Prediction of Boiling Point of Imidazolium-Based Ionic Liquid + Solvent Mixtures

Abstract

A wide application of ionic liquids in different separation processes, such as multiple reaction, extraction cycles, and azeotropic or close-boiling-mixture distillation processes, has been reported during the past decade. One of the most fundamental parameters commonly required in these processes is normal boiling point. Regarding this requirement and the complexities of ionic-liquid binary mixtures, a cascade artificial neural network was used to correlate the binary boiling point of ionic-liquid mixtures. The molecular weight and melting points of both components and the mole fraction of nonionic-liquid components for 425 collected experimental data points are used to define systems and discriminate between the different components. The obtained results demonstrate a good capability of the used cascade-artificial-neural-network model to correlate the binary normal boiling points of the mixtures with a total average absolute relative-deviation percent of 0.38%.