Cyclic carbonates as solvents in the dearomatization of refinery streams: Experimental liquid–liquid equilibria, modelling, and simulation

Abstract

Obtaining aromatic hydrocarbons from refinery streams is one of the most concerning issues that the petrochemical industry currently faces. Now, the presence of azeotropes between aromatic and non-aromatic compounds in gasoline fractions demands the employment of organic solvents such as sulfolane to obtain benzene, toluene, and xylenes by liquid–liquid extraction. Several alternatives have been evaluated during the last decades to overcome the drawbacks related to the use of these solvents, being their replacement with ionic liquids one of the most promising so far. Nonetheless, the ionic liquids’ complex synthesis routes and lack of availability at large production still limit their industrial implementation. Recently, cyclic carbonates have emerged as a double solution related to CO2 fixation and designer solvents availability, the latter showing interesting extractive and physical properties which could reduce the environmental impact associated with the use of conventional solvents and ionic liquids in aromatic extraction processes. However, the availability of experimental data is scarce and only predictive models are available. In this work, firstly, the extractive properties of five cyclic carbonates (propylene, vinylethylene, glycerol, ethylene, and vinylene carbonates) were experimentally determined for wide composition ranges of several ternary systems {aliphatic/methylcycloalkane + aromatic + cyclic carbonate}. Afterwards, the NRTL, UNIFAC, and COSMO-SAC models, were used to fit or predict the liquid–liquid equilibria of those ternary systems. Finally, the aromatic extraction process from reformer gasoline with propylene carbonate as solvent was simulated in Aspen Plus with the COSMO-SAC model, comparing its performance with that of sulfolane to check its suitability at process scale.