Optical computers have long held the promise of incredible computing power, and now a team of researchers led by the University of Rochester is getting a chance to put its collective research together to show how some of its cutting-edge work, like bringing light to a near halt, can carry that promise to fruition. The project is spurred on by a Defense Advanced Research Projects Agency (DARPA) award of $2.2 million initially, potentially worth up to $6.5 million over the next four and a half years, and includes researchers from Cornell University, Duke University, the University of California at Santa Barbara, and the University of Southern California.

"We've got an all-star team tackling some of the toughest problems in all-optical processing," says Robert Boyd, professor of optics and physics at the University of Rochester and principal investigator on the project. "We've already made significant headway individually, and now we're going to increase our success by combining forces. We're very excited to be working on this." Boyd notes that all but one of the six researchers involved is either an alumnus of the University of Rochester or a current faculty member. John Howell, Assistant Professor of Physics, is a member of the team and is working on all-optical switching at the single-photon level.

All-optical signal processing has implications far beyond all-optical computers; the telecommunications industry, for instance, shuttles tremendous amounts of data via fiber optics, but must convert most of it back and forth to the electrical domain in order to route or process it. That conversion costs time and money and restricts the rate at which information can be transmitted. A more efficient method would be to process the incoming photons directly, operating at the speed of light and without suffering the inefficiencies associated with the conversion from optical to electrical and back to optical.

One of the ways that Boyd and coworkers at the University of Rochester's prestigious Institute of Optics have already developed to manipulate light comes from their research into room-temperature devices that slow the speed of light to a comparative crawl. In order to process information, light or electricity must be stored as well as shuttled. Whereas electrical signals have proven simple to store, light wants to race along at speeds up to 186,000 miles per second. Researchers have slowed light before, but Boyd's recent work using lasers and a piece of ruby did so with an incredibly simplistic system. Instead of the complex, room-filling mechanisms previously used to slow light, the new apparatus is small and, in the words of Boyd, "ridiculously easy to implement." It uses a laser to "punch a hole" in the absorption spectrum of a common ruby at room temperature, and a second laser shines through that hole at just 127 miles per hour—5.3 million times slower than light's normal speed.

Boyd's device can act as a buffer, either storing optical information for a short time when used in a computer, or as a delay element in a fiber optic transmission line. When signals from separate fiber optic lines merge, the two signals may reach the merging router at the exact same moment and need to be separated slightly in time so they can be laid down one after another. Like two cars merging on a highway, where one may need to slow down to let another car into the lane, a light-slowing device could help ease congestion on fiber optic lines and simplify the process of merging signals on busy networks.

The DARPA project involves the development of additional means of slowing down the speed of light, some of which may prove more suitable than the ruby approach for specific applications. Two telecom engineers are included in the research team to make sure that advances in slow light research are quickly coupled into modern technology. Other researchers in the consortium will be focusing on developing all-optical switches—routing devices that can control light with light—again avoiding the drawbacks associated with conversion. Optical filters that can be "tuned" via light are also under development, as are detectors that can sense a single photon without destroying its properties.

This project, titled "Applications of Slow Light in Optical Fibers," is funded by DARPA through Major John R. Lowell, program manager in DARPA's Defense Sciences Office.