In the experimental part of our study, we lithiate graphite anodes in half-cell-assemblies with varying, uncritical, and critical current densities and analyse the impedance behaviour in order to find anomalies which can be used for the detection of lithium deposition (LD). Due to kinetic limitations of the desired intercalation of lithium ions into the lattice structure at high current densities and low temperatures, the ions will deposit metallically on the electrode surface instead. This parasitic side reaction leads to rapid loss of lithium inventory and may also lead due to dendrite formation to separator penetration, and therewith, complete cell failure.

In the last years much research effort has been spent on detection methods of LD to prevent this severe degradation mechanism. Numerous retrospective methods [1–5] have been published but only few detect LD online during the charging process itself [6–8]. The latter allow the detection during graphite lithiation which would be highly beneficial for online charge control. The most promising publications are based on electrochemical impedance spectroscopy (EIS) during charging on experimental or commercial full-cells, but the groups show contradicting results. In this study we apply EIS during lithiation of graphite half cells, in order to solely analyse the polarisation behaviour on the relevant electrode – the graphite anode.

For our experiments we extracted graphite electrode samples from commercial high-power cells and integrated them as working electrodes (WE) in experimental cells, with lithium-foil as counter electrodes (CE) and a lithium ring reference electrode (RE). In our approach we lithiate the graphite anode with varying, critical, and uncritical current densities via the CE. The measurement of the anodic potential and the half-cell impedance are conducted via the RE to ensure that effects from the CE are eliminated. Compared to the potential analysis, the impedance analysis offers the opportunity to separate single polarisation effects, like charge transfer or solid-state diffusion, and offers a more precise interpretation of the physicochemical behaviour. Therefore, the half-cell is initially characterised with electrochemical impedance spectroscopy and the distribution of relaxation times to identify the characteristic excitation frequencies f_c of the most dominant electrochemical processes.

LD is provoked on purpose by lithiating the anode from the complete delithiated state to a degree of lithiation of 80 % at a low temperature of 5 °C. During charging, the impedances measured at the frequencies f_c, enable the tracking of single polarisation effects. In parallel the anode potential is measured to exclude the occurrence of LD as long as the potential does not fall below 0 V vs. Li/Li⁺, the reduction potential of lithium ions. After the end of charge the anodic potential and the impedances at f_c are measured for 1 h. During this relaxation phase the state-of-the art differential voltage analysis [4] is used as a reference method to proof LD and the method of impedance relaxation [5] is applied firstly on graphite half-cells.

The results show a reproducible impedance drop for critical charging events, which is in line with the majority of other studies [7, 9, 10]. The most sensitive processes seem to be the charge transfer and migration through the solid electrolyte interphase. Tracking these processes increase the sensibility of the method – and knowing which processes are relevant enables the transfer of the method to other cell systems. Furthermore, the retrospective detection method using impedance relaxation was successfully applied and validated.

References

[1] 10.1149/2.0621506jes

[2] 10.1016/j.jpowsour.2015.11.044

[3] 10.1016/j.jpowsour.2021.230870

[4] 10.1016/j.jpowsour.2021.230449

[5] 10.1016/j.jpowsour.2021.230009

[6] 10.1016/j.xcrp.2021.100589

[7] 10.1016/j.jpowsour.2021.230508

[8] 10.1016/j.jpowsour.2021.229794

[9] 10.3390/batteries7030046

[10] 10.1016/j.jpowsour.2020.227798