Electrodeposition of honeycomb-shaped NiCo₂O₄ on carbon cloth as binder-free electrode for asymmetric electrochemical capacitor with high energy density

Abstract

Combining high-capacitive metal oxides and excellent conductive carbon substrates is a very significant strategy to achieve high-performance electrodes for electrochemical capacitors (ECs). Herein, the bimetallic (Ni, Co) hydroxide is uniformly grown on the electro-etched carbon cloth (CC) by a facile co-electrodeposition method, and then the honeycomb-shaped NiCo₂O₄/CC (HSNC) composite is formed by transforming the hydroxide precursor into its bimetallic oxides through the subsequent thermal treatment. The special structure of the HSNC as binder-free electrode is responsible for its excellent electrochemical performance with carbon-like power feature. The experimental results show that HSNC electrode exhibits a high specific capacitance with remarkable cycle stability (94.3% after 10 000 cycles at 10 A g⁻¹) in the three-electrode configuration. To evaluate further the capacitive performance of the as-prepared binder-free electrode in a full cell set-up, an asymmetric electrochemical capacitor (AEC) is assembled by using the HSNC as the positive electrode and reduced graphene oxide/carbon cloth (rGO/CC) as the negative electrode in KOH electrolyte. The as-assembled device presents an energy density as high as 32.4 W h kg⁻¹ along with power density of 0.75 kW kg⁻¹, comparing with nickel-metal hyoride battery (Ni-MH) batteries (30.0 W h kg⁻¹ at 0.35 kW kg⁻¹). Even at the power density of 37.7 kW kg⁻¹ (50-time increase, a full charge–discharge within 3.5 s), energy density still holds at 17.8 W h kg⁻¹, indicating an outstanding rate capability. Furthermore, the as-fabricated device exhibits a long cycle lifetime (76.5% after 10 000 cycles at 3 A g⁻¹) with a cell voltage of 1.5 V.

---

1. Introduction

At present, it is imperative to develop and scale up environmentally friendly, sustainable, high efficient energy conversion and storage devices. Among various energy storage devices, supercapacitors (SCs), also called electrochemical capacitors (ECs), have received considerable attention because of outstanding reliability, fast charge–discharge rates (with seconds) and excellent cycling stability (≥10 000 cycles, 10 times more than Li-ion batteries). It is even expecting that it becomes one of the well-deserved candidates for next-generation energy storage devices to bridge the gap between conventional capacitors and batteries.^1–4 According to the charge storage mechanism, SCs are usually classified into two categories, that is, electrical double-layer capacitors (EDLCs), in which energy storage is on the basis of ion adsorption, and pseudocapacitors where fast and reversible faradaic reactions and/or intercalation of ions occurs on the surface of electrode material. The EDLCs electrode is generally composed of carbon materials with high surface areas, and accordingly some metal oxides with reversible multi-electron Faraday reactions,^5–7 such as RuO₂,⁸ MnO₂,⁹ TiO₂,¹⁰ SnO₂ (ref. 11) and so on, are suitable for using as electrode materials for pseudocapacitors. Interestingly, the several spinel metal oxides, such as Co₃O₄ and NiCo₂O₄, not only deliver capacitive contribution since counterpart of spinel compound can undergo a second reversible redox reaction, but also exhibit battery-type behavior due to the phase transitions during the charge storage process.¹² The pseudocapacitive contribution is expected to emphasize when the structure of these materials is designed at the nanoscale because storage sites are on the surface or near-surface region. In this case, kinetic features of the materials will exhibit a capacitive behavior that the redox process is controlled by the surface reaction rather than semi-infinite diffusion, that is to say, the peak current i varies with scan rate (ν) linearly.^12–14

The spinel nickel cobaltite (NiCo₂O₄) has attracted much attention because of many intriguing superiorities, such as high electrical conductivity and various oxidation states of both nickel and cobalt to store more charges. More importantly, the energy storage property of NiCo₂O₄ results from both the reversible transformation between the parent spinel compound and nickel and cobalt oxyhydroxides and reversible fast second redox reaction of the cobalt oxyhydroxide.^12,15,16 Therefore, NiCo₂O₄ could hold high specific capacitance even at short charge–discharge times. In this respect, Lou and his team have done interesting works. They prepared several kinds of Ni–Co mixed oxide with unique morphology such as yolk-shelled nanoprisms,¹⁷ tetragonal microtubes¹⁸ and double-shelled nanocages,¹⁹ and showed the enhanced electrochemical performance in their application of hybrid supercapacitors, lithium ion batteries and electrocatalytic activity for the oxygen evolution reaction. Moreover, spinel NiCo₂O₄ nanomaterial, particularly, combining with conductive substrates, plays an important role in the energy storage fields. For example, Liu et al. prepared porous hetero structure NiCo₂O₄ array on nickel foam, which showed a specific capacitance of 619 F g⁻¹ even at 40 A g⁻¹.²⁰ Zhang and his co-workers fabricated mesoporous NiCo₂O₄ nanowire arrays (NWAs) grown uniformly on carbon textiles via a simple surfactant-assisted hydrothermal method, which can achieve a high specific capacitance (1283 F g⁻¹ at 1 A g⁻¹).²¹ Emphasizedly, NiCo₂O₄ nanosheet arrays, which were electrodeposited on flexible carbon fabric (mass loading: 0.6 mg cm⁻¹) by Du and his co-workers, were used as a high-performance electrode that exhibited an ultrahigh specific capacitance of 2658 F g⁻¹ at 1 A g⁻¹ in a potential range of −0.1 to 0.35 V.²² Obviously, NiCo₂O₄ is advantageous over other electrode materials in specific capacitance. However, the high specific capacitance of the electrode materials in three-electrode system does not mean the high energy density obtained in two-electrode device. We know that the energy density (E) of the device depends on the capacitance (C) and/or cell voltage (V). That is E = 1/2CV².²³ The C and V are determined by the integrated device rather than single electrode. Therefore, it is a challenge to match potential windows and specific capacitances of positive and negative electrodes to release optimal electrochemical performance in two-electrode system, and accordingly improve the energy density of the devices.

