| Resumo: | The general objective of this work is the study of the behaviour of the Simulated Moving Bed (SMB) processes, with applications to chiral separations. Briefly, SMB chromatography allows the continuous injection and separation of binary mixtures. The simulated countercurrent contact between the solid and liquid phases maximizes the mass transfer driving force, leading lo a significant reduction in mobile and stationary phases consumption when cornpared with elution chromatography. The problem of modeling a SMB separation process is analyzed by two different strategies: one, by simulating the system directly, taking into account its intermittent behaviour (the SMB model); other by representing its operation in terms of a true countercurrent system (TMB model). From the comparative study, we conclude that the prediction of the performance of a SMB operation, and so its flow-rate optimization, can be safely done by using the TMB approach. The effect of the model parameters and operating variables on the SMB performance is analyzed in terms of four parameters: purity, recovery, solvent consumption, and adsorbent productivity. This work shows the influence of the switch time interval, recycling flow-rate, and inlet and outlet flow-rates. Also studied is the influence of axial dispersion and mass transfer resistance on the 8MB performance. We conclude that the mass transfer resistance can reduce significantly the possible set of operating conditions that lead to the desired separation. Moreover, mass transfer resistance can affect the SMB operating conditions, being this influence emphasized when a higher purity requirement is desired. The experimental operation of a Simulated Moving Bed unit was implemented in a Licosep 12-26 SMB pilot unit (Novasep, France), available at the LSRE. Two chiral systems were considered: the bi-naphlhol and the chiral epoxide enantiomers. For the bi-naphthol system purities and recoveries higher than 95% were obtained for both extract and raffinate with an adsorbent productivity of 68 grams of racemic mixture processed per day and per liter of bed. The solvent consumption was 1.2 liter per gram ofracemie mixture processed. For the chiral epoxide system, a first set of experimental runs Ied to purities and recoveries higher than 90%. A productivity of 52 grams of racemic mixture per day and per liter of bed was achieved, with a solvent consumption of 0.4 liter of mobile phase per gram of racemic mixture processed. In a second seI of experimental runs, we obtained 98% pure extract and raffinate, with a productivity of 34 grams per day and per liter of bed, and a solvent consumption of 1.3 liter per gram. The experimental results obtained in the Licosep 12-2 6 SMB pilot unil are reported in terms of process performance, steady-state internal concentration profiles, and transient evolution of the concentration of both species in the extract and raffinate streams. The simulation packages developed in this work are used lo predict the SMB behaviour in reasonable agreement with the experimental results.
|
|---|