In this study, a wide literature survey has been conducted to gather an extensive set of thermal decomposition temperature (Td) data for ionic liquids (ILs). A data set consisting of Td data for 586 ILs was collated from 71 different literature sources. Using this data set, a reliable quantitative structure-property relationship has been developed. In order to consider the effects of the anion and cation on the Td of ILs, both anion-based and cation-based molecular descriptors were considered. Finally, a genetic function approximation method was used which selected 6 molecular descriptors for anions, and 6 molecular descriptors for cations to develop the model. The predictive capability of the 12-parameter model was evaluated using several validation techniques. Its applicability domain is discussed. The proposed model produces an acceptable average relative deviation of less than 5.2% taking into consideration all 586 experimental data values.
Chemical Engineering Science, 2012, Vol 84, p. 557-563