Slashdot Mirror
Deformable Liquid Mirrors For Adaptive Optics
eldavojohn writes "Want to make a great concave mirror for your telescope? Put a drop of mercury in a bowl and spin the bowl. The mercury will spread out to a concave reflective surface smoother than anything we can make with plain old glass right now. The key problem in this situation is that the bowl will always have to point straight up. MIT's Technology Review is analyzing a team's success in combating problems with bringing liquid mirrors into the practical applications of astronomy. To fight the gravity requirement, the team used a ferromagnetic liquid coated with a metal-like film and very strong magnetic fields to distort the surface of that liquid as they needed. But this introduces new non-linear problems of control when trying to sync up several of these mirrors similar to how traditional glass telescopes use multiple hexagonal mirrors mounted on actuators. The team has fought past so many of these problems plaguing liquid mirrors that they produced a proof of concept liquid mirror just five centimeters across with 91 actuators cycling at one kilohertz and the ability to linearize the response of the liquid. And with that, liquid mirrors take a giant leap closer to practicality."
Probably becaue you lose the biggest advantage of the "liquid mirror". With a liquid, you can make a very large, very thin spinning surface which will keep its perfect shape because of the motion. Now, freeze that, peel it off --- how do you keep the shape? If it is thick, it will pose the same problems as any large mirror - heavy, unwieldy, needs lots of time to come to equilibrium with the environment, etc. If it is thin - keeping the shape is probably hopeless.
Even if you could keep the shape somehow, freezing isn't a uniform process. As the temperature is lowered, crystals, lumps and whatnot starts forming in the melt. Some of these will inevitably go to the surface and spoil the figure of the resulting surface. And we're talking really, really small lumps here - on the order of less than quarter of the lightwave the surface is supposed to reflect. So, you'll need to work on the surface afterwards, just the same way you'd work on a surface of a "normal" mirror.
I am not sure enough effort will be saved by making the initial figure in this way vs. the traditional methods of preparing a surface for polishing to justify the spinning. Speaking from experience, "pregrinding" a piece of glass to a rough sphere with a piece of pipe (or, if you're hi-tech, a diamond saw) does a good enough job. And the professional mirror makers have more than that at their disposal.