Data And QSAR/QAPR Study For Toxicity Of Ionic Liquids

Ionic liquids (ILs) are gaining increasing interests due to both of physical and chemicalcharacteristics. They are considered as the sustainable “Green Product”. It is inevitable thatthere are always parts of ILs could be dispersed into the environment finally. Therefore, theviewpoint concerned with the security of ILs to the environment is being doubted graduallyby many scholars for their further researches. Many ILs have been synthesized and a greatnumber of data about their toxicity property have been reported. However, there is not yetone comprehensive database publicly available for the researchers to consult. Therefore, it isnecessary to collect toxicity data of ILs for developing databases.Generally, the properties of materials are mainly determined by their structures. Thestudy of the structure-property relationship is thus the important basis of design andapplication of the new green ILs. The knowledge about the structure-property relationship ofILs is very helpful to develop a new desirable database for researchers due to they can changethe structures of the available ILs to design the new compounds showing particular propertiesand low toxicity. The research of quantitative structure-property relationship will promote theprogress of ILs and provide a theoretical basis for the design of ILs.In this work, there are four parts of contents summarized as following:1. A comprehensive database on toxicity of ILs, which has been collected from69kindsof literature sources in the period from2004through2012, was established. The cations andanions were named and ordered according to their molecular weight. There are totally276kinds of cations,51kinds of anions and4749pieces of data. On the basis of the databasewhich has been established by Institute of Process Engineering, Chinese Academy ofSciences, we increased some constructures chart and CAS number of inoic liquids2. Base on our database, we got the qualitative analysis of ILs structure on the toxicityand studied the effects of cations, headgroups and anions on the toxicity of ILs too.3. In this paper, the QSPR/QSAR (quantitative structure-property/activity relationships)method was used to study the quantitative relationship between the toxicity and structure of the43kinds of imidazolium ILs,10kinds of substances were used to carry out an externaltest. The model contains six structural descriptors selected from Heuristic Method, and R~2,R~2CV, F and S~2of the model were0.9214,0.8942,70.35,0.0984respectively. Test set wasused to conduct the external validation, and the R~2was0.952. The result shows that thismodel has good reliability and can be used to predict the toxicity of Imidazolium IonicLiquids.4. QSPR/QSAR models were carried out to predict toxicities (EC50values) of a widevariety of ILs towards the Leukemia rat cell line IPC-81. Five parameters selected by theheuristic method (HM) were input to perform the studies of Multiple Linear Regression(MLR) and Support Vector Machine (SVM). The obtained two QSPR/QSAR models for thetraining set were very effective. The squared correlation coefficient (R~2) and root meansquare error (RMSE) of them were0.925and0.945,0.255and0.222, respectively. Theresults demonstrated that the prediction R~2and RMSE of test set by MLR model were0.941and0.231, and the prediction R~2and RMSE of test set by SVM model were0.931and0.290,respectively. For a few data, the MLR method showing clear physical and chemical meaningshould be more effective compared with the SVM method.