Electrochemical behaviour of vertically aligned carbon nanotubes and graphene oxide nanocomposite as electrode material

Page 1

Eg

TFa

b

Cc

a

ARR1AA

Kgfncvo

1

tcdvebsgpn

tav

0h

Electrochimica Acta 119 (2014) 114– 119

Contents lists available at ScienceDirect

Electrochimica Acta

jou rn al hom ep age: www.elsev ier .com/ locate /e lec tac ta

lectrochemical behaviour of vertically aligned carbon nanotubes andraphene oxide nanocomposite as electrode material

iago Almeida Silvaa, Hudson Zaninb,∗, Eduardo Saitob, Roberta Antigo Medeirosc,ernando Campanhã Vicentinia, Evaldo José Coratb, Orlando Fatibello-Filhoa

Department of Chemistry, Federal University of Sao Carlos, Rod, Washington Luís km 235, São Carlos, CEP: 13560970, SP, BrazilAssociated Laboratory of Sensors and Materials of the National Institute for Space Research, Av. dos Astronautas 1758, São José dos Campos,EP: 12227010, SP, BrazilCenter of Engineering and Exact Sciences, State University of West Paraná, Rua da Faculdade 2550, Toleto, CEP: 85903000, PR, Brazil

r t i c l e i n f o

rticle history:eceived 25 September 2013eceived in revised form6 November 2013ccepted 6 December 2013vailable online 18 December 2013

eywords:raphene oxide

a b s t r a c t

Vertically aligned carbon nanotubes/graphene oxide nanocomposite (VACNT-GO) has been preparedand applied as electrode material. First, dense packets of VACNT were prepared in microwave chemicalvapor deposition reactor, and then functionalized by oxygen plasma etching. We observed that oxy-gen plasma could exfoliate carbon nanotubes tips and provide oxygen group attachment on its surface,changing its wettability as well. This change in wettability of the VACNT is crucial for its electrochem-ical application, since as-grown VACNT is super-hydrophobic. After exfoliation and functionalization,the electrochemical tests were performed using potassium ferrocyanide. The cyclic voltammetry (CV)and impedance spectroscopy revealed fast electron transfer kinetics on this new material. The CV

ast electron transferanotubes exfoliationarbon nanotubesertically alignedxygen plasma etching

peak potential separation was 59 mV, suggesting ideal reversibility at the electrode. The Nyquist andBode plots were well-fitted as modified Randles equivalent electrical circuits with non-charge trans-fer impedance. This new highly porous nanostructures have been intensively characterized by scanningelectron microscopy, Brunauer–Emmett–Teller surface area, surface wettability, Raman and X-ray pho-toemission spectroscopy. Our results suggest this new material has a relevant potential for futureapplications in electrocatalysis and (bio)sensors.

© 2013 Elsevier Ltd. All rights reserved.

. Introduction

Carbon nanotubes have been extensively investigated in elec-rochemical devices development because of a set of intrinsicharacteristics of these nanomaterials. Their good electrical con-uctivity, chemical and mechanical stability, high surface area,ersatility for biological and chemical species immobilization,tc, are essential for these applications. Carbon nanotubes haveeen applied in the design of electrochemical sensors and biosen-ors [1–3], supercapacitors [4,5], fuel cells [6] etc. Currently,raphene is at the center of a significant research effort with theromise of owning properties as good as or even better than theanotubes [7].

The electrochemical performance of carbon nanotubes elec-

rodes depends directly on the orientation and chemical function-lization of these nanostructures [8]. In fact, electrodes havingertically aligned carbon nanotubes display a better electrocatalytic

∗ Corresponding author. Tel.: +55 16 33518098; fax: +55 16 33518350.E-mail address: [email protected] (H. Zanin).

013-4686/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.ttp://dx.doi.org/10.1016/j.electacta.2013.12.024

activity than those modified with randomly oriented nanotubes[8–10]. This observation is due to the presence of a large num-ber of carbon nanotubes containing their free tips per area unit,which in vertically aligned carbon nanotubes based electrodesincreases electron transfer rate. It has been demonstrated thatthe electron transfer of carbon nanotubes occurs preferentially attheir tips [8,11,12]. Furthermore, the electrochemical behaviourof carbon nanotubes electrodes is improved significantly after apre-treatment in order to insert oxygen functional groups on itssurface [8,11,12]. The oxygen functional groups onto the carbonnanotubes structure contributes to improve the electron transferfrom the nanotubes, which occur through the oxygenated sites[8,13]. However, Nugent and co-workers [14] showed ideal electrontransfer on carbon nanotubes microeletrodes without any pretreat-ments. Although such good results obtained by Nugent et al. [14],raw carbon nanotubes are normally hydrophobic, which means lowwettability and therefore high impedance for charge transfer.

