# Fileset

[Advanced Science - 2023 - Ono - Poor Cycling Performance of Rechargeable Lithium Oxygen Batteries under Lean‐Electrolyte (3).pdf](https://mdr.nims.go.jp/filesets/26c3a129-05c0-49e1-bdb7-dd98bc91718d/download)

## Creator

[Manai Ono](https://orcid.org/0000-0003-4406-4113), [Jittraporn Saengkaew](https://orcid.org/0000-0002-8285-8152), [Shoichi Matsuda](https://orcid.org/0000-0002-0640-3404)

## Rights

[Creative Commons BY-NC-ND Attribution-NonCommercial-NoDerivs 4.0 International](https://creativecommons.org/licenses/by-nc-nd/4.0/)

## Other metadata

[Poor Cycling Performance of Rechargeable Lithium–Oxygen Batteries under Lean‐Electrolyte and High‐Areal‐Capacity Conditions: Role of Carbon Electrode Decomposition](https://mdr.nims.go.jp/datasets/b70de235-dfca-4bb4-b16e-afb1e28402d8)

## Fulltext

Poor Cycling Performance of Rechargeable Lithium–Oxygen Batteries under Lean‐Electrolyte and High‐Areal‐Capacity Conditions: Role of Carbon Electrode DecompositionRESEARCH ARTICLEwww.advancedscience.comPoor Cycling Performance of Rechargeable Lithium–OxygenBatteries under Lean-Electrolyte and High-Areal-CapacityConditions: Role of Carbon Electrode DecompositionManai Ono, Jittraporn Saengkaew, and Shoichi Matsuda*There is growing demand for practical implementation of lithium–oxygenbatteries (LOBs) due to their superior potential for achieving higher energydensity than that of conventional lithium-ion batteries. Although recentstudies demonstrate the stable operation of 500 Wh kg−1-class LOBs, theircycle life remains fancy. For further improving the cycle performance of LOBs,the complicated chemical degradation mechanism in LOBs must beelucidated. In particular, the quantitative contribution of each cell componentto degradation phenomenon in LOBs under lean-electrolyte andhigh-areal-capacity conditions should be clarified. In the present study, themass balance of the positive-electrode reaction in a LOB underlean-electrolyte and high-areal-capacity conditions is quantitatively evaluated.The results reveal carbon electrode decomposition to be the critical factor thatprevents the prolonged cycling of the LOB. Notably, the carbon electrodedecomposition occur during charging at voltages higher than 3.8 V throughthe electrochemical decomposition of solid-state side products. The findingsof this study highlight the significance of improving the stability of the carbonelectrode and/or forming Li2O2, which can decompose at voltages lower than3.8 V, to realize high-energy-density LOBs with long cycle life.1. IntroductionThere is growing societal demand for energy storage deviceswith superior high energy density. Although lithium-ion batter-ies (LIBs) are extensively used as energy storage devices, theirenergy density is approaching the theoretical limit. Thus, theM. Ono, J. Saengkaew, S. MatsudaCenter for Green Research on Energy and Environmental MaterialsNational Institute for Material Science1-1 Namiki, Tsukuba, Ibaraki 305-0044, JapanE-mail: matsuda.shoichi@nims.go.jpS. MatsudaNIMS-SoftBank Advanced Technologies Development CenterNational Institute for Materials Science1-1 Namiki, Tsukuba, Ibaraki 305-0044, JapanThe ORCID identification number(s) for the author(s) of this articlecan be found under https://doi.org/10.1002/advs.202300896© 2023 The Authors. Advanced Science published by Wiley-VCH GmbH.This is an open access article under the terms of the Creative CommonsAttribution License, which permits use, distribution and reproduction inany medium, provided the original work is properly cited.DOI: 10.1002/advs.202300896development of next-generation recharge-able batteries. Lithium–oxygen batteries(LOBs) are specifically attracting signifi-cant attention owing to their high theo-retical energy density.[1–6] However, mostinvestigations on LOBs have been per-formed under excess-electrolyte and low-areal-capacity conditions, resulting in theircell-level energy density is lower thanthat of conventional LIBs.[5,6] Recently, a500 Wh kg−1-class LOB exhibiting stabledischarge/charge cycling behavior at roomtemperature was demonstrated, which wasachieved by minimizing the amount of theelectrolyte as well as using novel electrodeand electrolyte materials.[7–10] However, the500 Wh kg−1-class LOB showed stable cy-cling behavior for < 10 cycles. Thus, thecomplicated chemical degradation mecha-nism in LOBs must be elucidated, in orderto further improve their cycle performance.The limited reaction efficiencies of theoxygen-positive electrode and lithium-negative electrode are thought to beprincipally involved in the degradation mechanism ofLOBs.[1–4,11,12] Recent studies on the lithium negative elec-trode have revealed its unique degradation mechanism in lean-electrolyte systems, which involves phenomena such as electrodeexpansion, electrolyte depletion, and chemical crossover fromthe positive electrode.