Fabrication of 3D binder-free graphene NiO electrode for highly stable supercapattery

Electrochemical stability of energy storage devices is one of their major concerns. Polymeric binders are generally used to enhance the stability of the electrode, but the electrochemical performance of the device is compromised due to the poor conductivity of the binders. Herein, 3D binder-free electrode based on nickel oxide deposited on graphene (G-NiO) was fabricated by a simple two-step method. First, graphene was deposited on nickel foam via atmospheric pressure chemical vapour deposition followed by electrodeposition of NiO. The structural and morphological analyses of the fabricated G-NiO electrode were conducted through Raman spectroscopy, X-ray diffraction (XRD), field emission scanning electron microscopy (FESEM), and energy dispersive X-ray spectroscopy (EDS). XRD and Raman results confirmed the successful growth of high-quality graphene on nickel foam. FESEM images revealed the sheet and urchin-like morphology of the graphene and NiO, respectively. The electrochemical performance of the fabricated electrode was evaluated through cyclic voltammetry (CV), galvanostatic charge–discharge (GCD), and electrochemical impedance spectroscopy (EIS) in aqueous solution at room temperature. The G-NiO binder-free electrode exhibited a specific capacity of ≈ 243 C g−1 at 3 mV s−1 in a three-electrode cell. A two-electrode configuration of G-NiO//activated charcoal was fabricated to form a hybrid device (supercapattery) that operated in a stable potential window of 1.4 V. The energy density and power density of the asymmetric device measured at a current density of 0.2 A g−1 were estimated to be 47.3 W h kg−1 and 140 W kg−1, respectively. Additionally, the fabricated supercapattery showed high cyclic stability with 98.7% retention of specific capacity after 5,000 cycles. Thus, the proposed fabrication technique is highly suitable for large scale production of highly stable and binder-free electrodes for electrochemical energy storage devices.


Results and Discussion
Structural and morphological characterisations. Raman spectrum gives useful information on the quality of the as-synthesized graphene and number of layers. These properties are typically estimated based on the ratio of I 2D /I G band intensities as well as a shift in their peak positions. The comparative Raman spectra of the bare nickel foam (NF) and G-Ni electrode are shown in Fig. 1a. Bare NF maintained a straight line without any peaks. However, the G-Ni electrode showed the prominent graphitic characteristics of the sp 2 hybridized carbon atoms (G and D bands) well bonded on the NF (Fig. 1a). This is an evidence of highly crystalline, few to multi-layered graphene sheets deposited on the NF. The graphitic layers were observed at the wavenumber value of 1,380 cm −1 , 1577 cm −1 , and 2,720 cm −1 for D, G, and 2D bands, respectively. The D band showed a slightly Scientific RepoRtS | (2020) 10:11214 | https://doi.org/10.1038/s41598-020-68067-2 www.nature.com/scientificreports/ defective graphene structure, while the G and 2D bands depicted the first (tangential stretching mode of highly oriented pyrolytic graphite-HOPG) and second-order phonon scattering, respectively. These results agree with the previous studies [37][38][39][40] . X-ray diffraction determines the crystallographic structure of the electrodes. The X-ray diffractometer obtained the XRD patterns over the 2θ range from 5° to 90° with monochromatized Cu K-α radiation. Figure 1b shows the comparative XRD patterns of NF (with lattice parameters of (111) at 44.51°, (200) at 51.85°, and (220) at 76.37° for NF 41 (which served as baseline spectrum), G-Ni electrode and G-NiO electrode. All the diffraction peaks for Ni and NiO can be indexed as a cubical crystalline phase of NiO (JCPDS card no.   42 . In the NF deposited graphene, a distinct peak was observed at angle 2θ peak position 26.45°, which confirmed the presence of graphene 37,38 . Figure 1c shows the XRD pattern of the G-Ni electrode; inset depicts the expanded graphene peak, which corresponds to the lattice parameter of carbon (002). There was a reduction in the peak intensity due to the graphene layers on the NF (Fig. 1d). It could be seen that the NiO thin film formed homogenous oxide layers on the G-Ni electrode without any obvious impurities. A different spectra showing the expended angle peak positions and peak intensities in the XRD pattern of the G-Ni electrode are presented in the supplementary data online (Figs. S2, S3 and S4).
