Contents lists available at ScienceDirect

Materials Science in Semiconductor Processing

Materials Science in Semiconductor Processing 31 (2015) 195–199

http://d1369-80

n CorrMalek A

nn Corment, PTel.: +9

E-mMahdi.n

journal homepage: www.elsevier.com/locate/mssp

Short Communication

Graphene nanosheets as electrode materialsfor supercapacitors in alkaline and salt electrolytes

Mahdi Robat Sarpoushi a, Mohammad Reza Borhani b,n, Mahdi Nasibi a,c,nn,Behzad Eghdami a, Hanif Kazerooni d

a Technical Inspection Engineering Department, Petroleum University of Technology, Abadan, Iranb Department of Materials Engineering, Malek Ashtar University of Technology, Shahin Shahr, Isfahan, Iranc Health, Safety and Environment (HSE) Engineering Office, NIOPDC, Yazd Region, Yazd 89167-84395, Irand Faculty of Chemical Engineering, Amirkabir University of Technology (Tehran Polytechnic), Tehran, Iran

a r t i c l e i n f o

Keywords:Composite materialsElectronic materialsNanostructuresUltrasonic techniquesMicrostructureElectrochemical properties

x.doi.org/10.1016/j.mssp.2014.10.03401/Crown Copyright & 2014 Published by E

esponding author at: Department of Mashtar University of Technology, Shahin Shahresponding author at: Technical Inspectionetroleum University of Technology, Abadan,8 9113708480; fax: +98 3535253091.ail addresses: Moh_brh65@yahoo.com (M. Rasibi@gmail.com (M. Nasibi).

a b s t r a c t

Carbon materials have played a significant role in the development of alternative cleanand sustainable energy technologies. In particular, we will systematically discuss theapplications of graphene nanosheets as an electrode material for supercapacitors. Thisarticle summarizes, the effect of size and nature of ions on pseudocapacitance and doublelayer capacitance of graphene electrode using CV and EIS techniques. The morphology andnature of the prepared electrode was investigated employing a scanning electronmicroscope. The prepared electrode shows better double layer characteristics in NaOHelectrolyte in the potential range between �0.55 and 0.3 V (V vs. SCE) at a scan rate of100 mV s�1.

Crown Copyright & 2014 Published by Elsevier Ltd. All rights reserved.

1. Introduction

Electric double-layer capacitors (EDLCs), also called super-capacitors, are used in a wide range of energy capture andstorage applications, and are believed to provide cleanenergy with nearly zero waste emission [1,2]. Comparedwith conventional capacitors, EDLCs can store much moreenergy for two main reasons: (1) a very small distance ofcharge separation at the interface between electrode andelectrolyte and (2) larger amount of charges on the largersurface area of electrode [3]. The properties of the electrolytesystem, solvent plus solute salt, required for supercapacitors

lsevier Ltd. All rights reserv

terials Engineering,r, Isfahan, Iran.Engineering Depart-Iran.

eza Borhani),

determine their electrical behavior [4]. Thus, the electrolyteplays an important role.

In this paper, the effects of ions size and specific proper-ties on pseudocapacitance and double layer capacitance ofthe graphene electrode are investigated using CV and EIStechniques. The morphology and nature of the preparedelectrode was investigated employing a scanning electronmicroscope.

2. Experimental

2.1. Materials

Graphene nanosheets (60 nm Flakes, multi-layered) with aspecific surface area of 15m2/g and purity of 98.5% werepurchased from graphene supermarket and polytetrafluor-oethylene (o2 μm) from Aldrich company. All other che-micals used in this study were purchased from Merck. In orderto prepare the electrodes, a mixture containing 90wt%

ed.