In this paper, we proposed a simple strategy to electro-etch carbon cloth (CC), which can obtain wide space between fibers for growth of active components. Subsequently, we used modified Du's²² electrodepositing method to directly grow NiCo₂O₄ nanosheet arrays on electro-etched CC with honeycomb-shaped structure (donated as HSNC), which is severed as positive electrode directly. Results shown that the HSNC obtained an ultrahigh cycle performance (capacitance retention of 94.3% after 10 000 cycles at a specific current of 10 A g⁻¹), which outperforms Du's result (capacitance retention of 80% after 3000 cycles at a specific current of 10 A g⁻¹). To match the capacitive behavior of HSNC electrode, reduced graphene oxide/carbon cloth (rGO/CC) electrode, which was prepared by coating reduced graphite oxide on the electro-etched CC via a “dip and dry” method, was used as negative electrode without the use of other binders or conductive additives. As a result, the as-fabricated asymmetric electrochemical capacitor (AEC) delivered a high energy density (32.4 W h kg⁻¹ along with 0.750 kW kg⁻¹) in aqueous electrolyte solutions. Importantly, the energy density of the device still held at 17.8 W h kg⁻¹ (54.9% of the initial value) when its power density was up to 37.7 kW kg⁻¹ (a full charge–discharge within 3.5 s), indicating an excellent rate capability.

2. Experimental

2.1 Materials

All reagents used in our experiments are analytical grade without further purification. Cobalt nitrate (Co(NO₃)₂·6H₂O), nickel nitrate (Ni(NO₃)₂·6H₂O), absolute ethyl alcohol (C₂H₅OH), acetone (C₃H₆O), trisodium citrate dihydrate (99.0%, Na₃C₆H₅O₇·2H₂O) and deionized (DI) water.

2.2 Sample preparation

2.2.1 Electro-etching of carbon cloth. CC was purchased from Hong Kong Physicochemical Limited Company. Being used before, CC was first cleaned by continuous ultrasonication in acetone, DI water, and ethanol for 1 h, respectively, and then dried in an oven at 60 °C for 12 h. In order to improve the wettability of fibers and enlarge space interval between fibers, we properly electro-etched CC by a direct-current power with constant voltage at 4.0 V in 0.05 M Na₃C₆H₅O₇ solution at 40 °C for 12 h. Successively, the treated CC was soaked in acetone at room temperature for 24 h, washed with DI water and then dried in air. After electro-etching, the mass of CC decreased from 12 mg cm⁻² to 8 mg cm⁻².

2.2.2 Growth of NiCo₂O₄ nanosheets on electro-etched CC fibers. NiCo₂O₄ sheets were grown on the surface of the electro-etched CC by a facile electrodeposition method and subsequent thermal treatment.²² Typically, 10 mmol of Ni(NO₃)₂·6H₂O and 20 mmol of Co(NO₃)₂·6H₂O were dissolved in DI water (50 mL) with magnetic stirring for 30 min at room temperature as electrolyte. The electrodeposition was performed in a standard three-electrode configuration at room temperature of 25 ± 1 °C by using an electrochemical workstation (Cheng-hua, China). The electro-etched CC was directly used as the working electrode, a platinum sheet was used as the counter electrode and a saturated calomel electrode (SCE) was used as reference electrode, respectively. The electrodeposition potential is −1.0 V (vs. SCE). After electrodeposition, the samples were carefully rinsed several times with DI water and absolute ethanol with the assistance of ultrasonication, and finally dried in air. Then, the resultant samples were calcined at 300 °C for 2 h in air atmosphere to get the NiCo₂O₄/CC composites.

2.2.3 Synthesis of rGO/CC. Graphite oxide (GO) was prepared by the modified Hummers method.²⁴ The rGO/CC was synthesized by a simple “dip and dry” method like our research group previous report.¹¹ Typically, GO powder was dispersed by ultrasound bath for 4 h to obtain monolayer GO solution with concentration of about 5.0 mg mL⁻¹. The first step, the electro-etched CC (3 × 3 cm²) was immersed the GO dispersion for 30 min, ensuring GO completely cover the surface of the CC, and then dried at room temperature. This “dip and dry” process was repeated for several times to control the mass loading of the GO sheets. The second step, as-prepared sample was placed in a muffle and heated up to 300 °C at a rate of 1 °C min⁻¹ for 2 h to obtain rGO/CC.

2.3 Characterization of samples

The crystal of the samples were determined by using a powder X-ray diffraction (XRD) equipped with Cu Kα radiation (λ = 0.15418 nm) operating at 40 kV, 60 mA. The morphologies and microstructures of the samples were characterized by field emission scanning electron microscopy (FESEM; LTRA plus, Germany), high-resolution transmission electron microscopy (HRTEM) and selected area electron diffraction (SAED) (JEOL JEM 2100 system operating at 200 kV). X-ray photoelectron spectroscope (XPS) measurement was performed on Escalab X-ray photoelectron spectrometer with a monochromatic Al Kα (1486.6 eV) as an X-ray source. The atomic ratio was measured by inductively coupled plasma (ICP) optical emission spectroscopy (725-ESICPOES, NYSE: A, America).

2.4 Electrochemical measurement

2.4.1 Three-electrode system. The samples for different electrodeposition times and rGO/CC were directly acted as the working electrodes without any other conductive additives or binders for the following electrochemical tests, such as cyclic voltammetry (CV), galvanostatic charge–discharge (GCD) and electrochemical impedance spectroscopy (EIS) performed in 3 M KOH electrolyte. All measurements were carried out by using a CHI 760E electrochemical workstation (Cheng-hua, Shanghai, China), in which a platinum plate and an Hg/HgO were used as the counter electrode and the reference electrode, respectively. The specific capacitance (C, F g⁻¹) of the electrode is calculated from the GCD curves based on the following equation:^7,16where I (A g⁻¹), t (s) and ΔV (V) are the discharging specific current, the discharging time and the discharging potential range, respectively.

