Liquid crystal glycolipids - Carbohydrate Chemistry: Chemical and Biological Approaches - page 335

On heating compound 23 it appeared to melt at approximately
80-100 1C to form a cubic phase, which, in the polarized microscope,
gave a black isotropic texture. Rotation of the microscope stage con-
firmed the dark texture to be isotropic as expected for a cubic phase. The
material when heated further melted to a liquid at 227 1C. When cooled
from the isotropic liquid back into its self-organized state, compound 23
did so via a different route. Instead of forming a cubic phase, this time a
lamellar phase dominated. Focal-conic domains were observed along the
air-mesophase interfaces of the specimen indicating that the phase was
lamellar. This behaviour was rationalised in the following way. Com-
pound 23 melted from crystals (obtained from recrystallization from a
solvent) to a cubic phase, which melted again potentially via a columnar
phase near to the clearing point and then to the liquid state. On cooling
the material has had time to find a minimum energy structure in the
liquid, and when it transformed back to the condensed phase, it did so
this time via a lamellar phase. This study calls into question the
thermodynamic stability of such phases without understanding the
kinetics associated with molecular dynamics.
As materials are built with more complicated molecular architectures,
kinetic properties might be expected to become increasingly relevant,
with the solid state becoming less crystalline in nature, and being more
likely to be a glass. Dendritic systems are quite common in this aspect.
Sugar moieties can be used in the central scaffolds, and their stereo-
chemical structures can impart structural disorder to the overall
dendritic structure, resulting in the suppression of crystalline order, in
favour of amorphous, plastic-like states. Tetramer 24 has a glucose core
unit with cyanobiphenyl units attached to it by spacer chains. 36 The
spacer lengths were set to be either short (C 5 ) and thus with relatively
poor flexibility, or long (C 10 ) that had a much higher degree of
flexibility. In the first case the overall molecular structure could be
considered as integral, whereas for the longer spacer the biphenyl units