Synthetic Metals 200 (2015) 117–122

Graphite nanosheets/nanoporous carbon black/cerium oxidenanoparticles as an electrode material for electrochemical capacitors

Mahdi R. Sarpoushi a, Mahdi Nasibi a,b,*, Mohammad Reza Shishesaz a,Mohammad Ali Golozar c, Hamid Riazi c

a Technical Inspection Engineering Department, Petroleum University of Technology, Abadan 63187-14331, IranbHealth, Safety and Environment (HSE) Engineering Office, National Iranian Oil Refining and Distribution Company (NIORDC), Yazd, IrancMaterials Science and Engineering Department, Isfahan University of Technology, Isfahan 84156-83111, Iran

A R T I C L E I N F O

Article history:Received 1 May 2014Received in revised form 25 December 2014Accepted 27 December 2014Available online xxx

Keywords:Electronic materialsElectrochemical techniquesNanostructuresMicrostructureElectrochemical properties

A B S T R A C T

In this study, effect of morphology and pore size distribution on physicochemical properties of graphitenanosheets (GNSs)/nanoporous carbon black (NCB)/CeO2 nanoparticle electrodes were investigated in3 M NaCl electrolyte. 90:00:00:10 (GNS:NCB:CeO2:PTFE) electrodes show a flat and smooth surface and75:00:15:10 electrodes showed micro pores which opened on surface. In the presence of NCB particles,electrodes like 25:50:15:10 showed the macro, micro and nano pores, simultaneously. Total charge (q�T) of1.73, 37.88, 208.33 C g�1cm�2 and outer charge to total charge (q�O/) of 0.53, 0.43 and 0.24 were obtainedfrom 90:00:00:10, 75:00:15:10 and 25:50:15:10 electrodes, respectively. It is concluded that introducingnarrower and deeper pores on surface of electrodes increases the charge storage capability and decreasesthe current response and power delivering capability. Flat surface of GNSs showed 3.4 F/g and alsoexhibited good capacitance retention of more than 80%. Additionally, nanoporous structures increasedthe capacitance of 25:50:15:10 electrodes up to 16.2 F/g at 10 mV s�1 in 3 M NaCl electrolyte.

ã 2014 Elsevier B.V. All rights reserved.

Contents lists available at ScienceDirect

Synthetic Metals

journal homepage: www.elsevier .com/ locate /sy nmet

1. Introduction

To produce and store energy storage would dramaticallyincrease in future, however the problem of ensuring power qualityis our responsibility. Energy storage is being widely regarded asone of the potential solutions to deal with variations of variablerenewable electricity sources, and it is the key to unlocking thedoor of renewable energy. Among different energy storagesystems, ultracapacitors are recognized as attractive energystorage devices to satisfy the current and future needs [1].Electrochemical capacitors have received a great attention inrecent years for their high power capability, life cycle (more than500,000 cycles at 100% depth of discharge) high energy efficiencyranging from 85% up to 98%, and their good reversibility againstother power storage devices [2,3].

With respect to electrode materials there are three maincategories: carbon base, transition metal oxides and conductivepolymers [4]. Almost, common electrodes proposed for

* Corresponding author at: Health, Safety and Environment (HSE) EngineeringOffice, National Iranian Oil Refining and Distribution Company (NIORDC), Yazd, Iran.Tel.: +98 911 3708480; fax: +98 35 35253091.

