Since the first isolation of graphene in 2004, graphene-based supercapacitors have attracted remarkable attention due to the capability of energy storage and delivery [1-5]. A supercapacitor is a conventional capacitor with higher capacitance for energy storage by values more than hundreds of times per unit mass. Advantages of these devices such as long cyclic life, short charge/discharge time, high capacitance and power motivate researchers to explore their capacitance and stability [4-7]. Supercapacitors store energy by electrical double layer capacitance or pseudocapacitance and these devices are often fabricated based on carbon materials, including activated carbon, carbon nanotube and graphene due to their special properties such as high surface area, relatively low cost, non-toxicity, high chemical stability and wide temperature range [8-11].
Graphene with high surface to volume ratio (theoretical 2630 m2/g), high charge carrier mobility as well as its chemical stability has been subject of recent research for application in flexible electrodes [3, 12]. Composites of metal oxides and graphene or carbon nanotube are interesting materials for electrodes of pseudocapacitors and hybrid capacitors which contribute in redox reactions and increase specific capacitance [13-16]. Over the last few years, it has been reported that mixture of graphene and conductive polymers can enhance capacitance and energy density of electrodes considerably. For example, doping polyaniline with graphene quantum dot exhibit a specific capacitance value of ∼1044 F/g by a current density of 1 A/g as well as better stability . Also, fabrication of electrodes using functionalized microporous carbon shows high specific capacitance and good stability . Recently, binder-free graphene-based papers have been applied in electrodes of supercapacitors due to their low weight and flexibility [3-5]. Facile and cost-effective method for fabrication of a high-performance electrode is a challenge in the field of supercapacitors. Several methods have been proposed for fabrication of graphene-based supercapacitor electrodes . Exfoliation and then restacking of two-dimensional layered materials has attracted remarkable attention for fabrication of high performance electrodes [20, 21]. A common method is vacuum filtration deposition of graphene suspension to fabricate graphene-based electrode for supercapacitor . Intercalation of ions into layered materials is another approach for fabrication of nanodevices for energy storage, including capacitors and batteries [20, 21, 23]. The interconnecting of carbon nanotubes and graphene is a novel approach to fabricate high-performance materials . It has been reported that improvement in capacitance of supercapacitors based on reduced graphene oxide (RGO) is due to the adsorption of cations to the surface of sheets, while intercalation of carbon nanotubes improves the capacitance and cyclic stability in comparison with electrode based on the pure RGO . However, fabrication of a stable, high-performance electrode with low resistance by a one-step and low cost method is a challenge yet.
In order to use graphene as a primal material for fabrication of the electrode, it is important to be produced by a facile and cost-effective method. A typical and low-cost approach for mass production of graphene is chemical method which includes oxidation and exfoliations of natural graphite. This method produces graphene oxide which is a graphene with oxygen functional groups on its edges and basal planes . Another challenge in the fabrication of graphene based electrodes is the reduction process. Although several reducing agents such as hydrogen sulfide, hydrazine, NaBH4, dimethylhydrazine, hydroquinone were reported to reduce the graphene oxide, non-toxicity, efficiency and cost of the reducing agent should be considered. It has been reported that hydroiodic acid is an efficient reducing agent to obtain RGO paper with high conductivity . Hydroiodic acid reduces graphene oxide paper by replacing the oxygen functional groups with organohalides and produces high-quality RGO paper . A hindrance for application of graphene as an electrode of supercapacitor is restacking of the sheets. Reducing graphene oxide during electrode preparation prevents RGO sheets from crumpling and restacking, which improves the performance of supercapacitor. In addition, separator of a supercapacitor should be electrically dielectric, but ion permeable. It seems that using graphene oxide (GO) with 1012 Ω/sq resistance as a separator can make supercapacitor fabrication more cost-effective and facile with high performance . Another determining factor in performance, energy density and capacitance of supercapacitors is electrolyte. It has been suggested that higher capacitance could be achieved in aqueous alkaline or acid solutions compared with organic electrolytes . Alkaline electrolytes such as NaOH and KOH have several advantages such as high ionic conductivity, high thermal stability, low cost, inflammability and non-toxicity that make them suitable candidates as electrolytes for supercapacitors.
