Using Different Ions to Tune Graphene Stack Structures from Sheet- to Onion-Like During Plasma Exfoliation, with Supercapacitor Applications

In this article, we report a facile and simple approach for tuning graphene nanosheet structures (GNS) with different ions in the electrolytes through cathodic plasma exfoliation process in electrochemical reactions. We obtained sheet- and onion-like GNS when aqueous electrolyte NaOH and H2SO4, respectively, were present during plasma exfoliation in the electrochemical reactions, as evidenced from scanning electron microscopy and transmission electron microscopy images. Moreover, the onion-like GNS exhibited a specific surface area of 464 m2 g−1 and a supercapacitive performance of 67.1 F g−1, measured at a scan rate of 5 mV s−1 in 1 M NaCl; these values were much higher than those (72 m2 g−1 and 21.6 F g−1, respectively) of the sheet-like GNS. This new approach for efficiently generating tunable stacked graphene structures with different ions, in the cathodic plasma exfoliation process, has promising potentials for use in energy storage devices. Electronic supplementary material The online version of this article (10.1186/s11671-019-2963-5) contains supplementary material, which is available to authorized users.


Introduction
With the rapid depletion of fossil fuels and the rising controversy over their use, the development of energy storage and conversion devices-including batteries [1,2] and supercapacitors [3,4]-is becoming increasingly necessary for the global community. Supercapacitors have been studied extensively because of the excellent power densities and cycling lifes. They can be divided into two typespseudocapacitors and electrical double-layer capacitors (EDLCs)-based on their charge storage mechanisms. In pseudocapacitors, the charge storage depends on reversible Faraday reactions, for which the electrodes are often made of transition metal oxides and conducting polymers. For EDLCs, the electrical energy is stored through ion adsorption and desorption processes, using typical electrode materials (e.g., porous carbon materials). In recent years, two-dimensional (2D) materials, including graphene and transition metal dichalcogenides (TMDCs), have attracted considerable attention for their excellent physical and chemical properties. TMDCs of the type MX 2 , where M is a transition metal atom (e.g., Mo, W) and X is a chalcogen atom (S, Se, or Te), possess a sandwich-like structure and have adequate theoretical capacitance and catalytic activity [5][6][7]. Graphene, a 2D single-layered carbon material having honeycomb-like lattices, has also drawn much attention for its extraordinary electrical [8,9] and thermal conductivities [10], elasticity [11,12], transparency [13][14][15], and specific surface area [16].
Two forms of graphene are common: thin films and powders. Mechanical exfoliation [17] and epitaxial [18,19] and chemical vapor deposition [20][21][22][23] can be used to produce high-quality graphene thin films-displaying flexibility and high transparency-that can be applied in electrical devices, but high costs and careful control are required, due to the necessary production equipment and precursors. Powdery graphene is more widely used because of its greater mass-production, superior stability, large specific surface area, and high applicability to charge storage applications (e.g., capacitors [24], lithium-ion batteries [25,26]).
Many methods-including ball-milling, arc-discharging, and solution plasma and electrochemical exfoliation-have been reported for the production of powdery graphene materials. The ball-milling and arc-discharging methods involve physical exfoliation and can be used for industrialscale production [27,28]; nevertheless, expensive equipment and vacuum conditions are essential, resulting in higher costs. The solution plasma and electrochemical exfoliation methods can both generate graphene sheets from graphite electrodes using simple setups and readily accessible electrolytes [29,30]. Although solution plasma exfoliation requires high voltages (> 1000 V) between very close anodes and cathodes and, therefore, extra care, electrochemical exfoliation can be performed at relatively lower applied voltages (< 20 V), but higher oxidation levels are introduced into the products, due to anodic exfoliation during electrolysis. Furthermore, the ready re-aggregation of layer-like graphene, through strong π-stacking interactions, can weaken its performance.
