Chemistry Reference
In-Depth Information
structure, unless the species is unstable in air or under the X-rays. In this case, the
low temperature could be one of the ways to reduce the damage (the easiest but not
necessarily the best). Refinements of low temperature datasets are certainly easier
because the increased resolution limit enables a larger observations/parameters
ratio. However, if a crystal structure is severely affected by poor long range order
(static disorder), a large part of the scattering which concentrates outside the Bragg
position is often understood just as uninformative background noise. If diffuse
scattering due to static disorder is large, any attempt to decrease the temperature
with the purpose of obtaining more observable intensities would just be a desperate
move and unlikely to succeed. On the contrary, one should provide thermal energy
to increase long range order, for example through high temperature annealing.
The situation is different when the disorder is dynamical in nature, as it might
occur when molecules have peripheral functional groups with enough flexibility to
show large libration (in the gas phase and likely in the solid state as well). For
example, substituted methyls are often associated with a relatively flat potential for
rotation about the pseudo threefold axis. This implies very large displacements of
the three carbon atoms. Sometime, the flat potential might display two or more
minima and the dynamic disorder can somehow be “localized” with two or more
competing conformations (see Fig.
6
). By lowering the temperature the dynamic
disorder can be significantly reduced or suppressed. In fact, one molecular confor-
mation becomes favored over the other(s) either because the shape of the potential
is itself modified or because the thermal energy is much reduced and less stable
conformations become unavailable.
Other typical disorder conformations in the solid state are those of E-stilbenes or
azobenzenes, where the two atoms of the central double bond are often involved in a
complicated dynamic process (called
pedal motion
; see Fig.
6
). Many crystallo-
graphic studies have been dedicated to analyze this kind of structural feature,
including theoretical modeling of the dynamics; see, for example, Harada and
Ogawa [
54
] and references therein. It is important to stress that this kind of dynamics
severely affect the equilibrium positions of atoms refined from X-ray (or neutron)
diffraction data, especially if the disorder cannot be satisfactorily modeled. As a
consequence, geometrical parameters calculated from refined coordinates are typi-
cally quite incorrect (with severe underestimation of bond lengths). This is in general
true when the thermal motion of a given molecule in a crystal is large. Therefore, a
correction for libration, translation, and coupled translation/libration is necessary
[
13
] to extract reliable bond distances from a set of refined coordinates. Unfortu-
nately this correction is seldom applied and theoretical chemists often use uncor-
rected geometries as benchmark experimental results to test ab initio calculations.
For minerals and inorganic samples, low temperature is almost useless to
improve structure solution and only marginally relevant to improve the refinement,
unless dealing with host-guest materials like zeolites. In facts, for harder materials
ambient temperature is already quite comparable and sometimes lower than the
Debye temperature. Therefore, resolution is seldom a limitation for structure refine-
ment of minerals at ambient temperature. On the contrary, for macromolecules and
especially for proteins, the low temperature significantly increases the number of