In this research, we present a facile and one-step method to fabricate high performance RGO-based electrode for supercapacitor. Prepared graphene oxide sheets were reduced by chemical method using hydroiodic acid. The electrodes were fabricated by drying the carbon black (CB) nanoparticles-graphene solution. We investigated the effect of intercalation of CB nanoparticles between RGO sheets in electrochemical performance using electrochemical impedance spectroscopy (EIS) and cyclic voltammetry (CV). In addition, the application of graphene oxide paper as a separator was examined.
MATERIALS AND METHODS
GO was synthesized using Hummers’ and Offman method [28-30]. This method was performed by adding 46 mL H2SO4 to 1 g of natural graphite (Asbury Carbons, USA) and then was stirred at 80˚C for 1 h. Then 1 g NaNO3 was added to the solution under stirring condition in an ice bath. As the solution was stirring, 4g KMnO4 was added slowly. The solution was reached to room temperature, then was placed in an oil bath at 40˚C and was stirred for 1.5 h. The solution was heated up to 90˚C and then diluted by 280 mL deionized water. Finally, 6 mL H2O2 was added to reduce residual KMnO4. The residual acids and salts were removed by washing and filtering. The exfoliated sheets were separated by centrifugation at 4000 RPM for 20 min.
Electrode and GO paper fabrication
In order to reduce graphene oxide, 0.5 mL HI (57% Merck) was added to 5 mL GO (6 mg/mL) and was stirred for 30 min. Then the solution was coated on a steel sheet (0.1 mm in thickness and 2.6 cm in diameter) as current a collector and was dried at 90˚C for 2 h. After drying, the electrodes were washed with acetone and deionized water. In the case of pillared RGO, 5 wt%, 15 wt% and 20 wt% of CB nanoparticles (Degussa D-60287, Germany) were dispersed in dimethylformamide by the aid of 30 min sonication and were added to the GO solution under the stirring condition before adding HI. When the electrodes were fabricated by casting method with 5 mg mass loading, supercapacitors were assembled by prepared electrodes using two kinds of separators of filter paper (filtraTech) and GO paper with the thickness of 50 µm. The GO paper was prepared by vacuum filtration of 4 mL of GO (5mg/mL) through filter film (0.45μm pore size and 45mm in diameter). Following this, the GO paper on the filter was dried in oven for 2h at 60˚C and then peeled from the filter.
Electrochemical capacitance performance of supercapacitor electrodes was investigated by cyclic voltammetry (CV), galvanostatic charge/discharge (GCD) and electrochemical impedance spectroscopy (EIS) by the use of a potentiostat from PGSTAT Autolab in 30 wt% KOH electrolyte. EIS measurements were performed using a 0.1 V DC bias voltage with a sinusoidal signal with the amplitude of 5 mV in the frequency ranged from 1 mHz to 10 kHz. Raman spectroscopy was carried out at room temperature using Almega Thermo Nicolet Dispersive Raman Spectrometer with excitation wavelength of 532 nm. Fourier transform infrared (FTIR) analyses were performed using Bruker-tensor27. The microstructure and morphology of the RGO and RGO/CB electrodes before and after the electrochemical tests were characterized by scanning electron microscopy (SEM) using Tescan Mira II and transmission electron microscopy (TEM) using Zeiss - EM10C at voltage of 80 kV. In addition, structure of RGO/CB was studied by x-ray diffraction (XRD) using PHILIPS-binary diffractometer by a Cu anode in the range of 5°-80°. Size of CB nanoparticles was measured by dynamic light scattering (DLS) method using Malvern instrument. The N2 adsorption–desorption isotherms of the samples were measured at 77 K using Belsorp mini II (BEL, Japan) to determine the specific surface areas. The specific surface area was calculated from the Brunauer–Emmett–Teller (BET) plot of the nitrogen adsorption isotherm.
