高等学校化学学报 ›› 2012, Vol. 33 ›› Issue (02): 346.doi: 10.3969/j.issn.0251-0790.2012.02.024

• 物理化学 • 上一篇    下一篇

基于平均影响值的反向传播神经网络方法用于提高密度泛函理论计算Y—NO键均裂能精度

李鸿志1, 陶委2, 高婷1, 李辉1, 吕英华1,2, 苏忠民2   

  1. 1. 东北师范大学计算机科学与信息技术学院, 长春 130017;
    2. 东北师范大学化学学院, 功能材料化学研究所, 长春 130024
  • 收稿日期:2011-05-03 出版日期:2012-02-10 发布日期:2012-01-13
  • 通讯作者: 吕英华, 男, 博士, 教授, 博士生导师, 主要从事面向并行对象计算机系统的模拟和模式识别研究. 苏忠民, 男, 博士, 教授, 博士生导师, 主要从事量子化学和功能材料化学研究. E-mail:luyh@nenu.edu.cn;zmsu@nenu.edu.cn
  • 基金资助:

    国家"九七三"计划项目(批准号: 2009CB623605)、国家自然科学基金(批准号: 20703008, 20903020)和吉林省科技发展计划项目(批准号: 20100114)资助.

Improving the Accuracy of DFT Calculation for Homolysis Bond Dissociation Energies of Y—NO Bond via Back Propagation Neural Network Based on Mean Impact Value

LI Hong-Zhi1, TAO Wei2, GAO Ting1, LI Hui1, LV Ying-Hua1,2, SU Zhong-Min2   

  1. 1. College of Computer Science and Information Technology, Northeast Normal University, Changchun 130017, China;
    2. Institute of Functional Material Chemistry, College of Chemistry, Northeast Normal University, Changchun 130024, China
  • Received:2011-05-03 Online:2012-02-10 Published:2012-01-13

摘要: 将基于平均影响值(Mean impact value, MIV)的反向传播神经网络(Back propagation neural netowrk, BPNN)(MIV-BPNN)方法用于提高密度泛函理论(Density functional theory, DFT)计算Y—NO(Y=N, S, O及C)键均裂能的精度. 利用量子化学计算和MIV-BPNN联合方法计算92个含Y—NO键的有机分子体系的均裂能. 结果表明, 相对于单一的密度泛函理论B3LYP/6-31G(d)方法, 利用全参数下的BPNN方法计算92个有机分子均裂能的均方根误差从22.25 kJ/mol减少到1.84 kJ/mol, 而MIV-BPNN方法使均方根误差减少到1.36 kJ/mol, 可见B3LYP/6-31G(d)和MIV-BPNN联合方法可以提高均裂能的量子化学计算精度, 并可预测实验上无法获取的均裂能值.

关键词: Y—NO键, 均裂能, 密度泛函理论, 平均影响值, 反向传播神经网络

Abstract: The back propagation neural network(BPNN) approach based on mean impact value(MIV)(MIV-BPNN) was used to improve the accuracy of density functional theory(DFT) calculation for homolysis bond dissociation energies of Y—NO bond. Quantum chemistry calculations and MIV-BPNN were used jointly to calculate the homolysis bond dissociation energy(BDE) of 92 Y—NO organic molecular systems. The results show that compared to a single density functional theory B3LYP/6-31G(d) approach, full parameters BPNN approach reduces the root-mean-square(RMS) of the calculated homolysis BDE of 92 organic molecules from 22.25 kJ/mol to 1.84 kJ/mol and MIV-BPNN approach further reduces the RMS to 1.36 kJ/mol. It is clear that the combined B3LYP/6-31G(d) and MIV-BPNN approach can improve the accuracy of the homolysis BDE calculation in quantum chemistry and can predict homolysis BDE which can not be obtained experimentally.

Key words: Y-NO bond, Homolysis bond dissociation energy, Density functional theory, Mean impact value, Back propagation neural network

中图分类号: 

TrendMD: