Ordered platinum intermetallics push fuel cell catalysts from binary alloys to high-entropy designs

Fuel cells promise clean power from hydrogen, yet their efficiency has long been throttled by a single stubborn chemical step: the oxygen reduction reaction, in which oxygen molecules at the cathode must be stripped of four electrons and converted to water. This reaction is notoriously sluggish, and the platinum catalysts that force it along are scarce, expensive, and prone to dissolving, clumping and poisoning during operation. A comprehensive new review published in the journal Catal examines how a family of materials called platinum-based intermetallics has steadily chipped away at these limits, evolving from simple two-metal compounds to ternary systems and now to high-entropy structures containing five or more elements in a single ordered crystal lattice.

The review, led by Changlin Li and colleagues, frames the oxygen reduction reaction as the rate-limiting half-reaction in proton exchange membrane fuel cells, the technology favored for its high energy density, rapid start-up, quiet operation and low-temperature performance. In an acidic environment, the four-electron pathway is preferred because the alternative two-electron route produces hydrogen peroxide, a corrosive species that degrades catalysts. Whether oxygen is reduced through an associative mechanism, in which adsorbed oxygen molecules are hydrogenated step by step before the oxygen-oxygen bond finally breaks, or through a dissociative mechanism that cleaves the bond first, the reaction proceeds through adsorbed intermediates known as *OOH, *O and *OH. Because these species bind to catalyst surfaces in ways that are mathematically linked, strengthening one adsorption inevitably shifts the others, a constraint known as the linear scaling relationship that caps how good any single catalyst can be.

Conventional platinum alloys, in which transition metals such as nickel, cobalt, iron or copper are randomly mixed into the platinum lattice, were an early answer to the cost problem. Alloying tunes the electronic structure of platinum, weakening the grip on oxygen intermediates and lowering the energy barrier. The celebrated Pt3Ni(111) surface, for example, achieved an oxygen reduction activity ten times that of pure platinum(111). But random atomic arrangements come with a price: heterogeneous active sites with uneven adsorption strengths, and transition metals that leach out under the highly acidic, high-potential conditions inside a fuel cell, contaminating the membrane and steadily eroding performance.

Intermetallics take a fundamentally different approach. Rather than scattering atoms at random, they adopt long-range ordered crystal structures with well-defined stoichiometry, most commonly the L10 and L12 superlattice configurations. This atomic order anchors transition metal atoms in place, suppressing their dissolution and migration, while the precise arrangement of atoms offers a platform for systematically optimizing the platinum d-band center, the electronic feature that governs how strongly intermediates bind. Ordered L10-PtCo, L10-PtNi and L10-PtFe catalysts have demonstrated mass and specific activities superior to their disordered counterparts, with stability enhanced by orders of magnitude. The review emphasizes that this determinism also makes intermetallics ideal for connecting experimental observations with theoretical simulations, because the surface composition is no longer a statistical guess.

Getting atoms to order, however, is kinetically demanding. The disorder-to-order transformation requires atomic diffusion, which in practice means high-temperature annealing, often 600 degrees Celsius or higher for L12-type structures. Such heat sintering nanoparticles into larger, less active particles. One strategy pairs the annealing with strong metal-support interactions, using supports doped with sulfur, boron or nitrogen to bond tightly to the metal nanoparticles and hold them in place even at temperatures approaching 1000 degrees Celsius. Using sulfur-doped carbon, one team built a library of 46 platinum-based intermetallics and found that oxygen reduction activity rises monotonically with compressive strain across a wide range. A platinum-cobalt catalyst made this way, loaded at an ultralow 0.02 milligrams of platinum per square centimeter, matched the performance of a conventional platinum-carbon cathode carrying more than ten times the platinum, reaching a maximum power density of 1.08 watts per square centimeter in a hydrogen-air fuel cell.

Ordering degree itself has emerged as a tunable dial. Carbon-supported platinum-iron catalysts prepared with ordering degrees from 10 to 70 percent showed a clear positive correlation between order and both activity and durability, with the most ordered sample, at 74.3 percent, delivering a mass activity of 2.12 amperes per milligram of platinum, fifteen times that of commercial platinum-carbon, and retaining 76 percent of its activity after 30,000 potential cycles. At the other end of the size spectrum, shrinking particles below two nanometers makes ordering harder because surface energy climbs and atomic mobility drops. Amino-tethering and hydrogen-adsorption-assisted bond-weakening strategies have now produced ordered intermetallics as small as 1.25 nanometers, and a seed-densification approach pushed ordered Pt3Co nanocrystals down to 0.69 nanometers, all while preserving catalytic performance under ultralow platinum loadings.

