of the future, but the dirty secret of the hydrogen economy is how the gas is actually made. More than 99 percent of today’s supply comes from steam methane reforming and other fossil-fuel-dependent processes, releasing substantial carbon dioxide along the way. Water electrolysis offers a genuinely green alternative, splitting water into hydrogen and oxygen using renewable electricity. Yet the technology is haunted by a stubborn electrochemical bottleneck: the oxygen evolution reaction, or OER, the anodic half-reaction that gives up electrons so water can be dismantled. A comprehensive new review published in Advances in Industrial and Engineering Chemistry maps out how researchers are dismantling that bottleneck using cheap, earth-abundant 3d transition metals such as iron, cobalt, and nickel, and the picture it paints is one of a field racing toward industrial relevance.
The thermodynamics of water splitting look deceptively simple. Whether the electrolyte is acidic or alkaline, the minimum cell voltage needed to drive the reaction is 1.23 volts. In practice, electrolyzers demand considerably more, because the OER is a four-electron, multi-step process riddled with energy barriers, whereas the hydrogen evolution reaction at the cathode is a comparatively easy two-electron affair. That extra voltage, the overpotential, is pure waste heat and lost efficiency. In acidic systems, proton exchange membranes deliver low resistance and high gas purity, but the corrosive environment forces the use of scarce, expensive iridium and ruthenium oxides at the anode. Alkaline electrolysis, by contrast, opens the door to first-row transition metals that cost a fraction as much and can match, or even exceed, the performance of precious metal systems.
Understanding why the OER is so sluggish begins with its mechanism. The dominant framework, the adsorbate evolution mechanism, describes a sequence in which hydroxide species adsorb onto metal sites, convert into oxygen intermediates, and ultimately release molecular oxygen. Each step carries a Gibbs free energy cost, and the largest one sets the rate-determining step and thus the overpotential. The problem is that the adsorption energies of the intermediates are not independent. Because the *OOH and *OH species bind to the surface through a single oxygen atom, their energies scale linearly across virtually all oxide surfaces. The difference between the oxygen and hydroxide binding energies, roughly 3.2 electron volts on average, cannot be compressed below about 1.6 electron volts, while an ideal catalyst would need only 1.23. That mismatch imposes a theoretical overpotential floor of roughly 0.37 electron volts, and it turns catalyst design into a game of climbing a volcano-shaped activity plot, where the sweet spot lies at neither too strong nor too weak oxygen binding.
Some catalysts, however, refuse to obey the volcano. Their activity exceeds what the scaling relationship permits, and their kinetics shift with electrolyte pH in ways the adsorbate mechanism cannot explain. The resolution is the lattice oxygen mechanism, in which oxygen atoms belonging to the catalyst’s own crystal lattice participate directly in forming the O–O bond, bypassing the hydroperoxide intermediate entirely. The evidence is compelling: when researchers label lattice oxygen with the isotope oxygen-18 and run the reaction in an oxygen-16 electrolyte, differential electrochemical mass spectrometry detects mixed isotopic oxygen molecules, proof that lattice oxygen is being released as product gas. The mechanism is favored when the metal–oxygen bond is highly covalent, which is governed by the position of the oxygen 2p band relative to the metal 3d band. Raising the oxygen p-band center toward the Fermi level activates oxygen redox chemistry, promotes vacancy formation, and unlocks the lattice oxygen pathway, offering a route around the scaling-relation ceiling.
Armed with these mechanistic insights, catalyst designers have developed a toolkit of activity descriptors that predict performance without simulating every reaction step. Beyond the oxygen binding energy difference, the d-band center theory links adsorbate binding strength to the energy of the metal’s d electrons, while the occupancy of e_g orbitals in octahedrally coordinated transition metals provides a remarkably simple rule of thumb. Shao-Horn and colleagues showed that perovskite oxides perform best when e_g filling is near unity, with the perovskite Ba0.5Sr0.8Co0.2Fe0.2O3−δ sitting at the volcano apex. The oxygen p-band center itself has emerged as a descriptor that captures both activity and the instability that sometimes accompanies it. Perhaps the most striking demonstration of descriptor-guided design came from Sargent’s group, which gelled iron, cobalt, and tungsten into an atomically homogeneous oxyhydroxide. Iron binds oxygen too strongly, cobalt and tungsten too weakly, but interpolation among the three achieved near-optimal binding, delivering a record-low overpotential of 191 millivolts at 10 milliamperes per square centimeter and 500 hours of stable operation.
One of the field’s most consequential realizations is that many catalysts are not what they appear to be. Sulfides, selenides, nitrides, and phosphides that show impressive alkaline OER activity are frequently oxidized in situ under the high anodic potentials of operation, transforming into metal (oxy)hydroxides that are the true active phases. The as-synthesized materials are, in effect, precatalysts. Operando and in situ characterization techniques have confirmed that most transition metal cations undergo surface reconstruction during the OER, typically involving rising oxidation states and the leaching of specific cations or anions. Rather than fighting this transformation, researchers have learned to choreograph it. Nickel substitution in zinc cobaltate spinels pushes the oxygen p-band above the metal d-band, triggering reconstruction into an active oxyhydroxide surface and a sevenfold activity boost that surpasses the iridium oxide benchmark. Iron substitution in otherwise inert cobalt aluminate initiates reconstruction that then self-terminates, preserving the underlying structure.
