An expanding field of research involves a class of tunable solvents known as room temperature ionic liquids (ILs).  ILs are being investigated for a wide variety of reaction, separation and extraction processes involving liquid-liquid phase behavior.  For instance, certain ILs have been shown to selectively extract alcohols from fermentation broths and to recover amino acids from aqueous media [1,2].  Since the number of possible systems involving ILs is enormous, comprehensive coverage of ternary liquid-phase behavior via experimental observation is impossible.  Therefore, it is important to model the liquid-phase behavior of mixtures containing ILs.  

            The modeling of liquid-liquid equilibrium (LLE) is a problem that has been extensively studied; however, the macroscopic modeling of LLE in systems involving ionic liquids has just begun.  Experimental binary and ternary LLE data involving ILs can be correlated using conventional excess Gibbs energy models.  However, the predictive capability of these models in this context has not been widely studied.  The goal of this work is to develop an approach, based on excess Gibbs energy models, that can be used to predict ternary LLE from binary measurements and pure component data.  This is a stringent test of the suitability of various models for describing LLE in systems containing ILs.  

            Previously, we have tested [3] the suitability of various excess Gibbs energy models, including NRTL, UNIQUAC, and electrolyte-NRTL (eNRTL) [4], for describing LLE in ternary systems containing ILs.  In some cases, this provided a useful approach for predicting entire ternary diagrams using parameters determined [5] from binary measurements.  However, in other cases, these models proved inadequate.  In these previous studies, a symmetric model was used; that is, the same excess Gibbs energy models was used for all phases.  However, ILs may behave very differently in different types of phases, and this symmetric approach does not account for this.  For example, in an IL-rich phase, the IL may remain relatively associated, but in a water-rich phase the IL may become almost entirely dissociated.  Therefore, we present here an asymmetric modeling framework that uses different models to describe different types of phases.  This allows different phases to have different degrees of ionic dissociation.  As an example, we will use the NRTL model for organic- and IL-rich phases, thus treating the IL ions as associated in these phases, and the eNRTL model for aqueous phases, thus treating the IL as completely dissociated in these phases.  Results will be presented for several systems of interest, showing the extent to which this approach can be used to predict LLE of ternary systems using parameters determined from binary and pure component data.  Issues that arise in parameter estimation and computation of phase equilibrium when using the asymmetric framework will also be discussed.  

1.         Fadeev, A.G., M. M. Meagher, “Opportunities for ionic liquids in recovery of biofuels.” Chem. Commun., 2001, 3, 295-296.  

2.         Wang, J.J., Y.C. Pei, Y. Zhao, and Z.G. Hu, “Recovery of amino acids by imidazolium based ionic liquids from aqueous media.” Green Chemistry, 2005, 7, 196-202.  

3.         Simoni, L.D., Y. Lin, J.F. Brennecke, and M.A. Stadtherr, “Prediction of liquid-liquid equilibrium using excess Gibbs energy models for mixtures containing ionic liquids.  In preparation, 2007.  

4.         Chen, C.-C., H.I. Britt, J.F. Boston, and L.B. Evans, Local composition model for excess Gibbs energy of electrolyte solutions.  Part I:  Single solvent, single completely dissociated electrolyte systems.” AIChE J., 1982, 28, 588-596.  

5.         Simoni, L.D., Y. Lin, J.F. Brennecke, and M.A. Stadtherr, “Reliable computation of binary parameters in activity coefficient models for liquid-liquid equilibrium. Fluid Phase Equilibria, in press, 2007.