Chemistry Reference
In-Depth Information
transition point from the 16-water cell without extrapolation to the thermody-
namic limit. As discussed in Section II.A, the density dependence of the transi-
tion is carried by the
E
i
, and the
A
vib
have minor effect. Significantly, Umemoto
et al. [62] demonstrated that this is sufficient to describe the decrease in VII/VIII
transition temperature with increasing density. This finding is quite remarkable
because H-bonds are beginning to show effects of symmetrization in this range.
The vibrational free energy does have an important contribution to the pressure-
density relationship at very low pressure, although the effect diminishes at higher
pressures.
Despite the challenge posed by small energy variation among H-bond isomers,
our results qualitatively match the observed features of the ice VII/VIII phase
transition in several respects: (1) the calculated ground state is the known ice VIII
antiferroelectric structure [128], (2) the transition temperature, 228 K, is similar
to the experimental transition point measured in the range 263-274 K [47, 130],
and (3) the detectable partial ordering above the transition and partial disordering
below the transition, as measured by the entropy at the transition, is in agreement
with experiment.
2.
Site Disorder in Ice VII
The proton-ordering phase transition of ice VII to antiferroelectric proton-ordered
ice VIII is, in many respects, considered to be well characterized. What is not well
understood, however, is the apparent site disorder of the oxygen atoms suggested
by recent neutron diffraction studies [47, 48, 133]. Studying ice VII in its region
of stability, Kuhs et al. [47] achieved a good fit to neutron diffraction data treating
the thermal motion of the oxygens anharmonically. This model led to a surpris-
ingly short O-D distance of 0.89 A that lengthened by 0.13 A when this distance
was allowed to vary freely after another refinement with a rigid water geometry.
From this analysis, it appeared that there was significant motion along the set of
axes suggesting disorder of the oxygen atoms about their perfect lattice sites.
Here, the symbol
100
stands for all axes related by symmetry to the [100] axis,
for example: [100], [ 100], [010], [0 10], [001], and [00 1]. Jorgensen and Worlton
[132] investigated the structure of ice VII using a time-of-flight neutron diffraction
technique. They also found it difficult to separate the static and thermal displace-
ments of the oxygen atoms. Introducing a multisite model for the hydrogens did
elongate the O-D distance somewhat, but it suggested D-O-D angles that did not
seem probable. Thus, they also concluded that some displacement of the oxygen
along the
100
directions, away from the covalently bonded deuteriums, would
allow for the lengthening of the O-D distance.
More recently, an investigation by Nelmes et al. [48] refined neutron diffraction
data using various combinations of multisite models for both the oxygen and
hydrogens treating the thermal motion harmonically. A good fit of their models
was measured by the closeness of the water molecule geometry to that found in
100
