In this work, to accelerate the intrinsic modulation of catalysts in Li-S batteries for optimizing the adsorption capacity in local chemical environments, the macroscopic features of elemental ratios, crystal structures and facets were transformed into local coordination structures to fundamentally elucidate the mechanisms, and the machine learning model is employed to predict and interpret the interactions of catalyst-LiPS (Li2S6) and catalyst-solvent (DOL).
descriptor.py: calculate the designed hybrid multi-view descriptor (HMVD) that describes compositional, structural, and binding features.
model.py: train the model by using five-fold cross-validations.
Yeyang Jia#, Zhilong Wang#, Zhiyuan Han#, Junfeng Li, Mengtian Zhang, Zhoujie Lao, Yanqiang Han, Runhua Gao, Jing Gao, Zhiyang Zheng, An Chen, Hong Li, Rui Mao, Kehao Tao, Jinjin Li*, Guangmin Zhou*. Variable and Intelligent Catalyst Design Based on Local Chemical Environments in Sulfur Redox Reactions. Joule 101878 (2025). https://doi.org/10.1016/j.joule.2025.101878