Adding a third metal opens further levers. Ternary intermetallics retain the crystallographic fingerprints of their binary parents but shift the peaks slightly, reflecting subtle changes in electronic structure and lattice strain. In L10-CoNiPt, nickel-induced anisotropic compression shortened surface platinum-platinum bonds from 2.78 to 2.69 angstroms, weakening oxygen adsorption and yielding a mass activity of 3.1 amperes per milligram of platinum with only 15.9 percent decay after 30,000 cycles. Low-melting-point metals such as gallium, indium, tin and copper can migrate to the surface during annealing, leaving vacancies that accelerate ordering and allow synthesis at lower temperatures. Incorporating gallium into platinum-cobalt raised the ordering degree from roughly 17 to 65 percent, and the resulting catalyst delivered 1.07 amperes per milligram of platinum in a hydrogen-oxygen fuel cell, retaining 82.2 percent of its initial activity after accelerated durability testing.

High-entropy intermetallics represent the newest frontier. Defined as ordered compounds containing five or more elements in near-equimolar proportions, they combine the configurational entropy and sluggish diffusion effects of high-entropy alloys with the structural orderliness of intermetallics. Machine learning is proving essential to navigate the vast compositional space: one team used a crystal graph convolutional neural network trained on 538 sets of first-principles data to predict surface strain and formation energy with mean absolute errors of 0.003 and 0.011 electron volts per atom, guiding the synthesis of a low-platinum Pt-(FeCoNiCu)3 catalyst that lost only 11 millivolts of half-wave potential after 5,000 accelerated cycles. In ultrasmall PtFeCoNiCuZn nanoparticles, electronic effects shifted the platinum d-band center downward and activated non-noble metals as additional active sites, achieving a mass activity of 2.403 amperes per milligram of platinum, nineteen times that of commercial platinum-carbon.

Defects, too, are becoming design tools. A 1.7-nanometer PtZnFeCoNiCr high-entropy intermetallic with deliberate platinum antisite defects, in which oversized platinum atoms became trapped on base-metal lattice sites during annealing, introduced compressive strain that refined the electronic structure and activated zinc sites for synergistic catalysis. Another L10-type PtIrFeCoCu system converted the traditionally low-activity (001) crystal facet into an ultra-high-activity one by shifting electron transfer from iron, cobalt and copper toward platinum and iridium, minimizing the activation barrier of the rate-limiting step. Under heavy-duty vehicle conditions, a platinum-cobalt-nickel-iron-copper high-entropy intermetallic with a monolayer platinum shell maintained a current density of 1388 milliamperes per square centimeter at 0.7 volts even after 90,000 cycles, its stability reinforced by transition-metal-nitrogen bonds that pin the base metals in place.

The review closes by identifying the hurdles that stand between these laboratory successes and commercial fuel cells. Scalable synthesis with precise control over particle size, morphology and ordering degree remains difficult, and the authors call for operando X-ray absorption spectroscopy and identical-location electron microscopy to watch catalysts evolve under real operating conditions. Machine learning must move beyond composition screening toward predicting degradation pathways, and high-entropy designs must be validated in full membrane electrode assemblies under realistic swings of temperature, humidity and load. If those challenges are met, the authors argue, ordered platinum intermetallics could underpin a new generation of low-platinum, high-performance fuel cells, turning one of electrochemistry’s most stubborn bottlenecks into a solved problem.

Subject of Research: Ordered platinum-based intermetallic electrocatalysts for the oxygen reduction reaction in fuel cells

Article Title: Platinum-based intermetallics for oxygen reduction catalysis: from binary to high-entropy

Article References: Li, C., An, X., Qi, Y., Wang, X., Zhang, T., Liu, Y., He, L., Wu, F., Yang, W., Li, M., & Yu, Y. (2025). Platinum-based intermetallics for oxygen reduction catalysis: from binary to high-entropy. Catal, 1(1), Article 11. https://doi.org/10.1007/s44422-025-00011-9

Image Credits: AI Generated

DOI: 10.1007/s44422-025-00011-9

Keywords: platinum, intermetallics, oxygen reduction reaction, fuel cells, electrocatalysis, high-entropy materials, proton exchange membrane fuel cells, catalyst stability, core-shell structures, machine learning, strain engineering, membrane electrode assemblies