Highly Porous Reduced Graphene Oxide-Coated Carbonized Cotton Fibers as Supercapacitor Electrodes.

High-surface-area carbon-based capacitors exhibit significant advantages relative to conventional graphite-based systems, such as high power density, low weight, and mechanical flexibility. In this work, novel porous carbon-based electrodes were obtained from commercial cotton fibers (CFs) impregnated with graphene oxide (GO) at different dipping times. A subsequent thermal treatment under inert atmosphere conditions enables the synthesis of electrodes based on reduced GO (RGO) supported on carbon fibers. Those synthetized with 15 min and 30 min of dipping time displayed high specific capacitance given their optimal micro-/ mesoporosity ratio. Particularly, the RGO/CCF15A supercapacitor reports a remarkable specific capacitance of 74.1 F g-1 at 0.2 A g-1 and a high cycling stability with a 97.7% capacitive retention, making this electrode a promising candidate for supercapacitor design. Finally, we conducted a density functional theory study to obtain deeper information about the driving forces leading to the GO/CF structures.

Introduction

Nowadays, the search for new eco-friendly energy systems is a priority to mitigate the global impact associated with fossil fuel energy consumption.[1−4] At the same time, novel electronic devices entail the need to develop more efficient energy storage systems with higher capacity and longer average lifetime.[5] In this regard, electrical double-layer capacitors (EDLCs),[5] pseudocapacitors,[6] and flexible solid-state supercapacitors (FSSCs)[7,8] are able to cover the above-mentioned demands. Particularly, carbon-based capacitors exhibit significant advantages, such as high power density, low weight, and flexible packaging, in contrast to conventional graphite-based systems.[9] To achieve a better capacitance performance, electrodes based on porous carbon materials can be employed, given their appealing morphological features, like high specific surface area (SBET), defined porosity, and hierarchical arrangement.[10] Recently, the use of biomass as raw material has allowed the design of carbon-based storage systems with outstanding electrochemical and mechanical properties.[11,12] For instance, hierarchical porous carbon electrodes from melamine-treated cotton fibers (CFs) exhibiting a specific surface area (SBET) of 777 m2 g–1 and a specific capacitance (Cs) of 360 F g–1,[13,14] as well as KOH-treated natural fiber electrodes with 1435 m2 g–1 and 218 F g–1 at 0.1 A g–1, have been reported.[15]

The good performance of modified biomass-based supercapacitor electrodes made of graphene derivatives has been demonstrated by Wang and co-workers (2013), by means of a reduced graphene oxide (RGO) electrode supported on carbon cloth (CC) displaying a Cs of 79.2 F g–1, significantly higher in comparison to a pure CC electrode with a Cs of 1.35 F g–1.[7,16] Also, a RGO-based micro SC with an ultrahigh rate up to 1000 V s–1 and 78.9 mF cm–2 of capacitance has been recently obtained by Yang and co-workers (2017).[10]

In this work, we develop a novel carbon-based energy storage system from RGO supported on carbon fibers, and we study the influence of thermal treatment and dipping time on the structural and electrochemical properties. For the obtained microporous and meso-/microporous double-layer capacitors, an intrinsic relationship between superficial and electrochemical properties has been found. The optimal micro-/mesoporosity ratio, as well as high specific capacitance found for some electrodes, makes them promising materials for supercapacitor design.

