Volume 3, Issue 6 August 2003

2003
2002
2001

Merging Micromachines and Microelectronics by David Pescovitz Printer-friendly version Professor Tsu-Jae King also researches novel transistor designs to keep Moore's Law alive. Angela Privin photo From gears that are dwarfed by dust mites to Berkeley's own micron-scale radio components, amazing micromachines are emerging from laboratories around the world. But in order for many of these tiny devices to become practical, they must merge with traditional silicon circuits. Leading the charge at Berkeley to integrate micro-electromechanical systems (MEMS) with silicon electronics is Electrical Engineering and Computer Sciences (EECS) professor Tsu-Jae King. She and EECS professor Roger Howe are developing new standardized processes to make MEMS right on top of the conventional integrated circuits that control them. The ability to stack MEMS on top of electronics could lead to technology like true Smart Dust—cheap wireless sensors the size of sand grains under development at Berkeley—and wristwatches outfitted with mobile phones. MEMS are fabricated using processes similar to the way integrated circuits are manufactured. To create a three-dimensional MEMS structure, a sacrificial film is deposited on top of a silicon substrate and patterned as a sort of foundation for the structural layer that follows. Once the structural layer is deposited, the sacrificial layer is removed to leave the free-standing MEMS features. MEMS are traditionally fashioned from polycrystalline silicon, also known as polysilicon, because of the material's strength and resistance to fatigue. Today, MEMS like those in automobile airbag deployment sensors are then connected via wires to integrated circuits fabricated beside them. These interconnects, King says, can limit performance. In this conventional layout, a resonator is adjacent to its accompanying amplifier electronics (seen at the top of the image). Here, the resonator sits on top of the amplifier electronics, improving performance by eliminating the long interconnect wires between the two components. Stacking the MEMS and circuits is necessary to maximize performance and reduce the size of the device. The problem is that to obtain polysilicon's desirable properties, the material must be annealed, heated to a high temperature and then cooled. "Annealing burns out any electronics that are underneath the MEMS," says King, the director of Berkeley's state-of-the-art Microfabrication Laboratory and a member of the Center for Information Technology Research in the Interest of Society (CITRIS). While custom processes for integrating MEMS and electronics are available today, they're far too impractical for mass production. No semiconductor factory, King explains, is willing to pass their wafer to a MEMS foundry and then take it back again to complete the electronics. "How many products can you make with a boutique process?" King says. "Not many. If you rely on a specialized process for every MEMS product, it will never be cost effective." King's goal is to develop a process similar to the polysilicon technologies the MEMS industry is built upon. To do it, the researchers are exploiting a material in the same column of the periodic table of the elements as silicon. Silicon combined with germanium, King explains, provides the benefits of polycrystalline silicon but can be processed at temperatures hundreds of degrees lower. It can also be patterned using conventional MEMS fabrication tools. The Berkeley researchers have already built prototype devices using the silicon-germanium process, including an audio-frequency filter used in radio transceivers. In the future, King says, modularly integrated MEMS-electronics technology could be used to build low-power radio transceivers on a single chip. What kinds of new devices do you think could be developed by merging MEMS and microelectronics? We want to hear from you... "We're collaborating with mechanical engineering researchers who come up with novel designs that we can integrate using our technology," King says. Because the processes remain the same as those used by current commercial MEMS foundries, the factories do not need to be adapted for silicon-germanium nor does industry-standard MEMS design software need to be rewritten. "Instead of worrying about compatibility issues every time the electronics industry or MEMS industry updates its processes, you'll be able to take the latest and greatest the electronics industry has to offer and stick MEMS right on top of it," King says. Tsu-Jae King's Home Page Roger Howe's Home Page UC Berkeley Microfabrication Laboratory Center for Information Technology Research in the Interest of Society (CITRIS) Berkeley Breathes New Life Into Silicon by David Pescovitz (Forefront, Spring 2002) Lab Notes is published online by the Public Affairs Office of the UC Berkeley College of Engineering. The Lab Notes mission is to illuminate groundbreaking research underway today at the College of Engineering that will dramatically change our lives tomorrow. Editor, Director of Public Affairs: Teresa Moore Writer, Researcher: David Pescovitz Designer: Michele Foley Subscribe or send comments to the Engineering Public Affairs Office: lab-notes@coe.berkeley.edu. © 2003 UC Regents. Updated 7/31/03.