Several methods have been developed to improve the performance of active materials, including composite methods for energy storage devices [31][32][33], single-atomconfined 2D materials and surface regulation for catalysis [34,35], heterostructures in photovoltaics [36], and structural engineering for further technological applications [37]. These methods all result in synergistic effects of the heterogeneous materials, with more accessible active sites provided by heterojunctions and the larger surface areas of the modified materials. For graphene materials, sacrificial-template methods and aerogel fabrication are the most common processes used to overcome the issues of layered re-aggregation. The sacrificial-template methods normally employ aqueous solutions of graphene oxide and polymer microspheres that are subjected to high-temperature reduction, often yielding graphene materials of high specific surface areas [38,39]; nevertheless, the blending uniformity must be controlled well in terms of the nature of the surface charge, and such methods are time-consuming. Graphene aerogels can be synthesized at low temperatures [40][41][42] or under ambient conditions [43] to give outstanding specific surface areas; freeze gelation and solvent sublimation are, however, time-consuming and require toxic solvents to preserve the porous architectures of the aerogels. Hence, the opportunity remains to develop an easy-to-setup, inexpensive, one-step, and highly efficient method for the production of powdery graphene of high specific surface area under ambient conditions without using hazard chemicals.
Here, we report a facile method for regulating the morphologies of as-produced graphene nanosheet structures (GNS) through the use of various electrolytes during electrochemical cathodic plasma exfoliation under ambient conditions. Within a short time, we could produce GNS having high specific surface areas. We propose a mechanism for the cathodic plasma exfoliation and discuss the morphologies, structures, and supercapacitive performances of the GNS when used as electrode materials.

Materials
Sodium hydroxide and sulfuric acid were purchased from Sigma-Aldrich. Graphite rods (φ 7 mm; length 70 mm) and graphite (99.5%) were obtained from Toyo Tanso (Taiwan). Platinum plates (length 200 mm; width 30 mm; thickness 0.7 mm) were purchased from Guang Yi Eleciron Chemical Equipment (Taiwan). The GNS samples were prepared by the following procedures: a graphite rod and a platinum plate were used as the cathode and anode, respectively, during the exfoliation process; they were immersed at depths of 10-20 and 60-120 mm, respectively, under the electrolyte. Aqueous 2 M H 2 SO 4 and 4 M NaOH were used as electrolytes. Once a potential of 60 V was applied by the power supply (LinVac Tech, Taiwan), the graphite electrode was covered by the cathodic plasma and GNS were immediately exfoliated into the electrolyte. After vacuum filtration using a nylon filter (Millipore; pore size 0.2 μm), the as-produced samples were washed and dried at 70°C for 3 h. After dispersing the GNS in ethanol at a concentration of 0.3 mg mL −1 , the dispersion was subjected to centrifugation (2000 rpm, 30 min) to remove any un-exfoliated graphite flakes. The supernatant was subjected to further characterizations. The GNS obtained using H 2 SO 4 and NaOH are denoted herein as GNS_H + and GNS_Na + , respectively.

Characterization
The microstructures of the GNS samples were investigated using cold field emission scanning electron microscopy (SU-8010, Hitachi). A spherical aberration-corrected scanning transmission electron microscope (ARM200F, JEOL) was used to obtain transmission electron microscopy (TEM), high-resolution TEM (HRTEM), and fast Fourier transform (FFT) images. HRTEM was further used to determine the thickness distribution of the GNS samples. Raman spectra for determining the defects in the GNS samples were obtained using a Raman spectrometer (HORIBA, LabRAM HR) with a He/Ne laser source (laser excitation wavelength 632.8 nm); N 2 gas adsorption/desorption isotherms were recorded using the Brunauer-Emmett-Teller (BET) method (ASAP 2020, Micromeritics). Pore size distributions were determined using the Barret-Joyer-Halenda (BJH) model and the desorption branches of the isotherms. X-ray photoelectron spectroscopy (XPS; ULVAC, PHI Quantera SXM) and an Al X-ray source (1200 eV) were used to analyze the surface compositions of the as-produced GNS. A D2 X-ray diffractometer (Bruker), equipped with a Cu Kα tube and a Ni filter (λ = 0.1542 nm), was used for structural analysis of graphite and the as-produced GNS.

