The sales pitch for a hydrogen fuel cell has always been the tailpipe. Hydrogen goes in, air goes in, electricity comes out, and the only thing that reaches the pavement is water.
All of that is true. What doesn’t make the brochure is that the same machine shuts down if its insides ever dry out.
The membrane at the center of a conventional fuel cell only moves protons while it stays wet. That caps the stack at roughly 80 to 90 degrees Celsius and forces anyone packaging one into a vehicle to bolt on humidifiers, water plumbing, and a radiator wide enough to hurt the aerodynamics.
Engineers at Monash University in Melbourne built a membrane with no water in it at all. It ran a hydrogen fuel cell at 482 degrees Fahrenheit (250 Celsius) and put out 1,011 milliwatts per square centimeter, according to the paper in Science Advances, published May 15.
The number that matters here isn’t actually the power figure, and I’ll get to why. It’s the 166 millisiemens per centimeter of proton conductivity they measured with the thing bone dry.
A machine that makes water can’t run without it
Standard fuel cell membranes are perfluorosulfonic acid polymers, Nafion being the one everybody names. They work because protons hitch a ride on water molecules through the membrane, which is elegant right up until you want the cell to run hot.
Push past 100 Celsius and the water boils off. Conductivity collapses. The cell stops.
So every fuel cell vehicle on the road carries hardware whose entire job is keeping a machine that manufactures water from getting too dry, at a temperature low enough that the water stays liquid. The Army’s hydrogen Chevy Colorado, the Toyota Mirai, the cells in the long-endurance drones now attracting billion-dollar valuations, all of them are running that same balancing act.
The industry’s workaround has been phosphoric acid soaked into polybenzimidazole, which gets you to 140 or 180 Celsius. Better. Still not the range heavy machinery actually operates in, and the acid tends to leak out and dry up the hotter you go.
The radiator is the part that actually stops the truck
Here’s where this stops being a chemistry problem and starts being a packaging problem.
A fuel cell running at 90 Celsius has a small temperature gap between itself and a 40-degree summer day, and heat only moves across a gap. Small gap, big radiator. The Department of Energy’s own technical targets pin the metric at 1.45 kilowatts per degree Celsius, measured at 0.692 volts and 90 Celsius.
The DOE has been blunter than that in its funding paperwork. A solicitation for heavy-duty thermal management states outright that PEM-powered heavy vehicles need larger radiators than diesel equipment, with frontal areas big enough to cut into truck aerodynamics, plus bigger fans drawing more parasitic power.
A Class 8 truck already fights for every square inch of frontal area. Adding a radiator the size of a barn door to save fuel is the kind of trade that makes fleet managers laugh at you.
Which is why the DOE has spent years asking, in writing, for stacks that run above 100 Celsius purely for heat rejection. Cummins took federal money to chase exactly that. Run the stack hotter, the gap to ambient widens, the radiator shrinks, and the aerodynamics come back.
Monash swapped the river for a handrail
The Monash approach starts from a material nobody expected to conduct anything. Single-atom-thick sheets of graphene and hexagonal boron nitride let thermal protons pass through while blocking basically everything else, a property confirmed in the lab more than a decade ago.
The problem was always stacking them. Individual sheets conduct protons fine. Pile them up into something thick enough to be a membrane and the protons get stuck between the layers.
So the team, led by first author Dr Kaiqiang He with Professor Huanting Wang as corresponding author, filled the gaps between sheets with phosphoric acid held in nanoscale confinement, and coated the sheets with polyethylenimine to keep the acid anchored there. Protons then move two ways at once. Straight through the sheets, and by hopping down hydrogen bonds along the trapped acid layers.
Water carries protons the way a river carries a boat. This carries them the way a handrail carries your hand, one grip at a time, no current required.
The confinement does something else worth noting. Phosphoric acid normally starts breaking down around 165 Celsius in air. Locked into these 2D channels, the paper puts the decomposition onset at 225 Celsius for the acid, 437 for the polyethylenimine, and 802 for the graphene.
It also ran on concentrated methanol without falling apart, which opens a second door entirely, since methanol is a liquid you can pump at a truck stop today.
1,011 milliwatts is not the record, and that’s fine
This is the part the press coverage skipped, and it’s the part that tells you what the result actually is.
A team at the Korea Institute of Science and Technology got a fuel cell to 2.35 watts per square centimeter at 250 Celsius in dry hydrogen and oxygen, published in Nature Energy in 2024, with negligible degradation across 500 hours of thermal cycling. That is more than double the Monash number, two years earlier, with durability data attached.
Monash didn’t claim otherwise. The wording in the paper is that the membrane outperforms most previously reported anhydrous membranes, which is a narrower and more defensible claim than anything in the headlines that followed.
What’s new is the material class. The Korean membrane is a polymer with ceramic particles in it. Monash got comparable behavior out of stacked 2D nanosheets, a family that had been stuck for a decade on the interlayer transport problem.
You can see how much the filler chemistry matters from a paper published in Angewandte Chemie in June, which used the same graphene and boron nitride scaffold with polyethylenimine, but loaded it with ionic liquids instead of phosphoric acid. It reached about 10 millisiemens per centimeter at 240 Celsius. Same architecture, roughly one sixteenth the conductivity.
Three unrelated chemistries have now hit 250 Celsius. That convergence is the real signal, more than any single number.
Sinopec picked last month to say the quiet part
None of this arrives into a booming market.
Sinopec, the world’s second largest oil company and a firm with every commercial reason to talk hydrogen up, posted an assessment on its in-house news site on June 30 saying battery electric trucks are moving into the exact use case hydrogen was supposed to own. The company’s line, machine-translated from Chinese and reported by Electrek on July 2, is that charging and battery-swapping infrastructure has pushed electric heavy trucks into “medium-to-heavy loads and long range.” Sinopec still sees a long-haul niche for hydrogen. It also says finding that niche will be hard.
Meanwhile Toyota and Hyroad Energy signed a deal in May to put 40 hydrogen Class 8 trucks into Southern California freight lanes, and Toyota moved to join the Cellcentric fuel cell venture alongside Daimler Truck and Volvo. The money hasn’t left the room. It’s just gotten pickier.
And the sector’s biggest cost problem is upstream of any membrane, in moving the fuel from where it’s made to where it’s burned. A hotter membrane does nothing about that.
What Monash hasn’t shown yet
The gaps in this one are specific, and they’re the same gaps that separate a good lab membrane from a product.
There is no published stack. There’s no long-run durability figure of the kind the Korean team reported at 500 hours or the Chinese polybenzimidazole work reported at 250. There’s no disclosed membrane thickness in the public materials, which matters enormously because thickness drives resistance and cost. Monash has named no commercial partner.
Phosphoric acid also brings its own baggage. Acid migration into the electrodes is the chronic failure mode of every phosphoric-acid-doped high-temperature cell ever built, and confinement inside 2D channels is a promising answer to it rather than a proven one.
Still, the mechanism is the interesting bit, and mechanisms travel. The same team is already using the architecture for hydrogen purification, and the paper points at water splitting, CO2 reduction and ammonia synthesis. That’s the pattern with fuel cell research lately, where the useful discovery keeps turning out to be a transferable trick rather than a finished device, the way a Korean group recently improved the key reaction by changing the electric field instead of the catalyst.
If this holds up in a stack, a hydrogen truck gets to drop the humidifier, shrink the radiator, and stop bleeding drag. That’s a real prize. It’s also three or four hard engineering problems away from a lab bench in Melbourne.
