M-Ni-Co MOF (M=Zn, Fe, Mn) for high-performance supercapacitors by adjusting its morphology

Abstract

Metal-organic frameworks (MOF) have been wildly synthesised and studied as electrode materials for supercapacitors, and bimetallic MOF of Ni and Co has been broadly studied to enhance both specific capacitance and stability of supercapacitors. Herein, a best performance (about 320 F/g) of Ni–Co bimetallic MOF was found in a uniform preparation condition by adjusting the ratio of Ni to Co. Then tiny third metal ion was introduced, and we found that the morphology of material has a significant change on the original basis. Furthermore, certain ions (Zn, Fe, Mn) introduced make a huge improvement in capacitance based on Ni–Co MOF of 320 F/g. The result shows that Zn–Ni–Co MOF, Fe–Ni–Co MOF and Mn–Ni–Co MOF perform specific capacitance of 1135 F/g, 870 F/g and 760F/g at 1 A/g, respectively. Meanwhile, the asymmetric supercapacitor (ASC) was constructed by Zn–Ni–Co MOF as positive electrode and active carbon (AC) as negative electrode. The Zn–Ni–Co MOF//AC ASC possesses a energy density of 58 Wh/kg at a power density of 775 W/kg. This research provides a new methods to regulate the morphology of MOF and a novel viewpoint for assembling high-performance, low-price, and eco-friendly green energy storage devices.

1. Introduction

With the ceaselessly consumption of non renewable energy, like fossil energy, the reserves of them reduce continuously. Renewable energy like wind, solar, power, etc, rapidly develop in this condition [1]. Therefore, the study of supercapacitors as a sort of energy storage devices is one of the key points to enhance the utilization rate of renewable energy [2]. Batteries with Faraday reaction as the main energy storage mechanism have properties of high energy density but low power density, while traditional capacitors with double layers as the main energy storage mechanism have properties of high power density but low energy density. The study and fabrication about supercapacitors aim to break through the power limit of batteries and energy limit of capacitors [3]. In addition, supercapacitors are used into some instruments which need both high power density and energy density, like electric vehicles, to make the vehicles more stable when they start or climb [4]. Supercapacitors are classified into three categories according to its energy storage mechanism, which are electric double layer capacitors (EDLCs), pseudocapacitors (PCs) and battery type capacitors (BTCs) [5]. Briefly, in EDLCs, the charge was stored by the reversible ion absorption at the interface between electrode and electrolyte [6]; in PCs, the charge was stored by the redox reaction on or near the electrode surface [3]; in BTCs, redox reactions occurs in bulk phase of the active materials with the break and recombination process of chemical bond [3,7]. Materials about carbon was usually utilized into EDLCs, activated carbon (AC), carbon nanotubes (CNTs) and graphene included. They have disadvantage of low specific capacitance but advantages of high cycling stability, low cost, low toxicity and environmental friendliness [5,6]. On the contrast, PCs and BTCs can provide with high specific capacitance but a trivial cycling stability. RuO₂ is a classic material based on the surface redox reaction mechanism; MoO₃ is a typical material for BTCs, and MOF is a set of materials based on both redox reaction mechanism and double layer mechanism, which also be called hybrid supercapacitors (HSCs) [3,8,9].

MOF is a sort of materials that are constructed by the center and organic linkers. The center contains metal, and the strong bonds link the center and organic linkers. In this way, it grows into open crystalline frameworks with permanent porosity [10,11]. Varied metal salt and different organic linkers form thousands of types MOFs for different pore structures. High porosity makes MOF appropriate in adsorption, and varied pore structure make it high-performance in storage of different gas, like hydrogen, methane, acetylene, CO₂, etc [[12], [13], [14], [15], [16]]. High specific surface induced by the adjustable pore structures and ions in system supply MOF a high performance in supercapacitors [5,17]. The metal ions provide redox sites, while the porous structure and conjugated π electron cloud show the properties of double layers [18]. Notably, synergistic effect is a major factor to prove the electrochemical performance of electrodes [5]. For example, Garima Chaturvedi et al. synthesize a kind of nanoflower-like Ni-MOF material with a specific capacitance of about 815 F/g, and Yu Han et al. synthesize Co-MOF material with a specific capacitance of about 554.4 F/g, while Xiaoya Kang et al. Synthesize Ni, Co-MOF material with a specific capacitance of about 985.2 F/g [19,20]. Inducing a new metal ion may change the morphology of the material and increase the more binding sites, which make bimetallic MOF and trimetallic MOF widely be utilized into catalyzing [21].

