Introduction

Considering the graveness of the environment issues and the shortage of energy source, lithium-ion batteries (LIBs) get much more attention with its large specific capacity, environmental friendliness, and long cycle life [1]. However, the capacity of the traditional graphite (~ 370 mA h g−1) in theory is too low to meet the market demands. ZnO as one of the transition metal oxides is widely utilized in many fields, such as organic light [2], gas sensors [3], and smartphone [4]. Especially, ZnO as a promising anode material can be utilized in LIBs due to the high theoretical capacity (~ 978 mA h g−1) and good electrical conductivity. Nevertheless, the connatural high capacity of the ZnO is limited by a big volume change (the ZnO nanoparticles undergo 103% volume expansion after lithiation [5]) and the aggregation of the formed ZnO during the long cycles. Aimed at the above issues, some promising approaches are used to improve ZnO anode materials for LIBs, including the thermal decomposition of C4H4ZnO6 to ZnO@C [6], carbon/ZnO nanorod [7], ZnO-loaded/porous carbon [8], and ZnCo2O4-ZnO nanorod arrays [9].

On the other hand, compared with the present Li-ion technologies, Na-ion batteries are becoming the promising candidates to meet the larger energy market demands, since the sodium materials have many practical advantages, such as low-cost and natural abundance [10]. However, lots of materials that can be well represented in the LIBs have bad performance in the Na-ion battery mainly due to the larger radius of the Na+ than Li+. For example, graphite, as one of the frequently used electrode materials in LIBs, displays very poor capacity performance in the sodium-ion battery system [11]. In order to find the suitable electrode materials of the sodium-ion battery, it is believed that the bigger lattice constants and the open architectures of the as-prepared materials, such as the bowl-like hollow Co3O4 [12], porous CoFe2O4 nanocubes [13], porous CuO [14], and hierarchical durian-like NiS2 [15] are beneficial to carry out a broad channel for the embedding of Na+.

Metal–organic frameworks (MOFs), representing a kind of special skeletal structure, are widely applied in gas storage [16], catalysis [17, 18], drug delivery [19], chemical sensing [20], and energy storage [21]. The MOFs materials are also used as electrode materials for LIBs owing to large surface areas and organized pore sizes [22], which conduce to making the anode contact with the more electrolyte and shortening the ion transport path. Besides, it has been demonstrated that the MOFs as the sacrificial templates can improve the structural stability of the electrode material during insertion and extraction of the ions [22,23,24,25,26]. The graphene nanosheets are broadly applied with the transition metal oxide as good buffers during the volume change in the long cycles and the perfect electrical conductors in the whole battery system [27,28,29,30].

In this work, we make MOFs as sacrificial templates with introducing the Zn(NO3)2.6H2O and 2-methyl imidazole to fabricate ZnO-based nanoparticles. Additionally, GO and the glucose are imported as carbon sources. The construction of a suitable thin carbon layer on the ZnO/reduced graphene oxide (ZnO/rGO) composite via a simple two-step methodology can efficiently utilize the advantage of this two carbon resources. The as-prepared ZnO/rGO/C composite can prevent the aggregation of ZnO, enhance the electronic conductivity and offer a robust scaffold during electrochemical processes with excellent electrochemical performance in LIBs. More importantly, the electrochemical property of ZnO-based materials for Na-ion batteries is also proposed for the first time. We believe this simple methodology can be widely used in other metal oxides system.

Experimental section

Preparation of GO

GO was prepared by Hummers method. In details, 180 ml of concentrated sulfuric acid, 20 ml of concentrated phosphoric acid, and 1.5 g of flake graphite were added into three flasks with slow mechanical agitation. And then, 9 g of KMnO4 was slowly added into the mixed solution. The resulting mixture was further stirred at 48 °C for 12 h till the color of the mixture turned brown. H2O2 was added into the solution until the color became bright yellow and no more bubbles were formed. The mixture was centrifuged and washed with distilled water many times to remove excess acid and impurities, and the GO can be prepared by freeze-drying technique.