The areal capacitances of electrodes were calculated from the CVs according to the following equations:^25,26

Q = ∫idt (2)

where Q (C) is the electric quantity, i (A) is the current, t (s) is the time, ΔV (V) is the potential window, S (cm²) is the surface area of the working electrode and ν (V s⁻¹) is the scan rate.

2.4.2 Asymmetric electrochemical capacitor (AEC). An AEC was assembled based on the HSNC as positive electrode, rGO/CC as the negative electrode, and polymer filtering membrane as separator, respectively. The mass ratio of the electroactive materials loaded on positive electrode and negative electrode was decided according to the well-known charge balance theory (q₊ = q₋). Meanwhile, the charge stored on each electrode usually depends on the specific capacitance (C, F g⁻¹), the potential window (ΔV, V) and the mass of active material (m), as is shown in the eqn (4):^7,16,23,27

q = C × ΔV × m (4)

In order to obtain q₊ = q₋, the mass ratio will be expressed as the eqn (5):

C₊ and C₋ is the C, ΔV₊ and ΔV₋ is the potential range of the HSNC and rGO/CC electrodes, respectively. In the case of the two-electrode system, the expressions for specific capacitance (C, F g⁻¹), energy density (E, W h kg⁻¹) and power density (P, W kg⁻¹) are given as:where I (A g⁻¹), t (s) and ΔV (V) are the discharging specific current, discharging time and the cell voltage, respectively.

The AEC was measured with a two electrode configuration in 3.0 M KOH aqueous solution at room temperature.

3. Results and discussions

3.1 NiCo₂O₄ nanosheets/CC

3.1.1 Synthesis mechanism. Our synthetic strategy is briefly illustrated in Scheme 1. In the first step, the clean CC was electro-etched to improve the hydrophilicity because of the existence of oxygen functional groups and to reduce the diameter of the CC fiber. In the second step, the green bimetallic (Ni, Co) hydroxide precursors were co-electrodeposited on the electro-etched CC surface through reactions of (Ni²⁺, Co²⁺) ions with OH⁻ originated from the reduction of NO₃⁻ in cathodic.^6,22 The generation of OH⁻ ions at the cathode raises the local pH value, resulting in the uniform precipitation of mixed (Ni, Co) hydroxide on the CC surface because the solubility product constant (K_sp) at 25 °C of Ni(OH)₂ (2.8 × 10⁻¹⁶) is very close to Co(OH)₂ (2.5 × 10⁻¹⁶).²⁸ Then, NiCo₂(OH)₆ is thermally transformed into NiCo₂O₄ supported on the CC. The involved electrochemical reactions in whole process were given by the following reactions:^6,22,29

NO₃⁻ + H₂O + 8e⁻ = NH₄⁺ + 10OH⁻ (A)

Ni²⁺ + 2Co²⁺ + 6OH⁻ = NiCo₂(OH)₆ (B)

2NiCo₂(OH)₆ + O₂ = NiCo₂O₄ + 6H₂O (C)

Scheme 1 Schematic illustration of the formation of the samples.

3.1.2 Characterization of the samples. It is well known that hydrophilic surface is an indispensable factor for electrochemical capacitors electrode materials.^9,11 We measured the wetting properties of the varied materials (Fig. 1). It turned out that the contact angle for the untreated CC is about 137.3°, but for the electro-etched CC is nearly 0°, indicating the excellent wetting behavior after electro-etching, which is suitable for the growth of active materials. In addition, the contact angle for HSNC is approximate to 0°. The outstanding hydrophilicity is conducive to accessibility of the electrolyte to electroactive species and the effective solid–liquid contact during the electrochemical process.

Fig. 1 Contract angle measurement of (a) untreated CC, (b) electro-etched CC and (c) the HSNC composite.

The electro-etched CC sample and as-synthesized NiCo₂O₄/CC composites were characterized with SEM to investigate their microstructures. Fig. 2a and b show SEM images of the untreated CC and electro-etched CC (insets: the low magnifications), respectively. Compared with the pristine CC, the diameters of carbon fibers decrease from 10 to 6 μm after the electro-etching process, resulting in a wider distance between carbon fibers. This structure effectively facilitates the permeation of electrolyte ions. Fig. 2c–h show the images of samples for different electrodeposition times. When the electrodeposition time is 30 s (Fig. 2c), a certain amount of NiCo₂O₄ nanoparticles disorderly distribute on the surface of carbon fiber. With the increase of deposition time, NiCo₂O₄ nanoparticles as the building block gradually stack to form nanosheets through the “oriented attachment” growth process (Fig. 2d).^30,31 As time goes on 120 s, NiCo₂O₄ nanoparticles would spontaneously “land” on the as-formed sheets because of the minimization of facet free energy (Fig. 2e).^32,33 As the time is up to 180 s (Fig. 2f), the density NiCo₂O₄ nanosheets continues to increase and cross-links to form a honeycomb-shaped porous structure with many “electrolyte reservoirs”. The structure has several advantages in high-performance ECs. Firstly, the NiCo₂O₄ nanosheets are directly grown on carbon fibers via electrochemical deposition to sever as a seamless and integrated electrode, which facilitates the fast transport of electrons and decreases the resistance of electrode. Secondly, the “electrolyte reservoirs” can store electrolyte to provide sufficient ions for faradic reactions even at high specific current. Thirdly, the open pores formed through interconnected nanosheets could shorten the diffusion pathways of electrolyte ions. As the electrodeposition time is further increased to 240 s and 300 s (Fig. 2g and h), it can be found that the deposited layer become thicker and thicker, and the volume of “electrolyte reservoir” is decreasing with increasing time. In this case, the diffusion paths of electrolyte ions become narrow relatively and the exposed facets in the electrolyte decrease, which results in a low utilization of active species. HRTEM images of HSNC with the deposition time of 180 s is shown in Fig. 2i. Obviously, lattice spacings of 0.470 and 0.234 nm were observed, corresponding to the theoretical inter plane spacing of spinel NiCo₂O₄ (111) and (222) planes.^7,24 The selected-area electron diffraction (SAED) pattern (inset) shows well-defined diffraction rings, demonstrating their polycrystalline characteristics.