RESULTS AND DISCUSSION
Raman spectroscopy was utilized to explore the layer number and the crystal quality of GO sheets. There are two significant peaks in Raman spectra of the carbon materials so-called G and D band which appear at ~1580 cm-1 and ~1350 cm-1, respectively . The G band usually is attributed to the E2g phonon of graphitized structure and the D band is corresponds to breathing mode of κ-point phonons of A1g which is related to the disorder and defects [31-33]. In Raman spectra, peak position is dependent on graphene layers. G band of single layer graphene (1585 cm-1) shifts about 6 cm-1 to lower frequencies after stacking more graphene layers. Also, 2D band of single layer graphene located at 2679 cm-1 is sensitive to the number of graphene layers and shifts about 19 cm-1 into the larger wavenumbers [29, 31-33]. It is well known that the 2D/G intensity ratio of single, double, triple and multi (>4) layer graphene is typically greater than 1.6, 0.8, 0.3 and 0.07 respectively [32, 33]. Fig. 1 shows Raman spectra of synthesized graphene oxide and RGO with 15 %wt intercalated CB. The I2D/IG ratio of 0.19 shows that mean layer number of GO is about 4. According to the figure, the G and the D peaks appear at ~1630 cm-1 and ~1380 cm-1, respectively. The ID/IG ratio increases from ~0.97 to ~1.56 and I2D/IG ratio decreases from 0.19 to 0.16 by intercalation of CB nanoparticles, which illustrates an increase in sp3 hybridization due to the presence of CB nanoparticles and defect creation.
FTIR spectroscopy has been utilized to examine the reduction of GO by hydroiodic acid and the chemical components of the electrode materials before and after electrochemical tests. In a typical spectrum of GO, the stretching mode of hydroxyl, carbonyl (C=O) and epoxide (C–O) appears at ~3450, ~1728 and ~1061 cm-1 respectively[29, 34]. Fig. 2 demonstrates the FTIR spectra of GO and RGO. It is obvious that all the peaks related to functional groups have been eliminated in RGO spectra and this verifies the reduction of GO. FTIR spectra of RGO and RGO/CB electrodes before and after the electrochemical tests were presented in Fig. 3. No significant changes are seen in the number of peaks in the FTIR spectra of RGO electrode after the electrochemical tests. However, it is clear that three peaks are appeared in the spectra of RGO/CB electrode after the electrochemical tests in 661cm-1, which are attributed to O1CO2 bending modes, 702cm-1 related to C=O in-plane bending modes coupled with O…H stretching modes and the peak at 831 cm-1 is due to out-of-plane bending mode of the CO3 skeleton which are undoubtedly because of the effect of KOH on RGO/CB . According to the FTIR spectra of RGO/CB electrode after the test, it could be concluded that a KHCO3 structure is formed from the mixture of KOH and RGO/CB . This phenomenon could be explained by the presence of carbon black nanoparticles in RGO/CB electrode which not only play an important role in the absorption of KOH, but also could be an effective factor in the formation of new functional groups. XRD analyses of RGO/CB electrodes before and after the electrochemical tests are shown in Fig. 4. In the XRD pattern of the RGO/CB electrode before the tests there is a small peak at 12° which could be attributed to GO and might be because of functional group as a result of incomplete reduction. Also there is a sharp peak at 24° due to RGO/CB . However, XRD pattern for the RGO/CB electrode after tests confirms the structure of KHCO3. Table 1 and 2 present XRD data for the RGO/CB electrode after test and XRD data for KHCO3 crystal, respectively . The microstructure and morphology of RGO and RGO/CB electrodes before and after the electrochemical tests were characterized by SEM and TEM. According to SEM images in Fig. 5, it is clear that morphology of the electrodes becomes more porous by adding CB nanoparticles. This porous structure facilitates electrolyte ions absorption, causing to higher capacitance. Furthermore, it can be seen from the SEM images that the microstructure and morphology of the electrodes do not modify during the electrochemical tests. As shown in TEM image (Fig. 6(a)) there is a flat RGO sheet with some wrinkles on its surface, however in Fig. 6(b) nanoparticles were distributed on the RGO sheets, preventing them from restacking. By restacking RGO sheets in the electrode, effective surface is reduced and ions cannot move between the sheets. Consequently, restacking of RGO sheets affects performance of supercapacitor by an increase in charge/discharge time and a decrease in capacity. The prepared RGO paper and supercapacitor are presented in Fig. 6(c) and 6(d).