In this paper, we present a fast and efficient method for fab-rication of superhydrophilic vertically aligned carbon nanotubesand graphene oxide nanocomposite with excellent electrochemicalproperties as electrode material.

Page 2

mica A

2

2g

((dinVs5iodopaAc(

T.A. Silva et al. / Electrochi

. Experimental

.1. Synthesis of vertically aligned carbon nanotubes andraphene oxide nanocomposite

The VACNT films were produced using a microwave-plasmaMWCVD) chamber at 2.45 GHz [15]. Substrates were Ti sheets10 mm × 10 mm × 0.5 mm) covered with a 10 nm Ni layereposited by electron-beam evaporation. The Ni layer was heated

n a N2/H2 (10/90 sccm) plasma which caused it to ball up intoanoclusters that subsequently became the catalyst particles forACNT growth. The nanocluster formation took place as the sub-trate temperature increased from 350 to 800 ◦C over a period of

min. To grow the VACNT forest, CH4 (14 sccm) was introducednto the chamber for 1 min, maintaining a substrate temperaturef 800 ◦C. The reactor pressure was 30 torr during all proce-ures. The carbon nanotubes exfoliation and the incorporation ofxygen-containing groups (-OH, -COOH, =O) were performed in aulsed-DC plasma reactor with an oxygen flow rate of 1 sccm, at

pressure of 150 mTorr, −700 V, at a frequency of 20 kHz [16].

fter this functionalization the VACNT samples are named verti-ally aligned carbon nanotubes and graphene oxide nanocompositeVACNT-GO).

Figure 1. SEM images of the vertically aligned carbon nanotubes

cta 119 (2014) 114– 119 115

2.2. Materials characterization

The samples were characterized by X-ray photoelectron spec-troscopy (XPS), surface wettability, scanning electron microscopy(SEM), Energy dispersive and Raman spectroscopies. We employedRaman Scattering Spectroscopy (Renishaw 2000 system), withexcitation by Ar + -ion laser (�=514.5 nm) in backscattering geom-etry to analyze the structural changes on samples. The curvefitting and data analysis software Fityk assigned the peak loca-tions and corresponding fitting of all spectra. The surface areameasurements were carried out using Quantachrome NovaWinmodel 1000 for multi-point BET using the classical helium void vol-ume method. Morphological images were performed using a fieldemission scanning electron microscope (FEI Inspect F50) to eval-uate structural arrangements and monitor modifications detailson the surface morphology. A Krüss Easy Drop system in sessiledrop method measured the contact angle (CA) using high puritydeionised water at room temperature to evaluate the wettabilityof as-grown VACNT and superhydrophilic VACNT-GO films. X-rayphotoelectron spectroscopy (VSW - HA100, using Al (K�) radiation,

1486.6 eV) was used to identify the oxygen content of the sam-ples in chemical bonds. All measurements were conducted at roomtemperature.

films (a-e) after oxygen plasma treatment and (f) as-grown.

Page 3

116 T.A. Silva et al. / Electrochimica Acta 119 (2014) 114– 119

Figure 2. Optical microscopy images of the contact angle between deionized water (magnification 200 × ) of (a) as-grown VACNT and (b) carbon nanotubes and grapheneo

2e

lwicmecFar

3

3

vvso(giengtpettwolebw

rt

xide nanocomposite.

.3. Preparation of VACNT-GO working electrode andlectrochemical assays

The Ti substrate containing the VACNT-GO film was immobi-ized on a copper plate using conductive silver paste, and used as

orking electrode. A geometric area of 1 cm × 1 cm was employedn all electrochemistry assays. This material was electrochemicallyharacterized using the probe Fe(CN)6

4−/3−. The cyclic voltam-etry (CV) and electrochemical impedance spectroscopy (EIS)

xperiments were performed using an Autolab PGSTAT-30 (Eco-hemie) potentiostat/galvanostat controlled with the GPES 4.9 orRA software. All electrochemical experiments were carried out in

Pyrex three-electrode cell made, with an Ag/AgCl (3.0 mol L−1 KCl)eference electrode and a Pt foil as auxiliary electrode.

