[an error occurred while processing this directive]

高等学校化学研究 ›› 2023, Vol. 39 ›› Issue (2): 296-304.doi: 10.1007/s40242-022-2234-3

• Article • 上一篇    下一篇

Estimation of Enthalpy of Formation Using Benson's Group Addition and Functional Group Correction

LI Xianlan1,2, LUAN Yue1,2, LU Yanhua2, LI Wei2, MA Lihong3, ZHANG Qingyou1, PANG Aimin2   

  1. 1. Henan Engineering Research Center of Industrial Circulating Water Treatment, Henan Joint International Research Laboratory of Environmental Pollution Control Materials, Henan University, Kaifeng 475004, P. R. China;
    2. Science and Technology on Aerospace Chemical Power Laboratory, Hubei Institute of Aerospace Chemotechnology, Xiangyang 441003, P. R. China;
    3. Laboratory&Equipment Managements Service, Dalian University of Technology, Dalian 116024, P. R. China
  • 收稿日期:2022-07-13 出版日期:2023-04-01 发布日期:2023-03-16
  • 通讯作者: ZHANG Qingyou, PANG Aimin E-mail:qingyou@vip.henu.edu.cn;ppam@tom.com
  • 基金资助:
    This work was supported by the Open Research Fund Program of Science and Technology on Aerospace Chemical Power Laboratory, China(No.120201B01) and the National Natural Science Foundation of China(Nos.21875061, 21975066).

Estimation of Enthalpy of Formation Using Benson's Group Addition and Functional Group Correction

LI Xianlan1,2, LUAN Yue1,2, LU Yanhua2, LI Wei2, MA Lihong3, ZHANG Qingyou1, PANG Aimin2   

  1. 1. Henan Engineering Research Center of Industrial Circulating Water Treatment, Henan Joint International Research Laboratory of Environmental Pollution Control Materials, Henan University, Kaifeng 475004, P. R. China;
    2. Science and Technology on Aerospace Chemical Power Laboratory, Hubei Institute of Aerospace Chemotechnology, Xiangyang 441003, P. R. China;
    3. Laboratory&Equipment Managements Service, Dalian University of Technology, Dalian 116024, P. R. China
  • Received:2022-07-13 Online:2023-04-01 Published:2023-03-16
  • Contact: ZHANG Qingyou, PANG Aimin E-mail:qingyou@vip.henu.edu.cn;ppam@tom.com
  • Supported by:
    This work was supported by the Open Research Fund Program of Science and Technology on Aerospace Chemical Power Laboratory, China(No.120201B01) and the National Natural Science Foundation of China(Nos.21875061, 21975066).

摘要: The enthalpies of formation of solid organic compounds containing carbon, nitrogen, oxygen, and hydrogen were estimated using two suggested descriptor sets, separately, by machine learning methods. The two descriptor sets are both composed of descriptors of Benson’s groups and corrected groups. The main differences between them are that one is generated based on atoms and the other is based on bonds. An in-house program was specially written in Java to extract all the descriptors with a function to ensure that each atom(or bond) of a molecule is represented by Benson’s groups once for an atom-based(or bond-based) descriptor set. Multiple linear regression and partial least squares were used, separately, to build models to predict the enthalpy of formation for two descriptor sets. The combination of the models constructed by two descriptor sets based on the atoms and the bonds achieved the best-predicted results in this paper, and the corresponding results of the test set are better than that in the literature, from which the original data were retrieved. Further, a small data set of fluorinated molecules was collected, and satisfactory results were also obtained for these molecules containing fluorine with the assistance of the former data set.

关键词: Enthalpy of formation, Group contribution method, Benson's group, Multiple linear regression, Partial least square

Abstract: The enthalpies of formation of solid organic compounds containing carbon, nitrogen, oxygen, and hydrogen were estimated using two suggested descriptor sets, separately, by machine learning methods. The two descriptor sets are both composed of descriptors of Benson’s groups and corrected groups. The main differences between them are that one is generated based on atoms and the other is based on bonds. An in-house program was specially written in Java to extract all the descriptors with a function to ensure that each atom(or bond) of a molecule is represented by Benson’s groups once for an atom-based(or bond-based) descriptor set. Multiple linear regression and partial least squares were used, separately, to build models to predict the enthalpy of formation for two descriptor sets. The combination of the models constructed by two descriptor sets based on the atoms and the bonds achieved the best-predicted results in this paper, and the corresponding results of the test set are better than that in the literature, from which the original data were retrieved. Further, a small data set of fluorinated molecules was collected, and satisfactory results were also obtained for these molecules containing fluorine with the assistance of the former data set.

Key words: Enthalpy of formation, Group contribution method, Benson's group, Multiple linear regression, Partial least square