- Email: [email protected]
The effects of moisture contamination in the Li-O2 battery
Accepted Manuscript The Effects of Moisture Contamination in the Li-O2 Battery M.H. Cho, J. Trottie, C. Gagnon, P. Hovington, D. Clément, A. Vijh, C.-S. Kim, A. Guerfi, R. Black, L. Nazar, K. Zaghib PII:
S0378-7753(14)00851-9
DOI:
10.1016/j.jpowsour.2014.05.148
Reference:
POWER 19233
To appear in:
Journal of Power Sources
Received Date: 25 March 2014 Revised Date:
30 May 2014
Accepted Date: 31 May 2014
Please cite this article as: M.H. Cho, J. Trottie, C. Gagnon, P. Hovington, D. Clément, A. Vijh, C.-S. Kim, A. Guerfi, R. Black, L. Nazar, K. Zaghib, The Effects of Moisture Contamination in the Li-O2 Battery, Journal of Power Sources (2014), doi: 10.1016/j.jpowsour.2014.05.148. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
The Effects of Moisture Contamination in the Li-O2 Battery
RI PT
M.H.Cho1, J.Trottie1, C.Gagnon1, P.Hovington1, D.Clément1, A.Vijh1, C.-S.Kim1, A.Guerfi1, R. Black2, L. Nazar2 and K.Zaghib1*
1
Institut de Recherche d’Hydro-Québec (IREQ), 1800 boul. Lionel Boulet, Varennes,
Department of Chemistry, University of Waterloo, 200 University Avenue West,
Waterloo, Ontario, Canada N2L 3G1
Abstract
M AN U
2
SC
Québec, Canada J3X 1S1
The effects of moisture contamination in the Li-O2 battery system were investigated by
TE D
comparing the electrochemical performance, results from electrochemical impedance spectroscopy and post-mortem analysis of batteries prepared under different atmospheres: sealed container in an ambient atmosphere vs. sealed container in a dry-room or in a
EP
glove-box. The performance of the cells strongly depended on the atmosphere;
AC C
furthermore it was found that the performance degradation of the Li metal anode comes from moisture contamination from the feed lines. In sealed container in an ambient atmosphere, the cells showed higher 1st discharge capacity, higher impedance and significant increase of the cell weight owing to contamination of the oxygen by moisture. Under the same atmosphere, the discharge capacity and the impedance increase were significantly different for different cell connection sequences: cells closer to the gas outlet showed higher discharge capacity, and more pronounced increase in impedance
1
ACCEPTED MANUSCRIPT
and weight. The cells discharged in the dry room or glove box showed less discharge capacity than the cells discharged in ambient atmosphere, but the impedance exhibited negligible increase. Post-mortem analysis revealed that the deterioration of lithium metal
RI PT
anode leads to the cell failure mechanism, and this comes from the moisture contamination. This finding supports the idea that protection of the lithium metal electrode is indispensable to realize the practical application Li-O2 or Li-air batteries. It
SC
was found that the performance of Li-O2 batteries is very sensitive to even traces of moisture contamination and every single part of the cell design including the choice of
M AN U
fitting parts and water permeability of the fitting material should be verified in order to obtain credible and reproducible results.
Keyword: Li-O2 battery; O2 flow path; lithium metal; cycle life; moisture contamination
TE D
*Corresponding author.
AC C
EP
E-mail address: [email protected]
2
ACCEPTED MANUSCRIPT
Introduction The Li-air battery is considered one of the most desired technologies for electric vehicles because of its high theoretical specific energy (more than 3500 Wh kg-1,
RI PT
including oxygen), which could enable a longer driving distance of > 300 miles [1-3]. However, there are many technical challenges that hinder the practical use of lithium-air batteries. To address these technical challenges, significant research has been dedicated
SC
in the last 5 years on investigations of electrolytes [4-7], electrode materials and structures [8-9], catalysts [10-13], binders [14] and anodes [15-16]. Despite the
M AN U
significant progress in understanding the mechanisms associated with the electrode reactions, the current lithium-air technology is very far from the requirements for use in electric vehicles, especially with respect to the cycle life and the power capability. Recently, some research groups have reported improvements extending to a few hundred
TE D
cycles using a capacity or time controlled protocol [10, 17], but the life is still limited to less than tens of cycles in the full operating voltage range (2.0 ~ 4.3V). In this work, we examine the effects of moisture contamination on the Li-O2 battery.
EP
Although it is appreciated that the performance is profoundly sensitive to water, this study allows us to quantify the effects. It is found that the capacity and cell weight is
AC C
significantly increased by even traces of H2O contamination, leading to deterioration of the lithium metal (which serves as an indicator) and failure of the cell. It is also suggested that the seal-tightness should be well verified in all cell studies.
3
ACCEPTED MANUSCRIPT
2. Experimental 2.1. Electrode preparation and cell assembly Carbon black (Super-P, TIMCAL) was coated on carbon paper (TGP-H-030, Toray)
RI PT
and dried under vacuum at 100°C for 24 hrs. The loading of carbon coating was controlled to ca. 0.4 mg cm-2 after drying. Commercially available coin-type cells (Hohsen, punched-CR2032) were assembled in a He filled glove box. Cathode electrodes
SC
with a 16 mm diameter were used. Lithium foil (Hohsen, 200 µm) was used as anode, and 3 layers of porous polypropylene (PP, Celgard 3501)/glass filter (GF/C,
M AN U
Whatman)/PP (Celgard 3501) separators were used to facilitate the recovery of the electrode after disassembly of the cells. The power performance was not deteriorated by the addition of two PP layers at this low current rate of 50 mA g-1carbon. The electrolyte consisted of about 80 mg of LiCF3SO3 (lithium triflate, 99.995%, Sigma-Aldrich):
TE D
TEGDME (tetraethylene glycol dimethyl ether, ≥99%, Sigma-Aldrich) by 1:4 mole ratios, which was injected into the cell after drying to have the moisture content of < 20 ppm H2O. The cell weight was monitored during electrolyte injection and the subsequent
AC C
EP
electrochemical tests.
2.2. Electrochemical tests and characterization The coin cells were placed in the container (Bernardin, Mason Jar 250 ml with Standard 70 mm snap lid 1-PC) that has two fitting lines for gas inlet and gas outlet (Kimble Chase Life Science, Stopcock Nylon 3-way PK/10). Sealing resin (Torr Seal, Varian, Inc.) was applied on the joint parts to enhance the seal-tightness. High purity
4
ACCEPTED MANUSCRIPT
oxygen (>99.995%) was used at a flow rate of about 50 mL min-1. Up to six cell containers were connected in series using silicone tubes (Cole-Parmer Instrument co., Silicone Tubing (Platinum) L/S® 16) to flow the oxygen. The schematic of our
RI PT
experiment is shown in Fig. 1. Before testing, the cells were stabilized at open circuit for 10 hrs with flowing O2. Electrochemical tests were performed at room temperature using a VMP3 cycler (Bio-Logic SAS). The cycle life was monitored at a constant current of
SC
50 mA g-1carbon with voltage cut-off between 2.0 V and 4.3 V. The electrochemical impedance spectra were obtained in the frequency range from 200 kHz to 20 mHz with
M AN U
an amplitude of 5 mV. SEM observation was performed using a Hitachi S-4700 field emission gun scanning electron microscope (FEG-SEM). The Li sample was prepared in an Ar filled glove box and transferred to the FEG-SEM using a proprietary sample exchange holder that prevents air contamination between the glove box and the FEG-
TE D
SEM. XRD was performed using a Rigaku table top X-ray diffractometer running at 600 W (X-ray tube) using Co Kα radiation (1.79Å). The XRD sample was prepared in an Ar filled glove box and analysed in an air-tight sample holder. The moisture content in the
EP
cell containers and the atmosphere was measured by a dew point sensor (GE, M series
AC C
sensor/Moisture Target Series 6).
3. Results and discussion Fig. 2 shows the discharge profile of the 1st cycle according to the cell positions at a constant current for the cells sealed in the containers under flowing O2 as shown in Fig. 1. Some voltage disturbance of the 2nd cell in Fig. 2 is perhaps owing to a bad contact within the cell. The dew point of the first container was -25°C (ca. 500 ppm H2O) and that of the
5
ACCEPTED MANUSCRIPT
6th container was -15°C (ca. 1500 ppm H2O) under the environmental condition in the range of 0 ~ 20°C dew point. It is evident in Fig. 2 that cell-to-cell deviation is more than 100% and there is a trend with the order of O2 flow through the cells. The cells closer to
RI PT
the O2 outlet show higher discharge capacity than those closer to the O2 inlet-the 6th cell has 139% higher discharge capacity than the 1st cell. The inset of Fig. 2 shows that the ohmic resistance (Rs), contributed by the electrolyte, increases with the sequence of the
SC
cell connection and higher Rs values were found with the cell at the oxygen outlet. The weight change also shows the same trend-higher weight increase of the 6th cell (22 mg)
M AN U
cannot be explained by the oxygen uptake (3.3 mg) assuming all the discharge capacity comes from the formation of Li2O2. Similar results were reported regarding the increase of discharge capacity by some previous investigators [20-22]. In the gas containing 10% CO2 in O2, the 1st discharge capacity was about 60% higher than that of pure O2 in
TE D
dimethoxyethane (DME) [20]. In the other report, the 1st discharge capacity with 10% CO2 was about 100% higher than that with pure O2, and the capacity increased to 189% with 50% CO2 in ethylene carbonate (EC) and diethyl carbonate (DEC) [21]. Meini et al.