Crystallinity matters too. Amorphous materials, with their abundant defects and open structures, reconstruct far more readily than their crystalline counterparts. An amorphous NiFeMo oxide prepared by fast co-precipitation rebuilt itself into an oxygen-vacancy-rich oxyhydroxide layer within moments of the reaction starting, requiring only 280 millivolts of overpotential, while the crystalline version needed 358 millivolts because molybdenum ions diffuse sluggishly through the ordered lattice. Deliberate leaching has become a design strategy in its own right. Activating an inactive cobalt chromite spinel at high potential oxidizes chromium from the 4+ to the 6+ state, driving it out of the lattice and leaving behind defects and vacancies that let cobalt reconstruct into active oxyhydroxide, an activity gain of more than an order of magnitude. In another approach, molybdenum sulfide nanosheets coated with nickel-iron layered double hydroxide served as sacrificial templates; five cycles of voltammetry stripped the molybdenum into the growing oxyhydroxide, yielding a catalyst that needed just 242 millivolts at 10 milliamperes per square centimeter.
The review’s survey of recent multi-metal catalysts reveals how these principles converge in practice. Tensile lattice strain engineered into NiFe2O4 spinel ferrites flattened the electronic bands near the Fermi level and tuned e_g occupancy at iron sites, cutting the overpotential to 180 millivolts. Co-doping ruthenium and zinc into Co3O4 selectively activated lattice oxygen while preventing over-oxidation of ruthenium, achieving 172 millivolts and 100 hours of stability at a punishing 500 milliamperes per square centimeter. Vanadium leaching from Co2Fe0.25V0.75O4 left behind a stable vanadium oxide underlayer that compensated charge and prevented cobalt dissolution, sustaining 600 hours at 500 milliamperes per square centimeter. High-valence dopants such as molybdenum, chromium, niobium, and tungsten stabilize low-valence iron and nickel species, and the resulting Fe2+ fraction correlates directly with activity; NiFeMo oxyhydroxide reached 180 millivolts. Anchoring single molybdenum atoms onto NiFe oxyhydroxide nanoarrays strengthened metal–oxygen covalency and activated lattice oxygen, delivering 193 millivolts. Non-oxide systems joined the race as well: cobalt-nickel phosphide nanosheets hit 209 millivolts, vacancy-rich NiFe sulfide reached 213 millivolts and performed at 235 millivolts under quasi-industrial conditions of 6 molar potassium hydroxide at 60 degrees Celsius, and a hierarchical CoMnP/Ni2P array worked as a bifunctional catalyst for both half-reactions, splitting water at just 1.54 volts.
The newest frontier is high-entropy materials, in which five or more elements mix in near-equal proportions to form single-phase solid solutions. The compositional disorder stabilizes the phase, and the so-called cocktail effect of synergistic elemental interactions tunes adsorption energies toward the Sabatier optimum. A high-entropy sulfide of chromium, manganese, iron, cobalt, and nickel outperformed every unary through quaternary sulfide comparison, reaching 100 milliamperes per square centimeter at 295 millivolts, with charge migration from manganese and iron optimizing the cobalt active sites. High-entropy perovskites and spinels followed suit, and in the spinel (CrFeCoNiMo)3O4 the high-valence cobalt and molybdenum shifted the reaction pathway toward the lattice oxygen mechanism, verified by isotope labeling. An amorphous NiFeCoMnAl oxide, dealloyed into a nanoporous architecture, posted 190 millivolts. The authors of the review are candid about what remains: the synergistic interactions among multiple elements are still incompletely understood, stability at industrial current densities is not guaranteed, and bridging from laboratory electrodes to scalable electrolyzer architectures is the decisive next step. But the trajectory is unmistakable. By combining electronic structure engineering, dynamic surface reconstruction, and compositional complexity, researchers are turning the OER from the weakest link in electrochemistry into a solvable design problem, and with it, the prospect of hydrogen made cleanly, cheaply, and at scale moves measurably closer.
Subject of Research: Design strategies and recent advances in 3d transition metal-based multi-metal electrocatalysts for the alkaline oxygen evolution reaction
Article Title: An overview of design strategies and recent advancements in complex 3d transition metal-based electrocatalysts for alkaline oxygen evolution reaction
Article References: Lee, S., Shin, Y., Yeom, K., Shim, J., & Sung, Y.-E. (2025). An overview of design strategies and recent advancements in complex 3d transition metal-based electrocatalysts for alkaline oxygen evolution reaction. Advances in Industrial and Engineering Chemistry, 1(1), Article 2. https://doi.org/10.1007/s44405-025-00001-4
Image Credits: AI Generated
DOI: 10.1007/s44405-025-00001-4
Keywords: oxygen evolution reaction, electrocatalysis, water electrolysis, green hydrogen, transition metal catalysts, surface reconstruction, lattice oxygen mechanism, activity descriptors, high-entropy materials, layered double hydroxides, alkaline media, multi-metal catalysts
Bethany Barker. (October 4, 2026). Earth-Abundant Metal Catalysts Close In on the Oxygen Evolution Bottleneck. Scienmag.
