Liquid Crystals and Lasers

Wan2Be writes "A new kind of glass pane that quickly switches from transparent to diffracting and back again. The change is triggered by applying an electric field, so the pane could easily be controlled by the electric signals of a computer, offering a powerful new way to steer beams of light."

Comments

Re:Too slow for communications

[js7a]

given enough time, it could catch up with desirable data speeds

Liquid crystals are really slow. You really have to struggle to get 35 ms switching out of them, which is what the movie people want. These diffraction switches are a lot faster, but I doubt they'll ever get faster than 0.5 ms.

The basic problem is that you're dealing with long strands of polymers which orient themselves almost but predictably not quite parallel to each neighboring polymer strand, unless a current is flowing, in which case they just align strictly parallel to each other. It takes time, lots of time, for the reconfiguration to occur, because it is a fundamentally mechanical process.

The article cited did a horrible job IMHO of representing the underlyng science. The regularity of the crystal droplets has nothing to do with the new effect -- which is using a thinner layer of liquid crystals to difract instead of polarize, which requires thicker LCs. The droplet regularity is an artifact of the thinness, not a cause of the essential property as is purported.

Re:Too slow for communications

[shamino0]

LCs are very slow compared to what is nowadays the speed bits traveling along a glass fibre. I cannot see a useful way of using it directly to redirect or modulate laser light.

LCs are definitely too slow to be used for something like modulation or per-packet switching, given current data rates (an OC-192 fiber carries approximately 10Gbps, or one bit per 0.1ns).

But there are other uses. It is still useful for pure optical circuit-switching applications, where you want software to set up an optical circuit that will not change for a long time (hours, days, maybe even years) until it is explicitly reprovisioned or rerouted.

This is currently done with very small electro-mechanical parts that can optically redirect light from each input fiber to the appropriate output fibers. A system that can do it using a diffracting LC (such as that described in the article) would be able to do it with no moving parts - greatly increasing reliability. It would probably also be faster, although that's less important for this application, given the large amounts of time that the optical circuits are likely to remain established.

Depending on how precise they can control the diffraction, it might even be possible for one LC to diffract multiple wavelengths in different ways, and be able to control them individually. If this can be accomplished, it makes Wave Division Multiplexing (WDM - where multiple optical circuits are carried on one fiber using multiple wavelengths) easier to implement, since you would no longer have to separate the wavelengths into separate fibers before switching them.