Supercooled Water
December 22, 2011
I've used
chemical thermodynamics
very often to predict what will happen when
materials
are combined at a given
temperature
. When I presented my results, I was always careful to point out that
thermodynamics
gives answers for
systems in equilibrium
, and the
kinetics
of a
reaction
are another thing altogether.
You only have to mix
hydrogen
and
oxygen
together to see that this is the case. Nothing happens, unless an
electrical spark
or a
catalytic surface
is introduced into the system. When that happens, you're quite convinced that the predicted reaction takes place.
Water
has many unusual properties, so many that there's a web site dedicated to them.[1] These unusual properties are likely the result of water being small
molecule
, and the forces holding these molecules together in a
liquid
or
solid
coming from
hydrogen bonding
. There are sixteen solid phases of
water ice
, named
ice Ih
(our "normal" ice) to
ice XV
, existing at different temperatures and
pressures
, as shown in the figure.
Temperature-pressure
phase diagram
of water. Ice IV and Ice XII are
metastable
, so they are missing from this equilibrium diagram.
(Image by Cmglee, via Wikimedia Commons)
.
It's well-known that liquids can be
supercooled
; that is, they can be cooled below their
freezing points
without
solidifying
. Water can be supercooled, so a fair
scientific
question is how far can water be supercooled, and what are the factors that govern the degree of supercooling. These questions were addressed in a recent study by
Valeria Molinero
, an
assistant professor
of
chemistry
at the
University of Utah (Salt Lake City, Utah)
, and her
doctoral
student, Emily B. Moore.[2-3]
The properties of supercooled water have been investigated
experimentally
, but only at the lowest temperature of -42°F (232 K). This is the
homogeneous nucleation
temperature. Below this temperature,
crystallization
of water is too rapid for the properties of the remaining liquid to be measured.[3] For this reason, the Utah chemists used computers at the University of Utah's
Center for High Performance Computing
instead of
test tubes
.[3] Supercooled water was both
simulated
at the molecular level and
modeled
using experimental data.[3]
An important first step in their research was optimization of the
computer codes
used in such modeling and simulation. They were able to speed computations by a factor of 200, but computation still took thousands of hours of computer time on their simulation of freezing of a 32,768 molecule cluster. The number 32,768, of course, is 2
15
.
They calculated the
heat capacity
,
density
and
compressibility
of water as it is supercooled; and they also simulated the rate of ice crystallization.[3] Their calculations extended to -55°F, the temperature at which the maximum crystallization rate is obtained, and also the lowest possible supercooled temperature.[3]
Molecular simulation of water crystallization from supercooled water. Regular ice is shown in red, and "intermediate ice" is shown in red.
(University of Utah image)
.
So, exactly how does supercooled water freeze? As water cools, its structure becomes more and more like that of ice. As water approaches -55°F, there's a large increase in the proportion of water molecules that bond with others into a
tetrahedron
, an assemblage of molecules very much like that in ice. Molinero calls this phase "intermediate ice."[3] The calculations showed that
amorphous ice
, which is
glassy
ice formed in a cooling process too rapid to allow crystallization, is about 25% intermediate ice and small ice crystals.[3]
All of this molecular dynamics is reflected in the thermodynamics. At -55°F, the density decreases, and the heat capacity increases, as does the compressibility.[3] The heat capacity and compressibility of supercooled water would become
infinite
at ~225 K (-55.4°F), so there's a connection between these thermodynamic quantities and the crystallization rate of ice.[2]
References:
Martin Chaplin, "Anomalous properties of water," London South Bank University Web Site
.
Emily B. Moore and Valeria Molinero, "Structural transformation in supercooled water controls the crystallization rate of ice," Nature, vol. 479, no. 7374 (November 24, 2011), pp. 506-508
.
Lee J. Siegel, "Supercool - Utah Chemists: Water Doesn't Have to Freeze until -55 F," University of Utah Press Release, November 23, 2011
.