Field emission scanning electron microscopy equipped with energy dispersive X-ray analyser was used to characterize the morphology and elemental composition of the electrodes. Figure 2a shows the FESEM micrographs of NF revealing the clean surface without any contaminations. Similarly, Fig. 2b and Fig. 2c depict the thin film of graphene grown on the NF at different magnifications. The FESEM micrograph of the G-NiO electrode showing the sheet and urchin-like morphology is shown in Fig. 2d. To determine the percentage of the atomic composition for both G-Ni and G-NiO electrodes, EDS mappings were done ( Fig. 2e and Fig. 2f) and the results are presented in Table 1. The FESEM image and EDS spectrum of the NiO electrode are provided in the supplementary data online (Fig. S9).   Fig. 3a and Fig. 3b, respectively. The CV curve of the NiO electrode shows well-defined redox peaks, which is an indication of a battery electrode. The peaks indicated the redox transition of Ni ions due to noncapacitive faradaic behaviour 31 . The redox peaks are attributable to the diffusion of electrolyte in the material which suggested that the NiO electrode was showing battery-type behaviour 31,43 . The electrode redox behaviour was based on the Nernstian process as depicted by the peak-shaped CV curves. Distinct redox peaks were observed at high scan rates showing a high rate of capability and good reversibility of the working electrode 44 . From Fig. 3b, it is evident that, with increased scan rates, the redox peak current of the G-NiO electrode was significantly increased with better reversibility as compared to NiO electrode due to the presence of graphene. The introduction of graphene improved the exposure of active sites of the electrode for the faradaic reactions owing to the enhanced surface area. Additionally, the amorphous phase formed from the graphene and NiO thin film aids electroactivity. Figure 3c illustrates the comparative CV curves of NiO and G-NiO at the same scan rate of 3 mV s −1 revealing the higher specific capacity of the latter. G-NiO exhibited a high, and remarkable peak current compared to NiO. It is apparent that the electrochemical performance of the G-NiO was drastically improved, and this was due to the high surface area and conductive platform provided by graphene thin film. This high background current for the G-NiO electrode compared to NiO was as a result of the highly conductive platform of the graphene in the G-NiO electrode. The overlay of CV curves (at 3 mV s −1 ) for the NF and G-Ni electrode showing the significant difference in electrochemical performance is shown in the supplementary data online (Fig. S5). Therefore, the combination of NiO thin film with graphene improved its conductivity and capability performance. The specific capacity values of the electrodes, Q s (C g −1 ) from the CV curve were calculated from the relation as expressed in Eq. (1). where v and m represent the scan rate (mV s −1 ) at which the voltammogram is recorded and the mass loading of active materials (g) respectively. The integral term represents the area under the redox peaks. The value of the specific capacity as against different scan rates for NiO and G-NiO electrodes are presented in Table S1 of the supplementary data online. The maximum specific capacity achieved from the G-NiO electrode was 243 C g −1 which is significantly higher in value as compared to 96 C g −1 obtained from NiO electrode. Figure 3d represents the corresponding specific capacity of the two electrodes as a function of scan rate. It could be observed that an increase in the scan rate decreased capacity due to insufficient time for ions to penetrate the inner pores of the electrode, and thus, resulted in a low capacity and vice versa 45 .
