Stochastic modeling of the thermal and catalytic degradation of polyethylene using simultaneous DSC/TG analysis

Dissertação para obtenção do Grau de Mestre em Engenharia Química e Bioquímica

In the present work a stochastic model to be used for analyzing and predicting experimental data from simultaneous thermogravimetric (TG) and differential scanning calorimetry (DSC) experiments on the thermal and catalytic degradation of high-density polyethylene (HDPE) was developed. Unlike the deterministic models, already developed, with this one it’s possible to compute the mass and energy curves measured by simultaneous TG/DSC assays, as well as to predict the product distribution resulting from primary cracking of the polymer, without using any experimental information. For the stochastic model to predict the mass change as well as the energy involved in the whole process of HDPE pyrolysis, a reliable model for the cracking reaction and a set of vaporization laws suitable to compute the vaporization rates are needed. In order to understand the vaporization process, this was investigated separately from cracking. For that, a set of results from TG/DSC experiments using species that vaporize well before they crack was used to obtain a global correlation between the kinetic parameters for vaporization and the number of C-C bonds in the hydrocarbon chain. The best fitting curves were chosen based on the model ability to superimpose the experimental rates and produce consistent results for heavier hydrocarbons. The model correlations were implemented in the program’s code and allowed the prediction of the vaporization rates. For the determination of the global kinetic parameters of the degradation reaction to use in the stochastic model, a study on how these parameters influence the TG/DSC curves progress was performed varying those parameters in several simulations, comparing them with experimental data from thermal and catalytic (ZSM-5 zeolite) degradation of HDPE and choosing the best fitting. For additional improvements in the DSC stochastic model simulated curves, the thermodynamic parameters were also fitted. Additional molecular simulation studies based on quantum models were performed for a deeper understanding on the reaction mechanism and progress. The prediction of the products distribution was not the main object of the investigation in this work although preliminary results have been obtained which reveal some discrepancies in relation to the experimental data. Therefore, in future investigations, an improvement of this aspect is necessary to have a stochastic model which predicts the whole information needed to characterize HDPE degradation reaction.