Electrochemical impedance spectroscopy is a powerful tool to study electrode/electrolyte interfaces while they evolve, as measurements can be performed quickly and in-situ, i.e., in the battery cell without need for disassembly. To interpret electrochemical impedance spectroscopy (EIS) data is however not trivial, as all the diverse processes like electronic resistances, electrolyte ionic resistance, transport limitations within the electrodes, surface film resistances, and the actual charge transfer resistances of both electrodes contribute simultaneously to the EIS spectra. Recently, we developed a three-electrode Swagelok® type T-cell with a gold wire reference electrode (GWRE), which allows for a reliable deconvolution of anode and cathode impedance [1]. Using this GWRE, we were able to isolate the evolution of electrical contact resistance, pore resistance and charge transfer resistance of a lithium nickel manganese spinel (LNMO) cathode during cycling by impedance measurements at non-blocking and blocking conditions [2]. At the latter, the intercalation reaction is thermodynamically hindered, which increases the charge transfer resistance significantly and leads to a shift in its characteristic frequency, facilitating a separation of other impedance contributions.

In the present study, we apply the concept of blocking conditions to a graphite anode in LFP/graphite cells equipped with a GWRE. To monitor the SEI evolution during formation, we use a potential stepping procedure, which lowers the potential of the graphite anode gradually with every step down to 0.01 V vs. Li/Li⁺ and then back up. After each step, the graphite potential is moved back to blocking conditions (2 V vs. Li/Li⁺), where impedance spectra are acquired. As the charge transfer resistance in blocking conditions appears at very low frequencies, we can unambiguously ascribe an emerging additional resistance in the mid-frequency range to the formation of a growing solid-electrolyte interphase (SEI) resistance. The evolution of SEI resistance as a function of potential is now obtained by fitting the recorded spectra to a transmission-line based equivalent circuit model.

A comparison of the SEI resistance evolution with the reduction currents obtained by cyclic voltammetry (CV) in different electrolytes (pure LP57 or LP57 + 1 wt% vinylene carbonate (VC), fluoroethylene carbonate (FEC) or difluoroethylene carbonate (DiFEC)) confirms that the electrolyte/additive reduction potential coincides with an increase in the SEI resistance (see Figure 1). Lastly, we investigate how the SEI resistance develops during continuous cycling of LFP/graphite cells under different conditions.

References:

1. Solchenbach, D. Pritzl, E. Kong, J. Landesfeind and H. A. Gasteiger, J. Electrochem. Soc., 163, A2265 (2016).
2. Landesfeind, D. Pritzl and H. A. Gasteiger, J. Electrochem. Soc., 164, A1773 (2017).

Acknowledgements:

This work is financially supported by the BASF SE Battery Research Network. J. L. gratefully acknowledges the funding by the BMBF for its financial support under the auspices of the ExZellTUM II project, grant number 03XP0081.

Figure 1: Reduction currents during the 1^st and 2^nd negative-going CV scan (lines, upper panels) and graphite SEI resistance (symbols, lower panels) during successive potential steps of a pristine graphite electrode down to 0 V vs. Li/Li⁺ (arrows pointing to the left) and then back up to higher voltages (arrows pointing to the right) in various electrolytes: i) pure LP57 (left top panels); ii) LP57 + 1 wt% VC (right top panels); iii) LP57 + 1 wt% FEC (left bottom panels); and, iv) LP57 + 1 wt% DiFEC (right bottom panels). Cyclic voltammetry of graphite electrodes was performed in graphite/LFP T-cells (scan rate 0.03 mV/s, graphite capacity ~1.3 mAh/cm², 60 µL electrolyte) equipped with a lithium metal reference electrode. The SEI resistance was determined by fitting a transmission-line based equivalent circuit model to impedance spectra of graphite electrodes in blocking conditions (2 V vs. Li/Li⁺, 100 kHz-1 Hz, 10 mV excitation amplitude) after holding the respective lower potential for 30 min in graphite/LFP T-cells equipped with a GWRE. Data points represent the average of two identical cells. All experiments were performed at 25 °C.