Electrochemical Measurements
The electrochemical performances of graphite, GNS_H + , and GNS_Na + were determined in 1 M NaCl, using a Zahner Zennium electrochemical workstation in a three-electrode mode, including a standard calomel electrode as the reference electrode and a platinum foil as the counter electrode. The testing was conducted in the potential range of − 0.4 to + 0.6 V. Each electrode was prepared by mixing 90 wt% of the sample and 10 wt% of poly (vinylidene fluoride) with N-methylpyrrolidone (NMP) and then coating the slurry onto graphite papers and drying at 80°C overnight. The specific capacitance (C, F g −1 ) can be calculated from cyclic voltammetry (CV) curves using the equation: where v is the scan rate (mV s −1 ), Δm is the mass of the active material, ΔV represents the potential window, and Idv is the area under the CV curve (Q). Electrochemical impedance spectroscopy (EIS) was also conducted using an amplitude of the alternating voltage of 5 mV; data were collected in the frequency range from 100 mHz to 100 kHz. Figure 1 presents schematic and digital images of the cathodic plasma exfoliation setup and schematic images of the exfoliation process and the as-produced materials obtained using aqueous NaOH and H 2 SO 4 electrolytes. A graphite rod was used as the cathode, while a platinum plate was used as the anode. In contrast to traditional electrochemical exfoliation [44,45], here the ratio of the surface area in contact with the electrolytes between the cathode and anode was rather small (1:10), such that a higher current density could be created at the submerged part of the cathode [46,47]. With rapid heat accumulation arising from the higher current density, the vaporization of water at the cathode initially dominated over the H 2 gas produced from the normal electrolysis. When the voltage difference reached 60 V, the immersed part of the graphite was promptly covered by a stable and thin gas sheath, accompanied by a plasma, resulting in immediate exfoliation of graphene sheets [47,48]. Because of the cations and water in the electrolytes, orange and white plasmas formed when using the aqueous NaOH and H 2 SO 4 electrolytes, respectively. In previous studies [49,50], we found that sheet-like graphene nanosheets could be generated through cathodic plasma exfoliation when using NaOH or KOH as the electrolyte. In this present study, however, we obtained an onion-like morphology when the electrolyte was H 2 SO 4 , presumably because of a different type of cation bombardment during cathodic plasma exfoliation [51]. We discuss the mechanisms affecting the as-produced morphologies along with their materials characterization below. Figure 2a presents FESEM images of GNS_Na + at a magnification of 20 k; the sheet-like morphology indicates the successful exfoliation of typical graphene sheets from the cathodic plasma process, due to thermal stress and potential-driven ion bombardment. Interestingly, Fig. 2b displays a broccoli-like morphology with bunches on the surfaces of GNS_H + material. We suspect that this morphological difference was related to different ion bombardment effects during the cathodic plasma exfoliation: the smaller H + ions not only exfoliated the graphene sheets from the graphite electrode but also decorated the surfaces to create nanostructures, while the larger Na + ions had only an exfoliating effect [50,52]. Figure 2c and d display Fig. 1 Schematic representations of the experimental setup and the cathodic plasma exfoliation processes using NaOH and H 2 SO 4 as electrolytes, along with digital images of the experiments TEM images of GNS_Na + and GNS_H + , respectively. GNS_Na + had a typical graphene morphology with a smooth layer-like surface and fractured edges; in contrast, GNS_H + featured multiple stacked nanostructures. HRTEM revealed that the as-produced GNS_Na + (Fig. 2e) comprised four layers of graphene sheets; an onion-like