Datensatz

Zusammenfassung

Hybrid processes, i.e. processes integrated on the flowsheet level, are a possible method of choice if a certain task is too difficult to be resolved economically using only a single unit operation. Resolving racemates represents a rather difficult separation problem. Although "chiral" chromatography often is the method of choice for enantioseparations, it remains and expensive process. Chiral stationary phases are usually expensive and their performance is limited by their capacity and the separability of the enantiomers. It was shown previously [1,2] that the performance of SMB chromatography improves if the purity requirements can be lowered. This fact can be utilised by coupling the SMB with a second unit operation. In the example studied here, the SMB process supplies a subsequent enantioselective crystallisation with partially resolved mixtures (Fig. 1). Besides the possibility of saving investment costs by using less or lower-priced chiral stationary phase the hybrid process might allow for both higher productivity and lower solvent consumption compared to a stand-alone separation by SMB. The potential arising from the hybrid process can be evaluated on the basis of a few steady state calculations using an SMB model and the mass balances of the hybrid process. This approach is based on the assumption that the SMB process represents the limiting step in terms of separation performance and costs. In comparison to the stand-alone separation by SMB, the integrated process is more complex due to the additional units in the system, their specific dynamic behaviour and the introduction of additional discrete events (e.g., nucleation or seeding). An important issue for a practical realisation of the process is therefore its stability with respect to changes in the operating parameters. To systematically study the system's dynamic behaviour and its stability, a full dynamic mathematical model was developed using the process modelling tool PROMOT [3]. The model equations were solved by the simulation environment DIVA [4]. Essential parts of the model are a continuous chromatography unit modeled as a true-moving bed (TMB) process and the crystallisers assumed to be MDMPR units (mixed suspension - mixed product removal). Population dynamics are modeled using a simple moment formulation. To identify the main characteristics and the parameters with the highest impact on performance and behaviour, steady state continuation calculations and a sensitivity analysis were performed. Parameters revealing a strong impact on both stability and performance are, for example, the design parameters (flow rates) in the continuous chromatography, the reflux ratios of the mother liquor, crystallisation temperatures and enrichment ratios in the solvent treatment. As an example in Fig. 2 is shown how a model process corresponds to a consecutive activation of the recycle streams. The production rates of the crystallisers differ drastically with changes in the recycles. The interplay of nonlinear waves in the chromatographic process results in a complex transient behaviour of all other variables in the system. From the results of the studies above, guidelines for the practical implementation of the flowsheet-integration of SMB and enantioselective crystallisation are derived. [1] H. Lorenz et al., Journal of Chromatography A, 908 (2001), 201-214 [2] M. Kaspereit et al., Journal of Chromatography A, 944 (2002), 249-262 [3] M. Mangold et al., Computers & Chemical Engineering, 28 (2004), 319-332 [4] M. Mangold et al., Chemical Engineering Science, 55 (2000), 441-454