Fig. 2 FESEM images of (a) untreated CC, (b) electro-etched CC, (c–h) samples with different deposition times for 30 s, 60 s, 120 s, 180 s, 240 s and 300 s; insets: the high magnifications, and (i) HRTEM images, insets: SAED pattern of the HSNC.

Fig. 3a shows X-ray diffraction (XRD) patterns of the HSNC composite and pure CC. In XRD pattern of CC, the peaks at 26.2 and 43.0° are characteristic diffraction peaks of carbon materials. Compared with pure CC, the HSNC composite shows other diffraction peaks at 2θ values of 18.9, 31.2, 36.8, 38.4, 44.6, 55.4, 59.0, and 65.0°, which can be distinctly indexed to the (111), (220), (311), (222), (400), (422), (511), and (440) crystal planes of the spinel NiCo₂O₄ (JCPDS, no. 73-1702).^20,22 The elemental mapping of HSNC is shown in Fig. 3b. We can see that Ni, Co and O elementals are uniformly distributed on CC surface, indicating that NiCo₂O₄/CC nanosheets homogeneously coat the fibers of CC.

Fig. 3 (a) XRD pattern and (b) element mapping of the HSNC composites and (c, d, e and f) the XPS spectra of the HSNC composites, Co 2p, Ni 2p and O 1s, respectively.

The more detailed elemental composition and the oxidation state of metal elements in the HSNC composite was further investigated by XPS measurement, and the corresponding results are presented in Fig. 3c–f. The overall XPS spectrum of the HSNC reveals the presence of only C, Co, Ni and O elements in the HSNC composite (Figure 3c),³⁴ which is well consistent with the EDX analysis. By using a Gaussian fitting method, the Co 2p emission spectrum (Fig. 3d) was best fitted with two spin–orbit doublets, characteristic of Co²⁺ and Co³⁺, and two shakeup satellites (indicated as “Sat.”). The Ni 2p was also fitted with two spin–orbit doublets, characteristic of Ni²⁺ and Ni³⁺, and two shakeup satellites (Fig. 3e). Fig. 3f gives the high-resolution O 1s spectra. The O 1s spectrum can be fitted by means of following four peaks corresponding to different functional groups of oxygen atoms. (1) O²⁻ sitting at 529.2 eV is typical of metal–oxygen bonds (M–O). (2) Situate 530 eV is usually associated with oxygen in OH⁻ groups, and the presence of this contribution in the O 1s spectrum indicates that the surface of the NiCo₂O₄ is hydroxylated to some extent as a result of either surface oxyhydroxide or the substitution of oxygen atoms on the surface for hydroxyl groups. (3) The component corresponds to a higher number of defect sites with low oxygen coordination usually observed in materials with small particles. (4) The component H₂O can be attributed to multiplicity of physically and chemically absorbed water on or near the surface. These data explain that the surface of the NiCo₂O₄ nanosheets has a composition containing Co²⁺, Co³⁺, Ni²⁺, and Ni³⁺.^34,35 The chemical composition of the sample is further measured by using inductively coupled plasma (ICP) optical emission spectroscopy. According to the ICP analysis, the Co/Ni atomic ratio for the NiCo₂O₄ nanosheets sample is measured to be about 2.03 : 1, which is very close to 2 : 1. The result implies that the resultant material is NiCo₂O₄.

3.1.3 Electrochemical evaluation of the samples. In the electrochemical demonstration, the samples were directly used as the working electrodes to evaluation their electrochemical performances under a three-electrode configuration in 3 M KOH electrolyte. Fig. 4a presents CV curves of samples acquired in different deposition times at potential window of 0 to 0.6 V at a scan rate of 10 mV s⁻¹. For comparison, the pure CC was also tested. It is quite obvious that the CV integral area of pure CC is small enough so that the effective capacitive contribution of CC to the overall capacitance is negligible. Note that NiCo₂O₄ is generally regarded as a mixed valence oxide that adopts a spinel structure, in which all the nickel cations occupies the octahedral sites and the cobalt cations is distributed among the tetrahedral and octahedral interstices.^15,36 The electrochemical process in the alkaline electrolyte can be given by the following reactions:¹⁵

NiCo₂O₄ + OH⁻ + H₂O = NiOOH + 2CoOOH + e⁻ (D)

CoOOH + OH⁻ = CoO₂ + H₂O + e⁻ (E)

Fig. 4 (a) CV curves and (b) GCD curves of samples with different deposition times at a scan rate of 10 mV s⁻¹ and at a specific current of 1 A g⁻¹. And electrochemical properties of the HSNC (c) CV curves at different scan rates. (d) Variation of anodic and cathodic peak current with scan rate. (e) EIS measured of HSNC and CC at the open circuit potential in the frequency range from 0.1 to 10⁵ Hz (f) GCD curves at different specific currents. (g) Capacitance values versus specific current. (h) Cycling performance at 10 A g⁻¹ for 10 000 cycles.

Obviously, in the electrochemical process, the solid-state redox couples Co³⁺/Co⁴⁺ and M²⁺/M³⁺ (M = Co or Ni) are present in the structure, which can provide two kinds of redox active centers for releasing faradaic capacitance.^15,20 There is an oxidation peak because of an integration of two small peaks. The two reduction peaks correspond to the reversible reactions of Co³⁺/Co⁴⁺ and M²⁺/M³⁺ (M = Co or Ni) transitions (Fig. 4a), respectively. It is interesting that the position of oxidation peak barely shifts for the samples obtained under the deposition time from 30 s to 180 s. However, for the samples prepared under the deposition time more than 180 s, the position of oxidation peak shifts to a relatively high potential. Based on the above experimental phenomena, a reasonable interpretation may be that: when the deposition time is overlong, the volume of honeycomb, as an “electrolyte reservoir”, is relatively smaller that the electrolyte stored in it does not ensure steady supply of ions for redox reaction of NiCo₂O₄. On the other hand, the permeation of electrolyte ions from surface to bulk will encounter strong resistance for thick NiCo₂O₄ sheets, which can result in a polarization during the electrochemical process. These factors are negative for the capacitive behaviors of electrode materials.