morphology existed for GNS_H + (Fig. 2f), with diameters from 4 to 10 nm. Interestingly, the rings comprised four to six layers of graphene sheets. The spherical onion-like structure presumably resulted from the effect of H + ions during the bombardment, consistent with the FESEM data. Compared with the larger Na + ions, the smaller H + ions facilitated more thorough bond-breaking and dissociation of the graphene radicals produced during plasma exfoliation [53,54]. As soon as these radical species were sputtered outward from the plasma zone to the electrolyte, they were quenched because of the temperature gradient [55]; at the same time, due to the energy loss of the highly energized radicals, they recombined to form the spherical onion-like structure to minimize the surface energy. The insets to Fig. 2e and f present FFT images of GNS_Na + and GNS_H + , respectively; both reveal typical hexagonal diffraction patterns, suggesting that graphene lattices remained after cathodic plasma exfoliation. Because of the Fig. 2 a, b FESEM images of a GNS_Na + and b GNS_H + ; c, d bright-field TEM images and e, f HRTEM images with corresponding FFT images of c, e GNS_Na + and d, f GNS_H + many diffraction spots arising from the multiple crystal orientations of the onion-like graphene, additional ring patterns were evident in the image from GNS_H + . We could only roughly control the thickness of the as-produced graphene when varying the ions in the electrolytes, owing to the heterogeneous nature of ion intercalation and bombardment [50,56,57]. Additional file 1: Figure S1a reveals the thickness distributions of more than 30 GNS samples, as determined from their HRTEM images. Notably, GNS_H + and GNS_Na + samples having less than six graphene layers made up approximately 87 and 74%, respectively, of all the graphene nanosheets, while the percentage of graphene nanosheets comprising three to six graphene layers was larger for GNS_H + than it was for GNS_Na + , suggesting that the uniformity of the as-produced GNS was associated with the type of ions: the smaller H + ions could induce more bond-breaking and exfoliation than could the Na + ions during the electrochemical plasma process. We performed two additional sets of experiments to examine the reproducibility of the onion-like GNS produced by H + ions. Additional file 1: Figures S1b and S1c reveal the histograms of GNS_H + from the second and third batches; the thickness distributions were similar to those in the first batch, with the HRTEM images of the onion-like morphologies depicted in the insets to Additional file 1: Figures S1b and S1c, suggesting the consistent reproducibility of GNS_H + .

Results and Discussion
We used XPS to investigate the surface chemical composition of the as-produced GNS. Additional file 1: Figure  S2a displays the survey spectra of both samples, revealing the presence of carbon and oxygen species. The oxidation of GNS can be ascribed to the cathodic plasma exfoliation, which radicalized the graphene and electrolytes with water molecules. After radical reactions, quenching, and atomic rearrangement, oxygen-containing functional groups were created at the surfaces of the as-produced materials. Figure  3a and b present deconvoluted C 1s spectra of GNS_Na + and GNS_H + , respectively. In both spectra, we assign the peaks located at 284.5 and 285.1 eV to the bonding of sp 2and sp 3 -hybridized carbon atoms-namely C=C and C-C, respectively [58,59]. Peaks with higher binding energies of 285.7 and 287.1 eV are referenced to C-O and C=O units. The atomic percentages of carbon and oxygen for GNS_Na + were 97.9 and 2.1 at.%, respectively; GNS_H + had a higher oxygen percentage (up to 6.1 at.%). The higher degree of oxidation for GNS_H + presumably resulted from the extra nanostructures and defects introduced by the relatively smaller H + ions-a consequence of more radicalized sites appearing on the graphene during exfoliation.
