Contents lists available at ScienceDirect

Materials Science in Semiconductor Processing

Materials Science in Semiconductor Processing 20 (2014) 49–54

1369-80http://d

n CorrE-m

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

Short Communication

Graphite nanosheets as an electrode materialfor electrochemical double layer capacitors

Mahdi Nasibi a, Melika Irankhah b, Mehdi Robat Sarpoushi a,n,Mohammad Ali Golozar c, Masoud Moshrefifar d, Mohammad Reza Shishesaz a

a Technical Inspection Engineering Department, Petroleum University of Technology, Abadan, Iranb Chemistry Engineering Department, Amirkabir University of Technology, Tehran, Iranc Materials Science and Engineering Department, Isfahan University of Technology, Isfahan, Irand Materials and Mining Engineering Department, Yazd University, Yazd, Iran

a r t i c l e i n f o

Keywords:Electronic materialsNanostructuresElectrochemical measurementElectrical propertiesEnergy storage

01/$ - see front matter & 2014 Published byx.doi.org/10.1016/j.mssp.2014.01.002

esponding author. Tel.: þ98 9155725460.ail address: mehdi.sarpoushi@gmail.com (M

a b s t r a c t

In this paper, the effect of ion sizes of cations and anions on the charge storage capabilityof graphite nanosheets is investigated. Electrochemical properties of prepared electrodesare studied using cyclic voltammetry (CV) and electrochemical impedance spectroscopy(EIS) techniques, in 3 M NaCl, NaOH and KOH electrolytes. A scanning electron microscope(SEM) is used to characterize the microstructure and nature of prepared electrodes.SEM images and XRD patterns confirm the layered structure (12 nm thickness) of the usedgraphite with an interlayer distance of 3.36 Å. The electrochemical results and the ratio ofqn

O=qn

T confirm a better charge storage and charge delivering capability of preparedelectrodes in 3 M NaCl electrolyte. The charge/discharge cycling test shows a goodreversibility and confirms that the solution resistance will increase after 500 cycles.

& 2014 Published by Elsevier Ltd.

1. Introduction

Recently, electrochemical capacitors have attractedworldwide research interest. Depending on charge storagemechanisms, capacitors can be classified into three types:electrochemical double layer capacitors (EDLCs), faradaicpseudocapacitors and hybrid capacitors [1,2]. Especially inEDLCs and at high charge/discharge rates, since no chemicalreaction is involved, the effects are easily reversible withminimal degradation in deep discharge or overcharge andthe typical life cycle is hundreds of thousands of cycles [3,4].With respect to the electrode materials there are three maincategories: carbon based materials, transition metal oxidesand conductive polymers [5,6]. Among carbon based materi-als—due to good conductivity, superior chemical stability,

Elsevier Ltd.

.R. Sarpoushi).

large surface area-to-volume ratio and unique layered struc-ture—graphite has competed with carbon nanotubes, acti-vated carbon, etc. which are used as electrode material forECs. High surface area of graphite does not depend on thedistribution of pores in solid state, but comes from theinterconnected open channels between graphite layers dis-tributed in a two-dimensional architecture [7–9]. However,the major problem of such a material is that not all the BETsurface area is electrochemically accessible [10]. Ion sizes inelectrolyte and the charge/discharge rate are significantparameters affecting the ratio accessible to the total surface,energy storage and power capability of graphite electrodes.Therefore, choosing a proper electrolyte in accordance withthe morphology may be effective.

In this paper, the effect of ion size on charge storage,charge delivering capability and reversibility of graphiteelectrodes was investigated using cyclic voltammetry andelectrochemical impedance spectroscopy techniques. Themorphology and nature of the prepared electrodes wereinvestigated employing a scanning electron microscope.

Table 1Comparison of different carbon based materials in KCl electrolyte.

