Electrochemical Noise Spectral Signature of Commercial Li-Ion Batteries

Abstract

Electrochemical noise analysis (ENA) is performed on lithium cobalt oxide (LiCoO₂) ICR 18650 commercial lithium-ion batteries. The interest of ENA lies in its potential for passive diagnostics during the charging and discharging of the battery. The electrochemical noise of the batteries is characterized by two methods: power spectral density (PSD) and standard deviation (STD). Good reproducibility of the results has been demonstrated. All power spectral analyses highlight the same shape and robust signature. The obtained spectra show a classical fractional power frequency dependence (1/f ^α) at low frequency (0.1 < f < 1 Hz) and a plateau at very-low frequencies (less than 0.1 Hz). On one hand, the beginning (from SOC 100 to 84%) and end (SOC < 16%) zones are characterized by a PSD slope α ≈ 2. On the other hand, the middle zone (16 ≤ SOC < 84%) is characterized by a PSD slope α < 2. The obtained STD curves have a V-shape form with the minimum value occurring near 38% of SOC, corresponding to minimum slope α = 1.2. Furthermore, the end and beginning zones exhibit significantly higher PSD amplitude (PSD(f → 0)) and STD compared to the middle zone. There are promising results regarding the potential for determining the state of health of the battery, but aging campaigns still need to be carried out.