Electrochimica Acta 147 (2014) 1–8

A carbon powder-nanotube composite cathode for non-aqueouslithium-air batteries

P. Tan, W. Shyy, Z.H. Wei, L. An, T.S. Zhao *Department of Mechanical and Aerospace Engineering, The Hong Kong University of Science and Technology Clear Water Bay, Kowloon, Hong Kong SAR,China

A R T I C L E I N F O

Article history:Received 24 June 2014Received in revised form 7 August 2014Accepted 2 September 2014Available online 22 September 2014

Keywords:Lithium-air batteryDischarge capacityCarbon powderCathode structureOxygen transport

A B S T R A C T

Carbon powder has been predominately used to form cathode electrodes for non-aqueous lithium-airbatteries, mainly due to their large specific surface area. An issue, however, with carbon-powder basedcathodes is the large oxygen transport resistance due to limited pore spaces resulting from the packingwith nanosized spherical particles, leading to a practical discharge capacity much lower than thetheoretical value. The present work addresses this issue by proposing a composite cathode made of carbonpowder and nanotubes for non-aqueous lithium-air batteries. The discharge performance character-izations show that the discharge capacity of the cathode with mixed carbon materials increases with anincrease in the ratio of carbon nanotubes to powder. At the ratio of 1:1, the highest volumetric and thegravimetric capacity are achieved, which are respectively 67.2% and 36.3% higher than those with thecathode made of pure carbon powder. It is further demonstrated that the battery with the compositecathode at a fixed capacity of 1.0 mA h/cm2 exhibits a cycle life of up to 50 cycles, which is nearly twice thecycle number of the battery with its cathode made of pure carbon powder. The mechanism leading to theimproved performance can be mainly attributed to the improved oxygen transport as the result ofenlarged pore spaces with an appropriate composition of spherical carbon powder and cylindrical carbonnanotubes.

ã 2014 Elsevier Ltd. All rights reserved.

Contents lists available at ScienceDirect

Electrochimica Acta

journa l home page : www.e l sev ier .com/ loca te /e le cta cta

1. Introduction

Since the emergence of a rechargeable lithium-air batterybased on a non-aqueous polymeric electrolyte in 1996 [1], non-aqueous lithium-air batteries have attracted a great deal ofattention due to the large theoretical capacity (3.86�103mA h/g)and high voltage (2.96 V), which correspond to the energy densityof 1.14�104W h/kg. The capacity of non-aqueous lithium-airbatteries is still two to four times higher than that of Li-ionbatteries, even when taking into account of the entire batterysystem by volume [2]. This striking feature allows the lithium-airbattery to be one of the most promising power sources for next-generation electric vehicles and portable devices [3].

However, before this technology can be viably commercial-ized, a variety of technical hurdles must first be overcome. The

* Corresponding author. Tel.: +852 2358 8647E-mail address: metzhao@ust.hk (T.S. Zhao).

http://dx.doi.org/10.1016/j.electacta.2014.09.0740013-4686/ã 2014 Elsevier Ltd. All rights reserved.

main challenges include the instability of electrolyte andelectrode materials [4–7] caused by the active intermediatesfrom the oxygen reduction reaction during discharge [8,9], thelower discharge capacity than the theoretical value, and the poorreversibility. In addition to these stated issues, a key factor thatlimits the capacity is related to the formation of Li2O2 duringdischarge [10,11], which is insoluble in the non-aqueouselectrolyte and deposits in the void spaces of the porous cathode.The solid product continuously grows on the pore-solid surfaceswith an increase in the discharge capacity [12], and ultimatelyblocks the respective pathways for the transport of oxygen andlithium ions. Due to the low electric conductivity of Li2O2 [13], theelectron transport can also become resistant with the growth ofthe discharge product. As a result, the discharge process isterminated before the whole cathode volume is utilized, leadingto an actual capacity much lower than the theoretical value.

Carbon powder materials, e.g. Super P and Ketjen black, havebeen predominately studied as the cathode electrode for non-aqueous lithium-air batteries, mainly due to the large specific

Fig. 1. Morphology of the cathode before discharge (10000�) made of (a) purecarbon powder, (b) mixed carbon powder and nanotubes.

