Abstract The glass transition temperature (Tg) is a fundamental characteristic of an amorphous polymer. A quantitative structure-property relationship (QSPR) based on error back-propagation artificial neural network (ANN) was constructed to predict Tgs of 107 polystyrenes. Stepwise multiple linear regression (MLR) analysis was adopted to select an optimal subset of molecular descriptors. The chain segments (or motion units) of polymer backbones with 20 carbons in length (10 repeating units) were used to calculate these molecular descriptors reflecting polymer structures. The relative optimal conditions of ANN were obtained by adjusting various network paramters by trial-and-error. Compared to the model already published in the literature, the optimal ANN model with [4-7-1] network structure in this paper is accurate and acceptable, although our model has more samples in the test set. The results demonstrate the feasibility and powerful ability of the chain segment structures as representative of polymers for developing Tg models of polystyrenes.

In this work we have performed a QSAR study of atmospheric reactions with hydroxyl radical, at the B3LYP level of theory with 6-31G(d) basis set. Molecular descriptors selected by applying multiple linear stepwise regression (MLR) analysis were used to predict the reaction rate constants (-logkOH) of OH radicals with organic compounds in the atmosphere, including 98 alkenes and 80 aromatic compounds. For setting our reactions, we have calculated 98 alkenes and 80 aromatic compounds. A four-descriptor MLR model (rms = 0.102 and R²= 0.938) for 98 alkenes was developed based on the number of R=CHX functional group counts, E HOMO parameter, and Fukui indices of the double-bonded C atoms. We found vinyl chemicals with conjugated double bonds and electron-donor substituents are the most reactive systems; while alkenes with multiple halogen substitutions are the least reactive molecules. Additionally, a three-descriptor MLR model (rms = 0.282 and R²= 0.910) was built to predict OH radical rate constants for 80 aromatic compounds, which was dominated by the E HOMO parameter, a topological descriptor for steric hindrance, and the most positive net atomic charge on hydrogen atoms. Aromatics with electron-donor and electron-acceptor groups, respectively, possess high and low degradation rates. The halogen aromatics are less reactive, especially for aromatics with multiple halogen substitutions. In comparison to existing models, the two models obtained in this paper show better statistical quality.