Fig. 1. Electrical double layer showing inner Helmholtz, outer Helmhotz

M. Robat Sarpoushi et al. / Materials Science in Semiconductor Processing 31 (2015) 195–199196

graphene nanosheets (GNS) and 10wt% polytetrafluoroethy-lene (PTFE) was prepared by mixing well in paste form inethanol for about 30min using ultrasonic wave. Paste formwaschosen for homogenous dispersion of PTFE particles in thegraphene matrix and to prevent the agglomeration of PTFEparticles and graphene nanosheets. The main roles of a binderinside the electrode layer are: (1) to hold graphene nanosheetstogether forming a compacted porous layer; and (2) to help thiselectrode layer to adhere uniformly onto the current collector.Considering these properties, polytetrafluoroethylene (PTFE)has been recognized as one of the most popular choices.However, PTFE is a hydrophobic agent, and if too much PTFEis used, the porous electrode layer will become more hydro-phobic, causing difficulty in both electrolyte penetration andion mobility inside the matrix structure when an aqueouselectrolyte is employed [5]. After drying the paste and powder-ing it, the composite was pressed onto a 316 L stainless steelplate (50MPa) which served as a current collector (having asurface area of 1.4 cm2). A steel rod and a hollow cylinder ofepoxy were used for pressing. The composite was pressed ontothe epoxy properly by the steel rod. The typical mass load ofthe electrode material was 45mg. The electrolytes investigatedwere 3M NaOH and LiBr.

and diffuse layer (φ1¼ inner, φ2¼outer) [10].

3. Characterization

Electrochemical characteristics of the electrode wasdetermined by cyclic voltammetry (CV) and electrochemicalimpedance spectroscopy (EIS) in a three-electrode cell with3 M NaOH and LiBr as electrolytes. The potential was cycledat different scan rates using a potentiostat (PGSTAT302N,Netherland) in the range of �0.55 to 0.3 V. EIS measure-ments were also carried out in the frequency range of100–0.02 Hz at open circuit potential with an AC amplitudeof 10 mV. Surface morphology and cross-section image ofthe electrode were examined with the scanning electronmicroscope (SEM, TESCAN, USA).

The specific capacitance can be estimated from thevoltammetric charge surrounded by the CV curve accord-ing to the following formula [6,7]:

C ¼ qaþ qc��

��

2mΔVð1Þ

where qa and qc are the sums of anodic and cathodic volt-ammetric charges on positive and negative sweeps, respec-tively, m is the mass of active material (regardless of mass ofPTFE) and ΔV is the potential window of CV.

The real (C0) and imaginary (C″) capacitance of theelectrode are calculated using the following equations [8,9]:

C 0 ¼ Z″ðωÞω ZðωÞ��

��2 ð2Þ

C″¼ Z0ðωÞω ZðωÞ��

��2

ð3Þ

where Z0(ω) and Z″ (ω) are the respective real and imaginaryparts of the complex impedance Z (ω), ω is the angularfrequency and is given by ω¼2πf.

4. Results and discussion

In stern model, the overall capacitance in EDLC (Cdl),was considered as a series of capacitance, CH and Cdiff:

1Cdl

¼ 1CH

þ 1Cdiff

ð4Þ

where CH represents the compact Helmholtz layer forma-tion from solvated ions attracted electrostatically in theOHP layer, while diffusion capacitance (Cdiff) results fromthe ions transportation, caused by a gradient between bulkand interfacial concentration of the electrolyte ions. Repre-sentation of the Helmholtz double layer is illustrated inFig. 1. When the frequency increases or low conductiveions are used, the number of ions involved in the diffusionprocess can be reduced, therefore resulting in a decrease incapacitance. Graham's model presents the overall doublelayer capacitance which is composed of three components:adsorption capacitance (Cads), Helmholtz capacitance (CH)and diffusion capacitance (Cdiff) [11].

In Grahame's theory, it was recognized that dehydratedions in IHP region could reside on the electrode surfacewith specific adsorption processes. This phenomenonresults in adsorption capacitance, Cads. In a certain system,this Cads can be regarded as another capacitive elementwith some part of the electrochemical charge transferprocess. The total capacitance (CT) can be represented byCads and in a series combination with Cdl:

1CT

¼ 1Cads

þ 1Cdl

ð5Þ

It should be noted that two types of pseudocapacitance canarise in electrochemical processes: (1) adsorption pseudoca-pacitance and (2) redox pseudocapacitance [12]. However, itis necessary to differentiate Cads from another pseudocapa-citance which originates from Faradaic (oxidation/reduction

Fig. 2. SEM image obtained from graphene electrode.