EP
also reported that water had an effect on the discharge capacity [22]. They found that the cells exposed to water vapor and the cells that had a small leak to the ambient atmosphere
AC C
showed higher discharge capacity than sealed Li-O2 cells. Clearly, the accumulation of parasitic reactions is the reason for the increased discharge capacity according to the cell connection sequence.
Fig. 3 shows the discharge/charge profiles and the relevant impedance change observed at the end of the discharge and charge states for the first 2 cycles. The cells all showed significantly different impedance behavior arising from the cell connection
6
ACCEPTED MANUSCRIPT
sequences. In the 1st cell: Rs remains constantly while [Ri + Re] gradually increased with cycles, where Ri is the impedance contribution from the interface film and Re arises from charge transfer reactions. On the other hand, in the 2nd and 3rd cells, Rs significantly
RI PT
increased with cycling. Especially in the 3rd cell, both Rs and [Ri + Re] increase rapidly with cycling. This result explicitly indicates the increase of contaminant accumulation in subsequent cells with inducing the dry-out of electrolyte. The increase of cell weight is
SC
more significant with cycling than the initial values in Fig. 2. After 10 cycles, the weight increases by -3.1mg, 11.6mg and 46.8mg for the 1st cell, the 2nd cell and the 3rd cell
M AN U
respectively; the decreased weight of the 1st cell is probably due to the electrolyte evaporation surpassing the accumulated weight of by-products. The trend of weight increase indicates that the parasitic reactions are prolonged during the cycles. A more detailed impedance study was carried out on the 6th cell due to its higher
TE D
content of contaminants because of its place in the sequence. The evolution of the in-situ impedance spectra as a function of the depth of discharge (DOD) is shown in Fig. 4. The change of impedance spectra can be separated into 3 different zones categorized by the
EP
following DOD regions; 0~20%, 20~80% and 80~100%. In zone-A (Fig. 4b), [Ri + Re] increases without a change of the Rs, indicating that the electrolytes are relatively stable
AC C
in the beginning of discharge and the reformation of passivation film leads the change of impedance in this zone. The similar phenomenon was reported by Meini et al. [22]; The [Ri + Re] increase by the formation of LiOH layer on the surface of lithium metal anode in the electrolyte containing 1000 ppm water. Zone-B (Fig. 4c) shows a significant increase of Rs with a negligible increase of [Ri + Re], which implies that the electrolyte decomposition is accelerated by the Li2O2 generation in the cathode side. Zone-C (Fig.
7
ACCEPTED MANUSCRIPT
4d) shows the increase of Rs and [Ri + Re]. The increase of Rs is the extension of the electrolyte depletion in Zone-B, while the increase of [Ri + Re] represent the deterioration of electrode structure caused by the accumulation of the reaction products
RI PT
(Li2O2, LiOH, Li2CO3) on the electrodes.
In order to investigate the causes behind the trend of cell failure, post-mortem analysis was carried out on the six cells after the 1st discharge. Fig. 5 shows the optical
SC
photographs of the cell components (a) and lithium anodes (b). It is clearly seen that the Li surface is black in the 1st cell and the surface becomes covered with a thick white film
M AN U
in the 6th cell. The white film is believed to be responsible for the significant increase of the cell weight and the cell resistance. Assary et al. [23] reported the effect of oxygen crossover on the anode of a Li-O2 battery and showed the similar phenomena on the lithium metal. However, it cannot explain the trend according to the cell connection
TE D
sequences since all the cells are exposed to the same conditions of O2-crossover. Fig. 6 shows the SEM images of the cross-section of lithium metal anode after the 1st discharge. The thickness of the metallic lithium layer (compact black layer indicated by
EP
arrows) decreases with the cell connection sequences, and the Li metal surface becomes covered by porous and rough layers (which occupy more volume), as the cell is far down
AC C
the connection sequence. The pores in the anode absorb the electrolyte from the separator layer and are responsible for the severe increase of Rs in Fig. 4. To identify the surface film, XRD analysis was performed on the surface of this layer. Fig. 7 shows that LiOH is the main component that grows with the descending cell connection sequences, which implies that the moisture contamination is leading the parasitic reactions in this experimental condition.
8
ACCEPTED MANUSCRIPT
In this system, two kinds of parasitic reactions can be involved in the cell; the electrochemical reactions that can take place on the cathode electrode during the discharge, which contribute to the discharge capacity, and the chemical reactions that
RI PT
doesn’t contribute to the cell capacity but deteriorate the performance of lithium anode.
2H2O + 2Li+ + 2e- 2LiOH + H2
(1)
(2)
M AN U
2H2O + Li2O2 + 2Li+ + 2e- 4LiOH
SC
The additional capacity caused by H2O can be described as below reactions [25];
The weight increase based on the above reactions is about 1.3 mg (calculated from the excess discharge capacity), which cannot explain the excessive increase of cell weight of 22 mg of the 6th cell in Fig.2. This indicates that additional chemical reactions are
TE D
involved in the cells, leading to the increase of cell weight and the serious damage on the cell performance. As evidenced by the analysis data on the lithium metal anode in Fig. 6,
EP
the parasitic reaction caused by H2O is dominated by the following reaction [26];
(4)
AC C
2H2O + 2Li 2LiOH + H2
The amount of lithium to yield the weight increase of 22 mg corresponds to 35% of lithium anode (200 µm thickness, 2 cm2 dimension) based on the above reaction, which is reasonably in line with the result in Fig. 6. To verify this observation, we set up the test cells in the dry-room which was controlled under -40°C dew point. Fig. 8 represents the 1st discharge profile according to
9
ACCEPTED MANUSCRIPT
the cell positions, which is completely different from the previous results (Fig. 2). The weight of the cell and Rs impedance were found quite stable, while the discharge capacity does not show the correlation with the cell connection sequences.
RI PT
When the cells are cycled in the glove box, where the moisture is controlled under 90°C dew point, the 1st discharge profile is shown in Fig. 9. Unlike the previous tests where all the cell containers are connected in series with continuous flowing O2, the cells
SC
are individually isolated and tested in the closed system after filling O2. Under the O2 flow condition (50 ml min-1), the dew point in the containers was about -65°C, which
M AN U
means that ~5 ppm H2O is originated from the O2 feed line. The figure shows that the 1st discharge capacity is less than the capacity in the dry room, and without an increase of the cell weight and Rs. This result demonstrates that the moisture from the atmosphere and O2 feeding line is responsible for the significant increase of the cell weight, cell
TE D
resistance and 1st discharge capacity in Fig. 2.
Fig. 10 shows the voltage profiles for the 10 cycles with different current rate in the He filled glove-box. The cycle life is improved but still too short to be useful for practical
EP
applications. The impedance measured for 10 cycles are shown in Fig. 11. The plots measured at the end of each discharge (a) and at the end of each charge (b) are totally
AC C
different from those previous obtained under ambient conditions without moisture control (Fig.3-b, d and f). Even when the capacity retention is less than 20% after 10 cycles, the impedance profiles are well maintained in the frequency range between 200 kHz and 20 mHz. The lower frequency behavior is currently under investigation to further understand this phenomenon. To analyze the different behavior according to the cell connection environment, we did the post-mortem analysis again, and the optical photographs of the
10
ACCEPTED MANUSCRIPT
lithium surface are showed in Fig. 12. In both cases, after the 1st discharge in the dryroom (a) and after 10 cycles in the glove-box (b), the Li metal anodes remained shiny
RI PT
without surface changes during the test period of 100 hrs.
4. Conclusions
In this study, we examined lithium-air cells under different experimental conditions
SC
with respect to moisture contamination. We performed electrochemical measurements, electrochemical impedance spectroscopy and post-mortem analysis to understand the
M AN U
effects of moisture contamination by investigations of a sealed container in an ambient atmosphere vs. sealed container in a dry-room or in a glove-box. The cells tested in the sealed container in ambient atmosphere showed higher discharge capacity, higher impedance and significant weight gain after the 1st discharge to 2.0 V under O2 flow. On
TE D
the other hand, the cells tested in the dry-room or glove-box showed less discharge capacity, negligible impedance change and insignificant cell weight change. It is believed that the contaminated gas causes parasitic reactions in the downstream neighboring cells.
EP
Post-mortem analysis revealed that the deterioration of lithium metal anode leads to the cell failure, which is directly related to the moisture contamination. This finding supports
AC C
the idea that protection of the lithium metal electrode is crucial to realize the practical applications of Li-O2 or Li-air batteries. We confirmed that the performance of Li-O2 batteries is very sensitive to even trace amounts of moisture contamination. Furthermore, we found that every single part of cell design including the choice of fitting parts and water permeability of fitting material should be verified to obtain credible results.