2 P. Tan et al. / Electrochimica Acta 147 (2014) 1–8

surface area. For example, Tran et al. [14] and Park et al. [15]investigated the discharge performance of activated carbon andvarious types of carbon powder, and found that micro-pores andsome meso-pores could be blocked by the discharge product,suggesting that carbon powder with a high surface areaassociated with larger pores is needed. Hayashi et al. [16] foundthat the capacity was in proportion to the surface areas of thecarbon powder, and the meso-pores seemed to function as activesites during discharge. Gao et al. [17] used various carbon powdermaterials to form the cathode, including Super P, Vulcan-XC72,Ketjen black and activated carbon, and showed that the carbonsources and loadings had a large impact on the discharge capacity.Zhang et al. [18] studied the effect of two types of carbon powder,Super P and Ketjen black, on the discharge performance; they alsofabricated a novel composite cathode comprising these two typesof carbon powder to increase the discharge capacity by 20%.

Previous studies suggest that for the cathodes made of carbonpowder materials, although the specific surface area is large, smallpores can be easily blocked by the solid product, causing a highoxygen transport resistance [18]. As a result, the utilization of thevoid volume in the cathode is low, leading to a discharge capacitymuch lower than the theoretical value. To address this issue, in thiswork we propose to form a composite cathode of non-aqueouslithium-air batteries with a mixture of carbon powder andnanotubes. The morphology of the cathodes made of carbonpowder and nanotubes at different weight ratios before and afterdischarge was examined, and the product chemical compositionwas also detected. The rate and cycle performance of the batterywith the composite cathode were tested, and compared with thebattery with the cathode made of pure carbon powder. Thecorrelation between the discharge capacity, the pore spaces, andthe transport resistance of oxygen was studied.

2. Experimental

2.1. Cathode fabrication

The carbon powder applied in this work is Ketjen black EC-600JD with a diameter of approximately 20 nm (AkzoNobel Co.Ltd., China). The carbon nanotubes are multi-walled withdiameters ranging from 40 to 60 nm and lengths ranging from5 to 15 mm (Shenzhen Nanotech Co. Ltd., China). The specificsurface area and pore size distribution of carbon powder andnanotubes were analyzed by liquid nitrogen sorption measure-ments based on the Brunauer-Emmett-Teller (BET) and Barrett-Joyner-Halenda (BJH) methods, respectively.

The carbon nanotubes were added to the carbon powder withdifferent weight ratios, and mixed with polytetrafluorethylene(PTFE) as the binder at the dry weight ratio of 7:3. Ethanol wasadded as a dispersing agent and then the slurry was ultrasonicallystirred for 1 hour. After being air-dried, the mixture was rolled tothe thickness of 350 mm and cut into the film electrode with the

Table 1Geometrical properties of the two carbon materials

Carbonmaterial

Surface area,m2/g

Micropore surfacearea, m2/g

Mesopore surfacearea, m2/g

Macroarea, m

Carbonnanotubes

48.92 0.01 27.17 21.74

Carbonpowder

1334.40 12.10 648.55 673.75

diameter of 10 mm. The carbon loading of the cathode electroderanged from 7.5 to 11.5 mg. Finally, all the electrodes were baked at240 �C and sintered at 350 �C, both for 1 hour.

2.2. Battery assembling and test

The lithium-air battery contains a lithium metal foil as theanode, a glass-fiber separator (Whatman GF/C), and the carboncathode as prepared. 200 mL 0.25 M lithium perchlorate (LiClO4,Sigma-Aldrich, 98%) in tetraethylene glycol dimethyl ether(TEGDME, Sigma-Aldrich, 99%) was added in the battery to fullysaturate the separator and the cathode. The entire battery

pore surface2/g

Total pore volume,cm3/g

Mesopore volume,cm3/g

Macropore volume,cm3/g

0.16 0.08 0.08

2.30 1.32 0.98

Fig. 3. Capacity increment with the weight percentage of carbon nanotubes in thecomposite cathode.