Fig. 3. (a) CV curves and (b) Nyquist diagrams obtained from preparedelectrode in 3 M LiBr and NaOH electrolytes. (For interpretation of thereferences to color in this figure, the reader is referred to the web versionof this article.)

Fig. 4. Cations and anions alignment on the electrode surface.

M. Robat Sarpoushi et al. / Materials Science in Semiconductor Processing 31 (2015) 195–199 197

reactions) processes due to other sources such as metaloxide, conductive polymers or the functional group [10,13].

Fig. 2 shows the SEM images of GNS. Graphene layersform a continuous network where they are closely asso-ciated with each other. With this morphology, it seems thecharge storage depends directly on the charge separationon the upper part (as the most accessible surface of theelectrode) and ions hardly diffuse into the inner pores.Therefore, ions size (α), ion mobility (u) and ion conductiv-ity (Λ) directly influences the value of capacitance. The maindifferences in NaOH and LiBr electrolytes is the effectiveradius and ion mobility of their anions and cations. Liþ ,Naþ , OH� and Br� ions have effective radii of 76, 102, 153and 196 pm, and relative ion mobility of 0.525, 0.682, 2.690,and 1.603 respectively. The ratio of interlayer distance ofgraphene (3.36 Å) to ionic radius (α) for these ions would be4.33, 3.29, 2.2 and 1.71, respectively. Increasing the ionicradius would decrease the number of adsorbed ions storedon the outer Helmholtz layer of the electrode, therefore, itis expected that the double layer capacitance in LiBr beless than that of NaOH ððαLiþ =αNaþ Þ ¼ 0:745; ðαOH� =αBr� Þ ¼0:780; ðuLiþ =uNaþ Þ ¼ 0:77; ðuBr� =uOH� Þ ¼ 0:59; ðΛLiþ =ΛNaþ Þ¼ 0:77; ðΛBr� =ΛOH� Þ ¼ 0:39Þ. On the other hand, electro-chemical studies carried out in electrolytes have shown thatthe highest capacitances are obtained when the pore sizematches the diameter of the ionic species [14]. In order toinvestigate this phenomenon, EIS and CV tests wereemployed. Potential range is one of the predominant factorsinfluencing the capacitive properties of the electrochemicalcapacitor. Generally, prerequisite for selecting the potentialrange should guarantee the stabilization of the activematerial on the electrode and will not cause the decom-position of the electrolyte, which is the crucial point tosolve this problem. Therefore, we must locate a potentialwindow in which the oxygen and hydrogen evolutiondoes not occur in cathode [15]. As shown in Fig. 3(a), the

M. Robat Sarpoushi et al. / Materials Science in Semiconductor Processing 31 (2015) 195–199198

electrode in 3 M LiBr electrolyte possesses the highestsurface area which leads to a significant capacitance. Thecations or anoins attracted towards the electrode arelimited in their approach to the electrode because of thepresence of ions with opposite sign charge on the electrodesurface. Therefore, it was concluded that in addition to ionicradius, ion mobility and ion conductivity, other parameterssuch as ions alignment on the electrode surface may alsoaffect the capacitance (see Fig. 4). For an ideal double layercapacitor, the sign of the current changes immediately afterreversing the potential and the shape of the voltammetry isrectangular. For this electrostatic energy storage, the cur-rent is independent of the potential of the electrode. Fordouble layer capacitors with resistive losses, the shapechanges into a parallelogram. The blue curve in Fig. 3(a),is composed of resistive losses, double layer capacitanceand a small amount of faradaic pseudocapacitance. Forelectrodes with faradaic pseudocapacitance, the electricalcharge stored in the capacitor is strongly dependent on thepotential (Fig. 3(a), red curve) [16]. Considering Eq. (5) andFig. 3(a), it is obvious that Cads contributes more to the totalcapacitance in LiBr than in NaOH electrolyte. The Nyquistplots of the electrode at OCP (open circuit potential) and theequivalent circuits obtained from Nyquist plots for thesetwo electrolytes are shown in Fig. 3(b). The charge transferresistance, Rct, is mainly related to the gradients of poten-tials between the electroactive species (in this study,lithium ion (Liþ) and bromide (Br�) in the salt electrolyte,sodium ion (Naþ) and hydroxide (OH�) in the alkaline

