(169c) Estimation of Organic Solid Solute Solubility in the Supercritical Carbon Dioxide from Predictive Equations of State

The information of solid solute solubility is important for the design and optimization of process using supercritical carbon dioxide, such as rapid expansion of supercritical solution. Experimental measurement is the most reliable and straightforward way to generate such data, but may be costly and time-consuming. Thus, having a predictive thermodynamic model to provide such information with the input of molecular structure and a few thermophysical properties of the solid solute would be useful for the industry.

The PR+COSMOSAC equation of state (EOS) utilizes quantum mechanical calculation results of a molecule to estimate the energy and volume parameters of the Peng-Robinson EOS. Thus, it can predict fluid phase equilibrium without input of critical properties and acentric factor of pure substances. In this study, we investigated the accuracy of the PR+COSMOSAC EOS in predicting solid solute solubility in supercritical carbon dioxide for 94 organic solid solutes. The accuracy of the PSRK (Predictive Soave-Redich-Kwong) EOS, a well-known group contribution method, was investigated and considered as a reference. The required critical properties and acentric factor of organic solid solutes for the PSRK EOS were also determined by two selected group contribution methods, Marrero-Gani (MG) and Nannoolal-Rarey-Ramjugernath (NRR) models. However, because of the lack of functional group definition or missing group interaction parameters, the PSKR EOS can predict solubility for only about half of solid solutes (55 out of 94 solutes) considered in this study.

The overall average logarithmic errors (ALD-x) in solubility for 55 solid solutes from PSRK-NNR, PSRK-MG, and PR+COSMOSAC are 1.51, 1.28, and 1.03, respectively. PR+COSMOSAC EOS provides slightly better accuracy than group contribution methods. The accuracy of PSRK is affected by the method for estimating critical properties. Since the PR+COSMOSAC EOS does not have the missing parameter issue, it can be applied to predict all solid solutes considered in this study and given that ALD-x = 0.94. This study shows that the PR+COSMOSAC EOS could provide useful information to industry process design when no experimental data are available.