Most of the electronic gadgets being made today have accelerometers. By detecting motion, they protect data by stopping laptop hard drives and rotate iPhone images when the screen is turned. Accelerometers have become so ubiquitous because they have been integrated into processing chips as micro-electrical-mechanical systems, or MEMS. Otherwise, they would require too much power and cost a lot more.
By modeling MEMS, David Bindel is helping circuit designers integrate other mechanical components important for signal processing into chip production. "Then you could make something that has much lower power requirements, is much cheaper to fabricate because it's all built using the same processes, and is much more compact, so you could have your Dick Tracy watch that's essentially operating off of one chip," says Bindel. "People actually have cell phone watches now, but they've got a dreadfully short battery life."
Bindel writes his calculations into software that gives circuit designers an idea of how a design on paper will function in the real world without fabricating an expensive prototype. "These resonators are vibrating at high frequencies," says Bindel. "You want to be able to tell how much they interact with their environment both in terms of allowing forcing signals to come in and also in terms of how much accidentally leaks out to the environment."
MEMS resonators lose energy to damping, when vibrations travel through their base and into the substrate of the chip, which from the perspective of the resonator is an infinite expanse capable of absorbing all its energy. "You have to have very low losses, otherwise you're hosed before you even start," says Bindel. "So the design problem there is, 'How do you make something that has these extremely low losses given that the fabrication technology is not great?'"
Recently, Bindel has also been working on similar problems for quantum mechanical resonances, like an electron escaping a quantum corral. "There are all sorts of problems in nature where resonance problems come up," he says. "Anywhere you've got the possibility of radiation off to infinity, there's probably a resonance behind it."
Bindel, who prefers to do his calculations with a fountain pen, describes himself as "basically a numerical linear algebra guy." He has always liked math and programming. "I started programming when I was 8 on a Commodore 64 and never looked back," he says. "I'm one of the few people I know where everybody was able to predict my career path. I figure I'm lucky that way."
Before coming to Cornell, Bindel was a Courant instructor at New York University for three years, where he taught undergraduate and graduate students. "You can get a long way on an ounce of organization and a lot of enthusiasm," he says. "Students really pick up on enthusiasm and they seem to care a lot about whether you care about and are engaged in the subject."
Bindel chose Cornell because it was a place he could easily work at the intersection of math, computer science and engineering. "For somebody who is between fields like me, it's important to have an environment that's supportive of advising and collaboration across disciplines and in general supportive of faculty members who work in the way that I do," he says.