Voltage Mining for (De)Lithiation-Stabilized Cathodes and a Machine-Learning Model for Li-Ion Cathode Voltage

• Published in: Meeting abstracts (Electrochemical Society) Vol. MA2024-02; no. 3; p. 373
• Main Authors: Li, Haoming Howard, Chen, Qian, Ceder, Gerbrand, Persson, Kristin A.
• Format: Journal Article
• Language: English
• Published: The Electrochemical Society, Inc 22.11.2024
• ISSN: 2151-2043; 2151-2035
• DOI: 10.1149/MA2024-023373mtgabs

Advances in Li-metal anodes have inspired interests in discovery for Li-free cathodes, most of which can be more formally defined as “delithiation-stabilized”, as opposed to conventional Li-ion battery cathodes, most of which are “lithiation-stabilized”. In this study, we combine the cathode voltage information from the Materials Project consisting mostly of lithiation-stabilized structures, and from a newly available dataset with mostly delithiation-stabilized structures. The resulting voltage distributions with respect to redox pairs and anion types for both classes of compounds reemphasize known design principles for high-voltage cathodes, which favor later transition metals in their higher oxidation states and polyaion groups. In addition, a machine learning model for voltage prediction based on chemical formulae is constructed, and shows state-of-the-art performance when compared to two established composition-based ML models for materials properties predictions, Roost and CrabNet.