Results and Dicussion

Physicochemical Characterization

Commercial CFs were immersed in a home-made GO suspension, leading to darker-brown CFs (GO/CF, x: dipping times). GO/CF fibers display higher mechanical resistance than CF fibers and similar morphological features to those reported by Zhang (2016),[27] but without the use of the high-cost cryogenic treatment. According to the EDS test, CF and GO/CF show 49.1 and 63.9 C wt %, respectively, which indicates a feasible interaction of GO sheets on CF fibers. Moreover, thermally treated CF (CCFA) and GO/CF (RGO/CCF) fibers register 95.6 and 90.7 C wt %, respectively, mainly due to the loss of H2O, CO, and CO2. According to Tang and Bacon (1964),[28] the wt % increment of thermal treated fibers is due to the physical desorption of water (25–150 °C), dehydration (150–240 °C), thermal cleavage of glycosidic C–O (240–400 °C), and aromatization (above 400 °C) processes, until a graphite-like layered crystal structure is achieved. Thus, thermal treated samples display a fiber-like morphology (Figure a,b). In addition, RGO/CCF45A has a lower C wt % than CCF, in line with the major amount of oxygen released by the RGO sheet insertion. On the other hand, CCFA fibers turn from having a soft to a rough surface owing to the inserted RGO sheets (Figure c). The RGO/CCF samples display a porous surface, in contrast to that reported by S. Wang et al. (2013).[7] (Figure d). This porous structure can be attributed to gas evolution from the matrix during the thermal treatment.[29,30]

Figure 1

FE-SEM images of (a,b) CCFA and (c,d) RGO/CCF45A. (Inset) Relative carbon weight percent (C wt %) according to EDX analysis.

To describe GO/CF and RGO/CCF interactions, and their influence on the electrochemical properties, infrared and Raman spectroscopy analyses were conducted. The IR results (Table S1) identify the main chemical groups of CFs, that is, carbonyl (C=O, 1033.8 cm–1), methylene (CH2, 1322.2 cm–1), and cellulosic hydroxyl (O–H, 3325.3 cm–1). The intensity of these signals decreases with increasing impregnation time (Figure S1a).[31] The main CF infrared signal can be attributed to the GO sheet stacking process on the CF surface. After the thermal treatment, this single signal is replaced by two peaks at 979.8 and 1512.2 cm–1 associated with C–Cring and C=C bonds, respectively, which are characteristic of graphitic systems (Figure S1). According to Tang (1964),[28] the cellulose fiber structure experiences a glycosidic rupture at 250 °C as a consequence of the carbonyl group elimination with the C–C bond rupture, leading to a pronounced weight loss owing to H2O, CO, and CO2 elimination. From 400 to 700 °C, the cyclic hydrocarbons start an aromatization stage, losing hydrogen gas and generating C=C bonds until a polymeric structure is obtained, which turns into graphite-like layers at higher temperatures.[28] This affirmation could explain a similar infrared spectrum between carbonized samples (CCFA and RGO/CCF) and graphite powder (Figure S1c). Raman spectroscopy was used to quantify the interaction degree in the GO/CF and RGO/CCF systems, after the impregnation and carbonization stages, respectively (Figure S1b,d). From the I(D)/I(G) ratio, it is possible to determine the crystallite size and defect density on GO/CF and RGO/CCF composites. According to Robertson and O’Reilly (1987), the I(D)/I(G) ratio can be related to the C(sp3)/C(sp2) ratio, where C(sp3) is associated with network defects and C(sp2) with the crystalline region of a carbon system.[32,33] In turn, the crystallite carbon size (La) can be estimated through the following equation

where λl is the wavelength of the laser beam. Similarly, the defect distance (LD) and defect density (nD) can be estimated as follows[34,35]

Aqueous-suspended GO sheets show a decrease in the defect density from 45.5 × 1010 to 21.6 × 1010 defects cm–2 when they are supported on CF, and the impregnation time is increased (Table S2). Also, an increment in the GO crystallite size from 9.33 nm (GO) to 19.7 nm (GO/CF45A) is registered. Additionally, a linear tendency between the impregnation time and crystallite size is observed, which could be related to the number of GO sheets on the CF surface achieving a stacked arrangement (Figure S2). Similarly, this effect is evidenced as a decrease of the I(D)/I(G) value from 2.06 to 0.98 for aqueous-suspended GO sheets and GO/CF45A, respectively (Figure S1b). After the thermal treatment, both POS(D) and POS(G) confirm the existence of RGO sheets on the CCF surface (Figure S1d).[34] In addition, the decrease of the nD value implies an enhancement of crystallinity, in contrast to the GO/CF samples (Figure S2). Despite their high crystallinity, RGO-coated samples show a higher D band intensity (I(D)), in contrast to the multilayer graphene powder (ML graphene), because of the higher oxidation degree of supported RGO sheets.[34,36]

