Lithium-Ion Batteries containing a high-voltage positive electrode as LiNi_0.5Mn_1.5O₄ (LNMO) have a high operating voltage and are therefore considered for high energy batteries [1]. However, it’s high potential (4.75 V vs. Li/Li⁺) causes some severe side reactions, especially at elevated temperatures, such as electrolyte oxidation and transition metal dissolution [2]. A further failure mechanism of graphite/LNMO cells is the increase of the cell impedance during cycling [3].

In a previous study of our group [4] we introduced a novel impedance procedure where we measured the LNMO impedance at two distinct states of charge: at 4.4 V cell voltage (10% SOC, referred to as non-blocking conditions) and at 4.9 V cell voltage (100% SOC, referred to as blocking conditions). By fitting the two obtained spectra at the same time, we were able to deconvolute the cathode impedance (R_Cathode) into contributions of the charge transfer resistance (R_CT), the pore resistance (R_Pore) and the contact resistance between current collector and cathode coating (R_Anode). In the present study, we apply the same approach by measuring the anode impedance at 4.4 V cell voltage (10% SOC, non-blocking condition) and 3.0 V cell voltage (0% SOC, blocking condition for the anode)

This was examined by assembling graphite/LNMO or graphite/LFP Swagelok T-cells equipped with a gold-wire reference electrode (GWRE) [5] and using 1 M LiPF₆ in EC/EMC as electrolyte. The cycling procedure was carried out at 40 °C. The GWRE allows the measurement of the half-cell impedances (anode vs. RE and cathode vs. RE) during cycling, so that the impedance evolution of each electrode can be monitored independently.

First, we will show that blocking conditions – a 45° transmission line which represents 1/3 of the pore resistance – can be achieved for the graphite anode cycled versus an LFP cathode. Interestingly this transmission line is also observed for the blocking spectra of a graphite anode cycled versus an LNMO cathode, however upon extended charge/discharge cycling at elevated temperatures (40°C), the transmission line turns into a semi-circle in the impedance spectra (Nyquist representation). This means that after switching the temperature from 25°C (after formation) to 40°C (during charge/discharge cycling), a novel interface resistance evolves, which only appears for graphite anodes cycled versus an LNMO cathode.

We attribute this semi-circle evolving in the blocking spectra to occur at the separator/graphite anode interface: (i) temperature dependent impedance measurements suggest an activation energy of ~20 kJ/mol which lies in the range of typical values for ionic resistances [6] and, (ii) estimating the contributing surface area by the capacitance, a very small surface area of ~ 1cm² is obtained that fits to the interface separator/anode. Lastly, we support this hypothesis by adding defined amounts of a manganese (II)-salt into graphite/LFP cells, which in the absence of manganese show a blocking electrode behavior (transmission line) for the anode impedance upon cycling at 40°C. When the manganese salt is added to cell after formation, a semi-circle evolves in the blocking spectra. This leads us to the conclusion that manganese is reduced at the top of the graphite anode and leads to an enhanced SEI formation, which can be detected by our novel impedance procedure. By fitting the obtained anode impedance spectra at blocking and non-blocking conditions, we can deconvolute the anode impedance (R_Anode) into contributions of a charge transfer resistance (R_CT), the high frequency resistance (R_HFR) and the evolving interface resistance between anode and separator (R_{Contact, ionic}), as shown in Figure 1.

Figure 1: Fitting results of the anode impedance in blocking conditions (3V cell voltage) and non-blocking conditions (4.4V cell voltage). The Impedance is recorded at 40°C from 100 kHz to 0.1 Hz with a perturbation 15 mV.

References

[1] J-H. Kim, N. P. W. Pieczonka and L. Yang, ChemPhysChem. 15, 1940 - 1954, 2014.

[2] D. Lu, M.Xu, L. Zhou, A. Garsuch, and B. L. Lucht, J. Electrochem. Soc. 160 (5), A3138-A3143, 2013.

[3] D. Aurbach, B. Markovsky, Y. Talyossef, G. Salitra, H-K. Kim and S. Choi, J. Power Sources. 162, 780-789, 2006.

[4] J. Landesfeind, D. Pritzl, H. A. Gasteiger, J. Electrochem. Soc. 164 (7), A1773-A1783, 2017

[5] S. Solchenbach, D. Pritzl, E. Kong, J. Landesfeind, and H. A. Gasteiger, J. Electrochem. Soc. 163 (10), A2265-A2272, 2016

[6] N.Ogihara, S. Kawauchi, C. Okuda, Y. Itou, Y. Takeuchi, and Y. Ukyo J. Electrochem. Soc. 159 (7), A1034-A1039, 2012

Acknowledgement

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