Three decades ago, when the computer industry was in its infancy, its undisputed leader, IBM, had a single-word motto: “Think.”
As the 20th Century winds down, and computers are ubiquitous, an appropriate update might be: “Think Small.”
That is certainly the working slogan among Northwestern University researchers working to make new kinds of computer chips that can take the industry into the next century with technology that is far faster, cheaper and more powerful than anything now in sight.
These chips will use components 1,000 times smaller than today’s miniaturized electronic chips, and the new components will manipulate light instead of electricity to do their work.
Traditional electronic computer chips, which have been doubling in power and dropping in price and size for decades, now may be approaching the physical limits within which millions more transistors can be packed into ever-smaller spaces. Many chip developers believe new approaches are needed for information technology to meet market expectations.
At Northwestern, a research team is making devices that manipulate bits of light, called photons, instead of the electrons that are the basis for today’s computing industry.
Photons already are widely used in telecommunications to carry a million phone calls simultaneously on optical fiber, a single strand of glass as slender as a human hair.
Potential benefits of photonics are so great that at least two dozen labs across the country are working in the general field of opto-electronics, said Nick Holonyak, a professor of electrical engineering and physics at the University of Illinois at Urbana-Champaign who invented the light-emitting diode.
“New labs are popping up like mushrooms,” Holonyak said. “This is very hot.”
Using light to boost the efficiency of electronics is beguiling, but after 35 years in the field, Holonyak is pessimistic about any overnight breakthroughs.
“It’s a nice idea,” he said, “but a real dog to pull off. There are lots of technical problems to overcome.”
Instead of radical new advances, Holonyak expects a trickle of very useful niche applications to flow from the flurry of research into photonics now under way.
In Evanston, Northwestern researchers have produced photon-manipulating devices with the potential to boost the information-carrying capacity of today’s fiber by a factor of 100 to 1,000. Under the direction of Seng-Tiong Ho, a member of the electrical engineering and computer science faculty, they have created the world’s tiniest laser and more recently have produced a resonator or switch to go with it.
This is the world of nano–for nanometer, or billionth of a meter–instead of today’s world of micro, for micrometer, or millionth of a meter.
These achievements have attracted attention from outside investors who have formed a company with Ho to finance the world’s first laboratory dedicated to producing photonic devices on the scale of billionths of a meter.
Obtaining investments from venture capitalists for what amounts to basic scientific research is rare.
“These people contacted us after reading that professor Ho had created the world’s smallest laser,” said Ira Uslander, director of industry relations at Northwestern’s McCormick School of Engineering and Applied Sciences.
“They’ve formed a startup company that has raised money and fully intends to make a profit. They’ve placed no restrictions on professor Ho’s ability to publish his scientific findings, providing we get it patented first.”
One reason for the unusual commercial interest in basic research is industry’s need for some breakthrough technology to keep its frenetic pace of productivity running.
Ho is working at the outer limits of technology. He blends materials and uses the latest techniques for etching and buffing them to produce pathways for light that actually squeeze photons into a single dimension.
At their smallest scale, photons are both particles and waves at the same time, and when they are forced into spaces where there is little room to move up, down or sideways, the bits of light react in ways altogether different than they do in larger surroundings and bigger groupings.
Physicists refer to these changes as quantum effects, and studying these effects is the realm of Ho’s science.
The world’s smallest laser illustrates one of these quantum effects. Lasers are streams of light that have been concentrated into a coherent beam of a single color, or wavelength. In our human-size world, where trillions times trillions of photons have plenty of room to bounce around, only one out of 10,000 to 100,000 photons available may be recruited for a laser beam.
In the world’s smallest laser, where small groupings of photons are jammed into spaces comparable to a CTA subway car at 5 p.m., as many as seven of 10 available photons may be recruited for the laser beam, a mind-boggling leap in efficiency.
“People didn’t expect such a jump in efficiency,” Ho said. “They thought the gain would be so small from a tiny laser that it wouldn’t be worth making one.”
Ho’s latest development is a resonator, which is a small, circular photon racetrack that can be placed adjacent to a strand of fiber with a tiny opening or gap that allows certain colors of light to enter when the gap is open. The gap opens and closes as light or electricity is applied to the resonator.
This gives Ho a switch that grabs photons and sends them in a new direction at will.
Such devices are available commercially today and used in telecommunications systems. But their size is measured in microns instead of in nanometers.
Today’s photon circuits are relatively large and simple, and may pack together 4 to 10 devices to manipulate photons and cost $10,000, Ho said. His goal is a chip with 1,000 super-tiny components packed together that would use far less power than today’s circuits and cost far less as well.
“With today’s technology, you can only afford to use photon circuits in central locations,” Ho said. “When you get out in the network near where the users are, you have to switch from photonics to electronics because of the costs. We want to change that, to make it cheap to move the photonics deeper into the network.”
This revolution is what Ho’s financial backers are counting on to make their money, and their faith in his ability to deliver is manifested in their investment.
“I take early-stage gambles,” said Caisey Harlingten, founder of United States Integrated Optics Inc., the startup firm backing Ho.
“If you find a technology that’s already developed, then the guys will be there. I’m not Motorola or AT&T. I’m a small entrepreneurial risk-taker. I don’t mind raising $5 million or $20 million for something I believe in, and I believe in the story of nanophotonics.
“I believe this will be developed to the point of commercial viability within a few years.”
Although Ho’s achievements in nanophotonics make him a leader in the field, he has plenty of competition.
A team at the Massachusetts Institute of Technology last month presented results at a physics meeting in Baltimore on development of a photonic band gap similar to Ho’s resonator. The MIT team is using silicon as its building material of choice, while Ho’s team has used gallium arsenide and other more exotic semiconductors.
Ho has chosen materials with optimal qualities for generating and manipulating photons, while MIT has chosen silicon because it is the basis for today’s electronics industry.
“We’ve done work with gallium arsenide, but my focus is on silicon because that’s what the industry is most familiar with,” said MIT professor Lionel Kimerling. “We want to make photonic devices that look exactly like electronics in their size, process steps and performance. When you go to a designer with something new, he’s going to want to know how this fits into his current architecture, and if you’re talking silicon, you’re talking his language.”
Kimerling’s work with the photonic band gap was done in conjunction with Hewlett-Packard Corp., but he said virtually the entire semiconductor industry is supporting one phase of photonics research or another.
Within three or four years, Kimerling said, the industry expects to run out of tricks to extend the productivity of electronics and is looking for photonics to help out.
Kimerling doubts that photonics will displace electronics soon to produce all-optical computing.
“Electrons are good for logic and photons are good for interconnections, with their far higher information carrying capacity,” Kimerling said. “I think computing will still be electronic for some time to come, but with photons for interconnection.”