Table 1Electrochemical parameters obtained from electrochemical impedancespectroscopy measurement from graphene electrode.

Type ofelectrolyte

Rs(Ω)

Q (S sα), α Rct(Ω)

W(S s0.5)

CF(F)

Rl(Ω)

LiBr 9.455 0.071, 0.514 4.728 0.698 0.112 624NaOH 4.821 0.051, 0.244 5.806 – 0.084 589

Fig. 5. Cyclic voltammogram curves obtained from graphene electrode a

electrolyte) in electrolytes and the electrode surface, lead-ing to the charge transfer phenomena. If there is an electrontransfer reaction, Rct becomes smaller, otherwise chargetransfer resistance becomes very large and the electrode ispolarized with poorly defined potential [17]. As shown inTable 1, the diameter of the semicircle in high frequencies isslightly higher for NaOH electrolyte, which indicates thesmall increase of charge transfer resistance. Also CF obtainedfrom LiBr solution is larger than NaOH (this is confirmed bycomparing red curve and blue curve in Fig. 3(a), due to theElliptical and parallelogram shapes, respectively). Rct in LiBris smaller than NaOH which indicates that Faradic reactionstake place easily in LiBr. Warburg diffusion element (W)attributed to the diffusion of ions, is considered for LiBrwhich indicates that the electrode is under diffusion controlonly in LiBr electrolyte (as seen in Fig. 3(b) Warburg diffusionelement appears as a straight line with a slope of 451 atmoderate frequency).

In addition, the practical EDLC devices suffer from chargeleakage (defined as self-discharge) which results frompotential-dependent charge transfer reactions (see Table 1).These deviations from the ideal capacitive behavior of EDLCare attributed mainly to ionic chemical/physical adsorptionand diffusional impedance, incomplete polarization of theporous electrode and Faradaic charge transfer resistance cau-sed by the voltage differential across the electrode/electro-lyte interface [18–20].

Fig. 5 shows the cyclic voltammograms using differentscan rates in 3 M LiBr and NaOH electrolytes. The resultsindicated that by increasing scanning rate, the enclosingarea of curves became smaller and smaller. The parallelo-gram and rectangle-like CV curves were not maintainedin various scan rates, which is a major characteristic ofFaradaic processes in an EDLC.

It is known that a phase angle of 901 signifies an idealcapacitive behavior, otherwise the material would show apseudocapacitive behavior [21,22]. The phase angle val-ues for NaOH and LiBr electrolytes are �771 and �721,

t different scan rates in (a) 3 M LiBr and (b) 3 M NaOH electrolytes.

Fig. 6. (a) Phase angle vs. frequency and (b) imaginary capacitance vs.frequency in 3 M LiBr and NaOH electrolytes.

M. Robat Sarpoushi et al. / Materials Science in Semiconductor Processing 31 (2015) 195–199 199

respectively (Fig. 6).This phenomenon confirms further thepseudocapacitive nature of the electrode in LiBr electrolyte.

τ0 is the minimum time needed to discharge all theenergy from the device with an efficiency of greater than50% of its maximum value, and it can be derived from thefrequency at maximum C″ [23]. Values of τ0 for NaOH andLiBr electrolytes are 4.03 and 10.41 s, respectively. The lowion diffusion in LiBr electrolyte was further confirmed (dueto the presence of Warburg element in the equivalentcircuit) by its large relaxation time constant (τ0) (Fig. 6).