. Results and Discussions

.1. Material characterization

Figure 1 (a-e) show the typical SEM images of the exfoliatedertically aligned carbon nanotubes films. Figure 1 (f) shows a sideiew image of as-grown vertically aligned carbon nanotubes, pre-enting a highly dense forest of tubes, high porosity, film thicknessf around 40 �m with a carpet-like structure at top view. Figure 1a) is top view image of carpet-like carbon nanotubes after oxy-en plasma exfoliation. The Figures 1 (b-e) show higher resolutionmages of those carbon nanotubes, which reveal oxygen plasmatching as able to exfoliate its tips. It is well known that carbonanotubes are graphene sheets rolled into cylinders [17]. Also,raphene sheet rolling up into a tube is a standard picture to illus-rate carbon nanotube formation [17]. We observed that oxygenlasma exfoliated the carbon nanotubes tips, opening its walls andxposing its fundamental structure: graphene (Fig. 1 b-e). Nano-ube tips are defective and interact directly with plasma, causingips exfoliation. Depending on plasma conditions we could etch thehole nanotubes or simply attach oxygen groups to them with-

ut exfoliation. At higher plasma pressures (80-180 mTorr) or foronger process times erosion was complete. With diluted plasmaxfoliation do not show up but wettability increases significantlyecause of oxygen groups attachment. The VACNT-GO composites

ere achieved at quite specific conditions.

We observed the VACNT films have high tube crystallinity, lowesidues of amorphous carbon content, no catalyst metallic par-icles outside the tubes and high porous structure, as we already

showed in our previous works [15,16,18]. The BET surface area ofVACNT is around 930 m2/g, which is consistent with SEM images.

The comparison of surface wettability of as-grown VACNT andthe sample after oxygen plasma treatment using a polar liquid(water) is presented in the Figure 2 (a, b). Notice that the surfaceof as-grown VACNT films exhibited superhydrophobic behaviour(Fig. 2 a), presenting a contact angle of ∼157◦ for dropped water.After oxygen plasma functionalization, the nanocomposite haspolar characteristic and can interact strongly with water (Fig. 2 b).Based on these low contact angles, we show that the oxygen plasmachanges the hydrophilicity of the sample surfaces.

The oxygen plasma treatment cause chemical modification inthe VACNT sample as can be seen in Raman spectra. Figure 3 (a-c)show first- and second-order Raman scattering spectra of VACNTfilms before and after plasma treatment. The deconvolutions wereperformed using Lorentzian shapes for the D, G and G’ bands, andGaussian shape for bands around 1250 (#), 1480 (*) and 1611 cm−1

(D’ shoulder) [15–19]. The D band is usually attributed to the dis-order and imperfection of the carbon crystallites. The G band isassigned to one of the two E2g modes corresponding to stretch-ing vibrations in the basal plane (sp2 domains) of single crystalgraphene [15,16]. The high intensity G’ band reveals that thesematerials present high structural quality [18,19]. In the VACNT-GO first order Raman, for appropriated deconvolution fitting, twoGuassian peaks centered at around 1250 and 1480 cm−1 wereadded necessarily. Probably the shoulder has its origin in doubleresonance process, because its Raman shift (∼1200 cm−1 #) is apoint on graphene phonon dispersion curves [19]. The origin of the1480 cm−1 (*) Band is probably correlated with the polar groupsgrafting onto CNT surfaces [20].

