Assessment of ionic liquid electrolytes for high‐performance lithium‐sulfur batteries using machine learning

To realize the full potential of the lithium-sulfur (Li-S) batteries, the shuttling of the polysulfide (PS) intermediates between the electrodes should be prevented. Moreover, inactive material mass in the battery pack should be minimized to increase the system-level energy density. In this respect, ionic liquids (IL) are getting increasing attention as they can reduce the PS shuttle mechanism with their limited PS solubility and functionality in lean electrolyte conditions. In this work, a dataset is constructed from the experimental literature data, which uses ILs as their electrolytes, to analyze the important cell variables and promising ILs for high cell- and system-level performances using association rule mining. The study reveals that with the help of ILs, high-performance cells with high sulfur loadings and low electrolyte-to-sulfur ratios can be attained. It is found that 1-ethyl-3-methylimidazolium bis(trifluoromethanesulfonyl)imide (EMI_TFSI) or ionic liquids with the n-methyl-n-propylpyrrolidinium (P13) cation lead to higher cell capacities, whereas Li(tetraglyme) bis(trifluoromethanesulfonyl)imide (Li[G3]_TFSI) is found promising for high system-level energy densities. Highlights Machine learning method of association rule mining is used to investigate the effect of critical materials and cell design factors on the performance of Li-S batteries using ionic liquid electrolytes. ARM was employed first for the experimental peak discharge capacity data and then for energy densities and specific energies predicted by a proposed system-level performance model. IL electrolytes show promising results in peak discharge capacities at high sulfur loadings and low electrolyte-to-sulfur ratios; however, their cycling performance at high C-rates needs to be improved.