Degradation Diagnostics for Commercial Lithium-Ion Batteries

Tuesday, 28 July 2015: 16:00
Carron (Scottish Exhibition and Conference Centre)
C. R. Birkl (University of Oxford, Department of Engineering Science), E. McTurk (University of Oxford, Department of Materials), M. Roberts (University of Oxford), P. G. Bruce (University of Oxford, Department of Materials), and D. A. Howey (University of Oxford, Department of Engineering Science)
We present a method to diagnose electrode-specific degradation in commercial lithium ion (Li-ion) cells. The method is used to estimate the loss of active electrode material and lithium inventory resulting from long-term cycling of commercial cells.

Li-ion cells degrade due to a complex interplay of physical and chemical processes [1, 2], which lead to performance deterioration and, ultimately, cell failure. The various degradation mechanisms progress at different rates and have distinct effects on cell performance and lifetime. Some mechanisms, for instance lithium plating and dendrite formation, can result in internal short circuits and, in the worst case, fires or explosions [3]. The ability to identify the prevalent degradation mechanisms at any point in time is therefore crucial in order to anticipate the remaining lifetime of a cell accurately and to warn against imminent cell failure.

Many degradation mechanisms have a noticeable effect on the cell’s open circuit voltage (OCV) [4, 5]. The proposed method is designed to harness the information contained in the OCV by fitting a phenomenological OCV model [6] to the OCV of cells measured at different stages of their cycle life. The OCV model consists of parametric electrode sub-models and a full cell model, which combines the electrode OCVs by adjusting the loading ratio (the ratio of anode versus cathode capacity) and the offset (the portion of un-utilized cathode capacity), see Figure 1. Fitting the loading ratio and offset at various stages in the cell’s life allows estimation of the loss of active electrode material as well as the loss of lithium inventory. By tracking the rates of loss of active material and cyclable lithium, projections about the cell’s end-of-life can be made. Moreover, registering the loss of a critical amount of active anode material might prevent lithium plating and associated risks of cell failure.

The OCV model has been fitted to experimental data recorded on a 4 Ah, LiCoO2 pouch cell, which was cycled at 2C and 45°C, with pseudo OCV curves (C/20 charge and discharge) recorded every 100 cycles. Initial results have shown high accuracies of OCV model fits, with RMS errors < 15 mV (see Figure 2). In order to validate identified degradation mechanisms, the diagnostic procedure is repeated on commercial LiNiMnCoO2cells, which are subjected to post mortem analyses including electrochemical analyses on 3-electrode cells, scanning electron microscopy and XRD.


J. Vetter et al. , J. Power Sources, vol. 147, pp. 269-281, 2005.


P. Arora and R. E. White, J. Electrochem. Soc., vol. 145, no. 10, pp. 3647-3667, 1998.


Q. Wang et al., J. Power Sources, vol. 208, pp210-224, 2012.


M. Dubarry et al., J. Electrochem. Soc., vol. 9, no. 10, pp. A454-A457, 2006.


M. Dubarry et al., J. Power Sources, vol. 219, pp.204-216, 2012.


C. R. Birkl et al., Abstract no. 63, 227th ECS Meeting, 2015.