Turning wasted heat into a power source

Mercouri Kanatzidis envisions a refrigerator that not only would keep the Maytag repairman pining by a silent phone, but could put him out of business altogether.

Gone would be the noisy compressors, the environmentally dubious coolants, and the dust bunnies under the cooling coils. Instead, says the chemistry professor at Michigan State University, the unit would rely on electricity flowing through specially designed semiconductors to keep the inside of the icebox chilled. Those same semiconductors also could be used to convert wasted heat in auto exhaust pipes, power-plant smokestacks, or other sources into valuable electricity.

The problem: So far, it's been hard to develop semiconductors in big enough chunks and with the right characteristics to turn these hopes into affordable hardware that works.

Now, Dr. Kanatzidis and colleagues have hit on a semiconductor recipe that appears to move those hopes closer to reality.

The material is a combination of silver, lead, antimony, and tellurium. And the thermoelectric equivalent of a football quarterback's rating - its "figure of merit," or ZT - appears to be the highest yet achieved for bulk materials at high temperatures. In other words, the material seems to be more efficient at converting heat into electricity than any other similar material.

Indeed, its rating appears achingly close to being competitive with current power-generation and cooling technologies, scientists believe.

This stems not just from the material's relatively high figure of merit, but also from the way the silver and antimony arrange themselves at such a minuscule scale, where sizes are measured in a few billionths of a meter.

That's encouraging, says Arun Majumdar, a mechanical engineer at the University of California in Berkeley who specializes in energy conversion and transport at such small scales. Heating and cooling technologies using materials with a ZT of 3 or more begin to close the cost gap with conventional technologies, he points out. Until now, bulk materials reached ZTs in a range of only 0.6 to 1.0. Kanatzidis's team pushed their material to a ZT of 2.2.

For all the high-tech approaches to forming the new materials, the principle behind what they do is simple, researchers say. Join a pair of wires made from different materials, apply an electric current, and one of the two wires will heat, while the other cools.

If the two wires are joined at each end to form a closed loop, and one joint is subjected to a higher or lower temperature than the other, a current will flow through the loop. These traits were first discovered in the early 1800s, Kanatzidis says. The discovery remained a curiosity until the 1940s and '50s, when semiconductors were developed. Semiconductors could be fabricated to display the same thermoelectric traits as traditional conductors.

Indeed, researchers initially envisioned using semiconductors for large-scale thermoelectric cooling and power generation. Products using these approaches have found specialized niches, such as powering spacecraft instruments or ground equipment in remote locations.

But the hopes for more widespread use proved elusive. The field got a boost in the 1990s when researchers found that the performance of the material increased when it shrank in size. So they developed devices that used thin films of the material.

For consumers, this progression has led to small beverage coolers that plug into auto cigarette lighters or to collars that claim to help people keep warm in the winter or cool in the summer. In Finland, a company has devised a thermoelectric generator from semiconductors that uses heat from boiling water to power the pump that circulates that water through a home-heating system.

Still, the market has remained relatively small and specialized, with many thin-film devices being used in science equipment or to cool computer chips.

But for large-scale applications, "there's a big need for bulk materials," Kanatzidis says, "because thin films are hard to make and expensive."

"It took a lot of trial and error to make it," he adds. The result, however, is a material that in bulk makes twice as good a thermoelectric device as previous bulk materials.

The team is reporting its results in Friday's issue of the journal Science. While much has been made over the years of the maintenance-free fridge, Kanatzidis says he sees power generation as the most obvious winner if researchers can improve these materials.

In principle, if semiconductors can be formed with sufficiently high ZTs, it might be possible to build self-powering devices or to augment electricity.

More immediately, it may be possible to wring more electricity out of every ton of coal, cubic foot of gas, or barrel of oil burned. If 10 percent of that energy could be harnessed using semiconductors to convert heat to electricity, it could outstrip the contribution from the renewables, he says.