Speaker
Prof.
C. Austen Angell
(Arizona State University)
Description
The bizarre behavior of water in its low pressure,
supercooled, domain is now well known. Unless confined
nanoscopically, response functions and transport
properties
follow power laws in T that are characteristic of
approach to
spinodal limits on phase stability. Such behavior has
provoked a great deal of computer simulation-based
investigation, key results of which turn out to be in stark
conflict with the predictions of empirical (engineering)
equations of state when extrapolated into the domain of
isotropic tension (negative pressure). This domain is the
least
investigated in all of water research, despite the
abundance,
in nature, of samples for study (milky quartz, under the
microscope). We posit that the secrets of water lurk on
the
other side of the zero pressure line.
In this talk I compare some of the predictions of
currently
favored pair potential simulation models with the best
available data for real water, in both positive and
negative
pressure domains. We first apply an approach that is
well-
tested for models with critical points, to the best
available
positive pressure equation of state, but fail to find the
isochore-crossing "smoking gun" anywhere in the
positive
pressure domain. Only spinodal divergences are
indicated,
and the value at ambient pressure coincides with the
limit of
supercooling from recent short time μdroplet studies.
Spinodals without critical points at positive pressure,
imply an
L-L coexistence line at positive pressures, rather than a
LLCP.
Turning to isobaric data, we show how models that are
in
good accord with experiment at normal temperatures
deviate
dramatically at large supercooling. While TMDs from
certain
models agree quite well with experiment at positive
pressure,
for P < -100 MPa, they vary in opposite directions1.
However, all leading empirical equations of state, (based
on
positive pressure data) accord with the direct
observations -
implying that any critical point has merged with the gas-
liquid
spinodal. Heat capacity data are especially valuable
because
of availability over wide temperature ranges (including
near
Tg) and also from very fast measurements that
postpone
crystallization while yielding faster structural relaxation
data.
We compare them with thermodynamic constructions to
argue that a first order transition near 230K provides
the
most plausible rationalization.
Finally, we introduce new data from non-crystallizing
water-
rich solutions of a previously unstudied class. Discovered
as
an offshoot of protein folding studies, these are
aqueous
solutions of hydrazinium salts that, according to melting
point
depression data, dissolve to form ideal solutions. At 15
mole% salt they yield cooling thermograms with large
endothermic spikes. These mimic the diverging heat
capacity
of pure water but are not interrupted by crystallization
(perhaps due to a sort of wall-free confinement). This
allows
both sides of the anomaly to be seen before the glass
transition intercedes. We will report new TMD data to
argue
that these spikes are the manifestation of the first order
liquid-liquid transition that would occur in water if
crystallization did not intercede.
A phase diagram from the H-bond-modified van der
Waals
thermodynamic model of Peter Poole, provides a simple
rationalization for all of these observations.
[1] Meadley, S. L. & Angell, C. A. Water and its closest
relatives: insights from metastable state studies. Nuovo
Cimento, Enrico Fermi summer school on water (in press)
arXiv 1404.4031 (2014).
