As I combed through email over the holidays, I ran across an interesting article from MIT's Technology Review about the potential to tap waste energy in cars using "thermoelectrics", which convert heat directly to electricity. The article focused on the possibility of using these materials to capture heat, e.g. from the car's exhaust, to run accessories or power steering. It now occurs to me that the best use of this technology might be as an adjunct to hybrid systems, boosting the stored electricity available to drive the car. That could make hybrids more efficient and attractive in the long run, giving fuel cells an even tougher competitor to beat.
The "holy grail" of energy efficiency is converting most of the chemical energy of our fuels into useful power, and much less of it into waste heat to the environment. The internal combustion engines that power our cars do a terrible job of this. The theoretical maximum efficiency of a heat engine is around 40%. Most car engines are lucky to deliver half of that, in the real world. In other words, for every gallon of gas you buy, you are only benefiting from the energy content of about one-and-a-half pints. What if you could turn that into 3-4 pints?
Two of the most popular energy efficiency technologies today, the hybrid car and the combined cycle gas turbine, are aimed directly at capturing energy otherwise lost as heat. This is done directly in the case of the CCGT, by using the hot turbine exhaust to generate steam to run another turbine, and indirectly in the case of a hybrid car, which recycles some of the energy lost in braking. Marrying a hybrid car drivetrain with thermoelectric material designed to turn engine heat into extra power might effectively give you a "combined-cycle car" with thermal efficiency approaching that of a fuel cell but without the latter's need for hydrogen infrastructure.
The biggest challenge in making this practical--aside from whatever is involved in perfecting the thermoelectric material--is maximizing the temperature at which the engine runs. If that sounds paradoxical or even dangerous, remember that the laws of thermodynamics dictate that the useful energy you can extract from something is related to its temperature. A CCGT is only as efficient as it is, because the turbine exhaust comes out at a couple thousand degrees F. By comparison, your radiator, the main heat sink for your car, keeps the engine block at 150-200 degrees, which is pretty low-level as heat sources go. This might be where ceramic engine blocks--an idea that has been floating around for decades--might shine, by operating at temperatures high enough to provide a nice heat source for thermoelectrics.
Now, I have to admit this whole idea is pretty speculative. But even if what I described never ends up in a car you can buy, it illustrates the diversity of ways to skin the vehicle efficiency cat. It is still premature to proclaim fuel cells, conventional or plug-in hybrids, or anything else the clear winner in this race.