Molecular distance-edge vector (μ):: An extension from alkanes to alcohols

被引:34
作者
Liu, SS [1 ]
Liu, HL
Xia, ZN
Li, ZL
机构
[1] Chongqing Univ, Coll Bioengn, Chongqing 400044, Peoples R China
[2] Chongqing Univ, Coll Environm & Chem Engn, Chongqing 400044, Peoples R China
[3] Guilin Inst Technol, Dept Appl Chem, Guilin 541004, Guangxi, Peoples R China
[4] Chongqing Univ, Coll Environm & Chem Engn, Chongqing 400044, Peoples R China
[5] Xiangtan Teachers Coll, Dept Chem, Xiangtan 411100, Peoples R China
来源
JOURNAL OF CHEMICAL INFORMATION AND COMPUTER SCIENCES | 1999年 / 39卷 / 06期
关键词
D O I
10.1021/ci990011f
中图分类号
O6 [化学];
学科分类号
0703 ;
摘要
The oxygen atom in the hydroxyl group OH of an alcohol molecule is considered a pseudo-carbon atom, like a carbon atom in alkanes, for our present case of extension from alkanes to alcohols. The ratio of electronegativity of oxygen (ENO) to that of carbon (ENC) is defined as the relative electronegativity, REN ENO/ENC, of a oxygen atom under the condition of REN being 1 for a carbon atom; similarly, the value of bond length for the bond C-O (BLO) over that of the bond C-C (BLC) is regarded as the relative bond length, RBL = BLO/BLC, of the bond C-O with RBL being equal to 1 for C-C. The molecular distance-edge vector, i.e., vector mu for alkanes can be extended to describe the molecular structure of alcohols instead of alkanes. A quantitative structure-property relationship (QSPR) equation can be modeled by using multiple linear regression (MLR). The first model, M1, between the modified mu vector and boiling points of 106 alcohols is created with correlation coefficient being R = 0.9951 and root-mean-square error being rms = 3.222 degrees C between the values estimated by the above QSPR model and the experimental boiling points. With 30 compounds drawn from all 106 alcohols we construct an external prediction set, and the remaining 76 ones construct an internal calibration set, and the second model, M2, is developed with the statistics being R = 0.9946 and rms = 3.480 degrees C. The 30 boiling points in the external testing set can be predicted by model M2, and good prediction results are obtained with R being 0.9964 and rms being 2.586 degrees C.
引用
收藏
页码:951 / 957
页数:7
相关论文
共 14 条