【摘要】 以原子类型电拓扑状态指数(ETSI)表征19个酚类化合物的分子结构,应用基于预测的变量选择与模型化(VSMP)方法,建立了酚类化合物在产甲烷和脱氮两种厌氧环境下的降解效果与分子结构的定量相关模型.结果表明:在产甲烷环境中,影响酚类化合物厌氧降解的主要结构因素是由4个ETSI描述子对应的子结构碎片,即aCHa,aaC—,—NH2和—OH,其中子结构aCHa和aaC—与酚环母体骨架密切相关,而—NH2和—OH反映取代基的变化.通过多元线性回归法建立的产甲烷条件下酚类化合物的结构-厌氧降解性相关模型(QSBR)发现,其模型估计相关系数r=0.9173,LOO检验相关系数q=0.8461;反硝化条件下,影响酚类化合物厌氧降解的主要结构因素是C,aaC—,O和—Cl,其中子结构aaC—与酚环母体骨架相关,而C,O和—Cl反映支链或取代基的变化.其模型估计相关系数r=0.8953,LOO检验相关系数q=0.8488.以上结果均表明模型具有良好的估计能力与稳定性,进而从建模的角度证明了厌氧环境中,电子受体的选择影响酚类化合物的降解效果和机理.更多还原

【Abstract】 Electrotopological state index(ETSI)for atom types was used to describe the structures of 19 phenols. A quantitative linear relationship between the ETSI descriptors and the anaerobic biodegradability of phenols under methanogenic and denitrifying conditions was developed using the variable selection and modeling based on predicting(VSMP). It was found that some main structural factors influencing the anaerobic biodegradation of phenols under methanogenic condition are 4 substructures such as aCHa, aaC—(where"a"refers to a chemical bond in the aromatic ring),—NH2,and —OH, which were related to the molecular skeleton of phenol and substituent groups on phenyl ring. The best 4-variable model, with the calibrated correlation coefficient of r= 0.9173 and the validated correlation coefficient of q=0.8461, were built by multiple linear regression .Under denitrifying condition the main structural factors influencing the biodegradation of phenols are aaC—(where "a"refers to a chemical bond in the aromatic ring too),C,O,and—Cl, which were related to the molecular skeleton of phenol and substituent groups on phenyl ring. The model containing best 4-variable, with the calibrated correlation coefficient of r=0.8953 and the validated correlation coefficient of q=0.8488, was built by multiple linear regression. The two models with statistic parameters show a good estimation ability and stability and the results also indicate that the presence of different electron acceptors affect the initial mechanism of biotransformation under anaerobic condition.更多还原