Solubility of anthracene and phenanthrene in binary mixtures of ethanol þ 2,2,4-trimethylpentane at (298.2, 308.2, and 318.2) K are reported. Five numerical methods were used to predict the solubility of anthracene and phenanthrene in the studied binary solvent mixtures, and the mean relative deviation (MRD) was used as an error criterion. The MRD values for studied solvents are in acceptable range.
' INTRODUCTIONPolycyclic aromatic hydrocarbons (PAHs) are among environmental pollutants with very low aqueous solubility values. Anthracene (with aqueous solubility of 0.0076 mg 3 L -1 ) and phenanthrene (with aqueous solubility of 1.20 mg 3 L -1 ) are among the most commonly found PAHs in the environment. 1 Anthracene and phenanthrene are two isomeric forms of the simplest tricyclic aromatic hydrocarbon ( Figure 1) and were used as model compounds in this work. Phenanthrene has lower molecular total energy, and it is more stable than anthracene according to ab initio studies using molecular orbital method in the gas phase. 2 Solubility alteration of chemicals is required in many industrial applications, and the solvent mixing or cosolvency is one of the most frequent and feasible methods used in the industry. Using different ratios of the solvents, a wide range of solubility for a given compound can be achieved. The next parameter which could be employed in the chemical industry is the changing temperature of the system which make a significant contribution in solubilization or crystallization of a compound. Finding an optimum solvent composition and the appropriate temperature in solubility investigations is usually obtained by trial and error which is both time-consuming and costly.
' SOLUBILITY MODELSNumerous predictive models have been introduced to replace the trial-and-error approach or at least reduce the number of required experimental data. 1 One of these models, is the Jouyban-Acree model which correlates the solubility of a solute in binary solvent mixtures at various temperatures. 3where X m,T Sat is the solute mole fraction solubility in the mixed solvents at temperature T; x 1 and x 2 are the mole fractions of solvents 1 and 2 in the absence of the solute; X 1,T Sat and X 2,T Sat denote the mole fraction solubility of the solute in monosolvents 1 and 2; and J i terms are the solvent-solvent and solute-solvent interaction coefficients. Our previous results showed that the model could be trained using the experimental data at 298.2 K and be used to predict the solubility at other temperatures of interest with acceptable error. 4 These coefficients could also be computed using a minimum number of experimental solubility data in binary solvents and then be employed to predict solubility at other solvent compositions. To provide trained versions of the model to predict the solubility in binary solvents without employing solubility data in mixed solvents, a number of attempts were made. The binary interaction terms (J terms) have been correlated with the Hildebrand solubility parameters of the sol...