Impedance Analysis of the Side Reactions of Li Electrodeposition in a Quaternary Ammonium-Based Ionic Liquid Electrolyte

Abstract

Despite the high energy densities offered by secondary batteries with metallic lithium (Li) anodes, their commercialization is hindered by the problem of dendritic growth of metallic lithium at the anode. Therefore, we aim to address this problem by evaluating the morphology of the deposited lithium, which is affected by the surface film produced by various side reactions in the system. Considering the wide potential windows of quaternary ammonium ionic liquids, we study precipitation morphologies using these ionic liquids as electrolytes. In this study, we examine the side reactions of Li electrodeposition using impedance spectroscopy and transmission electron microscopy. Samples were prepared and observed by gradually decreasing the potential from 2.8 V, the open circuit potential. As a result, it was confirmed that surface films were formed at about 1.5 V (vs. Li reference electrode), even when ionic liquids with a wide potential window were used, although the direct cause of the surface film formation is not known. Importantly, the vertex frequency of the semi-circular arc derived from the surface film in the impedance measurement was 500 Hz, which is faster than the time constant of the semi-circular arc due to the dissolution and precipitation reaction of Li.

1. Introduction

The use of renewable energy is becoming increasingly important for realizing a sustainable society. The research and development of batteries for the effective storage of renewable energy are becoming increasingly active. Batteries for use in airplanes and automobiles are required to have long cruising time, which means that the installed batteries should have high energy density. Batteries using metallic lithium as the anode have been extensively studied because it can provide higher voltage and capacity than other anode materials. However, the problem of Li dendrite growth at the anode during charging has been a barrier to its practical use, since the dendritic growth of lithium metal can cause an internal short circuit and subsequent thermal runaway. Dendritic growth of metallic lithium at the anode during charging must be suppressed for practical commercialization.^1,2

It was reported that the morphology of the deposited lithium is affected by the surface film produced by side reactions, which occurs between the electrolyte and the electrode.² To suppress surface-film formation, the width of the potential window of the electrolyte is thought to be very important. Given their wide potential window,^3,4 quaternary ammonium ionic liquids were used as electrolytes to investigate the morphology of the precipitated lithium.^5–13 However, when metallic lithium was deposited on a nickel foil, a small side reaction was observed at a potential of approximately 1.5 V as an initial process, as seen in Fig. 6a in the referenced literature.¹³ This type of side reaction is usually confirmed during the initial stage of current flow before the electrodeposition of metallic lithium. A possible cause of side reactions is the presence of water or oxygen impurities in the electrolyte due to inadequate cell drying or insufficient cell sealing. It has been reported that the stability of the ionic liquid itself is reduced when trace amounts of water are present. However, the details of such reactions remain unclear.

The side reaction resulting in film formation has been examined by ac impedance measurement^14–19 and Fourier transform infrared spectroscopy measurements,^20–23 as well as their combination with potential step experiments. In these measurements, the experimental system was the same throughout the measurements; thus, the background spectra and noise were very similar for each spectrum, and changes in film formation are obvious and easy to analyze.

In the present study, we examined the side reactions of Li electrodeposition in quaternary ammonium-based ionic liquids by using impedance spectroscopy combined with a potential step experiment. Measurement of impedance while decreasing the potential in steps clearly shows at which potential the surface film formation occurs, and also shows the frequency component of the semicircular arc attributed to the surface film. To the best of our knowledge this analysis method was first introduced by Katayama’s group.^15–17 We also examined the differences in the state of the electrode surface using transmission electron microscopy (TEM). TEM analysis directly confirmed the surface film formation observed indirectly by impedance. This combination of impedance measurement and TEM analysis with stepwise changes in potential can contribute to the discrimination of some semicircular arcs that are difficult to attribute in impedance measurement.

