Observation of memory effects associated with degradation of rechargeable lithium-ion cells using ultrafast surface-scan magnetic resonance imaging

2021 
Lithium-ion cells represent the most wide-spread power storage devices in portable electronics and electric vehicles. Cells can follow various degradation pathways that cause safety concerns associated with flammable components and high energy density. With the growing demand for batteries, new screening methods are becoming critical in avoiding catastrophic failures. A cell that has undergone a hazardous life cycle event (e.g., overcharging) could be identified via magnetic field patterns associated with direct currents through the cell's electrodes. Accurate detection of these patterns has been impossible due to insufficient sensitivity of recent acquisition protocols. An operando magnetic resonance imaging (MRI) methodology, surface-scan MRI, has been developed to address this challenge. The magnetic field distribution produced by an operating Li-ion cell is measured in a thin solid-state detection medium layer placed in direct contact with the cell. A speedy purely phase-encoded two-dimensional acquisition is the key to accurate distortion-free visualization of rapid charge transfer processes. At rates of multiples of C, the method is highly sensitive to onset of degradative processes in the cell's electrodes and to hazardous states preceding internal short-circuits. Surface-scan MRI is a non-destructive technique suitable for diagnostics of Li-ion and other types of cells. The method is compatible with both research and commercial cell designs. Degradation of active electrochemical materials in overcharged Li-ion cells can manifest through a hysteresis-like behaviour of local current-induced magnetic fields. Such novel “memory” effects can be potentially observed in a variety of hazardous scenarios including formation of dendrites, damage of electrodes and decomposition of electrolyte.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    20
    References
    0
    Citations
    NaN
    KQI
    []