中文摘要：

运用Gaussian 03W对209种多溴联苯醚（PBDEs）在B3LYP/6-31G＊水平上进行结构优化.应用偏最小二乘法（PLS）建立PBDEs的相对色谱保留时间（RRT）与分子结构参数的定量结构与活性关系（QSAR）模型,并对83种未知的PBDEs的RRT进行预测.实验结果表明：模型的累计交叉有效判别系数Q2cum为0.931,拟合相关系数平方R2Y（cum）为0.960,预测值与实验值的残差范围为-0.083~0.168,模型具有较强的预测性;分子总能量、次低占有轨道能量、电子空间广度、最低空轨道能量等量子化学参数对模型的贡献较大,其变量投影性指标值（VIP）分别为1.858,1.741,1.719,1.712.

英文摘要：

In this paper,the molecular structure of all 209 polybrominated diphenyl ethers（PBDEs）was optimized by using Gaussian 03 Wsoftware at the B3LYP/6-31G＊level.Then quantitative structure-activity relationship（QSAR）between relative retention time（RRT）of PBDEs and the molecular structure parameters were correlated by partial least squares（PLS）.The RRT of the 83 unknown PBDEs was also predicted.The results showed that the cross validation（Q2cum）and correlation coefficient of fitting（R2Y（cum））of the optimized model RRT were 0.931 and 0.960 respectively.Residual range of predicted and experimental values was between-0.083 to 0.168,which indicates that the model has a strong predictive ability.Furthermore,there existed high correlation between molecular total energy,next lowest unoccupied molecular orbital,electronic spatial extent,lowest unoccupied molecular orbital and RRT of PBDEs,and their variable importance in projection（VIP）was 1.858,1.741,1.719 and 1.712 respectively,which took greater contribution to the model.