Comparison of classical population balance models of emulsion polymerization with experimental results and a stochastic extension

Population balance models are frequently used to predict the distribution of the particle sizes in emulsion polymerization. In the numerical simulation of these models care has to be taken to choose a reliable numerical simulation method otherwise the results are influenced strongly by numerical diffusion. As we show in this paper, the standard population balance model of emulsion polymerization does not match experimental data for styrene polymerization well, if numerical diffusion is avoided by use of a suitable discretization scheme. The observed particle size distributions are much broader than predicted by the model. It is shown that this is a structural inadequacy of the model and cannot be overcome by re-tuning of the parameters. It is therefore proposed to augment the model by a stochastic term in the growth kernel that represents the inhomogeneity of the evolution of the particles. With this term, the predictions of the model match experimental data considerably better.

► We examined the classical PBE models of emulsion polymerization.
► Significant model/plant mismatch in transient PSD predictions was observed.
► Source of the error was found to be the structural inadequacy of the PBE models.
► Addition of a stochastic term to the growth kernel improved the model prediction significantly.