Computational and Machine Learning-Assisted Discovery and Experimental Validation of Conjugated Sulfonamide Cathodes for Lithium-Ion Batteries

Conjugated sulfonamides (CSAs) stand out for their propitious electroactivity and notable stability in ambient conditions, making them suitable candidates for high-potential cathode materials in lithium-ion batteries (LIBs). This study employs a combination of machine learning, semi-empirical quantum mechanics, and density functional theory methods to evaluate a large library of 11 432 CSA molecules, focusing on material properties crucial for application in batteries, such as synthetic complexity, redox potential, gravimetric charge capacity, and energy density. After applying the thresholds for the synthetic complexity score at 2.62 and the redox potential at 3.25 V versus Li/Li+, we identify 50 CSA molecules that are easy to synthesize and suitable for the positive electrode in LIBs. By ranking on the basis of redox potential, 13 CSA molecules having potentials greater than 3.50 V versus Li/Li+ are identified. Through further investigations using molecular dynamics simulations on these reactant molecules and their lithiated products, a molecule is singled out for synthesis and electrochemical evaluation. This molecule, lithium (2,5-dicyano-1,4-phenylene)bis((methylsulfonyl)amide)(Li2-DCN-PDSA), demonstrates a redox potential surpassing those previously reported within the class of CSA molecules. Moreover, the study explores the quantitative structure-property relations of CSAs, yielding insights for the development of CSA-based LIB cathode materials, informed by the comprehensive data assembled. An outline of the study's methodology, where a vast molecular library is digitally cataloged and used for computational screening. Through the application of various property criteria, the promising candidates are shortlisted, leading to the selection of one for experimental validation. image