Assessing suitability of glycerol-derived green solvent for the separation of n-hexane + ethanol azeotropic mixture, accompanied by VLE studies using machine learning

The objective of this work is to assess the effectiveness of incorporating most commonly used glycerol based green solvent for the separation of an azeotropic mixture of interest in bio-diesel and chemical industry. To evaluate the viability of this application, the phase equilibrium for the ternary system (n-hexane + ethanol + deep eutectic solvent (DES)) was experimentally determined with vapor composition, activity coefficient and pressure using head space gas chromatography method. Within a wide range of DES concentrations, experimental vapor–liquid equilibrium data for binary system: n-hexane and ethanol with DES consisting of choline chloride and glycerol are recorded are reported at 328.15 K. The Non-Random Two-Liquid (NRTL) activity coefficient liquid model is applied to correlate its phase behavior. The machine learning algorithms for support vector machines and Gaussian process regression are used for correlation and is found to be satisfactory in binary system. Using machine learning surrogates, the coefficients required in NRTL can be avoided which requires large computational efforts, and thus can be employed for wide range of mixtures.