To explain the influence of RGO incorporation on the morphological characteristics of the electrodes, a N2 adsorption–desorption test was carried out. A hysteresis transition from H1 (CCFA) to H4 (RGO/CCF15A) is observed (Figure a). Specifically, RGO/CCF15A shows a hysteresis loop related to laminar pores, suggesting that the RGO cover structure does not collapse under thermal treatment (Figure b).[37] Furthermore, an increase in SBET from 1074 m2 g–1 in CCFA to 1458 m2 g–1 in RGO/CCF15A was registered, in line with an increment of the mesoporous surface area (Smeso) from 299 to 730 m2 g–1 (Table ).

Figure 2

N2 adsorption–desorption test. (a) N2 adsorption–desorption isotherms and (b) pore size distribution for CCFA and RGO/CCF.

Table 1

Specific Surface Area (SBET), Microporous Surface Area (Smicro), Mesoporous Surface Area (Smeso), Smicro/Smeso Ratio, BJH Micropore Diameter, and Micropore Volumea

samples/cm3 g–1	SBET/m2 g–1	Smicro/m2 g–1	Smeso/m2 g–1	Smicro/Smeso	BJH pore diameter/nm	micropore volume
CCFA	1074	775	299	2.6	2.1	0.40
RGO/CCF15A	1458	728	730	1.0	2.0	0.38
RGO/CCF30A	1207	784	423	1.9	2.0	0.41

The BJH micropore diameter and micropore volume were calculated from a N2 adsorption–desorption test.

Computational Study

We conducted a computational study in the framework of density functional theory (DFT) at the M06-2X/6-311G(d) level of theory to estimate the effect of the GO epoxy (−O−) and hydroxyl (−OH) groups on the anchoring of the cellulosic fibers. The choice of the M06-2X functional is based upon the inclusion of dispersion corrections, in line with recent works.[38,39] The results are presented in Figure , where a d-glucose molecule—representing a cellulose segment integrating a CF—interacts with graphene (I) and GO (II) molecular models. In both cases, we employed a 15 Å × 9 Å sheet. Multiple graphene oxide models have been proposed, given that the exact chemical structure has long been the subject of debate and no definitive model exists.[40] The solvent effects—water in our case—are included through the polarizable continuum model (PCM). In this approach, the solute is embedded in a cavity surrounded by a continuous medium modeling the solvent effects. For the d-glucose···graphene system, an interaction energy of −15.5 kcal mol–1 has been obtained, mainly associated with London forces between the fragments, whereas in the d-glucose···GO assembly, a major stabilization is revealed by a comparatively higher interaction energy of −23.3 kcal, which can be attributed mainly to the hydrogen bonds established between the d-glucose hydroxymethyl/hydroxyl groups and the hydroxyls belonging to GO.

Figure 3

Optimized structures of (a) d-glucose interacting with graphene and (b) d-glucose interacting with GO. Hydrogen bond distances and some selected interatomic distances are highlighted.

Electrochemical Characterization

According to the cyclic voltammetry (CV) test, the pure carbonized cotton-based electrode (CCFA) exhibits a double-layer capacitive profile from −0.6 to 0.6 V, with a slight internal resistance. At higher positive potentials, the carbon-based electrode is able to promote H2O-oxidation to H2O2; meanwhile, at more negative potentials, the electrode surface likely starts the proton adsorption, leading to H2 production from water reduction.[41,42] In contrast, the RGO-covered electrode (RGO/CCF15A) registers a redox couple from −0.2 to 0.2 V, which may be associated with Faradaic processes of oxygenated groups owing to the partially reduced network of RGO-based covering (Figure a).[42] Regarding the capacitive profile, RGO-covered electrodes (RGO/CCF5–45A) register a higher capacitive current than CCFA in the double-layer capacitive current domain from 0.2 to 0.4 V of potential range, taking into account that they have an open circuit potential (OCP) of ∼0.3 V (Figure b). Specific capacitances (CCV) were calculated from the voltammetry charge, according to eq . In this regard, CCFA reports 69.6 F g–1, whereas the RGO-covered carbonized-cotton electrodes register an enhanced specific capacitance, with the highest value of 197.8 F g–1 for RGO/CCF15A (Table ). The increase of CCV is attributed to the presence of RGO microsheets on the pure carbon-based fiber surface during the thermal treatment (Figure c). The increment of the capacitive current is associated with an enhancement of the specific surface area and, specifically, of the mesoporosity, as discussed above. Therefore, both the increased specific surface area (SBET) and the insertion of oxygenated groups of RGO covering add a pseudocapacitive current in the charge storage process.[5]