2. Experimental

A quaternary ammonium ionic liquid, Py14[TFSA] (Py14: N-butyl-N-methylpyrrolidinium; TFSA: bis(trifluoromethylsulfonyl)amide, Kanto Chemical), and Li[TFSA] (Kishida Chemical) were mixed to prepare the electrolyte after drying them under vacuum and heating at 85 °C for 3 hours and at 120 °C for 14 hours, respectively. A three-electrode cell was constructed following previous literature,^12,13 using a nickel foil as the working electrode and lithium foil as the counter and reference electrodes. The area of the two electrode was c.a. 2 cm². The potentials of the working electrode were maintained at 2.0, 1.6, 1.3, 1.0, 0.7, 0.4, 0.1, and −0.03 V vs. the Li-reference electrode (Li_ref) for 1 h each by a potential step experiment, and impedance measurements were performed at these potentials, immediately after 1 h equilibrium, in the frequency range of 10⁵ Hz to 10⁰ Hz with an amplitude of 10 mV at 25 °C. Here, during the electrode maintenance at −0.03 V, metallic lithium is thought to electrodeposited on the electrode.

For the TEM observation, a two-electrode cell was constructed following the previous literature,²⁴ with a thin Cu grid as the working electrode and a lithium foil as the counter electrode. The cell voltage was maintained at 2.0, 1.3, and 0.1 V for 1 h each. After 1 h equilibrium, the Cu grids were disassembled, washed with dimethyl carbonate in a dry chamber (dew point (dp) ≤ −50 °C), and transferred to a TEM chamber for observation of the surface film on the Cu grids.

3. Results and Discussion

Figure 1 shows the electrode potential and current applied during the potential step experiment and impedance measurements. The current applied during the impedance measurement was low so that the current flow is not observed in the figure at the shown scale. When the potential was stepped, for example, from 2.0 V to 1.6 V vs. Li_ref, rather large current was initially applied for a short time and then quickly decayed. When the potential was set to −0.03 V, the current was continuously applied; metallic lithium was thought to electrodeposit on the electrode. Evidently, impedance measurement was conducted when the current was sufficiently attenuated and the electrode was stable.

Figure 1.

Ni working electrode potential and current applied during the potential step experiment and impedance measurement.

Figure 2 shows the impedance spectra in the Nyquist plots and bode plot at each potential. For the cases of the open circuit potential and an electrode potential of 2.0 V, the Nyquist plot appears that of a blocking electrode. This result is reasonable because the nickel electrode is blocking for the Li-based electrolyte. When the potential was 1.6 V, a new arc was observed, the size of which remained similar until the electrode potential reached 0.1 V. The vertex frequencies for the arcs are around 500 Hz (see Fig. 2b). The vertex frequency and the size of the arc did not change from 1.6 V to 0.1 V. After electrodeposition at −0.03 V, the first arc remained unchanged while another arc appeared. The vertex frequency of the second arc (with a lower frequency) was 5 Hz (see Fig. 2b). The first arcs was considered to originate from the surface film, which is the solid electrolyte-like interface (SEI), and the second arc was considered to originate from the redox reaction of Li. The surface film was formed at a potential from 2.0 V to 1.6 V during the negative scan.

Figure 2.

Impedance spectra of the Ni electrode in (a) Nyquist plot and (b) bode plot at each potential. The legend shows the potential when the impedance was measured. “−0.03 V” means the measurement at open circuit potential after maintaining the electrode at the −0.03 V of electrode potential for 16.6 h.

Two arcs are commonly observed in the impedance spectra of the metallic lithium electrode–electrolyte interface; in such cases, the potential step analysis method can be used to determine the contribution of the two arcs.^15–17 A side reaction occurred at approximately 1.5 V in the same electrolyte reported in the literature,¹³ which reaction amount increased with increasing temperature, is thought to reductive decomposition reaction of the electrolyte forming surface film.

