Performance enhancement of asymmetric supercapacitors with bud-like Cu-doped Mn3O4 hollow and porous structures on nickel foam as positive electrodes

Cu-doped Mn3O4 hollow nanostructures supported on Ni foams as high-performance electrode materials for supercapacitors were successfully synthesized through a facile hydrothermal method and subsequent calcination. The morphology, structure, and electrochemical performance of the as-prepared Mn3O4 nanostructures can be tuned just by varying the Cu doping content. Benefiting from the unique bud-like hollow structure, the 1.5 at% Cu-doped Mn3O4 sample has a high specific capacitance of 257.6 F g−1 at 1 A g−1 and remarkable stability (about 90.6% retention of its initial capacitance after 6000 electrochemical cycles). Besides, an asymmetric supercapacitor (ASC) cell based on the 1.5 at% Cu-doped Mn3O4 exhibits a high specific capacitance of 305.6 F g−1 at 1 A g−1 and an energy density of 108.6 W h kg−1 at a power density of 799.9 W kg−1. More importantly, the ASC shows good long-term stability with 86.9% capacity retention after charging/discharging for 6000 cycles at a high current density of 5 A g−1.


Introduction
In recent years, supercapacitors (SCs) have been widely recognized in a wide range of energy storage applications due to their properties such as their faster charging-discharging, higher specic power density, and longer cycling life compared to batteries. 1,2 In particular, asymmetric supercapacitors (ASCs), which usually include a faradic electrode (as energy source) and a capacitive electrode (as power source) in one system, have been considered a promising strategy to increase the potential window and the energy density. 3 The capacitive electrode materials generally involve various metal oxides, 4-6 metal hydroxides, 7,8 metal sulphides, 9,10 metal carbonate hydroxides 11,12 and conductive polymers. 7,13 Among these materials, hausmannite (Mn 3 O 4 ) has shown promising potential in the eld of supercapacitors on account of its low-cost, low toxicity, environmental benignity 14 and satisfactory supercapacitive performance. [15][16][17] Compared with MnO 2 , Mn 3 O 4 has much superior properties, such as uniform structure, stable performance and easily obtained pristine phase Mn 3 O 4 . 18 However, the direct application of pure Mn 3 O 4 in supercapacitors is limited owing to its low electronic conductivity, low specic capacitance values and non-ideal rate performances. 7 Doping of various elements such as Sn, 19 Cr, 2 Co 2 and Ni 2 into Mn 3 O 4 compounds has been found to be effective way to promote its conductivity and capacitance. It is obvious that doping with the selective metal ions may overcome the limitations of electronic states and also generate new phenomena. Su et al. 20 reported the utilization of Cu-doped d-MnO 2 as electrode material for a supercapacitor electrode, and exhibited higher capacitance and better cycle performance than that of the undoped samples. To our best knowledge, we have also not found the report on the preparation of Cu-doped Mn 3 O 4 as electrode materials for supercapacitors. 2 Moreover, the capacitive behavior of Mn 3 O 4 electrode critically depends on its crystal structure and morphology. 16 Therefore, the majority of studies on Mn 3 O 4 electrode material for supercapacitors have focused on achieving the highest capacitance by adjusting growth conditions to obtain Mn 3 O 4 with desirable morphologies. These morphologies include nanospheres, 21 nanorods, 22 nanosheets, 23 and nanobers. 24 The unique bud-like morphology of Mn 3 O 4 will be reported in this paper.
In this work, we explore the effect of Cu doping on the supercapacitor performance of Mn 3 O 4 -based electrodes, synthesized by a one-pot hydrothermal method. The effects of Cu doping on the modication of structure, surface morphology and electrochemical performance have been studied exhaustively and discussed in detail.