Figure 4

CV test of (a) CCFA vs RGO/CCF15A from −0.6 to 0.6 V at 10 mV s–1 and (b) CCFA vs RGO/CCF5–45A from 0.2 to 0.4 V at 10 mV s–1 (OCP ∼ 0.3 V) in 1.0 mol L–1 of H2SO4 electrolyte. GCD of CCFA vs. RGO/CCF5 - 45A at (c) 0.4, (d) 0.8, (e) 1.6, and (f) 3.2 A g–1 in a three-electrode cell.

Table 2

Specific Capacitance from CV (CCV), the GCD Test (CGCD), and NLCF of EIS Spectrums (CEIS)a

technique	CV	GCD		EIS			BET
samples	CCV/F g–1	CGCD/F g–1	IRdrop/V	QEISα≈1.0/F	CEIS/F g–1	SEIS/m2 g–1	SBET/m2 g–1	RESA
CCFA	70	129	0.48	0.23	64	212	1074	0.20
RGO/CCF5A	117	185	0.33	0.44	108	359
RGO/CCF15A	198	219	0.15	0.63	172	573	1458	0.39
RGO/CCF30A	176	202	0.11	0.72	149	497	1207	0.41
RGO/CCF45A	115	165	0.35	0.31	111	369

QEISα≈1.0, internal constant phase element with a nonideal constant (α) that tends to unity, SEIS, electrochemical accessible surface, SBET, specific surface area, and RESA, electrochemically accessible surface ratio.

The galvanostatic charge/discharge (GCD) test for a three-electrode system provides us relevant information about the effect of RGO covering on the capacitive and resistivity profiles of the synthetized carbon-based electrodes. According to the N2 adsorption/desorption test, CCFA can be described as a microporous electrode because of its high micro-/mesoporosity ratio of 2.6. This characteristic can reduce the ion diffusion in/out the electrode/electrolyte interface, owing to its high adsorption free energy.[5,24,37] Thus, the discharge process exhibits a high ohmic drop, likely related to pseudocapacitive or intercalation processes and a high discharging slope (Figure c–f).[5,24] In contrast, RGO/CCF5A, RGO/CCF15A, and RGO/CCF30A electrodes show a significant enhancement of both capacitive and resistive behavior of pure carbonized cotton-based fibers (Figure c–f). Specifically, the ohmic drop is reduced from 0.48 to 0.15 (RGO/CCF15A) or 0.11 V (RGO/CCF30A) (Table ). Particularly, RGO/CCF15A and RGO/CCF30A display an appreciable reduction of the ohmic drop, a lower discharge slope, and a significant stability at high current densities. Both electrodes show a typical triangular profile related to electrochemical double layer (EDL) mechanisms at low current densities. However, a pseudocapacitive component according to their CV profiles and their discharge slopes at high current densities can be related (Figure c–f).[5,7] According to the GCD test, CCFA shows a specific capacitance (CGCD) of 63.5 F g–1, while RGO/CCF15A and RGO/CCF30A display an outstanding CGCD of 219 and 201.7 F g–1, respectively (Table ).

