Screening of cosolvents for a supercritical fluid: A fully predictive approach

Abstract

AbstractSolubility of solids was modeled in modified supercritical carbon dioxide as a predictive method. The Peng‐Robinson equation of state was used to characterize the ternary systems using the one‐fluid van der Waals rules. Ternary mixture parameters k12, k23, and k13 were obtained from the binary mixtures using Wong‐Sandler mixing rules or the linear combination of Vidal‐Michelsen mixing rules together with UNIFAC as required as a fully predictive tool for the liquid activity coefficients. When this method was applied to predict the solubility of β‐naphthol in CO2 modified with 6% and 10% toluene, values for the predicted solubility were very similar to the measured ones, with deviations of 9.6 and 26.3%, respectively. However, when the method was applied to the data of Simon et al. who examined six common cosolvents, the results were not as good in absolute terms, but the solvents were sorted in about the same order as found experimentally. The method provides quite an approximate guide to judge the most effective cosolvent from several candidates using a fully predictive methodology.