Join 3,512 readers in helping fund MetaFilter (Hide)


Liquid mirror telescopes
July 1, 2008 3:08 PM   Subscribe

A liquid mirror telescope is made by spinning a reflective fluid, such as mercury, at a constant rate. This rotation produces a parabolic surface, which is an ideal shape for a telescope mirror. (You can try this yourself.) While these mirrors can be built to be large and orders of magnitude cheaper than solid mirrors, they have the disadvantage that they can only look straight up. Creating mirrors this way is not new; they have a history [.ps] that dates back to Newton. However, they have recently regained attention as the technology behind proposals to build an enormous (20m+) telescope on the moon. (A less technical treatment here.)
posted by Upton O'Good (36 comments total) 2 users marked this as a favorite

 
They have a significant disadvantage: they can't track the sky. They always have to point straight up. So you can't see much, and you can't see any of it for very long.
posted by Class Goat at 3:23 PM on July 1, 2008


Also, mercury vapor is poisonous.
posted by Class Goat at 3:23 PM on July 1, 2008


Ah, but here's a neat trick - If you have a silver compound in solution, spin it up, and then precipitate out the silver with a standard redox reaction, you have a wonderful solid mirror.
posted by backseatpilot at 3:34 PM on July 1, 2008


The solution to that small problem is obviously to control gravity.
posted by Joakim Ziegler at 3:53 PM on July 1, 2008 [1 favorite]


If the mirror is spun, the liquid forms a parabolic surface. If the mirror is still and the 'room' is spun, it does not. Why is this? Isn't motion relative?

This simple question started a beautiful train of thought.
posted by Mei's lost sandal at 4:27 PM on July 1, 2008 [3 favorites]


Mercury freezes at -38.83 °C. Maybe you could spin it up then freeze it to keep its shape?
posted by Laen at 4:29 PM on July 1, 2008


As I understand it, the big problem with the proposed moon telescope is not that it points only straight up. At this point, if you could build, say, a usable 100-meter telescope -- which would have a light-collecting area about ONE HUNDRED TIMES the size of the biggest useful telescopes we currently have -- you can let the damn thing point wherever it wants.

However, the real problem is that the technology to build such a thing doesn't really yet exist. The biggest LMT being built on earth is about 6 meters, I believe, which is a respectable size, but hardly ground-breaking, and if it's not an improvement over current models, then the only-points-up thing really is a major pain.

And this idea is not just to build an LMT bigger than any that has ever been built, but to build it ON THE MOON. This requires some kind of reflective liquid material (not yet invented) which will not evaporate in the lack-of-atmosphere and will remain liquid at incredibly cold temperatures, and all materials to build it would then have to transported to the moon, where we haven't gone for, well, a while, and it would need to be constructed there, plus any repairs or upgrades would take further moon trips.

So, this would require using a material that doesn't yet exist to build something on a scale never before attempted in an area that is so damn difficult to get to that no one has bothered to go for decades. It's a problem.

This is not to say that it's necessarily a bad idea ... I know several of the astronomers who wrote the paper in the second-to-last link, and they're aware of the difficulties. The boon to science if this could be built would be amazing. But there's a reason the guy who is the main driving force behind this idea is considered a visionary by some and kind of a nutcase by others; there are a LOT of obstacles to overcome for this, so much so that only-points-up doesn't even really register as a big one.
posted by kyrademon at 4:58 PM on July 1, 2008


Backseatpilot, wouldn't all the precipitate just centrifuge to the edge of the circle and give you a ring?

Could you instead spin and freeze or chemically solidify a mold to cast or grind different mirrors?
posted by BrotherCaine at 5:09 PM on July 1, 2008


The thought crossed my mind, but you've already got the liquid being accelerated out to the edge of the disk, so I don't think the precipitate would go much further - especially if there's not much liquid to begin with. The problem would be drag induced on the particles as they separate out, but again, with a low enough Reynold's number (i.e. slow rotation) the effects would be negligible.
posted by backseatpilot at 5:49 PM on July 1, 2008


'They have a significant disadvantage: they can't track the sky. They always have to point straight up. So you can't see much, and you can't see any of it for very long'

It's certainly far from ideal, but it's not perhaps quite the disadvantage one might think at first. You could operate such a telescope in drift-scan mode. You take advantage of the fact that you read out CCDs basically along one edge, shunting charge across the surface as you go. If you match your readout speed to the speed the sky moves overhead, you can take continuous images of a stripe, a bit like having a film camera where the film moves through the camera at the same speed as the scene moves in front of the lens.

