(502b) Micellar Solutions in Catalysis and Separation: Prediction of Partition Coefficients

Micellar systems have been successfully applied as reaction media for hydrogenation reactions [1]. They offer the advantage to solubilize reactants and catalysts which do not dissolve well in the aqueous phase. Moreover, catalyst recovery as well as product isolation can be achieved simultaneously by using ultrafiltration. Since the partitioning of reactants, products and catalyst between the aqueous and the surfactant-rich phase has a substantial influence on the reaction kinetics as well as on the separation process, optimization of reaction conditions in micellar systems can be done by taking into account the partition coefficient. Micelle/water partition coefficients can be either measured [2] or adequately predicted using thermodynamic models such as the group-contribution method UNIFAC and the a priori Conductor like Screening Model for Real Solvents (COSMO-RS) based on quantum mechanics [3]. Both models have shown their applicability in the field of chemical engineering for more than 10 years [4]. The main goal of the present study is to show the applicability of these models to optimize the conditions of reactions taking place in micellar systems (tailor-made surfactants, favorable phase compositions, temperature etc.). The catalytic hydrogenation of itaconic acid and its derivatives, such as itaconic acid-dimethylester and ?diethylester, in micellar solutions and microemulsions is used as a model system. As a first step, partition coefficients of reactants, products and catalysts for different types of surfactants (nonionic, ionic) are predicted at infinite dilution. Further, the partitioning is also studied depending on the composition of the coexisting phases. Furthermore, the influence of different overall concentrations of solute and surfactant on the partition coefficient is studied via an iterative method based on liquid-liquid phase equilibrium conditions with the goal to find a relevant surfactant-to-solute ratio at which the partition coefficients are not limited by the surfactant amount. The predictions of both thermodynamic models are then compared regarding the optimization of reaction and separation conditions. Both the UNIFAC and COSMO-RS models have been shown to be able to predict partition coefficients depending on concentration, type of surfactant and composition of the phases. Based on this information, the reaction and separation conditions in micellar systems can be optimized. Those predictions minimize the experimental effort significantly. We thank DFG (grant AR 236/32-1 and SCHO 687/7-1) for the financial support.

[1] G. Oehme et al, Angew. Chem. Int. Ed. 44 (2005) 7174-7199 [2] M. Schwarze, R.Schomäcker, CIT 7 (2006) 931-936 [3] L. Mokrushina, M. Buggert, I. Smirnova, W. Arlt, R. Schomäcker, Ind. Eng. Chem. Res. 46 (2007), 6501-6509 [4] S. Maaßen, W. Arlt, A. Klamt; Chem.-Ing.-Tech. 67 (1995), 476; J. Gmehling. 20th ICCT, 3-8 August 2008