The relationship between the first arc and film formation was more detailly examined using TEM. Figure 3 shows the TEM images of the Cu grid sample before and after the electrode potential was maintained at 2.0, 1.3, and 0.1 V. For the pristine sample, no film was observed on the Cu surface. At 2.0 V, no obvious film was observed on the Cu surface, although a small amount of electrolyte residue remained. For the sample with the electrode potential maintained at 1.3 and 0.1 V, a 2–5-nm-thick film was observed on the Cu surface. Similar to the impedance measurement results, this indicates that no change occurred during the potential scan between 1.3 V and 1.0 V. These results support the impedance findings indicating that the film formed below 1.6 V.

Figure 3.

TEM images of the Cu grid sample (a) before and after being maintained at (b) 2.0, (c) 1.3, and (d) 0.1 V.

To study the chemical composition of the film, the samples were subjected to mapping using energy dispersive X-ray spectroscopy in scanning transmission electron microscopy (STEM-EDS). Figure 4 shows the STEM-EDS images of the Cu grid sample after the potential was maintained at 1.3 and 0.1 V. No differences were observed between the films of the two samples. In both cases, oxygen is a prominent chemical component of the film. Figure 5 shows the EDS spectra for the film on the Cu grid sample after the sample was maintained at potentials of 1.3 and 0.1 V. Each spectrum corresponds to each square in the images in the inset. No differences were observed between the two samples. In both cases, O was a prominent chemical component of the film, and no other elements were found. Li could not be detected using this method.

Figure 4.

STEM-EDS images for the Cu grid sample after the electrode potential was maintained at (a) 1.3 and (b) 0.1 V.

Figure 5.

EDS spectra for the film on the Cu grid sample after the electrode potential was maintained at (a) 1.3 and (b) 0.1 V. Each spectrum corresponds to each square in the inset images.

To investigate the state of Li in the films of the two samples, STEM combined with electron energy loss spectroscopy (ELLS) was employed. Figure 6 shows the STEM-ELLS Li-K edge spectra of each film on the Cu grid with the potential maintained at 1.3 and 0.1 V. No difference was observed between the two spectra, which were similar to those of Li₂O. The inset shows the overlay composition mapping image of the STEM-ELLS image of Li₂O and the Cu EDS map. Here we will consider the origin of Li₂O. Howlett et al. reported that the surface film consists of an inner compact layer of LiF and an outer more diffuse layer of reduction products of the TFSA anion (i.e., LiS₂O₄, LiF, LiSO₃CF₃, Li₂SO₃, Li₂S, Li₂O, LiOH, etc.).^5,6 Li₂O-derived film observed in this study may come from oxidation of the original surface film when being dismantled in the dry chamber for TEM observation. Considering the absence of organic decompositions, the surface film in this case may have been the compound of S element and Li element. TEM observation without exposure to air would be benefitable for more detailed explanation of the phenomena.

Figure 6.

STEM-ELLS (Li–K edge) spectra of each film on the Cu grid at 1.3 and 0.1 V. The inset shows the overlaid mapping image of STEM-ELLS image of Li₂O and the Cu EDS map. Each spectrum corresponds to each small square in the images in the inset.

4. Conclusion

We evaluated the side reactions occurring during Li electrodeposition in an [Py14][TFSA] + Li[TFSA] ionic liquid electrolyte via impedance measurements and analytical electron microscopy. During a side reaction at 1.5 V versus the Li reference electrode, a surface film with a thickness of approximately 2–5 nm was formed on the electrode. The film was confirmed to be composed mainly of Li₂O by analytical TEM; Li₂O may come from oxidation of inorganic decomposition products during dismantling of the cell for TEM observation. The impedance analysis applied in this study can also be used to analyze other electrolytes; the components of the impedance arc in Nyquist plots can be determined.

CRediT Authorship Contribution Statement

Hikaru Sano: Conceptualization (Lead), Data curation (Lead), Formal analysis (Lead), Investigation (Lead), Methodology (Lead), Validation (Lead), Visualization (Lead), Writing – original draft (Lead), Writing – review & editing (Lead)

Mitsunori Kitta: Data curation (Equal), Formal analysis (Equal), Investigation (Equal), Methodology (Equal), Validation (Equal), Visualization (Equal), Writing – review & editing (Equal)

Keigo Kubota: Investigation (Equal), Methodology (Equal), Validation (Equal), Visualization (Equal)

Conflict of Interest

The authors declare no conflict of interest in the manuscript.

