醇类化合物热值改进剂的定量结构-色谱保留相关性研究

Study on quantitative structure-rention relationships of alcohol compounds calorific value improver

计算表征了醇类化合物热值改进剂的部分分子连接性指数和5个量子化学参数,同时运用多元线性回归方法对参数进行了优化筛选,在2种不同极性强弱的固定相(PH-5、RTX-WAX)上,分别建立了9种和11种醇类化合物热值改进剂的气相色谱保留值(t)与分子连接性指数和量子化学参数的定量结构-色谱保留关系(QSRR)模型。其相关系数分别为R=0.997和R22=0.971,关系式可分别表示为模型1:t1=-5.972-0.047 ET-1.857~3X_p^v和模型2:t_2=-2.409-0.034 ET,在2种固定相上的QSRR模型的建模计算值复相关系数(R)、采用留一法(leave-one-out)交互检验复相关系数(R)分别为:R_1=0.998、R_(1CV)=0.999和R_2=0.985、R_(2CV)=0.999,结果表明:模型1和模型2与色谱保留指数显著相关的分别有总能量ET、三阶分子连接性指数~3X_p^v和总能量ET,所建模型具有良好稳定性和预测能力,分子连接性指数和量子化学参数能较好地反映化合物的结构特征。

Abstract： The relationships between quantitative structure-retention of alcoholic compounds calorific value improver were researched after some parameters including some molecular connectivity indexes and 5 quantum chemistry parameters of 11 preservatives con-taining alcoholic compounds were calculated and optimizationally selected by multiple linear regression method. On the two different polarity strength stationary phases （ PH-5RTX-WAX） set up respectively nine and eleven kinds of alcoholic compounds calorific value improver of the gas chromatographic retention values（t）of quantum chemistry parameter and molecular connectivity index of the quantitative structure-chromatographic retention relationships （ QSRR） model. The correlation coefficient R1^2= 0. 997 and R2^2 =0. 971, the relation can be expressed as model 1 ： l1^ = -5. 9 7 2 - 0 . 0 4 7 ET -1. 8 5 7 3X p a n d m o d e l 2 ： t2 = -2. 409-0. 0 3 4 ET. O n two kinds of stationary phase QSRR modeling of the model calculation value of the multiple correlation coefficient （R） , using one method （leave-one-out）interaction test multiple correlation coefficients（R） Respectively R2 =0 . 9 9 5.R2CV =0 . 9 9 9 an d R2 =0 0. 999. The results showed that model 1 and 2 of the chromatographic retention indexes were significantly associated with the total energy ET, third-order molecular connectivity index 3X a n d total energy ET. T h e mo d e l has a good stability a n d predictive validity. Molecular connectivity indices and quantum chemical parameters can better reflect the features of the compound structure.