Experimental
All chemicals reagents were of analytical-reagent grade and purchased from Zhanyun Chemical Reagent Co., Ltd (Shanghai, China) without further purication. Prior to use, the fresh Ni foam was slightly etched in an ultrasonic bath with 3 M HCl to remove oxides, and then washed thoroughly with deionized water and ethanol respectively. The Mn 3 O 4 (or Cu-doped Mn 3 O 4 ) samples were obtained by thermally annealing MnCO 3 (or Cu-doped MnCO 3 ) precursors in the presence of nitrogen gas ow, and the MnCO 3 (or Cu-doped MnCO 3 ) precursor was synthesized by a modied procedure according to reference. 25 Mn(CH 3 COO) 2 $4H 2 O (3.48 g), urea (CO(NH 2 ) 2 ) (5.07 g) and of NH 4 F (1.63 g) were dissolved one by one in 10 min interval with constant stirring in 70 mL of ethanol solution to form a homogenous solution. Doping concentration of cupric nitrate (Cu(NO 3 ) 2 $6H 2 O) was 0, 0.5, 1, 1.5 and 2 at%, which refers to the molar ratio of Cu to Mn in the reactants. The total volume was then brought up to 350 mL with distilled water and transferred into ten 50 mL Teon-lined autoclaves with 8 pieces of cleaned Ni foams (1 Â 1 cm 2 in sizes) in each autoclave. The autoclave was kept at 160 C for 16 h in an electric oven and cooled down to room temperature. Finally, the precursors grown Ni foam were collected by washing with distilled water and ethanol, then dried in a vacuum furnace at 60 C for 6 h and further calcined in nitrogen gas at 400 C for 3 h. The total mass loading on nickel foam was around 2 mg cm À2 for all the ve fabricated electrodes.
The crystalline structures of the samples was characterized by X-ray diffraction (XRD) with use of a Smartlab-9 X-ray diffractometer with Cu Ka radiation (l ¼ 1.5418Å). The microstructure and chemical composition of the Cu-doped Mn 3 O 4 samples was characterized by scanning electron microscopy (SEM, Zeiss Supra 35VP, USA) equipped with an energy-dispersive X-ray spectroscopy (EDS). The chemical composition of the Cu-doped Mn 3 O 4 samples was also were determined by means of an inductively-coupled plasma (ICP) mass spectrometer (ICP-AES, Horiba Jobin Yvon). Specic surface areas were measured with a Brunauer-Emmett-Teller (BET) sorptometer (SSA-4300, Beijing Builder, Inc.) using N 2 adsorption at 77 K. To evaluate the electrochemical performance of the fabricated electrodes, cyclic voltammetry (CV), galvanostatic charge-discharge (GCD), and electrochemical impedance spectroscopy (EIS) measurements were conducted in a three-electrode cell in N 2 -saturated 1.0 M Na 2 SO 4 aqueous solution. The electrochemical performances were recorded on a conventional CHI660E setup comprising a saturated calomel electrode (SCE) reference electrode, a counter electrode of Pt sheet, and a working electrode of Ni foam supported prepared Mn 3 O 4 -based electrode. The Cu/Mn atomic ratios of various Cu-doped Mn 3 O 4 were also determined by ICP spectrometer and EDS, and listed as in Table 1. The representative EDS microanalysis of 1.5 at% Cudoped Mn 3 O 4 is shown in Fig. 2(a), which also conrms that Cu dopants have been successfully introduced into the Mn 3 O 4 . The atomic ratio of Cu/Mn was found to be linearly proportional to the concentration of Cu in the initial reaction solution. The elemental mapping of 1.5 at% Cu-doped Mn 3 O 4 are shown in Fig. 2(b), which intuitively shows the distribution of Mn, Cu, O. It is suggested that the Cu is uniformly distributed in the sample.

