Introduction

In recent years, energy consumption and the fossil fuels crisis are forecast to have serious impacts on the world economy and ecology. Supercapacitors (also named electrochemical capacitors or ultracapacitors) have attracted much attention due to their fast energy delivery, short charging time, high power capability, long life cycle (>105 cycles), and environment-friendly features [1–3]. In general, supercapacitors can be divided into double-layer capacitors (EDLCs) and pseudocapacitors by their charge-storage mechanisms [4–6]. EDLCs using carbon-based materials usually have high power density but suffer from low capacitance, which charge is stored by rapid adsorption/desorption of electrolyte ions on high-surface-area carbon materials [7–9]. In contrast, pseudocapacitors, employing transition metal oxides/hydroxides, or conducting polymers can provide high specific capacitance, which charge is stored and released in Faradaic electron-transfer processes of a metal oxide/hydroxide or conducting polymer [10, 11]. Although metal oxides/hydroxides offer high specific capacitance, they deliver low power density, bad cycling stability, and poor rate capability (dramatic drop of specific capacitance with an increase at a high scan rate) because of poor electrical conductivity [12, 13]. As a result, the composites of metal oxides/hydroxides and carbon-based materials are greatly studied in order to get superior performance.

Graphene, a two-dimensional monolayer of carbon atoms packed into a honeycomb lattice, has drawn our attention because of its large surface area, high electrical conductivity, superior mechanical properties, and good electrochemical stability [14–16]. So, the utilization of graphene to prepare electrodes for supercapacitors becomes a focus [8, 17–19]. But the formation of irreversible aggregates of graphene during its preparation leads to very low EDLC capacitance from graphene-based supercapacitors. A strategy is to grow and anchor metal oxides/hydroxides onto graphene to give rise to composites, where graphene has high electrical conductivity to enhance electronic transfer and produce EDLC capacitance, and metal oxides/hydroxides mainly contribute higher capacitance to the whole capacitor and act as spacer to prevent the aggregation of graphene. Recently, based on this strategy, lots of graphene-based composites including graphene/Co3O4 [20], graphene/Ni(OH)2 [21], graphene/SnO2 [22], graphene/MnO x [23], graphene/ZnO [24], graphene/Fe3O4 [25], and so forth have been reported as electrode materials for supercapacitors and show better electrochemical performances in comparison with single component [17, 19]. Generally, Co3O4, which has very high theoretical pseudocapacitance of 3560 F/g [26, 27], and graphene is integrated into composites for supercapacitor through many expatiatory or high-cost methods, such as a two-step surfactant assisted method or by an in situ solution-based method under reflux condition [28, 29].

In this paper, we employed a facile one-step hydrothermal treatment of graphene oxide (GO) and cobalt acetate (Co(Ac)2) to synthesize reduced GO (rGO)/Co3O4 composites for supercapacitors. Co3O4 nanoparticle homogeneously coated onto rGO sheets due to the electrostatic interaction between Co2+ cations and GO sheets before the hydrothermal reaction [30]. The influences of mass ratios of GO and Co(Ac)2 on the performances of supercapacitors were investigated.

Experimental

Materials

Potassium permanganate (KMnO4), polytetrafluoroethylene (PTFE) aqueous solution (60 wt%), sulfuric acid (H2SO4), phosphoric acid (H3PO4), hydrochloric acid (HCl), graphite, hydrogen peroxide (H2O2, 30 wt%), and ammonia solution (28 wt%) were from Sinopharm Chemical Reagent Co., Ltd. Cobalt acetate (Co(Ac)2), and alcohol (95 wt%) were purchased from Shanghai Aladdin Reagent. All materials were commercially available and employed without further purification.