Comparatively, the GCD curves of samples present in Fig. 4b. According to the eqn (A), the specific capacitances of samples electrodeposited for different time (30, 60, 120, 180, 240 and 300 s) are calculated to be 211, 387, 514, 614, 815, 514 and 400 F g⁻¹ at specific current of 1 A g⁻¹, respectively. More significantly, the specific capacitances are increasing first (from 30 to 180 s) and then decreasing (from 180 to 300 s). The above phenomenon may result from the following factors. The short depositing time leads to incomplete honeycomb-shaped structure of NiCo₂O₄. As described above, the overlong depositing time will result in abundance and thick NiCo₂O₄ sheets, which block the transporting channel of electrolyte ions and reduce the utilization of active species. So the optimal depositing time is 180 s, which is in consistent with the analysis of SEM images (mass loading: 1.0 mg cm⁻²).

Fig. 4c shows the CV curves of the HSNC composites with the deposition time of 180 s at scan rates ranging from 5 to 50 mV s⁻¹. When increasing the scan rate, the curve shapes nearly remain unchanged. Specifically, the peak potential shifts only ca. 80 mV for a 10-time increase in the scan rate, indicating low polarization and fast response under electrode processes. This is because the CC as robust scaffold will provide excellent electric conductivity for the electrode. According to eqn (B) and (C), the areal capacitances of HSNC are calculated to be 8100, 7830, 7670, 7540, 7330, and 6960 F m⁻² at scan rates ranging from 5 to 50 mV s⁻¹, respectively.

Furthermore, the relationship between the peak current (i_p) and the scan rate (ν) can provide a better understanding on the kinetic of the electrode material. According to the CV curves at different scan rates (Fig. 4c), it is found that the peak current is linearly dependant of the scan rate (ν) rather than ν^1/2 (Fig. 4d). The standard least squares regression coefficients (R²) are 0.995 and 0.997 for the oxidation and reduction peaks, respectively. That is to say, the redox process is not limited by concentration diffusion and also reveals fast kinetics for the HSNC electrode.^37,38 To the best of our knowledge, such a feature was not reported previously for NiCo₂O₄ electrode. This may be because NiCo₂O₄ nanosheets vertically growing on the CC fiber surface construct an “electrolyte reservoir”, which is beneficial to store electrolyte to ensuring steady supply of ions for redox reaction of NiCo₂O₄ even at high specific current. The Nyquist plots of the HSNC and CC are shown in Fig. 4e. The intercept at the real axis of the plot corresponding to the equivalent series resistances (R_L) are very small for CC (0.44 Ω) and HSNC (0.52 Ω), respectively. The slope of a straight line in the low frequency region can reflect the diffusive resistance resulting from the diffusion of electrolyte in the active species.^9,39 Herein, the HSNC and CC show almost a vertical line in low frequency region, indicating the small diffusive resistance of the electrolyte in electrode pores. The negligible diameter of the semicircle in the high-frequency range in the EIS represents the faradic charge transfer resistance (R_ct) at the interface between the electrode and electrolyte. Here, the R_ct of the HSNC is only about 0.57 Ω, which is much smaller than those reported for NiCo₂O₄ samples, such as the flower-like NiCo₂O₄ (16 Ω),⁴⁰ and NiCo₂O₄ nanosheets (14.3 Ω).⁴¹

To evaluate the application potential of the HSNC as an electrodes for ECs, GCD measurements were carried out in 3 M KOH electrolyte within a potential window of 0–0.55 V (vs. Hg/HgO) at various specific currents ranging from 1 to 20 A g⁻¹, as shown in Fig. 4f. The nearly symmetric potential–time curves at all specific currents imply the high charge–discharge coulombic efficiency and low polarization of the unique electrode. The Fig. 4g shows the dependent of specific capacitance calculated based on the charge–discharge curves (Fig. 4f) with specific current for the HSNC electrode. It is seen from Fig. 4g that about 84.4% of the initial capacitance is still retained when the charge–discharge rate increases from 1 to 20 A g⁻¹. The excellent rate capability originates that the vertical honeycomb-shaped porous channels ensure efficient accessibility of electrolyte to electroactive NiCo₂O₄ nanosheets. Furthermore, the seamless connection between NiCo₂O₄ nanosheets and CC substrate avoids the use of polymer binder and conductive additive, which reduce contact resistance.

A long cycle life, especially at high specific current, is the required feature for ESs. As shown in Fig. 4h, the HSNC electrode exhibits a remarkable cycling stability that the specific capacitance has only 5.7% decay after 10 000 cycles even at a specific current of 10 A g⁻¹. This result is much better than the one in Du's report (20% decay after 3000 cycles). The inset in Fig. 4h shows XRD pattern of the HSNC after GCD for 10 000 cycles. The intensity and position of diffraction peaks nearly unchanged, indicating that no new substance or phase state forms after long cycle. In other words, HSNC possesses complete reversibility. Table 1 shows the comparison between NiCo₂O₄ electrodes reported in the recent years. As seen, the HSNC electrode shows a substantially satisfactory electrochemical performance in term of excellent rate capability and long cycle lifetime.

