Re: Can you crystallize glass?

The term "freeze" implies that atoms form into a regular crystaline lattice
arrangement when converting from a liquid to a solid.  Glasses, on the
other hand, are indeed solids, but have short-range order, but no regular
lattice structure.  An excellent discussion concerning whether glass is a
solid or a liquid is presented by 
Phil Gibbs

As it turns out, when a liquid is cooled, three things can happen.  First,
it can nucleate and crystalize (freeze).  Second, as the liquid is cooled
below it normal freezing point, it may have difficulty actually starting to
nucleate, and will remain as a liquid.    For example, pure glycerine has a
freezing point of 17C, however, almost no one has seen solid glycerine
because it generally remains in a supercooled state. Third, the liquid may
become progressively more viscous until it stops moving, forming a glass. 
Sometimes solids with no regular structure are called amorphous solids. 
Mixtures of atoms/molecules with different sizes will make it more
difficult to form a regular lattice arrangement, making it easier to form a
glass or amorphous solid.

Glasses can, in fact, crystalize.  The usual term for the crystalization of
glass is devitrification.  The normal process to crystalize or devitrify
glass is by very slow cooling from the molten state; essentially keeping
the temperature hot long enough so that the atoms in the glass can
rearrange themselves into a regular crystaline state.  I have watched
quartz tubes that are used in high temperature furnaces and become
contaminated, change from a clear glass into a white, translucent,
crystaline material.  

One of the myths about glass is that it continues to flow just like any
viscous liquid.  
Robert Brill from Corning Glass Museum has written an interesting
article about glass flow.  

According to some calculations, a one meter high pane of glass one
centimeter thick will thicken by 10 angstroms in approximately 10 billion
years.  Since fluid flow in glass at room temperature is pretty slow, the
speed of devitrification isn't going to go very fast, either.  It is more
likely that the glass in 50 bazillion year experiment may actually
devitrify or crystallize in about 5 billion years as the sun grows hot
enough to engulf the earth, providing a hot enviornment for glass
crystalization over an extended time.  Unfortunately, performing a 50
bazillion year experiment is rather difficult.

Now, as to your question, what happens if you take a hot blob of liquid
glass and cool it very rapidly.  Well, the first thing that you would
suspect is that there will be absolutely no time for crystalization;
viscosity is going to increase too rapidly for the atoms to rearrange
themselves into a regular structure.  In fact, if you want to make
amorphous or glassy metals, the trick is to cool them very, very rapidly.

So, if you put a hot blob of glass into a zero K container, what would
happen.  Turns out that the rate of cooling is more important than the
absolute temperature to which the glass is cooled.  For example, if your
hot blob of glass is suspended in a vacuum chamber at 0K, it will cool
relatively slowly, since radiation heat transfer is the only way that it
will cool down, and the emmisivity of glass is pretty low.  If the blob of
glass hits a metal wall at 0K, one side of the drop will cool rapidly,
while the other side cools more slowly.  The uneven stresses may make the
glass drop break.

Prince Rupert of Bavaria provided the most interesting answer to your
question in the 1640s.  He would let heated glass drip into water, where
the drops would cool very, very rapidly.  The outside surface of the glass
would solidify, while the center initially remained molten.  As the center
of the drops cooled and contracted, the outside surface would be in
compression, while the interior of the glass drop was in tension.  A
description of Rupert Drops is found at cmog.org,
while a really great photo can be found here.

Prince Rupert drops are strong enough that they can be hit with a moderate
blow with a hammer without breaking.  On the other hand, if the tail of the
drop is broken off, or if the drop receives a small scratch, the drop
bursts into powder.  The remains of the drop, while broken and perhaps
looking like powder, have not crystalized.  If one were to examine the
remains under a microscope, there would be irregular fracture patterns,
characteristic of glasses, rather than cleavage along regular crystaline
planes.  Since the outside of the drop is in compression, and the interior
in tension, perhaps it would be appropriate to say that the drops are
attempting to implode.

The important thing to remember is that the stresses in the drops are
caused by the difference in the rate of cooling of the outside of the drop
compared with the inside, not the absolute temperature to which the drop is
cooled.  Although the outside surface of the drop will contract more as it
cools from 1000C to -273C, compared with cooling from 1000C to 10C, it is
difficult to experimentally produce a more rapid rate of cooling than
simply dropping a hot blob of glass into a bucket of water because of
water's high heat capacity, high heat of vaporization and good conductive
heat transfer.   

The rate of cooling of a surface by a boiling liquid is relatively complex,
depending upon not only the nature of the liquid, but also upon the
temperature of the solid surface.  If the temperature difference is too
high, a film of gas separates the liquid from the solid, actually
decreasing the cooling rate.  As a result, for example, a molten glass drop
will probably cool more rapidly in a bucket of liquid water than in a
bucket of liquid nitrogen, despite the liquid nitrogen being almost 200C
colder than the water.

According to the last reference mentioned above, when making Prince Rupert
drops, many of them shatter while being made.  So, the answer to your
hypothetical question is that if molten glass drops were rapidly cooled in
a 0K container, some drops will shatter during cooling because of
differences in stresses between the outside and inside of the drops, while
others may end up as Prince Rupert drops made conventionally.  If the drops
are cooled more slowly to 0K, they will remain drop-shaped, clear and
glassy.  In either case, the glass drops will remain a glass, since rapid
cooling and low temperatures are the wrong way to try to induce crystalization.

Thanks for a very interesting question.