E-mail addresses: m_nasibi@yahoo.com, mahdi.nasibi@gmail.com (M. Nasibi).

http://dx.doi.org/10.1016/j.synthmet.2014.12.0310379-6779/ã 2014 Elsevier B.V. All rights reserved.

electrochemical double layer capacitors contain carbon basematerials, such as activated carbons, carbon black (CB), carbonaerogels, carbon nanotubes (CNTs), graphene, ordered mesoporouscarbons and hierarchical porous carbons, that have different specificsurface area and pore size distributions [5,6]. Additionally, conduc-tivity, porosityand morphologyare shape dependent and the uniquecharacteristics of these materials enhance their electrochemicalperformance [7–15]. It is reported that porous materials andnanoporous structures show a higher specific surface area. On theother hand, flake and sheet like structures almost show higheraccessibility but decreased specific surface area and they mayincrease the charge transfer rate during charging and dischargingprocesses. On these structures, almost always, less active surfacesand/or internal surfaces are prevented from forming double layer[16,17]. Different morphologies have been reported for carbon basematerials; for example, carbon blacks (CBs), CNTs and graphenenanosheets exhibit spherical, tubular and lamellar structures,respectively. On the other hand, oxides and hydroxides of transitionmetals like Ru [18], Co [19], Zr [20], Mn [21] and Ce [22], andconducting polymers, like polyaniline (PANI) [23] and polypyrrole[24] are of candidate materials for supercapacitors which storeenergy through pseudo mechanism. These materials possessmultiple oxidationstatesthatenable rich redoxandFaradicreactionson surface of carbon base materials.

Fig. 1. Schematic illustration of different pores prepared on the electrodes.

118 M.R. Sarpoushi et al. / Synthetic Metals 200 (2015) 117–122

The aim of this study is to investigate these morphologicaleffects on physicochemical properties of carbon containing nano-composites as electrode materials for electrochemical capacitors.In this respect, mechanical pressing was used to prepare electro-des. The electrochemical properties of the electrodes wereevaluated by cyclic voltammetry, electrochemical impedancespectroscopy and scanning electron microscopy.

2. Experimental

2.1. Material and electrode preparation

High purity (>99%) cerium oxide (CeO2) (<30 nm) andpolytetrafluoroethylene (PTFE) (<2 mm) as a polymeric binderwere purchased from Aldrich, USA. Multi-layered graphite nano-sheets (GNSs) with a purity of 98.5% purchased from graphenesupermarket, USA and nanoporous carbon black (NCB) (<10 nm inpore diameter and <2 mm in particle size) purchased fromDegussa, Germany. All the other chemicals used in this studywere purchased from Merck, Germany. In order to prepare theelectrodes, the mixture containing different wt% of CeO2 nano-particles, GNSs, NCB particles and 10 wt% PTFE were well mixedwith propeller stirrer at 1200 rpm for 30 min in ethanol bath toform a paste. Prepared mixtures were then sonicated for 90 min.Sonication process was performed at frequency of 20 kHz with aninlet ultrasound power of around 0.5 W mL�1 (Hielscher Ultra-sound Technology, Germany). Then the prepared pastes weredried, powdered and pressed onto a 316 L stainless steel plate(5 �107 Pa) which served as a current collector (surface area was1.22 cm2). The typical mass of used active material was 30 mg. Theused electrolyte was 2 M NaCl.

2.2. Characterization

Electrochemical behavior of prepared electrodes was charac-terized using cyclic voltammetry (CV) and electrochemicalimpedance spectroscopy (EIS) techniques. The electrochemicalmeasurements were performed using an Autolab (Netherlands)potentiostat Model PGSTAT 302N. CV tests were recorded at scanrates of 10, 20, 30, 50 and 100 mV s�1 and EIS measurements werealso carried out in the frequency range of 100 kHz to 0.01 Hz atopen circuit potential with ac amplitude of 10 mV. The specificcapacitance C (F g�1) of prepared electrodes was determined byintegrating either the oxidative or reductive parts of cyclicvoltammogram curves to obtain the voltammetric charge Q (C).This charge was divided by mass of active material m (g) ofelectrode and the width of potential window of the cyclicvoltammogram DE (V), i.e., C = Q/(mDE) [21]. Scanning electronmicroscopy (TESCAN, USA) was used for better understanding theeffect of surface morphology and its nature on physicochemicalproperties of prepared electrodes.