Preparation of ZnO, ZnO/rGO and ZnO/rGO/C

As for the MOFs and MOFs-GO, in briefly, 8 g of 2-methyl imidazole and the 3 g of Zn(NO3)2.6H2O were dissolved in 25 ml methanol and 20 ml methanol, respectively. And then, 50 mg of graphene oxide powder was dissolved in 75 ml solution (H2O:CH3CH2OH = 1:2). All of solutions were stirred for 0.5 h. Subsequently, the three kinds of solution were mixed together and stirred another 4 h. The resulting solution was centrifuged, washed with ethanol, and dried overnight at 80 °C in vacuum, and then the MOFs-GO was obtained. While the MOFs was synthesized by the same method without graphene, the ZnO was obtained by carbonizing MOFs at 650 °C for 2 h under N2 protection directly. ZnO/rGO/C was synthesized by making the MOFs-GO as the template, and the 10 wt% glucose was utilized as the carbon source to coat the MOFs-GO under the same condition as ZnO. The ZnO/rGO was formed in the same process without adding glucose. The synthetic routes of ZnO, ZnO/rGO, and ZnO/rGO/C are demonstrated in Fig. 1.

Figure 1
figure 1

The synthesizing processes of the ZnO, ZnO/rGO, and ZnO/rGO/C. (1) the bare ZnO was obtained by carbonizing MOFs. (2) MOFs-GO was carbonized to get the ZnO/rGO. (3) ZnO/rGO/C was obtained by carbonizing MOFs-GO and C6H12O6 together

Full size image

Sample characterizations

The thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) were recorded on a NETZSCH STA 409 PC/DSC. Using JSM, 6490LV and NavoNano Eml to obtain scanning electron microscope (SEM) images, transmission electron microscope (TEM) images were carried out on a JSM 64090LV transmission electron microscope. A Rigaku RINT 2400 X-ray diffractometer with Cu Kα radiation was used to collect X-ray diffraction (XRD) patterns. The Brunauer–Emmett–Teller (BET) and Barrett–Joyner–Halenda (BJH) were performed on a ASAP2020.

Electrochemical testing

The working electrode contains 80 wt% active materials, 10 wt% acetylene black and 10 wt% polyvinylidene difluoride (PVDF). For LIBs, a CR2032 coin-type half-cell was assembled with Celgard 2500 as the separator, the lithium foil as the counter electrode, and 1.0 M LiPF6 solution in ethylene carbonate (EC): diethyl carbonate (DEC): dimethyl carbonate (DMC) (1:1:1) as the electrolyte. For assembling sodium cells, the sodium foil was used as the reference electrode and 1 M NaPF6 in EC: DMC (1:1) as the electrolyte, and the glasses fiber as the separator.

The galvanostatic charge–discharge test was carried out with a LAND system with the cutoff potential window between 0.01 and 3 V for Li-ion batteries, between 0.01 and 2.5 V for Na-ion batteries, respectively. And cyclic voltammetry (CV) tests were measured on Chenghua CHI 660E for Li-ion batteries between 0.01 and 3.0 V with a scan rate of 0.1 mV s−1, while the Na-ion batteries between 0.01 and 2.5 V with a scan rate of 0.05 mV s−1. The electrochemical impedance spectroscopy (EIS) was performed on the frequency ranging from 100 kHz to 0.01 Hz on a Chenghua CHI 660E electrochemical station.