Similarly, the galvanostatic charge-discharge (GCD) studies were conducted to investigate the stability of the NiO and G-NiO electrodes. Figure 4a and Fig. 4b represent the charge-discharge plots for the two electrodes at different current densities in the range of 0.6 A g −1 to 2 A g −1 over a potential window of 0.6 V (− 0.2-0.4 V). The battery-like characteristics of the electrodes could be observed from the charge-discharge behaviour which was based on the non-capacitive faradaic reaction mechanism as the electrolyte was dispersed. The discharge time was decreased with an increase in current density which implied an inverse relationship (in both cases). Figure 4c is the comparison GCD plot for both electrodes revealing a higher specific capacity for the G-NiO electrode. The specific capacity of the electrodes, Q s (C g −1 ) were evaluated from the GCD curve using the relation as given in Eq. (2). www.nature.com/scientificreports/ where I, m, and t represent the discharging current (A), mass loading of active materials (g), and time taken to fully discharge the electrode (s) respectively. The value of the specific capacity against diverse current densities for NiO and G-NiO electrodes are presented in Table S2 of the supplementary data online. The specific capacity values were estimated to be 92 C g −1 for the G-NiO electrode and 44 C g −1 for the NiO electrode. The Q s value for the G-NiO is significantly higher compared to the results obtained in the previous studies that investigated transition metal oxides (TMOs) as electrodes 46,47 . The GCD measurements showed that the specific capacity value was highest at the lowest current density, and decreased with an increase in current density up to 2 A g −1 . This effect was well illustrated in the charge-discharge plots where the Q s value decreased from 92 to 26 C g −1 for G-NiO with a retention of 71.2%. Likewise, the NiO electrode retained 58.7% which was a reduction from 44 to 18 C g −1 . The tremendous enhancement in the specific capacity of the G-NiO electrode is ascribed to the higher specific surface area of graphene (which boosted charge mobility).
It is noted that the 3D nanostructure of the G-NiO electrode enabled excellent acceleration of the ionic diffusion into various directions and thus, minimized differential activation energy of the ionic species. Moreover, the NiO thin film contributed to the synergistic effect of the nanocomposite. The direct relationship of specific capacity with surface area applies to the related studies involving mesoporous materials such as NiO 48 and carbon 49 , however in the binder state. A larger current accommodation during the charge-discharge process indicates a good battery-grade material behaviour with respect to charge storage mechanism, thereby providing more active sites to generate a larger number of redox activities owing to the large specific surface area. This capability is demonstrated by the longer discharge time as recorded in Fig. 4c for the G-NiO electrode. Figure 4d depicts the comparative specific capacity as a function of current density for the NiO and G-NiO electrodes.
The electrochemical performance of the fabricated electrodes, in terms of charge storage kinetics, were studied through electrochemical impedance spectroscopy (EIS). EIS characterisation is associated with migration and diffusion of reactants towards or away from the electrode surface and thus, produces a peculiar frequency character known as the Warburg impedance 50 . The frequency character dependence on the impedance can demonstrate the underlying electrochemical phenomenon. EIS study was performed on the G-Ni, G-NiO, and NiO electrodes to examine, and compare the associated mechanism occurring on the electrode surface. A frequency range of 0.01-100 kHz was employed at an alternating signal of 10 mV (RMS). Figure 5 demonstrates the comparison of EIS spectra for the three electrodes. All spectra consisted of a typical semicircle, followed by a straight line. The high frequency region in the spectra gives the parametric information on the electrode resistance. The diameter of the semicircle in the spectrum is a measure of the resistance arising from the charge transfer kinetics which is associated with the electrode/electrolyte interface, and electrode geometry 51 . A straight line taken from the frequency knee point (real axis-Z re ) depicts electrode capacitive response. Ideally, a straight line response would be parallel to the imaginary axis (Z im ). The line gradient in the low frequency region measures the diffusion resistance called the Warburg impedance, W 12 . The illustration of the behaviour of the electrode in the high frequency region is depicted in Fig. 5 (inset) showing an expanded plot for the G-Ni, G-NiO, and NiO electrodes. As expected, the plots distinctly illustrated the semicircle in the high frequency region, and a straight line in the low frequency region.