[13–16] In contrast, the detailed mechanismfor the oxygen-positive electrode in lean-electrolyte systemsremains unclear. The degradation mechanism is believed to in-volve events such as decomposition of the electrolyte and redoxmediator, oxidation of the carbon electrode, and pore-cloggingin the carbon electrode due to the accumulation of solid-stateside products.[17–26] However, such knowledge was obtainedusing LOBs with excess electrolytes (> 50 μL cm−2) and/orlimited areal capacity conditions (< 1 mA cm−2) (Figure 1).Thus, the quantitative contribution of each cell component todegradation phenomenon in LOBs under lean-electrolyte andhigh-areal-capacity conditions remains unknown.Based on these research backgrounds, in the present study, thecomplicated chemical/electrochemical reaction in the positiveelectrode of LOBs under lean-electrolyte and high-areal-capacityconditions was comprehensively analyzed using several in situand ex situ techniques. Using a solid-state ceramic-based sep-arator to protect the lithium-metal negative electrode, the sideAdv. Sci. 2023, 10, 2300896 © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH2300896 (1 of 7)http://www.advancedscience.commailto:matsuda.shoichi@nims.go.jphttps://doi.org/10.1002/advs.202300896http://creativecommons.org/licenses/by/4.0/http://creativecommons.org/licenses/by/4.0/http://crossmark.crossref.org/dialog/?doi=10.1002%2Fadvs.202300896&domain=pdf&date_stamp=2023-06-20www.advancedsciencenews.com www.advancedscience.comFigure 1. Schematic illustration of LOB.reaction with the negative electrode was considerably sup-pressed, which helped exclusively assess the degradation phe-nomena occurring in the oxygen-positive electrode. The resultsindicated that carbon electrode decomposition was particularlydetrimental to the prolonged cycling of LOBs.2. Results and DiscussionIn our experiments, stacked-type LOB cells were utilized(Figure S1, Supporting Information). A self-standing binder-freesingle-walled carbon nanotube (CNT) membrane with mass load-ing of 8.8 mg (2.2 mg cm−2) and 100-μm-thick lithium foil wasused as the positive and negative electrodes, respectively. A solu-tion of 0.5 mol L−1 LiTFSI + 0.5 mol L−1 LiNO3 + 0.2 mol L−1LiBr in tetraethylene glycol dimethyl ether (TEGDME) was usedas the electrolyte. A ceramic-based solid-state separator was usedto protect the lithium-metal negative electrode and minimize itsundesired influence. Here, the ceramic-based solid-state separa-tor was sandwiched between polyolefin layers and the same elec-trolyte was used on the positive and negative sides of the cell.Discharge/charge performance tests were conducted at a currentdensity of 0.2 mA cm−2 capacity limitation of 2.0 mAh cm−2and cutoff voltages of 2.0–4.5 V. The amount of the electrolytein the LOB on the positive-electrode side was controlled at12.5 mg cm−2 Consequently, the ratio of the amount of the elec-trolyte to the areal capacity (E/C) was 6.25 g A−1h−1, which wassufficiently low for realizing cell-level high-energy-density LOB.Figure 2a shows representative voltage profiles of the LOB cell.During the discharge process, the cell exhibited a voltage plateauat ≈ 2.6 V. In charging, there can be seen the stable voltage plateauappeared at ≈ 3.5–3.6 V during middle part of the process. Thevoltage gradually increased up to 4.0 V at the end of the charg-ing. The final voltage in discharge and charge process for eachcycle was plotted against the cycle number (Figure 2b). Associ-ated with the progress of discharge/charge cycle, the overpoten-tial during both discharging and charging increased. As a result,the discharge voltage reached the cutoff voltage of 2.0 V at the20th cycle. In a Li/Li symmetric cell fabricated with the sameelectrolyte composition, repeated cycling was achieved for over30 cycles with an overpotential of < 100 mV (Figure S2, Support-ing Information). Thus, the increase in the overpotential of theLOB was considered to be originated in the positive-electrode re-action.One possible explanation for the increase in overpotentialduring repeated discharge/charge cycling is the accumulationof solid-state side products on the carbon electrode, such asLi2CO3, which decreases the effective electrochemical surfacearea of the electrode. To investigate this possibility, the carbonelectrodes that were taken out from the LOB cell after the 20thcycle were analyzed by scanning electron microscopy (SEM). Forthis experiment, the electrode removed from the LOB cell was,first, washed by TEMDME. After that washed with acetonitrileand then dried in vacuum. The SEM image of the electrodeindicated