Ever drive by a power plant, gaze at its obtrusive cooling towers and think, “There goes a monument to wasted thermal energy.” Well, RTI International is heading a project to push the fold in thermo-electric power conversion and turning the ecosystem’s frown upside down. Using DARPA funds, RTI is combining minds from institutions such as Cal Tech, North Carolina State University, Purdue, Iowa State, UC-Riverside, the University of Delaware, and the United Technologies Corp. to capitalize on the heat energy plants generate, rather than leave it to the ozone.
The objective is to develop new materials and devices that will achieve energy-conversion efficiencies around 30% (presumably 30% of Carnot-cycle efficiency at some delta-T.) The immediate application seems to be in stationary diesel generators, with an eye toward improving fuel efficiency for the U.S. Army—which isn’t too farfetched considering DARPA’s backing.
RTI’s previous research in nano-scale superlattice materials—another DARPA-supported venture—has yielded a spinoff of the thermo-electric technology via a company called Nextreme Thermal Solutions. This company samples thin-film thermo-electric modules with applications in optoelectronics, electronics cooling, and power generation; so I’d say this is a good place for the government to be investing taxpayer money.
Catching this news whet my appetite for more thermo-electric information, so I hit the Web. It appears that ordinary Seebeck/Peltier devices, while cheap, are a dead end efficiency-wise. They yield 10% of Carnot efficiency at best. Digging deeper for more about Carnot efficiency, I surfed my way to Cool Chips’ site. Cool Chips, plc. is a company out of Europe looking to achieve 55% Carnot efficiency using thermionic conversion.
“Thermionic” in this context was a new term to me. I knew thermionic emission is what happens in the cathode of a 6SN7, but conversion? According to Wikipedia, a thermionic converter consists of a hot electrode, which thermionically emits electrons over a potential energy barrier to a cooler electrode, producing a useful electric power output. Cesium vapor is used to optimize the electrode work functions and provide an ion supply (by surface contact ionization or electron impact ionization in a plasma) to neutralize the electron space charge.
Reading further I learned that a thermionic converter consists of a hot emitter electrode from which electrons are vaporized by thermionic emission and a colder collector electrode into which they are condensed after conduction through the interelectrode plasma. The resulting current, typically several amperes per square centimeter of the emitter surface, delivers electrical power to a load at a typical potential difference of .5 to 1 V and thermal efficiency of 5% to 20% (depending on the emitter temperature, ranging from 1500°K to 2000°K).
Moving past the theory, I also read that the U.S. and Soviet Union invested a lot in nuclear thermionic-power reactors in space—presumably in spy satellites—well into the 1990s. The research shows that improvements in converter performance can be obtained at lower operating temperatures by: adding oxygen to the cesium vapor; suppression of electron reflection at the electrode surfaces; and by hybrid-mode operation.
Back to Cool Chips. I didn’t have a lot of success working my way through their site, which is essentially a marketing tool, not a scientific paper. I got the impression that bismuth telluride is the material involved, but I also got the impression that the process involves nanostructures with a feature size on the order of the de Broglie wavelength of an electron, and that can’t be right. I must have confused two statements on two different sites.
It’s not that you can’t do interesting things at a very fine level. I was discussing Cool Chips with Charlie Osborne, whom I usually turn to whenever the physics gets weird, and he “reminded” me of SiOnyx (www.sionyx.com/technology.asp). The company makes really sensitive silicon photodetectors using a process that employs a powerful, femto-second laser that exposes the target semiconductor to high intensity pulses as short as one billionth of a millionth of a second. The target’s atomic structure becomes instantaneously disordered and new compounds are “locked in” as the substrate re-crystallizes.
Presumably, this is totally different than what’s needed for thermionic conversion—or not. I can only speculate at what the scientists at RTI are thinking.