Analyses of polarisation effects and operando detection of lithium deposition in experimental half- and commercial full-cells

Fast-charging of lithium-ion batteries provokes lithium deposition, a severe degradation mechanism of the graphitic anode that causes a rapid performance decay and may lead to safety-critical conditions. In this study, an extensive analysis of the polarisation behaviour during fast-charging is conducted for both graphite half-cells and commercial full-cells. Anomalies in the impedance and potential signals during charging and subsequent relaxation are shown, providing deeper insights into the cell behaviour in case of lithium deposition. The impedance measurement is conducted at characteristic frequencies to enable a physical interpretation. For this purpose, the cells are characterised beforehand by electrochemical impedance spectroscopy and a distribution of relaxation times analysis. For the comprehensive experiments, potential and impedance measurements are run during charging with varying C-rates to analyse the polarisation effects for critical and uncritical cases. A sensitivity analysis of the excitation frequency is conducted for the half-cell results which identifies the migration of lithium ions through the SEI as the most suitable process for the detection of lithium deposition. The same process is tracked in the full-cell experiments showing the same trends and demonstrating the transferability and reliability of the method. The comparison with state-of-the-art retrospective detection methods reveals a high sensitivity. A lumped electrochemical model is introduced which elucidates the observed phenomena and extends the understanding of cell behaviour during critical charging events.

Automated summary

This study provides a comprehensive analysis of polarization behavior during fast-charging of lithium-ion batteries, revealing insights into lithium deposition and cell behavior. Sensitivity analysis identifies the migration of lithium ions through the SEI as the most suitable process for detection. The results demonstrate high sensitivity and reliability of the proposed method.