that no apparent solid-state products accumulated onthe surface of the electrode (Figure 3a–d; Figure S3, SupportingInformation). X-ray photoelectron spectroscopic (XPS). analysisalso indicated that limited amounts of Li2CO3 accumulatedon the electrode (Figure S4, Supporting Information). Theseresults clearly revealed that the increase in overpotential duringrepeated discharging/charging originated from factors otherthan the accumulation of solid-state side products on the carbonelectrode.We investigated the weight change of the carbon electrode dur-ing discharge/charge cycling. The LOB cell was terminated atspecific durations, and the carbon electrodes were taken out. Theelectrodes were washed by water to completely remove the solid-state side products accumulated in the electrode, which was con-firmed by XPS analysis (Figure S5, Supporting Information). InFigure 2c, the weight of the CNT electrode was plotted against thecycle number. There can be seen that the weight of the carbonelectrode was found to linearly decrease with progress of cycle.Notably, the weight of the CNT electrode at the 20th cycle reached5.2 mg (1.3 mg cm−2), which was 60% of the initial weight. SEManalysis of the CNT electrodes (Figure 3e–h) revealed that themorphologies of the CNT electrodes that were removed from theLOB cell after the 1st, 5th, and 10th cycle was almost identicalto that of the pristine CNT electrode (Figure S3, Supporting In-formation). In contrast, the SEM image of the CNT electrode re-moved after the 20th cycle revealed aggregated structures. Photo-graphic images of the CNT electrodes (Figure 3i–l) revealed thatthe CNT electrodes remained self-standing up to the 10th cyclebut collapsed structurally and cannot sustain their self-standingnature after the 20th cycle (Figure 3l).Next, the amounts of the TEGDME solvent and redox mediatorspecies (LiNO3 and LiBr) in electrolyte were quantified. The so-lution taken out from positive electrode side in LOB cell was sub-jected to liquid chromatography–mass spectrometry (LC–MS)and ion chromatography (IC) analyses. In Figure S6 (Support-ing Information), the weight of each component was plottedagainst the cycle number. The amount of TEGDME decreasedwith progress of discharge/charge cycle. As a result, the amountof the electrolyte remaining after the 20th cycle was ≈ 24 mg(6.0 mg cm−2), which was almost half the initial amount ofTEGDME. It should be noted that the amount of electrolyte ex-tracted from the non-cycled cell is 47 mg (11.7 mg cm−2), whichcorresponds with 94% of the initially injected amount of elec-trolyte 50 mg (12.5 mg cm−2). The results also revealed that theamounts of the redox mediator species decreased with progressof cycle (Figure S6, Supporting Information). In our previousstudy, we experimentally confirmed the decomposition of anion(NO3− and Br−) proceeds during the operation of LOBs.[27] Evenin the present study, the decomposition of anion occurs, resultingAdv. Sci. 2023, 10, 2300896 © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH2300896 (2 of 7) 21983844, 2023, 24, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202300896 by National Institute For, Wiley Online Library on [18/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comFigure 2. a) Discharge/charge profile of LOB at selected cycle (black curve: 1st, red curve: 5th, green curve: 10th, blue curve: 15th, purple curve: 20th).b) The final voltage during discharge (black circle) and charge process (red circle) was plotted against cycle number. c) The weight of carbon electrodesthat were taken out from LOB cell at selected cycle was plotted against cycle number.Figure 3. a–h) SEM and i–l) photographic images of carbon electrodes that were taken out from LOB cell at selected cycle. (a–d) The electrodes werewashed by TEGDME and acetonitrile. (b–l) The electrodes were washed by TEGDME, acetonitrile, and water. Scale bars are (a–h)1 μm and (b–l) 1 cm.in a decrease in the number of anions with the progress of re-peated discharge/charge cycles.Overall, the results obtained by analyzing the LOB cell afterthe 20th cycle, in which the discharge voltage reached the cut-off voltage condition, highlighted the following aspects: (i) lim-ited accumulation of solid-state side products in the carbon elec-trode, (ii) decrease in weight of the carbon electrode, and (iii) de-crease in weight of the electrolyte (both solvent and redox me-diator species). Therefore, the bottleneck for the prolonged cy-cling of LOBs had to be subsequently determined, with a focus onidentifying the factor that decreased the discharge voltage. To thatend, the effects of adding the electrolyte to the LOB cell when thedischarge voltage reached the cutoff condition were