Ask A Scientist Chemistry Archive

name         Daniel G.
status       student
age          17

Question -   I am doing an experiment in which I measure the
temperature of distilled water as it freezes.  I have found consistently
in my results that, as the water freezes, it drops below 0.0 degrees C to
as low as -0.9, then works its way back up to around -0.2 before it
completely freezes (where it then continues its original descent in
temperature).  I understand the concept of latent heat capacity, but why
does the temperature overshoot the freezing point of water then work
itself back up?
The description does fit that of supercooled water, but there must be a
factor that is keeping the water from achieving the degree of organisation
necessary to freeze at 0 degrees Celsius, such as induced movement or
possibly an electrostatic field. Neither factors are such the case here.
What else could it possible be? Does supercooling occur to an extent in all
water phase changes, no matter how stable the environment?
 ----------------------------------------
You correctly identify supercooling as the probable cause of the overshoot
of the water as it reaches its melting point. The explanation for the
supercooling is that ice has a rather intricate crystal structure and water
does have some viscosity the inhibits it motions on the molecular scale. The
purer the water, the slower rate of cooling you use, and the cleaner your
glassware  actually favors supercooling. Under very controlled condition,
water can be supercooled to
-10 C or so. Electromagnetic fields will not have much effect. Just stirring
or even tapping the container is often enough to start the freezing process.
The final temperature of about -0.2 C is caused partly by freezing point
depression caused by dissolved air, and of course, check the calibration of
your thermometer.

A classic example of supercooling and glass formation is glycerin. It has a
m.p. of about 15C but almost no one has ever seen crystalline glycerin
because its high viscosity prevents crystal formation upon cooling. The
supercooled glass is most often seen.

Vince Calder
==============================================================
Please have a look at

http://www.phase-technologies.com/html/vol._5_no._5.html
http://hepweb.rl.ac.uk/ppUK/PhysFAQ/hot_water.html

for a reasonably complete answer to this question,
as well as to the ever popular question about why initially hot
water apparently freezes faster than initially cold water.
There are references included to the primary  literature.

My compliments to Daniel on his observations.

This problem is apparently why the freezing point of water is not used to
define the Celsius scale anymore. The triple point of water is used instead
(which is much more reproducible as it turns out).

As a water clusterologist, I must say that there is a lot more to water
than meets the eye!

R. Topper
==============================================================