Mathematical modeling and process investigation of melt grafting reactions in an intermeshing co-rotating twin screw extruder

Due to environmental and economical factors, the use of extruders as continuous chemical reactors has grown rapidly in the past two decades. Despite the importance of twin screw extruders in "reactive extrusion" there have been limited attempts to mathematically model polymerization and polymer modification in such devices. The objectives of this thesis are twofold: to study the melt grafting of glycidyl methacrylate (GMA) onto polyethylene in the presence of free radical initiators conducted in an intermeshing co-rotating twin screw extruder, and to develop a mathematical model to predict the extent of the reaction under different processing and reaction conditions.; In order to be able to predict the extent of reaction, kinetic studies were conducted in both low molecular weight analogs of polyethylene and in the polyethylene melt. Reaction kinetic parameters including the rate constants, the order of reaction with respect to monomer and initiator, and activation energies were estimated for the overall (homopolymerization and grafting) and grafting reactions respectively. Lupersol 130{dollar}sp{lcub}circler{rcub}{dollar} melt decomposition rates were determined and shown to be very close to the published data in low molecular weight solvents.; Processing and reaction parameters in a twin screw extruder were initially screened and subsequently optimized using an experimental design strategy. The objectives were to maximize grafted GMA (degree of grafting) onto the polymer backbone, suppress GMA homopolymerization (grafting efficiency) and improve reaction yield while minimizing polymer cross-linking. Increasing GMA concentration marginally unproved degree of grafting but lowered grafting efficiency significantly. Increasing initiator concentration had the effect of increasing both degree of grafting and grafting efficiency substantially, but at the same time the higher level of initiator caused considerable cross-linking in the base resin. Of the two peroxide initiators used, Lupersol 231 was less efficient but caused less cross-linking of the polyethylene. It was concluded that pre-mixing all the ingredients in the first feed zone of the extruder resulted in a slightly higher degree of grafting but injecting a mixture of monomer and initiator along the extruder barrel is more suitable in industrial practice. Other factors such as barrel temperature in the reaction zone and screw speed appeared to be significant under certain conditions.; A reactor model was developed for the twin screw extruder by breaking it into reactor cells consisting of plug flow and axial dispersion. Flow residence time and pressure profiles were computed using the "Bivis" simulation program, developed by the National Research Council of Canada. Actual melt temperatures measured in three locations down the length of the extruder were used as the reaction temperatures. The transport parameters: Peclet number and mean residence time for the axial dispersion model reactors were measured on the extruder under each set of reaction conditions. Deconvolution theory was applied to the RTD to determine the local transport parameters in a few runs.; Predictions of concentration profile using the reactor model indicated that most of the conversion occurs in regions with long residence time and high melt temperatures. This could be either the first or second filled region in the reactor zone. There was also a strong correlation between the local initiator concentration and the reaction conversion. The deviations between the predicted values and experimental data were varied. However, in general, runs which had small deviations were those closer in experimental conditions to those under which the kinetic data were extracted.