investigated.Here, the amount of the added electrolyte was set to be equivalentto that of the decomposed electrolyte. Consequently, the amountof the electrolyte in the LOB cell after the electrolyte addition wasidentical to that in the initial cell. Figure 4 shows voltage profilesof the LOB cell before and after the electrolyte addition. The dis-charge voltage initially decreased to the cutoff voltage even afterthe addition of the electrolyte, suggesting that the shortage of theFigure 4. Voltage profile of LOB cell during repeated discharge/charge cy-cles. a) standard LOB cell, b) cycle was stopped at 15th cycle and elec-trolyte was added to carbon electrode, and then re-start cycle test.electrolyte did not primarily impede the prolonged cycling of theLOB cell. Thus, the insufficiency of the carbon electrode was pre-sumably the crucial factor that increased the overpotential duringrepeated discharging/charging.Adv. Sci. 2023, 10, 2300896 © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH2300896 (3 of 7) 21983844, 2023, 24, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202300896 by National Institute For, Wiley Online Library on [18/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comFigure 5. a) Raman spectroscopic analysis of carbon electrodes that were taken out from LOB cell at selected cycle. b) The value of Id against Ig wasplotted against cycle number. c,e) TEM images of pristine CNT electrode and d,f) that were taken out from LOB cell after 20th cycle. Scale bar is (c,d)5 μm and (e,f) 10 nm.To get deep insight into the carbon electrode degradation phe-nomenon, we performed Raman analysis and transmission elec-tron microscopy (TEM) analysis of the CNT electrodes that wereremoved from the LOB cell after different cycles. In the Ra-man spectra of the CNT electrode subjected to various cycles(Figure 5a), peaks appeared at 1300 and 1600 cm−1, which wereassigned to the D-band and G-band, respectively. The ID/IG inten-sity ratio is typically used as an indicator of the defect density inCNTs.[28,29] A plot of ID/IG against the cycle number (Figure 5b)indicated that ID/IG gradually increased with increasing cyclenumber, suggesting an increase in the defect density and/or de-crease in the CNT length. Figure 5c–f shows TEM images of thepristine CNT electrode (Figure 5c–e) and the CNT electrode re-moved from the LOB cell after the 20th cycle. The length of theCNTs in the pristine specimen is known to be 100–600 μm. Thiswas confirmed by the low-magnification TEM image of the pris-tine CNT sample. The high-magnification TEM image revealedthe single-walled structure of the CNTs, which is consistent withpreviously reported results.[28–30] However, the CNTs in the cycledspecimen were < 20 μm long (Figure 5d). Additionally, the pres-ence of defect structures was confirmed by high-resolution TEM(Figure 5f). Therefore, the results of the Raman spectroscopy andTEM analyses clearly suggest the decomposition of the CNT elec-trode during repeated discharging/charging of the LOBs.To further clarify the mechanism of the reaction in the positiveelectrode, on-line MS analysis of the generated gas-phase com-pounds, such as O2, CO2, and H2O, was performed. Figure S7(Supporting Information) shows voltage profiles and time evolu-tions of the gas compounds generated during the 1st, 5th, 10th,and 20th charging cycles. In all cases, the yield of oxygen gen-erated during charging was ≈ 80%. The oxygen generation oc-curred during the entire charging process in the 1st cycle. At thefinal stage of the charging process, oxygen generation abruptlydecreased at a capacity of 1.9 mAh cm−2. CO2 generation wasinitiated at a capacity of 1.7 mAh cm−2 and increased rapidly to-ward the end of the charging. As a result, the total yield of oxygengenerated during the 1st charging was 115 μmol, which corre-sponded to 82% of the theoretical value. This is essentially consis-tent with previous results obtained using similar cell setups.[31,32]The yield of CO2 generated was 27.6 μmol. It should be noted thateven after the charging is completed with capacity limitation of2.0 mA cm−2, the generation of O2 and CO2 does not quickly de-crease, instead gradually decreases. We considered that it takescertain time for generated gas to completely get out from the in-side of pores in CNT electrode. The oxygen generated during the5th cycle started to decrease at a capacity of 1.7 mAh cm−2, and itstotal yield was 112 μmol. However, the yield of CO2 was 43 μmol,which was considerably higher than that of the 1st charging cy-cle. Gas-generation profiles similar to those of the 5th chargingcycle were obtained for the 10th and 20th cycles. The yields ofthe generated gases are summarized in Table 1. Ex situ weightmeasurements of the carbon electrode indicated that ≈ 0.025 mg( = 27 μmol) of carbon decomposed during each cycle. Thus, theresults of in situ MS analysis clearly revealed that more than halfof the CO2 generated during the charging originated from thecarbon electrode decomposition.The results of in situ MS experiments clearly revealed thatmost of the CO2 was generated at voltages higher than 3.8 V.Thus, we hypothesized that the carbon-electrode-decompositionreaction could have occurred predominantly at voltages higherthan 3.8 V. To confirm this hypothesis, experiments were per-formed using a LOB with a charging cutoff voltage of 3.8 V(Figure S8, Supporting Information). Although the stable volt-age profiles were acquired up to the 10th cycle, the overpotentialfor the discharge process started to increase at the 15th cycle. As aresult, the discharge voltage reached the cutoff voltage at the 20thAdv. Sci. 2023, 10, 2300896 © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH2300896 (4 of 7) 21983844, 2023, 24, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202300896 by National Institute For, Wiley Online Library on [18/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comTable 1. Amount of generated O2 and CO2 during charging process.Cycle O2 μmol−1 CO2 μmol−11 115.41 27.685 112.95 43.1310 104.63 45.3715 103.85 52.6820 105.06 49.63Charge start <1 1.68discharge cycle. We also performed the in situ MS experiment ofLOB cell with cutoff voltage of 3.8 V, revealing the limited amountof CO2 generation in this condition (Figures S9 and S10, Sup-porting Information). In addition, ex situ experiment of the elec-trode taken out from LOB cell after 20th cycle revealed that thesolid-state products in the electrode were uniformly distributedon the electrode surface (Figure S11a, Supporting Information),suggesting the amount of accumulated Li2CO3-like side prod-ucts keep increasing with progress of cycle. Figure S11b (Sup-porting Information) showed the SEM image of the CNT elec-trode that was taken out from LOB cell with cutoff voltage of3.8 V, after washing with water to completely remove the solid-state side products accumulated in the electrode. Photographicimages of the CNT electrodes revealed that the CNT electrodes re-mained self-standing condition even after 20th cycle (Figure S12,Supporting Information). These results clearly revealed that thestructure of CNT electrode remains even after 20th cycle, whichis sharp contrast with the case of cutoff voltage of 4.5 V.We also investigated the cutoff voltage dependency of car-bon electrode degradation (Figure S13, Supporting Information).First, there can be seen that ≈ 40% of carbon electrode is decom-posed after 20th cycle when cutoff voltage is 4.2 or 4.5 V. In suchcase, the CNT electrode does not remain their self-standing na-ture. In contrast, in case with cutoff voltage with 3.8 V, the de-composition of carbon electrode is suppressed to 20%. In thiscase, CNT electrode sustains their self-standing property even af-ter 20th cycle. These results clearly revealed that major part ofcarbon electrode decomposition proceeds at the voltage regionover 3.8 V. However, we also mentioned that certain amount ofcarbon decomposition reaction also proceeds at the voltage re-gion < 3.8 V. We also performed Raman spectroscopic analysis ofCNT electrode that taken out after repeated discharge/charge cy-cle with cutoff voltage of 3.8 V condition. The results revealed thatthe increase of ID/IG value keep increasing with progress of cycleeven in case with cutoff voltage of 3.8 V condition (Figure S14,Supporting Information), although the increase rate is smallerthan the case with cutoff voltage of 4.5 V condition.Finally, the detailed mechanism of the CNT-electrode-decomposition reaction had to be clarified. One possible can-didate is the direct electrochemical decomposition of the CNTelectrode. To investigate the contribution of the electrochemicaloxidation of the CNT electrode during charging, the fabricatedLOB cell was charged but not discharged. The acquired voltageand gas-generation profiles (Figure S15, Supporting Informa-tion) indicated that the voltage promptly increased and reachedthe cutoff voltage of 4.5 V at a capacity of 1 mAh cm−2 Moreover,CO2 generation—which was limited—was initiated at a voltageof 3.8 V. Because the amount of CO2 generated was significantlylower than that in the 1st charging cycle, these results revealedthat the CO2 generation during charging did not occur via sim-ple electrochemical oxidation.The other possible mechanism of carbon electrode decomposi-tion during charging at voltages higher than 3.8 V is the decom-position of CNT electrodes associated with the electrochemicaldecomposition of Li2O2 or Li2CO3.