Results and discussion

Structure and morphology

The as-prepared ZnO-based composites were identified by XRD. Figure 2 shows the XRD patterns of ZnO/rGO/C, ZnO/rGO, and ZnO. Clear diffraction peaks at 31.64°, 34.13°, 36.28°, 47.52°, 56.7°, 63.05°, 66.32°, 67.89°, and 69.29° can be observed for all the samples, corresponding to (100), (002), (101), (102), (110), (103), (200), (112), and (201) planes of hexagonal prisms ZnO (JCPDSNo.36-1451), respectively. It suggests the presence of ZnO crystals without obvious impurity in all three samples. An arc peak of ZnO pattern is observed at ~ 25°, which can be related to the carbon produced by the pyrolysis of MOFs [31]. Moreover, the typical diffraction peak at ~ 11° for graphene sheets cannot be obviously distinguished in the XRD patterns of the ZnO/rGO/C and ZnO/rGO, which may be ascribed to the reduction in the GO [32].

Figure 2
figure 2

XRD patterns of as-prepared ZnO, ZnO/rGO, and ZnO/rGO/C

Full size image

The morphology and microstructure of ZnO/rGO/C, ZnO/rGO, and ZnO composites were examined by SEM. The bare ZnO synthesized with the same procedure without GO and C6H12O6 is revealed in Fig. 3a, d. A great deal of homogeneous ZnO nanoparticles is dispersed on reduced graphene sheets as shown in Fig. 3b, f. These ZnO nanoparticles are around 80–150 nm in size with hexahedron frame structure and can form an alternating lamellar structure with rGO nanosheets. This unique stabilized frame structure can play a significant role in the process of insertion and extraction of Li+ or Na+, which can improve the electrochemical performance during the long cycles.

Figure 3
figure 3

SEM images of the ZnO-based samples. a, d the bare ZnO, b, e ZnO/rGO, and c, f ZnO/rGO/C composite with different magnifications

Full size image

In order to further explore the structural information of ZnO/rGO/C, the as-prepared material was identified by TEM and HRTEM characterizations. ZnO nanoparticles with the size about 180 nm (Fig. 4a) were coated with an amorphous carbon layer with the thickness of several nanometers, which can also be in well agreement with Fig. 4b. Besides, as shown in Fig. 4c, these nanoparticles were dispersed on reduced graphene nanosheets. It can be concluded that the structure: ZnO-based MOFs were coated with a layer of carbon and dispersed on rGO sheets successfully, and the special structure can correspond to what we originally designed (Fig. 4f). Furthermore, Fig. 4d, e shows the fringe spacing of the lattice is 0.28 and 0.19 nm, corresponding to the lattice plane of (100) and (102) of ZnO, respectively.

Figure 4
figure 4

a, b, c TEM images of the ZnO/rGO/C, d, e HRTEM images of ZnO/rGO/C, and f schematic diagram of ZnO/rGO/C

Full size image

The content of active materials in the composite was measured by thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) in the air. From Fig. 5a, there is a visible weight loss for ZnO/rGO over the temperature range from 100 to 450 °C. At the same time, there is a strong endothermic peak at 454.4 °C, which can be attributed to the gradual decomposition of the organic matter of ZnO/rGO. From 454.5 to 562.1 °C, a substantial loss of weight may correspond to the combustion of carbon in the air [8]. With the temperature increasing, the weight loss is finished at around 600 °C. Therefore, the weight percentage of ZnO in ZnO/rGO and ZnO/rGO/C were calculated to be 53.78 and 47.39%, respectively.

Figure 5
figure 5

DSC and TGA curves of a ZnO/rGO, b ZnO/rGO/C in the air, c N2 adsorption–desorption isotherms, and d the pore size distribution of the bare ZnO, ZnO/rGO, and ZnO/rGO/C

Full size image

The specific surface area and pore size distribution of these three samples were obtained by using Brunauer–Emmett–Teller (BET) and Barrett–Joyner–Halenda (BJH) methods at 77 K. The nitrogen absorption–desorption isotherms of Fig. 5c show that the isotherm curves of the ZnO/rGO display a obvious Langmuir IV behavior with the H3-type hysteresis loop at a relative pressures of 0.4–0.9, which indicates that the mesoporous exists in it [33]. While the nitrogen-adsorption curves of ZnO and ZnO/rGO/C show that both of them are microporous structure. The BET specific surface area of ZnO/rGO/C (128.33 m2 g−1) is lower than that of ZnO/rGO (151.08 m2 g−1), mainly due to the thin carbon layer coated on the composites. And the BET specific surface area of the ZnO is only 31.96 m2 g−1 due to the agglomeration of ZnO nanoparticles in the absence of GO nanosheets. The pore size distributions of ZnO, ZnO/rGO, and ZnO/rGO/C are shown in Fig. 5d.

