Graphite is the standard anode material in Li-ion cells due to its high capacity of 372 mAh g-1, good cycling stability and low price. Lithium ions can reversibly intercalate into the graphite host structure at low potentials of 80 – 120 mV vs. Li/Li+. While the low intercalation potential is favorable in terms of energy density, it can cause lithium plating if cells are charged at high rates or at low temperatures. Lithium metal plating adversely affects the cell life due to the loss of active lithium and electrolyte and also poses a serious safety hazard because of internal short circuiting of the cell. For the development of fast and safe charging protocols, a detailed understanding of the plating mechanism is necessary, as lithium plating is the rate limiting step during the charge of Li-ion cells. Unfortunately, there are no analytical techniques available to directly monitor the occurrence of lithium plating on graphite anodes during cell operation, expect for a recent operando neutron diffraction analysis.[1] In general, lithium plating is only indirectly observed based on specific features in the subsequent discharging profile[2] or on the observation of decreasing Coulombic efficiencies[3]. Both methods are elegant and simple, but also have specific limitations: the former fails to detect electrochemically inactive “dead” lithium, while the latter cannot distinguish whether the missing lithium is present in the metallic state (dead lithium) or in ionic form (SEI components). Additionally, microscopic techniques can yield valuable structural and morphological information regarding lithium metal deposits, but are limited to post-mortem analysis of cycled electrodes. In a previous publication,[4] we showed that operando electron paramagnetic resonance (EPR) spectroscopy can be used for the detection of mossy lithium formation on lithium metal anodes. In the present work we extend this concept and suggest to also use operando EPR spectroscopy for the time/voltage resolved and semi-quantitative detection of lithium metal plating on graphite anodes upon cell charge. The spectro-electrochemical cell design has been adapted to implement a reference electrode for the precise determination of the graphite electrode potential. Firstly, the good homogeneity of the electrochemical processes in the EPR cell design are demonstrated, exploiting the characteristic color change of the graphite electrode as a function of the state of charge. In the next step, lithium plating is forced on a fully lithiated graphite electrode. The upper panel in Figure 1 shows the typical voltage profile for a galvanostatically controlled lithium plating process. The exemplary EPR spectra (Figure 1b), measured at positions as indicated in Figure 1a, show two distinct features which differ in signal width and can therefore be discriminated. The broader signal can be assigned to lithium intercalated in graphit (LiCx) and the very narrow signal to metallic lithium. The nearly linear increase of the lithium metal signal intensity (lower panel in Figure 1a) nicely displays the semi-quantitative nature of the EPR technique. Then, the influence of the charging rate on both the amount and the reversibility of lithium metal plating is investigated. It is found, that the amount of plated lithium increases with the charging rate whereas the reversibility decreases. According to a comparison of the EPR results with the Coulombic inefficiency, the majority of the “missing” lithium is “dead” lithium, i.e., it remains trapped in the graphite anode in its metallic form without electrical contact to the electrode. Lastly, the lithium plating process is also investigated at sub-ambient temperatures and in the presence of electrolyte additives. The results are discussed in the context of designing a fast and safe charging protocol.
Read full abstract