For a better understanding of the electrochemical behavior of RGO/CCF electrodes, the electrochemical impedance spectroscopy (EIS) test and nonlinear complex fitting (NLCF) of their equivalent circuits are discussed (Figure ). A brief description of the Nyquist diagrams shows a mean electrolyte resistance (RS) of 2 Ω, approximately. Besides, the asymmetric growth between real and imaginary impedance from high to medium frequencies can be attributed to the presence of nonideal circuit elements, for example constant phase element (Q), Warburg resistance (W), and finite diffusion element, among others.[43,44]

Figure 5

(a) Nyquist diagrams of CCFA and RGO/CCF. Experimental (empty dots) and theoretical spectra (filled dots). (b) Equivalent circuits of CCFA, RGO/CCF15A, and RGO/CCF30A. The inserted values are calculated from NLCF of the EIS measurements.

RGO/CCF reports low electrical resistance, and this fact confirms the ohmic drop tendency shown in the GCD test. CCFA and RGO/CCF show a similar angle phase at low frequencies, around 60°, with a slight variation in the Rct values (Figure a). After resistive/capacitive loop and at medium frequencies, it shows a linear increment of imaginary impedance (-Z″) with a low-angle slope into short range of frequencies, for example, from 62.5 to 0.69 Hz (CCFA) or from 8.69 to 0.17 Hz (RGO/CCF15A), which could be related to a non-ideal circuit element of finite diffusion (M), whose behavior depends on the porous characteristic on the electrode surface, such as pore shape and size.[45−47] Both theoretical spectrum and equivalent circuits from NLCF modeling of the electrode/electrolyte interface are reported. Theoretical data are presented with experimental data, and a minimum χ-square is considered.[23] At high frequencies, NLFC reports a mean H2SO4 electrolyte resistance (Rs) of 1.46 Ω.

Similarly, the internal resistance increases from 5.81 (CCFA) to 29.52 Ω (RGO/CCF45A), whereas their external capacitance decreases from 2.99 × 10–5 (CCFA) to 2.35 × 10–5 F (RGO/CCF45A), together with a reduction of its ideal behavior (α1) from 1.0 to 0.95. From 5 (RGO/CCF5A) to 45 (RGO/CCF45A) minutes of impregnation time, a reduction of 50% of the internal resistance is registered, in contrast to N2-made samples (Figure S3). Conversely, the internal capacitance reports a progressive increment of its ideal behavior; for example, CCFA shows an internal capacitance (C2) of 0.23 F, whereas RGO/CCF15A presents a constant phase element (Q3) of 0.63 Ω–1 sα (α3, 0.99). As the RGO/CCF15A nonideal constant (α) is higher than 0.9, its internal capacitance can become close to 0.63 F. Likewise, RGO/CCF30A reports an internal capacitance of about 0.72 F (Table ).[45,46] In addition, all the samples report a Warburg impedance (W) in series with a constant phase element (Q). This internal circuit (W–Q) can be represented with a finite diffusion impedance (M), and it is associated with controlled diffusion of the ionic transport.[45] Remarkably, RGO/CCF samples report a laminar mesoporosity, suggesting that the ionic diffusion is controlled by the pore characteristics.[45,46,48] Specifically, CCFA reports a Warburg impedance (W) of 0.16 Ω–1 s0.5, while RGO/CCF15A and RGO/CCF30A show values of 0.25 and 0.32 Ω–1 s0.5, respectively. According to Klink (2013), a finite diffusion element (M) achieves a pure capacitive behavior at low frequencies, whereby the total capacitance can be estimated from this value (Table ).[45] Accordingly, the total capacity (CEIS) from the internal capacitive element (QEISα) with a nonideal constant (α) that tends to unity is reported. For the estimation of the electrochemically accessible surface (SEIS), a double-layer charge density (Qdlo) of the glassy carbon electrode of 3.0 × 10–5 F cm–2 is considered.[49,50]

Additionally, an electrochemically accessible surface ratio (RESA) registers the fraction of the interfacial surface area available (SEIS) relative to the total physical surface area (SBET).[50]

The difference between the CEIS and CGCD values is due to NLCF of EIS measurements and reports the total charge related to the EDL, whereas the GCD test also registers the Faradaic charge by the surface carbon and oxygen groups.[50]

In this sense, RGO/CCF electrodes register remarkable capacity, both RESA and SBET. RGO/CCF15A shows a CEIS of 171.90 F g–1, with 40% of electrochemical surface area available. These characteristics can be related to their laminar mesoporous surface, as mentioned above, and a controlled—not limited—ionic diffusion. Therefore, RGO/CCF electrodes show promising application as electrodes in the design of supercapacitors.