From the Nyquist plot (inset (i) of Fig. 5), the impedance behaviour of the G-Ni electrode is portrayed and shown by a vertical line in the low frequency region showing an ideal capacitive behaviour coupled with a low charge transfer resistance (R ct ). Similarly, the NiO electrode revealed a bigger semicircle (R ct = 0.45 Ω) than G-NiO (R ct = 0.16 Ω), with the least being G-Ni electrode (R ct = 0.12 Ω) in the high frequency region which implied that G-Ni electrode had the lowest charge transfer resistance in comparison with others. The equivalent www.nature.com/scientificreports/ circuit was used to obtain the R ct of the electrodes (inset (ii) of Fig. 5). Typically, three factors play a key role in the equivalent series resistance (ESR). They are; (a) discontinuity in ionic conductivity and electric conductivity occurring during charge transfer mechanism (b) resistance between NF (electron collector) and the connectors (leads), and lastly, (c) the intrinsic resistance of the NF 52 . The G-Ni electrode relatively recorded the best line which was parallel to the imaginary axis (Z im ) and thus, proved that it had a better storage capacity due to better conductivity. This result was expected as the G-Ni electrode had a uniform thin layer of graphene well distributed on NF. However, the G-NiO electrode which comprised graphene thin film and NiO thin film in the amorphous phase was considered highly conductive and hence, the pathways for ion transport to the nickel foam were increased. This phenomenon gave rise to the high specific capacity even at high current density and making it a more appropriate material for supercapattery development.
Asymmetric device assemblage and performance. Supercapattery is a hybrid device that has the features of both capacitor and battery. The characteristic parameter in supercapattery is the energy density which is enhanced by extending the operating voltage window. Since, energy density is a function of electrode capacity and cell voltage, a supercapattery device incorporating EDLC, and battery-grade utilises the combined potential window of the two electrodes leading to a higher potential window of the device. In this study, the supercapattery was configured by employing activated charcoal (AC) as the negative electrode and G-NiO as the positive electrode as shown in Fig. 6a. As the first step in the electrochemical studies, the individual cyclic voltammograms were run for both G-NiO and AC electrodes in a three-electrode system at room temperature. These preliminary analyses allowed for the accurate estimation of the maximum possible operating potentials by investigating the individual electrode properties. It was observed that G-NiO and AC electrodes operated optimally in the potential range of − 0.2-0.4 V and − 1.0-0 V, respectively (Fig. 6b). Note that, from Fig. 6b, the potential window for the fabricated G-NiO//AC supercapattery can be extended from 0 to 1.4 V by combining the potential range of www.nature.com/scientificreports/ both G-NiO and AC electrodes. Thus, the maximum stable working potential window for the assembled device was taken to be 0-1.4 V. At a fixed scan rate of 10 mV s −1 , voltammograms at different potential ranges were recorded to observe the rate capability of the device (Fig. 6c). The cyclic voltammograms of the assembled device over the potential window of 0-1.4 V recorded at diverse san rates (3-200 mV s −1 ) is represented in Fig. 6d. The capacitive carriage contribution of AC was observed in the potential range of 0-0.5 V. From Fig. 6d, it could be seen that until a potential of 0.5 V, the shape of the CV curve was rectangular which manifested that the charge storage was mainly due to the EDLC effect. However, at potentials beyond (above 0.5 V), redox peaks also appeared revealing that charge storage was due to the Nernstian processes. Hence, over the full potential window of 0 to 1.4 V, the charge storage mechanism was due to both, EDLC and non-capacitive faradaic reactions. At the potential range of 0-0.5 V, energy storage was contributed by capacitive behaviour while beyond 0.5 V (0.5-1.4 V), non-capacitive faradaic redox reactions were dominant in the energy storage mechanism. High rate capability and stability of the supercapattery had been demonstrated by the shape, constancy, and amplification of the CV curves at the various scan rates (3-200 mV s −1 ). An estimation of the specific capacity, Q d of the assembled asymmetric device was done using the relation in Eq. (3). The assembled asymmetric device operated at a specific capacity of 67.8 mA h g −1 at a current density of 0.2 A g −1 . Table S3 presents the specific capacity values against the diverse current densities of the device while Fig. S1 illustrates the device rate capacity (see supplementary data online). Figure 7a depicts the charge-discharge plots of the G-NiO//AC supercapattery (asymmetric device) measured over different potential window at the same current density to observe the device performance while Fig. 7b shows the charge-discharge plots at various current densities in the potential range of 0-1.4 V. The symmetrical charge-discharge plots illustrated the capacitive nature combined with high reversible redox reactions. To evaluate the efficiency of the assembled device, the energy density, E (W h kg −1 ), and power density, P (W kg −1 ) are the key parametric quantities of the fabricated supercapattery and therefore, were calculated by using the relation in Eqs. (4) and (5), respectively.
