Dependence of capacitive properties of an EDLC on exfoliation time of graphite electrodes

The scientific focus has been directed through the production and application of “wonder material-graphene” after its discovery in 2004. But the mass production cost has become a huge disadvantage towards commercializing graphene-based manufactures. As alternative low-cost material, exfoliated graphite (EG) has emerged to be a novel nanostructured carbon material with a potential for electrochemical energy storage device applications owed to its unique characteristics similar to graphene. In this study a series of EG samples were prepared by a surfactant-mediated liquid-phase exfoliation method by changing the exfoliation time. Electrochemical double-layer capacitors (EDLCs) were fabricated using different EG samples as an electrode material and a gel polymer electrolyte (GPE). They were characterized by electrochemical impedance spectroscopy (EIS), cyclic voltammetry (CV), and galvanostatic charge discharge (GCD) techniques. EDLC having EG electrodes of 10 h exfoliation time showed the highest results with single-electrode specific capacitance (Csc) of 4.12 F g−1, single-electrode specific discharge capacitance (Csd) of 1.10 F g−1, and relaxation time of 0.22 s from CV, GCD, and EIS, respectively.


Introduction
Exfoliated graphite (EG) denotes to graphite which has a degree of separation of carbon layers in the graphite.
'Exfoliation' is the process by which the separation occurs [1]. Several methods have been deployed to achieve graphite exfoliation, each with benefits and drawbacks [2]. Most commonly used methods include, mechanical exfoliation, chemical conversion, thermal exfoliation, electrochemical exfoliation and liquid phase exfoliation.
Liquid phase exfoliation is among the most inexpensive and eco-friendly processes for EG production. Graphite can be successfully exfoliated in liquid mediums by exploiting ultrasound to extract layers. The process comprises three main steps: dispersion of graphite in a solvent, exfoliation and purification [3].
Liquid phase exfoliation can be performed in various mediums such as organic solvents, aqueous surfactant solutions and ionic liquids [4]. Organic solvents like 1-methyl-2-pyrrolidinone (NMP), g-butyrolactone (GBL), N, N-dimethylacetamide (DMA), N-vinyl-2-pyrrolidone (NVP) and N, N-dimethylformamide (DMF) have surface energies similar to graphene. But the toxicity and high cost limited their use in mass production. Exfoliation of graphite in water with the aid of surfactants can avoid such complications. Surfactants support EG to remain suspended in aqueous mediums while providing a suitable solution for hydrophobic nature of EG sheets. Lotya et al.
reported a surfactant mediated liquid phase exfoliation using sodium dodecylbenzenesulfonate (SDBS) as surfactant [5]. Since then, a variety of cationic, anionic and non-ionic surfactants have been used. Nawaz et al. evaluated the effect of concentration of sodium cholate (SC) on the exfoliation of graphite in aqueous media [6]. These surfactants are not only used as intercalating agents but also as stabilizers to prevent EG from restacking in the solution.
EG has a higher volume than graphite due to porosity which occurs during local separation. And also, it exhibits ultra-light weight and high selection sorption capacity [7]. There are numerous research works demonstrating EG structure to be different, like a fluffy cellular structure or a worm-like structure, depending on its exfoliation degree [8]. Because of its uncommon structure, EG exhibits high specific surface area, improved porous structure, low density, high concentration of edges and surface defects and high thermal and chemical stability. These properties of EG that differ from conventional graphite have led to several applications in batteries and supercapacitors. EG and EG-based materials have been explored for use as electrode materials in practical applications, owing to their pristine and highly conducting nature [4].
Properties of EG will vary upon its synthesis method. Variables such as, exfoliation time, surfactant type, surfactant concentration and solvent will affect on the physical and chemical characteristics of EG. Considerable deviation of these factors can have a huge impact on the characteristic features of the devices that are fabricated with EG.
Most of the research works are based on commercially available graphite and/or graphene based supercapacitors while few studies are based on EG based supercapacitors. Out of previous studies, many are reported with liquid electrolytes [9][10][11]. This study mainly focuses on the surfactant mediated liquid phase exfoliation of Sri Lankan natural graphite and to use the resulting EG as an electrode material in an EDLC with a GPE while giving an estimation on the capacitance of the EG based EDLCs through a viewpoint of exfoliation time.

Preparation of Exfoliated graphite (EG) electrodes
Graphite was exfoliated in aqueous SDBS solution (5 mg ml -1 ) using ultrasonic homogenizer (Athena ATP 150) for different time periods (45 min, 2, 5, 10, 15 hrs) and filtered separately. The filtered EG samples were then dried and mixed with 1-methyl-2-pyrrolidinone before deposited onto fluorine-doped tin oxide (FTO) glasses using the doctor blade method as the electrodes. After coating, electrodes were dried to evaporate the solvent and also to improve the contacts. Area of the electrode was 1 cm 2 .

Preparation of gel polymer electrolyte (GPE)
Required amounts of PVDF, EC, PC and ZnTF were used to prepare a GPE as reported in our previous study [12].
The mixture was stirred well and heated at 130  C. The resultant hot mixture was pressed in between two wellcleaned glass plates. Upon cooling, it was possible to obtain a bubble-free thin film.

Fabrication of EDLCs
The EDLC was fabricated with a pair of EG electrodes and with the GPE of the configuration, EG/GPE/EG. Several EDLCs were fabricated using electrodes prepared with different exfoliation time.

Electrochemical Impedance Spectroscopy (EIS)
EDLCs fabricated with different electrodes were initially tested with Electrochemical Impedance Spectroscopy (EIS) technique using an impedance analyzer (Metrohm Autolab M101) at room temperature. Impedance data were gathered in the frequency range 400 kHz -0.01 Hz. Nyquist plots were drawn from the impedance data. The real part of the complex capacitance (C / (ω)) and the imaginary part of the complex capacitance (C // (ω)) were calculated based on the real (Z′(ω)) and imaginary (Z″(ω)) parts of the impedance and Bode plots were drawn [13].