You don't get a huge amount of time on any bit of the sky, but given the huge mirror you might get and the lower costs, it could work out alright.
posted by edd at 6:02 PM on July 1, 2008


mercury + parabolic + mirror + telescope + moon is automatically cool, BUT

A huge mechanical/analog system? On the moon? Is this 1968 or 2008? Synthetic aperture CCDs, baby.
posted by DU at 6:20 PM on July 1, 2008


No idea What the Hell DU said---
but I like it, babe!
posted by Dizzy at 6:24 PM on July 1, 2008


Now that I think about this... they did this on Mythbusters to build the solar-powered death ray. Except they spun plaster, which created a parabola and then hardened, which was used as a mold for a mirror. Or they lined it with foil or something, I forget.

I still think you could deposit silver on a flat plate using this method. Once it gets up to speed there's surprisingly little radial movement in the water disk. I think, due to the radial inertia and the frictional forces in the radial direction the precipitated silver particles would want to stay at the same "orbit" that they started at.
posted by backseatpilot at 6:29 PM on July 1, 2008


I still think you could deposit silver on a flat plate using this method.

I think you are going to have problems with uneven cooling/evaporation and whatnot because of the different thicknesses of liquid at different points. Keep in mind that you want the errors to be much less than a wavelength (at least, this is the case for small amateur scopes--I assume it is also true for monstro professional ones).
posted by DU at 6:58 PM on July 1, 2008


I believe most telescope mirrors actually are currently made by spinning and cooling. But I believe they spin it to get the parabolic shape, and then coat it with reflective material afterwards.
posted by kyrademon at 7:23 PM on July 1, 2008


They are first ground then polished
posted by hortense at 9:30 PM on July 1, 2008


"Liquid-Mirror Telescopes. An old idea for astronomical imaging is undergoing a technology-driven renaissance". (Paul Hickson, American Scientist, May-June 2007 Issue, Vol. 95, No. 3, p. 216)
posted by christopherious at 11:06 PM on July 1, 2008


You don't get a huge amount of time on any bit of the sky, but given the huge mirror you might get and the lower costs, it could work out alright.

But you can only see a circle of the sky a couple of degrees longitude across.

What you're describing is known as a "survey instrument" and there are other, better ways of doing that.
posted by Class Goat at 11:08 PM on July 1, 2008


Class Goat: I won't argue with the limited patch of sky you can survey. But Schmidt cameras don't really tackle the same issue of mirror construction and cost that the liquid mirror approach does.

I'd also point out that LMTs aren't the only kind of telescope that ends up having a fixed primary mirror (or equivalent) - SALT and HET have a primary mirror that has a fixed elevation angle, although the field of view can be moved to some extent by changing the position of the rest of the optics and instruments. Also, Arecibo, while radio rather than optical, is steered not by moving the dish but by moving the instrument package over head.

Of course all those use spherical mirrors and so using a mobile package with those is going to be a lot more straightforward than if you have a fixed parabolic dish, so LMTs are going to lose out there.
posted by edd at 4:00 AM on July 2, 2008


Wait, why can't they redirect columns of sky light to the lens by means of prisms and mirrors that would handle the sky tracking?
posted by odinsdream at 10:00 AM on July 2, 2008


Mei's lost sandal (7:27 PM on July 1), don't give up your day job doing stand-up for the glitz & glamour of physics just yet.

It doesn't matter how the liquid is spun (by itself, or with the room): it will still form a section of a parabolic solid of revolution. If the bucket of mercury is in the center of rotation, it will form a symmetric parabolic mirror; otherwise, an off-axis paraboloid.

Which, interestingly, brings me to the next idea: to scan the heavens, move the bucket off-axis (spin it about a point that is a movable distance from the center of the bucket). Then you can form a controllable off-axis paraboloid that focuses a desired, controllable center-of-focus angle of the sky onto the detector array.

Et voilà!
posted by IAmBroom at 1:06 PM on July 2, 2008


Odinsdream, because you'd have to grind a near perfect mirror or prism the same diameter as your focusing mirror. IANAA, but I think at that point you're pretty close to the same expense and hassle as grinding a focusing mirror of that diameter.
posted by BrotherCaine at 1:33 PM on July 2, 2008


IAmBroom, you need to simultaneously spin it around two seperate axes to get a focus point that isn't straight up. One to form the parabaloid, and another to control elevation. You'd still only get a time slice image of a given segment of sky as spinning it around the second axis (elevation control) would cause it to sweep the sky in a circular pattern slower when nearly straight up, and faster as you approached the horizon line.
posted by BrotherCaine at 1:40 PM on July 2, 2008


Since we are talking about building this on the Moon anyway, why not build it on an asteroid or other smaller body? Then you can control the pointing by shifting the entire body around. (You could even make it closer to the size of a satellite, but I assume a system this baroque would need a human attendant.)
posted by DU at 3:03 PM on July 2, 2008


Good point, BrotherCaine. The time slice isn't so bad, if the surface area makes up for the integration time required (or is enough to get what you want to see in one pass). But still tricky.