As known, integrated intensity ratio AD/AG in the Raman spec-trum should approximately correspond to the extent of disorderin the graphitic carbon, the AG′ /AG has direct relation with sam-ple crystallinity, the A*/AG has correlation with the polar groupsattached onto sample surfaces and A#/AG may be correlated tostructural changes. The Table 1 presents the comparison betweenthose Bands areas extracted from Raman spectra (514.5 nm) byFityk fitting using Gaussian and Lorentzian equations and pre-sented in Figure 3 (b & c). From Table 1 we observed that the AD/AGslightly increases from 0.54 (VACNT) to 0.58 (VACNT-GO), A*/AGincreases from 0 (VACNT) to 0.05 (VACNT-GO) and A#/AG increases

from 0 (VACNT) to 0.08 (VACNT-GO), after functionalization. Inaddiction the AG′ /AG slightly decrease from 1.78 (VACNT) to 1.54(VACNT-GO) after functionalization [19]. It means slight increaseof defects and loss of crystallinity, however all those parameters

Page 4

T.A. Silva et al. / Electrochimica Acta 119 (2014) 114– 119 117

Table 1Comparison between Bands areas extracted from Raman spectra (514.5 nm) by Fityk fitting using Gaussian and Lorentzian equations.

Sample AD/AG AG′ /AG A*/AG A#/AG

VACNT (21.83)/(40.17) = 0.54 (71.63)/(40.17) = 1VACNT-GO (28.64)/(48.56) = 0.58 (75.15)/(48.56) = 1

Figure 3. (a) First and second-order Raman spectra of vertically aligned carbonno

d(ss

GaTssp(ttOt

neous electron transfer rate constant to quasi-reversible systemscontrolled by diffusion [25]. In this method, the k0 value is obtained

anotubes films before (b) and after (c) oxygen plasma treatment exfoliation andxidation.

epend of the exposition time to oxygen plasma. The presence of*) Band reveals the carboxylic groups attached onto VACNT-GOurfaces [20]. The increase of (#) band may be attributed to thetructural change caused by the exfoliation to graphene sheets.

To better identify specific carboxylic groups attached on VACNT-O, the C1s and O1s XPS spectra were taken before (Fig. 4 a & b) andfter (Fig. 4 c & d). Deconvolutions were performed for both cases.he spectra were deconvoluted assuming a Lorentzian–Gaussianum of functions (20% Lorentzian maximum contribution) by Fitykoftware [20]. Figure 4 (a & c) present the C1s peaks decom-osed into five Gaussian components, referring to the bonds: C = C∼284.9 eV), C O (∼285.8 eV), C = O (287.5 eV), COO (289.3 eV), andhe last one at 290.8 eV assigned to the shake-up peak (�–�*

ransitions) [20,21]. Figure 4 (b & d) show deconvolution of the1s spectra using three peaks. First is in the range from 531.0

o 531.4 eV, attributed to oxygen doubly bound to carbon, the

.78 - -

.54 (2.68)/(48.65) = 0.05 (3.69)/(48.65) = 0.08

second in the range from 534.6 to 536.1 eV, attributed to oxygensingly bound to carbon and third at 534 eV, attributed to carboxyl[21]. With our approaches, the oxygen content increased from 4.9%(Fig. 4a) to 21.4%, after the oxygen plasma treatment (Fig. 4c), whichwas calculated as performed by Payne et al. [22]. In summary, wepoint out that the oxygen plasma is able to attach oxygen func-tional groups onto carbon nanotubes and changes its wettability.Furthermore, in the conditions shown the oxygen plasma is able toexfoliate carbon nanotubes into graphene and concurrently oxideit as well.

3.2. Electrochemical behaviour of VACNT-GO electrode

The electrochemical behaviour of VACNT-GO electrode wasstudied using as electrochemical probe the redox coupleFe(CN)6

4−/3−. Figure 5 presents the cyclic voltammogram obtainedfor a 1 × 10−3 mol L−1 K3Fe(CN)6 solution in 0.1 mol L−1 KCl usingthe VACNT-GO electrode. As can be seem, the redox couple pre-sented a perfect reversible behaviour, with a �Ep value equal to59 mV, as proposed for an ideal reversible process [23,24], indi-cating that the VACNT-GO electrode promoted a rapid electrontransfer. Cyclic voltammetry experiments at various scan rates(Fig. 6 a) were performed (10–500 mV s−1), in order to deter-mine the electroactive surface area and the heterogeneous electrontransfer rate constant at VACNT-GO electrode, and thus explorein more details the good electrochemical behaviour previouslyobserved for the proposed material. The electroactive surface areawas determined using the Randles-Sevcik equation (Eq. 1):

