New assessment of cyclohexane–ethanol vapor–liquid equilibrium using Peng–Robinson and UNIFAC models

Abstract Vapor–liquid equilibria (VLE) are fundamental to distillation, one of the most widely employed separation techniques in chemical and pharmaceutical industries. The presence of asymmetrical molecules in shape and size, combined with polar functional groups, results in strong intermolecular interactions, leading to significant deviations from ideal mixture behavior. Consequently, accurately modeling such non-ideal systems, particularly those forming azeotropes, remains a significant challenge requiring the development of robust thermodynamic models and precise parameter estimation. The binary cyclohexane–ethanol is an example of such system, with high importance for the ethanol chemical industry. In this study, group contribution parameters for the UNIFAC model were optimized using the Nelder-Mead algorithm (downhill simplex method), implemented in MATLAB, based on the use of literature isothermal VLE data ( $$\gamma $$ – $$\varphi $$ formalism) in temperature range between 278 and 338 K, with Peng–Robinson equation for describing the vapor phase and UNIFAC/UNIFAC-Do model for the liquid phase. It is worth mentioning that the present modeling strategy has not been applied to the cyclohexane–ethanol system yet. The isobaric VLE predictions (40–150 kPa) obtained with the estimated UNIFAC model were compared with those obtained using the UNIFAC-Do (Dortmund) model, with parameters extracted from the DDB database. Under isothermal conditions, the UNIFAC model performed better, but under isobaric conditions, UNIFAC-Do showed lower absolute deviations for both bubble temperature and vapor phase compositions. Regarding the azeotropic point prediction, both parameter sets enabled achievement of a similar proximity to experimental data at 101.3 kPa. Finally, in order to check the thermodynamic reliability of both the data-sets considered, chemical activities for the present binary and VLE data for two different binaries (cyclohexane-2-butanol and cyclopentane-ethanol), were accessed. The chemical activities calculated (278–338 K) were compared with COSMO-SAC data and results suggest that both data-sets have similar performances regarding the chemical activity of cyclohexane and a superior performance for the present six parameter data-set regarding the chemical activity of ethanol. On the other hand, the VLE predictions for the other binaries suggest an overall better performance with the UNIFAC-Do data-set, which can be understood, as for this data-set the energy parameters depend on temperature and were optimized considering different substances containing the same type of molecular groups, and not specifically for the cyclohexane–ethanol binary, as performed in present manuscript.