General reflections on the symposium on physical structure of the amorphous state

Abstract

Abstract Following a discussion of the historical perspective of the problem of order in amorphous polymers and a definition of terms, this paper reviews various topics in the literature. This problem and the related one of order in liquid alkanes (as a model for polyethylene) and amorphous polymers were omitted from or inadequately treated at this symposium. These topics relate both to short-range order (>50 Å) and long-range order (>100 Å) including supramolecular amorphous structure. Among the former are depolarized Rayleigh scattering and heats of mixing of normal alkanes; positron annihilation in polyisobutylene, x-ray scattering, dielectric loss data, and the constancy of the Tg/Tm ratio. Among the latter are the Mooney-Rivlin C2 constant and the general implications of short-range order for a supramolecular structure in long-chain polymer. Since local order exists in simple liquids, it must be present in amorphous polymers. A term, correlation of molecular orientations, CMO, introduced by Bothorel for low molecular weight liquids, is employed to express a feeling for this short-range order. The connectedness of an ensemble of polymer chains implies at once some supramolecular structure (even though being X-ray amorphous and, therefore, not having supramolecular order), since any given polymer chain can experience several CMO's not only with itself but with one or more neighboring chains. This supramolecular structure is dynamic above but frozen below Tg. This postulated supramolecular structure may result in the nodules seen by electron microscopy. The key unresolved problem of this symposium concerns the reality of, and the kinetics for, the assumed reversible transformation at Tm: random interpenetrating coils in melt chain-folded crystals in solid This dilemma is posed by neutron scattering on polyethylene showing (in some cases) random coils in the melt and by various types of physical evidence (not cited at the Symposium) for chain-folded crystals in the solid. A semicollapsed coil model with local order, proposed independently by Lindenmeyer and by Arzhakov et al., is consistent with a) ¯r0 2 1/2 values observed in the amorphous state, b) the existence of local order, and c) the reversible melting-recrystallization kinetics. Factors which favor a high CMO (local order) are 1) close-packing possibilities (no or small side chains), 2) polarity, 3) a stiff backbone, 4) regularity of backbone structure, and 5) a small temperature interval between the measuring temperature and Tg, i.e., small free volume. Elastomers as a class, especially when observed at T ≥ Tg + 100 °C, are least likely to exhibit local ordering but still are not completely free from it. It is observed that the ratio C2/C1 of the Mooney-Rivlin constants decreases with average area per chain molecule, suggestive of local order induced by orientation. Finally, suggestions are made for topics at a future symposium on amorphous order.