Ip = ±(

2.69 × 105)n3/2AD1/2Cv1/2 (1)

where Ip is the peak current (anodic or cathodic), n is the number ofelectrons of the redox process, A is the electroactive surface area, Dis the diffusion coefficient (7.6 × 10−6 cm2 s−1) [24], C is the concen-tration of the electroactive specie (1.0 �mol cm−3) and v1/2 is thesquare root of the scan rate. Figure 6 (b) shows that the anodic (Ia)and cathodic (Ic) peak currents varied linearly with the square rootof the scan rate, indicating that the mass transport of species ontothe electrode surface was controlled by diffusion, as expected forthe probe used. The Ip vs. v1/2 plots for the anodic and cathodic cur-rent peaks presented very similar slopes: −5.92 × 10−4 A (V s−1)−1/2

and 5.78 × 10−4 A (V s−1)−1/2. Thus, we calculated the electroac-tive surface area for the VACNT-GO electrode taking into accountboth Ip vs. v1/2 plots, obtaining an average value of 0.79 ± 0.01 cm2.Therefore, approximately 80% of the geometric area of the VACNT-GO electrode is available for the electron transfer processes. Thispercentage is relatively high, and very relevant for applications inelectroanalysis, where a high electroactive surface area can lead toa significant increase in the analytical signal (peak current) and soprovide better analytical parameters, such as high sensitivity andlow limits of detection and quantification.

The heterogeneous electron transfer rate constant (k0) to theVACNT-GO electrode was also calculated. For this, we used theNicholson method, widely used for the calculation of heteroge-

through the Eq. 2:

= k0[�DnvF/ (RT)

]−1/2(2)

Page 5

118 T.A. Silva et al. / Electrochimica Acta 119 (2014) 114– 119

tment. O1s XPS peak analysis before (b) and after (d) the oxygen plasma treatment.

wmpe

seDTtte(b1t(h

v

F(o

Figure 4. C1s XPS peak analysis before (a) and after (c) the oxygen plasma trea

here � is a kinetic parameter, and the other terms have their usualeanings. The � values can be easily obtained using the Eq. 3 pro-

osed by Lavagnini et al. [26], which relate � and the �EP value forach scan rate:

=(−0.6288 + 0.0021�Ep

)/(

1 − 0.017�Ep)

(3)

From the calculated � values, k0 parameter follows from thelope in the graphic of � vs.32.79 v−1/2 plotted. The 32.79 factor isquivalent to the term [� D n F/(R T)] −1/2, calculated considering

= 7.6 × 10−6 cm2 s−1, F = 96485 C mol−1, R = 8.314 J K−1 mol−1 and = 298.15 K. The linear regression equation gives a k0 value equalo 3.4 × 10−2 cm s−1. To get an idea of the magnitude of this elec-ron transfer rate constant, the k0 value obtained for the VACNT-GOlectrode was around 94 higher than for a carbon paste electrode3.6 × 10−4 cm s−1) [26], 33 times higher than for a glassy car-on electrode (1.02 × 10−3 cm s−1, determined in this work) and1 times higher than for vertically-aligned carbon nanotube elec-rode covalently attached to p-type silicon via a thioester linkage

2.98 × 10−3 cm s−1) [27]. This result proves VACNT-GO electrodeas great ability to transfer electron.

The stability of the VACNT-GO electrode was studied by cyclicoltammetry employing a 1 × 10−3 mol L−1 K3Fe(CN)6 solution in

igure 6. (a) Cyclic voltammograms obtained for 1.0 mmol L−1 K3Fe(CN)6 in 0.1 mol L−1

d) 40, (e) 50, (f) 75, (g) 100, (h) 150, (i) 200, (j) 250, (k) 300, (l) 350, (m) 400, (n) 450 anbtained for 1.0 mmol L−1 K3Fe(CN)6 in 0.1 mol L−1 KCl using the VACNT-GO electrode.

Figure 5. Cyclic voltammogram (v = 50 mV s−1) obtained for 1.0 mmol L−1 K3Fe(CN)6

in 0.1 mol L−1 KCl using the VACNT-GO electrode.