Table 1 Electrochemical performance of the HSNC in this study, compared with some other NiCo₂O₄ electrodes reported in previous literature

Electrode	Load mass	Electrolyte	Specific capacitance	Retention rate	Cycle stability	Ref.
NiCo2O4/CC	0.6 mg	3 M KOH	2658 F g^−1 (2 A g^−1)	70.0% (20 A g^−1)	80% after 3000 cycles	22
NiCo2O4 flowerlike	—	6 M KOH	658 F g^−1 (1 A g^−1)	78.0% (20 A g^−1)	80% after 3000 cycles	5
NiCo2O4 nanowires	—	6 M KOH	743 F g^−1 (1 A g^−1)	78.6% (40 A g^−1)	93.8% after 3000 cycles	2
NiCo2O4 nanosheets	0.7 mg	2 M KOH	560 F g^−1 (2 A g^−1)	71.0% (40 A g^−1)	95.2% after 5000 cycles	42
NiCo2O4 microsphere	2.4 mg	6 M KOH	1006 F g^−1 (1 A g^−1)	72.2% (20 A g^−1)	93.2% after 1000 cycles	32
Ni@NiCo2O4	1.54 mg	6 M KOH	899 F g^−1 (1 A g^−1)	67.8% (20 A g^−1)	84.9% after 5000 cycles	43
Urchin-like NiCo2O4	5 mg	6 M KOH	296 F g^−1 (1 A g^−1)	72.6% (5 A g^−1)	100% after 1000 cycles	16
NiCo2O4 nanorods	1 mg	2 M KOH	758 F g^−1 (1 A g^−1)	47.1% (20 A g^−1)	80% after 1500 cycles	44
NiCo2O4/CC	1 mg	3 M KOH	815 F g^−1 (1 A g^−1)	84.3% (20 A g^−1)	94.3% after 10 000 cycles	This work

---

3.2 rGO/CC electrode

3.2.1 Morphology of rGO/CC. Fig. 5a shows the low-magnification SEM image of rGO/CC. It can be seen that rGO sheets uniform coverage on electro-etched CC fibers surface and interconnection between fibers and rGO sheets. The high-magnification image of rGO/CC is shown in Fig. 5b. The curly and flexible morphology of graphene sheets can be clearly observed. The rGO does not obviously stack on the surface of carbon fibers, which is positive for the penetration of electrolyte ions.

Fig. 5 (a) and (b) low-magnification and high-magnification SEM of rGO/CC.

3.2.2 Electrochemical properties of rGO/CC. Fig. 6a shows the CV curves of rGO/CC at various scan rates in a potential window of −1.0 to 0.0 V. As the scan rate increases from 5 to 50 mV s⁻¹, all the CV curves shapes nearly keep unchanged, indicating an excellent capacitive behavior and fast diffusion of electrolyte ions within the electrode material. It is noting that there is a couple of broad redox pecks, implying the existence of residual oxygen functional groups on rGO sheets and CC fiber surface. Fig. 6b shows the GCD curves of rGO/CC at various specific currents. From 1 A g⁻¹ to 10 A g⁻¹, all GCD curves show a linear shape and nearly symmetric potential–time curves, suggesting the well electrochemical reversibility for rGO/CC electrode. Based on the eqn (A), a specific capacitance of 207 F g⁻¹ can be achieved at 1 A g⁻¹.

Fig. 6 (a) CV curves at various scan rates and (b) GCD curves at different specific currents of the rGO/CC.

3.3 Electrochemical properties of as-fabricated AEC. To further evaluate the capacitive performance of the HSNC composite in a full cell setup. As shown Scheme 2, we used HSNC as the positive electrode, rGO/CC as the negative electrode and aqueous KOH solution as electrolyte to assemble an AEC. In order to make the HSNC//rGO/CC AEC reaches an optimal performance, the charge of the HSNC and rGO/CC should be at balance. According to eqn (D) and (E), the optimal mass ratio of active materials between the two electrodes in the AEC is expected to be m/m₊ = 2.2. On the basis of CV curves (Fig. 7a) of the HSNC (0–0.6 V) and rGO/CC (−1.0 to 0 V), the theoretical maximum operation voltage of the HSNC//rGO/CC AEC can reach 1.6 V. The CV curves at 10 mV s⁻¹ and GCD curves at 1 A g⁻¹ were measured with different cell voltage for the AEC, and the results are shown in Fig. 7b and c, respectively. The HSNC//rGO/CC AEC demonstrates an ideal capacitive behavior without obvious oxygen evolution curves, even at the cell voltage as large as 1.5 V. The enhancement of the cell voltage will be a critical factor to improve the energy density of the HSNC//rGO/CC AEC.

Scheme 2 Schematic of the HSNC//rGO/CC AEC.

Fig. 7 (a) CV curves of individual the HSNC and rGO/CC electrodes at 10 mV s⁻¹. (b) CVs of the AEC device collected at different potential windows at 10 mV s⁻¹; (c) GCD curves of the AEC device collected at different potential windows at a fixed specific current of 1 A g⁻¹.

Fig. 8a shows the CV curves of the HSNC//rGO/CC AEC at various scan rates with a cell voltage of 1.5 V. It is obvious that the broad oxidation peaks and reduction peaks emerge on the CV curves and the peak current increases with the scan rate increasing from 5 to 50 mV s⁻¹. However, there is no obvious distortion in the CV curves even at a high scan rate of 50 mV s⁻¹, indicating a rapid I–V response and fast charge–discharge properties of the device. The GCD curves at different specific currents from 1 to 50 A g⁻¹ are shown in Fig. 8b and c. All the discharge curves are nearly linear and symmetric with the charge counterparts, suggesting an outstanding electrochemical reversibility. According to eqn (6), the AEC shows a high specific capacitance 103.6 F g⁻¹ at 1 A g⁻¹. Furthermore, the HSNC//rGO/CC AEC exhibits an excellent rate capability when the specific current increases from 1 to 50 A g⁻¹ as shown in Fig. 8d, which is comparable or superior for asymmetric supercapacitors (ASCs) in the literature. In addition, the HSNC//rGO/CC AEC shows an outstanding cycle stability at 3 A g⁻¹, as shown in Fig. 8e (76.5% of the initial C after 10 000 cycles).

Fig. 8 (a–c) CV and GCD curves of the AEC at various scan rates and different specific currents in 3 M KOH. (d) Specific capacitance of the AEC as a function of specific current. (e) Cycling performance of the AEC measured at 3 A g⁻¹ for 10 000 cycles. (f) Ragone plot of the HSNC//rGO/CC AEC.