Preparation of graphite oxide

Graphite oxide was prepared from graphite by an improved Hummers’ method [31]. Typically, flake graphite (1 g) and KMnO4 (6 g) were slowly added into 300 mL of concentrated H2SO4 and H3PO4 with volume ratio of 9:1 at room temperature. The suspension was stirred at room temperature for 30 min and then at 50 °C for 12 h. After the suspension was cooled down to room temperature, it was poured into 200 mL ice solution containing a certain amount of 30 wt% H2O2. The obtained suspension was separated using a centrifuge at the speed of 4800 rpm, and the remaining slurry was washed with 10 % HCl solution three times and deionized (DI) water three times. The collected sample was dried at 50 °C under vacuum condition to obtain solid state graphite oxide.

Preparation of rGO/Co3O4 composite

In a typical synthesis of the rGO/Co3O4 composite, firstly, 70 mg of graphite oxide was dispersed into 70 mL DI water by ultrasonication for 2 h to form uniform exfoliated GO solution. Secondly, the required amount of Co(Ac)2 was added into the GO solution. Thirdly, the pH value of solution was adjusted to 9.5 by ammonia solution. Finally, the reaction mixture was transferred to a Teflon-lined autoclave and hydrothermally treated at 180 °C for 15 h. The solid product was obtained by filtering and washing with water and ethanol for several times. Pure Co3O4 was also prepared by hydrothermal method under the same conditions in the absence of GO. In our experiments, the mass ratios of GO and Co(Ac)2 (r G/Co) were controlled as 0:1, 1:3, 1:2, and 1:1. And the as-prepared composites were named as Co3O4, and rGO/Co3O4-3, 2, and 1, respectively.

Preparation of electrode

The electrode material was composed of the active material, graphite and PTFE aqueous solution with the weight ratio of 85:10:5 [32]. At room temperature, alcohol was added to the electrode material and stirred to form homogeneous slurry. The viscous slurry was pressed by Decal method to form a thin sheet. Then, the thin sheet was pressed onto a nickel net. After being dried at 60 °C for 12 h, the working electrode was obtained.

Characterizations

Powder X-ray diffraction (XRD) analyses were performed on an X-ray diffractometer (D8-advance, Bruker, Germany) with a Cu Kα source. Raman spectra were obtained using a Raman spectrometer (Renishaw inVia, Germany) under a laser light at 532 nm. The morphologies and sizes of samples were characterized by using field emission scanning electron microscopy (FESEM, HITACHI S-4800, Japan) and transmission electron microscopy (TEM, HITACHI H-7650, Japan). Brunauer–Emmett–Teller specific surface areas were determined by N2 adsorption–desorption isotherms at 77 K on a 3H-2000PS1 specific surface and pore size analysis instrument (Beijing BeiShiDe Sci. and Tech. Co., Ltd.). The total pore volume was derived from the adsorbed amount at a relative pressure of P/P 0 = 0.988. Pore size distributions were estimated by Barrett–Joyner–Halanda method.

Two-electrode supercapacitors containing electrolytes of 2 M KOH aqueous solution were fabricated and tested on an electrochemical workstation system (CHI 660C, Shanghai Chen Hua Co., Ltd., China). The cyclic voltammetry (CV) measurements were done at a potential window of 0–0.8 V at various scan rates from 5 to 100 mV/s. The galvanostatic charge–discharge (GCD) tests were conducted between voltage of 0–0.8 V at different constant current densities of 0.2, 0.7, 1.2, and 2.0 A/g. The electrochemical impedance spectroscopy measurements were performed with the frequency ranging from 10 mHz to 100 kHz. The supercapacitor specific capacitance (C, F/g) and electrode specific capacitance (C s, F/g) were calculated from GCD curves according to the following equations [33]:

$$ C = \frac{I \times \Updelta t}{{\Updelta V \times m_{\text{ac}} }}, $$
(1)
$$ C_{\text{s}} = 4 \times C. $$
(2)

In addition, equivalent series resistance (ESR, Ω) was evaluated from the following equation [34]:

$$ {{ESR}} = \frac{{{{iR}}_{\text{drop}} }}{2 \times I}. $$
(3)

In these equations, I (A) is the discharge current, Δt (s) is the discharge time, ΔV (V) is the voltage change after a full charge or discharge, iR drop (V) is defined as the electrical potential difference between the two ends of a conducting phase during charging–discharging, and m ac (g) is mass of the electroactive material.