Capacitive performance of CCFA and RGO/CCF15A was tested in a home-made two-electrode cell by the GCD test. Briefly, a coin-like cell was built in a Teflon-based cylindrical cell to reduce ohmic resistance and to avoid electrolyte leakage (Figure a). In addition, stainless steel current collectors and a pressure regulator were used to guarantee good electrical contact (Figure b).

Figure 6

(a) Scheme of the home-made two-electrode cell and its components. (b) Image of the GCD test of carbon-based SSCs.

According to the GCD test, the CCFA symmetric supercapacitor (SSC) shows low potential values to full charge. Constant current discharge of CCFA registers a remarkable ohmic drop, presumably related to its highly microporous CCFA electrodes (Smicro, 775 m2 g–1) (Figure a). Thus, CCFA registers a high discharge rate and a high specific capacitance (Csp) of 31.2 F g–1 at 0.2 A g–1 (Table ). In contrast, RGO/CCF15A shows a low ohmic drop and an EDLC profile at each applied current (Figure b). At a constant current discharge, RGO/CCF15A reports a lower discharge rate than CCFA. Specifically, a discharge slope ratio of 3:1 between CCFA and RGO/CCF15A is reported. Thus, RGO/CCF15A shows a remarkable Csp of 74.1 F g–1 at 0.2 A g–1 (Table ). Unlike the RGO/CCF15A electrodes, which report a similar microporosity (Smicro, 728 m2 g–1) to the CCFA electrode, the SSC built from these electrodes shows an enhanced charge retention. This last fact can be related to the increase of the mesoporous surface area (Smeso) from 299 to 730 m2 g–1, increasing the EDL surface area.

Figure 7

GCD test of (a) CCFA and (b) RGO/CCF15A at 50, 60, 100, 150, and 200 mA g–1 of current density. (c) GCD test of RGO/CCF15A at 60 mA g–1 for 900 s. (d) Cycling stability test of RGO/CCF15A at 60 mA g–1 for 1000 cycles, (inset) RGO/CCF15A two-electrode system on a stainless steel current collector.

Table 3

Specific Capacitances (Csp) of Carbon-Based Electrodes Determined by the GCD Test (CGCD) in a Two-Electrode Cell

electrode	CCFA		RGO/CCF15A
applied current density	discharge slope	specific capacitance	discharge slope	specific capacitance
id/mA g–1	dV/dt/mV s–1	Csp/F g–1	dV/dt/mV s–1	Csp/F g–1
50	9.3	23.9	2.6	66.0
60	10.0	27.5	3.2	69.9
100	15.1	29.3	4.8	71.4
150	22.3	29.7	7.1	71.6
200	31.9	31.2	10.0	74.1

RGO/CCF15A shows an enhanced capacitive behavior, according to the above discussion and time stability tests, which were done by the GCD method (Figure c). RGO/CCF15A exhibits a plateau after 900 s of charged time but a linear profile at constant discharge current owing to EDLC and pseudocapacitive behavior.[5,24,51] As well, this device reports 67.8 F g–1 at 60 mA g–1 of applied current density. Thus, it means that the total electrochemical surface area is polarized, and the internal series resistance is exposed.[44] For long cycling conditions, RGO/CCF15A reports a capacity retention of 97.7% at 60 mA g–1 for 1000 cycles, a value that represents an electrochemical stability similar to that of carbon-based systems reported in the literature (Figure d).[11,16,25]

A RGO/CCF15A FSSC was built to test the bending angle effect on its capacitive profile. As such, the mechanical flexibility test of RGO/CCF15A FSSC was conducted at 0, 45, 90, and 180° of angle of bend by CV from −0.5 to 0.5 V at 100 and 200 mV s–1 (Figure a,b). According to the obtained results, the capacitive profile of the RGO/CCF15A FSSC does not drastically change with the angle of bend at 100 and 200 mV s–1 of scan rate. As well, the device shows a stable capacitive behavior within a significant potential window and a remarkable mechanical stability (Figure c).