11
ACCEPTED MANUSCRIPT
Reference [1] J. Hou, M. Yang, M. W. Ellis, R. B. Moore, B. Yi, Phys. Chem. Chem. Phys., 14 (2012) 13487-13501.
RI PT
[2] J. Christensen, P. Albertus, R. S. Sanchez-Carrera, T. Lohmann, B. Kozinsky, R. Liedtke, J. Ahmed, A. Kojic, J. Electrochem. Soc., 159 (2012) R1-R30.
[3] P. G. Bruce, S. A. Freunberger, L. J. Hardwick, J. M. Tarascon, Nature Mat., 11
SC
(2012) 19-29.
[4] W. Xu, J. Hu, M. H. Engelhard, S. A. Towne, J. S. Hardy, J. Xiao, J. Feng, M. Y. Hu,
M AN U
J. Zhang, F. Ding, M. E. Gross, J. G. Zhang, J. Power Sources, 215 (2012) 240-247. [5] B. D. McCloskey, D. S. Bethune, R. M. Shelby, T. Mori, R. Scheffler, A. Speidel, M. Sherwood, A. C. Luntz, J. Phys. Chem. Lett., 3 (2012) 3043-3047. [6] Y. Chen, S. A. Freunberger, Z. Peng, F. Bardé, P. G. Bruce, J. Amer. Chem. Soc., 134
TE D
(2012) 7952-7957.
[7] M. J. Trahan, S. Mukerjee, E. J. Plichta, M. A. Hendrickson, K. M. Abraham, J. Electrochem. Soc., 160 (2013) A259-A267.
EP
[8] Y. Cui, Z. Wen, X. Liang, Y. Lu, J. Jin, M. Wu, X. Wu, Energy Environ. Sci., 5 (2012) 7893-7897.
AC C
[9] H. D. Lim, K. Y. Park, H. Song, E. Y. Jang, H. Gwon, J. Kim, Y. H. Kim, M. D. Lima, R. O. Robles, X. Lepró, R. H. Baughman, K. Kang, Adv. Mat., (2013) 1-5.
[10] J. J. Xu, D. Xu, Z. L. Wang, H. G. Wang, L. L. Zhang, X. B. Zhang, Angew. Chem. Int. Ed., 52 (2013) 1-5.
[11] B. D. McCloskey, R. Scheffler A. Speidel, D. S. Bethune, R. M. Shelby, A. C. Luntz, J. Am. Chem. Soc., 133 (2011) 18038-18041.
12
ACCEPTED MANUSCRIPT
[12] H. G. Jung, Y. S. Jeong, J. B. Park, Y. K. Sun, B. Scrosati, Y. J. Lee, ACS NANO, 7 (2013) 3532-3539. [13] R. Black, J. H. Lee, B. Adams, C. A. Mims, L. F. Nazar, Angew. Chem. Int. Ed., 52
RI PT
(2013) 392-396.
[14] Y. Cui, Z. Wen, Y. Lu, M. Wu, X. Liang, J. Jin, J. Power Sources, 244 (2013) 614619.
SC
[15] A. Zhamu, G. Chen, C. Liu, D. Neff, Q. Fang, Z. Yu, W. Xiong, Y. Wang, X. Wang. B. Z. Jang, Energy Environ. Sci., 5 (2012) 5701-5707.
M AN U
[16] I. C. Jang, Y. Hidaka, T. Ishihara, J. Power Sources, 244 (2013) 606-609. [17] H. G. Jung, J. Hassoun, J. B. Park, Y. K. Sun, B. Scrosati, Nature Chemistry, 4 (2012) 579-585.
[18] Z. Peng. S. A. Freunberger. Y. Chen, P. G. Bruce, Science, 337 (2012) 563-566.
TE D
[19] H. K. Lim, H. D. Lim, K. Y. Park, D. H. Seo, H. Gwon, J. Hong, W. A. Goddard III, H. Kim, K. Kang, J. Am. Chem. Soc., 135 (2012) 9733-9742. [20] S. R. Gowda, A. Brunet, G. M. Wallraff, B. D. McCloskey, J. Phys. Chem. Lett., 4
EP
(2013) 276-279.
[21] K. Takechi, T. Shiga, T. Asaoka, Chem. Commun., 47 (2011) 3463-3465.
AC C
[22] S. Meini, M. Piana, K. Tsiouvaras, A. Garsuch, H. A. Gasteiger, Electrochem. SolidState Lett., 15 (2012) A45-A48.
[23] R. S. Assary, J. Lu, P. Du, X. Luo, X. Zhang, Y. Ren, L. A. Curtiss, K. Amine, ChemSusChem, 6 (2013) 51-55.
[24] M. M. O. Thotiyl, S. A. Freunberger, Z. Peng, P. G. Bruce, J. Am. Chem. Soc., 135 (2013) 494-500.
13
ACCEPTED MANUSCRIPT
[25] D. Pletcher, J. F. Rohan, A. G. Ritchie, Electrochim. Acta, 39 (1994) 1369-1376.
AC C
EP
TE D
M AN U
SC
RI PT
[26] K. Chen, Z. Zhang, E. Ni, J. Appl. Electrochem. 40 (2010) 197-204.
14
ACCEPTED MANUSCRIPT
Figure captions
Figure 1. Schematic diagram of O2 flow and cell connections. O2 line was serially
cells were independently controlled by a cycler.
RI PT
connected from the 1st cell to the 6th cell through a single pathway while the individual
SC
Figure 2. The 1st discharge profiles according to the cell positions at a constant current of 50 mA g-1carbon between 2.0 ~ 4.3 V in sealed containers in ambient atmosphere. The
M AN U
impedance profiles, the discharge capacity and the cell weight change after discharge for each cell are displayed in the insets, which are significantly influenced by the connection order of the cells.
TE D
Figure 3. Cycle behavior (a, c, e) and the relevant change of impedance spectra (b, d, f) according to the cell positions. (a) and (b) : the 1st cell, (c) and (d): the 2nd cell, (e) and
EP
(f) : the 3rd cell.
Figure 4. (a) The first discharge profile of the 6th cell indicating the points of
AC C
electrochemical impedance measurement. The pattern of impedance evolution can be categorized by (b) zone A showing the increase of Ri + Re without change of Rs, (c) zone B showing the increase of Rs with a negligible increase of Ri + Re, and (d) zone C showing the increase of Rs with a significant increase of Ri + Re (Rs: impedance from the ion conductivity in electrolyte, Ri: impedance from the interface film, Re: impedance from the charge transfer reactions).
15
ACCEPTED MANUSCRIPT
Figure 5. Photographs of (a) disassembled cells and (b) lithium metal anodes after the 1st discharge. The Li metal surface turns black in the 1st cell and then turns to a thick white
RI PT
film in the 6th cell.
Figure 6. SEM images of (a) ~ (c) the cross section of lithium metal anode after the 1st
SC
discharge and (d) the surface of lithium metal anode from the 6th cell. The arrows highlight the different thickness dimensions of the metallic lithium layer while the rough
M AN U
and porous surface layer grows according to the cell connection order.
Figure 7. XRD patterns of lithium metal anode after the 1st discharge. (*) LiOH, (■)
TE D
lithium metal, (●) austenite from sample holder.
Figure 8. The 1st discharge profiles according to the cell positions at a constant current of 50 mA g-1carbon between 2.0 ~ 4.3 V. Experiments conducted in the dry-room which was
EP
controlled to less than -40°C dew point. The impedance profiles, the discharge capacity
AC C
and the cell weight change after 1st discharge for each cell are in the inset.
Figure 9. The 1st discharge profiles according to the cell positions at a constant current of 50 mA g-1carbon between 2.0 ~ 4.3 V. Experiments conducted in the glove-box filled with helium which was controlled to less than -90°C dew point. The table lists the 1st discharge capacity and the weight change after the 1st discharge. The impedance profiles
16
ACCEPTED MANUSCRIPT
for each cell, the discharge capacity and the cell weight change after 1st discharge are displayed in the inset.
RI PT
Figure 10. Voltage profiles between 2.0 ~ 4.3 V for 10 cycles at current rates of (a) 50 mA g-1carbon and (b) 100 mA g-1carbon in the oxygen-filled container in the Helium-filled
SC
glove-box which are controlled to less than -90°C dew point.
M AN U
Figure 11. Nyquist plots after (a) discharge and (b) charge during 10 cycles.
Figure 12. Photographs of disassembled cells after (a) the 1st discharge in the dry-room and (b) the 10 cycles in the glove-box. All the lithium metal anodes remain shiny unlike
AC C
EP
TE D
the lithium metal anode discharged in the sealed container in the ambient atmosphere.
17
ACCEPTED MANUSCRIPT
Silica gel
2nd Cell
3rd Cell
4th Cell
O2 (99.995%)
6th Cell
M AN U
SC
Cycler
5th Cell
RI PT
1st Cell
(Oil)
Figure 1. Schematic diagram of O2 flow and cell connections. O2 line was serially connected from the 1st cell to the 6th cell through a single pathway while the individual cells were
AC C
EP
TE D
independently controlled by a cycler.
