There`s been widespread research of possible application for many Ionic Liquids (ILs) in recent years: as reactional media, catalyst enhancers, as entrainers for azeotrope separation, as solvent phase for mineral and organic... [ view full abstract ]
There`s been widespread research of possible application for many Ionic Liquids (ILs) in recent years: as reactional media, catalyst enhancers, as entrainers for azeotrope separation, as solvent phase for mineral and organic compounds extraction. In the wake of new potential applications, the shift from laboratory to industrial processes urges for knowledge of many thermodynamic properties, ranging from pure substance properties to in-equilibria mixture phase behaviour. Though experimental data for ILs is in continuous growth, those will never suffice for process project, operation and optimization. Since a grand number of mixtures containing ILs can be moulded, methods with good predictive capability are crucial for industrial applications. Models like UNIFAC(Do) and COSMO-RS, e.g., have this predictive power, yet with some well-known disadvantages. A more recent model alternative is the F-SAC model. Developed by our group, this model seeks to blend the molecule surface contact theory contained in the COSMO-RS model, with the correlation power from the empirically adjusted parameters from UNIFAC(Do). Studies into the use of the F-SAC model for Ionic Liquids have been promising. At the present moment, the F-SAC model is able to work with 1-alkyl-1-alkylimidazolium as a cation-base with eight possible different anions (tetrafluoroborate, bis(trifluoromethylsulfonyl)imide, tetracyanoborate, methanesulfonate, trifluoromethanesulfonate, trifluoroacetate, thiocyanate, chloride). Data regressions with infinite dilution activity coefficient (IDAC) data have been capable of finding correlation coefficients above 0.9 for mixtures of ILs with hydrocarbons (alkanes, alkenes, cycloalkanes, cycloalkanes and aromatics), with over 1500 different data points. Three anions have UNIFAC(Do) parameters available in the literature, and it was possible the compare the UNIFAC(Do) results with those obtained with F-SAC. For tetrafluoroborate, trifluoromethanesulfonate and bis(trifluoromethylsulfonyl)imide the F-SAC achieved a correlation coefficient (R2) of 0.9489, 0.9646 and 0.8171 and an Absolute Average Deviation(AAD) of 0.29454, 0.19552 and 0.22477 logarithm units. The UNIFAC(Do) parameters currently available produce higher deviations, with AAD of 1.53388, 0.32470 and of 0.40598 logarithm units for 404, 201 and 383 data points, respectively. Further, a lower parameter cost is perceived for the F-SAC model, where 20 parameters were needed for these Ionic Liquids, plus 33 parameters for hydrocarbons. UNIFAC(Do) required 44 group parameters and 96 binary parameters. The parameter demand for UNIFAC(Do) grows geometrically, thus the linear growth of parameters in the F-SAC model has been proven as an advantage for Ionic Liquid studies. In addition, with the parameter estimation based on IDAC values it is possible to calculate thermodynamic functions as the partial molar excess Gibbs energy, enthalpy and entropy at any composition. Such information is extremely valuable, since it provides an insight into the behaviour of liquid mixtures, allowing the prediction of azeotrope existence, to calculate the Henry constant, partition coefficients and estimate solubilities. It is also possible to screen solvents for extraction and extractive distillation. At the moment, as an improvement of the F-SAC application to Ionic Liquids, new anions and cations are being studied for F-SAC parameter determination.
