Early Detection of the Degradation at Li-Solid Electrolyte Interface During Stripping with Galvanostatic Electrochemical Impedance Measurements

Using lithium metal anode in solid-state batteries enables high energy density due to the high specific capacity of the lithium metal anode. However, Li|solid electrolyte (SE) interface suffers from instability during cycling, especially during stripping due to void formation at the interface. The accumulation of voids at Li|SE interface increases cell polarization and causes current focusing, which leads to poor efficiencies and potential cell failure. To develop preemptive strategies to mitigate the interfacial degradation, it is necessary to develop in-operando early detection strategies to identify the interfacial degradation based on easily characterizable descriptors. Although there have been many previous reports on the characterization of interfacial degradation at the Li|SE interfaces, including in-operando approaches, it is unclear whether those approaches could be integrated into battery cycling in commercial products or in industrial settings. Most of the previous approaches depend on data from sophisticated experimental apparatus (electron microscopy, optical microscopy) and analysis by human experts, which take long time and are difficult to scale up.
In this work, we developed a strategy based on galvanostatic electrochemical impedance measurements to determine the onset of instability at Li|SE interface during stripping. We demonstrated that using impedance values corresponding to a single frequency is enough to track interfacial degradation through stripping and develop a model detecting the instability conditions. The model based on impedance measurements enabled earlier detection of instability compared to the reference model based on cell polarization increase. Using a single frequency greatly reduces the system requirements of the characterization setup compared to the conventional electrochemical impedance spectroscopy (EIS) measurements done with a wide range of frequencies (MHz–mHz), which makes the strategy more applicable to commercial products. Moreover, our strategy only uses raw impedance data without fitting the data with equivalent circuit models, which makes the work easily applicable to autonomous systems.
The strategy developed in this work could be easily integrated into the current battery cycling setup, as it only requires an additional AC power supply with a single frequency. The additional energy required for the measurement is minimal since the characterization of degradation only requires a couple of seconds. Moreover, the strategy in this work is well suited to be integrated into the control setup to optimize the performance of solid-state batteries by minimizing interfacial degradation, since the analysis could be fully automated without the need for human intervention.