Fig. 2. Comparison in the discharge capacity among different cathode compositions(CP: Carbon powder, CNT: carbon nanotubes).

P. Tan et al. / Electrochimica Acta 147 (2014) 1–8 3

assembly was carried out in an argon-filled glove box (Etelux, Lab2000) at water and oxygen contents below 1 ppm.

The battery was rested by a constant flow of pure oxygen for2 hours, and then exposed to oxygen at the constant pressure ofabout 1 atm during the discharge process. The capacity of theprepared carbon cathode was examined at the discharge currentdensity of 0.4 mA/cm2 by the battery cycling system (Neware, CT-3008W) with the cut-off voltage of 2.0 V. The rate performance wasexamined at the discharge current densities of 0.4, 0.5, and 0.6 mA/cm2. The cycle performance was tested at the current density of0.4 mA/cm2 with a fixed capacity of 1.0 mA�h/cm2, and the cut-offvoltage of 2.0 V for discharge and of 4.8 V for charge. All testes werecarried out at room temperature (25 �C).

2.3. Product characterization

The cathode morphology before and after discharge wasobserved by a scanning electron microscope (SEM, JEOL Inc., JSM-6300) under an acceleration voltage of 20 kV. The structurecharacterization of the discharge product was analyzed by aPhilips high resolution X-ray diffraction system (XRD, model PW1825) using a Cu-Ka source operating at 40 keV. The productsformed during cycling were examined by Fourier-transforminfrared spectrometer (FT-IR, Vertex 70, Bruker) in the frequencyrange of 400-2000 cm�1 under an argon atmosphere. To obtain thecathode after test, the battery was disassembled in the argon glovebox, and the cathode was rinsed by pure TEGDME and then driedat room temperature under an argon atmosphere. For allmeasurements, a home-made gas container filled with argonwas used to transfer the discharged cathodes.

Table 2Gravimetric and volumetric capacities of each cathode

CNT in the cathode (wt%) 0 20

Volumetric capacity (mAh/cm3) 154.35 155.61

Gravimetric capacity (mAh/g) 571.54 555.23

3. Results and discussion

3.1. Cathode structure characterization before discharge

The surface areas and pore volumes of these two carbonmaterials are listed in Table 1. The specific area and total porevolume of carbon powder are much larger than those ofnanotubes, due to the small particle size and abundant meso-and micro-pores. Since the large geometrical differencebetween carbon nanotubes and carbon powder, the additionof carbon nanotubes would change the cathode microscopicstructure. As shown in Fig. 1a, For the cathode made of purecarbon powder, spherical particles accumulate closely to formlarge aggregations. Apart from the pores formed by theinterstices among single particles, the interstices among theaggregates also form as pores. The pores are around severalhundred nanometers and provide both the pathways foroxygen transport as well as the spaces for the dischargeproduct. Fig. 1b shows the composite cathode made of mixedcarbon materials at the weight ratio of 1:1. Carbon powder andnanotubes are well mixed, and the addition of carbon nanotubeschanges the gathering state of carbon powder. Compared withthe cathode made of carbon powder, the particle aggregationsize is reduced. Apart from the pores formed by carbon powder,the carbon powder and nanotubes also connect with each otherto form a framework with numerous large interconnected poresfor oxygen transport and product accumulation. Therefore, thepore spaces are enlarged in the composite cathodes, which mayinfluence the oxygen transport and thus improve the dischargecapacity.

40 50 60 80

187.36 258.06 224.61 151.15597.86 779.14 653.03 400.86

Fig. 4. XRD patterns of the cathodes before and after discharge and the referencepatterns according to JCPDS database.