Footnotes

A part of this paper has been presented in the 61st Batter. Symp. Japan in 2020 (Presentation #2C01).

H. Sano and K. Kubota: ECSJ Active Members

References

• 1)  R. R. Chianelli, J. C. Scanlon, and B. M. L. Rao, J. Solid State Chem., 29, 323 (1979).
• 2)  D. Aurbach, E. Zinigrad, Y. Cohen, and H. Teller, Solid State Ionics, 148, 405 (2002).
• 3)  H. Sakaebe and H. Matsumoto, Electrochem. Commun., 5, 594 (2003).
• 4)  H. Matsumoto, Electrochemical Aspects of Ionic Liquids, John Wiley & Sons, Inc., Hoboken, NJ, U.S.A., p. 35 (2005).
• 5)  P. C. Howlett, D. R. MacFarlane, and A. F. Hollenkamp, Electrochem. Solid-State Lett., 7, A97 (2004).
• 6)  P. C. Howlett, N. Brack, A. F. Hollenkamp, M. Forsyth, and D. R. MacFarlane, J. Electrochem. Soc., 153, A595 (2006).
• 7)  H. Sano, H. Sakaebe, and H. Matsumoto, J. Power Sources, 196, 6663 (2011).
• 8)  H. Sano, H. Sakaebe, and H. Matsumoto, J. Electrochem. Soc., 158, A316 (2011).
• 9)  H. Sano, H. Sakaebe, and H. Matsumoto, Electrochemistry, 80, 777 (2012).
• 10)  H. Sano, H. Sakaebe, and H. Matsumoto, Chem. Lett., 42, 77 (2013).
• 11)  H. Sano, H. Sakaebe, H. Senoh, and H. Matsumoto, J. Electrochem. Soc., 161, A1236 (2014).
• 12)  H. Sano, M. Kitta, and H. Matsumoto, J. Electrochem. Soc., 163, D3076 (2016).
• 13)  H. Sano, M. Kitta, M. Shikano, and H. Matsumoto, J. Electrochem. Soc., 166, A2973 (2019).
• 14)  I. Yamada, K. Miyazaki, T. Fukutsuka, Y. Iriyama, T. Abe, and Z. Ogumi, J. Power Sources, 294, 460 (2015).
• 15)  Y. Katayama, H. Kodake, N. Tachikawa, K. Yoshii, and N. Serizawa, 58th Batter. Symp. Japan, Fukuoka, 2B08 (2017).
• 16)  N. Tachikawa, K. Shima, N. Serizawa, and Y. Katayama, 59th Batter. Symp. Japan, Osaka, 3E05 (2018).
• 17)  R. Furuya, N. Serizawa, and Y. Katayama, Electrochemistry, 90, 057002 (2022).
• 18)  S. Kato, N. Serizawa, and Y. Katayama, J. Electrochem. Soc., 169, 076509 (2022).
• 19)  J. Inamoto, T. Fukutsuka, K. Miyazaki, and T. Abe, J. Appl. Electrochem., 47, 1203 (2017).
• 20)  K. Kanamura, T. Umegaki, M. Ohashi, S. Toriyama, S. Shiraishi, and Z. Takehara, Electrochim. Acta, 47, 433 (2001).
• 21)  M. Matsui, K. Dokko, Y. Akita, H. Munakata, and K. Kanamura, J. Power Sources, 210, 60 (2012).
• 22)  K. Kanamura, S. Toriyama, S. Shiraishi, M. Ohashi, and Z. I. Takehara, J. Electroanal. Chem., 419, 77 (1996).
• 23)  Y. Akita, M. Segawa, H. Munakata, and K. Kanamura, J. Power Sources, 239, 175 (2013).
• 24)  M. Kitta and H. Sano, Microscopy, 69, 227 (2020).