Design and preparation of sulfonated polymer membranes for Zn/MnO2 flow batteries with assistance of machine learning

Abstract

The performance of a zinc ion battery highly depends on the comprehensive properties of the battery electrolyte, especially its selective conductivity to zinc ions (Zn2+). In order to purposefully obtain a competent polymer electrolyte for superior performance of the battery, we construct the models of machine learning to predict the contribution of polymer functional groups to ionic conductivity of both Zn2+ ions and protons by using the gradient boosting decision tree (GBDT) algorithm. Following the predicting results by machine learning, a series of crosslinked polymers of poly (terphenyl methyl-piperidone bromo-trifluoroacetophenone) (PTPT) are synthesized and sulfonated to fabricate the sulfonic acid group containing membranes (SPTPT). The prepared membrane reaches a Zn2+ conductivity of 12 mS cm−1, and a proton (H+) conductivity of 22 mS cm−1 in water, respectively, at 30 °C. Using the proposed membrane as the electrolyte, the Zn/MnO2 flow battery working at room temperature delivers a peak power density of 150 mW cm−2, a specific capacity of 1.95 mAh cm−2, and a cycling capacity retention rate of 71% after 1000 cycles at 30 mA cm−2.