Correlating and predicting the solubilities of structurally similar organic solid compounds in supercritical CO2 using the compressed gas model and the reference solubilites

• Reference solubility of the solute was introduced into the compressed gas model.
• By using reference solubility, the sublimation pressure of the solute is avoided.
• The solubilities of some structurally similar compounds in sc-CO2 were correlated.
• In solubility correlation, the binary parameters for these compounds are close.
• The solubilities of these compounds were predicted by a specific binary parameter.

In this paper, the solubilities of some structurally similar organic solid compounds namely 9,10-anthraquinone derivatives, thioxanthone derivatives, pyrazolones and steroids were correlated and predicted using the compressed gas model combined with the Carnahan–Starling–van der Waals equation of state (CS–VDW EoS) and the Peng–Robinson equation of state (PR EoS), respectively. To avoid using the sublimation pressure of the solute and reduce the applied pressure range of the corresponding equation of state, a solubility datum was used as the reference solubility in the compressed gas model. By introducing the reference solubility, the simple van der Waals one-parameter (VDW1) mixing rules can be used for the supercritical solutions and the calculated solubilities are not sensitive to the binary parameter in the compressed gas model. So in solubility prediction, the binary parameter in the compressed gas model can be set constant for a series structurally similar organic compound. The prediction results showed that the compressed gas model combined with the CS–VDW EoS–VDW1 provide better prediction (average absolute relative deviations, AARD = 18.41%) for the compounds with similar structure and without intermolecular hydrogen bond. For the compounds with larger structural differences, the compressed gas model combined with the PR EoS–VDW1 is recommended. However, for the compounds with intermolecular hydrogen bond, neither of the two EoS can provide satisfactory prediction results.