ACCEPTED MANUSCRIPT
L-1st
3.0
L-2nd
L-3rd
L-4th
1500
Sample L-1st L-2nd L-3rd L-4th L-5th L-6th
1.0
0.5
∆Weight (g) 1 7 4 16 23 22
1200
SC
Capacity (mAh g -1 carbon ) 1905 2790 2550 3988 3452 4552
-Z'' (ohm)
1.5
L-6th
RI PT
2.0
900
M AN U
Voltage (V)
2.5
L-5th
600
300
0
0
0.0 500
1000
1500
2000
TE D
0
2500
300
3000
Capacity (mAh
600 900 Z' (ohm)
3500
4000
1200
4500
1500
5000
g-1carbon)
Figure 2. The 1st discharge profiles according to the cell positions at a constant current of 50 mA
EP
g-1carbon between 2.0 ~ 4.3 V in sealed containers in ambient atmosphere. The impedance profiles, the discharge capacity and the cell weight change after discharge for each cell are displayed in
AC C
the insets, which are significantly influenced by the connection order of the cells.
ACCEPTED MANUSCRIPT
4.5
400
(a)
(b) C1
4.0
D0
D1
C1
D2
3.0
D0
200
RI PT
3.5
-Z'' (ohm)
Potential (V)
300
2.5 100 2.0
D1
D2
-3.1 mg
1st Cell
0
1.5 0
50
100
150
200
250
0
300
100
200
300
400
Z' (ohm)
SC
Time (h) 4.5
400
(c)
(d)
4.0
D0
D1
C1
D2
M AN U
C1
3.0
-Z'' (ohm)
Potential (V)
300
3.5 D0
2.5
200
100
2.0
D1
D2
11.6 mg
2nd Cell 1.5
0
50
100
150
200
250
300
TE D
0
0
100
Time (h) 4.5
(e)
300
(f)
D0 C1
3.0 D0
AC C
2.5 2.0
46.8 mg
3rd Cell
100
150
1000
D2
0
1.5
50
1500
500
D1
0
D1 D2
2000
-Z'' (ohm)
EP
3.5
400
2500
C1
4.0
Potential (V)
200
Z' (ohm)
200
Time (h)
250
300
0
500
1000
1500
2000
2500
Z' (ohm)
Figure 3. Cycle behavior (a, c, e) and the relevant change of impedance spectra (b, d, f) according to the cell positions. (a) and (b) : the 1st cell, (c) and (d): the 2nd cell, (e) and (f) : the 3rd cell.
ACCEPTED MANUSCRIPT
300
(a)
(b)
Zone A
3.0
6%
250
17% 28% 39%
11% 22%
50
33% 44%
2.5
61% 72% 83% 56% 67%
78%
200
2.0
11% 33% 56% 78% 100%
17% 39% 61% 83%
400
100%
150
100
400
600
800
1000
Z' (ohm)
0
1.0 20
40
60
80
100
Time (h) 800
17%
50
100
150
200
250
300
Z' (ohm)
800
(c)
22% 56%
33% 67%
(d)
44% 78%
78%
83%
94%
100%
89%
Zone B
200
0 0
200
EP
TE D
400
400
Z' (ohm)
600
800
-Z'' (ohm)
600
600
-Z'' (ohm)
0
120
M AN U
0
SC
50 0 200
11%
Zone A
200
1.5 0
6%
RI PT
600
6% 28% 50% 72% 94%
-Z'' (ohm)
0% 22% 44% 67% 89%
0%
94%
89% -Z'' (ohm)
Potential (V)
Zone C
Zone B
Zone C
400
200
0 200
400
600
800
1000
Z' (ohm)
AC C
Figure 4. (a) The first discharge profile of the 6th cell indicating the points of electrochemical impedance measurement. The pattern of impedance evolution can be categorized by (b) zone A showing the increase of Ri + Re without change of Rs, (c) zone B showing the increase of Rs with a negligible increase of Ri + Re, and (d) zone C showing the increase of Rs with a significant increase of Ri + Re (Rs: impedance from the ion conductivity in electrolyte, Ri: impedance from the interface film, Re: impedance from the charge transfer reactions).
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Figure 5. Photographs of (a) disassembled cells and (b) lithium metal anodes after the 1st discharge. The Li metal surface turns black in the 1st cell and then turns to a thick white film in the 6th cell.
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
TE D
Figure 6. SEM images of (a) ~ (c) the cross section of lithium metal anode after the 1st discharge and (d) the surface of lithium metal anode from the 6th cell. The arrows highlight the different thickness dimensions of the metallic lithium layer while the rough and porous surface layer
AC C
EP
grows according to the cell connection order.
ACCEPTED MANUSCRIPT
18000 (a) L-1st
(b) L-3rd
(c) L-6th
9000
*
■
●
SC
6000
RI PT
●
*■
*
12000
*
Intensity (cps)
15000
0 20
30
M AN U
3000
40
50
60
70
2 theta (θ)
Figure 7. XRD patterns of lithium metal anode after the 1st discharge. (*) LiOH, (■) lithium
AC C
EP
TE D
metal, (●) austenite from sample holder.
ACCEPTED MANUSCRIPT
D-1st
3.0
D-2nd
D-3rd
D-4th
RI PT
2.0
1500
D-1st D-2nd D-3rd D-4th D-5th
1.0
0.5
∆Weight (g) -4 -4 -2 -2 -2
SC
Capacity (mAh g -1 carbon ) 1729 1444 1158 1503 2383
Sample
-Z'' (ohm)
1200
1.5
900
M AN U
Voltage (V)
2.5
D-5th
600
300
0
0
300
600
900
1200
1500
Z' (ohm)
0.0 500
1000
1500
2000
2500
TE D
0
3000
Capacity (mAh
3500
4000
4500
5000
g-1carbon)
Figure 8. The 1st discharge profiles according to the cell positions at a constant current of 50 mA
EP
g-1carbon between 2.0 ~ 4.3 V. Experiments conducted in the dry-room which was controlled to less than -40°C dew point. The impedance profiles, the discharge capacity and the cell weight
AC C
change after 1st discharge for each cell are in the inset.
ACCEPTED MANUSCRIPT
G-1
3.0
G-2
G-3
G-4
1500
G-1 G-2 G-3 G-4 G-5 G-6
1.0
0.5
∆Weight (g) -4 -2 -3 -3
1200
SC
Capacity (mAh g -1 carbon ) 1014 1040 1553 918 1118 1359
Sample
-Z'' (ohm)
1.5
G-6
RI PT
2.0
900
M AN U
Voltage (V)
2.5
G-5
600
300
0
0
0.0 500
1000
1500
2000
2500
3000
600 900 Z' (ohm)
3500
4000
1200
4500
1500
5000
TE D
0
300
Capacity (mAh g-1carbon)
Figure 9. The 1st discharge profiles according to the cell positions at a constant current of 50 mA
EP
g-1carbon between 2.0 ~ 4.3 V. Experiments conducted in the glove-box filled with helium which was controlled to less than -90°C dew point. The table lists the 1st discharge capacity and the
AC C
weight change after the 1st discharge. The impedance profiles for each cell, the discharge capacity and the cell weight change after 1st discharge are displayed in the inset.
ACCEPTED MANUSCRIPT
G-1st
(a)
4.5
G-2nd
RI PT
3.5
3.0
2.5
2.0 20
40
60
80
100
M AN U
0
SC
Potential (V)
4.0
120
140
Time (h)
4.5
(b)
120
140
TE D
G-5th
3.5
3.0
EP
Potential (V)
4.0
G-4th
2.5
AC C
2.0 0
20
40
60
80
100
Time (h)
Figure 10. Voltage profiles between 2.0 ~ 4.3 V for 10 cycles at current rates of (a) 50 mA g1
carbon
and (b) 100 mA g-1carbon in the oxygen-filled container in the Helium-filled glove-box
which are controlled to less than -90°C dew point.
ACCEPTED MANUSCRIPT
2000
1st D 3rd D 5th D 7th D 9th D
(a) After Discharge
RI PT
-Z'' (ohm)
1500
200
1000
SC
100
500
M AN U
0
0
0 0
2nd D 4th D 6th D 8th D 10th D
500
100
1000
200
1500
2000
Z' (ohm)
2000
1st C 3rd C 5th C 7th C 9th C
TE D
(b) After Charge
200
1000
EP
-Z'' (ohm)
1500
2nd C 4th C 6th C 8th C 10th C
100
AC C
500
0 0
0
0
500
100
1000
200
1500
2000
Z' (ohm)
Figure 11. Nyquist plots after (a) discharge and (b) charge during 10 cycles.
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
Figure 12. Photographs of disassembled cells after (a) the 1st discharge in the dry-room and (b) the 10 cycles in the glove-box. All the lithium metal anodes remain shiny unlike the lithium metal anode discharged in the sealed container in the ambient atmosphere.
ACCEPTED MANUSCRIPT
Li-O2 batteries have been tested under different atmospheres.
