Quantitative modelling predicts the MWs, structures, and yields of 55 volatile products from the pyrolysis of polyisobutylene

Abstract

Pyrolysis–GC shows that high MW polyisobutylene can degrade thermally to give more than 60 volatile products. This paper sets out to predict their molecular weights, detailed structures, and relative yields, by quantitatively modelling the detailed pyrolysis mechanisms. There are five stages in the development: (1) A preliminary statistical analysis of experimentally observed pyrolysis products, grouped into general oligomeric types up to nonamer, leads to the deduction that random scission is the predominant mechanism leading to the wide range of oligomeric products. (2) A preliminary model is set up assuming that random scission is the sole mechanism, and relative yields and detailed structures of all products are predicted on this basis. These predictions involve the consideration of all possible scission reactions and assignments of their relative probabilities. (3) The observed monomer, dimer, and trimer yields are found to exceed the predictions of (2) the preliminary model, and these products are therefore being supplemented to this extent by other processes. The latter are deduced to be depropagation plus intramolecular transfer (backbiting) mechanisms. Detailed modes and probabilities of backbiting attack are proposed, leading to predictions of relative yields and structures of the products from this source. The relative yield of monomer from depropagation is also deduced from its excess. (4) The backbiting predictions and relative monomer yield from (3) are then combined on the same scale with the random scission predictions from (2) in order to obtain a complete prediction of the structures and the relative yields of all products from the pyrolysis. This is the required predictive model for the pyrolysis. (5) In evaluating the model, it is first of all encouraging to note that it leads to the prediction of 55 pyrolysis products, which number compares well with the 63 products detected. Also, the range of predicted yields from the model may be compared throughout the oligomer range with the experimental chromatogram, and the characteristics are found to be similar. The model is then assessed more quantitatively by comparing the predicted weight fractions of all 55 products with those observed experimentally, and a good correspondence is obtained. (6) Objectives of this approach are: (a) to assess the relative importance of possible degradation mechanisms and to elucidate which structural features of polymers principally contribute towards their thermal instability, and (b) to develop general modelling methods which may be used to predict the degradation behaviour of other polymers, and in particular to predict the more complex thermal degradation of oils. The latter are more difficult to study experimentally, and give a much wider range of products because of the large number of ends present as a consequence of their low molecular weight.