Phase behavior of ternary mixtures {aliphatic hydrocarbon + aromatic hydrocarbon + ionic liquid}: Experimental LLE data and their modeling by COSMO-RS

• Several ionic liquids were studied as solvent to separate aromatic and aliphatic hydrocarbons.
• (Liquid + liquid) equilibrium data were experimental determined at T = 298.15 K and p = 1 atm.
• Selectivity and solute distribution ratio were calculated.
• Experimental data were correlated using NRTL and UNIQUAC thermodynamic models.
• COSMO-RS model was applied to predict the (liquid + liquid) equilibrium.

Separation of aromatic and aliphatic hydrocarbons is a complex process in the petrochemical industry due to overlapping boiling points and azeotrope formation. In this paper, liquid extraction of aromatic compounds (toluene and ethylbenzene) from aliphatic compounds (hexane and cyclohexene) using ionic liquids (1-butyl-3-methylimidazolium methylsulfate, BMimMSO4, 1-propyl-3-methylimidazolium bis{trifluoromethylsulfonyl}imide, PMimNTf2, and 1-butyl-3-methylimidazolium bis{trifluoromethylsulfonyl}imide, BMimNTf2) as solvent was studied. (Liquid + liquid) equilibrium (ELL) data for the ternary systems {hexane (1) + ethylbenzene (2) + BMimMSO4, or BMimNTf2, or PMimNTf2 (3)}, {hexane (1) + toluene (2) + BMimMSO4 (3)} and {cyclohexene (1) + ethylbenzene (2) + BMimMSO4 (3)} were experimentally determined at T = 298.15 K and atmospheric pressure. Moreover, an analysis of the influence of the structure of each compound on the phase behavior was also carried out. The ability of the studied ILs to separate aromatic from aliphatic compounds was evaluated in terms of the solute distribution ratio, β, and the selectivity, S. The Non Random Two-Liquid (NRTL) and UNIversal QUAsiChemical (UNIQUAC) thermodynamic models were used to correlate the experimental LLE data. Furthermore, the COnductor-like Screening MOdel for Real Solvents (COSMO-RS) was applied to predict the (liquid + liquid) equilibrium. The suitability of this model to describe the phase behavior of the studied mixtures was evaluated comparing the experimental and calculated data.