The reliable predicting refractive index for diverse polymers only from structural moieties in repeating unit structures

Abstract

The refractive index (RI) of polymers is used to understand polymers' behavior in the optical wavelength range, which is essential for designing optical devices. In contrast to available Quantitative Structure-Property Relationship (QSPR) approaches which need complex descriptors, this work introduces a simple method for predicting logRIof polymers based on elemental composition in repeating unit structures as well as two correcting structural parameters. The largest available experimental data of 221 diverse organic polymers are used to derive and test the new model by considering different statistical parameters as well as internal and external validations. Reliably the novel method is compared with the calculated outputs of four QSPR models. The value of Root Mean Square Error (RMSE) of the new model for the training set containing 154 polymers is 0.00623, which is smaller than four QSPR models, i.e., 0.00694, 0.00680, 0.00685, and 0.00691. RMSE values of the test set including 67 polymers also have the same trend, i.e., 0.00621 versus 0.00695, 0.00654, 0.00644, and 0.00640.