The power density (P) and energy density (E) are two important parameters that characterize the performance of energy storage devices. The GCD curves can be applied to evaluate the power and energy densities of the HSNC//rGO/CC AEC. With increasing specific current from 1 to 50 A g⁻¹, the Ragone plot of the AEC is shown in Fig. 8f. The E reaches 32.4 W h kg⁻¹ along with P of 0.75 kW kg⁻¹. When its power density increases by about 50 times compared with the initial value, that is to say, it is equal to 37.7 kW kg⁻¹ (a full charge–discharge within 3.5 s), energy density of the device still hold at 17.8 W h kg⁻¹, indicating an outstanding rate capability. It outperforms of some ASCs in literature such as NiCo₂O₄@MnO₂//AC (35 W h kg⁻¹ at 0.163 kW kg⁻¹),⁴⁵ NiCo₂S₄//AC (25.5 W h kg⁻¹ at 0.334 kW kg⁻¹),²⁷ NiCo₂O₄-NSs@HMRAS//AC (15.42 W h kg⁻¹ at 7.8 kW kg⁻¹),²⁹ NiCo₂O₄-rGO//AC (23.3 W h kg⁻¹ at 0.325 kW kg⁻¹),⁴⁶ ZnCo₂O₄//AC (16.63 W h kg⁻¹ at 2.561 kW kg⁻¹).⁴⁷ 3D-ICHA α-Ni(OH)₂//AC (14.9 W h kg⁻¹ at 0.14 kW kg⁻¹).⁴⁸

Based on the above overall experimental results, the outstanding electrochemical performances of the HSNC//rGO/CC AEC may be due to the following factors. Firstly, three-dimensional porous CC has high conductivity. Secondly, the active materials in positive and negative electrode bands together with CC current collector without conductive polymers, which minimizes the contact resistance so that electrochemical processes of the active species become easy. Thirdly, the honeycomb architecture with like-electrolyte reservoir shortens diffusion path of electrolyte ions and facilitates penetration of electrolyte form the surface to bulk. Lastly, the NiCo₂O₄ tightly attached on the surface of CC withstand the structure collapse upon long cycle.

4. Conclusion

In summary, we used two-step strategy to fabricate the honeycomb-shaped porous NiCo₂O₄ on electro-etched CC with strong adhesion. The HSNC as a bind-free electrode display fast kinetics and electrochemical behavior controlled by the surface reaction. The unique porous structure could provide rapid transport pathways for electrolyte ions, which is responsible for excellent rate capability of the HSNC electrode. The integration of the HSNC and rGO/CC into an AEC device is able to generate combination of comparable energy density with ultrahigh power density and outstanding structural stability. Importantly, the energy density of the device still hold at 17.8 W h kg⁻¹ when its power density is up to 37.7 kW kg⁻¹ (a full charge–discharge within 3.5 s), indicating an excellent rate capability.

Acknowledgements

The authors gratefully acknowledge the financial support offered by the National Natural Science Foundation of China (20963009, 21163017 and 21563027) and Specialized Research Fund for the Doctoral Program of Higher Education (No. 20126203110001).

