A Nanocarbon Cage Could Make Hydrogen Fuel Cells Cheap Enough for Data Centers

Data centers already draw roughly 4% of America's electricity, and that share could reach 9% by 2030. That growth pressure is serious enough that New York has already paused new hyperscale data center permits to study the strain. A new catalyst design out of Washington University in St. Louis tries to ease that pressure by making hydrogen fuel cells cheaper to build and longer-lasting.

KEY NUMBER
The new catalyst kept 85% of its performance after 150,000 stress cycles, roughly equivalent to 25,000 hours of continuous operation.

Why it matters

Fuel cells generate electricity by combining hydrogen and oxygen, producing only water and heat as byproducts, but they depend on platinum to speed the reaction. Platinum is expensive and, at the nanoparticle scale needed to make it efficient, prone to dissolving, clumping and losing performance over time. Every fuel cell design has to balance activity against durability, and until now, pushing one up has generally meant pushing the other down. As power-hungry data centers strain electrical grids, a catalyst that resists that trade-off could make on-site hydrogen power a more realistic option, alongside vehicles and other applications that need durable, low-platinum fuel cells: the same durability-versus-performance problem researchers are also chasing in next-generation battery chemistry.

What happened

A team led by chemical engineer Gang Wu, working with researchers at Brookhaven National Laboratory, Lawrence Berkeley National Laboratory, Northeastern University and the University of Pittsburgh, designed a new carbon scaffold made of porous, hollow spheres threaded with ordered channels radiating out from the center. Think of it as a honeycomb of tiny, branching tunnels: each channel gives an individual platinum-cobalt particle its own protected pocket, spaced far enough from its neighbors that heating the material to extreme temperatures does not cause the particles to touch and fuse.

That structure let the team heat platinum-cobalt nanoparticles to 1,000°C, hot enough to lock the metal atoms into a highly ordered, more stable arrangement, while the channels kept the particles smaller than 5 nanometers and evenly spread out. Most existing platinum-alloy catalysts are processed below 700°C to avoid clumping, a temperature too low to fully complete that ordering process. By solving the clumping problem at higher heat, Wu's group produced a catalyst that combined high activity with the durability that comes from a fully ordered structure. The open channels also helped the fuel cell's ion-conducting material spread evenly, easing the movement of protons, oxygen and water through the electrode. The results were published August 6 in Nature Nanotechnology.

What happens next

Wu's team says the next step is working with industry partners to move the catalyst design toward manufacturing and to resolve the remaining engineering problems around scaling up production. The university has already filed a patent on the technology. Whether it reaches commercial fuel cells will depend on whether the nanostructured carbon can be produced reliably and cheaply at the volumes an automaker or data-center operator would need.

Takeaway

The advance does not solve data centers' energy appetite on its own, but it chips away at one of the specific technical barriers, platinum cost and durability, that has kept hydrogen fuel cells on the margins of heavy-duty power generation. The next test is whether it survives the jump from a research lab to a factory floor.