Lithium storage performance

The CV and the initial three charge–discharge curves of the as-prepared three ZnO-containing samples between 0.01 and 3 V are shown in Fig. 6. Due to the aggregation and poor electronic conductivity of ZnO nanoparticles, the bare ZnO sample shows weak reversible peaks in its CV curve and poor electrochemical property as anode for Li-ion batteries (shown in Fig. 6a, d). Nevertheless, for the modified ZnO/rGO composite in Fig. 6b, during the first lithiation process, two distinct reduction peaks appeared at 0.27 and 0.68 V, which are related to the reduction of the ZnO to Zn, Zn to the Li-Zn alloy, and the formation of solid electrolyte interface (SEI) [34]. A reduction peak appeared at 1.40 V, but disappeared in the rest of cycles, which is assigned to the formation of the irreversible Li2O. In the first delithiation process, three peaks at 0.55, 0.69, and 1.21 V were observed, indicating the dealloying processes of Li-Zn alloys [35]. The CV curves of the ZnO/rGO/C in Fig. 6c were similar to those of ZnO/rGO. Moreover, the curves of CV are highly coincident over the next few cycles, demonstrating the good reversibility of the ZnO/rGO and ZnO/rGO/C. The first coulombic efficiency (CE) of ZnO (~ 18.6%) is lower than the ZnO/rGO (~ 58.3%) and ZnO/rGO/C (~ 53.1%). The rather low initial CE for bare ZnO is caused by the formation of SEI layer, its poor electronic conductivity as well as the aggregation. The CE of ZnO/rGO is slightly higher than the ZnO/rGO/C mainly due to the electrochemical reaction between carbon and lithium in the first cycle. Obviously, both the ZnO/rGO (Fig. 6e) and ZnO/rGO/C (Fig. 6f) can deliver better specific discharge/charge capacity than the bare ZnO, mainly due to the well-defined structure composed of rGO nanosheets that play a very favorable role during electrochemical processes [36].

Figure 6
figure 6

The CV curves of a ZnO, b ZnO/rGO, and C ZnO/rGO/C at a scan rate of 0.1 mV s−1. The charge/discharge profiles of d ZnO, e ZnO/rGO, and f ZnO/rGO/C during the first three cycles

Full size image

The cyclic behaviors of the ZnO, ZnO/rGO, and ZnO/rGO/C at 0.1C (1C = 978 mA g−1) are shown in Fig. 7a. The reversible capacity of ZnO/rGO/C decreased slightly in the initial 20 cycles and then increased gradually during the rest of cycles, which should be related the activation of the electrodes since a part of the capacity was released from the surface layer of the polymer [37, 38]. And then it reaches its highest value of 878 mA h g−1 in the 95th cycle, followed by a slow decline with 830 mA h g−1 of capacity retained after 100 cycles. The ZnO/rGO/C exhibits the superior cycling behaviors and higher reversible capacities (~ 830 mA h g−1) than the ZnO/rGO (~ 640 mA h g−1) and the bare ZnO (~ 100 mA h g−1), which are ascribed to its optimal morphologies and special structures. Namely, (1) the metal–organic frameworks of the composite can buffer the volume expansion during the insertion and extraction of Li+; (2) the reduced graphene can contribute to dispersing the composite nanoparticles and prevent them from aggregation; and (3) the amorphous carbon layers, which coat the hexahedron ZnO composites, not only can amortize the large volume change, enhance the electrical conductivity of the as-prepared materials, but also prevent the falling off of electrode capacity during the charge–discharge processes [8]. These favorable factors produce synergistic effects, resulting in the excellent charge–discharge performance of ZnO/rGO/C.

