Prediction of heat capacity of ionic liquids: A simple group contribution method

Abstract

A simple method based on the mathematical relational expression and group contributions is proposed to predict the heat capacity of ionic liquids (ILs). A comprehensive heat capacity database of 11873 measured data points for 310 ILs at pressure of one atmosphere, and the temperature and the heat capacity in the ranges of 167.70-524.87 K and 84-2040 J⋅mol−1∙K−1, respectively, is established. The optimal group contribution values and universal parameters are determined from fitting 10675 measured data with the Newton's central difference algorithm by the categorical method (M1) and the overall method (M2), where the difference is whether each class of ILs has independent group contribution parameters. Average absolute relative deviations (AARDs) between the predicted and the measured values are 2.36% and 3.20% for the M1 and M2 methods, respectively. The test set including 1198 data points for 62 ILs is used to test and evaluate the model and parameters which have been trained on the above training data set. The group contribution method proposed in this work achieves the accurate prediction of the heat capacity of a variety of ionic liquids over a wide temperature range, which provides new ideas for predicting the other physical properties of ionic liquid systems.