 |
Big Ideas for Small Devices
By Stewart Wolpin
| |
How small can cell phones get? The goal, expressed by numerous pop culture writers seeking to express the public's techno-desire, is Dick Tracy's two-way wrist radio. Unfortunately for lovers of science fiction, cell phones have pretty much reached a technological dead end. Cell phone components can't get any more compact using current chip technology, and lack of battery power limits a cell phone's data storage capacity, processing power and screen size. These limitations will be exacerbated once new third-generation cell networks are implemented. These higher-frequency, broadband digital networks will allow significantly faster data communication, but they'll also require better processing and demand a lot more power from your handset.
A solution to these technological roadblocks is MicroElectroMechanical Systems (MEMS), the technical term for a very simple idea — microscopic electrical motors or other transistorized mechanical devices that will fit on a microchip.
A long time coming
The idea behind MEMS isn't new. MEMS were conceived by legendary physicist Richard Feynman, who theorized in 1959 that size was not a barrier to advanced technology. But it took more than 30 years — and the development of powerful microprocessors etched in silicon — for MEMS physics to catch up to Feynman's theory. The first commercial MEMS appeared in cars in the early 1990s to control unpredictable airbags. Instead of expensive and dumb mechanisms that caused airbags to erupt at the wrong time, sometimes fatally, MEMS accelerators accurately sense the g-forces of a car's sudden deceleration, then determine if the airbag should be deployed. MEMS also exist in such diverse devices as ink jet printers, blood pressure monitors and digital light processor (DLP) digital video projection systems.
Despite it's relatively long history, MEMS research is still in its infancy. Researchers at such diverse institutions as U.S. Departments of Energy and Commerce, the National Institutes of Standards and Technology, NASA, Motorola, Intel, Lucent, Agilent, Seagate, Hitachi, Fujitsu, STMicroelectronics, the University of Michigan, University of South Florida, and Carnegie-Mellon University, are figuring out how to maximize the benefits of MEMS.
"This field is as open as the wild west," observes Dr. William Trimmer, founder and former editor of the IEEE/ASME Journal of MicroElectroMechanical Systems and currently president of Belle Mead Research, near Princeton, N.J. "It's like the pre-Cambrian age when all sorts of things were being tried before life figured things out."
Cell MEMS
Arguably the most compelling MEMS applications are aimed at solving the cell phone dilemma.
"Seventy-five percent of the components in cell phones are passive devices — filters, switches, inductors, variable capacitors, etc.," explains David Seeger, manager of IBM's silicon science and process technology department at the Thomas J. Watson Research Center in New York. "If you can find a way to take all those devices — and there can be more than 100 of them — and shrink them down and integrate them on the ICs (integrated circuits) in the phone, then you can be a very rich person."
For the user, MEMS means a cell phone with fewer parts — which means fewer things to go wrong — and greatly increase the battery life. "These devices are terrible power drains," Seeger says. "MEMS, with the ability to integrated direcly on-chip, need practically no power."
Additionally, so-called RF, or radio frequency, MEMS, shrink down the receiver inside the phone. An RF MEMS frequency resonator using microscopic tuning forks works to filter out unwanted frequencies, resulting in a stronger signal and a clearer conversation.
IBM's RF MEMS, which measure much smaller than an eighth of an inch square, can be built on silicon germanium, better known as SiGe (rhymes with "Ziggy"), pioneered by IBM in the early 1990s. SiGe is less-expensive alternative to the MEMS-on-chip solutions other companies are pursuing, such as gallanium arsenide (GA), a technology already used extensively in wireless chip technology. Gallium arsenide, however, though it delivers comparable speeds to SiGe, is far more expensive. IBM's MEMs-based SiGe solution is still in testing, but there could be MEMS-based cell phones on the market by the middle of this decade.
Storing big ideas in small places
| | 
Whereas the readback rate of an individual probe (above) is limited, high data rates can be achieved through the use of massive parallelism such as in the Millipede system (below).
|
Cell phones aren't the only promising MEMS development. IBM's Zurich Research Labs, in collaboration with researchers at IBM's Almaden, Watson and East Fishkill labs, are working on a MEMS data storage project code named Millipede, a memory system capable of holding 500 gigabits or more on a square inch of ultrathin organic polymer.
Millipede uses a revolutionary technology based on the Atomic Force Microscope (AFM), invented by IBM Zurich researcher Gerd Binnig in 1986. The AFM was an outgrowth of IBM's Scanning Tunneling Microscope (STM), which won Binnig and Heinrich Rohrer the Nobel Prize for Physics. In essence, STM and AFM technology enable the manipulation of actual atoms.
Millipede works a bit like a punch card puncher. The Millipede read/write head is embedded with 1,000 AFM probes fabricated on a single silicon MEMS chip. These probes creep across the media surface to either pinch nanomter-sized indentations, or pits, such as those on a CD or DVD, that represent digital bits of information, or read back or erase previously pinched nanoscale pits.
"We are aiming to replace flash," says Zurich Lab Millipede project manager Peter Vettiger, "But with a much higher capacity in the same shape and size as traditional flash memory media." Flash media, used primarily for file storage on PDAs, MP3 players and digital cameras, is a removable storage media like a floppy disk. Millipede is not a disc, but a removable system; the Millipede media and the motion mechanism is combined in one package about the same size as a flash media card. The Millipede advantage is not only more data in less space but also a lower cost-per-megabyte for consumers than flash. A 64 megabyte flash media card today costs around $100. A consumer could one day buy a postage-stamp-sized 5-10-gigabyte Millipede for potentially a lower price.
Data squeeze
To meet the challenge of increasingly data-ravenous desktop computer and server applications — such as digital video and large Web sites — there is a constant need to squeeze more data space out of existing magnetic hard drives.
Hard disk drives store data much like a vinyl record stores sound, in concentric grooves carved into blank media. These grooves are created or read by a read/write head mounted on a computerized tone arm, called a servo. The amount of storage on a hard disk is limited to the physical width of these grooves, or data tracks, and by the space between the tracks. One way to pack more data onto a disk is to narrow the data tracks and the space between data tracks. But the narrower the tracks and the space between, the more difficult it becomes for a conventional mechanical servo to accurately align the magnetic read/write head on top of a specific data track.
Enter researchers at IBM's San Jose Almaden Labs, who have created a MEMS technology called the micro-actuator. The silicon-based MEMS micro-actuator is attached to a conventional servo arm in order to move the read/write head with higher speed and better accuracy. It can be compared to your own body's movements: to pick up a large object, you move your shoulder and elbow to position your arm and fingers. But to pick up a smaller object, you need your wrist to position your hand and fingers more precisely. The micro-actuator acts as a microscopic wrist on a conventional servo arm to more accurately locate the thinner, more densely packed tracks.
According to Dr. Toshiki Hirano, senior engineer in charge of IBM's micro-actuator project, micro-actuator hard disk drives could store 30 to 40 gigabytes per square inch — quadruple the amount of storage space of today's average hard drive.
Even though the idea for MEMS is more than 40 years old, most of the important work has been done in the last decade. And most of the benefits of MEMS technology are just now reaching commercial viability, making MEMS believers out of dozens of companies and universities. "Over the last year or two, there's been a lot more industry interest in MEMS devices," observes Seeger. That heightened interest could result in a revolutionary generation of smaller, smarter and more efficient cell phones and disk drives before the end of this decade — and maybe even a two-way wrist phone.
Stewart Wolpin has been writing about technology for nearly 20 years, and is the author of "The Official Guide to Buying, Connecting and Using Consumer Electronics," published by the Electronic Industries Association.
