When Vivek Subramanian was a graduate student at Stanford University in the mid-1990s, he didn't have a big food budget. So he paid close attention to the expiration dates on the items in his refrigerator.
"You know, even if food had expired, I'd still eat it," he recalled. "Expiration dates are pretty conservative."
Now an assistant professor of electrical engineering at the University of California, Berkeley, Subramanian, 30, found his academic calling by pondering a high-tech way to track food freshness.
"What we really want is communication between the food and the supermarket," he said he remembered thinking, where "the food tells you when it's gone bad."
So, he set out to develop sensors and other electronic devices that could be built into food packages to provide and transmit information about inventory and freshness. Such devices would use organic rather than silicon-based transistors.
Although food tags and other commercial products based on organic transistors are two to five years away, researchers are making progress in some challenging areas - including cost.
"We have to make them so cheap that you're willing to throw them away," Subramanian said.
Today's cheapest electronic trackers, known as radio frequency ID tags, cost a dollar or more because they use conventional silicon circuits, which require expensive lithography and processing.
An organic or plastic circuit, however, could theoretically be printed directly onto a package's surface in one step by a souped-up inkjet printer, perhaps for less than a penny. Made of carbon and hydrogen as opposed to inorganic silicon, such circuits are soluble and can be attached to such organic substrates as plastic, paper or even cloth.
"They're inherently printable," said Subramanian, who expects his team's prototype inkjet printer to produce full circuits by summer. A supermarket radio frequency tag requires about 5,000 transistors for its processor, compared with 40 million for Intel's Pentium 4. It also requires power and wireless communication circuitry, consisting of inductors and capacitors.
"There are still things we have to solve," Subramanian said. "We're literally where the semiconductor industry was in the late '60s or early '70s - we don't quite understand them, but we can kind of make them work."
Organic transistors are much slower than silicon devices, and can become useless within days or weeks if exposed to air or water.
Depending on the application, however, performance may not matter so much. A sensor designed to detect the presence of E. coli bacteria in a package of meat, for example, may work fine at the slow speeds of organic transistors.
"It's OK if it takes a few minutes to derive that information," said Thomas Jackson, a professor of electrical engineering at Pennsylvania State University. "You're not asking it to control the launching of nuclear missiles or anything like that."
Researchers predict that organic transistors will first be widely used in display technology for cell phones, handheld organizers and computers because display circuitry doesn't require very high performance and the display market is so strong. Today's "active matrix" displays contain silicon transistors, mounted on a glass "backplane," that drive each pixel.
Other potential uses for organic transistors include cheaper, faster blood tests, biometric sensors, anti-counterfeiting circuits printed onto currency, apparel that resists theft, smarter luggage tags and products that take advantage of the transistors' ability to superconduct energy at low temperatures and to drive lasers.
In addition to small academic teams like Subramanian's, companies including IBM, Lucent, Mitsubishi Chemicals, Philips and Xerox are working on such devices.