In this work, a best ratio of Ni to Co was determined by a gradient experiment first. Base on the best ratio of Ni to Co, the third metal ion was introduced with different ratio to explore the influence on the capacitance and different morphology controlled by the third metal ion. And then, the best trimetallic MOF, Zn–Ni–Co with a certain ratio of Zn, was chosen to be the positive electrode of the ASC while AC was chosen to be the negative electrode. Zn–Ni–Co MOFs//AC ASC achieved a high energy density of 59.40 Wh/kg at a power density of 798 W/kg. Additionally, a method to control the morphology of MOFs was raised in this paper.

2. Experimental

2.1. Materials

Nickel chloride hexahydrate (NiCl·6H₂O), cobaltous nitrate hexahydrate (Co(NO₃)₂·6H₂O), zinc nitrate hexahydrate (Zn(NO₃)₂·6H₂O), manganese nitrate tetrahydrate (Mn(NO₃)₂·4H₂O), Ferric nitrate nonahydrate (Fe(NO₃)₃·6H₂O), terephthalic acid (PTA) were obtained commercially (Shanghai Aladdin Reagent Co., LTD). N,N dimethylformamide (DMF) and Ethylene glycol (EG) was obtained by Modern Oriental (Beijing) Technology Development Co. Ltd.

2.2. Synthesis of Ni–Co-MOF

Certain amount of NiCl·6H₂O, Co(NO₃)₂·6H₂O and 0.6 mmol PTA were dissolved into a mixed solution containing 20 mL DMF and 15 mL EG with the molar ratios (Ni: Co) of 0:1, 1:3, 1:1, 3:1 and 1:0. Samples of different molar ratio are presented by C-1 to C-5. The total amount of metal ions is 0.85 mmol. Subsequently, the solvend was ultrasonic dispersed for 10 min, followed by stirring for 2 h at room temperature to get a uniform bright red solution. Then, the solution was poured into reaction kettle, and the kettle was placed in oven at 170 °C for 6 h. After naturally cooling down to room temperature, the product was obtained by centrifugation. Finally, wash it with ethanol for three times and dry it at 60 °C for 12 h.

2.3. Synthesis of M-Ni-Co-MOF (M = Zn, Fe, Mn)

The third metal ion Mⁿ⁺ was added into Ni–Co-MOF in the first step. The total amount of Co and Ni was maintained, and the ratio (Co:Ni) of 1:1 was used in this synthesis process for its best electrochemical properties in the gradient ratio of bimetallic system. 0.425 mmol Co(NO₃)₂·6H₂O, 0.425 mmol NiCl·6H₂O, 0.6 mmol PTA and certain amount of M-salt were dissolved into a mixed solution containing 20 mL DMF and 15 mL EG with the molar ratios (Co:Ni:M) of 1:1:0.025, 1:1:0.05, 1:1:0.1, 1:1:0.15, 1:1:0.175. Samples of different ratio are represented by M − 1 to M − 5, respectively, where M = Zn, Fe, Mn. The rest process was the same as synthesis of Ni–Co-MOF's.

2.4. Electrochemical measurement

The electrochemical behavior of the samples as electrode materials for supercapacitors were performed on an electrochemical workstation (CHI660E, CH Instruments, Inc., Shanghai) with a representative three-electrode system at room temperature. The electrolyte was 2 M KOH aqueous solution; the counter electrode is platinum electrode, and reference electrode is Hg/HgO. In preparation of working electrode, the substance on the working contains 80 wt% active matter (about 4 mg Ni–Co or M-Ni-Co MOF), 10 wt% acetylene black as conductive agent, and 10 wt% polytetrafluoroethylene as binder. These matter were mixed into ethanol and coated onto a piece of nickel foam with an area about 1.0 × 1.0 cm². Finally, the nickle foam was pressed with a pressure of 10.0 MPa to gain the working electrode. The active matter's mass was obtained by the mass discrepancy of nickle foam before and after pressing the substance on. The electrochemical tests contains cyclic voltammetry (CV) at different scan rates from 5 mV/s to 50 mV/s, galvanostatic charging and discharging (GCD) at different current densities from 1 A/g to 20 A/g in voltage from 0 V to 0.4 V, and electrochemical impedance spectroscopy (EIS). According to the CV and GCD test, the specific capacitance (C, F/g) could be calculated by the following two equations [22,23]:

$$C=\frac{1}{2mv\Delta V}\int i(V)dV$$ (1)

$$C=\frac{I\Delta t}{m\Delta V}$$ (2)

Where C (F/g) is the specific capacitance of the working electrode, I (A) is the discharging current, t (s) is the discharging time, v (V/s) is the scan rate, and ΔV (V) is the potential window. To further understand the charge storage mechanism of a material as electrodes, chemical kinetics properties were analyzed through CV curve. It is well known that the correlation between different scanning rates v (V/s) and their corresponding peak current i (A) is shown in the following equation [24].

$$i=a{v}^b$$ (3)

Generally, the storage mechanism of charge is judged by b value. The charge storage process tend to diffusion control which behavior more like batteries if the b value close to 0.5, while the charge storage process tend to surface capacitance control which behavior more like pseudocapacitors if the b value close to 1.0. Take the log of both sides of equation (3), we get equation

$$\mathit{log}i=b\mathit{log}v+\mathit{log}a$$ (4)

Value of b can be got form the slop of the line log(i) to log(v). Furthermore, the ratio of diffusion control and surface capacitance control can be determined by following equation [25].

$$i(v)=k_{1}v+k_2{v}^{1/2}$$ (5)

Where k₁v represents the degree offered by surface capacitance control, and k₂v represents the degree offered by surface diffusion control at different scan rate.

2.5. Material characterization

The micromorphology of samples were tested by scanning electron microscopy (SEM, SU8010, Hitachi, Japan) with an X-ray energy dispersive spectrometer (EDS). The crystal structure and chemical valence of the prepared materials were characterized by X-ray diffraction (XRD, Ultima IV, Cu Ka), and X-ray photoelectron spectroscopy (XPS, Thermo Scientific K-Alpha). Fourier transform infrared spectroscopy (FTIR) was conducted on a Spectrum One instrument (Frontier FTIR, PerkinElmer, USA).

2.6. Assembly of asymmetric supercapacitor (ASC)

Zn-3 pressed on the nickel foam as the positive electrode and active carbon on the nickel foam as the negative electrode, filter paper infiltrated by 6 M KOH as exchange membrane, an asymmetric supercapacitor (ASC) was assembled. The mass ratio of active materials Zn-3 and active carbon can be determined by following charge balance equation:

$$\frac{m^{+}}{m^{-}}=\frac{C_s^{-}\Delta V^{-}}{C_s^{+}\Delta V^{+}}$$ (6)

where m (g) represents the mass of active materials in each electrode, C_s (F/g) is the specific capacitance of each active material as working electrode in a three electrode system, and V (V)is the potential window, the plus or minus sign represents that this physical quantity belongs to the positive electrode or the negative electrodes. Calculated by above equation, the ratio of active materials in positive electrode and negative electrode was 3.2: 1. The mass of Zn-3 was about 16 mg, and the mass of AC was about 5 mg in this work. After testing the electrochemical performance of ASC, the energy density (E, Wh/kg) and power density (P, W/kg) can be calculated by following equations:

$$E=\frac{1}{2\times 3.6}C\Delta V^2$$ (7)

$$P=\frac{3600\times E}{\Delta t}$$ (8)

where C (F/g) is the specific capacitance determined by equation (1) or (2) of the assembled ASC, V (V) is voltage window of the ASC, and t (s) presents the discharging time in GCD curves.

3. Result and discussion

3.1. Structure and morphology characterization

The main synthesis procedure of M-Ni-Co MOF (M = Zn, Fe, Mn) was displayed in Fig. 1. Method of solvothermal was used. The morphology of C-3, Zn-3, Fe-2 and Mn-3 were analyzed by SEM. The images shown in Fig. 2a and e, the morphology of C-3 was microrob made of some tiny fibers (∼700 nm), with a size about 5 μm long and 1.5 μm wide. Zn-3 (Fig. 2b and f) consist of boat-like unit with the size of about 6 μm long and 3 μm wide, and every unit was self-assembled by several sheets. As shown in Fig. 2c and g, the morphology of Fe-2 was microsphere made of some flakes (∼1 μm), and the size is about 4∼10 μm. Mn-3 (Fig. 2d and h) has a ellipsoidal particle structure, and every unit has a size of about 700 nm length and 350 nm width. The above description of dimensions was obtained by taking the average, and the statistical results were shown in Fig. S1 and S2. The introduction of a small amount of the third metal ion had a significant effect on changing the morphology of MOF. In addition, when the concentration of the third metal was appropriate, the morphology of the material tended to be uniform (Fig. 2 a, b and d). If the concentration of the third metal ion was too little, the morphology of the material tended to be complex. As shown in Fig. S3a, the morphology of Zn-1 included both microrob and boat-like structure. On the contrary, if the third metal ion is equivalent to Ni²⁺ and Co²⁺, the morphology will tend to be irregular (Fig. S3b), and there is a significant reduction in electrochemical performance. We believe that there must be a certain connection between morphology and electrochemical performance. Notably, in C-3, Zn-3 and Mn-3, the size of the units is similar to each other and separated from each other, while in Fe-2, the size of the microspheres is very different, and the spheres are easily cross-linked with others (Fig. S3c). This partly explains why C-3, Zn-3 and Mn-3 are powdery and Fe-2 is prone to agglomeration.