3. Result and discussion

Fig. 1 shows a schematic of surface morphology of electrodesfabricated. In the absence of NCB and CeO2 nanoparticles, GNSsdeposit on each other and make a flat and smooth surface on theelectrode. This type of morphology may act like a barrier againstthe electrolyte diffusion bath through the electrode material, andmay decrease the accessible surface area of the electrode. Asshown in Fig. 1, addition of CeO2 nanoparticles can crack distancesbetween the nanosheets and will make micro pores on the surfaceof the electrode. These micro pores are almost narrow and deep(about 1 mm in diameter and few mm in depth). On the other hand,transition metal oxides like CeO2 are non-conductive and electro-active materials that will take part in Faradic reactions. Addition of

nanoporous CB particles (<2 mm in diameter) into the electrodematerial can act as a mixer due to its large particle size andincreases distances between GNSs and re-arranges them indifferent directions (Fig. 1). These macro pores are almost lessthan 5 mm in diameter and about 10 mm in depth, which canincrease the specific surface area and the ability of electrolyte todefuse. So, in the presence of GNS, NCB and CeO2 nanoparticles,surface topography revealed by SEM images suggest that threetypes of macro, micro and nano pores formed on the surface. Asshown in Fig. 1, macro pores are made between the GNSs and NCBparticles, micro pores are made between the GNS and CeO2

nanoparticles and the surface of carbon black containing electro-des is nanoporous. As shown in Fig. 2, SEM images obtained fromthe surface of 90:00:00:10 (GNS:NCB:CeO2:PTFE) electrodesconfirm the flat and smooth surface of the electrode (Fig. 2(a)).SEM images obtained from 75:00:15:10 electrodes confirm thepresence of micro pores which open on the surface of the electrode(Fig. 2(b)), and NCB and CeO2 containing electrodes like25:50:15:10 shows macro, micro and nano pores, simultaneously(Fig. 2(c) and (d)). These will change the specific surface area,electrical resistance, diffusion characteristics and therefore, theycan alter the energy storage and rate capacity of the preparedelectrodes. Hence, these electrodes were selected for furtherinvestigations.

Fig. 3shows the second CV of prepared electrodes at scan rate of10 mV s�1 in 3 M NaCl electrolyte. CV curves exhibit a rectangularshaped profile which is characteristic of ideal capacitive behavior[25] and show almost potential-independent double layer capaci-tance over relatively wide range of potential window. Addition ofCeO2 nanoparticles into the electrode (75:00:15:10) increased thecharge stored on the electrode slightly (Fig. 3) which may be due toFaradic reactions and increase of specific surface area. The principlereaction involved in charging and discharging processes of ceriumdioxide in an aqueous electrolyte can be described by Reaction (1)[?]:

Fig. 2. Scanning electron microscopy images obtained from (a) 90:00:00:10, (b) 75:00:15:10, (c) 25:50:15:10 and (d) nanoporous surface of carbon black particles.

M.R. Sarpoushi et al. / Synthetic Metals 200 (2015) 117–122 119

CeIVO2 + l+ + e�! CeIIIOOl (1)

where l denotes K+ or H+. Additionally, two mechanisms can beproposed for supercapacitive charge storage in the presence ofCeO2 particles. The first mechanism is based on the intercalation/extraction of protons or alkali cations into the oxide particles(denoted as Reaction (2)), whereas the second mechanism involvesthe surface adsorption/desorption of proton or alkali cationsprobably (denoted as Reaction (3)):

CeO2 + M+ + e� = CeOOM (2)

and

CeO2)surface + M+ + e�= (CeOOM)surface (3)

where M+ denotes as K+ or H3O+.

Fig. 3. Cyclic voltammetry curves obtained from 90:00:00:10, 75:00:15:10 and25:50:15:10 electrodes at 10 mV s�1 in 3 M NaCl electrolyte.

But introducing narrower and deeper pores on the surface of75:00:15:10 electrodes decrease electrolyte ability to diffuse;therefore, this low capacitance may be due to the high non-accessible surface area in presence of nanoparticles.