Results and discussion
The morphologies of the samples were examined using SEM to investigate the effects of Cu doping on the nanostructure. In  structure. The formation of this hollow structure is assumed to take place through a self-assembly process similar to that described by Meher et al. 26 in the case of MnO 2 . Therefore, the incorporation of Cu dopants may have a signicant inuence on the hausmannite Mn 3 O 4 , leading to the formation of hollow structure. However, as doping concentration of Cu reached 2 at%, the hollow morphology disappeared, and coarse grains were produced. As for the bud-like hollow structure of moderate Cu-doped Mn 3 O 4 particles, it is assumed that the formation of micrometer sized hollow spheres is analogous to that reported previously. 25,27,28 It can be explained as a fundamental solidstate phenomenon, the so-called Kirkendall effect, 27,29 which deals with the movement of the interface between diffusion couples. Fig. 3(f) shows a typical low-magnication SEM image of the 3D hierarchical structure of the 1.5 at% Cu-doped Mn 3 O 4 grains growing on Ni foam. Obviously, the Mn 3 O 4 grains are uniformly grown on the Ni foam substrate while completely maintaining the 3D grid structure of the pristine Ni foam. The XRD and SEM investigations conrm that although Cu dopant has no inuence on the crystal structure the hausmannite Mn 3 O 4 , it indeed make apparent change in the morphological structure. The TEM image of the 1.5 at% Cu-doped Mn 3 O 4 further conrms the hollow bud-like morphology (Fig. 3(g)). The SAED pattern (inset in Fig. 3(g)) of the 1.5 at% Cu-doped Mn 3 O 4 shows a crystalline nature with the (1 0 1), (1 1 2), (2 0 0), and (2 1 1) planes of Mn 3 O 4 . Moreover, in the HRTEM image ( Fig. 3(h)), the lattice fringes are clearly seen and matches well with d-spacing of 0.25 and 0.31 nm, corresponding to (2 1 1) and (1 1 2) planes of Mn 3 O 4 , respectively. The results are consistent with the XRD results (Fig. 1).
The specic surface area and pore parameters of undoped and Cu-doped Mn 3 O 4 powders scraped from Ni foam were further investigated by measuring adsorption-desorption isotherms of nitrogen at 77 K. Fig. 4 illustrates the N 2 adsorption/desorption isotherms and corresponding Barrett-Joyner-Halenda (BJH) adsorption pore size distribution plots (inset). According to the IUPAC classication, all these samples behave type-IV isotherms. The increase of the Cu doping concentration has a signicant effect on the nal textural parameters, including the surface areas and pore volume. The BET surface areas are 95.2 m 2 g À1 for undoped Mn 3 O 4 , 97.8 m 2 g À1 for 0.5%, 103.4 m 2 g À1 for 1%, 114.5 m 2 g À1 for 1.5% and 109.2 m 2 g À1 for 2% Cu-doped Mn 3 O 4 , respectively. The specic pore volumes of undoped, 0.5 at%, 1 at%, 1.5 at% and 2 at% Cudoped Mn 3 O 4 are 248, 180, 239, 262 and 251 m 3 g À1 , respectively. Among the samples studied, 1.5 at% doped Mn 3 O 4 shows the largest BET surface area and special pore volume. It is worthy of note that Cu dopant has no signicant inuence on the pore size of the Mn 3 O 4 nanostructure. The pore size distribution (insets of Fig. 4) in these samples was similar, featuring pores between 2 nm and 100 nm, which mainly consisted of mesopores and macropores. These multiple mesopores pore size distribution would facilitate the electrolyte ions transfer during electrochemical charging/discharging by providing an efficient transport pathway and more adsorption sites in the interior of the electrodes, 30 which is helpful to enhance specic capacitance of the electrodes.
