Documenting the Coming Singularity

Wednesday, June 20, 2007

How to Catch Photons and Accelerate Computing Speed

Have you ever thought about catching a photon? Before you assume that I am insane, consider that a photon, a particle of electromagnetic radiation, moving at the speed of light (naturally), should be catchable. All you would need to do is close the entrance before the photon can get out, once it comes in. And have some way to keep it from being absorbed or otherwise interacting with the inner surface of the trap. Right?

Something like that has been just been done by researchers at Cornell University. To what end? you ask. Transmitting signals in computers can be done better and faster with photons than with electrons, but photons are more difficult to control and so routing becomes an issue. Today, this routing of photons has to be done by converting the light pulses to electrons and back again to photons. Not very efficient. So a practical way of controlling the light pulses was needed.

Enter the Cornell researchers, who have developed a way to trap pulses of light and thus route them efficiently.
The new device relies on an optically controlled "gate" that can be opened and closed to trap and release light. Temporarily storing light pulses could make it possible to control the order in which bits of information are sent, as well as the timing, both of which are essential for routing communications via fiber optics. Today, such routing is done, for the most part, electronically, a slow and inefficient process that requires converting light pulses into electrons and back again. In computers, optical memory could also make possible optical communication between devices on computer chips.

Switching to optical routing has been a challenge because pulses of light, unlike electrons, are difficult to control. One way to slow down the pulses and control their movement would be to temporarily confine them to a small continuous loop. (See "Tiny Device Stores Light.") But the problem with this approach is getting the light in and out of such a trap, since any entry point will also serve as an exit that would allow light to escape. What's needed is a way to close the entryway once the light has entered, and to do so very quickly--in less time than it takes for the light to circle around the loop and escape. Later, when the light pulse is needed, the entryway could be opened again.

The Cornell researchers, led by Michal Lipson, a professor of electrical and computer engineering at the university, use a very fast, 1.5-picosecond pulse of light to open and close the entryway. The Cornell device includes two parallel silicon tracks, each 560 nanometers wide. Between these two tracks, and nearly touching them, are two silicon rings spaced a fraction of the width of a hair apart. To trap the light in these rings, the researchers turned to some of their earlier work, in which they found that the rings can be tuned to detour different colors by shining a brief pulse of light on them.

Light of a certain color passes along the silicon track, takes a detour through one of the rings, and then rejoins the silicon track and continues on its way. However, if the rings are retuned to the same frequency the moment after a light pulse enters a ring, the light pulse will circulate between the rings in a continuous loop rather than rejoin the silicon track and escape. Tuning the rings to different frequencies again, such as by shining another pulse on one of the rings, allows the light to escape this circuit and continue on to its destination.
I don't know how the light feels about being trapped, but since they are quickly let go again I don't think they mind too much. Stay tuned.

Singularity & The Price of Rice is updated daily; the easiest way to get your daily dose is by subscribing to our news feed. Stay on top of all our updates by subscribing now via RSS or Email.