Re: when water freezes does it always freeze to the same size?

I think this question might be based on a misconception, so let me address that first, then get to the question itself. Gasses have a much higher compressibility (which relates pressure to volume) than liquids. To satisfy yourself of this, you could wrap a balloon over the neck of a soda bottle, mark the line where the surface of the liquid rests, then open the cap. Both the gas and the liquid are at the same pressure before opening the bottle. You'll obviously see the balloon fill with gas once the cap is opened, but you'll see that the soda volume hasn't changed at all (and if it appears to change, it's most likely because the bottle deformed in some way, raising the level.)

If you were to take an 'open' container deep under water, instead of expanding the contents would be compressed. The compressibility of the gas and liquid would not be changed. So, the gas would be much more compressed than the water. If the container were sealed underwater and brought to the surface, both the gas and liquid would be at the same pressure. When we let the two decompress, we would observe much more gas expand out of the container than water.

What about a solid? Solids are also compressible, but even less so than liquids. Here's where we really have to break down what's physically happening in your scenario into small pieces (and I'll have to skim over some parts because this is a complex problem...) First, we have a bowl of water at some depth - I'll arbitrarily specify that the depth results in exactly 10 atm pressure. The ice will freeze at a lower temperature at high pressure (graph at caltech.edu) 28.5 F at the surface, 28.4 F at ~100 m (10 atm) See the equation of state calculator here.

The molecules in the container are now frozen - and lower in density than the water around it (i.e. the ice expanded.) I'm not readily finding values for: density of the ice under pressure, expansivity of the ice under pressure, etc, so I can't give a value for how much it has expanded or how much it has been compressed from the pressure. However, compressibilities of solids are much smaller than for liquids so I expect it to have a minimal effect.

I'm not clear if in your question you wanted to bring this ice to the surface... If you were to do so, I expect you would probably have cracks form or enlarge, but again I have no numbers.

I don't think using ice for a telescope lens is desirable simply because too many complicating factors exist, and there are too many well-understood alternatives with better optical qualities. Consider this list of variables which would affect your 'lens' formation: salinity, temperature, pressure/depth, presence of sediment/algae/etc, speed of freezing (ice cubes in your freezer get that hazy look in the center because they freeze from the outside-in - layered freezing is one way to get the ice clear, as in an icicle). Plus, you would have to control temperature and pressure once it's made, and hope that pressure changes don't result in clouding. Existing materials don't have any of these problems, are usable (I suppose) in an aqueous environment, can be finely polished, and have better refraction/less wavelength range of absorption (quartz, for example.)