NEWS
Hydrogen-selective membranes could unlock new possibilities for high-temperature hydrogen production, fusion energy, and cleaner industrial processes.
Written by Annemarie Mannion
PALLADIUM HAS LONG OCCUPIED a quiet but critical role in hydrogen infrastructure. The silvery metal allows hydrogen to pass through while blocking almost every other gas. That selectivity makes palladium-based membranes one of the most effective ways to produce ultra-pure hydrogen, a necessity for industries ranging from semiconductor manufacturing to fertilizer production.
But palladium has a weakness. At temperatures above about 800 Kelvin, the thin films traditionally used in membranes begin to fail. They form holes, bead up into droplets, and lose their ability to block other gases. That limitation has kept palladium membranes out of many next-generation hydrogen technologies—systems like compact steam methane reforming, ammonia cracking, and even fusion energy, all of which operate at much higher temperatures.
At the Massachusetts Institute of Technology, a team of engineers set out to redesign palladium membranes so they could survive and function at temperatures approaching 1,000 Kelvin, without sacrificing hydrogen selectivity.
“We were thinking how to develop membranes that could withstand high temperatures encountered in fusion applications for recovery of deuterium and tritium,” said Rohit Karnik, the Jameel Professor of Mechanical Engineering at MIT. “Thin, continuous metal films tend to be unstable at high temperatures due to a phenomenon called de-wetting.”
The project emerged from a broader effort supported by the energy company Eni S.p.A. through the MIT Energy Initiative, focused on future fusion power plants. In these systems, hydrogen isotopes circulate at extreme temperatures and must be continuously separated and recycled. Cooling gases just to pass them through a membrane is costly and inefficient. A membrane that could sit closer to the reactor—and tolerate the heat—could be transformative.
This image illustrates a palladium “plug” composite membrane separating hydrogen (red) from helium (blue). The isolated plug architecture resists solid-state de-wetting and maintains stable performance even after exposure to 1000K. Source: Loyhun Kim/MIT
The team’s breakthrough came from rethinking the membrane itself. Instead of forming palladium as a continuous film on a support, they asked a different question: What if palladium could be placed in a configuration that was already thermodynamically stable?
“The key principle is surface energy minimization,” Karnik explained. “A thin metal film on a ceramic substrate has high surface area, and the adhesive energy between the metal and the substrate is weak. The thermodynamically stable state of the metal film is to ball up.”
That tendency—to shrink into droplets at high temperature—is exactly what destroys conventional membranes. But the team realized it could also be the solution. If palladium were embedded as tiny “plugs” inside the pores of a supporting material, each plug would already resemble a droplet with minimal surface energy. In that confined space, palladium would have little incentive to move or clump further, even under intense heat.
It was an intriguing idea—but proving it worked was another matter.
“We tested dozens of different prototypes with different ideas during the research process,” Karnik said. “Many membrane designs failed prior to this result.”
Lohyun Kim, who led much of the experimental work as a doctoral candidate, describes a painstaking fabrication process. The membranes began as porous silica supports with pores roughly half a micron wide. Palladium was deposited into those pores, and then the surface was mechanically polished to remove any remaining palladium film.
“The most challenging step was the final mechanical polishing,” Kim said. “Insufficient polishing can leave residual palladium film, which can compromise thermal stability, while excessive polishing can damage the palladium plug and the underlying support.”
But then came the moment when the plug-based membranes were pushed to extreme conditions and kept working.
“When it first demonstrated stable and promising performance at high temperatures, the focus shifted to verification,” Kim explained. “At that point, I felt relief and satisfaction, as so many membrane designs had failed prior to this result.”
“Thin, continuous metal films tend to be unstable at high temperatures due to a phenomenon called de-wetting.”
—Rohit Karnik, the Jameel Professor of Mechanical Engineering at the Massachusetts Institute of Technology
In tests lasting more than 100 hours at temperatures up to 1,000 Kelvin, the membranes continued to selectively separate hydrogen without degrading. Compared to traditional palladium films, the new design extended thermal resilience by at least 200 Kelvin.
That durability opens doors. Steam methane reforming, one of today’s primary hydrogen production methods, could be redesigned as compact membrane reactors that operate at high temperatures and directly extract hydrogen. Ammonia cracking, a promising way to transport hydrogen safely, could also benefit from membranes that function reliably around 800 Kelvin.
“The membrane concept illustrates a pathway to make membranes that are much more resistant to high temperatures,” Karnik said. “It also opens up the possibility of reducing the amount of precious metal used in membranes while retaining stability.”
Looking ahead, the team sees scaling the membrane as its most immediate hurdle.
“The main challenge is scaling up the membranes to wafer sizes and packaging them into usable modules,” Karnik said. The group hopes to partner with industry to validate the membranes under realistic operating conditions and integrate them into working reactors.
There are also other questions to explore: how gases interact at the interface between palladium plugs and pore walls, how the nanostructures evolve over time, and whether the approach can be extended to other metals and alloys.
After three years of work, the team pioneered a new way of thinking about materials under extreme conditions. By embracing palladium’s natural tendencies rather than fighting them, the team has moved one step closer to making high-temperature hydrogen production—and a hydrogen-powered future—more practical, efficient, and affordable.
Annemarie Mannion is a technology writer in Chicago.
© 2026 The American Society of Mechanical Engineers. All rights reserved.