Slashdot Mirror
Scientists Trap a Rainbow
An anonymous reader writes to tell us that Physicists from both the University of Surrey and Salford University have devised a method to trap a multi-colored rainbow of light inside a prism. "Previous attempts to slow and capture light have involved extremely low or cryogenic temperatures, have been extremely costly, and have only worked with one specific frequency of light at a time. The technique proposed by Professor Hess and Mr Kosmas Tsakmakidis involves the use of negative refractive index metamaterials along with the exploitation of the Goos Hänchen effect, which shows that when light hits an object or an interface between two media it does not immediately bounce back but seems to travel very slightly along that object, or in the case of metamaterials, travels very slightly backwards along the object."
It would be nice if the "journalists" bothered to mention there's an article in Nature.
Kosmas L. Tsakmakidis, Allan D. Boardman & Ortwin Hess 'Trapped rainbow' storage of light in metamaterials Nature 450, 397-401 (15 November 2007) | doi: 10.1038/nature06285. (See also summary comment box, doi 10.1038/450330a.) They propose a method that might. The meta-materials needed to do this with visible light don't exist yet. Your caution is quite correct. The paper is theoretical. An actual device has not yet been built. However this result is still significant because what they are showing is that the various results on "slow light" and "trapped light" can be realized in optical metamaterials. This is significant because metamaterials are in principle more amenable to technological deployment than the more exotic techniques of slowing light (ultra-cold condensates, etc.).
It's also worth noting that metamaterials at various wavelengths (e.g. microwave band and IR) have already been made. We are getting very close to optical metamaterials. For instance, see this review of the field:
Vladimir M. Shalaev Optical negative-index metamaterials Nature Photonics 1, 41 - 48 (2006) doi: 10.1038/nphoton.2006.49.
We already have prototype metamaterials at wavelengths of 780 nm, which is on the edge of the visible spectrum. Significantly, we already have metamaterials that operate in the IR band, which is what is used for modern fiber-optics, telecommunications, etc. The materials to date are not optimized, so it will of course be awhile before all these great applications of metamaterials are implemented in real telecom devices. But, still, we are getting quite close to these applications. In particular, I expect we'll see a commercial 'rainbow trapping' device for communications before we see a commercial 'invisibility cloak'!
The main problem with such techniques is losses. Even if your mirror is 99.9% reflective (and mirrors that good are expensive, by the way), you quickly lose all your signal intensity if you are reflecting thousands or millions of times. Your idea of using a photonic crystal is neat, but you would be hard-pressed to make a very long path length without making the crystal large, too. And if you cap the end with a mirror (to trap the light for longer), you run into losses from that.
That's one of the reasons the research mentioned in TFA is significant: in principle it allows a pulse to be trapped for an arbitrary amount of time with no losses (and for a broad range of wavelengths).
*creature. Never work, eat and /. at the same time! ;o
*bows in shame*
There are similar problems with refraction: the refractive index contrast is not infinite, so some amount of light is always transmitted. At glancing angles (below the critical angle), you theoretically get perfect 100% internal reflection. This is how fiber-optics work: by having a glancing-angle internal reflection, the losses at the boundary are quite low. However the beam is then propagating inside a material, and there is absorption from the material itself. (Even if the absorption was somehow zero, the refraction at the boundary would never be perfect: imperfections and evanescent waves would cause some amount of light to escape.)
So, while theoretically one could build a light-trap using reflection or refraction, using any known material would involve some imperfections or losses preventing long-term trapping.
is there a way to make nanoscopic light frequency shifters by moving masses closer/farther near light's path, perhaps shining in a vacuum channel? In theory, yes. By moving a mass you could shift the frequency, which would change how the light would interact with materials for instance. (Note that this shift wouldn't be "permanent"; that is the frequency shift that occurs to light as it passes near a mass is exactly "undone" when it passes away from the mass and leaves the gravitational field. So the shift only occurs while the light is nearby...)
However, it's important to keep in mind that the shift you would introduce by using a nano-sized chunk of mass would be very, very, very, very small. So small, that the light's frequency would be randomly shifted by other effects (e.g. nearby cars driving by) far more than the little mass could control. Even in an isolated environment, the frequency shift would be very small (things like thermal noise and Heisenberg uncertainty would be far larger). It's only really relevant on the scale of planets and stars. What if the mass weren't matter, but more light - could that make a photonic "transistor" that either deflects or frequency-shifts light with solely photonic control? Light doesn't interact with itself. Photons have no mass hence they don't generate gravitational fields (and don't attract each other), and photons have no charge hence they can't affect each other through electric/magnetic effects. So there is no way for light to interact with light in a vacuum (except of course for light that is "coherent"; that is two beams that the same frequency and are phase-locked to each other will interfere).
However, light can interact with light in a material if the material exhibits certain "non-linear" properties (actually all materials exhibit these effects but some are more pronounced than others). This idea of a "photonic transistor" has actually long been sought in photonics: it would make for amazing all-optical computing. The idea is to interact two beams of light in a non-linear material, so that one beam can switch the other beam on and off. One simple example is a "photo-refractive": a material where the refractive index actually changes when light passes through it. The idea being that you could alter the path of one "data beam" by using a "trigger beam" (when the trigger passes through, the refractive index goes above some critical value which deflects the data beam from one exit port to another...). Thanks for taking the time with this esoteric stuff My pleasure... it's always fun to get an actual scientifically-interesting conversation going on Slashdot!