0.1 mol L−1 KCl at a fixed scan rate, 100 mV s−1. Thus, 100 cycles

were taken and the result is presented in the Figure 6 (c). Thecyclic voltammograms took from Fe(CN)6

4−/3− on VACNT remainedalmost constant during the various potential cycles. The �Ep

value did not change and the anodic and cathodic peak currents

KCl using the VACNT-GO electrode in different scan rates (v): (a) 10, (b) 20, (c) 30,d (o) 500 mV s−1. (b) Ip vs. v1/2 curves. (c) Cyclic voltammograms (v = 100 mV s−1)

Page 6

T.A. Silva et al. / Electrochimica Acta 119 (2014) 114– 119 119

ned fo

srV

wTadtrlTcttiititfnotliaacsspalrwtadc

4

vpbetc

[

[

[

[

[[

[

[

[

[

[

[

[

[

[

[[[

[

Figure 7. (a) Nyquist plot (from 10 mHz to 100 kHz) and (b) Bode plot obtai

howed relative standard deviations (RSD) of only 0.33% and 0.29%,espectively. These results demonstrate the high stability of theACNT-GO electrode.

The electrochemical behaviour of the VACNT-GO electrode stillas assessed by electrochemical impedance spectroscopy (EIS).

he Nyquist and Bode plots obtained for the VACNT-GO electrodere shown in Figure 7 (a, b). Nyquist plot usually include twoistinct regions, a semi-circular at higher frequencies, related tohe electron transfer processes, where the semicircle diameter iselated to the electron transfer resistance, and a linear region atower frequencies related to diffusion controlled processes [28,29].he Nyquist plot recorded for the VACNT-GO electrode is practi-ally linear throughout the frequency range studied, showing thathe VACNT-GO electrode has high ability to transfer electrons, andhe process of the redox probe is controlled by diffusion. To betternterpret the Nyquist and Bode plots they have been fitted as mod-fied Randles equivalent electrical circuits [30] (inserted). It modelshe electrochemical electrode as an active electrolyte resistance RSn series with a limiting capacitance, and in parallel with a combina-ion of the double-layer capacitance Cdl, and the impedance of thearadaic reaction, called the Warburg impedance, ZW. Rs is the inter-al resistance of the system, which consists of the ionic resistancef the electrolyte, the intrinsic resistance of the active material andhe contact resistance at the electro-active material/current col-ector interface. ZW represents the hindrance to mass transfer ands related to the rate of diffusion to and from the electrode by ionsnd electrons. Because the electrode surface is not perfectly smoothnd infinitely large, it will not behave like a perfect capacitor, andharging of the electrical double layer will be non-faradaic. Thetandard way to account for the microscopic roughness and atomiccale inhomogeneity in the surface is to replace Cdl, with a constanthase element, CPE, given by CPE = Q1/n with Q being the chargend n > 0.9. Thus, CPE is the effective capacitance of the doubleayer for nanostructured electrodes. The interfacial charge-transferesistance, Rct, was expected in series with ZW. The impedance dataere fitted using the FRA software using RS, Cdl, ZW and Rct as fit-

ing parameters. The average results are RS = 118 �; Cdl = 0.63 mFnd ZW = 0.22 m � s−0.5 and Rct = 0, which is consistent with the CVata, where is clear a fast charge transfer and low impedance toharge transfer.

. Conclusions

In this work we showed the fast electron transfer kinetics onertically aligned carbon nanotubes and graphene oxide nanocom-osite as electrode material. The nanocomposite was synthesized

y CVD techniques. The oxygen plasma treatment provided anfficient functionalization of this new nanomaterial. The high elec-roactive surface area and heterogeneous electron transfer rateonstant demonstrates the excellent electrochemical behaviour of

[

[

r 1.0 mmol L−1 K3Fe(CN)6 in 0.1 mol L−1 KCl using the VACNT-GO electrode.

the VACNT-GO electrode. Future applications in electrocatalysisand (bio)sensors are underway in our groups.

Acknowledgments

The authors are very grateful to the financial support fromthe Brazilian funding agencies CNPq (202439/2012-7), FAPESP andCAPES.

References

[1] N. Punbusayakul, Procedia Eng. 32 (2012) 683–689.[2] J. Wang, Electroanalysis 17 (2005) 7–14.[3] J.J. Gooding, Electrochim, Acta 50 (2005) 3049–3060.[4] P. Lv, P. Zhang, F. Li, Y. Li, Y. Feng, W. Feng, Synth. Met. 162 (2012) 1090–1096.[5] R. Reit, J. Nguyen, W.J. Ready, Electrochim, Acta 91 (2013) 96–100.[6] N. Rajalakshmi, H. Ryu, M.M. Shaijumon, S. Ramaprabhu, J. Power Sources 140

(2005) 250–257.[7] T. Gokus, R.R. Nair, A. Bonetti, M. Boehmler, A. Lombardo, K.S. Novoselov, A.K.