We obtained further structural information using Raman spectroscopy. In Fig. 4a, the three distinguishing peaks at 1325, 1571, and 2648 cm −1 represent the D-, G-, and 2D-bands, respectively [60][61][62][63]. The D-band represents the A 1g mode, related to defects and structural disorder; the G-band is associated with in-plane bonding of a stretching pair formed from sp 2 -hybridized carbon atoms, as is the E 2g mode; related to the splitting of phonon bands or electron bands, the 2D-band is the secondary order of the D-band. The ratios of the peak intensities I D /I G for graphite, GNS_Na + , and GNS_H + were 0.08, 0.46, and 0.79, respectively. Because the cathodic plasma process enhanced the number of functional groups and edges and increased the structural disorder of the carbon lattice, the defect intensities of GNS_Na + and GNS_H + were higher than that of graphite. Furthermore, we ascribe the higher value of I D /I G for GNS_H + to the greater levels of oxidation and nanostructure formation, as evidenced using XPS, FESEM, and TEM. The relatively distinguishable D′ Fig. 3 High-resolution C 1s XPS spectra of a GNS_Na + and b GNS_H + (1610 cm -1 ) and D + G (2909 cm -1 ) bands confirmed the greater disorder of GNS_H + . Because of van der Waals forces, some agglomeration occurred in the powdery GNS samples, as evidenced from the FESEM images; from our Raman spectra in Fig. 4a, the symmetrical and downshifted 2D-bands of GNS_Na + and GNS_H + , relative to that of graphite, suggest that the few-layer graphene structures were mostly retained without large degrees of restacking to form graphite. Figure 4b also provides box charts of the intensity ratios of the D-and G-bands for GNS_Na + and GNS_H + , each collected from 15 samples. GNS_H + featured higher defect levels on average, suggesting that most of the crystallinity domains in the onion-like morphology were smaller than those of the sheet-like GNS_Na + . Interestingly, the statistical range of values of I D /I G for GNS_H + was narrower than that for GNS_Na + , indicating that GNS_H + had superior uniformity, associated with the smaller H + ions inducing more thorough bond-breaking and dissociation of graphene radicals. The XRD patterns of the graphite, GNS_Na + , and GNS_H + reveal (Additional file 1: Figure S2b) additional structural information. The pattern for graphite featured a sharp diffraction peak at 26.7°, which we assigned to the (002) diffraction, indicating a high degree of graphitization and an interlayer distance 0.334 nm; this peak for both GNS was downshifted by 0.1°, consistent with an increase in their interlayer distances. The full widths at half maximum (FWHMs) of the X-ray diffraction peaks near 26.7°for graphite, GNS_Na + , and GNS_H + were 0.11, 0.40, and 2.7°, respectively, suggesting that the crystal sizes decreased in that order and, thus, that the amorphous nature of the GNS samples increased after cathodic plasma exfoliation. We performed N 2 adsorption experiments to determine the BET surface areas and BJH pore sizes, with distributions, of the samples. Figure 5a reveals that the N 2 adsorption/desorption isotherms had typical H3 hysteresis loops, as classified by the International Union of Pure and Applied Chemistry (IUPAC). The hysteresis loops appeared at relatively low pressure (0.4-0.8), indicating the presence of mesopores in the as-produced materials [50]. The specific surface areas of graphite, GNS_Na + , and GNS_H + were 9, 72, and 464 m 2 g −1 , respectively. The higher specific surface area of GNS_H + is consistent with its onion-like nanostructures, confirming the participation of H + ions during the cathodic exfoliation. Theoretically, a higher surface area for graphene as an electrode material would provide more absorption sites for ions during electrochemical measurements. The BJH pore size distributions of the samples (Fig. 5b) revealed that the dominant pore sizes in the range 2-20 nm for graphite, GNS_Na + , and GNS_H + were 2.5, 14.9, and 9.2 nm, respectively. Thus, the dominant pore size of the as-produced GNS was smaller when fabricated using H + ions. Nevertheless, the pore volume of GNS_H + (0.928 cm 3 g −1 ) was larger than those of GNS_Na + (0.289 cm 3 g −1 ) and graphite (0.058 cm 3 g −1 ), suggesting that the porous structures between the nano-bunches of GNS_H + provided extra spaces for access of the electrolyte and for ion transport during electrochemical charging and discharging.
