Degradation Mechanism of Nickel Manganese Cobalt Oxide-Type Commercial Lithium-Ion Cells By Long-Term Cycle Tests

Abstract

Introduction Lithium-ion battery life is expected to increase, and the need for degradation analyses has grown [1]. Among the degradation mechanism, the change in the electrode surface state has significant influence on lithium-ion battery performance. In this study, a long-term cycle test was conducted using commercial 18650-type lithium ion cells for a current rate of 1 C at 25 °C. The electrode surface state was investigated by X-ray photoelectron spectroscopy (XPS), hard X-ray photoelectron spectroscopy (HAXPES), and transmission electron microscopy (TEM). Experimental The cycle test was performed using a 18650-type commercial lithium-ion cell equipped with a Li(Ni1/3Mn1/3Co1/3)O2 cathode and a graphite anode. After cycle test completion, the cell was analyzed at a state-of-charge of 50% (= 3.694 V) by electrochemical impedance spectroscopy at 25 °C. The degraded single electrode property was measured by preparing a half cell against lithium metal. The electrode surface state was characterized by XPS and HAXPES. HAXPES measurements were performed at the BL46XU beamline of SPring-8. The cathode surface state was also investigated by TEM. The electrolytic solution was analyzed by 1H- and 19F-NMR. Results and discussion During the cycle test, the discharge capacity decreased gradually and then drastically above 1500 cycles. The capacity retention measured at 1 C reached below 10% after 1700 cycles. Nyquist plots before and after cycle test showed an increase in electrolyte and charge transfer resistances (Fig. 1). To understand the degradation mechanism, the degraded cell was disassembled after the cell was discharged at C/20, and their electrodes and electrolytic solution were removed under argon atmosphere. Half cells were prepared, and measured electrodes were characterized. After the first cycle, the degraded cathode exhibited a coulombic efficiency of 186%, indicating that it was partially charged when the cell was opened. Conversely, the degraded anode displayed a coulombic efficiency of about 100%, showing it was fully discharged. This mismatch between cathode and anode state-of-charge may explain the decrease in cell capacity. In addition, charge/discharge curves in the second cycle showed a capacity loss of about 20% at the cathode compared to initial condition. XPS and HAXPES spectra of Mn 2p, Co 2p, and Ni 2p were measured to investigate the cathode surface state. XPS and HAXPES detection depths are approximately 6 and 30 nm, respectively. No spectral changes were observed for Mn 2p 1/2 and Co 2p 1/2. Figure 2 shows the XPS and HAXPES spectra of Ni 2p 1/2. Ni 2p 1/2 XPS peak shifted toward higher energies (around 1.9 eV), indicative of the higher valence state of Ni at the cathode. Ni 2p 1/2 HAXPES peak only slightly shifted (around 0.7 eV) as a potential result of the oxidation of Ni2+, consistent with the partially charged state of the cathode. The difference between XPS and HAXPES suggests that the cycle test changed the cathode surface state. The structural properties of the cathode surface were evaluated by TEM. The lattice fringes of the layered rock-salt Li(Ni1/3Mn1/3Co1/3)O2 structure disappeared from the active material surface, and the electron diffraction pattern of the surface was unclear. These results indicate that amorphization occurred at the cathode surface (about 3 nm), augmenting the charge transfer resistance. Moreover, electrolyte analysis revealed that LiPF6was decomposed, explaining the increase in electrolyte resistance. In summary, the long-term cycle test revealed that the capacity loss stemmed from the shift of cathode and anode reaction regions and cathode degradation. Furthermore, the increase in impedance spectrum resulted from cathode surface amorphization and LiPF6decomposition. Acknowledgment Synchrotron radiation experiments were performed at the BL46XU of SPring-8 with the approval of the Japan Synchrotron Radiation Institute (JASRI) (Proposal No. 2013A1234, 2014A1558, 2014B1015, and 2014B1594) [1] T. Matsuda et al. 226th ECS meeting A5-343 Figure 1