•
Li-O2 battery is very sensitive to even traces of moisture contamination.
•
Moisture leads to higher capacity with serious impedance increase.
AC C
EP
TE D
M AN U
SC
RI PT
•
S0378-7753(14)00851-9
DOI:
10.1016/j.jpowsour.2014.05.148
Reference:
POWER 19233
To appear in:
Journal of Power Sources
Received Date: 25 March 2014 Revised Date:
30 May 2014
Accepted Date: 31 May 2014
Please cite this article as: M.H. Cho, J. Trottie, C. Gagnon, P. Hovington, D. Clément, A. Vijh, C.-S. Kim, A. Guerfi, R. Black, L. Nazar, K. Zaghib, The Effects of Moisture Contamination in the Li-O2 Battery, Journal of Power Sources (2014), doi: 10.1016/j.jpowsour.2014.05.148. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
The Effects of Moisture Contamination in the Li-O2 Battery
RI PT
M.H.Cho1, J.Trottie1, C.Gagnon1, P.Hovington1, D.Clément1, A.Vijh1, C.-S.Kim1, A.Guerfi1, R. Black2, L. Nazar2 and K.Zaghib1*
1
Institut de Recherche d’Hydro-Québec (IREQ), 1800 boul. Lionel Boulet, Varennes,
Department of Chemistry, University of Waterloo, 200 University Avenue West,
Waterloo, Ontario, Canada N2L 3G1
Abstract
M AN U
2
SC
Québec, Canada J3X 1S1
The effects of moisture contamination in the Li-O2 battery system were investigated by
TE D
comparing the electrochemical performance, results from electrochemical impedance spectroscopy and post-mortem analysis of batteries prepared under different atmospheres: sealed container in an ambient atmosphere vs. sealed container in a dry-room or in a
EP
glove-box. The performance of the cells strongly depended on the atmosphere;
AC C
furthermore it was found that the performance degradation of the Li metal anode comes from moisture contamination from the feed lines. In sealed container in an ambient atmosphere, the cells showed higher 1st discharge capacity, higher impedance and significant increase of the cell weight owing to contamination of the oxygen by moisture. Under the same atmosphere, the discharge capacity and the impedance increase were significantly different for different cell connection sequences: cells closer to the gas outlet showed higher discharge capacity, and more pronounced increase in impedance
1
ACCEPTED MANUSCRIPT
and weight. The cells discharged in the dry room or glove box showed less discharge capacity than the cells discharged in ambient atmosphere, but the impedance exhibited negligible increase. Post-mortem analysis revealed that the deterioration of lithium metal
RI PT
anode leads to the cell failure mechanism, and this comes from the moisture contamination. This finding supports the idea that protection of the lithium metal electrode is indispensable to realize the practical application Li-O2 or Li-air batteries. It
SC
was found that the performance of Li-O2 batteries is very sensitive to even traces of moisture contamination and every single part of the cell design including the choice of
M AN U
fitting parts and water permeability of the fitting material should be verified in order to obtain credible and reproducible results.
Keyword: Li-O2 battery; O2 flow path; lithium metal; cycle life; moisture contamination
TE D
*Corresponding author.
AC C
EP
E-mail address: [email protected]
2
ACCEPTED MANUSCRIPT
Introduction The Li-air battery is considered one of the most desired technologies for electric vehicles because of its high theoretical specific energy (more than 3500 Wh kg-1,
RI PT
including oxygen), which could enable a longer driving distance of > 300 miles [1-3]. However, there are many technical challenges that hinder the practical use of lithium-air batteries. To address these technical challenges, significant research has been dedicated
SC
in the last 5 years on investigations of electrolytes [4-7], electrode materials and structures [8-9], catalysts [10-13], binders [14] and anodes [15-16]. Despite the
M AN U
significant progress in understanding the mechanisms associated with the electrode reactions, the current lithium-air technology is very far from the requirements for use in electric vehicles, especially with respect to the cycle life and the power capability. Recently, some research groups have reported improvements extending to a few hundred
TE D
cycles using a capacity or time controlled protocol [10, 17], but the life is still limited to less than tens of cycles in the full operating voltage range (2.0 ~ 4.3V). In this work, we examine the effects of moisture contamination on the Li-O2 battery.
EP
Although it is appreciated that the performance is profoundly sensitive to water, this study allows us to quantify the effects. It is found that the capacity and cell weight is
AC C
significantly increased by even traces of H2O contamination, leading to deterioration of the lithium metal (which serves as an indicator) and failure of the cell. It is also suggested that the seal-tightness should be well verified in all cell studies.
3
ACCEPTED MANUSCRIPT
2. Experimental 2.1. Electrode preparation and cell assembly Carbon black (Super-P, TIMCAL) was coated on carbon paper (TGP-H-030, Toray)
RI PT
and dried under vacuum at 100°C for 24 hrs. The loading of carbon coating was controlled to ca. 0.4 mg cm-2 after drying. Commercially available coin-type cells (Hohsen, punched-CR2032) were assembled in a He filled glove box. Cathode electrodes
SC
with a 16 mm diameter were used. Lithium foil (Hohsen, 200 µm) was used as anode, and 3 layers of porous polypropylene (PP, Celgard 3501)/glass filter (GF/C,
M AN U
Whatman)/PP (Celgard 3501) separators were used to facilitate the recovery of the electrode after disassembly of the cells. The power performance was not deteriorated by the addition of two PP layers at this low current rate of 50 mA g-1carbon. The electrolyte consisted of about 80 mg of LiCF3SO3 (lithium triflate, 99.995%, Sigma-Aldrich):
TE D
TEGDME (tetraethylene glycol dimethyl ether, ≥99%, Sigma-Aldrich) by 1:4 mole ratios, which was injected into the cell after drying to have the moisture content of < 20 ppm H2O. The cell weight was monitored during electrolyte injection and the subsequent
AC C
EP
electrochemical tests.
2.2. Electrochemical tests and characterization The coin cells were placed in the container (Bernardin, Mason Jar 250 ml with Standard 70 mm snap lid 1-PC) that has two fitting lines for gas inlet and gas outlet (Kimble Chase Life Science, Stopcock Nylon 3-way PK/10). Sealing resin (Torr Seal, Varian, Inc.) was applied on the joint parts to enhance the seal-tightness. High purity
4
ACCEPTED MANUSCRIPT
oxygen (>99.995%) was used at a flow rate of about 50 mL min-1. Up to six cell containers were connected in series using silicone tubes (Cole-Parmer Instrument co., Silicone Tubing (Platinum) L/S® 16) to flow the oxygen. The schematic of our
RI PT
experiment is shown in Fig. 1. Before testing, the cells were stabilized at open circuit for 10 hrs with flowing O2. Electrochemical tests were performed at room temperature using a VMP3 cycler (Bio-Logic SAS). The cycle life was monitored at a constant current of
SC
50 mA g-1carbon with voltage cut-off between 2.0 V and 4.3 V. The electrochemical impedance spectra were obtained in the frequency range from 200 kHz to 20 mHz with
M AN U
an amplitude of 5 mV. SEM observation was performed using a Hitachi S-4700 field emission gun scanning electron microscope (FEG-SEM). The Li sample was prepared in an Ar filled glove box and transferred to the FEG-SEM using a proprietary sample exchange holder that prevents air contamination between the glove box and the FEG-
TE D
SEM. XRD was performed using a Rigaku table top X-ray diffractometer running at 600 W (X-ray tube) using Co Kα radiation (1.79Å). The XRD sample was prepared in an Ar filled glove box and analysed in an air-tight sample holder. The moisture content in the
EP
cell containers and the atmosphere was measured by a dew point sensor (GE, M series
AC C
sensor/Moisture Target Series 6).
3. Results and discussion Fig. 2 shows the discharge profile of the 1st cycle according to the cell positions at a constant current for the cells sealed in the containers under flowing O2 as shown in Fig. 1. Some voltage disturbance of the 2nd cell in Fig. 2 is perhaps owing to a bad contact within the cell. The dew point of the first container was -25°C (ca. 500 ppm H2O) and that of the
5
ACCEPTED MANUSCRIPT
6th container was -15°C (ca. 1500 ppm H2O) under the environmental condition in the range of 0 ~ 20°C dew point. It is evident in Fig. 2 that cell-to-cell deviation is more than 100% and there is a trend with the order of O2 flow through the cells. The cells closer to
RI PT
the O2 outlet show higher discharge capacity than those closer to the O2 inlet-the 6th cell has 139% higher discharge capacity than the 1st cell. The inset of Fig. 2 shows that the ohmic resistance (Rs), contributed by the electrolyte, increases with the sequence of the
SC
cell connection and higher Rs values were found with the cell at the oxygen outlet. The weight change also shows the same trend-higher weight increase of the 6th cell (22 mg)
M AN U
cannot be explained by the oxygen uptake (3.3 mg) assuming all the discharge capacity comes from the formation of Li2O2. Similar results were reported regarding the increase of discharge capacity by some previous investigators [20-22]. In the gas containing 10% CO2 in O2, the 1st discharge capacity was about 60% higher than that of pure O2 in
TE D
dimethoxyethane (DME) [20]. In the other report, the 1st discharge capacity with 10% CO2 was about 100% higher than that with pure O2, and the capacity increased to 189% with 50% CO2 in ethylene carbonate (EC) and diethyl carbonate (DEC) [21]. Meini et al.