We used a three-electrode system to determine the electrochemical properties of graphite, GNS_Na + , and GNS_H + in 1 M NaCl at potentials from − 0.4 to + 0.6 V. Figure 6a presents the cyclic voltammograms of graphite, GNS_Na + , and GNS_H + recorded at a scan rate of 5 mV s −1 . We observe quasi-rectangular shapes of the measured curves with no obvious Faradaic reaction, suggesting that the electrode materials had excellent electrical double layer capacitance as a result of highly accessible ions [64]. Table 1 compares the specific surface areas, pore volumes, specific capacitances, and charge-transfer resistances of graphite, Fig. 4 a Raman spectra and b box charts of the intensity ratios of the D-and G-bands of graphite, GNS_Na + , and GNS_H + GNS_Na + , and GNS_H + . The specific capacitances, determined at a scan rate of 5 mV s −1 , for graphite, GNS_Na + , and GNS_H + were 11.4, 21.6, and 67.1 F g −1 , respectively. Because graphene structures of only a few layers were produced during the cathodic plasma exfoliation, the supercapacitive performances of both GNS were higher than that of the graphite; furthermore, the 3D-structured onion-like GNS_H + had even higher specific capacitance because its larger specific surface area and pore volume could accommodate extra ions during charging and discharging. Additional file 1: Table S1 compares the synthesis methods, times, and temperatures; the specific surface areas; and the electrochemical performances (in NaCl solutions at different scan rates) of various graphene materials that have been reported in the literature. Although the specific capacitances in this present study are lower than some of those reported previously, most other synthesis methods have required longer processing times and higher temperatures Fig. 5 a Typical N 2 adsorption/desorption isotherms and b BJH pore size distribution of graphite, GNS_Na + , and GNS_H + Fig. 6 a Cyclic voltammograms of the various samples, recorded at 5 mV s −1 . b Specific capacitances of the samples, determined at scan rates from 5 to 500 mV s −1 . c Galvanostatic charge/discharge curves recorded at 0.1 A g −1 . d Nyquist plots of the samples measured from 100 mHz to 100 kHz; inset: corresponding expanded high-frequency region of the plots to produce graphene materials with high specific surface areas. Hence, we believe that our facile, one-step, and green cathodic plasma exfoliation process is very competitive at producing good-quality GNS. Additional file 1: Figures S3a-c present the cyclic voltammograms of graphite, GNS_Na + , and GNS_H + recorded at various scan rates. Upon increasing the scan rate, the measured curves of each sample retained their quasi-rectangular shapes, suggesting that the electrode materials had excellent electrical double-layer capacitance without obvious Faradaic reaction. Notably, both GNS exhibited excellent electrochemical stability over a wide range of scan rates; the distinct enhancements in current density upon increasing the scan rate suggest superior rate abilities for both of these electrode materials. Figure 6b reveals that the specific capacitances of GNS_H + were relatively higher than those of GNS_Na + at all scan rates from 5 to 500 mV s −1 , presumably because of the higher specific surface area arising from the onion-like nanostructures of GNS_H + . Figure 6c displays the typical galvanostatic charge/discharge curves of the graphene materials GNS_Na + and GNS_H + at a current density of 0.1 A g −1 . Their highly symmetric triangular curves imply that both samples underwent no obvious potential (iR) drops, indicating low internal resistances within these electrode materials. Additional file 1: Figures S3d-f presents the galvanostatic charge/discharge curves of the samples at various current densities. The measured curves retain the notable triangular shapes in their symmetry upon increasing the current density, indicating the outstanding rate abilities for these as-produced GNS with different morphologies. We measured the cycling abilities of GNS_Na + and GNS_H + over 1000 cycles at a scan rate of 100 mV s −1 (Additional file 1: Figure S4). Every material displayed outstanding capacitance retention: 93, 91, and 88% for graphite, GNS_Na + , and GNS_H + , respectively. The slightly lower capacitance retention percentage of GNS_H + presumably arose from non-reversible ion trapping in the smaller pores of its nanostructures during the charging and discharging processes. EIS is a powerful method for measuring the electrical conductivities of carbon electrodes. Figure 6d displays the Nyquist plots of graphite, GNS_Na + , and