[33]Li2O2+C → CO2+2Li++2e− (1)2Li2CO3+C → 3CO2+4Li++4e− (2)The Li2O2 that did not decompose at voltages lower than 3.8 V,and the Li2CO3 generated via the chemical reaction betweenLi2O2 and the carbon electrode or TEGDME, were consideredto decompose electrochemically at voltages higher than 3.8 V.Consequently, the carbon electrode containing these solid-stateproducts also decomposed owing to the electrochemical decom-position of these products, forming CO2 at voltages higher than3.8 V. Such degradation of CNT electrode leads the increase ofover-potential during repeated discharge/charge process. For thedetailed mechanism, following four issues should be considered.(i) decrease of surface area of CNT electrode due to the decreaseof the mass of CNT, (ii) decrease of surface area of CNT elec-trode due to the decrease of electric conduction through CNTelectrode, (iii) increase of IR drop due to the decrease of elec-tric conduction through CNT electrode. In addition, the decreaseof micro-sized pore in CNT electrode results in inefficient oxy-gen transport through CNT electrode. These factors result in theincrease of over-potential during repeated cycle of LOBs.3. ConclusionThe degradation mechanism of LOBs under lean-electrolyte andhigh-areal-capacity conditions was investigated. The conclusionsof a series of analytical investigations conducted in this study areoutlined henceforth. (i) Although the electrolyte degradation in-tensified with increasing repeated discharge/charge cycling, thescarcity of the electrolyte minimally influenced the cycle life ofthe LOBs. (ii) Most of the Li2CO3-like solid-state side productsdecomposed during charging at voltages higher than 3.8 V. (iii)Severe decomposition of the carbon electrode occurred duringcycling, and an equivalent amount of CO2 was generated duringcharging at voltages higher than 3.8 V. Because the carbon elec-trode used in these experiments merely decomposed via simpleAdv. Sci. 2023, 10, 2300896 © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH2300896 (5 of 7) 21983844, 2023, 24, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202300896 by National Institute For, Wiley Online Library on [18/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comelectrochemical oxidation at voltages up to 4.5 V, the carbon elec-trode was considered to deteriorate through the decompositionof its solid-state side products. The findings of this study are an-ticipated to guide future investigations on the advancement ofLOBs. In particular, a detailed understanding of the reaction atthe interface between the carbon electrode and solid-state prod-ucts is crucial for realizing practical high-energy–density LOBswith long cycle life.4. Experimental SectionMaterials: Tetraethylene glycol dimethyl ether: Tetraglym (batterygrade) and lithium bis(trifluoro methanesulfonyl)imide : LiTFSI (Li bat-tery grade) were purchased from Kishida Chemical Co., Ltd. and used asreceived. Lithium nitrate (LiNO3, 99.99% trace metals basis) and lithiumbromide (LiBr, 99.995% trace metals basis) were obtained from Sigma–Aldrich Co., LLC and were dried under vacuo at 120 °C for > 3 days. Liq-uid electrolytes, 0.5 mol L−1 LiTFSI + 0.5 mol L−1 LiNO3 + 0.2 mol L−1LiBr/TEGDME were prepared in a dry condition. Their water contents wereconfirmed < 100 ppm by the Nittoseiko Karl Fischer Moisture Meter CA-31. A lithium foil (100 μm thick) was obtained from The Honjo ChemicalCorp. and cut into 20 mm2. The CNT (carbon nano fiber unwoven cloth)sheet was obtained from Zeon Corporation (100 μm thick, 2 mg cm−2)and cut into 20 mm2 and was used after dried in vacuo at 110 °C for 15 h.The carbon paper TORAYCA H-030 was purchased from Toray Industries,Inc. and cut into 20×22 mm then dried in vacuo at 110 °C for 15 h. A Liconducting solid electrolyte (LICGC-AG-01, 180 μm thick) was obtainedfrom Ohara, Inc. and cut into 25 mm2. Porous polyethylene membraneseparators (20 or 5 μm thick) obtained from Hosen, and Toray Industries,Inc., respectively, were cut into 22 mm2 and kept in the drying condition for> 1 week before use. All dried cell components were stored in a dry boothits dew point lowered to −50 °C for more than overnight before use.Cell Assembly and Discharge/Charge Cycling Tests: An original test cell(stainless steel, inner diameter: 45 mm, depth: 15 mm) equipped with agas inlet and outlet was assembled in a dry room. A Li metal electrode(100 μm thick, 20 mm square) was put onto a 1 mm thickness of thestainless spacer as a current corrector, a porous polyethylene membrane(20 μm thick, 22 mmsquare) infiltrate a 10 μL of electrolyte, a Li+ con-ducting solid electrolyte, a polyethylene membrane (5 μm thick, 22 mmsquare) with another 10 μL of electrolyte, an