Fundamental solubility study of polycyclic aromatic hydrocarbons in subcritical water and ethanol mixtures

Abstract

Water is increasingly being used as a processing medium as a consequence of the current emphasis for sustainable processes. At relatively high temperatures and pressures, water acts as a good solvent, resulting in a number of potential applications. A lack of fundamental data, however, has limited the development of subcritical water technologies. Hence, solubility studies are essential to quantify the solvating power of subcritical water and to determine the thermodynamic limit of a process.

A static analytical equilibrium method was used to measure the binary and ternary solubilities of anthracene and p-terphenyl in subcritical water and ethanol mixtures between 393 K and 473 K, and at 50 bar and 150 bar. Temperature was found to have the most significant effect on the solubility of polycyclic aromatic hydrocarbons (PAHs) in subcritical water. The effect of pressure, and the combined effect of temperature and pressure on solubility were found to be insignificant, particularly when the range of pressure considered is relatively small. It was also found in this work that the solubilities of PAHs are governed primarily by sublimation pressure, and only secondarily by the dielectric constant of water. The use of small amounts of ethanol in subcritical water systems was found to greatly enhance the solubility of hydrophobic solutes and thus, is able to expand the range of applications of subcritical water technologies, while enabling relatively mild operating conditions to be maintained.

The UNIQUAC, O-UNIFAC, and M-UNIFAC models were used to correlate the solubilities of PAHs in subcritical conditions. The UNIQUAC model best represents the solubilities of anthracene and p-terphenyl in binary and ternary systems while the O-UNIFAC and M-UNIFAC models perform poorly. The poor performance of the O-UNIFAC and the M-UNIFAC models were mainly due to the inadequacy of the residual component of the activity coefficient. All three models show increasing deviations from experimental data as ethanol concentration increases.