萘系衍生物核磁共振碳谱模拟研究

Spectroscopic Simulation of ^(13)C Nuclear Magnetic Resonance for Derivatives of Naphthalene

从分子二维拓扑结构出发,应用原子电性作用矢量(AEIV)和原子杂化状态指数(AHSI)对35种萘系衍生物共计375个等价碳原子进行结构表征,分别以多元线性回归和逐步回归方法建立13C核磁共振化学位移定量结构波谱关系模型,通过严格检验,所得2个回归模型的复相关系数R分别为:0.978、0.978,留一法交互检验复相关系数Rcv分别为:0.977、0.978。结果表明,AEIV和AHSI具有表征能力强、物化意义明确等优点,所建模型具有良好的稳定性和估计能力。

Based on two dimensional topologic characterization,both atomic electronegativity interaction vector（AEIV） and atomic hybridization state index（AHSl） were developed for expression of chemical microenvironment and atomic hybridization state. By applying AEIV and AHSI, we successfully predicted 13C chemical shift of 375 equivalent carbon atoms in 35 naphthalene derivatives. Two quantitative models were established using the multiple linear regression（MLR） and stepwise multiple regression（SMR）. The correlation coefficients R and leave-one-out cross-validation Rcv were 0. 978, 0. 978 and 0. 977, 0. 978, respectively. The results show that both AEIV and AHSI are excellent topological indexes with satisfactory estimation stability and favorable predictability.