In the presence of NCB particles charge storage ability of25:50:15:10 electrodes increases more (Fig. 3) that may be due tothe porous structure of used CB particles and the random orientationof GNSs. Obtained Nyquist diagrams are plotted in Fig. 4. Allimpedance curves described the locus of lumped and distributedconstant-types in low and high frequency ranges, respectively(Fig. 5). As shown in Fig. 4, addition of CeO2 into the electrode shiftsthe distributed to lumped type transition frequency to the lowerfrequency sides and this will increase if the NCB particles areintroduced into the electrode. These changes confirm that thecurrent distribution becomes non-uniform and the currentresponses at end potentials are lower [26]. On the other hand,decreasingtheabsolutevalueof impedanceinlumpedconstant-typerange (Fig. 4) of 25:50:15:10 electrode confirms that the accessiblesurface area of inner wall of electrodes increases in the presence ofmacro (CeO2 particles) and nano pores (NCB particles) [26].Therefore, it is expected that in the presence of NCB particles,current response and electrolyte diffusion through the electrodewillbe lower and the accessible inner wall surface area will be increasedandthe currentdistributionontheelectrodeinthepresenceofsuchananoporous morphology will be non-uniform. The point ofintersecting with real axis of Nyquist curves (Fig. 5) in range ofhigh frequency is the equivalent series resistance (ESR). It indicatesthe total resistance of electrode, the bulk electrolyte resistance andthe resistance at electrolyte/electrode interface (Fig. 5) [27]. ESRsobtained from 90:00:00:10 and 25:50:15:10 electrodes are identical(Fig.4)andhigherthantheESRobtainedfrom75:00:15:00electrode.This may be due to addition of transition metal oxide nanoparticles

Fig. 4. Nyquist plots obtained from 90:00:00:10, 75:00:15:10 and 25:50:15:10 elec-trodes in 3 M NaCl electrolyte.

Table 1Electrochemical and operational characteristics obtained from different electrodein 3 M NaCl solution.

Electrode type(GNF:CB:CeO2:PTFE)

90:00:00:10 75:00:15:10 25:50:15:10

C (F/g) (10 mV s�1) 3.4 4.5 16.2C (F/g) (100 mV s�1) 2.9 2.1 4.8C (F/g) (200 mV s�1) 2.5 2 4.3Cfading (10 ! 200 mV s�1) 73% 56% 26%q�T (C g�1 cm�2) 1.73 37.88 208.33q�O (C g�1 cm�2) 0.91 16.32 49.95q�O=q

�T 0.53 0.43 0.24

Cretention (200th cycle, 10 mV s�1) 85% 81% 76%

120 M.R. Sarpoushi et al. / Synthetic Metals 200 (2015) 117–122

on graphene sheets orientations on prepared electrodes. Therefore,two counter acting parameters will act simultaneously as the CeO2

content of the electrode increases: increasing the specific surfacearea and increasing the electrical resistance of the electrodes.

Finally, 90:00:00:10 and 75,00:15:10 electrodes show acapacitance of as high as 3.4 and 4.5 (F/g) at scan rate of 10 mV s�1

in 3 M NaCl electrolyte (Table 1). NCB utilization during electrodefabrication increases the specific capacitance up to 16.2 (F/g) at10 mV s�1 in 3 M NaCl electrolyte.

As expected for porous electrodes, increasing the charge/discharge rate almost decreases the charge storage capability ofthe electrode due to an increase in the non-accessible inner surfacearea of the pores and the efficiency which increases the energy lossat high sweep rates. Therefore, CV curves obtained from preparedelectrodes at various scan rates between 10 and 100 mV s�1 wereobtained and depicted in Figs. 6 and 7. As the scan rate increasesthe current vs. potential relation of CV deviates from its classicalsquare waveform, expected for a pure capacitor. As discussed by

Fig. 5. Schematic of electrochemical impedance of electrochemical capacitors,lumped constants equivalent circuit and distributed constants equivalent circuit.