CV has been used to evaluate the electrochemical performances of Cu-doped Mn 3 O 4 -based electrodes (Fig. 5(a)). It is clearly illustrate that Cu dopant can affect the CV curve area. It is well known that the larger CV curve area is, the greater specic capacitance will be obtained. 31 It can be seen that the CV curve area rstly increases and then decreases with the increase of the Cu doping concentration. It is clearly indicated that the specic capacitance of the 1.5 at% Cu-doped Mn 3 O 4 sample is the maximum. The CV curves of 1.5 at% Cu-doped Mn 3 O 4 sample with the scan rate ranging from 5 to 100 mV s À1 were performed in the potential window of 0 to 0.8 V, as is shown in Fig. 5(b). The CV curves at different scan rates exhibit  an approximately rectangular shape, which is an indication of ideal capacitive behavior with excellent reversibility of this electrode material. 32 The specic capacitance (C) was calculated from CV curves according to the following equation: 33 where I is the response current, V is the potential, v is the potential scan rate, m is the mass of active electrode material.
The specic capacitance is calculated as 194.5 F g À1 at a scan rate of 5 mV s À1 and 133.3 F g À1 at 100 mV s À1 , which is comparable to the reported values of Mn 3 O 4 based electrodes in literature. [34][35][36] It keeps a comparable high specic capacitance retention even at high scan rate of 100 mV s À1 , showing a good electrochemical behavior of the device. The energy storage mechanism of Mn 3 O 4 electrode materials has been accepted as the proton-electron mechanism. [37][38][39] While the doped Cu ions may incurs the following redox reactions: 40 Cu 2+ + e À # Cu + . GCD experiments were carried out to test the capacitive performance of the undoped and doped Mn 3 O 4 samples. The GCD experiments were carried out from 0 to 0.8 V with different current densities from 1 to 10 A g À1 . The inuence of discharge current density on the specic capacity of the undoped and doped Mn 3 O 4 electrodes was summarized in Fig. 5(c). And Fig. 5(d) shows the representative charge-discharge curve of 1.5 at% Cu-doped Mn 3 O 4 sample. The ideal charge-discharge characteristic and good reversibility are further conrmed by the linear and symmetric charge-discharge proles. 41 The specic capacitance of the electrode was calculated according to the following equation: where I, t, m and DV are constant discharge current (A), discharging time (s), weight (g) of electroactive material and discharging potential range (V), respectively. Using eqn (2), the specic capacitances of undoped, 0.5 at%, 1 at%, 1.5 at% and 2 at% Cu-doped Mn 3 O 4 samples are calculated to be 180.3, 222.5, 238.2, 257.6 and 169.6 F g À1 at the current density of 1 A g À1 , respectively. The specic capacitance increase with increasing doping concentration of Cu from 0 to 1.5 at%, and the 1.5 at% Cu-doped Mn 3 O 4 sample shows a maximum specic capacitance of 257.6 F g À1 at 1 A g À1 , a 43% increase in specic capacitance over the undoped one. However, further increase of the Cu content to 2 at% results in a fade electrochemical performance. The specic capacitances have the same order as that of the variation of the CV result with respect to the doping concentration. As can be seen in Fig. 5(c), similar tendency in the specic capacitances for Cu-doped Mn 3 O 4 and undoped electrodes are maintained at different current densities ranging from 1 to 10 A g À1 . It clearly shows that the 1.5 at% Cu-doped Mn 3 O 4 sample depicts the highest specic capacitances among all the prepared samples at all tested current densities, apparently revealing its excellent capacitive performance. The specic capacitance obtained from the discharging curves ( Fig. 5(d)) is calculated to be 183.0 F g À1 at a current density of 2 A g À1 , which is almost comparable with the specic capacitance 194.5 F g À1 calculated from the CV measurements at a