Predicting reaction rate constants of ozone with organic compounds from radical structures

Xinliang Yu,Bing Yi,Xueye Wang,Jianfang Chen

DOI: https://doi.org/10.1016/j.atmosenv.2012.01.037

IF: 5

2012-01-01

Atmospheric Environment

Abstract:The reaction rate constants of ozone with organic compounds in the atmosphere were predicted by a quantitative structure–activity relationship (QSAR) model. Density functional theory (DFT) calculations, for the first time, were carried out on the radicals from organic compounds, at the UB3LYP level of theory with 6-31G(d) basis set. A set of quantum chemical descriptors calculated from the radicals, the energy of the highest occupied molecular orbital of beta spin states (EβHOMO), the molecular average polarizability (α), and the total energy (ET), were used to build the general QSAR model for aliphatic compounds, applying the genetic algorithm (GA) technique and support vector machine (SVM) regression. The root mean square errors (RMSE) are 0.680 for the training set (68 compounds), 0.777 for the validation set (36 compounds) and 0.709 for the test set (35 compounds). Investigated results indicate that the SVM model given here has good predictivity for aliphatic compounds.