some researchers this may be due to the resistance effects downthe pores and diffusion characteristics of the electrolyte [28]. Asthe sweep rate increases, energy loss will increase and the chargestored on the electrode decreases and causes the capacitance todecrease. Figs. 6 and 7 confirm that the scan rate has a strongereffect on charge storage ability as the NCB content of the electrodesincreases due to the presence of narrower pores and lowerdiffusion of electrolyte ions within these nanopores. According toprevious studies [29,30], carbon base materials with porousstructures are attractive materials for supercapacitors. Large poresare beneficial to provide favorable and quicker pathway for smoothelectrolyte transportation, whereas the small pores are useful inincreasing the specific surface area. But as the pores dimensionshifts to nano-sized dimensions, this makes it difficult forelectrolyte ions to diffuse quickly. From Fig. 8, it can be seen thatCeO2 and NCB free samples are beneficial for ion transportation athigh scan rates, so high capacitance retention percent of 74% couldbe observed at the scan rate of 200 mV s�1. CeO2 containingelectrodes (75:00:15:10) and, CeO2 and NCB containing samples(25:50:15:10) show low capacitance retention percent of 44% and27% at the scan rate of 200 mV s�1, respectively. This confirms thereverse effect of nanoporous structures on capacitance retentionsof nanoporous electrodes for supercapacitors at high scan rates. Inorder to gain quantitative information on the charge storagecapability and power delivering ability of prepared electrodes,obtained voltammograms were analyzed as a function of scan rate,according to procedure reported by Ardizzone et al. [31]. The scanrate dependence of the capacitance can be related to the lessaccessible surface area which becomes excluded as the ratereaction is enhanced. In charge and discharge cycles, the totalcharge can be written as a sum of an inner charge from the lessaccessible reaction sites and an outer charge from the more

Fig. 6. Cyclic voltammetry curves obtained from 75:00:15:10 electrodes at differentscan rates.

Fig. 9. (a) Extrapolation of q to v = 0 from the q�1 vs. v0.5 plot given the total chargeand (b) extrapolation of q to v = 1 from the q vs. v�0.5 plot given the outer charge fordifferent electrodes.

Fig. 7. Cyclic voltammetry curves obtained from 25:50:15:10 electrodes at differentscan rates.

M.R. Sarpoushi et al. / Synthetic Metals 200 (2015) 117–122 121

accessible reaction sites, i.e., q�T ¼ q�I þ q�O, where q�T,q�I and are the

total charge and charges related to the less accessible (inner) andmore accessible (outer) surfaces, respectively. As shown inFig. 9((a) and (b)), extrapolation of q* to v = 0 from 1/q* vs. v1/2

plot obtained from 75:00:15:10 electrode gives the total charge of37.88 C g�1 cm�2which is the charge related to entire active surfaceof 75:00:15:10 electrodes. In addition, extrapolation of q* to v = 1(v�1/2 = 0) from the q* vs. v�1/2 plot obtained from 75:00:15:10electrodes gives the outer charge, 16.32 C g�1 cm�2, which is thecharge due to redox processes on the most accessible active surface[31,32]. As represented in Table 1, and obtained from 25:50:15:10electrodes are 208.33 and 49.95 C g�1 cm�2, respectively. There-fore, it should be mentioned that NCB containing electrodes show ahigher charge storage capacity due to their nanoporous characterand higher charge separation ability. The free NCB electrodes showa high ratio of outer to total charge (/) of 0.53 and 0.43 obtainedfrom 90:00:00:10 and 75:00:15:10 electrodes, respectively. Thisratio was 0.24 for NCB containing electrodes (Table 1) whichconfirms the low current response of NCB electrodes on voltagereversal.

For practical applications the cyclic stability of supercapacitorsis a crucial parameter. As for pseudo-capacitive materials, the lifecycle of both conductive polymers and metal oxides is muchshorter than carbon base materials, because of their dissolutionand thus loss of active materials during redox reactions [33]. Onthe other hand, almost re-deposition of these dissolved metaloxides and electrolyte ions during the cycling on the pore walls

Fig. 8. Capacitance fading percent obtained from different electrodes by changingthe scan rate from 10 to 200 mV s�1.

