Microelectromechanical systems (MEMS) are poised to leverage the economies of scale that have driven the semiconductor industry, enabling a bright future of new devices with lower cost, superior performance and increasingly diminutive size. So says MEMS expert Kaigham (Ken) Gabriel, the co-founder of Akustica Inc. (Pittsburgh) and a veteran MEMS developer.
With MEMS accelerometers making it into million-unit consumer devices such as Nintendo's Wii controller, and with MEMS microphones breaking into the cell phone market, Gabriel predicts billion-dollar MEMS markets over the next few years.
"We see the total available market for Akustica's CMOS MEMS technology at greater than $40 billion," he said, counting acoustic, inertial (accelerometers and gyroscopes) and RF (radio-frequency oscillators) MEMS devices "and accompanying IC functionality which we can address with our products."
The use of CMOS "will allow Akustica to uniquely drive cost-performance trends in MEMS products that will be reminiscent of the cost-performance trends we've seen for decades in microelectronics industry," Gabriel added.
Today Akustica parades its "world's first" status as the only maker of single-chip MEMS microphones. Rival Knowles Acoustics (Itasca, Ill.) has had success with what it touts as the world's first surface-mount MEMS microphone, albeit a two-chip solution in a single package. And rival, Sonion MEMS A/S (Roskilde, Denmark), is primed to begin volume shipments of MEMS microphones from its new production facility in Ho Chi Minh City, Vietnam.
"Akustica is the only company with a single-chip MEMS microphone," he said. "But because there are now two strong players, we've already seen an increase in the adoption of MEMS microphones."
Akustica puts all the functions of a miniature electret condenser microphone--including a discrete FET, a separate operational preamplifier chip and an A/D converter chip--onto a single CMOS MEMS microphone chip. But in the cell phone market, the company is at a disadvantage vis-à-vis Knowles, said Marlene Bourne, president and principal analyst at Bourne Research LLC (Scottsdale, Ariz.). That's because Knowles' analog MEMS mic doesn't require as large a redesign as a digital microphone.
"Of course, there is voice-over-IP [Internet Protocol] and the PC market to sell digital MEMS microphones into, and there are all sorts of applications, like hearing aids and automotive and some really interesting industrial applications," Bourne said. "But MEMS microphone chip makers have to go after a volume market like cell phones in order to lower the cost of their chips."
Although it has research projects under way for other types of MEMS chips, such as inertial and RF devices, Akustica plans to concentrate on microphones for a few years. Its single-chip mic has landed several design wins in PCs, including Fujitsu's Lifebook PC and the TabletKiosk Sahara Slate. Next, Akustica plans to diversify into consumer devices.
Analyst Bourne predicts that by 2010, three quarters of all cell phones will benefit from MEMS. The total could be as high as 600 million units, according to forecasts by Yole Development (Lyon, France).
By then, the VoIP and Bluetooth markets will be maturing, said Stephan Ohr, research director of analog semiconductors at Gartner Dataquest (Stamford, Conn.). According to Ohr, 580 million phones will have Bluetooth headsets by 2010. Knowles has the lion's share of today's more than 140 million-unit market.
"Knowles has been selling to cell phone makers for five years--they are by far the biggest player in MEMS cell phone mics," said Bourne. "MEMS Tech of Malaysia is a close second, and then you have companies making good progress, like Akustica, Sonion MEMS and Infineon."
"Near the end of the year and into 2008, we expect to see our digital microphone family of products move into the cell phone market," said Gabriel.
In the beginning
Akustica has built its MEMS development on five years of accumulated intellectual property licensed from Carnegie Mellon University and invented by Gabriel, formerly an EE professor there. Earlier, Gabriel researched MEMS at AT&T Bell Labs, spent five years managing MEMS development at the Defense Advanced Research Projects Agency (Darpa) and, in between, scoped out Japan's MEMS efforts at the University of Tokyo.
MEMS first caught Gabriel's attention after he joined the machine perception group at AT&T's Bell Labs in 1985, where he worked on teleoperating micron-size robots.
"I needed a way to make very small sensors, mechanisms and actuators for teleoperators that worked at very small micron scales," said Gabriel. "When I found out about bulk micromachining silicon, I started etching pincers and other mechanisms for teleoperated robots, such as force-feedback systems so that teleoperators could feel the tiny objects they were manipulating."
Gabriel quickly discovered the limitations of MEMS techniques in those early days and was eager to explore other people's solutions. So in 1990 he became a visiting professor at the University of Tokyo, managing separate MEMS development projects at IBM's Tokyo Lab, Ricoh Central Research and Toyota Central Research.
"At that time, the U.S. was afraid the Japanese were going to get ahead of us technologically," Gabriel said. "So when I left Bell Labs in 1991 to teach MEMS at the Naval Research Lab, Darpa became aware of my experience with Japan's MEMS efforts, and in 1992 I was recruited by Darpa to start their own MEMS effort, which was specifically targeted at keeping ahead of the Japanese."
Japan's Ministry of International Trade and Industry (since renamed the Ministry of Economy, Trade and Industry) had just started its 10-year "MEMS and Micromachines" initiative. But Japan never became the MEMS powerhouse that the United States feared. Gabriel claims that was because Japan was on the wrong track.
"The Japanese saw MEMS as an extension of microminiaturization--their purpose was just to attain higher precision," he said. By contrast, "I saw MEMS as important because it was a means of building electromechanical structures that used semiconductor fabrication techniques and thus could leverage the global semiconductor industry with all its associated fabrication, materials and control knowledge."
Gabriel left Darpa for a joint Carnegie Mellon appointment in electrical engineering and robotics. It was there that he developed the intellectual property on which Akustica bases its MEMS devices .
Gabriel co-founded Akustica in December 2001 to develop the microphone design he had begun at the university into commercial products. Even though Gabriel did the development, the initial four patents on MEMS microphones belong to Carnegie Mellon. Akustica licenses them.
Future directions
Besides microphones, Akustica sees its future in other MEMS chips for consumer products, such as inertial, navigation and RF devices. Because its MEMS chips are fabricated on standard CMOS lines available at most foundries, the company claims all of them will enjoy the economies of scale that result from scaling down device sizes as fabs migrate to more advanced nodes.
"The total market will be huge, and microphones are just the beginning," said Gabriel. "By leveraging the global semiconductor industry--from design tools to foundries and automated test--Akustica's CMOS MEMS technology platform is changing the way MEMS sensors will be developed and delivered to the market."
Consumer products will likely incorporate multiple MEMS devices, some of which will be used in different ways on one platform. For instance, on a cell phone an accelerometer can be used to scroll the display by tipping it, for dead-reckoning navigation when out of radio contact, for image stabilization of the built-in camera and for auto-shutdown sequences if the handset is dropped.
"Inertial MEMS sensors [mainly triaxis accelerometers] are being incorporated into mobile phones for use in image stabilization, gaming, natural user interfaces, position tracking [augmenting GPS] and system protection," said Gabriel. "MEMS devices will continue to find homes in all portable electronic devices. As these devices get smaller and at the same time require additional functionality, more opportunities will become available for MEMS product solutions."
Gabriel cites lower implementation costs, better performance and higher reliability as MEMS' advantage. For example, "in the PC microphone array market, Akustica can show the voice quality benefits for VoIP and other speech applications, with easier system design and manufacturing, lower implementation cost and higher reliability that can be gained by using our microphones," he said. "It's a very compelling demonstration that speaks for itself."
Moreover, he said, "Because Akustica also works closely with codec manufacturers as well as PC market leaders, we are able to bring a whole-product solution to our customers--this is an important part of what OEMs are looking for."
