Slashdot Mirror
First Experimental Demonstration of a Trapped Rainbow Using Silicon
KentuckyFC writes Back in 1947, a pair of physicists demonstrated that when a beam of light reflects off a surface, the point of reflection can shift forward when parts of the beam interfere with each other. 60 years later, another group of physicists discovered that this so-called Goos-Hanchen effect could sometimes be negative so the point of reflection would go back toward the source rather than away from it. They even suggested that if the negative effect could be made big enough, it could cancel out the forward movement of the light. In other words, the light would become trapped at a single location. Now, physicists have demonstrated this effect for the first time using light reflected off a sheet of silica. The trick they've employed is to place a silicon diffraction grating in contact with the silica to make the interference effect large enough to counteract the forward motion of the light. And by using several gratings with different spacings, they've trapped an entire rainbow. The light can be easily released by removing the grating. Until now, it has only been possible to trap light efficiently inside Bose Einstein Condensates at temperatures close to absolute zero. The new technique could be used as a cheap optical buffer or memory, making it an enabling technology for purely optical computing.
It's a wave guide. Light travels along it, but because of the diffraction grating on the top, light gets shifted backwards as it bounces off the top. The trick is different frequencies get shifted back by different amounts, so they change the spacing of the grating over the length of the device, so that first one color gets stalled, while the rest move on and then the next stalls, until they're all stopped. However, they aren't trapped forever, the waveguide has "ohmic loses", which means light is absorbed by guide and so it eventually is lost.
The paper is entirely numerical simulation, despite what the linked blog post says. I quote: "In this paper, we numerically demonstrate an approach..". I'm not denigrating numerical simulations: I'm a computational physicist. Just, you know, RTFA?
It means an identifier that is not the actual name or not an actual literal description. Nothing is implied about accuracy.