Fig.10. Nyquist plots obtained from 75:00:15:10 electrode before and after 1, 10, 50and 100 CV cycles in 3 M NaCl electrolyte at scan rate of 10 mV s�1.

Fig. 11. Nyquist plots obtained from 25:50:15:10 electrode before and after 1, 10, 50and 100 CV cycles in 3 M NaCl electrolyte at scan rate of 10 mV s�1.

122 M.R. Sarpoushi et al. / Synthetic Metals 200 (2015) 117–122

decreases electrical conductivity of the electrodes and theirreversibility [34,35]. Therefore, cyclic performance of preparedelectrodes was evaluated by repeating the CV tests for 100 cycles atscan rate of 10 mV s�1. Simultaneously, EIS tests were used toevaluate surface changes (Figs. 10 and 11). As shown in Fig. 10,75:00:15:10 electrode shows no significant changes during first100 charging/discharging cycles. The absolute value of impedancein lumped constant-type range increases slightly which may bedue to the decrease of active surface area of inner walls of poresduring cycling [26]. In the presence of NCB particles (Fig. 11),increasing this absolute value of impedance is much more andconfirms the sever re-deposition of dissolved ions on inner walls ofnanopores which will decrease the electrolyte diffusion and willdecrease the accessible surface area of the NCB containingelectrodes. At lower frequencies, the straight sloping lines alongthe imaginary axis represent the Warburg diffusive resistance ofelectrolyte along pores and the proton diffusion in the hostmaterials. Transition frequencies (ftr) obtained from90:00:00:10 and 75:00:10:15 electrodes shift to lower frequenciesby charging and discharging of the electrodes after 100 cycles,slightly. This confirms that charging/discharging of electrodeshelps current to become non-uniform. Additionally, re-depositionof electroactive ions like Na and Ce (in CeO2 containing systems) onthe electrode during cycling makes preferred charge storage siteson the surface of electrode and makes the current non-uniform. Onthe other hand, almost always the ftr obtained from nanoporous

morphologies will be lower than the ftr obtained from smoothsurfaces but deposition of electroactive ions on the nanopores ofthe NCB containing electrodes can block them and decrease thedown pore effect of the structure and shift the ftr to the higherfrequencies by cycling. This phenomenon helps to make a uniformcurrent on the surface of nanoporous electrodes. Table 1 shows thecyclic electrochemical performance of prepared electrodesobtained by galvanostatic charging/discharging at a sweep rateof 10 mV s�1 for 100 cycles. More than 76% of the initial capacitancewas preserved by 25:50:15:10 electrode after 100 charge/dis-charge cycles.

4. Conclusion

Total charge () of 1.73, 37.88, 208.33 C g�1 cm�2 and outer chargeto total charge (/) of 0.53, 0.43 and 0.24 were obtained from90:00:00:10, 75:00:15:10 and 25:50:15:10 (GNS/CeO2/CB) electro-des, respectively. As the morphology becomes narrower anddeeper, the charge storage ability was increased and decreases thecurrent response and power delivering capability of the electrodes.Flat surface of GNSs shows a capacitance as high as 3.4 F/g andnanoporous structure obtained by carbon black containingelectrodes increases the capacitance up to 16.2 F/g at 10 mV s�1

in 3 M NaCl electrolyte.

References

[1] K.T. Chau, Y.S. Wong, C.C. Chan, Energy Convers. Manage. 40 (1999)1021–1039.

[2] H. Ibrahim, A. Ilinca, J. Perron, Renew. Sustainable Energy Rev. 12 (2008)1221–1250.