EP
also reported that water had an effect on the discharge capacity [22]. They found that the cells exposed to water vapor and the cells that had a small leak to the ambient atmosphere
AC C
showed higher discharge capacity than sealed Li-O2 cells. Clearly, the accumulation of parasitic reactions is the reason for the increased discharge capacity according to the cell connection sequence.
Fig. 3 shows the discharge/charge profiles and the relevant impedance change observed at the end of the discharge and charge states for the first 2 cycles. The cells all showed significantly different impedance behavior arising from the cell connection
6
ACCEPTED MANUSCRIPT
sequences. In the 1st cell: Rs remains constantly while [Ri + Re] gradually increased with cycles, where Ri is the impedance contribution from the interface film and Re arises from charge transfer reactions. On the other hand, in the 2nd and 3rd cells, Rs significantly
RI PT
increased with cycling. Especially in the 3rd cell, both Rs and [Ri + Re] increase rapidly with cycling. This result explicitly indicates the increase of contaminant accumulation in subsequent cells with inducing the dry-out of electrolyte. The increase of cell weight is
SC
more significant with cycling than the initial values in Fig. 2. After 10 cycles, the weight increases by -3.1mg, 11.6mg and 46.8mg for the 1st cell, the 2nd cell and the 3rd cell
M AN U
respectively; the decreased weight of the 1st cell is probably due to the electrolyte evaporation surpassing the accumulated weight of by-products. The trend of weight increase indicates that the parasitic reactions are prolonged during the cycles. A more detailed impedance study was carried out on the 6th cell due to its higher
TE D
content of contaminants because of its place in the sequence. The evolution of the in-situ impedance spectra as a function of the depth of discharge (DOD) is shown in Fig. 4. The change of impedance spectra can be separated into 3 different zones categorized by the
EP
following DOD regions; 0~20%, 20~80% and 80~100%. In zone-A (Fig. 4b), [Ri + Re] increases without a change of the Rs, indicating that the electrolytes are relatively stable
AC C
in the beginning of discharge and the reformation of passivation film leads the change of impedance in this zone. The similar phenomenon was reported by Meini et al. [22]; The [Ri + Re] increase by the formation of LiOH layer on the surface of lithium metal anode in the electrolyte containing 1000 ppm water. Zone-B (Fig. 4c) shows a significant increase of Rs with a negligible increase of [Ri + Re], which implies that the electrolyte decomposition is accelerated by the Li2O2 generation in the cathode side. Zone-C (Fig.
7
ACCEPTED MANUSCRIPT
4d) shows the increase of Rs and [Ri + Re]. The increase of Rs is the extension of the electrolyte depletion in Zone-B, while the increase of [Ri + Re] represent the deterioration of electrode structure caused by the accumulation of the reaction products
RI PT
(Li2O2, LiOH, Li2CO3) on the electrodes.
In order to investigate the causes behind the trend of cell failure, post-mortem analysis was carried out on the six cells after the 1st discharge. Fig. 5 shows the optical
SC
photographs of the cell components (a) and lithium anodes (b). It is clearly seen that the Li surface is black in the 1st cell and the surface becomes covered with a thick white film
M AN U
in the 6th cell. The white film is believed to be responsible for the significant increase of the cell weight and the cell resistance. Assary et al. [23] reported the effect of oxygen crossover on the anode of a Li-O2 battery and showed the similar phenomena on the lithium metal. However, it cannot explain the trend according to the cell connection
TE D
sequences since all the cells are exposed to the same conditions of O2-crossover. Fig. 6 shows the SEM images of the cross-section of lithium metal anode after the 1st discharge. The thickness of the metallic lithium layer (compact black layer indicated by
EP
arrows) decreases with the cell connection sequences, and the Li metal surface becomes covered by porous and rough layers (which occupy more volume), as the cell is far down
AC C
the connection sequence. The pores in the anode absorb the electrolyte from the separator layer and are responsible for the severe increase of Rs in Fig. 4. To identify the surface film, XRD analysis was performed on the surface of this layer. Fig. 7 shows that LiOH is the main component that grows with the descending cell connection sequences, which implies that the moisture contamination is leading the parasitic reactions in this experimental condition.
8
ACCEPTED MANUSCRIPT
In this system, two kinds of parasitic reactions can be involved in the cell; the electrochemical reactions that can take place on the cathode electrode during the discharge, which contribute to the discharge capacity, and the chemical reactions that
RI PT
doesn’t contribute to the cell capacity but deteriorate the performance of lithium anode.
2H2O + 2Li+ + 2e- 2LiOH + H2
(1)
(2)
M AN U
2H2O + Li2O2 + 2Li+ + 2e- 4LiOH
SC
The additional capacity caused by H2O can be described as below reactions [25];
The weight increase based on the above reactions is about 1.3 mg (calculated from the excess discharge capacity), which cannot explain the excessive increase of cell weight of 22 mg of the 6th cell in Fig.2. This indicates that additional chemical reactions are
TE D
involved in the cells, leading to the increase of cell weight and the serious damage on the cell performance. As evidenced by the analysis data on the lithium metal anode in Fig. 6,
EP
the parasitic reaction caused by H2O is dominated by the following reaction [26];
(4)
AC C
2H2O + 2Li 2LiOH + H2
The amount of lithium to yield the weight increase of 22 mg corresponds to 35% of lithium anode (200 µm thickness, 2 cm2 dimension) based on the above reaction, which is reasonably in line with the result in Fig. 6. To verify this observation, we set up the test cells in the dry-room which was controlled under -40°C dew point. Fig. 8 represents the 1st discharge profile according to
9
ACCEPTED MANUSCRIPT
the cell positions, which is completely different from the previous results (Fig. 2). The weight of the cell and Rs impedance were found quite stable, while the discharge capacity does not show the correlation with the cell connection sequences.
RI PT
When the cells are cycled in the glove box, where the moisture is controlled under 90°C dew point, the 1st discharge profile is shown in Fig. 9. Unlike the previous tests where all the cell containers are connected in series with continuous flowing O2, the cells
SC
are individually isolated and tested in the closed system after filling O2. Under the O2 flow condition (50 ml min-1), the dew point in the containers was about -65°C, which
M AN U
means that ~5 ppm H2O is originated from the O2 feed line. The figure shows that the 1st discharge capacity is less than the capacity in the dry room, and without an increase of the cell weight and Rs. This result demonstrates that the moisture from the atmosphere and O2 feeding line is responsible for the significant increase of the cell weight, cell
TE D
resistance and 1st discharge capacity in Fig. 2.
Fig. 10 shows the voltage profiles for the 10 cycles with different current rate in the He filled glove-box. The cycle life is improved but still too short to be useful for practical
EP
applications. The impedance measured for 10 cycles are shown in Fig. 11. The plots measured at the end of each discharge (a) and at the end of each charge (b) are totally
AC C
different from those previous obtained under ambient conditions without moisture control (Fig.3-b, d and f). Even when the capacity retention is less than 20% after 10 cycles, the impedance profiles are well maintained in the frequency range between 200 kHz and 20 mHz. The lower frequency behavior is currently under investigation to further understand this phenomenon. To analyze the different behavior according to the cell connection environment, we did the post-mortem analysis again, and the optical photographs of the
10
ACCEPTED MANUSCRIPT
lithium surface are showed in Fig. 12. In both cases, after the 1st discharge in the dryroom (a) and after 10 cycles in the glove-box (b), the Li metal anodes remained shiny
RI PT
without surface changes during the test period of 100 hrs.
4. Conclusions
In this study, we examined lithium-air cells under different experimental conditions
SC
with respect to moisture contamination. We performed electrochemical measurements, electrochemical impedance spectroscopy and post-mortem analysis to understand the
M AN U
effects of moisture contamination by investigations of a sealed container in an ambient atmosphere vs. sealed container in a dry-room or in a glove-box. The cells tested in the sealed container in ambient atmosphere showed higher discharge capacity, higher impedance and significant weight gain after the 1st discharge to 2.0 V under O2 flow. On
TE D
the other hand, the cells tested in the dry-room or glove-box showed less discharge capacity, negligible impedance change and insignificant cell weight change. It is believed that the contaminated gas causes parasitic reactions in the downstream neighboring cells.
EP
Post-mortem analysis revealed that the deterioration of lithium metal anode leads to the cell failure, which is directly related to the moisture contamination. This finding supports
AC C
the idea that protection of the lithium metal electrode is crucial to realize the practical applications of Li-O2 or Li-air batteries. We confirmed that the performance of Li-O2 batteries is very sensitive to even trace amounts of moisture contamination. Furthermore, we found that every single part of cell design including the choice of fitting parts and water permeability of fitting material should be verified to obtain credible results.