where Q d is the specific capacity of the device (C g −1 ), V is the operating voltage window (V), and t is the discharge time (s). The specific energy and power densities were measured as 47. To evaluate the viability of the assembled supercapattery employing G-NiO as the positrode, a prolonged cycling stability studies were performed over charge-discharge of 5,000 cycles. Figure 8 represents the cyclic stability performance of the assembled asymmetric device showing 98.7% specific capacity retention after 5,000 cycles. From the stability test, it could be observed that the activation process of the electrodes in the electrolyte www.nature.com/scientificreports/ was stable. Initially, the specific capacity increased until 100 cycles, and then decreased slightly (≈ 1%) until 500 cycles, and then gradually and slowly decreased until the 5,000th cycle. Fortunately, 97.8% capacity retention was achieved after 5,000 cycles for the G-NiO//AC supercapattery. The high cyclability of the assembled supercapattery using G-NiO electrode assures its viability for practical usage. The charge-discharge plot for the first 10 cycles (during cycling) is shown in the supplementary data online (Fig. S6). Figure 8 (inset) depicts the overlay of the Nyquist plot for the assembled device showing the EIS spectra of initial (before) and after life cycle test, and portrays the cyclic stability of the device. As expected, in the high frequency region, a semicircle was apparent whereas, in the low frequency region, a straight line was observed for the device. The semicircle diameter measurements denoted the low cell resistance before cycling which suggested a short path travelled by an ion/electron. However, after cycling, the EIS spectrum revealed that the ion/ electron transport pathway was longer as indicated by the higher charge transfer resistance. Similarly, before cycling, the low frequency region of the plot showed the high line gradient in the Nyquist plot indicating low interfacial diffusion resistance (fast ion diffusion and mass transport at electrode/electrolyte interface). Accordingly, the Nyquist plots reported the charge transfer resistances of the device to be R ct = 0.175 Ω (before) and R ct = 0.182 Ω (after life cycle test). As an investigation of the morphological properties of the G-NiO electrode after cycling, a post-mortem analysis was done through FESEM to observe the morphological changes after the life cycle test. The micrographs for the freshly fabricated G-NiO is depicted in Fig. S7 while the micrograph showing the structural and morphological degradation of the electrode after the life cycle test is shown in Fig. S8 (see supplementary data online).

Conclusions
A three-dimensional (3D) graphene electrodeposited nickel oxide thin film (G-NiO) nanocomposite as a binderfree electrode has been successfully fabricated by a two-step route: APCVD and electrochemical deposition. The binary binder-free nanocomposite exhibited an excellent electrochemical performance owing to the synergistic effect of NiO thin film, and a highly conductive graphene platform. Moreover, the fabricated supercapattery in the configuration of G-NiO//activated charcoal supercapattery demonstrated a remarkable electrochemical performance. The assembled supercapattery (G-NiO//AC) portrayed a good balance between the parametric quantities of the device in terms of its energy density and power density. Furthermore, the evaluation of the device cyclic stability showed an excellent cycle life with only a 1.3% loss of its initial specific capacitance after 5,000 cycles. Hence, the unique properties of the supercapattery characterize its excellent electrochemical performance and thus, ensures sustainability.
