高等学校化学学报 ›› 2001, Vol. 22 ›› Issue (9): 1466.

• 研究论文 • 上一篇    下一篇

镧系金属离子Ln3+定量构效关系的神经网络研究

杨兴华1,2, 印春生2, 潘忠孝2   

  1. 1. 湖南省怀化师范专科学校化学系, 怀化 418008;
    2. 中国科学技术大学应用化学系, 合肥 230026
  • 收稿日期:2000-08-28 出版日期:2001-09-24 发布日期:2001-09-24
  • 通讯作者: 潘忠孝(1945年出生),男,教授,博士生导师,主要从事计算分析化学研究.
  • 基金资助:

    国家自然科学基金(批准号:No.29775001)资助

Studies on Quantitative Structure-property Relationship of Lanthanide Metal Ions Using an Artificial Neural Network

YANG Xing-Hua1,2, YIN Chun-Sheng2, PAN Zhong-Xiao2   

  1. 1. Department of Huaihua Teachers' College, Huaihua 418008, China;
    2. Department of Applied Chemistry, University of Science and Technology of China, Hefei 230026, China
  • Received:2000-08-28 Online:2001-09-24 Published:2001-09-24

摘要: 以Ln3+离子的半径、电负性、4f电子排布周期因子q(本文自定义)以及基态L值等为参数,使用函数连接型神经网络(FLN)对Ln3+离子的EDTA等络合常数lgK、水化能ΔH、水解常数pK1及氢氧化物溶解度pKm分别进行非线性关联和预测,获得了非常好的结果.与对应的线性模型比较表明,神经网络关联效果是较理想的.

关键词: 镧系金属离子, 定量构效关系(QSPR), 函数连接型神经网络(FLN), 非线性关联

Abstract: Systematical studies were made on the influencing factors concerned to the complexation, hydration, etc of lanthanide metal ions. Aset of parameters such as ionic radii, electronegativity, base state Lvalues, and periodic factors, defined in this work, were used to nonlinearly correlate the stability constants lg Kof EDTA, NTA and HTTAcomplexes, hydration energy Δ H, hydrolysis constants p K1 and pKm of hydroxides of the lanthanide metal ions with the functional link net(FLN). The obtained results are satisfactory, and better than that obtained by using linearly statistic method. Some of the absent constants values were predicted by FLN.

Key words: Lanthanide metal ions, Quantitative structure property relationship, Functional link net, Nonlinear correlation

中图分类号: 

TrendMD: