Journal of Power Sources 288 (2015) 20e25
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
High-performance anode-supported solid oxide fuel cell with impregnated electrodes D.A. Osinkin a, *, N.M. Bogdanovich a, S.M. Beresnev a, V.D. Zhuravlev b a b
Institute of High Temperature Electrochemistry, Ural Branch of the Russian Academy of Sciences, St. Academicheskaya 20, Yekaterinburg, 620137, Russia Institute of Solid State Chemistry, Ural Branch of the Russian Academy of Sciences, St. Pervomayskaya 91, Yekaterinburg, 620990, Russia
h i g h l i g h t s The powders for bi-layer Ni-cermet anode was produced by combustion synthesis. High-performance SOFC have been produced using simple methods. Impregnation of the electrodes has increased power of SOFC by seven times. The SOFC with impregnation electrodes revealed the power of 2.5 W cm2 at 900 C. Impregnation reduced resistances of all electrode reactions except the gas-diffusion.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 12 February 2015 Received in revised form 7 April 2015 Accepted 16 April 2015 Available online
The 61%NiO þ 39%Zr0.84Y0.16O1.92 (NiO-YSZ) and 56%NiO þ 44%Zr0.83Sc0.16Ce0.01O1.92 (NiO-CeSSZ) composite powders have been prepared using two-steps and one-step combustion synthesis, respectively. The Ni-YSZ anode substrate with a low level of electrical resistance (less than 1 mOhm cm) and porosity of about 53% in the reduced state was fabricated. The functional layer of the anode with the high level of electrochemical activity was made of NiO-CeSSZ. The single anode-supported solid oxide fuel cell with the bi-layer Ni-cermet anode, Zr0.84Sc0.16O1.92 film electrolyte and the Pt þ 3% Zr0.84Y0.16O1.92 cathode was fabricated. The power density and the UeI curves of the fuel cell at initial state and after impregnation of the cathode and anode by praseodymium and cerium oxides, respectively, have been measured at different temperatures. The maximum of power density of the initial fuel cell was 0.35 W cm2 at conditions of wet hydrogen (air) supply to the anode (cathode) at 900 C. After the electrodes were impregnated, the value of power density increased by seven times and was approximately 2.4 W cm2 at 0.6 V. It was suggested that after the electrodes impregnation the polarization resistance of the fuel cell was determined by the gas diffusion in the supported anode. © 2015 Elsevier B.V. All rights reserved.
Keywords: Ni-cermet supported SOFC Impregnation CeO2 Combustion synthesis
1. Introduction Solid oxide fuel cells (SOFCs) are perspective sources of electric power [1e3]. One way to increase the power output of SOFC is to reduce the thickness of the electrolyte which is the main contributor of ohmic resistance. However, as the thickness of the electrolyte decreases, its mechanical strength also decreases and, when the electrolyte thickness is less than 200e300 mm, it can no longer function as a support. Consequently, the SOFC development has advanced by using cathode-supported [4e6] or, more frequently,
* Corresponding author. E-mail address:
[email protected] (D.A. Osinkin). http://dx.doi.org/10.1016/j.jpowsour.2015.04.098 0378-7753/© 2015 Elsevier B.V. All rights reserved.
anode-supported [7e9] SOFCs with a thin film of electrolyte. The anode of the anode-supported SOFC should have the necessary level of thermal expansion coefficient (TEC), porosity, mechanical strength, electrical conductivity and electrochemical activity. These requirements can be achieved by using bi-layer anodes, the substrate of which possesses a high mechanical strength, porosity and electrical conductivity and the functional layer with a TEC value, possibly close to that of the electrolyte film, a high level of electrochemical activity and a level of porosity that is appropriate to the film electrolyte preparation. Applying different synthesis methodologies for the preparation of the anode powders and methods for formation of the anode layers [1,10] could give the possibility for obtaining the necessary anode layers characteristics, but the majority of these techniques is
D.A. Osinkin et al. / Journal of Power Sources 288 (2015) 20e25
complicated and requires specialized equipment. Combustion synthesis is the simplest and most perspective method to produce fine powders. Application of the combustion synthesis method implies the combustion of nitrate solutions with organic compounds such as: glycine, urea and others, which serve as a complexing agent and a fuel. Both simple and complex oxides and composite powders can be obtained using this method [11e16]. A high reaction rate and abundant gas emission prevent the growth of particles during the combustion process. A high level of chemical activity and the dispersion of powders obtained through this method make their use more profitable in the formation of SOFC components. Other simple way to improve the SOFC power characteristics is the impregnation (infiltration) of the electrodes. Impregnating the porous electrode by nano-sized particles of the metals [17,18] or/ and mixed conductors [19,20] leads to decreasing the electrode polarization resistance. This phenomenon is caused by several factors: improvement of the adsorption kinetic of the gas, increase of the triple-phase boundary length, improvement of the chargeetransfer reaction and etc. The purpose of this study is to reveal the possibility of producing the high-performance single SOFC by simple techniques. To achieve this goal, the combustion synthesis for preparation of the electrode powders, brush painting for applying the functional layer of the anode and electrolyte film, impregnation of the electrodes for improvement their electrochemical activity were used. 2. Experimental 2.1. Synthesis of powders Hexahydrates of Ni(NO3)2, Y(NO3)3, Sc(NO3)2, Ce(NO3)2 and ZrOCO3 * nH2O (purity 99%) were used as the initial substances, C2H5NO2 (glycine) was used as a fuel. The 61wt.%NiO þ 39wt.% Zr0.84Y0.16O1.92 (NiO-YSZ) composite powder have been prepared using two-steps combustion method. The YSZ powder was preliminary obtained by combustion synthesis. Then YSZ powder with the nickel nitrate and glycine solution was mixed and burned. The 56wt.%NiO þ 44wt.%Zr0.83Sc0.16Ce0.01O1.92 (NiO-CeSSZ) powder was prepared by one-step combustion synthesis. After combustion the obtained powders were annealed at 900 C in air for 5 h to remove carbon traces formed during the glycine combustion. Then NiO-YSZ and NiO-CeSSZ powders were grinded in a planetary ball mill PM100, Retsch for 2 h. The graphite powder (specific surface area is 12.4 m2 g1) in the amount of 10 wt.% was added to the NiO-YSZ as a pore-former. The graphite powder was obtained by grinding an industrial graphite electrode (with a purity of 99.9%) in a mill. Then the graphite- NiO-YSZ powder was grinded in a mill for 30 min. 2.2. Samples The samples for measuring the TEC and electrical conductivity were prepared in the shape of bars by a dry isostatic pressing at 150 МPa cm2 for 10 s. The thin electrodes (thickness is 20 mm) for measuring the polarization resistance were applied by the brush painting onto the dense commercial Zr0.84Y0.16O1.92 plates. All samples and electrodes were sintered for 2 h in air atmosphere at 1400 C. 2.3. Fabrication of SOFC The NiO-YSZ anode substrate was prepared in the shape of a disc with 12 mm diameter and 1 mm thickness by dry isostatic pressing at 150 МPa cm2 for 10 s and sintered for 2 h in air at 1400 C. The slurry for functional layer was produced by mixing NiO-CeSSZ
21
powder with polyvinylbutyral and isopropyl alcohol. The slurry was applied by the brush painting onto one side of the sintered NiO-YSZ disk. When the slurry dried (thickness of functional layer about 20 mm) the first layer of the commercial Zr0.84Sc0.16O1.92 (SSZ) electrolyte powder was applied by painting onto the functional layer. Sintering of the half-cell was performed at 1390 C with isothermal exposure for 2 h in air. After sintering the second SSZ layer was applied onto the first SSZ layer and then sintered at 1360 C with isothermal exposure for 5 h in air. The total thickness of the SSZ electrolyte was approximately 30 mm. The self-made Pt þ 3% Zr0.84Y0.16O1.92 (Pt-YSZ) cathode paste was applied onto SSZ electrolyte and sintered at 1100 C for 1 h. Thickness of the cathode was about 30 mm. The photos of anode-supported half-cell, single SOFC and sketch of SOFC are presented in Fig. 1. To reduce the electrodes polarization resistances they were impregnated by the saturated water solutions of praseodymium (cathode) and cerium (anode) nitrates. After thermal decomposition of the nitrates (at 600 C in wet hydrogen for 1 h) the weight of the electrode about 20% increased. 2.4. Measurements The method of lowetemperature nitrogen adsorption (by means of the SORBI N4.1 device) was used for measuring the Nicermet powders specific surface area. Before the measurement of the specific surface the degasification of powder in a helium atmosphere at 200 C for 1 h was done. The powder diffractograms were obtained using the X-ray Rigaku D/MAX-2200VIPC diffractometer with Ni-filtered CuKa radiation in the range of 25 2q 70 at angular scanning rate of 2 min1. For micrographs the epoxy was used for filled the samples under for vacuum with subsequent polishing in the Struers Labopol device. X-ray spectral analysis was taken using the scanning electron microscope Jeol JSM 5900LV with INCA energy 200. The microphotographs were taken with the scanning electron microscope TESCAN MIRA 3 LMU. The TEC measurements were carried out by the automated set with the programmable “Thermodat-16” thermo regulator, quartz dilatometer and “Tesatronic TT-80” digital meter with the proper TESA GT 21HP measuring probe in the temperature range of 100e900 C at the constant heating rate of 2 C min1 in air. The porosity (ε) of Ni-cermet bars was calculated as follows:
ε¼ 1
m *100 ; V*ðx1 *r1 þ x2 *r2 Þ
(1)
where m and V are the mass and volume of the sample, x and r are the mass and density of the fraction, respectively. The H2 of high purity not less than 99.9 vol.% was used for the fuel gas mixtures preparation. The H2O content in hydrogen was set by the temperature of the water evaporator, over which the H2 flow passed. The air compressor for supply of the ambient air as oxidizer on the cathode was used. Gases flow rate at the anode and cathode was 5 l h1. The electrical conductivity measurement was carried out by the DC four-probe method using the electrochemical interface EI-1287 (Solartron Instruments Inc.) in the current range of 0e750 mA cm2 (with step of 50 mA cm2) at 700e900 C. The SOFC electrochemical characteristics and thin electrodes were studied by means of impedance spectroscopy in the frequency range of 105e102 Hz using Frequency Response Analyzer FRA-1260 with the electrochemical interface EI-1287 (Solartron Instruments Inc.). AC measurements were carried out at galvanostatic mode with AC amplitude of 20 mA, the number of points per decade (log scale) was 20 at 700e900 C. The electrochemical cell was connected to an electrochemical interface by two-electrode four-wire mode
22
D.A. Osinkin et al. / Journal of Power Sources 288 (2015) 20e25
Fig. 1. Photo of half-cell and SOFC (left) and sketch of SOFC (right).
which permits the exclusion of the impedance of current-supplying cables from the overall impedance. Pt mesh and wires were used as current collectors (Fig. 1). 3. Results and discussion 3.1. Characterization of powders and samples The diffractograms of the NiO-YSZ and NiO-CeSSZ powders are illustrated in Fig. 2. An absence of any diffraction peaks specific to the components of the reaction demonstrated that the chosen synthesis conditions are sufficient for a complete reaction. It can be seen from Fig. 3 that the powders obtained by the combustion method have loose, sponge-like structure without clear boundaries between particles. The specific surface area of powders is 6.2 and 13.6 m2 g1 for NiO-YSZ and NiO-CeSSZ, respectively. The data concerning the porosity (ε) of bars after sintering in oxidizing condition and after reduction and the TEC values in the oxidized state is given in Table 1. The NiO-YSZ sample with 10 wt.% of graphite is seen to have a noticeably higher porosity value than that of the NiO-CeSSZ. The lower values of porosity and TEC of the NiO-CeSSZ samples should have a positive effect to fabrication of the film electrolyte on the NiO-CeSSZ functional layer. The temperature dependence of the electrical resistance for the Ni-YSZ and Ni-CeSSZ samples in wet hydrogen is presented in Fig. 4. For comparison, the electrical resistance of pure Ni wire (at least 99.99 wt.%) [21] is plotted on the figure. The electrical resistance of
Ni-YSZ sample is approximately 4 times lower than that for NiCeSSZ, despite their greater porosity. This was achieved by slightly higher content of the nickel oxide and by applying twosteps method of the NiO-YSZ powder preparation which improves contacts between Ni particles after reduction [21]. The polarization resistance of thin electrodes before and after impregnation is given in Fig. 5. The Ni-CeSSZ electrode's activity is seen to be higher by the order of magnitude than that of the Ni-YSZ (the polarization resistance is lower) at all temperatures studied. The high activity of the Ni-CeSSZ electrodes is achieved by a joint combustion of solutions during the NiO-CeSSZ powder preparation, which results in the formation of a well-developed triple-phase boundary. In addition, the usage of the CeSSZ electrolyte with the higher level of ionic conductivity in comparison to YSZ also leads to increased electrochemical activity. After impregnation of the Nicermet electrodes by cerium oxide their polarization resistance reduced approximately by 2 orders of magnitude at 900 C (at 700 C approximately by 3 orders of magnitude) and was about 0.03 (0.06) for Ni-CeSSZ and 0.3 (1.33) Ohm cm2 for Ni-YSZ at 900 (700) C, respectively. The impregnation of Pt-YSZ cathode by praseodymium oxide also resulted in decreasing the polarization resistance approximately by 1 order of magnitude at all temperatures studied. It should be noted that after impregnation the cathode polarization resistance remained high in comparison with high-activity cathodes [22]. The activation energies of the specific polarization resistance are listed in Table 2. After impregnation the activation energy of the Ni-CeSSZ and Pt-YSZ electrodes decreased from 110 and 132 to 41 and 110 kJ mol1, respectively. It can be suggested that the reaction mechanism changes. At the same time, the activation energy of the Ni-YSZ electrode has not changed and was about 70 kJ mol1. Authors of [23,24] also noted the decrease of the activation energy of the electrode polarization resistance after impregnation, nevertheless, in Ref. [22] almost has no change in activation energy after impregnation was reported. 3.2. Half-cell and SOFC investigation
Fig. 2. Diffractograms of NiO-YSZ (top) and NiO-CeSSZ (bottom) powders.
A cross section microphotograph of the Ni-YSZ/Ni-CeSSZ/SSZ half-cell is presented in Fig. 6. The absence of a distinct boundary between the Ni-YSZ anode substrate and Ni-CeSSZ functional layers can be explained by their similar nature, low thickness of the functional layer and the penetration of some amount of slurry into pores of the substrate during brushing procedure. Even at the higher magnification the functional layer remains unseen (B). It is marked only by the presence of scandium in the area adjacent to the electrolyte (C), which concentration decreases with distance from the electrolyte layer. From the micrograph one can determine the thickness of the SSZ electrolyte, which is about 30e35 mm. The
D.A. Osinkin et al. / Journal of Power Sources 288 (2015) 20e25
23
Fig. 3. Microphotographs of NiO-YSZ (left) and NiO-CeSSZ (right) powders.
Table 1 Porosity (ε) of samples before and after reduction and TEC in the oxidized state. Sample
εox (%)
εred (%)
TEC 106 ( C1)
NiO-YSZ NiO-CeSSZ
41 21
53 35
12.81 ± 0.03 12.04 ± 0.01
presence of the macro pores (size of about 10 mm) in Ni-YSZ substrate should also be mentioned. It can be explained by the NiO-YSZ powder microstructure (Fig. 3). The air is captured at pressing these powders, which does not allow making the electrode with isotropic microstructure. Since the electrochemical reaction is substantially localized in the thin electrode layer in contact with the electrolyte, the presence of macro pores in Ni-YSZ anode substrate should not negatively impact the characteristics of SOFC. The UeI and the SOFC power dependence at different temperatures before and after the electrodes impregnation at the wet hydrogen (air) supply to the anode (cathode) is given in Fig. 7. The characteristics of the initial SOFC were very low at 700 C (power density is not visible in the figure). After the cathode impregnation, the SOFC power increased and amounted to about 0.5 W cm2. After impregnation of both SOFC electrodes the maximum power
Fig. 5. Temperature dependence of the polarization resistance before and after (empty symbols) impregnation. (Ni-cermet in wet H2, Pt-YSZ in air. Dashed lines were added to aid visual perception).
was about 1.3 W cm2. At 900 C the power of the initial cell was about 0.4 W cm2. After impregnation of the cathode the SOFC power increased and amounted to about 1 W cm2. After impregnation the both electrodes the value of power about 2 W cm2 was obtained at 0.75 V, but as it is seen from Fig. 7, the maximum SOFC power was estimated after fitting the power curve, it was 2.5 W cm2 at 0.55 V. As it was mentioned earlier, the significant changes in the SOFC power after impregnation of the electrodes by mixed-conducting oxide is most likely associated with the changes of the electrode reaction mechanisms and the rate determining steps. The testing SOFC represents a complex electrochemical system, unlike the cells
Table 2 The activation energies (Ea) of the specific polarization resistance. Electrode
Fig. 4. Temperature dependence of the electric resistance in wet H2. (Dashed lines were added to aid visual perception).
Ni-CeSSZ Ni-YSZ Pt-YSZ
Ea (kJ mol1) Before impregnation
After impregnation
110 ± 3 70 ± 2 132 ± 4
41 ± 4 69 ± 1 110 ± 6
24
D.A. Osinkin et al. / Journal of Power Sources 288 (2015) 20e25
demonstrated in Fig. 8. The imaginary part peak of the initial SOFC is observed at 103 Hz. After impregnation of the cathode the maximum was shifted to 10 Hz and after the cathode and anode impregnation - to 0.1 Hz. Intersection of the high-frequency part of the spectra with the abscissa axis corresponds, mainly, to the resistance of the SSZ film. The resistivity of SSZ electrolyte calculated from impedance spectra before and after impregnation of the cathode was about 16e17 Ohm cm2 (at film thickness 30 micron) at 900 C. This value is significantly overstated [1,3]. After impregnation of the cathode and anode, the resistivity becomes equal to 5 Ohm cm2, which corresponds to the literature data. This may due to the area of boundary of the anode with SSZ film. Since the cerium and presidium oxides exhibit ionic and electronic conductivity, after impregnation the contact area of electrode/electrolyte increases, Fig. 6. Cross-section microphotographs of the Ni-YSZ/Ni-CeSSZ/SSZ half-cell at different magnification (A and B) and the result of the X-ray microanalysis of scandium concentration (C).
with symmetrical electrodes. An analysis of the impedance spectra of such SOFC and the definition of the partial resistances is quite difficult by both NLLS [25] and DRT [26,27] methods, therefore in this paper we give only a qualitative evaluation of the received data. The impedance spectra and the imaginary part of the SOFC at 900 C before and after impregnation of the electrodes are
Fig. 7. Power density and IeU dependencies of SOFC at different temperatures. (wet hydrogen (air) supply to the anode (cathode)).
Fig. 8. Impedance spectra of SOFC and their imaginary parts at 900 C before and after impregnation of the electrodes.
D.A. Osinkin et al. / Journal of Power Sources 288 (2015) 20e25
and SSZ resistivity becomes close to the true value. Usually, on the impedance spectra of such object at high frequencies the reaction of the charged transfer, at middle frequencies adsorption or/and surface diffusion process and at low e gas diffusion reaction is registered [28e33]. We assumed that after the electrodes impregnation the SOFC resistance can be determined by the gas diffusion resistance (in the electrode pores and/or under its surface) as the resistance of other stages of the electrode reaction becomes insignificant. 4. Conclusions The high-performance SOFC can be produced using simple methods. The anode-supported SOFC being fabricated revealed the power density of 1.3 W cm2 and 2.5 W cm2 at 700 and 900 C, respectively. It was suggested that the impregnation of the electrodes by mixed-conductor oxides significantly reduced the contribution of all stages of the electrode reaction except the gasdiffusion stages. The SOFC being investigated can be easily improved by reducing the anode and electrolyte film thickness and using more active cathode. Acknowledgments This work has been partly done using facilities of the shared access center “Composition of compounds” IHTE, UB RAS. The research was partly supported by the Russian Foundation for Basic Research (N 14-08-31030). References [1] J. Fergus, R. Hui, X. Li, D.P. Wilkinson, J. Zhang, Solid Oxide Fuel Cells: Materials Properties and Performance, CRC Press, New York, 2008. [2] S.P. Jiang, Y. Yan, Materials for High-temperature Fuel Cells, Wiley-VCH, 2013. [3] B. Viswanathan, M.A. Scibioh, Fuel Cells Principles and Applications, CRC Press, New York, 2008. [4] L. Zhao, X. Ye, Z. Zhan, J. Power Sources 196 (2011) 6201e6204. [5] S.M. Beresnev, O.F. Bobrenok, B.L. Kuzin, N.M. Bogdanovich, A.A. Kurteeva,