4 P. Tan et al. / Electrochimica Acta 147 (2014) 1–8

3.2. Discharge performance

The discharge voltage curves of all the cathodes are presentedin Fig. 2. The voltage plateau for the cathode made of pure carbonpowder is at about 2.50 V, with the capacity at 4.24 mA�h. For the

Fig. 5. Morphology of the cathode after discharge (10000�) made of (a) pure carbon powoxygen inlet and (2) separator.

composite cathodes with mixed carbon materials, the voltageplateaus are around 2.50 V. Even though the addition of carbonnanotubes may decrease the specific surface area due to the smallspecific surface area, the influence on the discharge voltage is notremarkable at such discharge condition. The discharge capacityincreases with an increase in the ratio of carbon nanotubes topowder, and the highest capacity (7.09 mA�h) is achieved at theratio of 1:1. The gravimetric and volumetric capacities of eachcathode are listed in Table 2. The percentage increment from thecapacity (both volumetric and gravimetric) of the pure carbon-powder cathode to that of the composite cathode is presented inFig. 3. It is seen that with an increase in the weight percentage ofcarbon nanotubes in the composite cathode, both the volumetricand gravimetric capacities increase, reach the peak at thepercentage of 50% (or at the weight ratio of carbon nanotubesto powder of 1:1), and then quickly decreases. It is significant tonotice that with the composite cathode made of half carbonpowder and half carbon nanotubes, the volumetric and gravimet-ric capacities increased, respectively, by 67.2% and 36.3%. It is alsoworth noting that the XRD characterizations, shown in Fig. 4,suggest that the discharge product in all the cathodes is Li2O2.

3.3. Cathode structure characterization after discharge

To provide insight into why the composite cathode with mixedcarbon materials led to an improved capacity, we examined themorphology of the cathodes after discharge. As shown in Fig. 5a,for the cathode made of pure carbon powder, at the side facing the

der, (b) mixed carbon powder and nanotubes at the weight ratio of 1:1 facing the (1)

Fig. 6. Comparisons in the (a) discharge voltage profile and (b) discharge capacity of the carbon powder and composite cathode at various current densities.

P. Tan et al. / Electrochimica Acta 147 (2014) 1–8 5

oxygen inlet, the surface is fully covered by the solid Li2O2 layer;while at the side facing the separator, the morphology is similar tothat before discharge, with many pores left empty. Thisphenomenon indicates that although the cathode made of purecarbon powder has a large specific surface area, the pores can beeasily occupied by the solid product, which further blocks thepathway for oxygen to transport and leads to the high oxygentransport resistance and a lower discharge capacity than thetheoretical value. When adding carbon nanotubes to the carbonpowder to form the composite cathode, as shown in Fig. 5b, atthe oxygen inlet side, although the surface is covered by solidLi2O2, near the carbon nanotubes there are still pores for oxygen todiffuse into the cathode. Hence, at the side facing the separator, thereaction area is also utilized with solid product covering thesurface. Therefore, for the cathodes made of mixed carbonmaterials, even though the specific surface area is reducedcompared with the cathode made of pure carbon powder, theoxygen transport improves as the result of enlarged pore spaces. Itis also worth noting that with too small amount of carbonnanotubes, the improvement is quite limited; while with too manycarbon nanotubes, the improvement in oxygen transport cannotcompensate the decrease of specific reaction area, leading to the

decrease of capacity. As a result, with an appropriate mixing ratioof carbon nanotubes to carbon powder in the cathode, the porespaces are optimum to improve oxygen transport, leading to theimproved discharge capacity.

3.4. Rate performance

To further evaluate the rate performance of the battery with thecarbon powder-nanotubes composite cathode, we measured thedischarge capacities by increasing the current density from 0.4 to0.6 mA/cm2, and compared with those of the battery with thecathode made of pure carbon powder. The results are shown inFig. 6. As shown in Fig. 6a, the discharge voltage plateau of thecomposite cathode is lower than that of the carbon powder cathode,and the difference becomes larger with an increase in the dischargecurrent density. This behavior is caused by the decrease in thespecific surface area of the cathode due to the addition of carbonnanotubes. As shown in Fig. 6b, the composite cathode exhibits thedischarge capacities of 7.09 mA�h at 0.4 mA/cm2, 5.74 mA�h at0.5 mA/cm2, and 4.28 mA�h at 0.6 mA/cm2, all of which are largerthan the carbon powder cathode. It is worth noting that with thedischarge current density increasing from 0.4 to 0.6 mA/cm2, the

Fig. 8. The (a) discharge and (b) charge capacity as a function of cycle number forthe cathode made of pure carbon powder and mixed carbon powder and nanotubesat the weight ratio of 1:1.

