MOLECULAR MOBILITY IN GLASSES AND GLASS STRUCTURE

Abstract

We have been engaged in a study by dielectric relaxation of molecular mobility in glasses made from molecular and fused Salt liquids [1, 2]. Our frequency range, 102-105 Hz, is sensitive to molecular rearrangements in the highly supercooled condition, in the vicinity of the glass transition temperature. In glasses composed of organic high polymers, in addition to the main molecular rearrangements whose slowing down on cooling is the origin of the glass transition, one often observes other molecular processes taking place at higher frequencies at any given measuring temperature. These have been attributed to the presence in the polymer molecule of internally hindered motions of specific chemical groups such as side-chains which slow down at a different rate from the main chain motions and thus can be resolved from them [3]. The molecular glasses we studied were composed of rigid molecules, lacking either side chains or any other internal degrees of freedom capable of giving rise to a relaxation by the mechanism traditionally invoked for the polymers. The results revealed a surprising and very uniform pattern : some three-quarters of the molecular glasses show a second relaxation region (the β-process) that consistently occurs at a temperature about 0.7 to 0.8 of Tg for a measuring frequency of 1 kHz, is more spread out in frequency than the main (α) molecular process, has about 1/5 to 1/10 the strength of the a-process, and an activation energy in the range of 0.2-0.4 eV. Some three quarters of amorphous high polymers show a process with characteristics so similar that it is not possible by examination of the experimental dielectric data to tell whether the molecule in question is a polymer bearing side chains or a small rigid molecule. Our conclusion was that the secondary (β) process is a characteristic feature of amorphous packing. This conclusion has not so far been widely accepted. Evidence against it is as follows : (1) Silicate glasses have not been shown to possess a β-relaxation ; (2) we have no explanation for the fact that some 25 % of our rigid molecule systems fail to show such a relaxation [3]. A few metal alloy glasses have been studied by the analogous method of mechanical spectroscopy : The results have been inconclusive. Chen et al. have studied a Pd-Au-Si alloy and not found a secondary relaxation [4] ; Eisenberg and Reich have studied Pd-Si and Pd-Au-Si alloys and found relaxations whose presence or absence was highly sensitive to sample history [5]. In spite of these facts we feel the weight of the evidence is on the side of our basic conclusion. Structural studies on glasses have most often been interpreted in terms of the random network model. Recent evidence on the stability of small clusters of particles [6, 7, 8] has raised an alternate possibility that glasses are composed of amorphous clusters : aggregates of high density and low energy that lack the symmetry needed for long range periodic packing. Such a structure would require a somewhat looser packing of atoms or molecules in the interstices between the amorphous clusters, a feature graphically referred to as connective tissue. In contrast to the random network model, in which all the atoms or molecular units are in essentially equivalent local environments, the amorphous cluster model requires two different environments in the structure. The usual tools of structure analysis have not so far given a definitive answer to the question of whether there are one or two types of local environments in amorphous solids, and it is not yet clear whether the resolving power of these tools is adequate to the problem. The results we have cited above on molecular mobility in glasses provides evidence for the existence of two types of environments, and may be plausibly interpreted therefore as support for the amorphous cluster model.