November 2004

Power on a Chip

Continued from page 1

By David H. Freedman

Grounded
Epstein started thinking about building a jet engine on a chip nearly a decade ago. At the time, microelectromechanical systems (MEMS) were picking up speed. Techniques had emerged for carving new types of features into the surfaces of slabs of silicon, including sealed chambers and pipes and moving parts like spinning wheelsmost of the parts needed for a gas turbine engine. Less clear at first was what one would do with a miniature fuel-burning engine. We thought wed be able to get the cost way down if we could figure out a reason for needing a lot of them, says Epstein. But the only thing we could see doing with tiny engines was flying tiny airplanes, and that seemed stupid. Of course, we hadnt counted on the DoD.

Sure enough, the U.S. military was suddenly gung ho over the idea of 15-centimeter-long planes that could carry small cameras for surveillance. The engineers at Epsteins lab were somewhat less enthusiastic; they suspected that getting jet chips that were airworthy would take a couple of decades. Then Epstein latched onto a more immediate military need: freeing soldiers from the batteries that many of them have to lug  around to power radios, GPS receivers, night-vision goggles, and other gadgets. And unlike a miniature aircraft engine, a battery-replacing jet chip would have enormous commercial potential.

Other materials scientists and engineers were already beginning to work on ways to shrink power-producing machines to supplement or replace batteries, creating a new field called power MEMS. The most popular approach involved shrinking fuel cells, which typically pass hydrogen through a membrane that pulls electrons out to create an electric current. But Epstein was convinced gas turbines were a better way to go, because of their unmatched ability to wring power out of hydrocarbon fuels. The technology becomes even more appealing where minimizing weight and volume is critical, as with portable devices. A jet chip would be at most half the size of a micro fuel cell of equal energy capacity. A gas turbine should also be relatively easy to fabricate, figured Epstein, because it could be built entirely out of silicon, using standard fabrication techniques.

Though Epstein envisioned his micro version working roughly the same way a conventional gas turbine does, much about micro jet engines was a mystery. Would silicon crumble under 1,300 °C temperatures? Could microscopic bearings handle a million-plus revolutions per minute? With funding from the U.S. military, Epstein tapped into the expertise of neighboring MIT labs in fluid mechanics, materials science, structural engineering, and microfabrication. The project team eventually swelled to dozens of researchers, including Mark Spearing, a materials engineer  charged with finding ways to keep the silicon microstructures intact under furious heat and pressure. Most MEMS chips involve etching small structures up to 10 microns tall, says Spearing. We were going to parts that are hundreds of microns tall.