Fig. 9. FT-IR characterization of the cathodes before and after cycle.

Fig. 7. Discharge and charge voltage curves of the cathode made of (a) pure carbonpowder, (b) mixed carbon powder and nanotubes at the weight ratio of 1:1.

6 P. Tan et al. / Electrochimica Acta 147 (2014) 1–8

capacity retention of the composite cathode is 60.4%, which is higherthan that of the carbon powder cathode (52.8%), showing the betterrate performance.

3.5. Cycle performance

The cycle performance of the battery with the carbon powder-nanotubes composite cathode was tested at a fixed capacity of1.0 mA h/cm2, and compared with that of the battery with thecathode made of pure carbon powder. As shown in Fig. 7a, thedischarge voltage plateau of the carbon powder cathode keepsincreasing in the first 5 cycles, which is caused by the increasedoxygen concentration inside the cathode, due to the releasedoxygen during the charge process (Li2O2! 2Li + O2). After the 5th

cycle, the discharge voltage plateau starts to decrease with cycling,and the charge voltage increases with cycling. Finally, the chargevoltage reaches to 4.8 V at the 22nd cycle before fully charged, andthe discharge capacity starts to decay at the 29th cycle. For thebattery with the composite cathode, as shown in Fig. 7b, there is nosignificant change in the discharge and charge curves up to40 cycles, indicating the good reversibility. In addition, the nearlyconsistent discharge curves indicate the improved oxygentransport in the cathode. The charge capacity starts to decreaseat the 41st cycle, and the discharge capacity decays at the 51st cycle.

Fig. 10. Morphology of the cathode after cycle (10000�) made of (a) pure carbon powder, (b) mixed carbon powder and nanotubes at the weight ratio of 1:1 facing the (1)oxygen inlet and (2) separator.

P. Tan et al. / Electrochimica Acta 147 (2014) 1–8 7

As presented in Fig. 8, the battery with the composite cathodecould run for 50 and 40 cycles in discharge and charge,respectively, with maintaining the initial capacity. While thebattery with the carbon powder cathode only runs for 28 and21 cycles in discharge and charge, respectively. Consequently, thebattery with the composite cathode achieves nearly twice the cyclenumber of the battery with its cathode made of powder cathode,exhibiting the better cycling stability.

The cathodes after cycling were further investigated by FT-IRto examine the reaction products, as shown in Fig. 9. Theformation of Li2CO3 and Li carboxylates (HCO2Li and CH3CO2Li)was detected in both carbon powder and carbon powder-nanotubes composite cathodes, which is ascribed to thedecomposition of the electrolyte [19–21] and the carbonelectrode [7,22]. Since the dominant reactions in non-aqueouslithium-air batteries are the formation and decomposition ofLi2O2 during discharge and charge processes, those byproductscan hardly be completely decomposed during the chargeprocess. The accumulation of byproducts upon cycling couldnot only cover the reaction sites, but also occupy the void spacefor Li2O2 deposition. For these reasons, the battery undergoes anirreversible process, leading to the decay in the capacity duringcycling.

The morphology of the cathodes after cycle was examined bySEM and presented in Fig. 10. For the carbon powder cathode, asshown in Fig. 10a, the oxygen side is covered by thick-layerbyproducts, while the separator side is relatively clean, with fewbyproducts accumulation. This phenomenon indicates that for thecathode made of carbon powder, due to the small oxygen transport

pathways, oxygen is hard to transport into the deep of the cathodeduring cycling, resulting in a waste of reaction sites. Forthe composite cathode, as shown in Fig. 10b, both sides are coveredby the byproducts, showing the increased utilization of thecathode. In addition, some pores in the cathode are still open,suggesting that the decrease of the capacity is mainly caused by thecoverage of byproducts on the reaction sites. Therefore, to furtherimprove the cycle performance, suppressing the side reactionsbetween the electrolyte and the carbon cathode [23,24], orapplying more stable electrolytes [25,26] and electrodes[11,27,28], are in great need in future research.