And, DU - why would you need a human attendant? To kick-start it, or to aim it? ;) No human required. It's pretty easy to control, actually (and an orbiting satellite is actually better for this - nearly zero gravity).
posted by IAmBroom at 3:13 PM on July 2, 2008


DU / IAmBroom ... How could you build an LMT on something with nearly zero gravity? Think about it.
posted by kyrademon at 3:26 PM on July 2, 2008 [1 favorite]


But Schmidt cameras don't really tackle the same issue of mirror construction and cost that the liquid mirror approach does.

The cost-per-visible-sky is a lot lower for a Schmidt.

why not build it on an asteroid or other smaller body? Then you can control the pointing by shifting the entire body around.

Because it's relying on substantial gravity in order to form the shape, so it's only possible on objects at least a quarter the size of the moon, and they would all be too heavy to shift around like that.

If you did this on something like Phobos, surface tension of the mercury would prevent it from ever forming the required paraboloid shape. (And shifting around even something as small as Phobos would be quite a trick.)
posted by Class Goat at 5:34 PM on July 2, 2008


As long as it isn't exactly zero, it should be OK.

Wait a minute--the Moon is rotating 30 times slower than the Earth. That's .5 degrees/hour or a mere 1.4x10-4 degrees/second. That should be slow enough to do some decent imaging of nearish objects, especially with such a huge mirror.

The human attendant would be there to pick hairs out of the liquid mercury and remove dead space rats from the gearing.
posted by DU at 6:11 PM on July 2, 2008 [1 favorite]


You could build it as a tumbling tube in space. Spin it on two axis at the same time, one to provide microgravity, and the other to form the parabaloid. You'd get a 360 degree continuous pan. I'm pretty sure adjusting the angle of view beyond that would disrupt the image for a while, and you'd have to do something about the temperature and/or pressurization (does mercury evaporate in vacuum?)
posted by BrotherCaine at 8:48 PM on July 2, 2008


does mercury evaporate in vacuum?

Of course it does. That's why mercury vapor lamps work.
posted by Class Goat at 10:16 PM on July 2, 2008


Put the liquid mirror telescope in a giant centrifuge, with a high speed digital camera.
posted by yohko at 10:21 PM on July 2, 2008


Of course it does. That's why mercury vapor lamps work.

Let me rephrase the question. Is there a temperature at which mercury or a reflective mercury alloy will remain liquid at near zero pressure or is there no liquid stage in the phase diagram at that point? Also, could a tube be economically lifted into orbit that when spun up sufficientlly fast could maintain air pressure with one end open to space?
posted by BrotherCaine at 1:57 AM on July 3, 2008


(Sorry, that 1.4x10-4 degrees/second is bugging me. .5 degrees/hour is just .5 arcseconds/second which is not only more a more traditional unit but also a simpler calculation.)
posted by DU at 4:48 AM on July 3, 2008


DU / IAmBroom ... How could you build an LMT on something with nearly zero gravity? Think about it.
posted by kyrademon at 6:26 PM on July 2 [1 favorite +] [!]


D'oh!

Um, ... a magnet & a magnetic liquid?
Electrostatic attraction?

Is there a temperature at which mercury or a reflective mercury alloy will remain liquid at near zero pressure or is there no liquid stage in the phase diagram at that point?
posted by BrotherCaine at 4:57 AM on July 3 [+] [!]


Yes. Mercury at room temperature remains liquid in a vacuum - as it does in a thermometer.

The human attendant would be there to pick hairs out of the liquid mercury and remove dead space rats from the gearing.
posted by DU at 9:11 PM on July 2 [1 favorite +] [!]


Again, d'oh! Forgot about the space rats.
posted by IAmBroom at 11:16 AM on July 3, 2008


'Mercury at room temperature remains liquid in a vacuum - as it does in a thermometer.'
A thermometer's a sealed system though. It's not at zero pressure. Any evaporating liquid mercury will, once it's in equilibrium, be balanced by mercury vapour in the thermometer condensing. The same will not be true in space. It's not obvious to me it won't all just boil off, and I strongly suspect that's exactly what would happen.
posted by edd at 3:16 PM on July 3, 2008


I've read that almost all elemental substances have no liquid stage at near zero pressure, but I can't find a non-subscription phase diagram on the Internet. Remember though, if you are spinning the tube to keep the mercury at one end it will be at very low pressure rather than zero pressure.
posted by BrotherCaine at 6:05 PM on July 3, 2008


« Older The Tunguska Event....  |  Bringing Down Bear Stearns,... Newer »


This thread has been archived and is closed to new comments