11
ACCEPTED MANUSCRIPT
Reference [1] J. Hou, M. Yang, M. W. Ellis, R. B. Moore, B. Yi, Phys. Chem. Chem. Phys., 14 (2012) 13487-13501.
RI PT
[2] J. Christensen, P. Albertus, R. S. Sanchez-Carrera, T. Lohmann, B. Kozinsky, R. Liedtke, J. Ahmed, A. Kojic, J. Electrochem. Soc., 159 (2012) R1-R30.
[3] P. G. Bruce, S. A. Freunberger, L. J. Hardwick, J. M. Tarascon, Nature Mat., 11
SC
(2012) 19-29.
[4] W. Xu, J. Hu, M. H. Engelhard, S. A. Towne, J. S. Hardy, J. Xiao, J. Feng, M. Y. Hu,
M AN U
J. Zhang, F. Ding, M. E. Gross, J. G. Zhang, J. Power Sources, 215 (2012) 240-247. [5] B. D. McCloskey, D. S. Bethune, R. M. Shelby, T. Mori, R. Scheffler, A. Speidel, M. Sherwood, A. C. Luntz, J. Phys. Chem. Lett., 3 (2012) 3043-3047. [6] Y. Chen, S. A. Freunberger, Z. Peng, F. Bardé, P. G. Bruce, J. Amer. Chem. Soc., 134
TE D
(2012) 7952-7957.
[7] M. J. Trahan, S. Mukerjee, E. J. Plichta, M. A. Hendrickson, K. M. Abraham, J. Electrochem. Soc., 160 (2013) A259-A267.
EP
[8] Y. Cui, Z. Wen, X. Liang, Y. Lu, J. Jin, M. Wu, X. Wu, Energy Environ. Sci., 5 (2012) 7893-7897.
AC C
[9] H. D. Lim, K. Y. Park, H. Song, E. Y. Jang, H. Gwon, J. Kim, Y. H. Kim, M. D. Lima, R. O. Robles, X. Lepró, R. H. Baughman, K. Kang, Adv. Mat., (2013) 1-5.
[10] J. J. Xu, D. Xu, Z. L. Wang, H. G. Wang, L. L. Zhang, X. B. Zhang, Angew. Chem. Int. Ed., 52 (2013) 1-5.
[11] B. D. McCloskey, R. Scheffler A. Speidel, D. S. Bethune, R. M. Shelby, A. C. Luntz, J. Am. Chem. Soc., 133 (2011) 18038-18041.
12
ACCEPTED MANUSCRIPT
[12] H. G. Jung, Y. S. Jeong, J. B. Park, Y. K. Sun, B. Scrosati, Y. J. Lee, ACS NANO, 7 (2013) 3532-3539. [13] R. Black, J. H. Lee, B. Adams, C. A. Mims, L. F. Nazar, Angew. Chem. Int. Ed., 52
RI PT
(2013) 392-396.
[14] Y. Cui, Z. Wen, Y. Lu, M. Wu, X. Liang, J. Jin, J. Power Sources, 244 (2013) 614619.
SC
[15] A. Zhamu, G. Chen, C. Liu, D. Neff, Q. Fang, Z. Yu, W. Xiong, Y. Wang, X. Wang. B. Z. Jang, Energy Environ. Sci., 5 (2012) 5701-5707.
M AN U
[16] I. C. Jang, Y. Hidaka, T. Ishihara, J. Power Sources, 244 (2013) 606-609. [17] H. G. Jung, J. Hassoun, J. B. Park, Y. K. Sun, B. Scrosati, Nature Chemistry, 4 (2012) 579-585.
[18] Z. Peng. S. A. Freunberger. Y. Chen, P. G. Bruce, Science, 337 (2012) 563-566.
TE D
[19] H. K. Lim, H. D. Lim, K. Y. Park, D. H. Seo, H. Gwon, J. Hong, W. A. Goddard III, H. Kim, K. Kang, J. Am. Chem. Soc., 135 (2012) 9733-9742. [20] S. R. Gowda, A. Brunet, G. M. Wallraff, B. D. McCloskey, J. Phys. Chem. Lett., 4
EP
(2013) 276-279.
[21] K. Takechi, T. Shiga, T. Asaoka, Chem. Commun., 47 (2011) 3463-3465.
AC C
[22] S. Meini, M. Piana, K. Tsiouvaras, A. Garsuch, H. A. Gasteiger, Electrochem. SolidState Lett., 15 (2012) A45-A48.
[23] R. S. Assary, J. Lu, P. Du, X. Luo, X. Zhang, Y. Ren, L. A. Curtiss, K. Amine, ChemSusChem, 6 (2013) 51-55.
[24] M. M. O. Thotiyl, S. A. Freunberger, Z. Peng, P. G. Bruce, J. Am. Chem. Soc., 135 (2013) 494-500.
13
ACCEPTED MANUSCRIPT
[25] D. Pletcher, J. F. Rohan, A. G. Ritchie, Electrochim. Acta, 39 (1994) 1369-1376.
AC C
EP
TE D
M AN U
SC
RI PT
[26] K. Chen, Z. Zhang, E. Ni, J. Appl. Electrochem. 40 (2010) 197-204.
14
ACCEPTED MANUSCRIPT
Figure captions
Figure 1. Schematic diagram of O2 flow and cell connections. O2 line was serially
cells were independently controlled by a cycler.
RI PT
connected from the 1st cell to the 6th cell through a single pathway while the individual
SC
Figure 2. The 1st discharge profiles according to the cell positions at a constant current of 50 mA g-1carbon between 2.0 ~ 4.3 V in sealed containers in ambient atmosphere. The
M AN U
impedance profiles, the discharge capacity and the cell weight change after discharge for each cell are displayed in the insets, which are significantly influenced by the connection order of the cells.
TE D
Figure 3. Cycle behavior (a, c, e) and the relevant change of impedance spectra (b, d, f) according to the cell positions. (a) and (b) : the 1st cell, (c) and (d): the 2nd cell, (e) and
EP
(f) : the 3rd cell.
Figure 4. (a) The first discharge profile of the 6th cell indicating the points of
AC C
electrochemical impedance measurement. The pattern of impedance evolution can be categorized by (b) zone A showing the increase of Ri + Re without change of Rs, (c) zone B showing the increase of Rs with a negligible increase of Ri + Re, and (d) zone C showing the increase of Rs with a significant increase of Ri + Re (Rs: impedance from the ion conductivity in electrolyte, Ri: impedance from the interface film, Re: impedance from the charge transfer reactions).
15
ACCEPTED MANUSCRIPT
Figure 5. Photographs of (a) disassembled cells and (b) lithium metal anodes after the 1st discharge. The Li metal surface turns black in the 1st cell and then turns to a thick white
RI PT
film in the 6th cell.
Figure 6. SEM images of (a) ~ (c) the cross section of lithium metal anode after the 1st
SC
discharge and (d) the surface of lithium metal anode from the 6th cell. The arrows highlight the different thickness dimensions of the metallic lithium layer while the rough
M AN U
and porous surface layer grows according to the cell connection order.
Figure 7. XRD patterns of lithium metal anode after the 1st discharge. (*) LiOH, (■)
TE D
lithium metal, (●) austenite from sample holder.
Figure 8. The 1st discharge profiles according to the cell positions at a constant current of 50 mA g-1carbon between 2.0 ~ 4.3 V. Experiments conducted in the dry-room which was
EP
controlled to less than -40°C dew point. The impedance profiles, the discharge capacity
AC C
and the cell weight change after 1st discharge for each cell are in the inset.
Figure 9. The 1st discharge profiles according to the cell positions at a constant current of 50 mA g-1carbon between 2.0 ~ 4.3 V. Experiments conducted in the glove-box filled with helium which was controlled to less than -90°C dew point. The table lists the 1st discharge capacity and the weight change after the 1st discharge. The impedance profiles
16
ACCEPTED MANUSCRIPT
for each cell, the discharge capacity and the cell weight change after 1st discharge are displayed in the inset.
RI PT
Figure 10. Voltage profiles between 2.0 ~ 4.3 V for 10 cycles at current rates of (a) 50 mA g-1carbon and (b) 100 mA g-1carbon in the oxygen-filled container in the Helium-filled
SC
glove-box which are controlled to less than -90°C dew point.
M AN U
Figure 11. Nyquist plots after (a) discharge and (b) charge during 10 cycles.
Figure 12. Photographs of disassembled cells after (a) the 1st discharge in the dry-room and (b) the 10 cycles in the glove-box. All the lithium metal anodes remain shiny unlike
AC C
EP
TE D
the lithium metal anode discharged in the sealed container in the ambient atmosphere.
17
ACCEPTED MANUSCRIPT
Silica gel
2nd Cell
3rd Cell
4th Cell
O2 (99.995%)
6th Cell
M AN U
SC
Cycler
5th Cell
RI PT
1st Cell
(Oil)
Figure 1. Schematic diagram of O2 flow and cell connections. O2 line was serially connected from the 1st cell to the 6th cell through a single pathway while the individual cells were
AC C
EP
TE D
independently controlled by a cycler.