4. Conclusions

In summary, we proposed, fabricated, and tested a compositecathode made of carbon powder and nanotubes for a non-aqueous lithium-air battery. The discharge performance charac-terizations show that the discharge capacity of the cathode withmixed carbon materials increases with an increase in the ratio ofcarbon nanotubes to powder. The highest capacity was achievedat the ratio of 1:1; the respective volumetric capacity and thegravimetric capacity are 67.2% and 36.3% higher than those withthe cathode made of pure carbon powder. It is furtherdemonstrated that the battery with the composite cathode ata fixed capacity of 1.0 mA h/cm2 exhibits a cycle life of up to50 cycles, which is nearly twice the cycle number of the batterywith its cathode made of pure carbon powder. The mechanismleading to the improved performance can be mainly attributed tothe improved oxygen transport as the result of improved pore

8 P. Tan et al. / Electrochimica Acta 147 (2014) 1–8

spaces with an appropriate composition of spherical carbonpowder and cylindrical carbon nanotubes.

Acknowledgements

The work described in this paper was fully supported by a grantfrom the Research Grants Council of the Hong Kong SpecialAdministrative Region, China (Project No. 622712).

References

[1] K. Abraham, Z. Jiang, A polymer electrolyte-based rechargeable lithium/oxygenbattery, J. Electrochem. Soc. 143 (1996) 1–5.

[2] P. Tan, Z. Wei, W. Shyy, T.S. Zhao, Prediction of the theoretical capacity of non-aqueous lithium-air batteries, Appl. Energy 109 (2013) 275–282.

[3] G. Girishkumar, B. McCloskey, A.C. Luntz, S. Swanson, W. Wilcke, Lithium-AirBattery: Promise and Challenges, Journal of Physical Chemistry Letters 1(2010) 2193–2203.

[4] S.A. Freunberger, Y. Chen, Z. Peng, J.M. Griffin, L.J. Hardwick, F. Barde, et al.,Reactions in the Rechargeable Lithium-O2 Battery with Alkyl CarbonateElectrolytes, J. Am. Chem. Soc. 133 (2011) 8040–8047.

[5] S.A. Freunberger, Y. Chen, N.E. Drewett, L.J. Hardwick, F. Barde, P.G. Bruce, TheLithium-Oxygen Battery with Ether-Based Electrolytes, Angewandte Chemie-International Edition 50 (2011) 8609–8613.

[6] H. Wang, K. Xie, Investigation of oxygen reduction chemistry in ether andcarbonate based electrolytes for Li-O2 batteries, Electrochim. Acta 64 (2012)29–34.

[7] B.D. McCloskey, A. Speidel, R. Scheffler, D.C. Miller, V. Viswanathan, J.S.Hummelshoj, et al., Twin Problems of Interfacial Carbonate Formation inNonaqueous Li-O2 Batteries, Journal of Physical Chemistry Letters 3 (2012)997–1001.

[8] C.O. Laoire, S. Mukerjee, K.M. Abraham, E.J. Plichta, M.A. Hendrickson,Elucidating the Mechanism of Oxygen Reduction for Lithium-Air BatteryApplications, Journal of Physical Chemistry C 113 (2009) 20127–20134.

[9] C.O. Laoire, S. Mukerjee, K.M. Abraham, E.J. Plichta, M.A. Hendrickson,Influence of Nonaqueous Solvents on the Electrochemistry of Oxygen in theRechargeable Lithium-Air Battery, Journal of Physical Chemistry C 114 (2010)9178–9186.

[10] H. Jung, J. Hassoun, J. Park, Y. Sun, B. Scrosati, An improved high-performancelithium-air battery, Nature Chemistry 4 (2012) 579–585.

[11] Z. Peng, S.A. Freunberger, Y. Chen, P.G. Bruce, A Reversible and Higher-Rate Li-O2 Battery, Science 337 (2012) 563–566.

[12] S.S. Zhang, D. Foster, J. Read, Discharge characteristic of a non-aqueouselectrolyte Li/O2 battery, J. Power Sources 195 (2010) 1235–1240.

