Accurately Predicting HOMO Energy of Linear Fluorinated Carbonates via Simply Counting the Number of Fluorine Atoms

Abstract

The HOMO energy levels of fluorinated carbonates are pivotal to their stability in high-voltage batteries and are thus, a key parameter to determine the potential of the target molecules as electrolyte solvents. The typical determination of HOMOs by professional quantum-chemistry calculations is challenging for commercial enterprises. Thus, the development of a convenient and rapid strategy becomes increasingly demanding. Herein, based on density functional theory (DFT) calculations on 192 fluorinated carbonate derivatives and data analysis, a multiple linear regression (MLR) was deduced to correlate HOMO energy with the simple structural parameters, i.e., the number of fluorine atoms on α-, β-, and γ-C sites of the carbonates, and the number of β-alkyl (denoting the branching of the α-alkyl group) of the carbonate group, as descriptors. There is excellent agreement between the predicted HOMO energy and the DFT calculated one, with a linear correlation coefficient of R-Square (R²) = 0.946, mean absolute error (MAE) = 0.076 eV, and root mean squared error (RMSE) = 0.095 eV. Using the convenient MLR model and the easily accessible descriptors, a platform was developed for predicting the HOMO levels of fluorinated carbonates promptly.