GNS_H + measured in the frequency range from 100 mHz to 100 kHz. In general, the first intercept to the real axis of the measured curves in the high-frequency region, known as the electrochemical series resistance (R s ), can be related to the ionic conductivity of the electrolyte in the electrode materials, the intrinsic resistance of the electrode materials, and the contact resistance of the materials to the current collector. Small values of R s were evident for both of our samples in their measured curves, indicating the superior conductivity of these as-produced graphene materials. The impedance plot also featured semicircles in the high-and medium-frequency regions, related to chargetransfer processes at the interfaces between the active materials on the electrode and the electrolyte; a 45°inclined line in the low-frequency region suggested Warburg impedance correlated to mass transport [65,66]. The charge-transfer resistances (R ct ) of graphite, GNS_Na + , and GNS_H + were 3.5, 3.9, and 4.6 Ω, respectively, revealing highly conductive graphite and graphene materials that facilitated ionic diffusion during charging and discharging. The slightly greater value of R ct for GNS_H + can be correlated to its smaller crystallinity area and greater level of oxidation, compared with those of graphite and GNS_Na + . Hence, although different morphologies formed in the presence of Na + and H + ions, the conductivities of the GNS produced through cathodic plasma exfoliation were not compromised, due to the preservation of the crystallinity of the graphene lattice. Because of the curvature of the onion-like structures, GNS_H + might have featured more structural dislocations and, thus, more active sites in the graphene hexagonal lattices in the curved basal plane [67][68][69]. Therefore, the ion absorption for the onion-like GNS_H + could have occurred at more edges and basal plane, rather than only at the basal plane as in the case for the sheet-like GNS_Na + . Moreover, the higher oxidative state of GNS_H + could provide a more hydrophilic surface for interaction with the aqueous electrolyte and, thus, allow more efficient transport and accessible micro-tunnels for ions during charging and discharging, resulting in higher supercapacitive performances.

Conclusion
We have produced typical sheet-like GNS when using aqueous NaOH as the electrolyte in a cathodic plasma exfoliation process performed within a short period of time under ambient conditions. When the electrolyte was replaced by aqueous H 2 SO 4 , an onion-like morphology was introduced to the as-produced GNS, which featured a superior specific surface area (464 m 2 g −1 ) and pore volume (0.928 cm 3 g −1 ). We suspect that the involvement of H + ions facilitated more thorough bond breaking and dissociation of radical species than did the Na + ions during the cathodic plasma exfoliation. Measurements of supercapacitive performance at a scan rate of 5 mV s −1 in 1 M NaCl indicated that the GNS featuring the onion-like nanostructures had a specific capacitance (67.1 F g −1 ) higher than that of the GNS having the sheet-like morphology (21.6 F g −1 ). Thus, it is possible to produce GNS with different morphologies and supercapacitive performances when using readily accessible electrolytes in a facile cathodic plasma exfoliation process; furthermore, the unique onion-like GNS structure, with some retained crystallinity and curvature, created by the H + ions, exhibited extraordinary conductivity and a high specific surface area, suggesting greater potential (relative to that of the corresponding sheet-like GNS) for use in energy storage devices.

Additional file
Additional file 1: Figure S1. (a) The thickness distribution of more than 30 GNS determined from HRTEM of GNS_Na + and GNS_H + (the first batch); (b) and (c) are GNS_H + produced from the second and the third batch of the cathodic plasma exfoliation, respectively (scale bar 4 nm). Figure S2. (a) XPS survey spectra of GNS_Na + and GNS_H + ; (b) XRD diffraction patterns of graphite, GNS_Na + , and GNS_H + ; inset: corresponding expanded region of the plots. Table S1. Synthesis methods, times, and temperatures; specific surface areas; and electrochemical performances, in NaCl solution at various scan rates, of graphene materials reported in the literature. *rt room temperature. Figure S3. (a-c) Cyclic voltammograms and (d-f) galvanostatic charge/ discharge curves of (a, d) graphite, (b, e) GNS_Na + , and (c, f) GNS_H + , recorded at various scan rates. Figure S4. Cycling tests of graphite, GNS_Na + , and GNS_H + over 1000 cycles at a scan rate of 100 mV s −1 .