Subscriber access provided by Warwick University Library
Energy, Environmental, and Catalysis Applications
On the Stability of NaO2 in Na-O2 Batteries Chenjuan Liu, Marco Carboni, William R Brant, Ruijun Pan, Jonas Hedman, Jiefang Zhu, Torbjörn Gustafsson, and Reza Younesi ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.8b01516 • Publication Date (Web): 04 Apr 2018 Downloaded from http://pubs.acs.org on April 4, 2018
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Applied Materials & Interfaces
On the Stability of NaO2 in Na-O2 Batteries Chenjuan Liu, Marco Carboni, William R. Brant, Ruijun Pan, Jonas Hedman, Jiefang Zhu, Torbjörn Gustafsson, Reza Younesi*
Ångström Advanced Battery Centre (ÅABC), Department of Chemistry - Ångström Laboratory, Uppsala University, Box 538, SE-75121, Uppsala, Sweden. KEYWORDS: metal-air battery, in operando X-ray diffraction, sodium superoxide, NaO2, decomposition
ABSTRACT: Na-O2 batteries are regarded as promising candidates for energy-storage. They have higher energy efficiency, rate capability and chemical reversibility than Li-O2 batteries and sodium is cheaper and more abundant compared to lithium. However, inconsistent observations and instability of discharge products have inhibited the understanding of the working mechanism of this technology. In this work, we have investigated a number of factors that influence the stability of the discharge products. By means of in operando powder X-ray diffraction, the influence of oxygen, sodium anode, salt, solvent and the carbon cathode were investigated. The Na metal anode and the ether-based solvent are the main factors that lead to the instability and decomposition of NaO2 in the cell environment. This fundamental insight brings new information on the working mechanism of Na-O2 batteries.
INTRODUCTION Alkali metal-oxygen batteries, such as Li-O2 and Na-O2, have been of great research interest over the past decade as high specific energy alternatives to current state-of-the-art lithium-ion batteries. Particularly attractive is the emerging Na-O2 battery with NaO2 as the discharge product, which has the potential to provide a higher energy efficiency, rate capability and chemical reversibility compared to Li-O2. Sodium is also cheaper and more abundant than lithium. However, both systems are facing great challenges, such as the instability of all cell components during cycling and the large overpotential originating from the insulating discharge products. To advance these two technologies, a deep mechanistic understanding of the electrochemical processes is crucial. In contrast to nonaqueous Li-O2 batteries, a great deal of debate has centred on the cell chemistry of Na-O2 batteries due to reports of a wide range of discharge products. Thus, a single main discharge product has not been observed and agreed upon.1 With their similar standard potential of formation (U° NaO2 = 2.27 V vs. Na+/Na, U° Na2O2 =2.33 V vs. Na+/Na) but different number
of electrons transferred, both NaO2 and Na2O2 as well as Na2O2 · 2H2O have been identified as discharge products.2-6 Unfortunately, the underlying reasons for these divergent results still remain unclear. The conversion of NaO2 to Na2O2 · 2H2O was observed experimentally by the influence of moisture, oxygen or air exposure.7,8 However, other studies show that the as formed NaO2 can be converted to Na2O2 · 2H2O under cycling or in resting the cell.5,9,10 Since NaO2 and Na2O2 are very sensitive to air or moisture, ex situ characterization methods could easily be affected by artefacts introduced during the cell disassembly and/or measuring procedure. Alternatively, in operando analysis can provides direct insight into the real discharge products and the factors that affect the formation of different products.11 In this work, we investigated reaction products formed during discharge of Na-O2 cells and their stability over relaxation (pause) time. In operando and ex situ XRD were used to quantify the amount of the main reaction products, i.e. NaO2, and to understand the possible factors that affect the stability of the discharge products.
1 Environment ACS Paragon Plus
ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 2 of 10
Figure 1. (a) Waterfall plot of the diffraction patterns collected in operando and the corresponding discharge voltage profile during the first discharge with a 0.5 M NaOTf/G2 electrolyte at a current density of 0.1 mA cm-2 and the low cut-off potential of 1.8 V. The mass evolution of NaO2 as a function of time at in the Na-O2 cells: (b) galvanostatically discharged without relaxation/pause, and (c) in the cell with applied pauses during the discharge.