Fig. 1.

Schematic illustration of the fabrication process of M-Ni-Co MOFs (M = Zn, Fe, Mn).

Fig. 2.

The SEM images of (a, e) C-3, (b, f) Zn-3, (c, g) Fe-2 and (d, h) Mn-3.

To explore the elements contained in samples, EDS mapping of C-3, Zn-3, Fe-2 and Mn-3 was performed and shown in Fig. 3. The images (Fig. 3ã d) shows that atomic ratio of Ni to Co was close to 1, and the report (Fig. S4c) also confirms this. Uniform dots in mapping images proved that Ni and Co indeed well doped in each sample. In C-3 (Fig. 2a), element O was also uniform and much more abundant than Ni and Co, which was one of the evidences proving that PTA was present in C-3 as organic ligands. In Zn-3, Fe-2, Mn-3 (Fig. 2b ∼ d), The third metal atoms (Zn, Fe, Mn) was much less than Ni and Co, which was consistent with the synthesis procedure of M-Ni-Co-MOF. By Controlling for other variables in synthesis procedure, we believe that it is the third metal ion that caused the changes in morphology. As shown in Fig. S4, the atomic ratio of Ni, Co and M in sample Zn-3, Fe-2, Mn-3 were 8.86: 5.23: 0.01, 16.36: 14.83: 0.11 and 15.54: 14.10 and 0.22. The doping level of Zn is significantly lower than the other two group, which indicated that the doping level not only determined by the dose of ingredient. We speculate that Zn plays a role similar to that of a catalyst in the reaction process, participating in the process of changing morphology, but the actual number of coordination involved is very small.

Fig. 3.

The EDS mapping images for Ni, Co, O of C-3 (a), for Ni, Co, Zn of Zn-3 (b), for Ni, Co, Fe of Fe-2 (c) and for Ni, Co, Mn of Mn-3 (d).

The surface elemental composition and chemical state of C-3, Zn-3, Fe-2, Mn-3 were investigated by XPS. There are obvious peaks in the survey spectrum (Fig. 4a), proving that these four samples contain elements C, O, Ni and Co. The high-resolution XPS spectrum of C 1s (Fig. 4b) could be fitted with three characteristic peaks. Peak located at about 284.7 eV assigned to C–C, and the peaks at 286.3 and 288.4eV assigned to C–N and C O, respectively [26,27,28], proving that the C is from organic ligands (C O and C–C), and DMF (C–N and C–C). From the fitting result, the high-resolution XPS spectrum of Ni 2p (Fig. 4c) has two characteristic peaks which belong to Ni 2p_3/2 and Ni 2p_1/2 and two satellite peaks. According to the binding energy of 855.6 eV and 873.43 eV, the chemical state of Ni can be determined as Ni²⁺ [[29], [30], [31], [32]]. Similarly, from the fitting result, the high-resolution XPS spectrum of Ni 2p (Fig. 4d) has two characteristic peaks which belong to Co 2p_3/2 and Co 2p_1/2 and two satellite peaks. According to the binding energy of 781.3 eV and 797.1 eV, the chemical state of Co can be determined as Co²⁺ [[27], [31], [32], [33], [34], [35]]. Trace elements Zn, Fe, Mn were also analyzed in high-resolution XPS spectrum (Fig. 4e). In high-resolution XPS spectrum of Zn, peaks at 1021.3 and 1044.3 eV attribute to the typical characteristic peaks of Zn²⁺ [36,37]. As for Fe, peaks at 712.3 eV and 723.4 eV can attribute to Fe 2p_3/2 and Fe 2p_1/2 orbitals, revealing that the valence state of Fe is Fe³⁺ [31,34,[38], [39], [40]]. The high-resolution XPS spectrum of Mn 2p could be fitted with four characteristic peaks, which located at 654.8 eV, 647.2 eV, 643.1 eV and 636.4 eV, and they can attribute to Mn 2p_1/2, Mn²⁺, Mn⁵⁺ and Mn³⁺ [41,42], indicating that Mn is unstable during high-temperature reaction, and some of it undergoes oxidation.

