Correlation and Estimation of Aqueous Solubilities of Polycyclic Aromatic Hydrocarbons

Quantitative structure‐activity relationships have been studied for the aqueous solubilities of 25 polycyclic hydrocarbons (PAH) at 25°C. Multiple regression correlations are reported between aqueous solubility (X) and molecular size descriptor (molecular surface area, molecular volume or molecular weight), melting point temperature (MP), and a newly‐defined molecular symmetry and structure descriptor (R · r d ). A generalized correlation takes the form of log X = a 0 + a 1 (Molecular size descriptor) + a 2 (MP‐25) + a 3 R · r d in which a i refers to the ith regression coefficient, R is a molecular redundancy index and r d is the summation of interatomic distances in a molecule. These correlations enable the aqueous solubility to be estimated from melting point temperatures and structural information. The 3‐parameter expressions best predict the experimental solubilities and are superior to the 2‐parameter expressions (a 3 = 0) suggested in the past. For the majority of compounds considered, agreement between observed and estimated solubility is quite good. The correlations may also be useful in screening experimental solubility data for which serious discrepancies exist in the literature.