No abstract

Atomic mean-square displacements for frequencies above 10" Hz have been determined from neutron time-of-flight data in the glassy, liquid and crystalline phase of selenium. As in many other substances, the difference between the values in the disordered and the ordered phase shows a strong increase with increasing temperature, setting in at about the glass temperature. The inverse of that difference turns out to be linearly related to the logarithm of the viscosity. The relation holds from below the glass transition temperature up to above the melting temperature, covering twenty decades in the viscosity.

The glassy dynamics of poly(propylene glycol) (PPG) and poly(dimethyl siloxane) (PDMS) confined to a nanoporous host system revealed by dielectric spectroscopy, temperature-modulated DSC and neutron scattering is compared. For both systems the relaxation rates estimated from dielectric spectroscopy and temperature-modulated DSC agree quantitatively indicating that both experiments sense the glass transition. For PPG the segmental dynamics is determined by a counterbalance of adsorption and confinement effect. The former results form an interaction of the confined macromolecules with the internal surfaces. A confinement effect originates from an inherent length scale on which the underlying molecular motions take place. The increment of the specific-heat capacity [Formula: see text] at the glass transition vanishes at a finite length scale of 1.8 nm. Both results support the conception that a characteristic length scale is relevant for glassy dynamics. For PDMS only a confinement effect is observed which is much stronger than that for PPG. Down to a pore size of 7.5 nm, the temperature dependence of the relaxation times follows the Vogel-Fulcher-Tammann dependence. At a pore size of 5 nm this changes to an Arrhenius-like behaviour with a low activation energy. At the same pore size [Formula: see text] vanishes for PDMS. Quasielastic neutron scattering experiments reveal that also the diffusive character of the relevant molecular motions --found to be characteristic above the glass transition-- seems to disappear at this length scale. These results gives further strong support that the glass transition has to be characterised by an inherent length scale of the relevant molecular motions.