While much research has focused on the electrodes and electrolyte materials, identifying suitable electrolyte additives is also crucial for enhancing the performance of Li-ion batteries (LIB).A functional additive can help stabilize the electrolytes, preventing degradation and enhancing the cycle life of LIB. It can also maintain a stable solid electrolyte interphase film on the anode without side effects of the electrolyte reduction on the anode and oxidation on the cathode. Up to now, there are several common used electrolyte additives, namely vinylene carbonate (VC), ally ethyl carbonate (AEC), 1,3-propane sultone (PS), fluoroethylene carbonate (FEC), etc. However, with a vast number of molecules awaiting testing, calculating or experimentally measuring them is not an efficient approach. As an alternative, this study has established a search map for evaluating electrolyte additives based on artificial neural networks (ANNs). An ANN model has been developed to estimate the electron affinity and ionization energy, which are crucial parameters for determining suitable electrolyte additives. This generative model can serve as a starting point for further investigations into other necessary properties, not only for electrolyte additives but also for electrolyte molecules. The results of this study are expected to provide a better materials design for LIBs and be useful for experimental fabrications.

Key words: Li-ion batteries, electrolyte additives, neural networks