scan rate of 5 mV s À1 . The improvement in the specic capacitance for 1.5 at% Cu-doped Mn 3 O 4 sample could be attributed to the bud-like hollow nanostructure as well as the enhanced electrical conductivity, which will be illustrated in following EIS part. The addition of Cu dopants can induce the formation of porous hollow nanostructure with larger surface area and higher pore volume. The mesoporous structures can promote the mass transport of electrolyte within the electrode during the redox reaction process and provide a large number of exposed active sites. Moreover, the direct contact of the deposited material with the underlying conductive Ni foam is highly favorable for electron collection and avoids the use of a polymer binder and conductive additives, and substantially reduces the "dead volume" in the electrode. In conclusion, the higher pore volume of 1.5 at% Cu-doped Mn 3 O 4 sample may also be useful to facilitate the diffusion of active species and thus contribute to the specic capacitance enhancement. The cycle stability is a crucial parameter for fast energy storage device. The cycle stability tests of the 1.5 at% Cu-doped Mn 3 O 4 sample was performed at the current density of 2 A g À1 for 6000 times, with undoped Mn 3 O 4 electrode as reference, as shown in Fig. 5(e). For the 1.5 at% Cu-doped Mn 3 O 4 sample, it is worth noting that the two curves present a short-termed increase of specic capacitance in the initial hundreds cycles due to an improvement in the surface wetting of the electrode by the electrolyte during extended cycling. 42 Aer 6000 chargedischarge cycles, 1.5 at% Cu-doped Mn 3 O 4 electrode materials show 90.6% of capacitance retention. For undoped Mn 3 O 4 , the specic capacitance presents a slight increase in the initial 750 cycles and thereaer decreased gradually, just 77.4% of initial capacitance can be maintained aer 6000 cycles. The results proved that Cu-doped Mn 3 O 4 materials can be served as promising electrode materials for supercapacitors.
The EIS analyses were carried out using an AC voltage of 5 mV in a frequency range from 0.01 Hz to 100 kHz to better understand the electrochemical properties of undoped and Cudoped Mn 3 O 4 samples. Fig. 6 shows the Nyquist plots for all samples, two regions of distinct electrochemical response can be seen. At high frequency region, a distorted semicircle is observed; at low frequency, the response is indicative of diffusion process represented by a straight line. It is well known that the power output capability of electrochemical capacitors depends largely upon the equivalent series resistance (ESR). 43 The ESR, related with the electrode material's conductivity, has been estimated from the X-intercept of the Nyquist plot. 26,44 The ESR values of undoped, 0.5 at%, 1 at%, 1.5 at% and 2 at% Cudoped Mn 3 O 4 sample were found to be 1.95, 1.71, 1.62, 1.49 and 1.90 U, respectively. The ESR value of samples rstly decreases and then increases with increasing Cu dopant content is clearly observed. The 1.5 at% Cu-doped Mn 3 O 4 sample exhibits the lowest ESR, which suggested that the electrode has an easy access to ions for the intercalation and deintercalation. 45 The straight sloping line at low frequency represents the diffusion of ions in the electrode material, and the slopes of these straight sloping lines are different. It was observed that the Nyquist plots of the 1.5 at% Cu doped Mn 3 O 4 sample displays more steeply rising behavior in the low frequency region as compared to the other samples, which demonstrates that its capacitive performance is much closer to an ideal supercapacitor. Therefore, the highest specic capacitance of the 1.5 at% Cu-doped Mn 3 O 4 sample can be attributed to the minimum ESR. The above results suggest that the 1.5 at% Cu-doped Mn 3 O 4 sample is a better candidate for supercapacitor application, as it has been already conrmed by the CV and GCD results.