RESULTS AND DISCUSSION In operando analysis of the discharge product. In operando XRD was carried out to track the formation of discharge products using a newly designed electrochemical cell (details in the experimental section). This analysis provides real time quantitative determination of the crystalline discharge products. Figure 1a shows the in operando diffraction patterns and the corresponding galvanostatic discharge voltage profile of a Na-O2 cell with the electrolyte of 0.5 M NaOTf in diglyme (G2) at the current density of 0.1 mA cm-2 and the low cut-off potential of 1.8 V. The cell provided a discharge capacity of around 2.2 mAh cm-2. The appearance of the diffraction peaks at 2θ ~32.5º and 46.6º, corresponding to the (200) and (220) reflections of NaO2 (ICSD # 87176), revealed that the electrochemically formed NaO2 is the sole crystalline discharge product. The peaks from metallic aluminum in the pouch and the current collector are also marked in the diffraction patterns. Silicon, added as an internal intensity standard, shows a peak at 2θ ~28.6º, and enables us to quantify the amount of NaO2 formed during the discharge reaction. Thus, we are able to discuss the ratio between main and parasitic reactions. In order to quantify the efficiency of formation of NaO2, the amount of NaO2 was determined from the weight ratio with the
constant, known quantity of Si reference, via a Rietveld refinement using the FullProf program. The details of the Si reference and the weight ratio calculation can be found in supporting information. When the cell is discharged/charged with a constant current, a linear growth/decomposition of NaO2 is expected in the absence of any side reactions. If all of the exchanged electrons are used to generate NaO2 during discharge, the Na+ + e- +O2 ⇋ NaO2 net reaction is assumed, and the theoretical amount of NaO2 generated via the reaction would be proportional to the discharge capacity (i.e., 2.05 mg NaO2 per mAh would be expected). In Figure 1b, the theoretical growth of NaO2 is presented as a green dot at each 15 min or 30 min interval during discharge. The detected amount of NaO2 (red dots in Figure 1b) is highly efficient up to around 50% of discharge, however, it slows down as the discharge voltage reaches below 2.0 V. The lower efficiency during the end of discharge indicates that i) side reactions occurred and/or ii) produced NaO2 underwent chemical decomposition. Considering the limitation of XRD, we cannot rule out the possibility that the formation of amorphous discharge products during the discharge.
2 Environment ACS Paragon Plus
Page 3 of 10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Applied Materials & Interfaces
Figure 2. Time-resolved in operando XRD showing the evolution of the discharge products of Na-O2 cells that rest/paused (a) inside the cell environment with 0.5 M NaOTf/G2 electrolyte and (b) inside the cell environment with 1.5 M NaOTf/G2 electrolyte. (c) and (d) are the quantity of NaO2 corresponding to (a) and (b), respectively.
To investigate the stability of the NaO2 formed during discharge, a galvanostatic intermittent titration technique (GITT) combined with the in operando XRD method was introduced. Figure 1c, shows another Na-O2 cell discharged with the same electrolyte and current density, while it was paused for 15 min after every 1 h of discharge during which XRD patterns were obtained. As shown in Figure 1c, the discharge capacity is reduced and the NaO2 formation efficiency is much lower in the cell with pauses during the discharge. The discharge capacity of the continuously discharged cell (Figure 1b) is around 2.2 mAh cm-2, while the cell with pauses during the discharge (Figure 1c) delivered a lower capacity of about 1.6 mAh cm-2, which means about 30% of the discharge capacity is lost. A smaller quantity of NaO2 formed in the GITT experiment compared to a pure galvanostatic run indicates that NaO2 is not stable and undergoes decomposition. Therefore, the detected amount of NaO2 results from a competition between the main electrochemical reaction (i.e. formation of NaO2) and chemical parasitic reactions resulting in decomposition of NaO2. However, formation of Na2O2 ·2H2O or similar side products on the cathode is not observed from XRD (Figure S1). Instability of NaO2 during resting of a Na-O2 cell has previously been observed.6,9,10,12 However, the main factors that lead to the decomposition of NaO2 is not yet known. At least five different possible pathways for spontaneous decomposition of the NaO2 could be considered: (1) NaO2 reacts with the electrolyte: salt, solvent, or both
