In this study, wide‐line NMR and x‐ray diffraction have been used in conjunction to study the crystal structure of poly(vinylidene fluoride). Drawn poly(vinylidene fluoride) film was found to contain two crystal phases, the relative amounts of each depending on the draw temperature. Drawing at 50°C. yields a single phase, designated as phase I, while drawing at temperatures between 120 and 160°C. yields a mixture of phase I and a second phase (phase II). The fraction of phase II increases with increasing draw temperature, but this phase was never obtained without some phase I. A tentative orthorhombic unit cell is proposed for phase II. The structure of phase I has been determined from x‐ray data. The unit cell is orthorhombic, space group Cm2m, having lattice constants a = 8.47, b = 4.90, and c (chain axis) = 2.56 A. There are two polymer chains in this unit cell. The conformation of the polymer chains is planar zigzag. The details of this structure have been confirmed by experimentally determining at −196°C. the change in the NMR second moment with the angle between the magnetic field and the draw direction of phase I (drawn at 50°C.), and by comparing these results with a theoretical calculation of the second moments, based on the atomic positions obtained from the proposed structure.

Measurements of the NMR second moment of a uniaxially oriented sample of polyethylene single crystals in the range of temperatures from −196°C to 130°C and its dependence on the alignment angle γ between the orientation axis (preferential direction of the molecular chains) and the NMR magnetic field are presented. The experimental results are discussed mainly with respect to the high temperature relaxation, called the α process, in polyethylene. They are compared to theoretical predictions made for a number of mechanisms of molecular motion in Part I of this work. Only one of the mechanisms considered is found to be in quantitative agreement with experiment, the mechanism here referred to as flip‐flop motion. This consists of thermally activated rotational jumps of the crystalline chain segment between folds around its axis between two equilibrium sites in the lattice. Each rotational jump through 180° is accompanied by a shift of the molecule along its axis by one CH2 group. The discussion of the low‐temperature relaxation of polyethylene, the γ process, is based partly on the above measurements and partly on measurements of second moments for unoriented polyethylene samples varying widely in morphology and noncrystalline content. The decrease of the second moment observed with these samples between −196°C and 20°C is taken as a measure of the intensity of the γ process. A linear correlation is found between the decrease in the second moment, designated ΔS, and the noncrystalline content, 1 − αm; this can be represented by ΔS = 1.4 + 22.1(1 − αm). It is shown that neither the crankshaft mechanism not the kink mechanism is able to account quantitatively for this result. The model of a chain end moving in a vacancy fails to adequately describe the angle dependence of ΔS in oriented polyethylene single crystals. The “sandwich model” of a polyethylene single crystal, in which the crystalline core is covered by noncrystalline surface layers, is in better agreement with observations.

Raman scattering from the longitudinal-acoustical vibration of single crystals of polyethylene has been observed using an argon laser and an iodine filter. The frequencies observed are in the range 10-40 cm-! and are inversely proportional to the thickness of the single crystals which varied between 90 and 250 A.

A Raman band of low frequency, arising from an accordionlike vibration of all‐trans \documentclass{article}\pagestyle{empty}\begin{document}$ \rlap{‐‐} ({\rm CH}_2 \rlap{‐‐})_n $\end{document} segments and previously observed in normal paraffins and in polyethylene single crystals, has now also been found in bulk and in cold‐drawn polyethylene, both linear and branched. The accordionlike vibration, or longitudinal acoustic mode (LAM), in polyethylene is compared with the LAM in normal paraffins. Whereas the Raman bands corresponding to the third (LAM‐3) and higher modes are quite intense in a long‐chain paraffin such as n‐C94H190, they are so weak in polyethylene as to be unobservable with the apparatus used. This is attributed to the presence of the chain fold in polyethylene. Of the two extreme structural models of the fold here discussed, namely the models of “tight folds” and of “loose loops,” only the latter seems capable of accounting for the weakness of LAM‐3 and higher modes in polyethylene. A quantity called “nominal Raman length” is defined as the length of that all‐trans n‐paraffin that would have the same LAM‐1 frequency as the polyethylene sample under consideration. The nominal Raman length is always greater than the average long spacing, deduced from discrete x‐ray scattering at small angles after applying a Lorentz correction, and, after allowing for chain tilt, is found equal to the segment length between folds. This can be accounted for by both of the models mentioned. As a test of the theory of surface melting the frequency of the accordion vibration of annealed polyethylene single crystals was measured as a function of temperature up to the melting point; no frequency change with temperature was observable. On the basis of the naive idea that there is complete decoupling of the vibrations in the all‐trans chain segment from the disordered (molten) surface layer, one would predict that upon surface melting and the concomitant shortening of the all‐trans segment, the LAM‐1 frequency should increase. A more careful analysis, taking into account the existence of coupling of the LAM to the surface layer, shows that the outcome of this experiment does not necessarily invalidate the idea of surface melting. Bulk polyethylene samples exposed to 60Co γ‐radiation for doses up to 100 Mrad show a slight shift of the Raman band to lower frequencies, whereas no such shift was observed upon absorption of a swelling agent. A search, without success, was made for a longitudinal acoustic mode in polypropylene, poly(vinylidene fluoride), nylon 66, and polyoxymethylene.