To study the capacitive performance for practical application, the asymmetric supercapacitor (ASC) device was fabricated by 1.5 at% Cu-doped Mn 3 O 4 sample as the positive electrode, graphene (G) coated Ni foams as the negative electrode, poly-tetrauoroethylene (PTFE) membrane separator and 1 M Na 2 SO 4 aqueous solution as the electrolyte. This assembled ASC was denoted as Mn 3 O 4 (1.5Cu)//G. To obtain the maximum performance of the Mn 3 O 4 (1.5Cu)//G ASC, it is crucial to keep the charges balanceable with the relationship q + ¼ q À . In order to get the charge balanceable, the optimum loading mass of AC was decided by the following equation: where m is the mass of activated material, C m represents the specic capacitance and DV is the potential window in the three-electrode test system. The GDC curve of the G electrode at 1 A g À1 is depicted in Fig. S2 † and the calculated C m value of the G electrode is 171.7 F g À1 . The C m value of the Mn 3 O 4 (1.5Cu) electrode is 257.6 F g À1 at 1 A g À1 (Fig. 5(c)). The optimal mass loading of G in Mn 3 O 4 (1.5Cu)//G was 3 mg. Fig. 7(a) (Fig. 7(b)), further indicating an ideal capacitive behaviour. Therefore, the 1.6 V potential was selected as the suitable operational voltage range for further electrochemical measurements. Fig. 7(c) shows the CV curves of the assembled ASC at different scan rates ranging from 10 to 500 mV s À1 within the potential window of 0-1.6 V. The shape of CV curves can be well retentive with increasing the scan rates, suggesting the ideal electron-transfer kinetics and good rate capability of the asymmetric supercapacitor. Fig. 7(d) shows the typical GCD curves of the ASC measured at various current densities in a potential window of 0-1.6 V. The GCD curves at various current densities from 1 A g À1 to 50 A g À1 manifest approximately symmetrical triangular shapes, demonstrating a good reversibility between the charge and discharge processes. 46 The specic capacitance calculated from galvanostatic discharge curves of the ASC device in Fig. 7(d) is 305.6 F g À1 at 1 A g À1 . Even at a high current density of 50 A g À1 , the specic capacity remains at 175.0 F g À1 (about 57.3% of the capacitance retained), showing the good rate capability. Long-term cycling performance is also one of important factors for supercapacitor applications. The cycling stability of the 1.5 at% Cu-doped Mn 3 O 4 //G ASC is evaluated by the repeated charging/discharging measurement at a constant current densities of 5 A g À1 between 0 and 1.6 V. As shown in Fig. 7(e), the specic capacitance of the device decreases gradually with increasing cycling times, and still retains 86.9% of the initial specic capacitance even aer 6000 cycles. Meanwhile, its coulombic efficiency maintained high values over 99% at the whole cycles. These results suggest that the 1.5 at% Cu-doped Mn 3 O 4 //G ASC delivers good electrochemical stability. Moreover, to evaluate the energy storage ability of the 1.5 at% Cudoped Mn 3 O 4 //G ASC, the Ragone plots based on the galvanostatic discharge curves are displayed in Fig. 7(f). The energy and power density of the ASC devices were calculated based on these equations: 47 where E (W h kg À1 ) is the average energy density; C m (F g À1 ) is the specic capacitance of the ASC device; DV (V) is the voltage window; P (W kg À1 ) is the average power density and Dt (s) is the discharge time. As shown in Fig. 7(

Conclusion
In summary, we have successfully fabricated hierarchical Cudoped Mn 3 O 4 hollow nanostructures on Ni foam by a simple hydrothermal method. The existence of Cu can not only induce the formation of the unique bud-like hollow nanostructure, but also effectively increase the conductivity of the electrode. It was found that Cu doping plays a promotional role in enhancing Mn 3 O 4 behavior in electrochemical capacitors. Electrochemical test conrms that 1.5 at% Cu-doped Mn 3 O 4 could achieve a high specic capacitance (257.6 F g À1 at 1 A g À1 ) and longterm cycling stability (90.6% capacitance retention aer 6000 times repetition at current density of 2 A g À1 ), holding a promise to be applied in high-performance asymmetric supercapacitor. Meanwhile, a 1.5 at% Cu-doped Mn 3 O 4 //G ASC is further fabricated and exhibits a good electrochemical performance with high specic capacitance, excellent long-term cycling stability, and good electrochemical reversibility of the faradaic reaction. As a result, the obtained bud-like Cu-doped Mn 3 O 4 can be a promising electrode material for future energy storage devices.

Conflicts of interest
The authors declare no conict of interest.