[3] Y. Zhang, Int. J. Hydrogen Energy 34 (2009) 4889–4899.[4] P. Sharma, T.S. Bhatti, Energy Convers. Manage. 51 (2010) 2901–2912.[5] E. Frackowiak, E.F. Beguin, Carbon 39 (2001) 937–950.[6] C. Liu, F. Li, L.P. Ma, Adv. Mater. 22 (2010) 28–62.[7] J.T. Hu, T.W. Odom, C.M. Lieber, Acc. Chem. Res. 32 (1999) 435–445.[8] G. Centi, S. Perathoner, Eur. J. Inorg. Chem. 26 (2009) 3851–3878.[9] A.G. Pandolfo, A.F. Hollenkamp, J. Power Sources 157 (2006) 11–27.

[10] G. Salitra, A. Soffer, L. Eliad, J. Electrochem. Soc. 147 (2000) 2486–2493.[11] D.Y. Pan, S. Wang, B. Zhao, Chem. Mater. 21 (2009) 3136–3142.[12] X.X. Wang, J.N. Wang, H. Chang, Adv. Funct. Mater. 17 (2007) 3613–3618.[13] M. Nasibi, M.A. Golozar, G. Rashed, Mater. Chem Phys. 139 (2013) 12–16.[14] D.W. Wang, H. Fang, F. Li, Adv. Funct. Mater. 18 (2008) 3787–3793.[15] W. Lv, D. Tang, Y.B. He, ACS Nano 3 (2009) 3730–3736.[16] K. Zhang, L.L. Zhang, X.S. Zhao, Chem. Mater. 22 (2010) 1392–1401.[17] D.W. Wang, F. Li, J.P. Zhao, ACS Nano 3 (2009) 1745–1752.[18] J.Y. Kim, K.H. Kim, S.H. Park, K.B. Kim, Electrochim. Acta 55 (2010) 8056–8061.[19] D. Yang, J. Power Sources 198 (2012) 416–422.[20] M. Nasibi, M.A. Golozar, G. Rashed, J. Power Sources 206 (2012) 108–110.[21] B. Babakhani, D.G. Ivey, Electrochim. Acta 56 (2011) 4753–4762.[22] L.S. Aravinda, K. Udaya Bhat, B.R. Bhat, Mater. Lett. 112 (2015) 158–161.[23] Y.G. Wang, H.Q. Li, Y.Y. Xia, Adv. Mater. 18 (2006) 2619–2623.[24] E. Frackowiak, V. Khomenko, K. Jurewicz, J. Power Sources 153 (2006) 413–418.[25] B. Babakhani, D.G. Ivey, J. Power Sources 195 (2010) 2110–2117.[26] M. Itagaki, S. Suzuki, I. Shitanda, K. Watanabe, H. Nakazawa, J. Power Sources

164 (2007) 415–424.[27] M. Nasibi, M.A. Golozar, G. Rashed, J. Power Sources 206 (2012) 108–110.[28] B.E. Conway, W.G. Pell, J. Power Sources 105 (2002) 169–181.[29] D.W. Wang, F. Li, M. Liu, G.Q. Lu, H.M. Cheng, Angew. Chem. Int. 47 (2008)

373–376.[30] H. Yamada, H. Nakamura, F. Nakahara, I. Moriguchi, T. Kudo, J. Phys. Chem. C

111 (2007) 227–233.[31] S. Ardizzone, G. Fregonara, S. Trasatti, Electrochim. Acta 35 (1990) 263–267.[32] J.Y. Kim, K.H. Kim, S.H. Park, K.B. Kim, Electrochim. Acta 55 (2010) 8056–8061.[33] P. Sharma, T.S. Bhatti, Energy Convers. Manage. 51 (2010) 2901–2912.[34] D. Choi, G.E. Blomgren, P.N. Kumta, Adv. Mater. 18 (2006) 1178–1182.[35] M. Nasibi, M. Irankhah, M.R. Sarpoushi, M.A. Golozar, M. Moshrefifar, M.R.

Shishesaz, Mater. Sci. Semicond. Process. 20 (2014) 49–54.