Chem. Pharm. Bull. 53(3) 290—295 (2005)

maceutical industry for production of fine drug particles with narrow size distribution, separation of active ingredients and preparation of microemulsions and sustained drug delivery systems. Supercritical carbon dioxide (SC-CO2) is a very popular fluid for pharmaceutical applications, as it is nontoxic, non-flammable and of low cost. Furthermore, because of its low critical temperature and pressure, SC-CO2 is ideal for thermal-liable compounds. In addition to the pharmaceutical applications, supercritical fluids are very important solvents in many other industrial processes, including decaffeination of coffee, extraction of hops, spices and seed oils, as well as decontamination of environmental wastes. However, the solubility of most pharmaceutical/chemical compounds is very low in SC-CO2 and an entrainer is usually added to enhance the solubility. Solubility data are required for designing SCF processes but experimental measurement of solubility is time-consuming and costly. Only a limited number of reports has been published on the solubility of pharmaceutical/chemical compounds in entrained supercritical fluids. Mathematical modelling of solubility data in SC-CO2 entrainer could provide useful information for pharmaceutical/chemical engineers to speed up the process. Thermodynamic models employing equations of state, lattice gas equations or expanded liquid models have been used for modelling such data. These methods require complicated computations and a number of physico-chemical properties which are often unavailable. The aim of this study was to present a simple and easy to use empirical model to calculate the solubility in SC-CO2 entrainer with respect to the concentration of the entrainer, pressure, temperature and density of pure SC-CO2. The accuracy of the proposed model is evaluated employing published experimental data sets and compared with that of a model from the literature. Reported experimental data sets containing solubility data of pharmaceutical/chemical compounds at different temperatures, pressures and entrainer concentrations were included in this study. Data sets containing the solubility at only one temperature, pressure or entrainer concentration, as well as data sets with data points numbering less than 30 (five points for each independent variable) were excluded from this study. Computational Methods An empirical equation has been reported to correlate the mole fraction solubility (y2) of a solute in pure SC-CO2 with respect to the density of pure SC-CO2 (r), pressure (P) and temperature (T):