Thermoelectric Research Heating Up at Penn State

Thermoelectric materials have the ability to turn heat into electricity. Long used in niche applications such as providing electrical energy on long distance spaceflights or as refrigeration in portable coolers, thermoelectrics are beginning to find more prominent applications as their performance improves.  

Thermoelectric materials have the potential to utilize the vast amounts of waste heat lost to the environment through industrial processes. The Department of Energy estimates that between 5-13 quadrillion BTUs of energy a year are generated by waste heat, enough to power some two million homes. Even capturing a portion of that heat to generate electricity could save companies an estimated 5 percent in energy costs. In vehicles, a thermoelectric device to capture and convert exhaust heat could eliminate the need for an alternator.    

The field of thermoelectrics and thermally functional materials was a backwater of research at Penn State and other research universities for most of the past 50 years, stalled by a failure to develop new materials with improved functionality. But in the last decade, new materials with improved thermoelectric efficiency have reinvigorated the search for functional thermal materials and devices, making this one of the emerging research fields for Penn State and the broader materials community.

The thermoelectric effect has been known for generations. The German physicist Thomas Johann Seebeck discovered thermoelectricity in the 1820s when he discovered that two metals with differing temperatures at their junction could deflect a magnet. It was later realized that the temperature difference was actually producing an electrical current. The Seebeck effect is defined as the electrical potential produced by a temperature difference. Since Seebeck’s time, the connection between heat and electric current has been put to use to scavenge useful electricity from waste heat, or reversing the effect and applying an electric current to a thermally functional material, to produce heating or cooling, a process called the Peltier effect. Penn State researchers are pursuing both paths.

In 1996, one of the more cited papers in the field of thermoelectrics was published by two Penn State  physicists, Jerry Mahan and his onetime post-doc and current colleague Jorge Sofo. Mahan is a theoretical physicist who along with his university post spent 30 years consulting for the General Electric Research Lab, primarily modeling devices, including thermal devices. Their paper was about figuring out what the best thermoelectric material would look like if it could be designed perfectly, a forerunner of the current Materials Genome Project, which is an attempt to cut down the time it takes to produce useful materials by using computational techniques.

“The most generally useful materials have to have three properties: one is low electrical resistance; another is low thermal conductivity; the third is a very high Seebeck coefficient,” Mahan explained. “We said if you want to have the perfect material, it has to have these properties. And then people tried to find those materials. It has led people in the right direction using the Mahan-Sofo algorithm.

 “The field of thermoelectrics has blossomed in the last few years with the advances in power conversion materials,” Mahan added. “We’ve found good high temperature materials, good for waste heat power generation. Now the great need is for materials that work at room temperature or below. All my theoretical work has never found a reason why you couldn’t have one. Someday someone will stumble across a material and the field will take off.”

 The figure of merit for comparison of the properties Mahan discussed is called zT.

Jorge Sofo, professor of physics and professor of materials science and engineering, said, “You could say that at the beginning in the early 1950s, the best materials had a zT just below 1. Now we are at 1.1 and what we need is 4. If you go above 2 with that number at room temperature, you can start to get real life applications. At 4, thermoelectrics would be everywhere.”

Sofo explained that the standard device hasn’t changed since Abraham Iofe developed the first semiconductor thermoelectric device in 1953. The device is always a negative semiconductor that carries electrons and a positive semiconductor that carries holes. In an n-type semiconductor the electrical current and heat current travel in the same direction. In a p-type semiconductor, the electrical current and heat current go in opposite directions. Hook them up and you have a thermoelectric device. The design hasn’t changed in 60 years, so what is needed is a material with a seriously higher zT.

That material may have been discovered here at Penn State, or at least that’s what Mahan thinks. “When I saw Clive Randall’s data, I said, ‘That’s spectacular, I hope you patented it.’ The material is strontium-barium niobate. It’s an alloy with a very complicated crystal structure. It’s the type of solid we didn’t know about in 1955.”