Fig. 4.

(a) The XPS survey spectrum of C-3, Zn-3, Fe-2 and Mn. (b) C 1s, (c) Ni 2p, (d) Co 2p high-resolution XPS spectra of C-3, Zn-3, Fe-2 and Mn-3. (e) Zn 2p, Fe 2p and Mn 2p high-resolution XPS spectra of Zn-3, Fe-2 and Mn-3, respectively. (f) the XRD patterns of C-3 and standard cards. (g) the XRD patterns of C-3, Zn-3, Fe-2 and Mn-3. (h) The FTIR spectra of C-3, Zn-3, Fe-2 and Mn-3.

The crystallographic phase of C-3, Zn-3, Fe-2 and Mn-3 was characterized by XRD shown in Fig. 4f and g. As could be seen in Fig. 4f, the diffraction peaks of CCDC 153067 (Co-MOF) and CCDC 985792 (Ni-MOF) were very similar, which indicated that they have a very similar crystal structure. Besides, the characteristic peaks of C-3 located at 2θ = 8.8°，15.7° and 17.8° were highly consist with the CCDC 153067 (Co-MOF) and CCDC 985792 (Ni-MOF) [43,44]. This indicated that the crystal structure of C-3 is similar to Ni-MOF and Co-MOF, which can be a strong evidence proving successful synthesis of Ni–Co MOF. Furthermore, Fig. 4g displayed the comparison of XRD patterns among C-3, Zn-3, Fe-2 and Mn-3. C-3, Zn-3 and Mn-3 had characteristic peaks located at the same degrees, indicating that they had similar crystal structure, while Fe-2 had different characteristic peaks, indicating that the crystal structure had changed. In order to further explore the changes in crystal structure of several MOFs, monometallic MOFs were synthesised at the same condition using the same ligand and tested in XRD, Zn-MOF, Fe-MOF and Mn-MOF included. As shown in Fig. S5a, Zn-MOF's characteristic peaks, Zn-3's characteristic peaks and C-3's characteristic peaks were highly matched indicating that they have similar crystal structures. The similarity of Zn-MOF's and C-3's characteristic peaks proves that changing the metal ion center of MOF from Ni²⁺ and Co²⁺ to Zn²⁺ has little impact on the crystal structure of MOF materials. This further explains the similarity between C-3's crystal structure and Zn-3's crystal structure. The XRD patterns of Fe-MOF, Fe-2 and C-3 were very different (Fig. S5b). The valence state of Fe is different from that of Ni and Co, which may cause the difference of coordination number, resulting in the changing of crystal structure. The differences among Fe-MOF's XRD pattern, Fe-2's XRD pattern and C-3's XRD pattern indicated that there can be certain cross action between Ni²⁺, Co²⁺ and Fe³⁺ in Fe-2, leading to that the crystal structure of Fe-2 is different from that of Fe-MOF and that of C-2. Fig. S5c showed the XRD patterns of Mn-MOF, Mn-3 and C-3. Compering Mn-MOF and Mn-3 respectively to C-3, they have the same characteristic peaks at 2θ = 8.8°，15.7° and 17.8°, while are different at interval from 2θ = 20°–35°. We believe that the complexity of valence state of Mn ion leads to this results. From the analysis of Fig. 4e, Mn²⁺, Mn³⁺ and Mn⁵⁺ all existed in Mn-3 and Mn-MOF. Mn²⁺ as the metal ion center may maintain its original crystal structure, while Mn³⁺ and Mn⁵⁺ very likely change the crystal structure, introducing new peaks into XRD pattern. According to the above analysis, we proposed two hypotheses regarding the mechanism by which the third metal ion affects morphology. First, the introduce of the third ion directly change the crystal structure of the MOF materials, resulting in the changing of the morphology. Second, the third metal ions do not change the crystal structure but change the lattice constant due to different atom size, affecting the self-assembly process, and then results in the change in morphology.

The surface chemical functional groups of C-3, Zn-3, Fe-2 and Mn-3 were analyzed by FTIR. As shown in Fig. 4h, the adsorption peak at 3402 cm⁻¹ was a stretching vibration peak of O–H [45,46], representing the presence of coordinated H₂O molecules in these four samples. The adsorption peaks at about 1572 cm⁻¹ and 1382 cm⁻¹ can attribute to asymmetric and symmetric stretching modes of coordinated (–COO–) group [46], respectively, which was from organic ligands PTA. Additionally, The peak at 751 was derived from the bending vibrations of C–H, further confirming the existence of organic compounds in the materials.

