Computing Sciences masthead Berkeley Lab Computing Sciences Berkeley Lab logo

Exascale for Energy


Turning waste heat into electricity

Automobiles, industrial facilities, and power plants all produce waste heat, and a lot of it. In a coal-fired power plant, for example, as much as 60 percent of the coal’s energy disappears into thin air. One estimate places the worldwide waste heat recovery market at one trillion dollars, with the potential to offset as much as 500 million metric tons of carbon per year.

One solution to waste heat has been an expensive boiler-turbine system; but for many companies and utilities, it does not recover enough waste heat to make economic sense. Bismuth telluride, a semiconductor material, has shown promise by employing a thermoelectric principle called the Seebeck effect to convert heat into an electric current, but it’s also problematic: toxic, scarce, and expensive, with limited efficiency.

Computed charge density in alloy
Figure 14. Contour plots of charge density in a zinc selenide/zinc oxide alloy. Source: J. Wu

What’s the solution? Junqiao Wu, a materials scientist at the University of California, Berkeley and Lawrence Berkeley National Laboratory, believes the answer lies in finding a new material with spectacular thermoelectric properties that can efficiently and economically convert heat into electricity. (Thermoelectrics is the conversion of temperature differences, that is, differences in the amplitude of atomic vibration in a solid, into an electric current.)

Wu knew from earlier research that a good thermoelectric material needed to have a certain type of density of states, which is a mathematical description of distribution of energy levels within a semiconductor. “If the density of states is flat or gradual, the material won’t have very good thermoelectric properties,” Wu says. “You need it to be spiky or peaky.”

Wu also knew that a specialized type of semiconductor called highly mismatched alloys (HMAs) could be very peaky because of their hybridization, the result of forcing together two materials that don’t want to mix atomically, akin to mixing water and oil. He hypothesized that mixing two semiconductors, zinc selenide and zinc oxide, into an HMA, would produce a peaky density of states.

Beginning in late 2008, Wu collaborated with other researchers to run computations on the Franklin supercomputer at NERSC. After months of number crunching, Wu’s idea held up (Figure 14). Theoretically, at least, mixing the two materials enhanced their thermoelectric performance considerably, producing a new, highly efficient, potentially low-cost thermoelectric material.

Wu is now working on synthesizing this material in the lab, a process that could take three to five years. If Wu’s experiments are successful, it might be another five years before his material is adapted into a thermoelectric device and sent out to recover waste heat.

Adapted from “Thermoelectrics: A matter of material” by Rachel Shafer, in Innovations: Research & News from Berkeley Engineering, Vol. 4, No. 5, June 2010.


<< Previous page