above-mentioned electrolyteimpregnated CNT electrode, a carbon sheet as a gas diffusion layer, and a1 mm thickness of stainless spacer with springs to fasten at 60 kPa werestacked in the cell. The dead volume of the cell was ≈ 24.5 mL, for excessamount of O2 inside space. Here, a solid electrolyte was used for sep-arating cathode and anode, in order to avoid poisoning electrodes withdecomposition and/or redox shuttling on the Li surface, and two types ofporous polyethylene membranes were for preventing direct contact withelectrodes and solid electrolyte. The sealed cell was purged with dry O2flow at 20 ml min−1. The discharge/charge tests were carried out by theHOKUTO DENKO HJ1020mSD8 at 0.2 mA cm−2 for 10 h at ambient tem-perature followed by the resting time for > 30 min at each interval betweendischarge and charge.Analysis of Gas Phase Components: For on-line MS analysis, the gen-erated gases were directly to the MS detector by the Canon AnelvaQuadrupole Mass Spectrometer M-401GA-DM equipped with a capillarytube (internal diameter: 0.05 mm, length: 7 m). After discharge, the testcell was purged with excess He (50mLmin−1) for 1 min to remove the re-maining O2. He as a carrier gas was flowed at 5 ml min−1 and keep it for 2h before charge. The former on-line measurement was carried out at 100μV applied voltage, the m/z range from 11 to 110 at ambient condition.Analysis of Liquid Phase Components: Liquid chromatography for theorganic component, and ion chromatography for the ions including in-organic materials. After the above measurement, the loss amount of theliquid electrolytes in cathode were extracted with water and analyzed by theAcquity H-class Ultra High-Pressure Liquid Chromatography (UPLC, XevoG2-S QTof, Waters) system coupled with a mass spectrometer and Ionchromatography (ICS-2100, Dionex). The carbon paper, cathode electrode,and cathode separator were immersed in ultrapure water and sonicated for10 min, then filtrated. The obtained cathode samples were diluted prop-erly for each measurement. The liquid TEGDME volume was measured byLC-MS by extracting the electrolyte from the positive electrode in the cellafter the measurements.Analysis of Carbon Electrodes: The electrochemically cycled cathodeelectrode was compared with the pristine one by weight loss measure-ment, morphological observation, chemical composition analysis, andstructural investigation of CNTs. The samples for weighing were obtainedby washing with water to remove electrolyte and solid-state products incycled cathodes. The cathode electrodes after the discharging/chargingcycles were immersed in excess ultrapure water at 40 °C for > 6 h andwiping surplus water, and this process was repeated several times. Theelectrodes were dried in vacuum and put into the Ar-filled glovebox atleased overnight before measurements. The surface morphological mea-surement of the positive electrode was carried out using a field emis-sion scanning electron microscope (JSM-7800F, JEOL) equipped with anenergy-dispersive X-ray spectrometer (X-MaxN 50, Oxford), and the sur-face chemical species were analyzed using X-ray photoelectron spectrausing a VersaProbe II Scanning XPS Microprobe (ULVAC-PHY). The sam-ples after the electrochemical measurements were cut and utilized fromthe weight measurement samples described above. Additionally, in orderto analyze the solid-state products, SEM and XPS measurements werealso performed using the samples washed as follows. The cell was dis-assembled in glove box after electrochemical measurements. The carbonelectrodes were first washed by TEMDME. After that electrodes were im-mersed in excess dry acetonitrile at 40 °C for > 8 h and wipe surplus sol-vent, and this process was repeated several times. The carbon electrodewith remaining solid products was obtained after dried in the glovebox.Sample preparations for SEM and XPS were performed in a glove boxand measured using an unexposed chamber or vessel, respectively. Thestructural defects in CNTs were investigated using microscopic Ramanspectrometer (Raman Touch-vis–NIR, Nanophoton) with a 532 nm laser,and the defects and the length of CNTs were observed using a Transmis-sion Electron Microscope (JEM-ARM200F, JEOL). The samples for Ramanspectroscopy were used the same samples as the weight measurementdescribed above. The TEM samples as dispersed solution was preparedusing pristine or electrochemically cycled carbon electrodes of the weightmeasurement described above. The samples were obtained by sonicationwith a piece of electrode into 1-propanol for > 4 h. Then dispersion wasdropped on a grid and air-dried before measurement.Supporting InformationSupporting Information is available from the Wiley Online Library or fromthe