25
D.A. Osinkin, G.K. Vdovin, D.I. Bronin, Rus. J. Electrochem. 48 (2012) 969e975. [6] M. Chen, J. Luo, K.T. Chuang, A.R. Sanger, Electrochim. Acta 63 (2012) 277e286. [7] A.E. Richards, N.P. Sullivan, Fuel cells 13 (2013) 470e475. [8] T.S. Li, W.G. Wang, H. Miao, T. Chen, C. Xu, J. Alloys Comp 495 (2010) 138e143. [9] A.A. Kurteeva, S.M. Beresnev, D.A. Osinkin, B.L. Kuzin, G.K. Vdovin, V.D. Zhuravlev, N.M. Bogdanovich, D.I. Bronin, A.A. Pankratov, I.Yu Yaroslavtsev, Rus. J. Electrochem. 47 (2011) 1381e1388. [10] H.J. Kitchen, S.R. Vallance, J.L. Kennedy, N. Tapia-Ruiz, L. Carassiti, A. Harrison, A.G. Whittaker, T.D. Drysdale, S.W. Kingman, D.H. Gregory, Chem. Rev. 114 (2014) 1170e1206. [11] W. Yan, H. Wen, Y. Chen, Y. Wang, Y. Liu, J. Power Sources 277 (2015) 76e83. [12] A.S. Mukasyan, P. Epstein, P. Dinka, Proc. Combust. Inst. 31 (2007) 1789e1795. [13] N. Dai, Z. Wang, T. Jiang, J. Feng, W. Sun, J. Qiao, D. Rooney, K. Sun, J. Power Sources 268 (2014) 176e182. [14] S.T. Aruna, A.S. Mukasyan, Solid State Mater. Sci. 12 (2008) 44e50. [15] E. Chinarro, J.R. Jurado, M.T. Colomer, J. Eur. Ceram. Soc. 27 (2007) 3619e3623. [16] W. Sun, Z. Shi, J. Qian, Z. Wang, W. Liu, Nano Energy 8 (2014) 305e311. [17] Y. Zhou, C. Yuan, T. Chen, X. Meng, X. Ye, J. Li, S. Wang, Z. Zhan, J. Power Sources 267 (2014) 117e122. [18] S. Huang, S. Feng, H. Wang, Y. Li, C. Wang, Int. J. Hydrogen Energy 36 (2011) 10968e10974. [19] N. Xu, X. Li, X. Zhao, H. Zhao, K. Huanga, Electrochem. Solid State Lett. 15 (2012) B1eB4. [20] S.P. Jiang, S. Zhang, Y.D. Zhen, W. Wang, J. Am. Ceram. Soc. 88 (2005) 1779e1785. [21] D.A. Osinkin, D.I. Bronin, S.M. Beresnev, N.M. Bogdanovich, V.D. Zhuravlev, G.K. Vdovin, T.A. Demyanenko, J. Solid State Electrochem 18 (2014) 149e156. [22] P. Hjalmarsson, M. Mogensen, J. Power Sources 196 (2011) 7237e7244. [23] T. Chen, M. Liu, C. Yuan, Y. Zhou, X. Ye, Z. Zhan, C. Xia, S. Wang, J. Power Sources 276 (2015) 1e6. [24] A.P. Khandale, S.S. Bhoga, J. Power Sources 268 (2014) 794e803. [25] B.A. Boukamp, Solid State Ionics 20 (1986) 31e44. [26] A. Leonide, SOFC Modelling and Parameter Identification by Means of Impedance Spectroscopy, KIT Scientific Publishing, 2010. e, J. Appl. Elec[27] H. Schichlein, A.C. Müller, M. Voigts, A. Krügel, E. Ivers-Tiffe trochem 32 (2002) 875e882. [28] S.P. Jiang, S.P.S. Badwal, J. Electrochem. Soc. 144 (1997) 3777e3784. [29] N. Dai, Z. Lou, Z. Wang, X. Liu, Y. Yan, J. Qiao, T. Jiang, K. Sun, J. Power Sources 243 (2013) 766e772. [30] P. Kim, D.J.L. Brett, N.P. Brandon, J. Power Sources 189 (2009) 1060e1065. [31] D. Chen, C. Chen, F. Dong, Z. Shao, F. Ciucci, J. Power Sources 250 (2014) 188e195. [32] D.A. Osinkin, B.L. Kuzin, N.M. Bogdanovich, Rus. J. Electrochem 45 (2009) 483e489. [33] H.P. Dasari, S.Y. Park, J. Kim, J.H. Lee, B.K. Kim, H.J. Je, H.W. Lee, K.J. Yoon, J. Power Sources 240 (2013) 721e728.