By Bob Haavind
Editor in Chief
While development of microelectromechanical systems (MEMS) has been going strong all over the world for at least a decade, Europe has been in the vanguard.
European companies undoubtedly see a market like MEMS, with specialized processing technology and many specialized niches, as more their cup of tea than commodity devices like CMOS chips or DRAMs. Just two to three years ago there were great hopes for optical MEMS devices to knit vast fiber networks into the premiere 21st century broadband delivery vehicle. But that dream fizzled with the dot-com bust along with the financial meltdown of the telecom carriers funding the build-up, and the subsequent worldwide economic downturn.
So now what?
That was the issue addressed by speakers at an international MEMS/MST Industry Forum in Munich before the SEMI Europa 2002 event this spring. While a wide array of potential uses were discussed, the main hopes for major growth appear to be RF/wireless devices and biomedical applications. Already there are markets for sensors, for acceleration, inertial measurements, process controls, and micro-fluidics.
Air bag sensors is the first of what is hoped will be many applications in automobiles, and MEMS have also found wide application in ink-jet printers, image projection, and heads for disk drives. The military and commercial satellite developers, which are willing to pay well for high performance and miniaturization, remain a core market. But with over 100 European companies making MEMS devices, wide support at research institutions, and a few MEMS foundries taking root in Europe, hopes are rising for some more high-volume markets.
Growth of 10 to 15% per year is projected by Yole Development, Lyon, France, and a Nexus market survey. The market is expected to reach more than US$50 billion by 2004 for over 70 applications, according to Eric Mounier of Yole. In 2000, Europe employed over 4,000 workers on MEMS projects, processing over 500,000 150mm equivalent wafers, Mounier estimated.
An on-going analysis of the potential for RF MEMS, reported by Jeremie Bouchaud, Wicht Technology Consulting, Munich, Germany, is looking at the promise for miniature, low power, high performance devices such as tunable capacitors, FBAR resonators, high-Q inductors, and microantennas.
Bouchaud feels the RF switches hold the most promise of all of them, mainly because there are no reliable alternatives at the higher frequencies (20-100 GHz). PIN diodes have high losses and mechanical switches are too bulky, and both are very expensive. He projected that RF switches will rise from near zero units in 2002 to more than 800 million by 2007. In the near term, however, RF MEMS are not an enabling technology for 3G handsets, so less costly alternatives will probably be used there.
Instead, their application will more likely be in base stations, he believes. He projects a total market for RF MEMS of a little more than US$1 billion by 2006. Packaging costs are critical, and perhaps wafer-level packaging could lower the costs for commercial devices, Bouchaud suggested.
There are promising signs that an initial MEMS marketplace is beginning to take shape, suggested Henne van Heeren, OnStream MST, Eindhoven, Netherlands. These include the entry of some big companies into the field, more attention to quality, some real investment, and increasing activity in the Far East with a rise in the a number of new foundries, especially in Taiwan and South Korea. In 1999/2000, the emphasis was on acquisitions, but now it has shifted over to expanding infrastructure, van Heeren said.
“The optical hype is over, and the life sciences and RF areas have become hot,” he commented.
ASML has just introduced a MEMS stepper, he pointed out, and many foundries and some semiconductor fabs are being equipped to provide enabling process technologies. Specifically, he mentioned the need for equipment for LIGA high aspect ratio lithography, bulk micromachining, surface micromachining, wafer bonding, deep reactive ion etch (DRIE), and electroplating.
The wide range of applications is requiring diverse, often customized, packaging, according to Doug Feicht, TechSearch International. The devices are often delicate, with micromechanical structures that must interact with external environments. These must be hermetically protected, often on alumina bases with metal housings, although vendors often strive for non-hermetic alternatives.
Still, they must be low cost: Accelerometers and gyros sell for less than US$5. Feicht detailed many of the innovative devices used in autos, projection displays, optical systems, liquid/gas pressure sensing systems, and biomedical applications. In the life sciences, he described a disposable blood pressure sensor for use in an intravenous line. About 17 million were sold with an ASP of about US$10. Other medical applications include implantable devices, for endoscopes for example, and laboratory-on-a-chip devices.
In Europe, the microsystems field has now grown to over 5,000 workers, with 2,500 in research, according to Marcel Hugen of the European Commission. There are over 80 companies now involved, with 150 factories, and 100 semiconductor foundries with 800,000 wafer capacity, he said, and there is a rising number of CAD vendors and design houses to support commercial development.