author.AcknowledgementsThe present work was partially supported by the ALCA-SPRING (AdvancedLow Carbon Technology Research and Development Program, the Spe-cially Promoted Research for Innovative Next Generation Batteries) projectof the Japan Science and Technology Agency (JST Grant Number JPM-JAL1301). This work also received support from the National Institute forMaterials Science (NIMS) Battery Research Platform.Conflict of InterestThe authors declare no conflict of interest.Data Availability StatementThe data that support the findings of this study are available from the cor-responding author upon reasonable request.Adv. Sci. 2023, 10, 2300896 © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH2300896 (6 of 7) 21983844, 2023, 24, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202300896 by National Institute For, Wiley Online Library on [18/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comKeywordscarbon nanotube, lean electrolyte, mass balance, metal–air batteriesReceived: February 9, 2023Revised: May 23, 2023Published online: June 20, 2023[1] Z. Wu, Y. Tian, H. Chen, L. Wang, S. Qian, T. Wu, S. Zhang, J. Lu,Chem. Soc. Rev. 2022, 51, 8045.[2] H. F. Wang, Q. Xu, Matter 2019, 1, 565.[3] T. Liu, J. P. Vivek, E. W. Zhao, J. Lei, N. Garcia-Araez, C. P. Grey, Chem.Rev. 2020, 120, 6558.[4] W. J. Kwak, D. Sharon, C. Xia, H. Kim, L. R. Johnson, P. G. Bruce, L.F. Nazar, Y. K. Sun, A. A. Frimer, M. Noked, S. A. Freunberger, D.Aurbach, Chem. Rev. 2020, 120, 6626.[5] M. Ue, K. Sakaushi, K. Uosaki, Mater. Horiz. 2020, 7, 1937.[6] S. Matsuda, M. Ono, S. Yamaguchi, K. Uosaki, Mater. Horiz. 2022, 9,856.[7] S. Matsuda, E. Yasukawa, T. Kameda, S. Kimura, S. Yamaguchi, Y.Kubo, K. Uosaki, Cell Rep. Phys. Sci. 2021, 2, 100506.[8] S. Matsuda, S. Yamaguchi, E. Yasukawa, H. Asahina, H. Kakuta, H.Otani, S. Kimura, T. Kameda, Y. Takayanagi, A. Tajika, Y. Kubo, K.Uosaki, ACS Appl. Energy Mater. 2021, 4, 2563.[9] J. O. Park, M. Kim, J. H. Kim, K. H. Choi, H. C. Lee, W. Choi, S. B. Ma,D. Im, J. Power Sources 2019, 419, 112.[10] J. Saengkaew, T. Kameda, M. Ono, S. Matsuda, Mater. Adv. 2022, 3,3536.[11] P. Zhang, Y. Zhao, X. Zhang, Chem. Soc. Rev. 2018, 47, 2921.[12] D. Wang, X. Mu, P. He, H. Zhou, Mater. Today 2019, 26, 87.[13] W. Liu, D. Lin, A. Pei, Y. Cui, J. Am. Chem. Soc. 2016, 138, 15443.[14] C. P. Yang, Y. X. Yin, S. F. Zhang, N. W. Li, Y. G. Guo, Nat. Commun.2015, 6, 8058.[15] K. H. Chen, K. N. Wood, E. Kazyak, W. S. L. Page, A. L. Davis, A. J.Sanchez, N. P. Dasgupta, J. Mater. Chem. A 2017, 5, 11671.[16] S. A. Rajeev, L. Jun, D. Peng, X. Luo, X. Zhang, Y. Ren, A. C. Larry, K.Amine, ChemSusChem 2013, 6, 51.[17] J. W. Jung, S. H. Cho, J. S. Nam, I. D. Kim, Energy Storage Mater. 2020,24, 512.[18] H. Kim, H. G. Jung, D. Aurbach, Y. K. Sun, J. Electrochem. Soc. 2018,165, A2274.[19] J. B. Park, S. H. Lee, H. G. Jung, D. Aurbach, Y. K. Sun, Adv. Mater.2018, 30, 1704162.[20] S. Yang, P. He, H. Zhou, Energy Environ. Sci. 2016, 9, 1650.[21] B. D. McCloskey, A. Speidel, R. Scheffler, D. C. Miller, V. Viswanathan,J. S. Hummelshøj, J. K. Nørskov, A. C. Luntz, J. Phys. Chem. Lett. 2012,3, 997.[22] Z. Zhao, J. Huang, Z. Peng, Angew. Chem., Int. Ed. 2018, 57,3874.[23] S. Zhang, M. J. Nava, G. K. Chow, N. Lopez, G. Wu, D. R. Britt, D. G.Nocera, C. C. Cummins, Chem. Sci. 2017, 8, 6117.[24] B. M. Gallant, R. R. Mitchell, D. G. Kwabi, J. Zhou, L. Zuin, C. V.Thompson, Y. Shao-Horn, J. Phys. Chem. C 2012, 116, 20800.[25] M. M. Ottakam Thotiyl, S. A. Freunberger, Z. Peng, P. G. Bruce, J. Am.Chem. Soc. 2013, 135, 494.[26] N. B. Aetukuri, B. D. McCloskey, J. M. García, L. E. Krupp, V.Viswanathan, A. C. Luntz, Nat. Chem. 2015, 7, 50.[27] M. Ono, S. Matsuda, J. Phys. Chem. C 2023, 127, 6117.[28] X. Xiao, P. Tan, X. Zhu, Y. Dai, C. Cheng, M. Ni, ACS Sustainable Chem.Eng. 2020, 8, 9742.[29] A. Ahmadiparidari, R. E. Warburton, L. Majidi, M. Asadi, A. Chamaani,J. R. Jokisaari, S. Rastegar, Z. Hemmat, B. Sayahpour, R. S. Assary,B. Narayanan, P. Abbasi, P. C. Redfern, A. Ngo, M. Vörös, J.Greeley, R. Klie, L. A. Curtiss, A. Salehi-Khojin, Adv. Mater. 2019, 31,1902518.[30] Z. Zhang, W. L. Bai, K. X. Wang, J. S. Chen, Energy Environ. Sci. 2020,13, 4717.[31] M. Ue, H. Asahina, S. Matsuda, K. Uosaki, RSC Adv. 2020, 10, 42971.[32] S. Matsuda, M. Ono, H. Asahina, S. Kimura, E. Mizuki, E. Yasukawa,S. Yamaguchi, Y. Kubo, K. Uosaki, Adv. Energy Mater. 2023, 13,2203062.[33] D. Cao, C. Tan, Y. Chen, Nat. Commun. 2022, 13, 4908.Adv. Sci. 2023, 10, 2300896 © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH2300896 (7 of 7) 21983844, 2023, 24, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202300896 by National Institute For, Wiley Online Library on [18/06/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.com