Isothermal phase equilibria in the binary C₃H₈ + 1-phenylethanol system at high temperatures and pressures: Experimental and modelling

Backgrounds Accurate vapor–liquid equilibrium (VLE) data are indispensable for the design and optimization of processes involving supercritical-fluid extraction. Such data prevent excessive equipment sizing, enhance energy efficiency, and support reliable modeling of separation systems—particularly in oil sludge treatment, where light alkanes like propane serve as effective supercritical solvents. Propane is especially favorable due to its moderate critical parameters ( T _C = 369.89 K, P _C = 4.25 MPa), which allow operation under milder conditions than CO₂ while retaining strong solvating power for heavy hydrocarbons. Methods Isothermal VLE measurements ( Pxy ) for the binary mixture of supercritical propane (C₃H₈) and 1-phenylethanol were carried out at (403.15, 423.15, and 443.15) K, with pressures ranging from (1.03 to 8.76) MPa, using a high-pressure optical cell equipped with gravimetric phase sampling. Significant Findings Critical point data and the critical locus for the C₃H₈ + 1-phenylethanol system were determined based on the measured PTxy data using the analytical extrapolating technique. Critical point data and the critical locus for the C₃H₈ + 1-phenylethanol system were extracted from the experimental VLE results. The Krichevskii parameter, a thermodynamically rigorous quantity with microscopic relevance, was evaluated from the initial slope of the critical curve. The predictive and correlative capabilities of three equations of state—Peng–Robinson (PR), PC-SAFT, and CP-PC-SAFT—were assessed against both pure-component properties and mixture VLE behavior. While PR showed an inferior performance in predicting the 1-phenylethanol’s pure-compound properties and the investigated VLE in C₃H₈ + 1-phenylethanol system without applying binary adjustable paramteres, it provided the most accurate prediction of the reported critical points. Besides that, PR with optimized binary parameters was the most accurate model in correlating VLE. CP-PC-SAFT was found to be superior in predicting the VLE without the binary parameters, while their optimization resulted in comparable performance of both SAFT models. It was also demonstrated that PC-SAFT’s overestimation of the pure-component critical temperatures and pressures leads to improvement of its accuracy in estimating critical points in the investigated temperature range.