Active material Electrolyte Specific capacitance (F g�1)

Carbon black [22] 3 M KCl 33.58Activated carbon [23] 1 M KCl 29Multilayer graphene [24] 1 M KCl 15.6Few layer graphene [24] 1 M KCl 14.9Single layer graphene [24] 1 M KCl 10.9

M. Nasibi et al. / Materials Science in Semiconductor Processing 20 (2014) 49–5450

2. Experimental

2.1. Materials

Graphite nanopowder (12 nm Flakes, multi-layered)with the specific surface area of 15 m2/g and purity of98.5% was purchased from graphite supermarket and poly-tetrafluoroethylene (o2 μm) from Aldrich company. Allother chemicals used in this study were purchased fromMerck. 90 wt% graphite nanopowder and 10 wt% polytetra-fluoroethylene (PTFE) were well mixed by ultrasonic wavein ethanol in paste form for about 60 min. Paste form waschosen for better dispersion of PTFE in graphite nanoflakes.After drying the paste and powdering, the composite waspressed onto a 316L stainless steel plate (5�107 Pa) whichserved as a current collector (surface area was 1.22 cm2).A steel rod and a hollow cylinder of epoxy were used forpressing. The composite is pressed onto the epoxy properlyby the steel rod. A Teflon paper was used at the bottom ofthe rod because of its very low adhesion, in order to stickthe composite material to the stainless steel substrate. Thetypical mass load of the electrode material was 45 mg. Theelectrolytes investigated were 3 M NaCl, NaOH and KOH.

3. Characterization

Electrochemical behavior of prepared electrodes wascharacterized using CV and EIS tests. Electrochemicalmeasurements were performed using an Autolab (Nether-lands) Model PGSTAT302N. CV tests were performedwithin the range of �0.55 and þ0.3 V (vs. SCE). EISmeasurements were carried out in the frequency rangeof 100 kHz–0.02 Hz at OCP with an AC amplitude of 10 mV.The specific capacitance C (F g�1) of the active material wasdetermined by integrating either the oxidative part or thereductive part of the cyclic voltammogram curve (Q(C)). Thischarge was subsequently divided by the mass of the activematerial m (g) and the width of the potential window ofcyclic voltammogram ΔE (V), i.e.,C ¼ Q= mΔVð Þ [11–13].

4. Results and discussion

Ion size and diffusion of anions and cations are effectiveparameters of specific capacitance. When the KCl and NaClelectrolytes are compared, the pronounced difference is ofcation nature. Ionic radii of Kþ and Naþ are 146 and135 pm, respectively. In water solutions hydroxide ions arebigger than protons (ionic radius of Hþ and OH� is 35 and153 pm, respectively) which may lead to hindrance in thediffusion of potassium or sodium ions and lead to non-capacitive behavior. On the other hand, ionic mobility ofKþ is larger than Naþ but ionic radius is smaller. However,electrochemical nature of KCl electrolyte is close to that ofNaCl electrolyte. In addition to the ionic characteristics,nature of the active material is an effective parameter ofcapacitance. The SSA (specific surface area) of graphiteused in this paper is about 15 m2/g but graphene (in allforms) almost shows a higher SSA. The higher SSA ofgraphene than that of graphite is the main reason forincreasing the capacitance (Table 1). Additionally, CB ischaracterized by high porosity and small particle size.

These porosities are also varied from mild surface pittingto the actual hollowing out of particles. The surface area ofCB is considered to be more accessible generally than theother forms of high surface area carbons; this fine highlybranched structure of CB particles makes them ideallysuitable for filling inter-particle voids created betweencoarse particles which improves the electrical contactbetween them and increases the capacitance (Table 1).

Specific surface area and conductivity are two importantparameters to prepare highly efficient electrodes for capa-citors. But only a part of the surface is always accessible byelectrolyte ions to be adsorbed. This would increase thesolution resistance. Fig. 1(a) shows the SEM image ofprepared electrodes, confirming the 2D graphite nanosheetsconsisting of several carbon atom layers with a totalthickness of about 12 nm. Graphite layers interact witheach other to form open pore systems through which ionseasily access the surfaces between the graphite nanosheetsto form an electric double layer. Distance between thesenanosheets was measured using XRD and was about 3.36 Å(Fig. 1(b)). The graphite used was perpendicular to thesenanosheets, showed no porosity and was completely flat(Fig. 1 (a)). Thus, the used material is in its 2D porous and1D flat surfaces completely. With this morphology, it seemsthat the charge storage depends directly on the chargeseparation on flat part (which is the most accessible surfaceof electrode) and on open pore systems (which are lessaccessible and ion size dependant). Ion size and ion diffu-sion through these pores would affect the activation ofthese less accessible surfaces, especially at high scan rates[14]. Increasing the ionic radius would decrease the numberof adsorbed ions on the unit surface area of the electrode.This would decrease the charge stored on the outerHelmholtz layer. Therefore, for further investigations 3 Melectrolytes of KOH, NaOH and NaCl were employed. Themain difference in these electrolytes is the effective radiusof their anions and cations. Naþ , Kþ , Cl�and OH� ions haveeffective radii of 102, 138, 181 and 153 pm, respectively. Theratio of interlayer distance of graphite 3.36 Å to ionic radius(α) for these ions would be 3.29, 2.43, 1.86 and 2.20,respectively.

