The continuing quest to covert automotive heat to electric power

Thermoelectric devices respond to heat by generating electricity. Such systems can generate electricity when heated by an external source, or quickly cool or heat their environment when powered with electricity. So why doesn't every car have a thermoelectric power generator?

"The reason thermoelectric devices have so far been limited to niche markets is that their efficiency is still too low," explains graduate student Jane Cornett, the project’s lead researcher at the A. James Clark School of Engineering of the University of Maryland. "The goal of our work is to design thermoelectric materials that convert energy from one form to another more efficiently so we can promote the widespread use of products that recycle waste heat and effectively reduce our consumption of fossil fuels."

Some cars manufactured or retrofitted with a thermoelectric device around the exhaust pipe, use waste heat to generate electricity thus improving mileage, especially when a power-robbing system like air conditioning is used. However, if the unit is too bulky it becomes inefficient, consuming more energy than it returns.

Consequently researchers, looking for a more efficient means of converting heat directly to electrical power, have been concentrating on what is known as nanotechnology or the science of manipulating matter on a molecular or atomic scale.

What they are finding, however, runs contrary to the common assumption in nanotechnology: "better" does not always mean "smaller." That simple fact may change the way engineers develop future thermoelectric devices.

In their approach to the problem Cornett and her advisor, Professor Oded Rabin, of the Dept. of Materials Science and Engineering, plus the Institute for Research in Electronics and Applied Physics, had to challenge some popular theories.

"Previous models told us that the use of nanomaterials at small dimensions would lead to an improvement in power generation efficiency," said Cornett. "The models also predicted the smaller the nanostructure, the more significant the improvement would be. In practice, people weren't seeing the gains they thought they should when they designed thermoelectric devices with nanoscale components, which indicated to us that there might be an issue with the interpretation of the original models."

The two have proposed a revised thermoelectric model confirming smaller is not always better. Using advanced computer modeling to predict the potential of thermoelectric nanowires only 100 to 1000 atoms thick (roughly 1000 times smaller than a human hair), they demonstrated in the tiniest nanowires decreasing their radii doesn’t result in the increased thermoelectric performance previous models predicted. In nanowires over 17 nanometers in radius, however, an improvement is seen as the radius increases.

"The surprising behavior in the larger size range demonstrates that a different physical mechanism, which was overlooked in previous models, is dominant," says Cornett.

"People were looking for solutions in the wrong places," says Rabin. "We've created a better understanding of how to search for the best new materials."

Thermoelectric devices are currently used in a few consumer products, including refrigerators and CPU coolers in computers. They could eliminate the need for fluorocarbon refrigerants, giving rise to fluid and compressor-free cooling systems that pose fewer health and environmental hazards.

Cornett and Rabin's research is supported in part by the Minta Martin Foundation and the ARCS Foundation.

Tech-minded? Click here.