The New MEMS and Their Killer Apps
Roger H. Grace
The rapidly changing market for MEMS-based technology is on the brink of a new era. In a guest editorial last year (The Commercialization of MEMS/MST: Where Do We Go from Here? - Sensors, February 1999), I presented a MEMS product evolution timetable which showed that pressure sensors, the earliest manifestation of MEMS technology (discovery period 1954-1985), had reached full commercialization by 1990 through the excellent efforts of such companies as Ford, Honeywell, Lucas NovaSensor, Motorola, and Sensym. And MEMS accelerometers (discovery period 1974-1985), manufactured by Analog Devices, Motorola, Nippondenso, and SensoNor, reached full commercialization status by 1997.
Since reaching commercialization, the growth of these technologies has been pretty impressive. From 1995 to 2000, MEMS accelerometers, thanks to their rapid and wide acceptance in air bag safety systems, have grown in the 20%25% compounded annual growth rate (CAGR) range. The CAGR of MEMS-based pressure sensors has been in the low teens.
Growth rates for these products are not expected to climb appreciably beyond these ranges, however. I submit that these are "old MEMS" technologies, and that the major growth in the MEMS market as in the stock market will be in the arena of "new MEMS": devices that will enable such "killer applications" as extremely high volume telecommunications and biomedical innovations (see Figure 1).
Figure 1. Microsystems Technology Worldwide Market Application Sector 2000 2004 CAGR(%) IT/Peripheral 8,700 13,400 11.5 Medical/Biochemical 2,400 7,400 32.5 Industrial/Automation 1,190 1,850 11.6 Telecommunications 130 3,650 128.1 Automotive 1,260 2,350 16.9 Environmental Monitoring 520 1,750 35.4 Total 14,200 30,400 21.0
Microsystems Technology Worldwide Market
In fact, the market for MEMS parallels the market at large in this respect. Youve heard stock market pundits recently comparing old-economy and new-economy stocks. Old-economy stocks are those of traditional and typically less technology-driven companies including Polaroid (photography), Chubb (insurance), and General Electric (industrial), and these accounted for little of the markets recent growth. By contrast, new-economy stocks, such as EMC (network/storage), Intel (semiconductors), JDS Uniphase (fiber-optic communications), and Amgen (biotechnology), soared in 1999 and accounted for a disproportionate share of overall market growth.
Over the past few years, a large number of bioMEMS companies have hung out their shingles, including Aclara, Affymetrix, Caliper, Cepheid, and Orchidall of which will have gone public by the time you read this. These companies have brought together an interdisciplinary group of experts who have created highly innovative products that are truly revolutionary and whose killer applications support major societal issues, including DNA sequencing; drug discovery; and food, water, and environmental quality monitoring.
In the realm of communications MEMS, the most significant products are integrated fiber-optical MEMS, used for optical switching of signals, multiplexers, filters, modulators, detectors, attenuators, and equalizers. Commercialization of these devices is expected to unclog the bottleneck currently impeding the connection of fiber-based systems. These MEMS provide optical-to-optical transfer of signalsa step up from the current optical-to-electrical-to-optical approachwhich is expected to greatly enhance fiber communications system performance.
I expect that these devices will achieve full commercialization by 2004. The great promise that these truly enabling optical MEMS devices offer has been validated recently by the acquisition of optical MEMS companies. Among these is XROS (a 95-person company not currently in volume production), acquired by Nortel Networks for $3.25 billion in stock; Intellisense, 67% of which was acquired for $500 million by Corning; and Cronos Technologies, acquired for $750 million in stock by JDS Uniphase, itself a new economy company. Cronos, a 1999 spin-off of MCNC, has been working with Lucent Technologies to produce an optical fiber switch. In addition, Cronos has developed a MEMS-based switch for nonoptical communications systems applications.
In yet another potentially highly explosive growth opportunity, a number of companies including Infineon, Raytheon, and Rockwell are exploring the use of radio frequency (RF) MEMS for cellular phone and other wireless applications (more than 200 million cell phones were sold in 1999). Here again, MEMS are enabling new economy applications. The University of Michigan, led by Professor Clark Nguyen, has pioneered this technology, which includes circuit tuning elements (capacitors/inductors), resonators, filters, and switches. These low-loss, ultraminiature and highly integrative microwave functions can and will eventually replace classical microwave circuit elements/functions and enable a new generation of RF transceivers. I am not currently aware of any startup RF MEMS companies, although Raytheon is shipping products containing RF MEMS devices. But I expect that the development of RF MEMS will likely come from intellectual property companies or existing RF integrated circuit/wireless companies. Why? Because these circuits will be required to work integral (monolithic) to an RF circuit rather than from a hybrid or discrete approach for them to be optimally efficient. RF MEMS are expected to achieve full commercialization status by 2005.
Infrastructure in Place
As with new-economy stocks, 1999 was a major turning point in the MEMS industry, where unprecedented growth, wealth creation, and venture capital/angel/corporate funding took place. The infrastructure barrier that existed just a year ago is crumbling rapidly, and this progress indicates a true maturing of the MEMS industry.
No longer is it necessary to build your own fab to create a MEMS product or invest hundreds of millions of dollars to become a MEMS player. The newly created infrastructure is readily available to support both emerging and proven MEMS-based applications. For instance, more than a dozen MEMS wafer foundries exist today, among them Onstream, Sony, and Standard MEMS. High-volume/low-cost advanced packaging foundries such as Teledyne and Microassembly Tech nologies have been created. Reliability analysis support is available from such organizations as Exponent Technologies and Sandia Labs. So is high-volume testing, from ETEC. Microcosm Tech nol ogies, Inc., through its strategic partnership program with infrastructure providers, serves companies requiring MEMS by orchestrating a complete design-to-delivery planor by providing whatever pieces of the plan are needed.
Figure 2 is a MEMS industry report card, wherein Ive graded the market in various key areas.
A MEMS industry roadmap, sponsored by the Semiconductor Equipment and Materials International (SEMI) organization, has been initiated to share pre-competitive information on the processes, technologies, applications, and markets for MEMS. Interested parties can find out more at www.roadmap.nl/. You can also get further details on the state of this market by attending the Commercialization of Microsystems 2000 Conference scheduled for September 59 in Santa Fe, NM. Sponsored by the newly created Micro and Nanotechnology Education Foundation, the conference will provide a forum for MEMS developers, foundries, equipment suppliers, venture capitalists, and users of this technology to discuss their common interests. The event is expected to draw more than 300 people worldwide, including technologists from many of the new MEMS companies who will be on hand to share their stories. For more on this, visit http://asm.unm.edu/mot/coms/COMS2000.htm.
For further explanation of these grades and subjects, see Commercialization Issues of MEMS/MST/Micromachines: An Updated Industry Report Card on the Barriers to Commercialization, Sensors Expo Anaheim 2000 Proceedings (also obtainable from www.rgrace.com).