Figure 8

Mechanical flexibility test of RGO/CCF15A FSSC at 0, 45, 90, and 180° of angle of bend by CV from −0.5 to 0.5 V at (a) 100 and (b) 200 mV s–1. (c) Image of RGO/CCF15A FSSC bent to 45°.

Concluding Remarks

Herein, an easy method to obtain carbon-based porous electrodes from commercial CFs is reported. The results indicate that RGO sheets are successfully supported on cotton-based carbon fibers. We demonstrated that in both the pure carbon-based fiber surface and RGO ones, the morphological and electrochemical properties are strongly dependent on the RGO load. Moreover, the RGO/CCF electrodes exhibit an increase of their laminar mesoporous surface area with the impregnation time, which implies an increase of their electrochemical accessible surface and capacitance. In addition, DFT studies on a model system reveal that hydroxyl groups belonging to the oxide graphene sheet act as anchoring points to the carbon fiber. In summary, this work reports the optimal micro-/mesoporosity ratio and high specific capacitance of novel porous carbon-based electrodes obtained from CF and impregnated with GO for 15 and 30 min of impregnation time, making them promising candidates for supercapacitor design. In this regard, RGO/CCF15A and RGO/CCF30A electrodes report a high specific surface area of 1458 and 1207 m2 g–1 with a remarkable specific capacitance of 219 and 202 F g–1, respectively, in a three-electrode system. Meanwhile, the RGO/CCF15A-based supercapacitor shows a noticeable specific capacity of 74.1 F g–1 at 0.2 A g–1 with a capacitive retention of 97.7% and a nondependent capacitive behavior with the bending angle.

Experimental Procedure

Reagents

Pure graphite sheet powder, sodium nitrate (NaNO3, 99%), sulfuric acid (H2SO4, 96.6%), potassium permanganate (KMnO4, 97%), hydrogen peroxide (H2O2, 30%), and hydrochloric acid (HCl, 5%) for the GO synthesis were used. In addition, home-made GO, commercial CFs, and high purity argon gas (Ar, 99.99%) were employed in the synthesis of RGO-modified carbon fiber electrodes. Commercial carbon felt, pure GO (GO*, 10% oxidized sheets), and multilayered graphene powder (ML graphene) were used as standards for spectroscopic and electrochemical analyses.

RGO-Coated Cotton Carbon Fibers

A 3.0 g L–1 of GO suspension was prepared from 1.06 g graphite powder according to the Hummers–Offeman method (1957).[17] Pure graphite sheet powder was oxidized by a 0.51 g NaNO3 and 40.0 mL of H2SO4 mixture under constant stirring in an ice bath for 30 min. Afterward, 3.0 g KMnO4 was slowly added under vigorous stirring for 2 h. The homogeneous suspension reacts with 100 mL of H2O at 95 °C for 30 min and 10 mL of H2O2 at 40 °C for 12 h, under stirring. The obtained solid was washed with 120 mL of 5% HCl aqueous solution, and—after a period of 48 h—it was washed three times with 400 mL of deionized water and left to rest for 24 h. The neutralized powder was dried for 72 h at room temperature and stored. Finally, 0.68 g home-made GO was dispersed in 100 mL of H2O by an ultrasonic treatment for 3 h to obtain a stable GO suspension.

Commercial CFs were dipped into the home-made GO suspension (3.0 g L–1) for 5, 15, 30, and 45 min and dried at room temperature for 72 h. The dried samples were named as GO/CF (x = 5, 15, 30, 45) where “x” is the dipping time, GO/CF0 being the nondipped sample. GO/CF samples were thermally treated at 800 °C (3 °C min−1) under an Ar atmosphere for 30 min in a tubular furnace (Nabertherm R120/500/13). Synthetized electrodes were labeled as CCFA for uncoated carbonized CFs and RGO/CCF for RGO-coated carbonized carbon fibers, respectively.

