Ice may melt when the pressure is increased, as you can find out from
other questions on the site. You may then expect that reducing the
pressure will raise the freezing temperature, and you'd be right (water
takes up more space when it freezes, and so if the pressure is lower,
it is easier for the water to expand, and so freezing is easier and
happens at a higher temperature). The only problem is that the effect
is very tiny for only one atmosphere's worth of pressure, which is the
most you can take away (there's no real limit to how much pressure you
can add, though).
One atmosphere corresponds to 30 inches of mercury in these units.
20 inches of vacuum gets us to 10 inches of mercury's pressure. At one
atmosphere, the freezing temp. of water is zero degrees Celsius. At 4.6
torr (at the triple point of water, which is at about 0.2 inches of
mercury) the freezing temp is 0.01 degrees Celsius. We went about 2/3
of the way there in pressure, so I'd interpolate the answer to be
T_water freezing(20 inches of vacuum) = 0.007 degrees Celsius, which is
very close to its freezing temperature at atmospheric pressure, but
ever so slightly higher.
Tom
(published on 10/22/2007)