Fundamental Study of Ionic Liquids Melting Point Structure-Property Using Machine-Learning Method

For the application of ionic liquids (ILs), the intense interest has been due to the various novel properties, such as tunable ionic conductivity, negligible vapor pressure, liquid phase within a wide temperature range, non-flammability at ambient condition, etc. Among these properties, the melting point (Tm) of ionic liquid (IL) is very important in various applications. However, the (Tm) can change considerably depending on the molecular structures of the anion and cation. Recent years have seen a huge rise in the successful application of the machine or statistical learning type approaches to the discovery and in silico design of new novel materials. Deep-Learning algorithms can take structure-properties from extremely large data sets and use them to create a predictive tool based on hidden patterns and correlations. In this study, we will explore the use of various machine learning and deep learning methods to predict the melting points of various ILs that consist of several different cation and anion classes. From this preliminary study, we hope some important molecular descriptors that dictate the melting temperature of ionic liquids can be found, and subsequently can be used as new design rules.