Results and discussion

Characterizations of electrode materials

The phase structures of as-synthesized samples were determined by XRD. Figure 1 shows XRD patterns of graphite oxide, rGO/Co3O4-2 composite and Co3O4. As shown in Fig. 1a, oxidation of graphite through an improved Hummers’ method makes the (002) diffraction peak move to about 9.2° with an interlayer distance of 0.96 nm, which is similar to that in many other graphite oxides, resulting from abundant oxygen functional groups locating at graphene sheets [31, 35–37]. The graphite oxide was exfoliated by the assistance of ultrasonication to form highly stable GO solution. Co(Ac)2 was added into the GO solution and then hydrothermally treated at 180 °C in the presence of ammonia. In Fig. 1b, the presence of weak broad (002) peak at about 25.5° displays that most oxygen functional groups bonded to graphene sheets have been eliminated, and the other peaks at 18.8, 31.0, 36.8, 38.7, 45.9, 59.4, and 65.2° fully correspond to the reflections from (111), (220), (311), (222), (400), (511), and (440) crystal planes of well-crystallized Co3O4 with a face-centered cubic structure (JCPDS Card No. 43-1003), evidently implying that rGO/Co3O4 composite can be prepared by facile one-step hydrothermal method. Figure 1c shows that pure Co3O4 can be prepared in the absence of GO under the same condition.

Fig. 1
figure 1

XRD patterns of (a) graphite oxide, (b) rGO/Co3O4-2 composite, and (c) Co3O4

Full size image

Raman spectroscopy is a suitable technique to characterize carbonaceous materials, particularly for distinguishing their ordered and disordered crystal structures [20]. In their Raman spectra, D band associating with the defects of the material and G band corresponding to the first-order scattering of the E2g mode that is related to the order of the material appear at around 1350 and 1580 cm−1, respectively [38]. The ratio between the intensity of D and G bands has been widely used to evaluate the quality of graphene materials coarsely [31]. Figure 2 shows the Raman spectra of rGO/Co3O4-2 composite and graphite oxide. In both spectra, there are D peak (~1350 cm−1) and G peak (~1590 cm−1). However, the I D/I G ratio of rGO/Co3O4 composite is smaller than that of GO, indicating that the backbone of graphene has been recovered to some extent. Three additional peaks at 470, 522, and 670 cm−1 for rGO/Co3O4-2 composite are ascribed to the Eg, F2g, and A1g modes in Co3O4, respectively [39]. These results further demonstrate the existence of both rGO and Co3O4 in the as-prepared composite.

Fig. 2
figure 2

Raman spectra of (a) rGO/Co3O4-2 composite and (b) graphite oxide

Full size image

SEM and TEM images of rGO/Co3O4-2 composite are shown in Fig. 3. The SEM image shows that rGO sheets are fully covered by Co3O4 particles in situ produced during the hydrothermal treatment. According to the TEM image, it is observed that 10–30 nm-sized Co3O4 nanoparticles are well dispersed on the rGO sheet surface despite the long-time treatment of sample by ultrasonication for the TEM characterization.

Fig. 3
figure 3

a SEM and b TEM images of rGO/Co3O4-2 composite

Full size image

Figure 4 shows the N2 adsorption–desorption isotherms of Co3O4 and different rGO/Co3O4 composites. All samples are found to have typical IUPAC type IV sorption behavior with hysteresis loops, indicating the existence of mesoporous structure, which is further demonstrated by the pore size distributions of all samples (Fig. 5). As shown in Table 1, the specific surface area increases from 42.65 to 124.89 m2/g with the increment of r G/Co from 0 to 1:2, resulting from that the rGO not only acts as support for deposition of Co3O4 nanoparticles but also prevents Co3O4 from aggregating. However, when the r G/Co further increases to 1:1, the specific surface area decreases to 99.23 m2/g, maybe due to the formation of more rGO aggregates.