Figure 7
figure 7

a The specific capacities of ZnO, ZnO/rGO, ZnO/rGO/C, rGO, and rGO/C at 0.1 C, b the rate capability of ZnO, ZnO/rGO, and ZnO/rGO/C at different current density

Full size image

To exclude the contribution of rGO and carbon layer to the capacity of Li-ion batteries systems, we prepared rGO, rGO/C under the same conditions to assemble Li-ion batteries without the Zn(NO3)2.6H2O and 2-methyl imidazole. From Fig. 7a, the capacity of rGO was only 22 mA h g−1 after 100 cycles, which can be ascribed to the amorphous structure of the as-prepared rGO. This amorphous form of rGO mainly prevents the agglomeration of nanoparticles with ignorable capacity contribution during charging and discharging processes. As shown in Fig. 7a, the rGO/C displays a capacity of ~ 90 mA h g−1 after one hundred cycles, and the carbon layer can sustain the structure during the long cycles [8].

The rate capability of ZnO, ZnO/rGO, and ZnO/rGO/C was evaluated at 0.1–1.6 C as shown in Fig. 7b. Compared to the ZnO, the ZnO/rGO composite shows a better rate performance, which is related to the good electronic conductivity and the shorter Li-ion transport path of ZnO/rGO. Notably, the ZnO/rGO/C composite displays the best electrochemical property. To be specific, the reversible capacity of ZnO/rGO/C was ∼ 154 mA h g−1 at 1.6C, much higher than those of ZnO/rGO (∼ 93 mA h g−1) and ZnO (∼ 17 mA h g−1). In addition, ZnO/rGO/C can recover a capacity of 830 mA h g−1 when the current rate is back to 0.1 C, exhibiting an excellent rate performance, which is mainly due to its optimal morphologies and special carbon-coating structure. In other words, the reducing graphene and carbon not merely promote the electronic conductivity, but increase the contact area with the electrolyte, and short the Li+ diffusion pathway.

To get a better insight into electrochemical behaviors of Li-ion system for the ZnO-based composites, the EIS of the three samples electrodes before and after 100 cycles is tested in Fig. 8. An arc shape at the high frequency which is related to the charge transfer resistance (Rct) and SEI film resistance (Rsf) at the interface between electrode and electrolyte. In the low frequency, the line can correspond to the resistance of the process of Li+ diffusion with electrodes. The resistance sums(Rsf+ct) of the ZnO/rGO/C, ZnO/rGO, and ZnO in Li-ion half-cell were estimated to be 96, 100, and 114 Ω, and shifted to 68, 153, and 419 Ω after 100 cycles, respectively, indicating the ZnO/rGO/C composite displays the highest electrochemical activity and the excellent reaction kinetics. Moreover, the unique structure made up of rGO and carbon layers can enhance electronic conductivity and shorten Li-ion transport pathway.

Figure 8
figure 8

EIS of the fresh ZnO, ZnO/rGO, and ZnO/rGO/C electrode. (inset is the electrode after 100 cycles for the three samples)

Full size image

Sodium storage performance

To study the role of these materials in other types of batteries, the ZnO/rGO/C, ZnO/rGO and ZnO were also newly applied as anode materials for Na-ion batteries. The cutoff voltage is between 0.01 and 2.5 V. As shown in Fig. 9a, the first CE of ZnO/rGO/C is 32.9%, which may be related to the bigger radius of sodium and the irreversible damage of electrode in the first sodiation process [39]. Besides, the first CE of ZnO/rGO/C is higher than that of ZnO/rGO (~ 22.8%) and the ZnO (~ 3.6%) in sodium battery systems, and it should be ascribed to the strong frameworks and the well-defined transport channels of ZnO/rGO/C.