Geim, A.C. Ferrari, A. Hartschuh, Acs Nano. 12 (2009) 3963–3968.[8] A. Chou, T. Bocking, N.K. Singh, J.J. Gooding, Chem. Commun. (2005) 842–844.[9] J.J. Gooding, A. Chou, J. Liu, D. Losic, J.G. Shapter, D.B. Hibbert, Electrochem.

Commun. 9 (2007) 1677–1683.10] F.C. Moraes, T.A. Silva, I. Cesarino, S.A.S. Machado, Sens. Actuators, B 177 (2013)

14–18.11] C.E. Banks, T.J. Davies, G.G. Wildgoose, R.G. Compton, Chem. Commun. (2005)

829–841.12] C.E. Banks, R.R. Moore, T.J. Davies, R.G. Compton, Chem. Commun. (2004)

1804–1805.13] P.J. Britto, K.S.V. Santhanam, A. Rubio, J.A. Alonso, P.M. Ajayan, Adv. Mater. 11

(1999) 154–157.14] J.M. Nugent, K.S.V. Santhanam, A. Rubio, P.M. Ajayan, Nano Lett. 1 (2001) 87–91.15] H. Zanin, E. Saito, H.J. Ceragioli, V. Baranauskas, E.J. Corat, Mater. Res. Bull. 49

(2014) 487–493.16] A.O. Lobo, H. Zanin, I.A.W.B. Siqueira, N.C.S. Leite, F.R. Marciano, E.J. Corat, Mater,

Sci. Eng., C 33 (2013) 4305–4312.17] Kaiyuan Li, Gyula Eres, Jane Howe, Yen-Jun Chuang, Xufan Li, Zhanjun Gu, Litong

Zhang, Z. Sishen Xie, Pan Sci. Rep. 3 (2013) 1–4.18] J. Tsukada, H. Zanin, L.C.A. Barbosa, d.G.A. Silva, H.J. Ceragioli, A.C. Peterlevitz,

R.F. Teófilo, V. Baranauskas, J. Electrochem. Soc. 159 (2012) 159–161.19] M.S. Dresselhaus, A. Jorio, M. Hofmann, G. Dresselhaus, R. Saito, Nano Lett. 10

(2010) 751–758.20] A.O. Lobo, S.C. Ramos, E.F. Antunes, F.R. Marciano, V.J. Trava-Airoldi, E.J. Corat,

Mater. Lett. 70 (2012) 89–93.21] V. Datsyuk, M. Kalyva, K. Papagelis, J. Parthenios, D. Tasis, A. Siokou, I. Kallitsis,

C. Galiotis, Carbon 46 (2008) 833–840.22] B.P. Payne, M.C. Biesinger, N.S. McIntyre, J. Electron Spectrosc. Relat. Phenom.

184 (2011) 29–37.23] C.M.A. Brett, A.M.O. Brett, Electrochemistry: Principles, Methods, and Applica-

tions Oxford Science Publications, Oxford, 1993.24] A.J. Bard, L.R. Faulkner, Electrochemical Methods: Fundamentals and Applica-

tions, JOHN WILEY; & SONS, INC, New York, 2001.25] R.S. Nicholson, Anal. Chem. 37 (1965) 1351–1355.26] I. Lavagnini, R. Antiochia, F. Magno, Electroanalysis 16 (2004) 505–506.27] Z. Poh, B.S. Flavel, C.J. Shearer, J.G. Shapter, A.V. Ellis, Mater. Lett. 63 (2009)

757–760.28] R. Pauliukaite, M.E. Ghica, O. Fatibello-Filho, C.M.A. Brett, Anal. Chem. 81 (2009)

5364–5372.29] R. Pauliukaite, M.E. Ghica, O. Fatibello-Filho, C.M.A. Brett, Electrochim, Acta 55

(2010) 6239–6247.30] H. Ghadimi, R.M.A. Tehrani, A.S.M. Ali, N. Mohamed, S. Ab Ghani, Anal. Chim.

Acta 765 (2013) 70–76.