(2) NaO2 is itself unstable at room temperature and a disproportionation reaction takes place (NaO2 + NaO2 → O2 + Na2O2) (3) NaO2 reacts with water from the electrolyte to form NaOH or Na2O2·2H2O (4) NaO2 reacts with the carbon electrode to produce NaCO3 (5) NaO2 can be dissolved into the electrolyte and the asformed O2− can migrate to and react with some parts of the cell, most likely the Na anode. NaO2 stability in the cell environment. In order to analyze the factors influencing the decomposition of NaO2 in the cell environment during relaxation, the discharge products were characterized under different conditions. First, the decomposition of NaO2 in electrolytes with two different salt concentrations was quantitatively demonstrated. Na-O2 cells with 0.5 M and 1.5 M NaOTf dissolved in G2 were fully discharged and then rested at room temperature. Without disassembling the cell, the amount of NaO2 was tracked by in operando XRD, as shown in Figure 2. The NaO2 was stable for about 1.5 days in the 0.5 M cell and about 2 days in the 1.5 M cell. However, as the pause continued, a linear decomposition of NaO2 was observed in both cells with similar decomposition rates (Figure 2c, 2d). Thus, the concentration of the electrolyte salt does not af-
3 Environment ACS Paragon Plus
ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 4 of 10
Figure 3. Time-resolved in operando XRD showing the evolution of the discharge products from Na-O2 cells at rest (a) inside the fresh G2 solvent (H2O < 1ppm) and (b) without Na metal inside the cell environment with 1.5M NaOTf/G2. (c) and (d) are the quantity of NaO2 corresponding to (a) and (b), respectively.
-fect the decomposition rate of NaO2 in the cell environment. Recently, Kim et al. have reported that NaO2 could completely convert to Na2O2·2H2O in an electrochemical cell due to a reaction with the electrolyte solvent, as determined by XRD.9 However, their XRD measurements were carried out after disassembling the cell and rinsing the cathode with a fresh solvent (acetonitrile). The water content in the acetonitrile was not specified. However, our in operando XRD data show that without disassembly of the cell, Na2O2·2H2O was not detected after allowed the cell to rest for more than 20 days. As shown in Figure 2a and 2b, neither NaOH nor Na2O2·2H2O was detected as the side product. Instead, a new unknown phase started to appear after about 4 days resting and at the same time NaO2 started to disappear. Unfortunately, due to insufficient crystallographic data available, identification of this unknown phase was not possible. In this case, we abbreviate this phase as X. As the salt concentration does not affect the decomposition rate of NaO2, it is then worth considering whether the solvent promotes the decomposition. The instability of NaO2 in the G2 solvent was thus investigated. A Na-O2 cell was discharged for 16 h and then disassembled inside a glovebox. The extracted cathode was rinsed by G2 (less than 1 ppm H2O content) to remove the residual salt and solvent. The cathode was then cut roughly in half, one part was dried and sealed in a pouch and the other half was sealed in a pouch containing 150 µL of G2 (the same as electrolyte volume added to Na-O2 cells). The decomposition of the NaO2 in these two samples was followed by XRD, as shown in Figure 3a. The X phase was detected
directly after disassembly, cleaning and placing the cathode in contact with the electrolyte solvent inside a pouch. The decomposition of NaO2 resting in fresh G2 solvent was observed and it is slower than that in the cell environment. In addition, the X phase was reduced in intensity over time. In this regard, once the X phase was formed during the washing procedure, the subsequent decomposition of NaO2 in the pure G2 solvent did not produce X phase. The evolution of the different morphologies of the discharge products, which were stored in a sealed pouch and under G2 solvent (the same electrode used in Figure 3a), respectively, was revealed by SEM. As shown in Figure S2, the discharge products kept the same cubic morphology when stored in the