By Alan Leo
Coming soon: nanoelectronics for infotech and medicine.
Three years ago, when Rice University chemist James Tour pitched his nanotechnology startup to investors, he had a hard time getting anyone to listen--despite his track record as one of the world's most accomplished experts in nanoscience. Today, Tour says those same investors are all ears. "After working in this area for 13 years and having people say, 'That's pie in the sky. It'll never work,' it's gratifying to see some validation from the investment community," he says.
To call it "some validation" is putting it mildly. The company Tour cofounded in 1999, Molecular Electronics, was one of the first to seek to commercialize scientific breakthroughs in nanoelectronics. But in the last year alone, with advances coming faster than almost anyone had predicted--and with venture capitalists suddenly interested--dozens of nanotech companies have formed, backed by hundreds of millions in investments.
While Molecular Electronics plans to build computer memory using individual molecules to store bits of information, others are taking aim at ultrasensitive biological sensors, or flat-panel displays, or nanoscopic lasers. What these efforts have in common is an ambition to use components mere nanometers (billionths of a meter) in size to replace conventional electronics. "Things have gone crazy in the last year," says Paul Weiss, a chemist at Pennsylvania State University. "We're a lot further than we thought we'd be a year ago."
Nanotechnology will likely affect vast sectors of the economy, from biotechnology and health care to energy. But if scientists like Tour and Weiss are correct, the biggest impact will come from nanoelectronics. For electronics manufacturing, the promise is smaller, faster and cheaper products than conventional approaches could ever achieve. And advances have come with remarkable speed. In 1998, researchers struggled to rig up a single nanoelectronic component: a molecule that acted as a rudimentary switch. Research teams now are connecting dozens of these nanoscale components and are looking to the next step: how to assemble entire devices, such as memory chips.
Today, silicon microchips have features as small as 130 nanometers. But continuing to shrink silicon chips is getting expensive and difficult. "At some point, silicon is going to run out of steam," says John Rogers, director of nanotechnology research at Lucent Technologies' Bell Labs and member of the 1999 TR100. "You're going to need something else." Something, Rogers says, like transistors the size of single molecules. Although still at least a decade from commercialization, chips built using these molecular transistors are the industry's best hope for building faster, cheaper computers well into this century.
"With the electronics we're talking about, we're going to make a computer that doesn't just fit in your wristwatch, not just in a button on your shirt, but in one of the fibers of your shirt," says Philip Kuekes, a computer architect at Hewlett-Packard Laboratories. Kuekes and his colleagues are designing circuits based on perpendicular arrays of tiny wires, connected at each intersection by molecular transistors. By the middle of the decade, Kuekes says, Hewlett-Packard will demonstrate a logic circuit about as powerful as silicon-based circuits circa 1969. "We're trying to reinvent the integrated circuit--with its logic and memory and interconnects--with a consistent molecular manufacturing process," Kuekes says.
Well before the first shirt-thread computer boots up, however, companies will begin to integrate nanoelectronic components, including tiny wires and ultradense computer memory, into conventional silicon electronics. Hewlett-Packard and Molecular Electronics, for example, both plan to have prototype memory devices ready as early as 2004. Devices that store a bit of data in a single molecule could provide thousands of times more storage density than the electronic memory currently used in computers.
Researchers are also working with nanoelectronics to develop new biological and chemical sensors not possible with conventional technology. University of California, Berkeley, chemist Peidong Yang is one researcher developing such sensors from silicon nanowires. Yang explains that contact with even a single molecule changes the wires' electronic state. Researchers can measure that change to identify unknown molecules for purposes of diagnosis or pathogen detection.
To fully realize the possibilities of nanoelectronics, however, researchers must clear several major hurdles. First, they must build robust nanoelectronic components that function as fully, reliably and efficiently as silicon--no small task, given the semiconductor's 50-year head start. Last fall, Bell Labs' Hendrik Schon made significant strides toward that goal by fabricating a molecular transistor that matches its silicon cousins in one key characteristic: gain, or amplification of current as it passes through the transistor. Without this amplification, the electrical signal quickly fades, and multiple devices can't work together as complex logic circuits. "We can not only switch with this device but amplify the current; therefore these transistors are suited to be building blocks of larger circuits," says Schon.
But these "little test pieces" are only half the battle, says Nobel laureate Richard Smalley, professor of physics at Rice. "One has to be able to develop ways of having the [pieces] go of their own volition to where you want them." Billions, even trillions, of molecular transistors could fit on a chip--far too many to arrange one by one. Adds Mark Ratner, professor of chemistry at Northwestern University, "You want this to become so automatic that any bozo can do it."
One of the most promising approaches is called "self-assembly" and hearkens back to biology. "Nature already does a wonderful job" of assembling molecules and other nanoscale components in complex patterns, says Angela Belcher , a chemist at the University of Texas at Austin. Belcher is growing multiple generations of viruses and bacteria, seeking to evolve traits such as "protein handles" that would bind with carbon nanotubes--pipelike molecules prized for their strength and electrical properties--and deposit them in patterns useful for nanoelectronics. Learning how to grow nanoelectronics this way may take a while, says Belcher. But a functional nanoelectronic device "seems a lot closer than it was supposed to be a couple of years ago."
It's this new promise that has sparked the rash of startups in the field. "A lot of VCs and investors are looking for the next great waves," says Steven Jurvetson, a venture capitalist at San Francisco-based Draper Fisher Jurvetson and member of the 1999 TR100. "Nanotechnology is one of the great technology opportunities with wide applicability." Jurvetson counts in his company's portfolio three nanoelectronics concerns. And his firm isn't alone. According to VentureSource, venture capitalists invested over $100 million in nanotech-related startups in 2001. But, Jurvetson says, investors should beware. "The prefix 'nano' shouldn't follow the same blind enthusiasm as the '.com' suffix did," he says.
Indeed, for the moment, conventional microelectronics companies have nothing to be afraid of. But the recent advances in nanotech have many researchers convinced that they have a fundamentally new technology on their hands, one that will greatly expand the possibilities of electronics. One important thing to remember, says Bell Labs' Rogers, is that the most eagerly awaited applications may not be the ones that eventually help change how people live. "The people who invented the transistor probably did not imagine a laptop computer," he says. "It's just hard to anticipate these things."
Nanoelectronics is very much in its infancy, and researchers like Schon and Tour freely acknowledge that they are still uncertain where it will have its initial impact. But at least people are now paying attention to--and even investing in--the possibilities.
Alan Leo is a staff editor at technologyreview.com.