It's kind of a tricky concept I assume, on one side you got those neat shared vertices of SiO2, on the other (water) you don't really have shared vertices, only kinda (but they still want to align themselves)

But what truly defines a glass is the existence of an alternative-form crystal that doesn't have time to form during the solidification, due to high viscosity and stuff.

Is ice the true crystal? If not it's a glass, right?


A crystalline substance doesn't necessarily have to be a single crystal to be deemed as such. An amorphous solid such as glass doesn't exhibit a crystalline structure even at very high levels of magnification. Glassy substances have a glass transition phase that is lower than the melting temperature. The melting point of ice formed under ordinary circumstances is ice Ih, with a melting point of 0° C. In other words, the ice we typically see is a crystalline form of ice. There are a number of other forms of ice; fortunately Kurt Vonnegut's ice nine (Cat's Cradle) is not one of them. Ice IX (as opposed to Vonnegut's ice nine) exhibits a tetragonal crystalline structure and only forms below liquid nitrogen temperatures. The ordinary ice Ih we typically encounter exhibits a hexagonal (think snowflakes) crystalline structure.

Amorphous ice can be made, but it's rather hard to prepare. Extremely pure water is needed, and the water needs to be supercooled very rapidly to liquid nitrogen temperatures. One can supercool liquid water, but any substantial disturbances or vibrations (e.g., looking at it cross-eyed) will result in that water instantaneously freezing into ordinary, everyday crystalline ice. If you quickly cool pure water far below its nominal freezing point and somehow avoid invoking any significant disturbances, you can indeed create a glassine form of ice.

  • $\begingroup$ Interesting. I wonder if there's also a way of transitioning from the crystalline ice to the glassy one without passing through water? Maybe by increasing the pressure while cooling or something... $\endgroup$
    – guillefix
    Oct 14 '14 at 15:58
  • $\begingroup$ @VolkerSiegel - Yes, that was slang for no disturbances whatsoever. I'll edit my answer to make that clearer. $\endgroup$ Oct 14 '14 at 16:15
  • $\begingroup$ I already imagined someone would mention super-cooling super-fast. Interesting that you need super pure water, want to specify the reason? $\endgroup$ Oct 14 '14 at 17:15
  • 1
    $\begingroup$ @fullmooninu: impurities in the water tend to act as nuclei and trigger crystallisation. Eliminating impurities retards crystallisation and makes it easier (though it's still hard!) to make an amorphous phase. $\endgroup$ Oct 14 '14 at 17:27
  • 1
    $\begingroup$ @fullmooninu - To add to what John Rennie wrote, impurities create heterogeneous nucleation sites that act as a catalyst that trigger crystallization. Eliminating those sites is one necessary step to making amorphous ice. You also need to eliminate disturbances such as vibrations (e.g., from the motor that circulates the coolant), and you also need to eliminate my apocryphal cross-eyed looks. As John wrote, it's hard to make amorphous ice. $\endgroup$ Oct 14 '14 at 17:45

Ice, as in the ice you see in cold weather, is most certainly a crystal. In fact ice has at least 11 different crystal forms depending on the temperature and pressure. The form we see on a cold day is ice 1h.

A glass, i.e. an amorphous solid, is a material that has no long range order and shows a glass transition rather than a distinct melting point. It is possible to make amorphous ice but it's hard, and it normally exists only in outer space.

From your question I wonder if you're mixing up a glass phase with a metastable phase. The 1h form of ice is the most stable under usual terrestial cold conditions, but there are other forms that are metastable under these conditions and will slowly turn into ice 1h. However being metastable does not make a phase glassy.

  • $\begingroup$ I dunno, "metastable" is such a meta word :). Things tend to deal with instability in an array of proccesses. Glass is just one example. You got others: stuff like military transitions, like twinning, and stuff with in-between dispositions like vitroceramics. I would define metastable as something that forms differently "elsewhere" and keeps being stable at NTP. So no, I was talking about glass. And if ice is really as it would like to be, or not. Sigh... categorization is so dangerous. Sorry about the ambiguity. $\endgroup$ Oct 14 '14 at 17:24

Not the answer you're looking for? Browse other questions tagged or ask your own question.