evacuated pouch for 60 days, while cubic particles could no longer be observed in the reference sample kept in G2 solvent. Thus, when the carbon fiber surface became exposed to the solvent the discharge products largely dissolved. Elemental analysis shown in Figure S3a further supports this finding. For the sample kept in G2 solvent, no Na and O could be detected on the carbon fiber. Instead, we observed the formation of an unknown Na-O compound in the pores of the carbon fiber. It has been reported that the as-formed NaO2 can dissolve in fresh electrolyte.9,13-15 If we assume that all of the missing NaO2 in this cathode had dissolved into the fresh G2 solvent, the concentration of the dissolved NaO2 would be ~200 mM. This value of dissolved O2- is slightly higher than the value measured by Kim et al. using electron spin resonance spectroscopy.9 As shown in Figure 3a, after a resting period in the fresh solvent that was 50 times longer than that reported by Kim et al, Na2O2·2H2O was finally detected by XRD. In this regard, our result indi-
4 Environment ACS Paragon Plus
Page 5 of 10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Applied Materials & Interfaces
cates that the liberated O2- from NaO2 can attack the G2 solvent to form Na2O2·2H2O as the side product, which agrees with the observation from Kim et al.9 Based these experiments, it can be concluded that Na2O2·2H2O was more likely formed after the cell disassembly and rinsing the cathode with a fresh solvent. It is worth noting that the disappearance of NaO2 in fresh G2 occurred with a rate of ~ 0.171 mg/day (Figure 3c), which is much slower than the decomposition of NaO2 in the whole cell environment. Thus, the factor, which promotes the fast decomposition of NaO2 in a cell at rest, needs therefore to be understood. As reported by Hartmann et al, O2- that is formed at the cathode surface can migrate within the cell. In this case, O2- could migrate to the anode side and react with Na metal. Therefore, to probe the influence of Na metal, a Na-O2 cell was disassembled directly after a full discharge, and then the Na metal anode was extracted from the cell and the rest of the cell was sealed in a pouch without adding any new electrolyte or washing procedures. After the Na metal was removed, the NaO2 phase remained in the cell after 25 days of resting (Figure 3b). Only a small decomposition of NaO2 was observed at
the very beginning of pause with a rate of 0.333 mg/day (Figure 3d, which is about half of the rate that is observed in the full cell environment). As shown in Figure 3b, without rinsing the electrode, the X phase did not seem to appear at the beginning. However, after about 3 days of resting, the NaO2 begins to decompose very slowly with a rate of about 0.015 mg/day and the X phase appears and grows in intensity during the pause procedure. From this observation, it can be concluded that the presence of Na metal promotes the decomposition of NaO2. It is worth noticing that even with the presence of a similar volume of electrolyte, the decomposition rate of NaO2 in the used electrolyte (without the presence of Na metal) is much slower compared with that in the fresh G2. In addition, the as formed NaO2 maintains a cubic shape inside the cell without Na metal for 80 days (Figure S3 and Figure S4), if the cathode is not rinsed by G2. However, the cubes vanished after resting in the whole cell environment for 80 days, and only some flake-shaped compounds remain on the carbon surface (Figure S4). This suggests that the used electrolyte reaches O2- saturation during the discharge process. However, the decomposition rate of NaO2 in the
Figure 4. The evolution of rinsed and dried NaO2 inside a sealed pouch. The time-resolved XRD of the cathode extracted from a discharged cell with (a) 0.5 M NaOTf/G2 and with (b) 1.5 M NaOTf/G2. (c) and (d) are the evolution of the NaO2 corresponding to (a) and (b), respectively.
whole cell environment is still faster than both of these cases (discharged cathode in fresh solvent and the cell without Na metal). This means, the Na metal in the cell environment is directly (Na reacts with O2-) or indirectly (Na reacts with G2 and the product of this reaction could in turn react with O2-) contributed to the migration of O2- to the anode side, which further promotes the dissolution process of the NaO2 into the electrolyte.
