欢迎访问《环境科学学报》编辑部网站！

快速检索			高级检索

研究论文

Prediction of K_OC of polar organic compounds based on molecular connectivity indices

摘要：研究了分子连接性指数与极性有机化合物吸附系数(以有机碳计)的关系。通过对建模组中211种极性化合物分子连接性指数与logKOC进行回归分析,发现仅根据分子连接性指数不能准确地计算其KOC值。当在模型中引入极性校正因子后,计算的准确度明显提高。据此建立了估算极性有机化合物KOC的定量模型。对所研究的231种极性化合物,平均估算误差为0.31个对数单位,45%的估算误差在0.2个对数单位之内。

Abstract：Relationahip beteen the molecular structure and K_OC of polar organic compounds were investigated based on molecular connectivity indices. The total 231 polar compounds were divided randomly into two sets: the training set and the testing set. Using ¹χ, ¹χ^v, ²χ ,and ²χ^v variables, the coefficient of determination (R₂) of the regression equation was only 0.564. The mean residual of the calculated K_OC of the compounds in the training set was 0.44 log units. The results showed that the K_OC of polar compunds can not be estimated by MCI alone due to the existence of the hydrophobic groups. However, by introducing thirteen polar factors into the regression equation, the precision of the model was improved greatly. To most polar compounds in the training set, their K_OC can be calculated by ¹χ^v, ²χ and the polar factors. The model was verified by a set of independent data. The final model was built on the basis of the total 231 measured K_OC values. The mean residual of the estimated K_OC of the 231 polar compounds is 0.31 log units and 45% of the residuals are less than 0.2 log unit.