MEMS: following in the footsteps of the Internet?

There are parallels and contrasts between the ubiquitous Internet and microelectromechanical systems (MEMS), which represent an emerging technology that promises to extend microelectronics into new sensing and actuating applications. The scales at this moment are weighted toward the contrasts.

By John Rhea

WASHINGTON — Not yet.

There are parallels and contrasts between the ubiquitous Internet and microelectromechanical systems (MEMS), which represent an emerging technology that promises to extend microelectronics into new sensing and actuating applications. The scales at this moment are weighted toward the contrasts.

MEMS are tiny machines that blend electrical and mechanical components. They can sense, control, actuate, and function individually or in arrays. Complex applications of MEMS, which can integrate simple elements into more complex systems, include accelerometers, pressure, chemical, and flow sensors, micro-optics, optical scanners, and fluid pumps.

The underlying technological parallel between MEMS and the Internet centers on the Defense Advanced Research Projects Agency (DARPA) in Arlington, Va., whose scientists developed both — despite Vice President Gore's claim in the case of the Internet. Yet the two have since taken divergent paths.

The Internet is a classic case of "market pull." Originally known as the ARPANet and intended only as an in-house hardwired computer network for academic and government scientists working on advanced military projects, the Net became such an obvious means for universal communications that its phenomenal growth was inevitable.

MEMS to date have taken the opposite path, which is "technology push." About 65 percent of companies that are involved in MEMS today, in fact, entered the field because the technology seemed compelling. This is according to a recent study by the New Jersey Institute of Technology. A DARPA support services contractor, System Planning Corp. in Arlington, Va., cited the New Jersey Tech study in a market survey last year.

"Also, many of the companies getting involved in MEMS are doing so in a way that is outside their core competencies," the System Planning Corp. study notes. "That is, they were often working with manufacturing techniques in which they had little experience." The System Planning Corp. survey is available on the World Wide Web at http://memsmarket.sysplan.com/.

The Internet, on the one hand, represents full-circle commercial off-the-shelf (COTS) technology. This is something the military developed, that private industry commercialized, and now that military designers use as COTS components.

MEMS, on the other hand, have yet to close this loop.

This is not to say that commercial designers have not implemented MEMS widely; they have, albeit in somewhat low-technology applications. The highest-volume user is probably Hewlett Packard of Palo Alto, Calif., where engineers use MEMS devices in inkjet printers to spit out little globules of ink. In addition, virtually all of the major automakers use MEMS as accelerometers in their airbag restraints.

These are price-sensitive markets, notes Jane Alexander, deputy director of DARPA and a vocal proponent of MEMS. It logically follows that if the companies in these markets can save money by replacing the more expensive mechanical parts of the past — and perhaps improve reliability in the bargain — their choice is clear.

The challenge before DARPA now is to get military users to do some pulling of their own. Don Garvick, program manager at the Naval Surface Warfare Center in Indian Head, Md., has a program to develop MEMS devices as "exploders" (safing, arming and fuzing) for a new class of anti-torpedo weapons. He has been working with MEMS since 1995 and expects the program to progress to engineering manufacturing development in 2002.

Garvick is talking about MEMS that are COTS parts produced by, among others, Analog Devices in Norwood, Mass. Still, Garvick is candid about the daunting economics of MEMS, which he says must "be batch fabricated to be inexpensive and that means a minimum of a million parts a year. We buy at best a thousand torpedoes a year."

Officials of the U.S. Army, meanwhile, are working on a short-range assault weapon in which MEMS rate sensors and accelerometers at the Army Aviation and Missile Command (AMCOM) at Redstone Arsenal, Ala. MEMS in this application would provide a third axis of measurement to improve accuracy. Similar efforts at the Army's Aberdeen Proving Ground in Maryland would use MEMS sensors in gun-launched instrument systems that would make dumb munitions smart.

The Army buys tens of millions of millimeter-sized shells every year, so this market is potentially as price-sensitive as inkjet printers and airbag restraints. As Garvick puts it, "The military's number-one need is a MEMS IMU [inertial measuring unit] ... It's the best mousetrap in the world."

Richard Singer, head of the MEMS transition group at the Institute for Defense Analyses, a federally funded research and development center in Alexandria, Va., also puts IMU at the head of his list. In fact, he will be briefing industry and military leaders on this opportunity at a workshop at the Army's Picatinny Arsenal in New Jersey Oct. 31 to Nov. 2.

MEMS could do the IMU job more cheaply than today's ring laser gyros and fiber optic gyros, he says. Another advantage for MEMS is protecting the global positioning system (GPS) receivers from hostile jamming. Other possibilities, Singer notes, include use of MEMS in power-constrained applications such as unmanned aerial vehicles and on individual soldiers under the Army's Land Warrior program. More information is available on his organization's web site, www.ida.org/mems.

Karen Markus, vice president of technology at Cronos Integrated Microsystems in Research Triangle Park, N.C., has a couple mousetraps of her own in mind: biotechnology and telecommunications. These areas are clearly of interest to the military, she says, and should be a logical COTS path for adapting technology now succeeding in the automotive and computer printer industries.

Cronos represents a significant accomplishment for DARPA in its eight-year involvement with MEMS because it represents a transition of the technology into the commercial marketplace, Alexander says. In effect MEMS would help create the infrastructure that military users will need, she says Alexander's mission statement spells out that role: "In 1992, there was little industry involvement and virtually no MEMS fabrication infrastructure anywhere in the world. DARPA's MEMS investments have generated that infrastructure."

Cronos fits that description as a spinoff of the Microelectronics Center of North Carolina, also located in Research Triangle Park, and is working with what Markus calls "hundreds of companies and universities" to do prototyping and later manufacturing of MEMS devices. These include such industry leaders as Lucent Technologies in Murray Hill, N.J., she says. More information about Cronos is available on the World Wide Web at http://www.memsrus.com.

Unlike the Internet, which rapidly generated its own momentum, MEMS remains a technology sitting on the shelf waiting for the big rush of customers. DARPA's role, in addition to advocacy, is to serve as a clearinghouse for information. That's the job of William Tang, MEMS program manager in the agency's Microsystems Technology Office.

To find out what he has on his shelf, get on — where else — the Internet and look at his web page: http://www.darpa.mil/MTO/MEMS.

More in Communications