Disproportionation of NaO2. If O2- migration and reaction at the anode surface is the main process for the decomposition of NaO2 in the cell environment, the origin of the slow decomposition of the NaO2 in the absence of Na metal must be due to other reasons. Considering the thermodynamic stability of solid NaO2 produced during discharge, one could question the storage of NaO2 at room temperature. To investigate this possibility, two Na-O2 cells were discharged to the same capacity
5 Environment ACS Paragon Plus
ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
(1.6 mAh cm-2) with the same current density of 0.1 mA cm-2 as used in previous test. The cathodes were directly extracted from the cells after discharge, rinsed with “water free” solvent (< 1 ppm), dried under vacuum and stored in sealed pouch. All these procedures were performed at room temperature. As shown in Figure 4a, 4b, due to the disassembly process, the X phase appears in both cells with 0.5 M and 1.5 M NaOTf/G2, which is the same species that appears after pausing the cell for a long time. When the cleaned NaO2 cathode is stored in an evacuated pouch, the amount of NaO2 reduces slowly over time with a decomposition rate of 0.01 mg/day, while the X phase increases slowly, as shown in Figure 4c, 4d. This rate is about 70 times slower than the case of a cathode resting in the whole cell environment. In addition, after the NaO2 cathode is stored for more than 4 months, Na2CO3 phase was not be detected by XRD, which may original from the reaction between NaO2 and the carbon cathode.
Figure 5. SEM images of cathodes from Na-O2 cells. (a) and (b) were captured from the cell directly after discharge; (c) and (d) after 40 days of storage in the evacuated pouch; (e) and (f) after 135 days of storage in the evacuated pouch.
SEM images revealed the morphology evolution of the discharge products during storage. The cathode was prepared under the same conditions as in Figure 4a. The extracted cathode was rinsed by dry G2 to remove the salt and then dried under vacuum. One part of the cathode was transferred into the SEM directly and the other part was sealed in a pouch under vacuum for future SEM studies. The SEM images revealed that after discharge the cathode consisted of a cubicstructure discharge product, which is reported as NaO2. Figure 5a, 5b indicate that the cubes have smooth surfaces and sharp edges, when the cell is disassembled directly after discharge. However, after 40 days of storage at room temperature, the surface of the cubes show a number of small holes (Figure 5c, 5d). After a rest for more than 4 months, the surface of the NaO2 became much more porous and many larger holes were observed (Figure 5e, 5f). In addition, the pouch that the electrode was stored in had inflated slightly. These results imply a potential disproportionation of NaO2 according to pathway (2).
Page 6 of 10
Thus, NaO2 undergoes self-decomposition and can decompose to the X phase and release gas during storage. The origin of Na2O2·2H2O and X. Na2O2·2H2O has been reported as the discharge product in a number of studies.3,4,16 However, the origin of this product is still not clear. One reason for the formation of Na2O2·2H2O could be the reaction between NaO2 and water in the electrolyte, which is pathway (3) for the decomposition of NaO2. However, if one assumes that all NaO2 reacts with water to from Na2O2·2H2O, then the water content in the electrolyte must be more than 900 times higher than what was measured (Calculation details can be found in SI). In addition, this is also in conflict with experimental observations where a large amount of water was introduced in the electrolyte and only NaO2 was detected as the discharge product.8,17 Therefore, this pathway is not likely to be the main reason for formation of Na2O2·2H2O phase as the discharge product. Another possible source of water contamination could be from O2 gas and/or the tubing from the O2 gas cylinder to the cell.7 To examine this possibility, a clean and dry cathode discharged to a capacity of 1.0 mAh cm-2 was cut into roughly two halves. One half was sealed in a pouch and characterized by XRD directly. The other half was exposed to O2 gas flow using the same gas tube as used for all the cells for one hour and then it was kept under an O2 atmosphere for one day (about twice the time needed to discharge a cell into 1 mAh cm-2). As shown in Figure 6a, Na2O2·2H2O was not detected from the as-discharged cathode. This indicates that the amount of water introduced from the gas tube did not cause the formation of Na2O2·2H2O during discharge. However, when the cathode was removed from the cell and exposed to the same O2 stream, Na2O2·2H2O was detected. A small amount of water may have thus been introduced by the O2 stream, reacting with NaO2 to from Na2O2·2H2O. Interestingly, the X phase, which appeared during the disassembly process, disappeared after exposure to the O2 flow. However, the unknown X phase appears to be more sensitive to water since after exposure 57% of the NaO2 was consumed while all of the X phase disappeared. It is possible that the X phase is one, possible two or more different intermediate transition phases between NaO2 and Na2O2, which could be correlated with the deficient sodium peroxide phase, Na2-xO2 (0