Figure 9
figure 9

a The first charge–discharge curves of ZnO, ZnO/rGO, and ZnO/rGO/C, b the CV curves of ZnO/rGO/C anodes at 0.05 C in sodium-ion systems, c the discharge–charge capacities , and d rate behavior of all three anodes

Full size image

To understand the reaction mechanism of ZnO in Na-ion battery, Fig. 9b shows the first three CV curves of the ZnO/rGO/C electrode at room temperature between 0.01 and 2.5 V at a scan rate of 0.05 mv s−1. In the first sodiation process, the three distinct cathodic peaks were observed at voltages of 0.23, 0.99, and 1.23 V, which is related to the electrochemical conversion of ZnO to Zn, Zn to NaZnx, and the formation of Na2O and SEI layer. In the subsequent cycles, these cathodic peaks were shifted higher to 0.9, 1.28, and 1.74 V due to the polarization effect, respectively. The electrochemical process and the reformation of ZnO can be demonstrated by the following electrochemical conversion reaction [39]:

$$ x{\text{ZnO}} + \left( {2x + 1} \right){\text{Na}}^{ + } + \left( {2x + 1} \right)e^{ - } \to x{\text{Na}}_{2} {\text{O}} + {\text{NaZn}}_{x} $$

In the initial desodiation process, three distinct peaks can be observed at 1.52, 1.87, and 2.06 V, which can be related to the dealloying of NaZnx to Zn, re-oxidation of Zn to ZnO, and decomposition of Na2O. In following CV cycles, the overlap of both oxidation curves and reduction curves confirms the high reversibility of Na2O and NaZnx again.

ZnO/rGO/C composites exhibit the best behavior performance among these three materials in Fig. 9c. Compared with the Li-ion batteries, the first discharge capacity (1579.6 mA h g−1) of Na-ion batteries is lower, which should be ascribed to bigger radius of Na+ than Li+. However, even the radius of Na+ is 34.2% bigger than that of Li+, the unique structure of ZnO/rGO/C still accommodates the insertion of Na+. The capacity retention can be maintained at 300 mA h g−1 after 100 cycles is higher than that of ZnO (~ 13 mA h g−1) and ZnO/rGO (~ 164 mA h g−1), thus outperforms many of the transition metal oxide materials for SIBs [12,13,14, 40, 41].

In addition, this is the first time using the as-prepared material as anode materials of SIBs as far as we know. The rate performance of the as-prepared materials was also evaluated in Fig. 9d. Among three kinds of anode materials, the ZnO/rGO/C has the best rate capability and optimal reversibility, the reversible capacity of ZnO/rGO/C was ∼ 162 mA h g−1 at 0.2C, much higher than those of ZnO/rGO (∼ 101 mA h g−1) and ZnO (∼ 60 mA h g−1) in SIBs. When the current density revers back to 0.05C, the capacity of the ZnO/rGO/C still remains 342 mA h g−1 after 100 cycles. Furthermore, this new synthesis method can be further extended to other secondary batteries, and the related work is still ongoing.

Conclusions

In summary, a novel ZnO/reduced graphene oxide/carbon composite is synthesized successfully via a simple pyrolysis of Zn-based metal–organic framework. This special structure is made of ZnO which is homogeneously dispersed on rGO sheets and coated with a carbon layer. It exhibits superior rate capability, especially reversible capacity (~ 830 mA h g−1 after 100 cycles, approximately 85% of theoretical capacity) as anodes for LIBs. Additionally, it is the first time to use ZnO materials as anodes for Na-ion battery, and it demonstrates an outperformance cyclic stability and considerable reversible capacity (~ 300 mA h g−1 after 100 cycles). The special structure, herein, not only can amortize the large volume change, enhance the electrical conductivity, but also prevent the falling off of electrode capacity. This simple method of synthesizing this unique structure can be widely applied in energy storage materials and extended to other important applications.