Novel descriptors from main and side chains of high-molecular-weight polymers applied to prediction of glass transition temperatures

New descriptors of main and side chains for polymers with high molecular weight are presented in order to predict the glass-transition temperature (Tg) by means of Tg/M ratio. They were obtained by molecular modeling for the middle unit in a series of three repeating units (trimer). Taken together with other classic descriptors calculated for the entire trimeric structure, the ones that correlated better with the property were selected by using a variable selection method. Only three descriptors were chosen: main chain surface area (SAMC), side chain mass (MSC) and number of rotatable bonds (RBN), where the first two descriptors belong to the set of the new ones proposed. By means of a multi-layer perceptron (MLP) neural network a good prediction model (R2 = 0.953 and RMS = 0.25 K mol/g) was achieved and internally (R2 = 0.964 and RMS = 0.41 K mol/g) and externally (R2 = 0.933 and RMS =0.47 K mol/g) validated. The dataset included 88 polymers. The selected descriptors and the quality of the obtained model demonstrate the advantages of capturing through computational molecular modeling the structural characteristics of the polymers’ main and side chains in the prediction of Tg/M.

Figure optionsDownload high-quality image (124 K)Download as PowerPoint slideHighlights
► A novel set of descriptors to predict glass transition temperatures for polymers was proposed.
► They were obtained by molecular modeling for the middle unit in a trimeric structure.
► A neural network prediction model with only 3 descriptors was developed.
► The good quality and robustness of the model for predicting Tg were shown.
► A structural explanation of the model descriptors and its relation to Tg was presented.