Physicochemical Characterization

Morphological properties of noncarbonized (CF and GO/CF) and carbonized fibers (CCFA and RGO/CCF) were characterized by an SU8230 field emission scanning electron microscope, Hitachi, and their relative compositions were registered by a Quantax flat-QUAD energy dispersive X-ray spectrometer, Bruker. The specific surface area (SBET), microporous surface area (Smicro), and BJH micropore diameter of CCFA and RGO/CCF were calculated from the N2 adsorption–desorption test at 77 K on a Micromeritics Gemini VII 2390 surface analyzer. The chemical interaction between GO and CF, or RGO and CCF, was analyzed by infrared spectroscopy from 700 to 4000 cm–1 (Prestige, Shimadzu) and Raman spectroscopy from 1100 to 2800 cm–1 at 532 nm (Scientific XploRA, Horiba).

Computational Details

All DFT calculations were carried out by using the Gaussian 09 code,[18] tightening self-consistent field convergence thresholds (10–10 a.u.). A 6-311G(d) basis set and the hybrid functional M06-2X were employed.[19] Remarkably, this functional includes dispersion effects, thus representing a suitable approach to calculate the interaction energies between the CFs and graphene/GO. The solvent effects of water were included according to the PCM, where the solvent is approximated as a structureless continuum whose interaction with the solute is mediated by its permittivity, ε.[20−22]

The carbon fibers electrodes were tested by CV, a GCD test, and EIS. Each test was carried out in a three-electrode cell, where a high-surface-area piece of CCFA was used as the counter electrode, a reversible hydrogen electrode (RHE) as the reference electrode, a piece of CCFA or RGO/CCF as the working electrode, and 1.0 mol L–1 of H2SO4 as the supporting electrolyte. To eliminate dissolved oxygen, N2 gas was bubbled for 5 min. CV measurements were performed in a potential range of 100 mV at 1.0 mV s–1, whereas the GCD test was recorded at 1.5, 3.0, 6.0, and 12.0 mA of applied current. Moreover, EIS analysis was carried out at an OCP (VOCP) range from 104 to 10–2 Hz. Each electrochemical test was performed in an AUTOLAB Modular PGSTAT302N Potentiostat/Galvanostat, Metrohm. NLCF of EIS measurements and the equivalent circuit modeling were analyzed by the ZMAN 2.3.2 software, Zivelab.[23]

The specific capacitance of each working electrode in a three-electrode system was determined by a CV test (CCV) and a GCD test (CGCD), both at several applied currents. In this sense, CCV (F g–1) was calculated according to the total voltametric charge.[7,11]

where is the total voltametric charge; (E2 – E1) is the potential range; υ is the potential scan rate; and mwe is the mass of the working electrode. On the other hand, the CGCD values were determined from the discharge slope, according to the equation[11,24]

where Id is the discharge current; Δt is the discharging time; and ΔV is the potential drop on the discharge process.

To test the capacitive performance of carbonized CFs as supercapacitor electrodes, a symmetrical supercapacitor (SSC) was built in a coin cell where two cylindrical CCFA (or RGO/CCF15A) electrodes were used as the working electrode and counter electrode, separated with a glass-based membrane, in 0.5 mol L–1 of H2SO4 as the supporting electrolyte. The GCD test of CCFA and RGO/CCF15A was conducted at 50, 60, 100, 150, and 200 mA g–1 of current density.[13] The cycle stability of RGO/CCF15A was tested at 60 mA g–1 for 1000 cycles, and its capacitive retention was recorded (%).where mT = mWE + mCE represents the total mass of the electrodes, and dV/dt, the slope of the discharge curve after the ohmic drop.

In order to test the capacitive profile versus the bending angle, a solid-state SSC was built from two covered RGO/CCF15A electrodes with a PVA–KOH gel electrolyte (1.0 g PVA, 0.56 g KOH, and 10 mL of ultrapure water) placed on a polypropylene (PP) film. The RGO/CCF15A electrode couple was placed on top of each other and sealed by mechanical pressure.[14,25,26] The capacitive profile versus the bending angle was tested at 0, 45, 90, and 180° from −0.5 to 0.5 V of potential range at 100 and 200 mV s–1.