A quantum chemical study of the degradation and the maximum polyene length in PVC

Abstract

The semi-empirical MINDO/3, AM1 and MNDO-PM3 methods were employed to study the stability of polyenes, polyenyl cations and polyenyl radicals. Both cations and radicals were subjected to computation since the elimination of HCl from PVC might proceed according to an ionic or radical mechanism. We subsequently investigated the effect of additional alkane chain ends, as well as PVC-like chain ends to both ends of the polyene sequence, in order to mimic degraded PVC better. The results are discussed in relation to the experimentally observed polyene length distribution in degraded PVC, as revealed from UV-Vis and Resonance Raman Spectroscopic data. The following conclusions may be drawn from the data presented. The MINDO/3 method has to be rejected. From the calculations, which basically refer to the molecules being in the gas phase, it was concluded that the formation of radical intermediates during polyene formation is much preferred (energetically) as compared to the corresponding ionic species. However, several researchers have suggested that the ionic process is preferred in solid PVC (‘solvation effect’); for solid PVC hard experimental evidence is still lacking with respect to either of the two mechanisms. The introduction of saturated chain ends markedly influences the energetics and the stability of the polyenes. HCl abstraction from an intact PVC chain was calculated to be about as probable as HCl abstraction adjacent to a polyene sequence. Calculations on polyenes with PVC-like chain ends do not give any indication that there is a maximum polyene length after which further elimination of HCl becomes less probable from an energetic point of view. The calculated data presented show no evidence for a finite polyene length, and thus do not support the unzipping mechanism. We propose a new one-step unzipping mechanism.