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.

中文翻译：

---

通过机器学习辅助衍射断层扫描探测疲劳层状阴极材料中的颗粒碳/粘合剂降解行为

了解锂离子电池（LIB）电化学循环过程中反应异质性如何影响正极材料对于揭示其电化学性能至关重要。然而，通过实验验证这些反应已被证明是一个挑战。为了解决这个问题，我们采用扫描 μ-XRD 计算机断层扫描来检查富镍层状 LiNi0.6Co0.2Mn0.2O2 (NCM622) 和富锂层状 Li[Li0.2Ni0.2Mn0.6]O2 (LLNMO)。通过利用机器学习 (ML) 技术，我们仔细检查了广泛的 μ-XRD 图案数据集（每片约 100,000 个图案），以揭示晶体结构和微应变的空间分布。我们的实验结果明确揭示了这些材料的独特行为。 NCM622 表现出结构退化和晶格应变，这与二次颗粒的尺寸密切相关。较小的颗粒和与碳/粘合剂基体接触的较大颗粒的表面在长期循环后会经历加剧的结构疲劳。相反，LLNMO 颗粒的表面和本体在高电压循环过程中都会承受严重的应变引起的结构退化，导致显着的电压衰减和容量衰减。这项工作具有微调先进层状材料的微观结构和操纵复合电极结构的潜力，以提高锂离子电池及其他性能。