The capacitance of each electrode was calculated fromthe CV curves using

Cs ¼Z

i dV=msΔV ð1Þ

where Cs is the specific capacitance,Ri dV is the integrated

area of the CV curve, m is the mass of the active material(mass of electrode material regardless of mass of PTFE),ΔV is the potential range, and s is the scan rate.

Table 2Capacitance (100 mV s�1) in different electrolytes.

Electrolyte Capacitance (F g�1 at 100 mV s�1)

NaCl 1.99KOH 1.09NaOH 1.84

Fig. 1. (a) Scanning electron microscopy image, (b) XRD pattern obtainedfrom graphite electrodes, (c) cyclic voltammetry curves and (d) Nyquistdiagrams obtained from graphite electrodes in different electrolytes.

M. Nasibi et al. / Materials Science in Semiconductor Processing 20 (2014) 49–54 51

Fig. 1(c) shows the first cyclic voltammetry curves ofthe prepared electrodes at the scan rate of 100 mV s�1 indifferent electrolytes. All CVs exhibit a rectangular shaped

profile which is a good characteristic of ideal capacitivebehavior. All the electrode/electrolyte systems exhibited apotential almost independent of the double layer capaci-tance. Prepared electrodes showed a low current densityin 3 M KOH electrolyte which increased in 3 M NaOH andNaCl electrolytes (Fig. 1(c)). Increasing the ionic radiuscould decrease the ionic diffusion through the poresand cause a lower current reversal at the final potentials(Fig. 1(c)). Table 2 represents capacitance at the scan rateof 100 mV s�1 in all three electrolytes. NaCl solutionpossessed the maximum capacitance among these elec-trolytes. The point intersecting the Nyquist curves with thereal axis in the range of high frequency shows theequivalent series resistance (ESR) (Fig. 1(d)). It indicatesthe total resistance of electrode, the bulk electrolyteresistance and the resistance at the electrolyte/electrodeinterface [15–18]. Therefore, two counter-acting para-meters would act simultaneously as the ionic radiusincreases: increase in electrical resistance and decreasein charge adsorption on the surface of prepared electrodes.In order to obtain quantitative information on the utiliza-tion of prepared graphite electrodes, voltammogramswere analyzed as a function of scan rate, using theprocedure reported by Ardizzone et al. Fig. 2(a), (b), and(c) exhibits cyclic voltammetry curves obtained usingdifferent scan rates in 3 M KOH, NaOH and NaCl electro-lytes, respectively. Table 3 represents capacitance in dif-ferent scan rates in NaCl solution (optimum solution). Incharge and discharge cycles, the total charge can bewritten as the sum of an inner charge from the lessaccessible reaction sites and an outer charge from themore accessible reaction sites, i.e.,qn

T ¼ qnI þqn

O, where qnT, q

nI

and qn

O are the total charge and charges related to innerand outer surfaces, respectively [19,20]. Extrapolation of q*to s¼0 from 1/q* vs. s1/2 plot (Fig. 2(d)) gives the totalcharge qn

T which is the charge related to the entire activesurface of the electrode. In addition, extrapolation of q* tos¼1 (s�1/2¼0) from q* vs. s�1/2 plot (Fig. 2(e)) gives theouter charge qn

O, which is the charge on the most acces-sible active surface. Prepared electrodes show the ratio ofouter to total charge qn

O=qnT

� �of 0.005, 0.008 and 0.239 in

KOH, NaOH and NaCl electrolytes, respectively. Theseratios confirm the low current response of graphite inKOH and NaOH electrolytes which is due to the largeradius of Kþ ions and low mobility of hydroxyl ions.Among these electrolytes, NaCl shows the lowest electricalresistance and therefore was chosen to investigate theeffect of charge/discharge cycle.

