Prediction of liquid−liquid equilibria of (aromatic + aliphatic + ionic liquid) systems using the Cosmo-SAC model

Highlights: ► Aromatic/aliphatic hydrocarbons separation by ionic liquids. ► Cosmo-SAC thermodynamic model for prediction of LLE by quantum chemical calculations. ► Predictions successful for some ionic liquids. ► Correlation by single parameter for other ionic liquids. ► Preliminary screening of additional ionic liquids for aromatic/aliphatic separation. - Abstract: The extraction of aromatic hydrocarbons from aliphatic hydrocarbons is an important problem. This process can be used to enhance the capacity of ethylene crackers by prior removal of aromatics that cannot be cracked. Ionic liquids have been investigated extensively for liquid–liquid extraction of aromatics from aliphatics. The choice of a suitable ionic liquid may be made by measuring liquid−liquid equilibria. However, the large number of ionic liquids, formed by the various cation and anion combinations, makes the experimental measurements expensive and time consuming. Hence, a predictive thermodynamic model called Cosmo-SAC that uses quantum chemical calculations for calculating liquid−liquid equilibria has been evaluated. A priori predictions are accurate for some ionic liquids and inaccurate for some ionic liquids. However, it has been shown that even when a priori predictions are inaccurate, the data can be correlated using a single parameter that is characteristic of the ionic liquid and accurate predictions can be made for additional aromatic/aliphatic combinations for the same ionic liquid. A comparison with the NRTL and UNIQUAC models has also been carried out. In addition, a preliminary screening of ionic liquids for aromatic/aliphatic separations has been carried out using the Cosmo-SAC model. Finally, the "Cosmo" files for eight cations and sixteen anions corresponding to 128 potential ionic liquids have been provided for the use of the general scientific community to predict any thermodynamic equilibria involving ionic liquids without the use of any molecular modeling software.