Machine learning-based design of electrocatalytic materials towards high-energy lithium||sulfur batteries development.

The practical development of Li | |S batteries is hindered by the slow kinetics of polysulfides conversion reactions during cycling. To circumvent this limitation, researchers suggested the use of transition metal-based electrocatalytic materials in the sulfur-based positive electrode. However, the atomic-level interactions among multiple electrocatalytic sites are not fully understood. Here, to improve the understanding of electrocatalytic sites, we propose a multi-view machine-learned framework to evaluate electrocatalyst features using limited datasets and intrinsic factors, such as corrected d orbital properties. Via physicochemical characterizations and theoretical calculations, we demonstrate that orbital coupling among sites induces shifts in band centers and alterations in the spin state, thus influencing interactions with polysulfides and resulting in diverse Li-S bond breaking and lithium migration barriers. Using a carbon-coated Fe/Co electrocatalyst (synthesized using recycled Li-ion battery electrodes as raw materials) at the positive electrode of a Li | |S pouch cell with high sulfur loading and lean electrolyte conditions, we report an initial specific energy of 436 Wh kg⁻¹ (whole mass of the cell) at 67 mA and 25 °C. The atomic-level interactions among electrocatalytic sites in Li | |S batteries remain unclear. Here, authors propose a multiview machine-learned framework to evaluate electrocatalyst features using limited datasets and intrinsic factors, thus enhancing the understanding of electrocatalytic sites. [ABSTRACT FROM AUTHOR]