As scan rate increases (Fig. 3(a)) the capacitance versuspotential relation will deviate from the classical squarewaveform, expected for a pure capacitor. As discussed by

Fig. 2. Cyclic voltammograms curves obtained using different scan ratesin 3 M (a) KOH, (b) NaOH and (c) NaCl electrolytes. (d) Extrapolation of q*to s¼0 from the 1/q* vs. s0.5 plot, which gives the total charge and (e)extrapolation of q* to s¼1 from the q* vs. s�0.5 plot, which gives theouter charge for graphite electrodes in different electrolytes.

Table 3Capacitance at different scan rates in NaCl solution.

Scan rate (mV s�1) Capacitance (F g�1)

5 3.6230 2.6750 2.38l00 1.99200 1.40300 1.05500 0.71

M. Nasibi et al. / Materials Science in Semiconductor Processing 20 (2014) 49–5452

some researchers this is due to the resistance effects downthe pores [4]. In addition, efficiency is another importantparameter affecting the capacitance at high sweep rates.As sweep rate increases, loss of energy increases and thecharge stored on the electrode surface decreases causingthe capacitance to decrease (Fig. 3(a)). Graphite electrodeshows a low capacitance of 3.62 F g�1 in 3 M NaCl electro-lyte. It seems that the low ionic diffusion due to smallinterlayer distances which is one of the unique propertiesof graphite nanosheets may be the main reason for thislow capacitance.

Regarding the practical applications, the cyclic stabilityof supercapacitors is a crucial parameter. The life cycles ofboth conducting polymers and metal oxides, as candidatesfor pseudocapacitive materials, are much shorter thanthose of the carbon based materials because of the lossof active materials. In the case of graphite nanosheets, thecyclic stability was evaluated by repeating the CV at a scanrate of 100 mV s�1 for 500 cycles (Fig. 3(b)). Simulta-neously, EIS tests were used to evaluate the electrodechanges (Fig. 3(c)). Graphite was found to exhibit excellentstability over the entire cycle numbers (Fig. 3(b)). Theanodic and cathodic currents decreased, but the cyclicvoltammetry curves remained in their rectangular shapedprofiles (Fig. 3(b)). Conversely the charge stored on theelectrode decreased only by �23% of the initial chargestored on the electrode (Table 4). These behaviors demon-strated a good cycling stability. Most capacitors, exceptthose of vacuum or air type, do not exhibit ideal purecapacitive behavior according to the criterion that the realand imaginary components of their impedance should beout of phase by 901 independent of the frequency betweenperiodic changes of current and voltage when addressedby a sinusoidally alternating voltage. Most practical capa-citors deviate from this requirement because their impe-dance is not that of a pure capacitance but of a capacitancelinked in series with an element exhibiting ohmic behaviorthat in combination with the capacitance gives rise to aphase angle that is less than 901 and is usually frequencydependent. This ohmic component is referred to as theequivalent series resistance [21]. In all three electrolytesthe phase angle is less than 901 because of the ESRinvolved in all three electrolytes. Fig. 4 shows that thephase angle (φ) between E (voltage) and I (current)depends on the frequency of NaCl electrolyte.

In this case, a good coating attributed to the properapplied pressure was one of the reasons for good cyclicstability. Although applying a higher pressure may increase

Fig. 3. (a) Cyclic voltammograms obtained at different scan rates, (b)cyclic voltammograms obtained over 500 cycles, and (c) Nyquist plotsafter different cycles for graphite nanosheets, in 3 M NaCl electrolyte.

Table 4Capacitance over 500 cycles in NaCl solution.

Cycle number Capacitance (F g�1 at 100 mV s�1)

1 cycle 1.9950 cycles 1.96200 cycles 1.73500 cycles 1.53

Fig. 4. Bode plot (phase angle/degree).

M. Nasibi et al. / Materials Science in Semiconductor Processing 20 (2014) 49–54 53

the cyclic stability due to the better adhesion of coating onthe substrate, capacitance would decrease dramatically.During the 500 charge/discharge cycles the equivalent seriesresistance increased (Fig. 3(c)) and Nyquist plots shifted tohigher values which can be attributed to the electrolytedecomposition on the electrode surface.