3.2. Electrochemical performance of Ni–Co and M-Ni-Co MOF (M = Zn, Fe, Mn)

Before exploring the electrochemical performance of M-Ni-Co MOF materials, electrochemical performance of Ni–Co MOF materials containing different ratio of Ni to Co was test in a three-electrode system, 2 M KOH solution as electrolyte, in order to find a ratio of the best electrochemical performance. To emphasize the influence from different Ni, Co ratio, a working electrode used Ni–Co MOF prepared under the same condition with different ratio of Ni to Co. The galvanostatic chargedischarge (GCD) curves (Fig. S6a) of Ni–Co MOF (C-1 to C-5) showed that C-3 had the longest discharge time, which proved it had the highest charge storage capacity among Ni–Co MOF. Sample C-1 to C-5 have specific capacitance of 59.5 F/g, 185 F/g, 315 F/g, 268 F/g and 238 F/g respectively, according to equation (2). Notably, the charging and discharging parts shows a nonlinear process with a plateau, which means that redox reactions occur in Ni–Co MOFs. As for C-3, cyclic voltammogram test was done with different scan rates from 5 to 50 mV/s (Fig. S6b). It is worth noting that obvious peaks existed in all CV curves, which can attribute to the redox reactions, and the plateaus in charging and discharging process of GCD curves also proved this reversible Faraday reactions. In addition, the oxidation peaks' position shift to positive orientation, and the reduction peaks positions shift to negative orientation with the increasing scan rate. This phenomenon can be ascribed to the influence of electrode polarization [17]. To further study the storage mechanism of charge, the relation between the logarithm of the current peaks and logarithm of the scanning rates is linearly fitted (Fig. S6c). The value of b was 0.79 for anodic current and 0.78 for cathodic current. Furthermore, the ratio of capacitance caused by diffusion control mechanism and caused by capacitance control mechanism was calculated by equation (5). Fig. S6d displayed this ratio at 5 mV/s to 50 mV/s In C-3, capacitance control dominated the whole charge storage process. As scanning rate increasing, surface-control's contribution lifted by degrees, while diffusion-control's contribution gradually declined. This result can attribute to the increased moving speed and of ions in the solution and ions' decreased diffusion path. In addition, as the scanning rates increasing, the ions in the solution was harder to reach internal part of the active materials, just staying at the interface between electrolyte and electrode. This process collaborated with polarization and ohmic effect to impeded the diffusion-control process and facilitated the surface-control process leading to the dominance of the suface-control process at a high scanning rate [23]. Fig. S6e was obtained by GCD testing of C-3 at different current densities ranging from 1 A/g to 20 A/g. By calculating, the specific capacitance values of C-3 were as large as 315 F/g, 302 F/g, 265 F/g, 220 F/g, 170 F/g at current densities of 1 A/g, 2A/g, 5 A/g, 10 A/g and 20 A/g, respectively. As the current density increased by 20 times, the capacitance of C-3 retained 52%. Furthermore, the GCD curves had a good symmetry with a pair of plateaus at different current densities proving that the Faraday reactions in charge storage process was highly reversible. Notably, when the current density reach to 20 A/g, GCD curve tended to linear indicating that redox reactions were unconspicuous, which proved that capacitance control dominated the charge storage process at huge charge density.

