Correlation between the Onset Temperature and Molecular Descriptors of Organic Peroxides

• Published in: Russian Journal of Physical Chemistry A 2023-11, Vol.97 (11), p.2550-2558
• Main Authors: Yuting, Liao, Fangrui, Jia, Huoyu, Rao, Zhanggao, Le, Zhenzhen, Xu
• Format: Article
• Language: English
• Subjects: Chemistry; Chemistry and Materials Science; Chemoinformatics and Computer Modeling; Correlation coefficients; Dependent variables; Independent variables; Molecular structure; Organic peroxides; Peroxides; Physical Chemistry; Test sets; Thermal decomposition; Training
• ISSN: 0036-0244; 1531-863X
• DOI: 10.1134/S0036024423110195

QSPR modeling has been performed on 38 organic peroxides against onset temperature. The molecular structures were optimized and frequencies were computed with DFT/b3lyp and 6-311g( d,p ) basis set. Avogadro 1.2 was used to Convert gaussian output files to MDL SDFile format. 5666 molecular descriptors were calculated by Alvadesc software based on MDL SDfiles. The data set is randomly divided into training and test sets, and the number of samples contained in the training and test sets are 30 and 8, respectively. A seven-parameter relationship was established by the stepwise multiple linear regression method. The correlation coefficient reaches 0.9843, and the RMSE for the training and test sets are 5.47 and 5.57, respectively. Student t -value, VIF value and Durbin–Watson value indicate that all the independent variables have significant impact on dependent variable, and there is no multicollinearity between independent variables and no correlation between the residuals, respectively. LOOCV, LMOCV, external validation, Application domain analysis show that the resulting QSPR model is robust and reliable and can be used to predict the thermal decomposition temperature of organic peroxides.