ACCEPTED MANUSCRIPT
L-1st
3.0
L-2nd
L-3rd
L-4th
1500
Sample L-1st L-2nd L-3rd L-4th L-5th L-6th
1.0
0.5
∆Weight (g) 1 7 4 16 23 22
1200
SC
Capacity (mAh g -1 carbon ) 1905 2790 2550 3988 3452 4552
-Z'' (ohm)
1.5
L-6th
RI PT
2.0
900
M AN U
Voltage (V)
2.5
L-5th
600
300
0
0
0.0 500
1000
1500
2000
TE D
0
2500
300
3000
Capacity (mAh
600 900 Z' (ohm)
3500
4000
1200
4500
1500
5000
g-1carbon)
Figure 2. The 1st discharge profiles according to the cell positions at a constant current of 50 mA
EP
g-1carbon between 2.0 ~ 4.3 V in sealed containers in ambient atmosphere. The impedance profiles, the discharge capacity and the cell weight change after discharge for each cell are displayed in
AC C
the insets, which are significantly influenced by the connection order of the cells.
ACCEPTED MANUSCRIPT
4.5
400
(a)
(b) C1
4.0
D0
D1
C1
D2
3.0
D0
200
RI PT
3.5
-Z'' (ohm)
Potential (V)
300
2.5 100 2.0
D1
D2
-3.1 mg
1st Cell
0
1.5 0
50
100
150
200
250
0
300
100
200
300
400
Z' (ohm)
SC
Time (h) 4.5
400
(c)
(d)
4.0
D0
D1
C1
D2
M AN U
C1
3.0
-Z'' (ohm)
Potential (V)
300
3.5 D0
2.5
200
100
2.0
D1
D2
11.6 mg
2nd Cell 1.5
0
50
100
150
200
250
300
TE D
0
0
100
Time (h) 4.5
(e)
300
(f)
D0 C1
3.0 D0
AC C
2.5 2.0
46.8 mg
3rd Cell
100
150
1000
D2
0
1.5
50
1500
500
D1
0
D1 D2
2000
-Z'' (ohm)
EP
3.5
400
2500
C1
4.0
Potential (V)
200
Z' (ohm)
200
Time (h)
250
300
0
500
1000
1500
2000
2500
Z' (ohm)
Figure 3. Cycle behavior (a, c, e) and the relevant change of impedance spectra (b, d, f) according to the cell positions. (a) and (b) : the 1st cell, (c) and (d): the 2nd cell, (e) and (f) : the 3rd cell.
ACCEPTED MANUSCRIPT
300
(a)
(b)
Zone A
3.0
6%
250
17% 28% 39%
11% 22%
50
33% 44%
2.5
61% 72% 83% 56% 67%
78%
200
2.0
11% 33% 56% 78% 100%
17% 39% 61% 83%
400
100%
150
100
400
600
800
1000
Z' (ohm)
0
1.0 20
40
60
80
100
Time (h) 800
17%
50
100
150
200
250
300
Z' (ohm)
800
(c)
22% 56%
33% 67%
(d)
44% 78%
78%
83%
94%
100%
89%
Zone B
200
0 0
200
EP
TE D
400
400
Z' (ohm)
600
800
-Z'' (ohm)
600
600
-Z'' (ohm)
0
120
M AN U
0
SC
50 0 200
11%
Zone A
200
1.5 0
6%
RI PT
600
6% 28% 50% 72% 94%
-Z'' (ohm)
0% 22% 44% 67% 89%
0%
94%
89% -Z'' (ohm)
Potential (V)
Zone C
Zone B
Zone C
400
200
0 200
400
600
800
1000
Z' (ohm)
AC C
Figure 4. (a) The first discharge profile of the 6th cell indicating the points of electrochemical impedance measurement. The pattern of impedance evolution can be categorized by (b) zone A showing the increase of Ri + Re without change of Rs, (c) zone B showing the increase of Rs with a negligible increase of Ri + Re, and (d) zone C showing the increase of Rs with a significant increase of Ri + Re (Rs: impedance from the ion conductivity in electrolyte, Ri: impedance from the interface film, Re: impedance from the charge transfer reactions).
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Figure 5. Photographs of (a) disassembled cells and (b) lithium metal anodes after the 1st discharge. The Li metal surface turns black in the 1st cell and then turns to a thick white film in the 6th cell.
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
TE D
Figure 6. SEM images of (a) ~ (c) the cross section of lithium metal anode after the 1st discharge and (d) the surface of lithium metal anode from the 6th cell. The arrows highlight the different thickness dimensions of the metallic lithium layer while the rough and porous surface layer
AC C
EP
grows according to the cell connection order.
ACCEPTED MANUSCRIPT
18000 (a) L-1st
(b) L-3rd
(c) L-6th
9000
*
■
●
SC
6000
RI PT
●
*■
*
12000
*
Intensity (cps)
15000
0 20
30
M AN U
3000
40
50
60
70
2 theta (θ)
Figure 7. XRD patterns of lithium metal anode after the 1st discharge. (*) LiOH, (■) lithium
AC C
EP
TE D
metal, (●) austenite from sample holder.
ACCEPTED MANUSCRIPT
D-1st
3.0
D-2nd
D-3rd
D-4th
RI PT
2.0
1500
D-1st D-2nd D-3rd D-4th D-5th
1.0
0.5
∆Weight (g) -4 -4 -2 -2 -2
SC
Capacity (mAh g -1 carbon ) 1729 1444 1158 1503 2383
Sample
-Z'' (ohm)
1200
1.5
900
M AN U
Voltage (V)
2.5
D-5th
600
300
0
0
300
600
900
1200
1500
Z' (ohm)
0.0 500
1000
1500
2000
2500
TE D
0
3000
Capacity (mAh
3500
4000
4500
5000
g-1carbon)
Figure 8. The 1st discharge profiles according to the cell positions at a constant current of 50 mA
EP
g-1carbon between 2.0 ~ 4.3 V. Experiments conducted in the dry-room which was controlled to less than -40°C dew point. The impedance profiles, the discharge capacity and the cell weight
AC C
change after 1st discharge for each cell are in the inset.
ACCEPTED MANUSCRIPT
G-1
3.0
G-2
G-3
G-4
1500
G-1 G-2 G-3 G-4 G-5 G-6
1.0
0.5
∆Weight (g) -4 -2 -3 -3
1200
SC
Capacity (mAh g -1 carbon ) 1014 1040 1553 918 1118 1359
Sample
-Z'' (ohm)
1.5
G-6
RI PT
2.0
900
M AN U
Voltage (V)
2.5
G-5
600
300
0
0
0.0 500
1000
1500
2000
2500
3000
600 900 Z' (ohm)
3500
4000
1200
4500
1500
5000
TE D
0
300
Capacity (mAh g-1carbon)
Figure 9. The 1st discharge profiles according to the cell positions at a constant current of 50 mA
EP
g-1carbon between 2.0 ~ 4.3 V. Experiments conducted in the glove-box filled with helium which was controlled to less than -90°C dew point. The table lists the 1st discharge capacity and the
AC C
weight change after the 1st discharge. The impedance profiles for each cell, the discharge capacity and the cell weight change after 1st discharge are displayed in the inset.
ACCEPTED MANUSCRIPT
G-1st
(a)
4.5
G-2nd
RI PT
3.5
3.0
2.5
2.0 20
40
60
80
100
M AN U
0
SC
Potential (V)
4.0
120
140
Time (h)
4.5
(b)
120
140
TE D
G-5th
3.5
3.0
EP
Potential (V)
4.0
G-4th
2.5
AC C
2.0 0
20
40
60
80
100
Time (h)
Figure 10. Voltage profiles between 2.0 ~ 4.3 V for 10 cycles at current rates of (a) 50 mA g1
carbon
and (b) 100 mA g-1carbon in the oxygen-filled container in the Helium-filled glove-box
which are controlled to less than -90°C dew point.
ACCEPTED MANUSCRIPT
2000
1st D 3rd D 5th D 7th D 9th D
(a) After Discharge
RI PT
-Z'' (ohm)
1500
200
1000
SC
100
500
M AN U
0
0
0 0
2nd D 4th D 6th D 8th D 10th D
500
100
1000
200
1500
2000
Z' (ohm)
2000
1st C 3rd C 5th C 7th C 9th C
TE D
(b) After Charge
200
1000
EP
-Z'' (ohm)
1500
2nd C 4th C 6th C 8th C 10th C
100
AC C
500
0 0
0
0
500
100
1000
200
1500
2000
Z' (ohm)
Figure 11. Nyquist plots after (a) discharge and (b) charge during 10 cycles.
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
Figure 12. Photographs of disassembled cells after (a) the 1st discharge in the dry-room and (b) the 10 cycles in the glove-box. All the lithium metal anodes remain shiny unlike the lithium metal anode discharged in the sealed container in the ambient atmosphere.
ACCEPTED MANUSCRIPT
Li-O2 batteries have been tested under different atmospheres.
•
Li-O2 battery is very sensitive to even traces of moisture contamination.
•
Moisture leads to higher capacity with serious impedance increase.
AC C
EP
TE D
M AN U
SC
RI PT
•