Prediction of Dielectric Constants of Binary Solvents at Various Temperatures

Abstract

A combined form of the Jouyban-Acree model and the Abraham solvation parameters was used for predicting the dielectric constant of binary solvent mixtures at various temperatures. The proposed model was trained with the numbers of the experimental data including aqueous and nonaqueous binary miscible mixtures, and then this trained form of the model was used to predict the dielectric constant of the left-out data. The mean relative deviations (MRDs) between predicted and experimental dielectric constants are used as an accuracy criterion. The overall MRD for all aqueous and nonaqueous mixtures is 22 To, and the corresponding value for aqueous binary mixtures is 9 To. Including the experimental dielectric constants of monosolvents, the prediction capability of the proposed models is improved, and the overall MRDs for all solvent mixtures and aqueous mixtures are 7 % and 3 %, respectively. A reduced form of the proposed model could be used to compute the dielectric constants of monosolvents at various temperatures with the MRD of 26 %.