Graphite/LiNi_0.5Mn_1.5O₄ full cells show severe capacity fading, especially at elevated temperatures, frequently related to: (i) electrolyte oxidation, (ii) transition metal dissolution, and (iii) damage of the Solid Electrolyte Interphase (SEI) on the graphite anode [1]. All these effects can lead to an increase in the full-cell impedance. However, to understand the degradation mechanism of high-voltage Li-Ion batteries, anode and cathode impedances have to be investigated separately. To deconvolute the contribution of anode and cathode from the full-cell impedance, micro-reference electrodes can be used [2, 3]. This type of reference electrodes allows to record the impedance of anode and cathode during cycling without the need of full-cell disassembly to determine anode and cathode impedance in symmetric cells [4].

Commonly the semi-circles in a Nyquist plot are interpreted as combination of a charge transfer resistance (R_CT) and a surface film resistance (R_SF), e.g., the solid electrolyte interphase (SEI) on graphite anodes [5]. However, the contributions of the ionic resistance (R_Ionic) in the pores of the electrode and also the contact resistance (R_Contact) between the active material/current collector interface are often neglected, which can lead to misinterpretations in the impedance analysis [6]. In this study, we use a novel impedance approach utilizing a gold-wire reference electrode (GWRE) [2], which allows for the recording the half-cell impedances in its deconvolution into contributions from: (i) contact resistance (R_Contact), (ii) ionic resistance in the pores (R_Ionic), and (iii) charge transfer resistance (R_CT). This is illustrated for a cathode without surface film formation.

Therefore, graphite/LiNi_0.5Mn_1.5O₄ (LNMO) cells are assembled, which are equipped with a gold-wire reference electrode (GWRE) [2] and contain 1 M LiPF₆ in EC/EMC (3/7 by volume) as electrolyte. The impedance is measured at two distinct points during cycling: (i) at 50% state of charge (SOC) and (ii) at 100% SOC. Both impedance spectra are analyzed using Matlab employing a transmission line model.

Figure 1 exemplarily presents the results of the impedance analysis for more than 80 cycles of a LNMO electrode in a full-cell. We can show that the R_CT has only minor contribution to the half-cell impedance of the LNMO cathode and that R_Contact and R_Ionic are the dominating resistances, which are slightly increasing during cycling. We explain the higher ionic resistance in the pores with electrolyte oxidation products, causing pore clogging.

Furthermore, we will show results for the negative electrode (Graphite). Therefore, the anode is analyzed analogously and these results are compared with Graphite/LFP cells, where the upper cut-off potential is lower by 0.8 V (compared to a graphite/LNMO cells). Hence, electrolyte oxidation products should not affect cathode or anode.

Figure 1: Fitted resistances from the Nyquist plots at 50% and 100% SOC. The Impedance is recorded at 40°C from 100 kHz to 1 Hz with a perturbation of 15 mV. The Loading of the anode was 6.6 mg_Graphite/cm² (0.95 cm² area) and 13 mg_LNMO/cm² (0.95 cm² area) for the cathode. A Swagelok T-cell with 60 ml 1M LiPF₆ in EC/EMC (3/7 by volume) and two glass fiber separators with a gold-wire reference electrode (GWRE) was used.

References:

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

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

[3] D. P. Abraham, S. D. Poppen, A. N. Jansen, J.Liu, and D. W. Dees, Electrochim. Acta. 49, 4763-4775, 2004.

[4] R. Petibon, C. P. Aiken.N. N. Sinha, J. C. Burns, H. Ye, C. M. VanElzen, G. Jain, S. Trussler, and J. R. Dahn, J. Electrochem. Soc. 160(1), A117-A124, 2013.

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

[6] M. Gaberscek, J. Moskon, B. Erjavec, R. Dominko and J. Jamnik, Electrochem. Solid. St., 11 (10), A170-A174, 2008.

Acknowledgements:

This work is financially supported by the BASF SE Battery Research Network. Sophie Solchenbach and Morten Wetjen (both TUM) are acknowledged for helpful discussions.