If entering a field before most people have even heard of it boosts your chances of making a mark in it, things look good for Gang Chen.

In the early ‘90s, Chen, now an MIT mechanical engineering professor, was a grad student at Berkeley, and had landed a spot in the lab of the then chancellor. Chen’s specialty is heat-transfer engineering — a staple of his discipline — and his advisor urged him to interpret that category adventurously.

Chen did. He probed topics from how to cool sophisticated lasers to the use of thin films — almost infinitesimally shallow coatings of selected materials — in electronic devices. “It was a lot of fun,” he says.

Some of the work involved dealing with materials at nanoscales, where dimensions are measured in billionths of a meter. Back then, of course, nanotechnology was a new concept — the word itself had only come into use in the mid-’80s. But Chen’s foray into what’s now one of the hottest areas in technology gave him a good background for work on a fast-growing approach to energy saving.

That approach involves thermoelectricity, which is based on a long-ago finding that some metals and, especially, semiconductors (the best known of which is the silicon used in computer chips) can generate a voltage when heated. The system also works in reverse: one common use of thermoelectricity relies on juice from the battery to rapidly cool the seats in some luxury car models.
Prof. Gang Chen

To those familiar with thermoelectricity in the mid- to late ‘90s, Chen’s interest might have seemed odd. The technology was a niche player in the energy arena. But he’d worked in closely related areas: in his Berkeley studies, he notes, “I showed that heat does not travel well in nanostructures.” About the same time, the first intellectual seeds for a vastly expanded role for thermoelectricity were being planted. Some ideas from that era, in fact, have combined with society’s energy worries to propel thermoelectricity into the limelight.

How promising is it? Chen says thermoelectricity has “game-changing” potential. One likely application: harvesting waste heat in cars by converting it into electricity. “Cars are about 20 percent efficient,” notes Chen, “and turning some of the energy wasted into electricity could increase that figure by as much as one-third.”

But that’s just for starters. The U.S. government has predicted thermoelectric generators could replace conventional engines in some cars before mid-century. And Chen’s striving to further such advances.

MINI-POWER PLANT

Thermoelectric devices are energy converters. When they’re producing electricity, this puts them in the same broad category as power plants and solar-generating systems. When outputting heat or its opposite, meanwhile, they’re like heat pumps and air conditioners respectively.

In design terms, thermoelectric devices have key pluses. For one, they’re solid state: no liquid fuels, no moving parts. They’re also easily scalable up or down.

This last feature explains many of thermoelectricity’s current uses. “If you need a small-scale device,” says Chen, “you don’t really have any other choices.” That’s why many deep-space probes use radioactivity-driven thermoelectric generators.

There have been efforts to make the technology more mainstream. “In the ‘40s and ‘50s,” notes Chen, “there was a lot of interest in solid-state refrigeration. The goal was to create full-sized thermoelectric refrigerators.” But while thermoelectric mini-fridges are increasingly common, the dreams of those early enthusiasts came to naught.

Why? It’s mainly an efficiency issue. And a key reason this dam began to give way in the early ‘90s is that MIT physicist Mildred Dresselhaus and a colleague had an idea: instead of simply testing a long list of different materials, why not change the materials themselves by structuring them internally such that performance improves?

The pair specifically proposed creating nanoscale substructures in the materials. And what made the concept intriguing is that the ideal thermoelectric device is one which is great at conducting current and an abject failure at conducting heat.

That’s a rare combination. “Nature,” notes Chen, “doesn’t provide many examples of materials that are great electrical conductors and also good thermal insulators.” But technical staff member Ted Harman at MIT’s Lincoln Laboratory — building in part on Chen’s earlier, unrelated work — showed that by using nanostructures, you can create materials that outdo nature: some of Harman’s materials, thanks to their unique heat-impeding qualities, are twice as efficient as their conventional cousins. It’s an astonishing advance — roughly equivalent, if on a drastically smaller scale, of turning a one-megawatt power plant into a two-megawatt one.

Of course, it’s tough to turn advances in tiny experimental devices into commercial winners: don’t expect whole-house thermoelectric air-conditioning systems to start turning up at your local HVAC dealer’s anytime soon.

On the other hand, Chen says innovations like an exhaust-mounted energy-mining device for vehicles needn’t wait until you hit Lincoln Lab realms of efficiency. “If you can reach a 10-to-15 percent conversion efficiency,” he notes, “that would be attractive for many applications.” In fact, results he’s had at that level are already drawing interest from companies.

This not only gives Chen hope that thermoelectricity’s time may have truly come, it also resonates with the goals he’s set for himself as a researcher. “I like to explore things that are fundamentally new and different,” he says, “and then see how I can use those findings to make an impact on the real world.”