For decades, electronics have become smaller and faster year by year. This size reduction and concurrent power increase have been dependent on the electronics industry consistently finding ways to increase the number of microcomponents carried by an average microchip. Lately the pace of optimization has slowed so much that some experts think we may have reached the limit of scaling down devices. This would mean the end of the hotly anticipated, ultra-expensive, new product iteration.
“If you’re in the integrated circuit business and you don’t have a better technology in three years, then competitors will catch up with you eventually. And your product will become a commodity,” says Amit Lal, Electrical and Computer Engineering.
“There’s no new performance edge that allows you to charge more by making technology smaller or by increasing performance,” Lal continues. “If you can’t fit more functionality—more transistors—on a chip, then what do you charge for and how do you maintain a healthy profit margin?”
A SonicMEMS Breakthrough
Lal is an expert on microelectromechanical systems (MEMS)—microscopic devices composed of components tens of nanometers to hundreds of micrometers in size. He and his group, SonicMEMS, focus on micromachining technologies for making ultrasonic transducers for ultrasonic applications. Through micromachining, thin-film piezoelectric transducers can be integrated directly with complementary metal–oxide–semiconductor (CMOS) electronics; so all the components—electronics and transducers—can be fabricated on one silicon chip. This mass production of CMOS/piezoelectric thin film wafers gives CMOS transistor technology a boost by making new applications of piezoelectricity available to electronics developers.
In the course of their research, Lal and his colleagues made an important breakthrough. They came up with the idea of putting a thin film of aluminum nitride directly on CMOS electronics, then applying voltage at gigahertz frequencies across the film to create ultrasonic waves. The small, gigahertz-frequency waves have many potential applications—from microchip security to acoustic storage of computer memory to ultrasonic imaging of small body parts such as individual blood cells.
“With ultrasound, even if your finger is wet, the sensor can still tell the difference between skin and water.”
SonicMEMS is pursuing many of these avenues with United States government funding from the Intelligence Advanced Research Projects Activity (IARPA), Defense Advanced Research Projects Agency (DARPA), and National Science Foundation (NSF). The SonicMEMS researchers are currently working on projects in robotics and digital agriculture, which has led them to new collaborations with colleagues in Mechanical and Aerospace Engineering and Electrical and Chemical Engineering.
Geegah, Commercializing Ultrasonic MEMS for Fingerprint Recognition and Other Applications
Recently Lal founded a company called Geegah (a meme word for Gigahertz) to commercialize the ultrasonic MEMS breakthrough, which won a Prototyping Award from the College of Engineering to transition to a startup. He is now the chief executive officer. His former graduate student Justin C. Kuo, PhD’18, Electrical and Computer Engineering, is chief technology officer. After licensing the relevant SonicMEMS patents from Cornell Technology Licensing, Geegah moved into the new Cornell Praxis Center for Venture Development.
One of Geegah’s main thrusts is to develop ultrasonic MEMS for fingerprint recognition. “Right away, we realized we could sense things ultrasonically on the backside of the silicon chip,” Lal says. “We can get very high resolution of greater than 1500 DPI (dots per inch) of anything touching the surface. That means we could make a state-of-the-art fingerprint sensor, and there’s a very large market for that.”
Smart phones currently have fingerprint sensors, based on measuring the difference in capacitance between air and human tissue. They work well until you add water to the mix. If you’ve ever tried to unlock your smart phone with a sweaty finger, you’ve experienced the problem firsthand. The sensor can no longer read your fingerprint. “With ultrasound, even if your finger is wet, the sensor can still tell the difference between skin and water,” Lal explains.
As an added bonus, the ultrasound sensors will work when the microchip is attached behind the phone’s screen glass, unlike current fingerprint sensors which must be accessed through the touch-button on the front of the phone. “The smart-phone world wants to eliminate that touch button,” Lal says. “They see it as a waste of potential screen space. I think we have a shot at succeeding in the smart phone area because we have a complete solution to the fingerprint problem, not just the sensor chip.”
Fingerprint Sensors for Credit Cards
Geegah will also take a stab at supplying the credit card market with fingerprint sensors—that is, if Lal and his cohorts can convince the banks that issue the credit cards to invest in the technology. “I would argue that credit cards need sensors more than smartphones do,” Lal says. “If they had them, the banks could be sure the card user was the authorized user. A lot of fraud could potentially disappear. It’s one of these extreme opportunities because if we can get a bank to adopt our technology, everyone will adopt it. People will feel more secure all around.”
Geegah needs to lower the barrier to the credit card industry’s adaptation of its fingerprint sensors by addressing a number of issues, Lal explains. “First, we’ve already addressed wet and dry finger compatibility, meaning the ultrasound sensor can still image your finger, whether it’s wet or dry. Then, the sensor also needs to be very thin to fit in a credit card and very low power. All of these design challenges are just the sorts of projects Geegah can address.”
Sensors for Nematode Detection in Soil
Being able to image fingerprints at very high resolution means the gigahertz ultrasound sensors can also image things as small as bacteria. A recent grant from the United States Advanced Research Projects Agency-Energy (ARPA-E) gives Geegah the support to develop sensors for use in agriculture. Also, the company is currently in talks with high-tech data companies like Google and Microsoft interested in processing the data the sensors would generate. “The ARPA-E grant funds our work on imaging soil and little worms called nematodes,” Lal explains. “Nematodes eat the roots of plants, so farmers need to apply pesticides whenever they are present.”
The Geegah researchers want to put their sensors in small cubes that could be planted throughout the field, near the plants’ roots. The sensors would send out data via remote radio frequency waves that could be picked up at a central location where data companies would process it. When nematodes or other pests or bacteria are found, the data companies would alert the farmer immediately. The farmer could then spread pesticide exactly where and when it is needed, boosting crop yields. “Food growth rates are nowhere near population growth rates. So by 2030 we may run out of water and food,” Lal says. “We have to do something to prevent that from happening and keep the cost of food cheap. This technology may be the answer.”