In order to measure the size of carbon black nanoparticles, DLS analysis was performed. Based on Fig. 6(c), the mean particle size of the CB is about 90 nm which is more than the TEM results. Hydrodynamic radius of nanoparticles in DLS measurements is typically greater than the real size of nanoparticles and therefore we expect the greater size for CB nanoparticles by DLS. Nevertheless, as shown in the TEM image most of the CB nanoparticles have a size of about 12 nm. However, the increase in the DLS particle size cannot only be due to the hydrodynamic radius. Therefore comparing particle size for TEM and DLS results indicates that few nanoparticles can agglomerate on graphene sheets.
CV is a powerful method to examine the performance and capacity of the prepared supercapacitor. As shown in Fig. 7, CV diagrams were broadened by increasing the scan rate. At higher scan rates more ions move toward the electrodes, resulting in higher current and wider CV diagram. As presented in Fig. 7, there is no Faradic peak in CV diagrams and charge storage occurs electro-statically. However, the CV diagram in Fig. 7(d) was not broadened by increasing the scan rate which indicates the weak capacitive behavior of sample G+20% CB. To discuss quantitatively, the specific capacitances were calculated by use of CV curve via the following formula :
where V1 and V2 are the first and the second potential (V1=0 and V2=1V) and I, m and are current, mass of each electrode (~5 mg) and scan rate, respectively. The calculated specific capacitances are presented in Fig. 8(a). As shown in this figure, at the scan rate of 10 mV/s the specific capacitance was calculated to be 118, 129, 139 and 125 F/g for intercalated RGO with 0, 5, 15, 20 wt% CB, respectively. Furthermore, the specific surface area measured by BET. Results show that RGO/CB is increased by 823%.This indicates that CB particles as nanoscale spacers can significantly decrease the agglomeration of graphene with high electrochemical utilization of graphene layers.
We also obtained 142 F/g by use of GO paper as a separator. Results revealed that the concentration of CB and GO paper separator are important factors for specific capacitance. The highest value of specific capacitance obtained for 15 wt% CB intercalated into RGO sheets. it means that CB nanoparticles act as spacers between the RGO sheets and also increase the conduction of electrodes with no rule in the formation of double layer, while adding more amount of CB decreases the specific capacity. In addition, the charge-discharge diagrams of the prepared samples were measured to explore the time of charging and discharging. Fig. 8(b) indicates that the time of charging for RGO-based supercapacitor is about twice that of the RGO/CB and the supercapacitor with GO paper separator has the lowest charge time. This can be due to a decrease in the electrode resistance by CB intercalation which also increases the porosity of electrodes. The intercalation of CB nanoparticles into RGO sheets makes a porous structure which facilitates the electrolyte ion movement, but also increases conductivity of the electrodes significantly, leading to decrease in measured resistance.
Resistivity behaviors of fabricated electrodes were examined by EIS measurements. Fig. 9 demonstrates Nyquist plots, i.e. the relationship between real (Z’) and imaginary (-Z”) part of impedance for supercapacitor. In the Nyquist plot, there is a semicircle at high frequencies and an approximately a straight line at low frequencies. The first point of the semicircle is the resistance of electrolyte and the diameter of the semicircle is charge transfer resistance. The straight line is related to Warburg resistance steamed from diffusion of ions [37, 38]. As explained, intercalation of CB nanoparticles into RGO sheets causes more conductivity between the electrodes. The semicircle parts of the Nyquist plot for electrodes with different percent of CB are compared in Fig. 9(b). Results show that by increasing the amount of CB, diameter of semicircle becomes smaller, meaning more conductivity in the electrodes.
We presented a facile method to improve performance of RGO based supercapacitor. Prepared RGO sheets were reduced chemically using hydroiodic acid. RGO paper was fabricated by drying and reducing the CB nanoparticles-RGO solution. We showed that the intercalation of CB nanoparticles into RGO sheets and use of GO as separator improves electrochemical performance of RGO paper and highest capacitance is obtained for RGO+15%wt CB. The electrochemical behavior of prepared electrodes was tested in 30wt% KOH solution and results verified that specific capacitance can increase from 118 to ~139 F/g and charge-discharge time becomes roughly half. In addition, EIS showed that electrode resistance decreases from 18 to ~6 and also ions diffusion occurs easily in the prepared electrodes.
CONFLICT OF INTEREST
The authors declare that there are no conflicts of interest regarding the publication of this manuscript.