A newly engineered carbon structure keeps tiny platinum-cobalt catalysts stable under extreme conditions, potentially helping fuel cells deliver cleaner, longer-lasting power for data centers, transportation, and other demanding applications.
The rapid growth of data centers is creating an enormous new appetite for electricity. These facilities require continuous power not only for computing but also for cooling and other supporting infrastructure.
The Electric Power Research Institute has estimated that U.S. data centers could consume as much as 9% of annual electricity generation by 2030, compared with about 4% of total electricity demand in 2023. Researchers are therefore searching for technologies that could generate reliable power closer to where it is needed and reduce some of the pressure on electrical grids.
A team led by Gang Wu at Washington University in St. Louis has now developed a nanostructured catalyst support that could improve the performance and durability of low-temperature fuel cells.
The design uses hollow carbon spheres filled with carefully organized radial nanochannels. These tiny channels confine platinum-cobalt nanoparticles, allowing the catalyst to withstand extremely high manufacturing temperatures without the particles clumping together and losing useful surface area.
The study, published in Nature Nanotechnology on August 6, 2026, involved researchers from Washington University, Brookhaven National Laboratory, Lawrence Berkeley National Laboratory, Northeastern University, and the University of Pittsburgh.
Why data centers are putting pressure on the grid
Modern data centers consume huge amounts of electricity because thousands of processors, storage systems, networking devices, and cooling systems may operate continuously.
The expansion of artificial intelligence and cloud computing is increasing that demand further, making access to dependable electricity an increasingly important constraint on where new facilities can be built.
Fuel cells offer one possible way for some facilities to generate electricity on-site.
Rather than producing electricity through combustion, a hydrogen fuel cell electrochemically combines hydrogen and oxygen. The process produces electricity along with water and heat.
Wu argues that if a data center could supply part of its electricity through fuel cells, the technology could reduce its direct dependence on the surrounding electrical grid.
But making fuel cells sufficiently efficient, durable, and economical remains a major engineering challenge.
Platinum is powerful โ but expensive
One of the most important components inside a proton-exchange-membrane fuel cell is the catalyst that accelerates the oxygen reduction reaction at the cathode.
Platinum is exceptionally effective for this job.
Unfortunately, it is also a precious metal.
Fuel-cell developers therefore want to obtain as much catalytic performance as possible from very small amounts of platinum.
A common strategy is to divide the metal into nanoparticles.
Nanoparticles expose far more surface area relative to their mass than a solid piece of platinum. Since catalytic reactions take place primarily at exposed surfaces, reducing the particle size allows more of the precious metal to participate directly in the reaction.
But extremely small platinum particles introduce another problem.
During prolonged fuel-cell operation, nanoparticles can dissolve, migrate, merge, or grow. As their structure changes, the catalyst can gradually lose activity.
That creates one of the central trade-offs in fuel-cell catalyst design: smaller particles provide high catalytic surface area, while larger or more strongly stabilized particles tend to survive longer.
Platinum-cobalt catalysts offer another advantage
Researchers have increasingly investigated ordered platinum-based intermetallic compounds as an alternative to conventional platinum alloys.
In these materials, platinum atoms and atoms of another metal occupy highly organized positions in the crystal structure.
Wu's team focused on platinum-cobalt, or PtCo.
A highly ordered PtCo intermetallic structure can provide strong catalytic performance and improved stability, but producing that atomic order requires heating the material to high temperatures.
And that creates another dilemma.
High-temperature annealing encourages platinum and cobalt atoms to arrange themselves into the desired ordered structure. At the same time, however, intense heat encourages nanoparticles to move and fuse together.
The catalyst may therefore gain atomic order while losing the tiny particle size needed for efficient platinum utilization.
Keeping both properties at once has been difficult.
Tiny radial channels solve the heat problem
The researchers designed a new support known as a radial nanochannel-array carbon sphere.
The structure resembles a hollow carbon sphere filled with mesoscopic channels extending through it in an organized radial arrangement.
These channels act as nanoscale compartments for the platinum-cobalt particles.
Because the nanoparticles are physically confined inside the carbon structure, they cannot move and merge as easily when the material is heated.
That allowed the team to anneal the catalyst at temperatures above 1,000ยฐC while still maintaining platinum-cobalt particles smaller than 5 nanometers.
At the same time, the intense heating produced a highly ordered platinum-cobalt intermetallic phase.
The Nature Nanotechnology study reports an ordering degree above 80% while maintaining particles below 5 nanometers even when platinum represented 40% of the catalyst's weight.
In other words, the carbon channels helped overcome the usual conflict between atomic ordering and nanoparticle size.
Why atomic order matters
An intermetallic catalyst is not simply a random mixture of two metals.
Its performance depends strongly on where different atoms sit relative to one another.
When platinum and cobalt become highly ordered, their electronic and structural interactions can change the way oxygen-containing molecules interact with the catalyst surface.
The goal is to create a structure that accelerates the oxygen reduction reaction efficiently while also resisting degradation during extended operation.
High-temperature processing helps produce this ordered arrangement.
Without an effective support, however, that same heat can destroy the nanoscale structure that makes the catalyst useful.
The radial nanochannels effectively provide tiny cages that allow atomic reorganization within individual particles while limiting large-scale particle migration.
The catalyst survived 150,000 severe voltage cycles
Durability testing produced one of the study's most important results.
