A greyscale erosion algorithm for tomography (GREAT) to rapidly detect battery particle defects

Particle micro-cracking is a major source of performance loss within lithium-ion batteries, however early detection before full particle fracture is highly challenging, requiring time consuming high-resolution imaging with poor statistics. Here, various electrochemical cycling (e.g., voltage cut-off, cycle number, C-rate) has been conducted to study the degradation of Ni-rich NMC811 (LiNi0.8Mn0.1Co0.1O2) cathodes characterized using laboratory X-ray micro-computed tomography. An algorithm has been developed that calculates inter- and intra-particle density variations to produce integrity measurements for each secondary particle, individually. Hundreds of data points have been produced per electrochemical history from a relatively short period of characterization (ca. 1400 particles per day), an order of magnitude throughput improvement compared to conventional nano-scale analysis (ca. 130 particles per day). The particle integrity approximations correlated well with electrochemical capacity losses suggesting that the proposed algorithm permits the rapid detection of sub-particle defects with superior materials statistics not possible with conventional analysis.