Cyclic Voltammetry (CV)
For the cyclic voltammetry (CV), a three electrode electrochemical cell was used where one electrode of the EDLC was used as the working electrode and the other one as the reference and counter electrodes. Cycling was done using a computer controlled potentiostat (Metrohm Autolab M101). In one of our previous studies, we have found that the potential window within which the EDLC could operate without any distortion was as -1.0 to 1.0 V [14]. Same potential window was selected for the present study. The scan rate selected was 10 mVs -1 . The effect of exfoliation time on single electrode specific capacitance (Csc) was determined by cycling EDLCs having EG electrodes with different exfoliation times. Csc was calculated using the equation 1.

V mS
In the equation 1, ʃ IdV is the integrated area of the cyclic voltammogram, m is the mass of a single electrode, ΔV is width of the potential window and S is the scan rate [15]. The optimum exfoliation time was determined considering the Csc values. To observe the durability of the device, continuous cycling was then carried out for all EDLCs for 500 cycles.

Galvanostatic charge discharge (GCD) test
Galvanostatic charge discharge (GCD) test was carried out within the potential window, 0.1 to 1.0 V at a constant current, I of 0.05 mA using a Metrohm M101 potentiostat. Single electrode specific discharge capacitance, Csd was calculated using the equation 2.
dt dV m where, m is the mass of an electrode, dV / dt is the rate of potential drop excluding internal resistance [16].

Preparation of exfoliated graphite (EG) electrodes
In the exfoliation process, when graphite powder is mixed in aqueous SDBS solution, a graphitic collidal suspension is formed. An uniform suspension is obtained mainly due to surfactant nature of SDBS. Under the ultra-sonication, the solvated ions (Na + and DBS -) will be co-intercalated into graphite layers. This intercalation permits to induce expansion of the graphite interlayers while weakening the van der waals forces between graphitic layers. When there are sufficient amount of intercalated ions, the inflated graphite will generate ultimate expansion enabling separation of the graphitic layers ( fig.1) [17]. Ultimately, it is possible to obtain an EG suspension consists of single (graphene) and multilayer sheets [18]. When the exfoliation time increases, the number of collisions between graphitic particles increases while giving more time for layer seperation. Once the exfoliation time rises, the exfoliated suspension may contain graphene and few layer EG. Nyquist plots of all EDLCs mainly consist of a semi-circle at mid frequency area followed by a tilted vertical line in the low-frequency region as depicted in fig. 2. Insert of fig. 2 illustrates the high frequency region. Ideal capacitance is represented by a vertical line parallel to imaginary axis (-Z // ). However, the tilted line suggests weaker capacitive features of the EDLC. This might arise due to problems of electrode surfaces as well as non uniform nature of electrodes. Semicircle in the mid frequency range occurs due to the charge transfer resistance at the electrode-electrolyte interface. Smaller semi-circle radius/diameter implies low charge transfer resistance [19]. Absence of a semi-circle that represent bulk electrolyte resistance at high frequency region might be due to the unavailability of required frequency values. When the exfoliation time increases the number of EG layers decreases creating more voids thus increasing the surface area. But at the same time, rather than separating into fewer layers, breaking into smaller pieces without decreasing the number of layers could happen. This decreases the accessibility of ions on the EG electrode surface. Therefore two main outcomes can be highlighted as the effect of varying exfoliation time: (i) separation of graphitic layers into fewer EG layers creating more spaces (pores) (ii) breaking of stacked graphitic sheets into small pieces (weaken the approachability for ions to inner layers) [20]. Combination of these two aspects influences the capacitive features of the EDLCs.   The highest Csc can be found using maximum C / in fig. 3a. Values are given in table 1. into inner layers resulting a decrease in the capacity [21].
Another important parameter is the relaxation time (τ0) which is the quantitative measurement of the response time of the EDLC. τ0 can be calculated by analyzing the Bode plots ( fig. 3b) using the equation 3 [22].
They are shown in Table 1 with the corresponding frequency values. Accordingly, EDLCs with the EG electrodes having an exfoliated time of 10 h shows the minimum τ0 further validating its higher capacitive properties.

Cyclic Voltammetry (CV)
The cyclic voltammograms obtained are shown in the fig. 4. They all exhibited a parallelogram shape without any redox peaks implying their good capacitive behavior. This also suggests that there were no redox reactions and the charge storage mechanism was mostly based on the electrostatic mechanism. Thus, capacitive properties are solely due to ion accumulation instead of ion intercalation process. The highly reversible charge discharge process is confirmed by the almost mirror image symmetry around zero current line [23]. All the CV curves show a rapid current response on voltage reversal at each end potential. The single electrode specific capacitance (Csc) was determined using the equation 1. As per fig. 4, area of the cyclic voltammograms with the exfoliation time till 10 h increases and decreases thereafter. This well agrees with the EIS results explained before.   This ohmic loss arises due to internal resistance of the device. Even under such variations, all EDLCs have been able to maintain a more or less constant Csd. This might be due to the property of the GPE which is called as self-healing property [24].

Conclusion
EG samples with various exfoliation times were synthesized using a surfactant mediated liquid phase exfoliation.
EDLCs were fabricated with the prepared EG samples. The electrochemical performances of these EDLCs were evaluated via EIS, CV and GCD techniques. EG sample with exfoliation time of 10 h showed maximum Csc of 4.12 F g -1 and maximum Csd of 1.10 F g -1 by CV and GCD techniques respectively.