5. Conclusions

In summary, studies confirmed the presence of flat andporous (having about 3.36 Å interlayer distance) surfaces,and showed ion size and charge/discharge rate dependenton energy storage and power capability of the usedmaterial. It showed a good cycling performance stemmedfrom controlled thickness and subsequent changing of thesubstrate reactive state. The relatively high ratio of qn

O=qnT

(0.236) confirms the high current response on voltagereversal in 3 M NaCl electrolyte. The rectangular shapedprofile which is a good characteristic of ideal capacitivebehavior is obvious over the entire cycle numbers. Chargestored on the electrode decreased only by �23% of theinitial charge stored on the electrode which representsgood reversibility. The proposed electrodes provided adouble layer capacitance and showed a capacitance ashigh as 3.62 F g�1 at 5 mV s�1 in 3 M NaCl electrolyte. Thelow specific surface area of graphite nanosheets wasresponsible for this low capacitance.

References

[1] B.E. Conway, W.G. Pell, J. Power Sources 105 (2002) 169–181.[2] I. Hadjipaschalis, A. Poullikkas, V. Efthimiou, Renew. Sustain. Energy

Rev. 13 (2009) 1513–1522.[3] J. Yan, T. Wei, B. Shao, F. Ma, Zh. Fan, et al., Carbon 48 (2010)

1731–1737.[4] V. Singh, D. Joung, L. Zhai, S. Das, S.A. Khondaker, et al., Prog. Mater.

Sci. 56 (2011) 1178–1271.[5] M. Nasibi, M.A. Golozar, Gh. Rashed, Mater. Chem. Phys. 139 (2013)

12–16.[6] B. Babakhani, D.G. Ivey, Electrochim. Acta 56 (2011) 4753–4762.[7] A. Burke, J. Power Sources 91 (2000) 37–50.[8] S. Ardizzone, G. Fregonara, S. Trasatti, Electrochim. Acta 35 (2000)

263–267.[9] Y. Honda, T. Haramoto, M. Takeshige, H. Shiozaki, T. Kitamura,

M. Ishikawa, Electrochem. Solid-State Lett. 10 (2007) 106–110.[10] I.-H. Kim, K.-B. Kim, J. Electrochem. Soc. 153 (2006) 383–389.[11] Z. Fan, J. Chen, K. Cui, F. Sun, Y. Xu, Y. Kuang, Electrochim. Acta 52

(2007) 2959–2965.[12] M.-S. Wu, Y.-A. Huang, C.-H. Yang, J.-J. Jow, Int. J. Hydrog. Energy 32

(2007) 4153–4159.

M. Nasibi et al. / Materials Science in Semiconductor Processing 20 (2014) 49–5454

[13] X.-H. Yang, Y.-G. Wang, H.-M. Xiong, Y.-Y. Xia, Electrochim. Acta 53(2007) 752–757.

[14] Z.J. Lao, K. Konstantinov, Y. Tournaire, S.H. Ng, G.X. Wang, H.K. Liu, J.Power Sources 162 (2006) 1451–1454.

[15] D.-D. Zhao, S.-J. Bao, W.-J. Zhou, H.-L. Li, Electrochem. Commun. 9(2007) 869–874.

[16] K. Okajima, A. Ikeda, K. Kamoshita, M. Sudoh, Electrochim. Acta 51(2005) 972–977.

[17] W. Xing, S.Z. Qiao, R.G. Ding, F. Li, G.Q. Lu, Z.F. Yan, et al., Carbon 44(2006) 216–224.

[18] D. Choi, P.N. Kumta, J .Electrochem. Soc. 153 (2006) 2298–2303.

[19] I. Danaee, M. Jafarian, F. Forouzandeh, F. Gobal, M.G. Mahjani, Int. J.Hydrog. Energy 34 (2009) 859–869.

[20] R. Kotz, M. Carlen, Electrochim. Acta 45 (2000) 2483–2498.[21] B.E. Conway, Electrochemical Capacitors: Scientific Fundamental and

Technological Applications, Kluwer Academic/Plenum, New York,1999.

[22] M. Nasibi, M.A. Golozar, Gh. Rashed, Mater. Lett. 91 (2013) 323–325.[23] H. Nakagawa, A. Shudo, K. Miura, J. Electrochem. Soc. 147 (2000)

38–43.[24] Ch.H. Chen, et al., J. Power Sources 205 (2012) 510–515.