References

1. S. Khalid, C. B. Cao, A. Ahmad, L. Wang, M. Tanveer, I. Aslam, M. Tahir, F. Idrees and Y. Q. Zhu, RSC Adv., 2015, 5, 33146–33154.
2. H. Jiang, J. Ma and C. Z. Li, Chem. Commun., 2012, 48, 4465–4467.
3. G. P. Wang, L. Zhang and J. J. Zhang, Chem. Soc. Rev., 2012, 41, 797–828.
4. B. Dyatkin, V. Presser, M. Heon, M. R. Lukatskaya, M. Beidaghi and Y. Gogotsi, ChemSusChem, 2013, 6, 2269–2280.
5. H. C. Chen, J. J. Jiang, L. Zhang, T. Qi, D. D. Xia and H. Z. Wan, J. Power Sources, 2014, 248, 28–36.
6. C. Z. Yuan, L. Yang, L. R. Hou, L. F. Shen, X. G. Zhang and X. W. Lou, Energy Environ. Sci., 2012, 5, 7883–7887.
7. C. Z. Yuan, J. Y. Li, L. R. Hou, X. G. Zhang, L. F. Shen and X. W. Lou, Adv. Funct. Mater., 2012, 22, 4592–4597.
8. Y. Y. Yang, Y. R. Liang, Y. D. Zhang, Z. Y. Zhang, Z. M. Li and Z. A. Hu, New J. Chem., 2015, 39, 4035–4040.
9. L. Li, Z. A. Hu, N. An, Y. Y. Yang, Z. M. Li and H. Y. Wu, J. Phys. Chem. C, 2014, 118, 22865–22872.
10. Y. S. Luo, D. Z. Kong, J. S. Luo, S. Chen, D. Y. Zhang, K. W. Qiu, X. Y. Qi, H. Zhang, C. M. Li and T. Yu, RSC Adv., 2013, 3, 14413–14422.
11. Y. D. Zhang, Z. A. Hu, Y. R. Liang, Y. Y. Yang, N. An, Z. M. Li and H. Y. Wu, J. Mater. Chem. A, 2015, 3, 15057–15067.
12. V. Augustyn, P. Simon and B. Dunn, Energy Environ. Sci., 2014, 7, 1597–1614.
13. P. Simon, Y. Gogotsi and B. Dunn, Science, 2014, 14, 1210–1211.
14. W. B. Yan, J. Y. Kim, W. D. Xing, K. C. Donavan, T. Ayvazian and R. M. Penner, Chem. Mater., 2012, 24, 2382–2390.
15. Z. B. Wu, Y. R. Zhu and X. B. Ji, J. Mater. Chem. A, 2014, 2, 14759–14772.
16. T. Wu, J. Y. Li, L. R. Hou, C. Z. Yuan, L. Yang and X. G. Zhang, Electrochim. Acta, 2012, 81, 172–178.
17. F. X. Ma, L. Yu, C. Y. Xu and X. W. Lou, Energy Environ. Sci., 2016, 9, 862–866.
18. L. Yu, B. Y. Guan, W. Xiao and X. W. Lou, Adv. Energy Mater., 2015, 5, 1500981.
19. H. Hu, B. Y. Guan, B. Y. Xia and X. W. Lou, J. Am. Chem. Soc., 2015, 137, 5590–5595.
20. X. Y. Liu, Y. Q. Zhang, X. H. Xia, S. J. Shi, Y. Lu, X. L. Wang, C. D. Gu and J. P. Tu, J. Power Sources, 2013, 239, 157–163.
21. L. F. Shen, Q. Che, H. S. Li and X. G. Zhang, Adv. Funct. Mater., 2014, 24, 2630–2637.
22. J. Du, G. Zhou, H. M. Zhang, C. Cheng, J. M. Ma, W. F. Wei, L. B. Chen and T. H. Wang, ACS Appl. Mater. Interfaces, 2013, 5, 7405–7409.
23. J. Yan, Q. Wang, T. Wei and Z. J. Fan, Adv. Energy Mater., 2014, 4, 1300816.
24. B. W. S. Hummers Jr and R. E. Offeman, J. Am. Chem. Soc., 1958, 80, 1339.
25. X. H. Lu, G. M. Wang, T. Zhai, M. H. Yu, J. Y. Gan, Y. X. Tong and Y. Li, Nano Lett., 2012, 12, 1690–1696.
26. Y. D. Zhang, Z. A. Hu, Y. F. An, B. S. Guo, N. An, Y. R. Liang and H. Y. Wu, J. Power Sources, 2016, 311, 121–129.
27. Z. B. Wu, X. L. Pu, X. B. Ji, Y. R. Zhu, M. J. Jing, Q. Y. Chen and F. P. Jiao, Electrochim. Acta, 2015, 174, 238–245.
28. J. W. Xiao and S. H. Yang, RSC Adv., 2011, 1, 588–595.
29. X. F. Lu, D. J. Wu, R. Z. Li, Q. Li, S. H. Ye, Y. X. Tong and G. R. Li, J. Mater. Chem. A, 2014, 2, 4706–4713.
30. J. M. R. Herrera, M. Terrones, H. Terrones, S. Dag and V. Meunier, Nano Lett., 2007, 7, 570–576.
31. Y. Liu, Y. Jiao, Z. L. Zhang, F. Y. Qu, A. Umar and X. Wu, ACS Appl. Mater. Interfaces, 2014, 6, 2174–2184.
32. Y. Lei, J. Li, Y. Y Wang, L. Gu, Y. F. Chang, H. Y. Yuan and D. Xiao, ACS Appl. Mater. Interfaces, 2014, 6, 1773–1780.
33. Y. N. Xia, Y. J. Xiong, B. K. Lim and S. E. Skrabalak, Angew. Chem., Int. Ed., 2009, 48, 60–103.
34. Y. Lei, Y. Y. Wang, W. Yang, H. Y. Yuan and D. Xiao, RSC Adv., 2015, 5, 7575–7583.
35. R. J. Zou, K. B. Xu, T. Wang, G. J. He, Q. Liu, X. J. Liu, Z. Y. Zhang and J. Q. Hu, J. Mater. Chem. A, 2013, 1, 8560–8566.
36. H. W. Wang, Z. A. Hu, Y. Q. Chang, Y. L. Chen, H. Y. Wu, Z. Y. Zhang and Y. Y. Yang, J. Mater. Chem., 2011, 21, 10504–10511.
37. H. W. Wang, H. Yi, C. R. Zhu, X. F. Wang and H. J. Fan, Nano Energy, 2015, 13, 658–669.
38. N. An, Y. F. An, Z. A. Hu, Y. D. Zhang, Y. Y. Yang and Z. Q. Lei, RSC Adv., 2015, 5, 63624–63633.
39. S. L. Zhang and N. Pan, Adv. Energy Mater., 2015, 5, 1401401.
40. L. L. Li, Y. Cheah, Y. Ko, P. Teh, G. Wee, C. L. Wong, S. J. Peng and M. Srinivasan, J. Mater. Chem. A, 2013, 1, 10935–10941.
41. X. H. Lu, X. Huang, S. L. Xie, T. Zhai, C. S. Wang, P. Zhang, M. H. Yu, W. Li, C. L. Liang and Y. X. Tong, J. Mater. Chem., 2012, 22, 13357–13364.
42. A. K. Mondal, D. Su, S. Q. Chen, K. Kretschmer, X. Q. Xie, H. J. Ahn and G. X. Wang, ChemPhysChem, 2015, 16, 169–175.
43. G. X. Gao, H. B. Wu, S. J. Ding, L. M. Liu and X. W. Lou, Small, 2015, 11, 804–808.
44. E. Jokar, A. I. Zad and S. Shahrokhian, J. Solid State Electrochem., 2015, 19, 269–274.
45. K. B. Xu, W. Y. Li, Q. Liu, B. Li, X. J. Liu, L. An, Z. G. Chen, R. J. Zou and J. Q. Hu, J. Mater. Chem. A, 2014, 2, 4795–4802.
46. X. Wang, W. S. Liu, X. H. Lu and P. S. Lee, J. Mater. Chem., 2012, 22, 23114–23119.
47. B. K. Guan, D. Guo, L. L. Hu, G. H. Zhang, T. Fu, W. J. Ren, J. D. Li and Q. H. Li, J. Mater. Chem. A, 2014, 2, 16116–16123.
48. C. D. Gu, X. Ge, X. L. Wang and J. P. Tu, J. Mater. Chem. A, 2015, 3, 14228–14238.
49. R. R. Salunkhe, B. P. Bastakoti, C. T. Hsu, N. Suzuki, J. H. Kim, S. X. Dou, C. C. Hu and Y. Yamauchi, Chem.–Eur. J., 2014, 20, 3084–3088.
50. S. E. Moosavifard, M. F. E. Kady, M. S. Rahmanifar, R. B. Kaner and M. F. Mousavi, Appl. Mater. Interfaces, 2015, 7, 4851–4860.

---