Characterisation techniques. The experimental results validation was done using various characterization techniques. The quality of the graphene (in terms of purity) as well as the properties of the G-NiO electrode were investigated using the WITec's ALPHA 300 M+ at an excitation wavelength of 514 nm (Raman spectroscopy). The surface morphology, elemental composition, and mappings were studied using a JEOL: JSM-7800F microscope fitted with energy dispersive X-ray analyser (FESEM-EDS). The crystalline nature and phase www.nature.com/scientificreports/ identification of the fabricated electrodes were recorded using a PANalytical-X'Pert MPD X-ray diffractometer equipped with Cu K-α radiation (λ = 1.5418 Å) at a scan rate of 0.2 s −1 ; step 0.05° over a 2θ range of 5° to 90°.

Electrochemical (EC) measurements.
The electrochemical workstation, VersaSTAT-3F model, was used to examine the electrochemical performance of the fabricated electrodes at room temperature. A 6 M KOH, saturated calomel electrode (SCE), and a platinum wire were used as the electrolyte, reference, and counter electrodes, respectively. The procedure involved the dipping of the sample electrode, with a dimension of 1 × 1 cm (working electrode) into the aqueous electrolyte. Cyclic voltammetry (CV), galvanostatic charge-discharge (GCD), and electrochemical impedance spectroscopy (EIS) techniques were used to determine the electrochemical parameters of the electrodes. The CV voltammograms were recorded within certain potential windows at different scan rates, while the GCD characteristics were measured at different potential windows, and various current densities. The EIS measurements were recorded in the frequency range of 0.01 Hz to 100 kHz at alternating current voltage of 10 mV (RMS).
Fabrication of 3D binder-free graphene electrode. Firstly, pre-treatment of the nickel substrate was done using HCl, DI water, and ethanol. HCl (0.1 M) was used to wash the substrate-NF (size: 1 cm × 2 cm) under ultrasonic bath for 10 min to remove the native oxide layer and dirt. The substrate was further cleaned with DI water in the bath sonicator for 10 min at room temperature. Finally, the substrate was rinsed with ethanol and dried in a vacuum oven at 80 °C for 6 h. The pre-treated substrate was then placed in a clean glass boat and loaded into a tube furnace. Atmospheric pressure CVD was used to grow graphene on the substrate under optimal synthesis conditions. The nickel foam served two purposes in the synthesis; firstly, as a catalyst in lowering the activation energy of the gaseous precursor and secondly, as a support material for the growth of graphene thin film. Methane (CH 4 ) was employed as the precursor gas. Hydrogen (H 2 ) was used as the etching gas, while Argon (Ar) was used to de-oxygenate the furnace and to maintain an uncontaminated synthesis environment throughout the synthesis period. Figure 9 depicts is the experimental setup for the preparation of the 3D graphene electrode (G-Ni).
Fabrication of 3D graphene supported NiO electrode. A solution of NiCl 2 ·6H 2 O (0.50 M) was prepared which served as the electrolyte and precursor for the growth of NiO thin film onto the G-Ni electrode.
The electrodeposition was carried out in a three-electrode cell configuration in which the G-Ni electrode was designated as the working electrode, while platinum wire and SCE served as the counter and reference electrodes, respectively. Electrodeposition was conducted through chronoamperometry for 20 min at a constant potential of − 1.2 V to grow NiO onto the G-Ni electrode. After electrodeposition, the working electrode was removed, washed with DI water, and dried in an oven at 80 °C for 12 h. Then, it was re-weighed to determine the mass loading of NiO on the G-Ni electrode. The fabricated nanocomposite was then designated as the G-NiO electrode.