(761a) Design of a Simulated Moving Bed Plant for the Purification of Flavors From Complex Mixtures

As a highly selective separation technology chromatography is almost an inevitable step in downstream processing schemes of mixtures containing components with similar physical properties. For a separation of a mixture in two fractions the continuous counter-current chromatography, i.e. Simulated Moving Bed (SMB), outperforms the conventional batch chromatography in terms of yield, productivity and eluent consumption. Currently, the SMB production scale applications involve mainly binary mixtures separations i.e. separations of enantiomers.

The objective of this work is the development of an approach for designing SMB purifications of multi-component mixtures. This is a particularly difficult task when dealing with feed stocks of plants, since the feed often contains hundreds of components, its composition is not fully characterized, and the physical and chemical properties of a high number of components present in the mixture are unknown. One of the particularities of these separation tasks is that high purities of a single feed component is often not of ultimate importance, more critical is ensuring the absence of certain critical components in the final product.

The lecture informs about the model-based approach supported by a restricted number of systematically selected experiments, which are combined with simulations of the SMB unit operation. The first step in the design of the SMB purification process was a simplification of the separation task, done by the assumption that the starting mixture can be presented as a model mixture containing the target component and the most critical impurities.

Furthermore, the chromatographic column hydrodynamics parameters, adsorption equilibrium and mass transfer parameters of the model mixture were determined experimentally. These parameters were used for the first selection of the operating parameters of a laboratory scale SMB unit for the separation of the model mixture. After an experimental validation of the simulated SMB unit performances, the SMB purifications of the complex mixture were performed. The results of these separation experiments were used to extend the SMB model and tune its parameters. The applicability of the extended model was confirmed experimentally.