Fig. 4
figure 4

N2 adsorption–desorption isotherms of Co3O4 and different rGO/Co3O4 composites

Full size image
Fig. 5
figure 5

Pore size distributions of Co3O4 and different rGO/Co3O4 composites

Full size image
Table 1 Pore structure parameters of Co3O4 and different rGO/Co3O4 composites
Full size table

Electrochemical tests for supercapacitors

Figure 6 shows the GCD behaviors of two-electrode supercapacitors based on Co3O4 and different rGO/Co3O4 composites tested from 0 to 0.8 V at a constant current density of 0.2 A/g. All curves exhibit a good symmetric shape, implying an ideal capacitor character. And it is obvious that the charge–discharge duration increases with increasing r G/Co from 0 to 1:2. When the r G/Co further rise to 1:1, the charge–discharge duration decreases. On the other hand, by GCD tests, the capacitances of the electrodes can be evaluated according to Eqs. (1) and (2). The C s of Co3O4, and rGO/Co3O4-3, 2 and 1 are 22.6, 193.8, 263.0, and 144.3 F/g, respectively. The change trend of C s is in accordance with that of charge–discharge duration. Besides, by GCD tests, the ESR can be obtained according to Eq. (3). Their ESR values are 27.45, 10.01, 8.65, and 13.02 Ω, respectively. These results show that the rGO can reduce the ESRs of supercapacitors evidently, resulting in the enhancement of C s. But when the GO content is too high, the C s decreases, mainly resulting from production of lower pseudocapacitance from Co3O4.

Fig. 6
figure 6

GCD curves of supercapacitors based on Co3O4 and different rGO/Co3O4 composites at a constant current density of 0.2 A/g

Full size image

Figure 7 shows CV curves of supercapacitors based on Co3O4 and different rGO/Co3O4 composites under the potential from 0 to 0.8 V at a scan rate of 20 mV/s. The CV curves are close to rectangular shape and there is no visible redox peak along the current–potential axis, illustrating that all samples have an ideal capacitive behavior [21, 40]. Obviously, the rGO/Co3O4-2 composite has a higher electrode specific capacitance than Co3O4 and other rGO/Co3O4 composites at the same scan rate, which is same as the result from GCD tests. Figure 8a shows the CV loops for the supercapacitor based on rGO/Co3O4-2 composite at various scan rates from 5 to 100 mV/s. It is observed that the rectangular area obviously increases and there was no obvious distortion in the CV loops with the increment of the scan rate, meaning a desirable rate capability and good capacitive behavior for the rGO/Co3O4-2-based supercapacitor [41]. Meanwhile, as is known that current density directly affects the capacitive behavior of supercapacitor. Figure 8b shows the GCD curves of the two-electrode supercapacitors based on rGO/Co3O4-2 composite tested from 0 to 0.8 V at different current densities of 0.2, 0.7, 1.2, and 2.0 A/g. The charging curves are highly linear and symmetric with their corresponding discharging counter parts at various current densities from 0.2 to 2 A/g [42]. This suggests that the supercapacitor based on rGO/Co3O4-2 composite has good electrochemical reversibility and charge–discharge performances [43]. And the C s calculated from these GCD curves tested at current densities of 0.2, 0.7, 1.2, and 2 A/g are 263.0, 245.6, 233.3, and 229.0 F/g, respectively. It can be found that the C s obtained at a high current density of 2.0 A/g retains 87.07 % of the value obtained at a low current density of 0.2 A/g, testifying the high rate capability of supercapacitor based on rGO/Co3O4-2 composite.

