(180e) Mathematical Model for the Devolatilization Stage in the Continuous High-Impact Polystyrene Process: Prediction of Swelling and Melt Flow Indexes

In the continuous high-impact polystyrene (HIPS) process, a devolatilization stage is carried out in order to remove the unreacted monomer and other volatile species (solvent, oligomers). High temperature reactions (chain scission, de-propagation, crosslinking) take place in the devolatilizer, affecting the molecular structure and quality variables of the final product. In this work a model for the HIPS devolatilization process is developed. The model consist of two modules: the first kinetic module considers a two-phase tubular reactor and simulates the devolatilizer from the mass balances for each species, based on the complex high temperature reactions kinetics and the interfacial mass transfer phenomena. Inputs for this module are HIPS composition, molecular weight distributions of the different species (free polystyrene, residual polybutadiene and graft copolymer), crosslinking degree, and devolatilizer operating conditions and geometry. Outputs to the module are the monomer and oligomer content, solvent and mineral oil in the melt, molecular weight distributions and rubber molecular weight between crosslinking points. The second module allows calculation of the Swelling Index (SI) and the Melt Flow Index (MFI) of the devolatilized product. SI is theoretically calculated by a modified Flory-Rehner equation (Karam and Tien, J. Appl. Pol. Sci., 30, 1969, 1985), considering: thermodynamic equilibrium between the swollen gel and the solvent, equivalent spherical gel, simplified strain energy function and negligible contributions from interfacial tension forces to the overall free energy. Osmotic pressure is considered in order to include the effect of polystyrene occlusions in the rubber particles. The MFI calculation is based on the model by Seavey et al. (Ind. Eng. Chem. Res., 42, 5354, 2003) combined with and independent estimation of the apparent viscosity (Luciani et al., Ind. Eng. Chem. Res., 44 (22), 8354, 2005). The model was adjusted and validated using experimental data, a good agreement between measurements and estimations is observed.