Evaluating cryogenic grinding variables to maximize valuable metal liberation from spent lithium-ion batteries

Abstract

This study explores the liberation of valuable metals from spent Lithium Ion Batteries (LIBs) following cryogenic grinding with varying grinding times, frequencies, and particle sizes. We characterize the physicochemical changes in lithium metal oxide particles focusing on the mineral phases, semi-quantification, and liberation of Co, Ni, and Mn. A total of 18 cathodic material samples underwent cryogenic grinding. Additionally, one sample was processed through conventional grinding for comparison. Grinding variables included frequencies of 10, 20, and 30 Hz, and durations of 3, 5, and 7 min. Subsequently, all tests were analyzed to determine their D₈₀ value. The response surface methodology was used to ascertain the frequency and time variables with the greatest influence on D₈₀ values. A predictive regression model was then applied to find the best D₈₀ value. This statistical analysis was also used to choose the samples for X-ray Diffraction (XRD), Scanning Electronic Microscope − Energy Dispersive Spectroscopy (SEM-EDS), and Quantitative Evaluation of Minerals by Scanning Electron Microscopy (QEMSCAN) characterization. Characterization results revealed that LiCoO₂ and Co₃O₄ particles have distinct spatial distributions compared to (Li_0.65Ni_0.05)(NiO₂) and Li_1.27Mn_1.73O₄ particles, which exhibited similar spatial distributions. LiCoO₂ and Co₃O₄ particles achieved a high degree of liberation at 20 and 30 Hz frequencies and were liberated even at coarser particle size distributions of approximately < 450 μm when the frequency was equal to 10 Hz. In contrast, (Li_0.65Ni_0.05)(NiO₂) and Li_1.27Mn_1.73O₄ particles demonstrated a much lower degree of liberation tending to aggregate and be locked, but they were liberated at < 38 μm size distributions at all frequencies analyzed. Results show that cryogenic grinding is superior to the traditional grinding method. Furthermore, the specific results of the predictive regression model indicate the optimal D₈₀ value of 55.82 μm can be achieved with a grinding frequency of 39 Hz and 7 min of constant grinding, enhancing the overall recovery efficiency for all target metals.