5. Conclusion

SEM image confirmed the presence of flat and poroussurfaces (large flakes with different orientation). It seems

the charges are stored mainly on the outer surface (themore accessible reaction sites) of the electrode. TheGraphene electrode in presence of an electrolyte is com-posed of resistive losses, a small amount faradaic pseudo-capacitance and double layer capacitance. The electrode inNaOH electrolyte showed a lower amount of resistivelosses and faradaic pseudocapacitance. The ideal behaviorof the electrode in NaOH electrolyte confirmed more, ascompared with LiBr electrolyte, by increasing phase anglevalues from �721 to �771.

References

[1] S. Trasatti, P. Kurzweil, Platin. Metal Rev. 38 (1994) 46.[2] B.E. Conway, W.G. Pell, J. Solid State Electrochem. 7 (2003) 637.[3] A.G. Pandolfo, A.F. Hollenkamp, J. Power Sources 157 (2006) 11.[4] B.E. Conway, Electrochemical Capacitors Scientific Fundamental and

Technological Applications, Kluwer Academic/Plenum, New York, 1999.[5] Keh-Chyun Tsay, Lei Zhang, Jiujun Zhang, Electrochim. Acta 60

(2012).[6] C.C. Hu, C.C Wang, Electrochem. Commun. 4 (2002) 554.[7] X.Y. Wang, X.Y. Wang, W.G. Huang, P.J. Sebastian, S. Gamboa, J. Power

Sources 140 (2005) 211.[8] J.M. Soon, K.P. Loh, Electrochem. Soild-Sate Lett. 10 (2007) A250.[9] V. Ganesh, S. Pitchumani, V. Lakshminarayanan, J. Power Sources

158 (2006) 1523.[10] B.E. Conway, V. Birss, J. Wojtowicz, J. Power Sources 66 (1997) 1–14.[11] Jinhee Kang, John Wen, Shesha H. Jayaram, Aiping Yu, Xiaohui Wang,

Electrochim. Acta 115 (2014) 587–598.[12] Zaki Ahmad, Principles of Corrosion Engineering and Corrosion

Control, first ed. Elsevier, Oxford, UK, 2006.[13] B.P. Bakhmatyuk, B.Y. Venhryn, I.I. Grygorchak, M.M. Micov,

S.I. Mudry, Rev. Adv. Mate. Sci. 14 (2007) 151–156.[14] C. Largeot, C. Portet, J. Chmiola, P. L Taberna, Y. Gogotsi, P. Simon,

J. Am. Chem. Soc. 130 (2008) 2730.[15] Y. Zhao, Y.Y Wang, Q.Y. Lai, L.M. Chen, Y.J. Hao, X.Y. Ji, Synth. Metals

159 (2009) 331–337.[16] Elzbieta Frackowiak, Francois Beguin, Carbon 39 (2001) 937–950.

(PERGAMON).[17] A.J. Bard, L.R. Faulkner, Electrochemical Methods, Fundamentals and

Applications, J. Wiley & Sons, New York, 2001.[18] J. Kowal, E. Avaroglu, F. Chamekh, A. S’Enfelds, T. Thien, D. Wijaya,

D.U. Sauer, J. Power Sources 196 (2011) 573–579.[19] PascalVenet, Gerard Rojat, Yasser Diab, Proceedings of the 3rd

European Symposium on Supercapacitors and Applications, Italy,2008.

[20] B.W. Ricketts, C. Ton-That, J. Power Sources 89 (2000) 64–69.[21] L. Yuan, X.H. Lu, X. Xiao, T. Zhai, J. Dai, F. Zhang, et al., ACS Nano 6

(2012) 656–661.[22] K. Krishnamoorthy, G. Veerasubramani, S. Radhakrishnan, S.J. Kim,

Mater. Res. Bull. 50 (2014) 499–502.[23] C. Du, N Pan, J. Power Sources 160 (2006) 1487–1494.