SCI和EI收录∣中国化工学会会刊

Chinese Journal of Chemical Engineering ›› 2013, Vol. 21 ›› Issue (4): 420-426.doi: 10.1016/S1004-9541(13)60483-8

• 化工热力学 • 上一篇    下一篇

A Quantitative Structure Property Relationship for Prediction of Flash Point of Alkanes Using Molecular Connectivity Indices

Morteza Atabati, Reza Emamalizadeh   

  1. School of Chemistry, Damghan University, Damghan 36716-41167, Iran
  • 收稿日期:2011-07-03 修回日期:2012-09-26 出版日期:2013-04-28 发布日期:2013-04-26
  • 通讯作者: Morteza Atabati E-mail:atabati@du.ac.ir

A Quantitative Structure Property Relationship for Prediction of Flash Point of Alkanes Using Molecular Connectivity Indices

Morteza Atabati, Reza Emamalizadeh   

  1. School of Chemistry, Damghan University, Damghan 36716-41167, Iran
  • Received:2011-07-03 Revised:2012-09-26 Online:2013-04-28 Published:2013-04-26

摘要: Many structure-property/activity studies use graph theoretical indices, which are based on the topological properties of a molecule viewed as a graph. Since topological indices can be derived directly from the molecular structure without any experimental effort, they provide a simple and straightforward method for property prediction. In this work the flash point of alkanes was modeled by a set of molecular connectivity indices (χ), modified molecular connectivity indices (mχth) and valance molecular connectivity indices (mχv), with?mχv?calculated using the hydrogen perturbation. A stepwise Multiple Linear Regression (MLR) method was used to select the best indices. The predicted flash points are in good agreement with the experimental data, with the average absolute deviation 4.3 K.

关键词: quantitative structure property relationship, flash point, molecular connectivity indices, hydrogen perturbation, alkane

Abstract: Many structure-property/activity studies use graph theoretical indices, which are based on the topological properties of a molecule viewed as a graph. Since topological indices can be derived directly from the molecular structure without any experimental effort, they provide a simple and straightforward method for property prediction. In this work the flash point of alkanes was modeled by a set of molecular connectivity indices (χ), modified molecular connectivity indices (mχth) and valance molecular connectivity indices (mχv), with?mχv?calculated using the hydrogen perturbation. A stepwise Multiple Linear Regression (MLR) method was used to select the best indices. The predicted flash points are in good agreement with the experimental data, with the average absolute deviation 4.3 K.

Key words: quantitative structure property relationship, flash point, molecular connectivity indices, hydrogen perturbation, alkane