The researchers subjected the catalyst to an accelerated stress test involving 150,000 voltage cycles, designed to simulate severe degradation conditions.
According to the peer-reviewed study, the catalyst retained 82.5% of its performance after the test.
Washington University described the result as roughly comparable to about 25,000 hours of operation, although real-world lifetime will ultimately depend on how fuel cells are operated and integrated into practical systems.
The catalyst also delivered a current density of 2.12 amperes per square centimeter at 0.70 volts under testing conditions designed for heavy-duty vehicle applications.
Together, those results suggest that the carbon architecture can support both high catalytic activity and long-term stability โ two properties that have traditionally been difficult to maximize simultaneously.
Open pores improve movement through the electrode
The new support does more than stop nanoparticles from clumping.
Its pore architecture also helps materials move through the fuel-cell electrode.
A working fuel cell depends on the efficient transport of oxygen, protons, water, and electrically charged species through a complicated network of catalyst particles, carbon, and ion-conducting material.
If pores are too small or poorly connected, reactants may struggle to reach active catalyst sites. Water can also accumulate and interfere with performance.
The radial carbon structure contains open mesopores with carefully controlled dimensions and volume.
According to the researchers, these channels promote more favorable distribution of the ionomer โ the ion-conducting material within the electrode โ while improving pathways for oxygen, protons, and water.
The design therefore addresses both catalyst stability and the movement of materials necessary for efficient electrochemical reactions.
A difficult four-way catalyst challenge
The study describes fuel-cell catalyst design as requiring several desirable properties simultaneously.
Researchers want very small metal nanoparticles to maximize useful surface area.
They want a highly ordered intermetallic structure for activity and durability.
They want enough platinum distributed through the catalyst to deliver high power.
And they need a porous carbon support that allows ionomers and reactants to move efficiently through the electrode.
Optimizing one characteristic can often damage another.
The radial nanochannel structure is significant because it integrates all four requirements into a single catalyst architecture.
That combination could make the approach especially relevant to demanding fuel-cell applications requiring both high power and long operating life.
Fuel cells could provide local power for data centers
The research itself focuses on catalyst science rather than the construction of a complete data-center power system.
Still, improved fuel-cell catalysts could make distributed electricity generation more practical.
A data center equipped with fuel cells could convert hydrogen or compatible fuels directly into electricity on-site rather than drawing all of its power from the electrical grid.
Such systems could potentially provide continuous generation while reducing transmission demands.
However, the environmental benefits would depend heavily on how the hydrogen or other fuel is produced.
Hydrogen made using low-carbon electricity can offer very different lifecycle emissions from hydrogen produced from fossil fuels without effective carbon capture.
The new catalyst does not solve hydrogen-production or infrastructure challenges by itself. Its contribution is to improve the device that converts the fuel into electricity.
The technology could extend far beyond computing
Data centers are only one possible application.
Proton-exchange-membrane fuel cells are also being investigated for heavy-duty transportation, backup electricity, stationary power systems, and other applications requiring compact and efficient energy conversion.
The high current density demonstrated under heavy-duty vehicle conditions makes the new catalyst particularly relevant to transportation research.
Long-lasting catalysts are important in such systems because replacing degraded fuel-cell stacks can be expensive.
Reducing the quantity of precious metal required while maintaining durability could also help make fuel cells more economically attractive.
A carbon structure built to protect precious metal
The underlying idea of the research is deceptively simple.
Platinum is valuable and highly effective, so researchers want every tiny particle to remain useful for as long as possible.
But producing the most desirable platinum-cobalt atomic structure requires temperatures high enough to threaten those nanoparticles.
The radial carbon channels provide a way around that conflict.
They allow researchers to heat the material intensely enough to create a highly ordered intermetallic catalyst while physically restricting the nanoparticles so that they remain extremely small and evenly dispersed.
This nanoconfinement strategy could potentially be adapted to other catalysts in which high-temperature processing and particle stability are in tension.
More development is still needed
The results represent a catalyst breakthrough rather than a finished commercial fuel-cell system.
Scaling synthesis, integrating the material into industrial manufacturing, reducing costs, confirming durability during years of realistic operation, and ensuring dependable fuel supplies will all be necessary before the technology could be deployed widely.
Wu's team plans to continue developing the catalyst and working with industrial partners.
The researchers have also filed patent applications through Washington University covering technology related to the radial nanochannel carbon spheres and the supported platinum-cobalt intermetallic catalysts.
Still, the work demonstrates that careful nanoscale engineering can overcome one of the most stubborn compromises in fuel-cell design.
By giving tiny platinum-cobalt particles a protective carbon architecture, the researchers were able to combine extreme processing temperatures, high atomic order, small particle size, high platinum loading, efficient transport, and strong durability in one material.
As electricity demand from data centers and other energy-intensive technologies continues to grow, advances like this could help expand the range of tools available for producing reliable power more efficiently.
Journal reference
Lei Gao, Sooyeon Hwang, Xiaorui Li, Jiamao Zheng, Kwanpyung Lee, Shuo Liu, Dominik Wierzbicki, Jialu Li, Jinghua Guo, Bingzhang Zhang, Honghong Lin, Qing Zhao, Guofeng Wang, Chaochao Dun, and Gang Wu. โRadial nanochannel-array carbon enables high-performance intermetallic fuel cell catalysts.โ Nature Nanotechnology, published August 6, 2026. The DOI was verified against the official Nature Nanotechnology article and resolves to this exact publication.