PHOTO: Lance W. Clayton

INNOVATOR: GTI’s Francis Tsang at his Idaho office.

What if you could make a miniature, superefficient nuclear power plant that’s simple to build and doesn’t get much hotter than a kitchen oven? That could be the result of an innovation being developed by scientists at Global Technologies, in Idaho Falls, Idaho.

GTI’s president, Francis Tsang, and colleagues are working on a nuclear voltaic cell consisting, basically, of a semiconductor and an amount of radioactive material [see photo, “Innovator”]. The semiconductor sits between two conductors to form a Schottky diode, and it is bombarded by particles from uranium, plutonium, or some less dangerous radioactive material.

Radiation has essentially the same effect on a semiconductor that light has on a typical solar cell. In a solar cell, the impact of a photon with the semiconductor crystal creates an electron and a positively charged particle called a hole. Because the cell’s semiconductor has been doped with chemical impurities, it has a natural polarization that draws the electron to one electrode and the hole toward the other, thereby producing current.

If a nuclear version of a solar cell sounds like one of the old atomic battery concepts from the 1950s, it is, but with a potentially all-new twist.

Some of those early concepts sought to harvest semiconductor energy from alpha radiation (positively charged helium nuclei) or beta radiation (electrons). Although they offered the hope of efficient batteries that would last for decades, they were limited by what seemed to be insurmountable problems associated with their radiation sources.

“Beta cells are restricted to low-energy beta, and they can’t use alpha,” says Jake Blanchard, a professor of engineering physics at the University of Wisconsin, Madison, who develops MEMS-based radioisotope batteries [see “The Daintiest Dynamos,” IEEE Spectrum, September 2004]. Alpha particles and other high-energy radiation “will trash the semiconductor by displacing the atoms,” Blanchard says. That has kept this class of nuclear battery from housing enough radioactive material to produce more than mere milliwatts of power.

Tsang, a former U.S. Energy Department researcher, was well aware of the beta cell’s problems. “Shoot a bullet into a block of ice, and the ice will shatter and can’t go back into its original form,” Tsang says. “But if you shoot a bullet into water, the water repairs itself.” So he began experiments replacing solid semiconductors with molten selenium and molten sulfur, both of which become semiconductors in their liquid state and melt at less than 300 °C. Because liquids don’t suffer any structural damage, Tsang’s nuclear battery could run on much more powerful radiation than a beta cell, and therefore generate more electricity.

A liquid nuclear diode could catch energetic alpha and beta particles, gamma rays, and even the new atoms left over from the fission of larger atoms, Tsang says. Fissile fragments could be a particularly good source of energy. In the fission of U-235, for example, the fragments carry 85 percent of the energy released. Because the fragments are heavy, as they plow through the semiconductor they “make a shower of electron-hole pairs along the path,” he says.

Tsang’s idea is not widely known, and for now, that’s the way he likes it. The U.S. Patent and Trademark Office posted GTI’s key patent application only in November. Tsang has not published data in a peer-reviewed journal (though some of the experiments were replicated at Lawrence Berkeley National Laboratory, in California), and he would reveal no hard numbers for this article.

The University of Wisconsin’s Blanchard, reserving final judgment until he sees published data, thinks the concept of a liquid nuclear battery is a good one. “It’s a clever idea,” he says. “It’s not totally crazy.”

At the moment, GTI’s battery is far from useful, not having quite reached 1 percent efficiency. Its development has gone far enough, however, to make an impression at the U.S. Defense Advanced Research Projects Agency. The agency gave Tsang’s company funds to support its Liquid Electronics Advanced Power System (LEAPS) program: first, US $1.4 million to prove the concept by producing current in a test cell, with a provision that would have allowed for additional funding of up to $26.6 million for over four and a half years. With submarine power plants in mind, DARPA wanted GTI to run full speed toward proving that a reactor of the 100- to 1000-kilowatt scale could be built.

But in October, Tsang’s group rejected the additional work, figuring that the effort envisioned by DARPA would overwhelm GTI’s resources. Tsang says GTI ultimately will make more progress by going after small-scale power sources first. So for now, nuclear reactors will have to take a backseat to nuclear batteries.