Low-Frequency Electrochemical Impedance Spectroscopy to Diagnose the Effect of the Aging Process on Mass Transfer in Cathode Particles

Abstract

Lithium ion batteries are widely used for small scale electronic applications and for large scale applications like electronic vehicles and grid energy storage. Hence, a key requirement to better predict battery performance is the development of advanced analysis techniques. Electrochemical impedance spectroscopy is a powerful tool to monitor the evolution of kinetic and mass transfer effects during cycling, as measurements can be performed without need to dissemble the battery cell. Moreover, measurements performed at very low frequencies (<1 mHz) can provide insightful information on mass transfer occurring at low rates, i.e. solid-state lithium ion diffusion. In the present study, we modelized impedances including finite-length internal diffusion for LiNixMnyCo1-x-yO2 of different ratios in NMC/graphite cells equipped with a gold wire reference electrode. This three-electrode Swagelok® type T-cell developed by Solchenbach and al. [1] allows for a reliable deconvolution of anode and cathode impedance. Thus, the evolution of the lithium ion diffusion coefficient into NMC secondary particles was monitored throughout the first charge and discharge of the battery cell. For this purpose, impedance analyses at frequencies from 200 kHz to 0.1 mHz were acquired at each ~5-10% state-of-charge gained (or lost during discharge). At first, the Warburg element of these spectra displayed a straight line of 45° corresponding to the semi-infinite diffusion. At lower frequencies the spectra subsequently showed a vertical line which corresponds to the diffusion layer thickness being much larger than the sphere radius. This indicated that the diffusion arrived at the very center of the particle and could not proceed any further (no dc current flowing). The electrical equivalent circuit including finite-length internal linear and spherical diffusion was used to fit the obtained spectra [2]. This model was previously applied by Rajabloo and al. [3] to LiFePO4 to develop a semi-empirical performance model and it was here modified to allow the diagnostic of NMC cathode material. This enabled estimation of the lithium ion diffusion coefficients at each state-of-charge. Therefore, we were able to observe the evolution of the solid-state diffusion coefficient during the very first charge/discharge cycle of the battery cell. This evolution was compared to those after a few charge/discharge cycles to determine the aging effect on mass transfer into cathode particles. We then accessed the characteristic behavior of the NMCs studied at various cycling conditions (i.e. cutoff potential and temperature).