Diagnosis Method to Detect the Incorporation of Metallic Particles in a Lithium Ion Battery
Test cells were assembled with positive electrodes of LiCoO2 mixed with acetylene black as conductive filler and PVDF binder, electrolyte solution of ethylene carbonate (EC)/diethylcarbonate (DEC) (1:1 vol.) containing 1 M LiPF6, and negative electrodes of graphite mixed with PVDF binder. Lithium metal was used as a reference electrode.
Cyclic voltammetry was carried out with positive electrodes of iron plate and copper foil as working electrodes. Current onset potentials originating in the anodic dissolution of iron and copper were around 2.4 V and 3.6 V (vs. Li+/Li), respectively. Hence those metal dissolve during charging of a cell. The anodic current for copper was larger than that for iron．
Iron particle with a diameter of 150 μm, and copper particle with a diameter of 100 μm were adhered to the positive electrode surface. The fully charged cell by 1C with the iron particle exhibited no rapid voltage drop for one week whereas the cell with the copper particle showed rapid voltage drop before full charge, resulting in the short circuit much earlier than iron in agreement with the larger anodic dissolution current shown by the cyclic voltammetry. SEM and EDX observations of the surface of the negative electrode after charging showed iron and copper depositions in torus-shape.
We also repeatedly maintained the cells with those metal particles under fully charged state (SOC:100%) for 24 hours before a discharging/charging cycle to assess the time variation of impedance spectra. As a result, we find characteristic change of the Bode plot for the negative electrode in the case of iron particles between 10-1 Hz as shown in Fig. 1. Since time variation in the impedance spectra was detected, electrochemical impedance spectroscopy is effective diagnosis method to detect the metal particle incorporation.
 C. Iwakura, Y. Fukumoto, H. Inoue, S. Ohashi, S. Kobayashi, H. Tada, and M. Abe, J. Power Sources, 68, pp. 301-303 (1997).