Probing Particle‐Carbon/Binder Degradation Behavior in Fatigued Layered Cathode Materials through Machine Learning Aided Diffraction Tomography

• Published in: Angewandte Chemie International Edition Vol. 63; no. 30; pp. e202403189 - n/a
• Main Authors: Hua, Weibo, Chen, Jinniu, Ferreira Sanchez, Dario, Schwarz, Björn, Yang, Yang, Senyshyn, Anatoliy, Wu, Zhenguo, Shen, Chong‐Heng, Knapp, Michael, Ehrenberg, Helmut, Indris, Sylvio, Guo, Xiaodong, Ouyang, Xiaoping
• Format: Journal Article
• Language: English
• Published: Germany Wiley Subscription Services, Inc 22-07-2024
• Edition: International ed. in English
• Subjects: Binders (materials); Carbon cycle; Cathodes; Chemical reactions; Composite materials; Computed tomography; Cycles; Degradation; Electrochemical analysis; Electrochemistry; Electrode materials; fatigue mechanism; Heterogeneity; Lattice strain; layered cathodes; Layered materials; Learning algorithms; Lithium-ion batteries; local phase distribution; Machine learning; Materials fatigue; Microstrain; Particle decay; Particulate composites; reaction heterogeneity; Spatial distribution; Tomography; Voltage; X-ray diffraction; μ-XRD tomography; reaction heterogeneity, local phase distribution, fatigue mechanism; layered cathodes; μ-XRD tomography
• ISSN: 1433-7851; 1521-3773; 1521-3773
• DOI: 10.1002/anie.202403189

Understanding how reaction heterogeneity impacts cathode materials during Li‐ion battery (LIB) electrochemical cycling is pivotal for unraveling their electrochemical performance. Yet, experimentally verifying these reactions has proven to be a challenge. To address this, we employed scanning μ‐XRD computed tomography to scrutinize Ni‐rich layered LiNi0.6Co0.2Mn0.2O2 (NCM622) and Li‐rich layered Li[Li0.2Ni0.2Mn0.6]O2 (LLNMO). By harnessing machine learning (ML) techniques, we scrutinized an extensive dataset of μ‐XRD patterns, about 100,000 patterns per slice, to unveil the spatial distribution of crystalline structure and microstrain. Our experimental findings unequivocally reveal the distinct behavior of these materials. NCM622 exhibits structural degradation and lattice strain intricately linked to the size of secondary particles. Smaller particles and the surface of larger particles in contact with the carbon/binder matrix experience intensified structural fatigue after long‐term cycling. Conversely, both the surface and bulk of LLNMO particles endure severe strain‐induced structural degradation during high‐voltage cycling, resulting in significant voltage decay and capacity fade. This work holds the potential to fine‐tune the microstructure of advanced layered materials and manipulate composite electrode construction in order to enhance the performance of LIBs and beyond.
A unique combination of machine learning and μ‐XRD CT has been used to unveil the spatial distribution of structure and microstrain in the fatigued layered materials. Smaller NCM622 particles and the surface of larger NCM622 particles in contact with the carbon/binder matrix experience intensified structural fatigue after long‐term cycling. Serious structural fatigue occurs in both the surface and bulk of LLNMO particles during extended cycling.