C-3 had the best electrochemical performance among C-1 to C-5, thus different ratio of ion Mⁿ⁺ (M = Zn, Fe, Mn) were introduced based on C-3, and they were called M − 1 to M − 5 according to different M concentration. Similarly, the electrochemical properties of each sample were studied by a three-electrode system in 2 M KOH aqueous electrolyte, and the working electrode used M-Ni-Co MOF. The CV curves of Zn–Ni–Co MOF (Zn-1 to Zn-5) (Fig. S7a) at 5 mV/s showed that Zn-3 had the largest integrated area, which indicated that Zn-3 had the largest specific capacitance among Zn–Co–Ni MOF, and GCD curves (Fig. S7b) at 1A/g also confirmed this by the longest discharge time for Zn-3. Calculated by equation (2), sample Zn-1 to Zn-5 had specific capacitance of 301 F/g, 610 F/g, 1135 F/g, 713.5 F/g and 488.5 F/g, respectively. As for Zn-3, cyclic voltammogram test was done with different scan rates from 5 to 50 mV/s (Fig. 5a). According to equation (4), value of b equals 0.72 for anodic current and 0.68 for cathodic current by linear fitting of log (i) and log (V) (Fig. 5b). Fig. 5c was obtained by GCD testing of Zn-3 at different current densities ranging from 1 A/g to 20 A/g. By calculating, the specific capacitance values of Zn-3 were as large as 1135 F/g, 1044 F/g, 863 F/g, 692 F/g, 522 F/g at current densities of 1 A/g, 2 A/g, 5 A/g, 10 A/g and 20 A/g, respectively. As the current density increased by 20 times, the capacitance of Zn-3 retained 46%. The CV curves (Fig. S8a) and GCD curves (Fig. S8b) of Fe–Ni–Co MOF (Fe-1 to Fe-5) showed that Fe-2 had the largest integrated area and longest discharging time, which proved that Fe-2 had the largest specific capacitance among Fe–Co–Ni MOF. According to equation (2), sample Fe-1 to Fe-5 have specific capacitance of 161 F/g, 835F/g, 167F/g, 107F/g and 52.5F/g, respectively. As for Fe-2, CV test was done with different scan rates from 5 to 50 mV/s (Fig. S8c). According to equation (4), value of b equals 0.67 for anodic current and 0.74 for cathodic current by linear fitting of log (i) and log (V) (Fig. S8d). Fig. S8e was obtained by GCD testing of Fe-2 at different current densities ranging from 1 A/g to 20 A/g. By calculating, the specific capacitance values of Fe-2 were as large as 835 F/g, 743 F/g, 635 F/g, 534 F/g, 426 F/g at current densities of 1 A/g, 2A/g, 5 A/g, 10 A/g and 20 A/g, respectively. As the current density increased by 20 times, the capacitance of Fe-2 retained 51%. Similarly, CV and GCD tests for Mn–Ni–Co was done. As shown in Fig. S9a and S9b, Mn-3 had the largest integrated area and longest discharging time, which proved that Mn-3 had the largest specific capacitance among Mn–Ni–Co MOF. According to equation (2), sample Mn-1 to Mn-5 have specific capacitance of 181.5 F/g, 318 F/g, 760 F/g, 297.5 F/g and 56.5 F/g, respectively. As for Mn-3, CV test was done with different scan rates from 5 to 50 mV/s (Fig. S9c). According to equation (4), value of b equals 0.60 for anodic current and 0.58 for cathodic current by linear fitting of log (i) and log (V) (Fig. S9d). Fig. S9e was obtained by GCD testing of Mn-3 at different current densities ranging from 1 A/g to 20 A/g. By calculating, the specific capacitance values of Mn-3 were as large as 760 F/g, 486 F/g, 319 F/g, 251 F/g and 122 F/g at current densities of 1 A/g, 2 A/g, 5 A/g, 10 A/g and 20 A/g, respectively. As the current density increased by 20 times, the capacitance of Mn-2 retained 16%. Additionally, the contribution of surface and capacitance controlled of Zn-3 (Fig. 5g), Fe-2 (Fig. S8f) and Mn-3 (Fig. S9f) at 5-50 mV/s was calculated by equation (5). Apparently, diffusion control parts of Zn-3, Fe-2 and Mn-3 were larger than that of C-3, and similar to C-3, the ratio of surface controlled gradually improved as the increase in scanning rates.

Fig. 5.

(a) The CV curves at different scan rates of Zn-3. (b) Linear relation between logarithm of rates and logarithm of current peaks. (c) The GCD curves of Zn-3 at different current densities. (d) The CV curves at 5 mV/s and (e) the GCD curves at 1A/g of C-3, Zn-3, Fe-2 and Mn-3. (f) Nyquist plots of C-3, Zn-3, Fe-2 and Mn-3. (g) Different contribution ratios at different scan rates of Zn-3. (h) The cycling stability performance at 10 A/g for C-3, Zn-3, Fe-2 and Mn-3.

