A critical comparison of procedures for evaluating rate constants in thermal degradation, illustrated by pyrolysis-gc results from four polymers

When studying the thermal degradation of macromolecular compounds that give volatiles, several methods are available for verifying the extent to which the rate of evolution of volatiles is first order in the mass of remaining sample. Five such methods are discussed, and each has its own demands for experimental data, and its own way of plotting the data. Each of the methods also provides a value for the first-order rate constant. For a given sample, it is shown that the value of the rate constant obtained by each of the methods is not necessarily the same. A principal reason for this is the different way in which the asymptotic limit of product yield is explicitly or implicitly determined in each of the methods. This problem is exacerbated when the evolving volatiles leave a residue that becomes progressively different from the original sample. A further problem arises when the method requires that the same sample is repeatedly heated and cooled in order to obtain the kinetic data. The evidence then suggests that the data become more unreliable as the heating cycles progress. Pyrolysis-g.c. results for PMMA, PIE, PSt and PBuAc are used to illustrate the utilisation and limitations of the methods. The general conclusion is that the 'normal log plot method' is best for obtaining absolute values of rate constants, and that sequential methods may be convenient for comparing relative rates. (C) 1998 Elsevier Science Limited. All rights reserved.