Optimized calculations of 75 PCDDs and their parent DD were carded out at the B3LYP/6-31G＊ level by density functional theory （DFT） method. The structural parameters were obtained and significant correlation betwee...Optimized calculations of 75 PCDDs and their parent DD were carded out at the B3LYP/6-31G＊ level by density functional theory （DFT） method. The structural parameters were obtained and significant correlation between the C1 substitution position and some structural parameters was found. Consequently, the number of C1 substitution positions was taken as theoretical descriptors to establish two novel QSPR models for predicting lgKow and -lgSw of all PCDD congeners. The two models achieved in this work contain two variables （Na and Nβ）, of which r = 0.9312, 0.9965 and SD = 0.27, 0.12 respectively, and t values are all large. The variation inflation factors （VIF） of variables in the two models herein are both less than 5.0, suggesting high accuracy of the lgKow and -lgSw predicting models, and the results of cross-validation test also show that the two models exhibit optimum stability and good predictive power. By comparison, the correlation and predictive ability of the present work are more advantageous than those obtained using semi-empirical AM1 and GC-RI methods.展开更多
基金This work was supported by the National Natural Science Foundation of China (No. 20737001)(2003CB415002)China Postdoctoral Science Foundation (No. 2003033486)
文摘Optimized calculations of 75 PCDDs and their parent DD were carded out at the B3LYP/6-31G＊ level by density functional theory （DFT） method. The structural parameters were obtained and significant correlation between the C1 substitution position and some structural parameters was found. Consequently, the number of C1 substitution positions was taken as theoretical descriptors to establish two novel QSPR models for predicting lgKow and -lgSw of all PCDD congeners. The two models achieved in this work contain two variables （Na and Nβ）, of which r = 0.9312, 0.9965 and SD = 0.27, 0.12 respectively, and t values are all large. The variation inflation factors （VIF） of variables in the two models herein are both less than 5.0, suggesting high accuracy of the lgKow and -lgSw predicting models, and the results of cross-validation test also show that the two models exhibit optimum stability and good predictive power. By comparison, the correlation and predictive ability of the present work are more advantageous than those obtained using semi-empirical AM1 and GC-RI methods.
