There's a hydrogen-powered fuel cell car in the fleet at work. It's a modified Toyota Prius, and from the looks of it, the hydrogen tank takes up all the space a battery would, plus the space the gas tank would, plus the trunk space. Part of this is due to it being a retrofit, part of it is due to the fundamentals of hydrogen.
One major barrier is the high price of most catalysts that one would use. The anti-smog device in a usual car contains a fair amount of similar stuff (platinum or palladium), but only a small fraction of what most designs of hydrogen fuel cell would require. To replace the whole fleet would probably take more of those precious metals than are available. Surface chemists are making great strides toward good catalysts made from less-scarce raw materials.
I think people have tried to work with fuel cells as an
energy storage device for hybrid vehicles. Generally, a NiMH battery (like comes standard in a Pius or Insight) is doing the same thing, electrolysing water to make hydrogen, then re-oxidizing that hydrogen to make electricity. But NiMH batteries are more practical all around, especially if space is a factor or large numbers of them might be made. Fuel cells are great for hybrid
solar aircraft, and a few other niches.
As to primary fuel cells, an old professor of mine, Lutgard
De Jonghe, works on solid oxide fuel cells. Rather than harness the flow of hydrogen ions to generate electricity, this type of fuel cell harnesses the flow of oxygen ions. It can run on any liquid or solid fuel with a low
enough sulfur content, from straight vegetable oil or kerosine to propane or hydrogen. They really shine when used as the combustion chamber of a gas turbine system; in that case, they can extract 80% of the fuel's energy as electricity, well over what's theoretically possible for heat engines (though, as a heat engine, usual rules apply to the gas turbine part). They're made of ceramics, so adapting them for use in vehicles has been a matter of tweaking the chemistry so that they operate at lower temperature, and can be packaged in less-expensive, less-brittle materials. I think they might be best-suited to locomotive use, ultimately.
SOFC