Fig. 7
figure 7

CV curves of supercapacitors based on Co3O4 and different rGO/Co3O4 composites at a scan rate of 20 mV/s

Full size image
Fig. 8
figure 8

a CV curves of supercapacitors based on rGO/Co3O4-2 composites at different scan rates from 5 to 100 mV/s, and b GCD curves of supercapacitors based on rGO/Co3O4-2 composite at different current densities of 0.2, 0.7, 1.2, and 2.0 A/g

Full size image

Figure 9 shows the Nyquist plots for the supercapacitors based on Co3O4 and different rGO/Co3O4 composites. Their impedances were measured in the frequency range of 100 kHz–10 mHz at open circuit potential with an AC perturbation of 5 mV. For supercapacitor, its ideal electrochemical capacitance behavior exhibits a small depressed semicircle at higher frequency and the imaginary part of the impedance in the low frequency region being perpendicular to the real part [44]. The electrode inter resistance (R s) is calculated from the high frequency intersection of the Nyquist plot in the x axis [20]. It is observed that pure Co3O4 and rGO/Co3O4 composites have the similar inter-resistance of about 1.33 Ω. The charge transfer resistance (R ct) is counted from the span of the single semi-circle along the x axis from high to low frequency region [20]. No semi-arc is found in the impedance spectrum for pure Co3O4 electrode. In contrast, a small semi-arc is observed for the rGO/Co3O4 composite electrodes. The charge transfer resistances (R ct) of rGO/Co3O4-3, 2, and 1 composite electrodes were estimated to be 14.2, 10.7, and 21.8 Ω, respectively. Clearly, the performances of the supercapacitor are heightened obversely after compositing the pure Co3O4 and rGO.

Fig. 9
figure 9

Nyquist plots of supercapacitors based on Co3O4 and different rGO/Co3O4 composites

Full size image

It is a crucial problem to solve the long cyclic durability for supercapacitor. Figure 10 depicts the cyclic performances of the supercapacitors based on Co3O4 and different rGO/Co3O4 composite examined by GCD tests under a current density of 2 A/g for 1000 cycles. All supercapacitors exhibit a little capacitance decay during the tests and the introduction of rGO can improve the cyclic durability of supercapacitor. For supercapacitor based on rGO/Co3O4-2 composite, the supercapacitor presents the best cyclic durability and the capacitance retains 92.09 % of the original values after 1000 cycles, suggesting that the supercapacitor has long cyclic durability.

Fig. 10
figure 10

Cyclic performances of supercapacitors based on Co3O4 and different rGO/Co3O4 composites at a current density of 2 A/g

Full size image

Based on the results of structural characterizations and electrochemical tests of electrode materials, rGO/Co3O4-2 composite exhibiting the best electrochemical performances is considered to be attributed to the following reasons. Firstly, the rGO not only increase the electrical conductivity of electrode, but also act as support for deposition of Co3O4 nanoparticles to increase the effective interfacial area between Co3O4 and the electrolyte. Secondly, the nanoscale size of Co3O4 particles can greatly reduce the diffusion length over which K+ must transfer during the charge/discharge process, leading to the improvement of the electrochemical utilization of Co3O4. Finally, when the mass ratio of GO and Co(Ac)2 is 1:2, the obtained rGO/Co3O4-2 composite presents the highest specific surface area and lowest charge transfer resistance.

Conclusions

This work has demonstrated that rGO/Co3O4 composites as electrodes for supercapacitors were prepared by a facile one-step hydrothermal treatment of GO and Co(Ac)2. 10–30 nm-sized Co3O4 nanoparticles are well dispersed onto the rGO sheet. In comparison with pure Co3O4-based supercapacitor, the introduction of rGO can increase the electrical conductivities and the specific surface areas of electrode materials, leading to the enhancement of their electrochemical performances. When the mass ratio of GO and Co(Ac)2 is 1:2, rGO–Co3O4 composite electrode exhibits the highest capacitance of 263.0 F/g. C s obtained at a high current density of 2.0 A/g retains 87.07 % of the value obtained at a low current density of 0.2 A/g, indicating that the supercapacitor has high rate capability. 92.09 % of initial capacitance is retained after 1000 charge/discharge cycles test under a current density of 2 A/g, showing that the supercapacitor exhibits long cyclic durability.