[13] V. Viswanathan, K.S. Thygesen, J.S. Hummelshoj, J.K. Norskov, G. Girishkumar,B.D. McCloskey, et al., Electrical conductivity in Li2O2 and its role indetermining capacity limitations in non-aqueous Li-O2 batteries, J. Chem.Phys. 135 (2011) 214704.

[14] C. Tran, X. Yang, D. Qu, Investigation of the gas-diffusion-electrode used aslithium/air cathode in non-aqueous electrolyte and the importance of carbonmaterial porosity, J. Power Sources 195 (2010) 2057–2063.

[15] C.K. Park, S.B. Park, S.Y. Lee, H. Lee, H. Jang, W.I. Cho, ElectrochemicalPerformances of Lithium-air Cell with Carbon Materials, Bulletin of the KoreanChemical Society 31 (2010) 3221–3224.

[16] M. Hayashi, H. Minowa, M. Takahashi, T. Shodai, Surface Properties andElectrochemical Performance of Carbon Materials for Air Electrodes ofLithium-Air Batteries, Electrochemistry 78 (2010) 325–328.

[17] Y. Gao, C. Wang, W. Pu, Z. Liu, C. Deng, P. Zhang, et al., Preparation of high-capacity air electrode for lithium-air batteries, Int J Hydrogen Energy 37 (2012)12725.

[18] Y. Zhang, H. Zhang, J. Li, M. Wang, H. Nie, F. Zhang, The use of mixed carbonmaterials with improved oxygen transport in a lithium-air battery, J. PowerSources 240 (2013) 390–396.

[19] V.S. Bryantsev, J. Uddin, V. Giordani, W. Walker, D. Addison, G.V. Chase, TheIdentification of Stable Solvents for Nonaqueous Rechargeable Li-Air Batteries,J. Electrochem. Soc. 160 (2013) A160–A171.

[20] C.J. Barile, A.A. Gewirth, Investigating the Li-O2 Battery in an Ether-BasedElectrolyte Using Differential Electrochemical Mass Spectrometry, J. Electro-chem. Soc. 160 (2013) A549–A552.

[21] V.S. Bryantsev, F. Faglioni, Predicting Autoxidation Stability of Ether- andAmide-Based Electrolyte Solvents for Li-Air Batteries, Journal of PhysicalChemistry A 116 (2012) 7128–7138.

[22] M.M. Ottakam Thotiyl, S.A. Freunberger, Z. Peng, P.G. Bruce, The CarbonElectrode in Nonaqueous Li-O2 Cells, J. Am. Chem. Soc. 135 (2013) 494–500.

[23] J. Lu, Y. Lei, K.C. Lau, X. Luo, P. Du, J. Wen, et al., A nanostructured cathodearchitecture for low charge overpotential in lithium-oxygen batteries, NatureCommunications 4 (2013) 2383.

[24] Z. Jian, P. Liu, F. Li, P. He, X. Guo, M. Chen, et al., Core-Shell-StructuredCNT@RuO2 Composite as a High-Performance Cathode Catalyst for Recharge-able Li-O2 Batteries, Angewandte Chemie-International Edition 53 (2014) 442–446.

[25] L. Cecchetto, M. Salomon, B. Scrosati, F. Croce, Study of a Li-air battery havingan electrolyte solution formed by a mixture of an ether-based aprotic solventand an ionic liquid, J. Power Sources 213 (2012) 233–238.

[26] F. Li, T. Zhang, H. Zhou, Challenges of non-aqueous Li-O2 batteries: electrolytes,catalysts, and anodes, Energy Environ. Sci. 6 (2013) 1125–1141.

[27] G. Zhao, Z. Xu, K. Sun, Hierarchical porous Co3O4 films as cathode catalysts ofrechargeable Li-O2 batteries, Journal of Materials Chemistry A 1 (2013) 12862–12867.

[28] Z.H. Wei, P. Tan, L. An, T.S. Zhao, A non-carbon cathode electrode for lithium-oxygen batteries, Appl. Energy 130 (2014) 134–138.