Molecular motions, the α relaxation, and chain transport in polyethylene crystals

Abstract

AbstractConformational energy calculations are used to investigate molecular motions in polyethylene crystals. From these a model is derived for the motion that accomplishes the net rotation translation that is believed to underlie the nuclear magnetic resonance (NMR) and dielectric α processes in polyethylenes and paraffins (and their dipole decorated derivatives). The resulting model is found to incorporate features of a number of previous models but differs significantly from all of them. The rotation is accomplished by means of a twisted (by 180°) region that propagates smoothly along the chain across the crystal. It differs from previous rotational models in that the twisted region is found to be rather localized (to ∼12 CH2 units). A dependence of activation energy on chain length (paraffins) or crystal thickness (polyethylenes), with the activation energy becoming independent of thickness in thick crystals, results not from the rotational lattice mismatch of the twisted region per se but from the translational lattice mismatch induced by the 180° rotation of one stem relative to the other. The twist differs from a stable point‐defect twist previously proposed (Reneker defect) in that the chain torsion is relatively uniform through the twist and there is no shortening of the chain accompanying it. Further, the twist propagates smoothly without local barriers to its advance. Thus the propagating twist is to be through of as a transition state rather than a hopping defect. Detailed atomistic conformational energy calculations on C22H46 crystals were combined with a simplified elastic theory for translational stem mismatch in longer chains. From the combined calculations the activation parameters for twist propagation as a function of chain length or crystal thickness could be calculated. The results were compared with experiment for the dielectric α relaxation in paraffins containing dissolved ketones and polyethylenes containing carbonyl groups. The agreement is quite good, especially considering the paucity of adjustable parameters in the model. There is only some slight uncertainty in the calculated entropy of activation and a scattering correction was made a posteriori to account for a slow continued drop‐off in relaxation rate in very thick crystals.