Fig. 5d showed CV curves of C-3, Zn-3, Fe-2 and Mn-3 which perform the best electrochemical properties in each group. Zn-3 had the largest integrated area consistent with its largest capacitance, and the longest discharging time (Fig. 5e) also confirmed this. To study the influence of various components of the electrode on its electrochemical performance, electrochemical impedance spectroscopy (EIS) of C-3, Zn-3, Fe-2 and Mn-3 was performed, and Nyquist plots of these four materials and equivalent circuit was shown in Fig. 5f. R2 in the equivalent circuit represented charge transfer resistance. By fitting, the R2 value of C-3, Zn-3, Fe-2 and Mn-3 were 4.07 Ω, 0.97 Ω, 0.67 Ω and 0.87 Ω. By contrast, the resistance of Zn-3, Fe-2 and Mn-3 is significantly smaller than C-3, which indicated that Zn-3, Fe-2 and Mn-3 had much faster charge transfer rates explaining their better electrochemical performance. Compared to some other Ni, Co-MOF or MOF derived LDT, R2 values of our samples still has advantages [16,17,22]. Table. S1 listed their R2 values. The cyclic stability of C-3, Zn-3, Fe-2 and Mn-3 was tested by charging and discharging cycles at current density of 10 A/g. As shown in Fig. 5h, after 3000 cycles, the specific capacitance retention rates of C-3, Zn-3, Fe-2, Mn-3 were 93 %, 63 %, 53 % and 60 %, respectively. The first and last 10 cycles of cycling stability test were displayed in Fig. S10. C-3 is more stable than other three. We speculated that the structure formed by introduction of new ions is not as stable as the structure of Ni–Co MOF, resulting in a lower stability than C-3's. Furthermore, numerous studies have shown that Ni-MOF and Co-MOF are unstable in alkaline solutions and form corresponding hydroxides [[47], [48], [49], [50], [51]]. This is one of the reasons why the cycling stability performance of Ni–Co MOF and M-Ni-Co MOF is much worse than Ni–Co layered double hydroxide (LDT) [22]. The redox reaction mechanism of Ni Co hydroxides was inferred from the above electrochemical characteristics and literature [22], and can be described as the following chemical equation:

Co(OH)₂(Ⅱ) + OH⁻ ↔CoOOH(Ⅲ) + e⁻ (9)

Ni(OH)₂(Ⅱ) + OH⁻ ↔NiOOH(Ⅲ) + e⁻ (10)

3.3. Electrochemical properties of the hybrid supercapacitor device assembled by Zn–Ni–Co MOF (Zn-3) and active carbon

Zn-3 had the best electrochemical performance in a three-electrode system. Therefore, Zn-3 was used as the positive electrode and AC as the negative electrode to study the electrochemical performance of Zn-3 in a hybrid supercapacitor (HSC), 6 M KOH aqueous as electrolyte (Fig. 6a). The electrochemical properties of AC in a three electrode system was performed in Fig. S11. Fig. S11a shows the CV cures at scan rates from 5 mV/s to 50 mV/s, and Fig. S11b shows the GCD curves at current densities from 1 A/g to 20 A/g. A voltage window from −1.1 V to 0 V was selected in AC, and a voltage window from 0 V to 1.5 V in HSC was roughly estimated from this, since there is no overlap between AC's voltage window and Zn-3's voltage window (Fig. 6b). Fig. 6c presented the CV curves of Zn-3//AC HSC at scan rates from 5 mV/s to 50 mV/s over a voltage range from 0 V to 1.5 V. redox peaks present at about 1.1V and a smooth curves in 0V–0.9 V confirms that the energy storage mechanism of this HSC contains both double electric layer and pseudocapacitance. Fig. 6d shows the GCD curves of Zn-3//AC HSC at current densities of 1 A/g to 20 A/g. The GCD curves of the HSC maintain a good symmetry under high current densities, which prove that the redox reaction in energy storage process is reversible and that the device has a good Coulombic efficiency.

Fig. 6.

(a) Schematic illustration of Zn-3//AC hybrid supercapacitor. (b) The CV curves of Zn-3 and AC electrodes at a scan rate of 5 mV/s in a three-electrode mode. Electrochemical performance of Zn-3//AC hybrid supercapacitor: (c) the CV curves at different scan rates from 5 mV/s to 50 mV/s, (d) the GCD curves at current densities from 1A/g to 20 A/g, (e) the cycling stability performance of Zn-3//AC HSC at a current density of 10 A/g. (f) The ragone plot correlating E and P compared with other reported hybrid devices.

The cyclic stability and the Coulombic efficiency of Zn-3//AC HSC was tested under the current density of 10 A/g. As shown in Fig. 6e, After 3000 cycles, the retention of HSC's capacitance maintain about 60 %, and the Coulombic efficiency was about 100 %. The first and the last 10 cycles of the cycling stability performance test was shown in Fig. S11c and S11d respectively. Furthermore, to verify the practical application performance of Zn-3//AC HSC, we make it as the power of a small LED bulb, and the bulb stay on for several hours. Additionally, the power density and the energy density of this HSC was calculated by equations (7), (8). This device has a energy density of 58 Wh/kg at a power density of 775 W/kg. And it still retains an energy density of 12 Wh/kg at a power density of 7750 W/kg, which was better than graphite nanosheets decorated with Ni-MOF [28], Ni–Co MOF onto porous support [52], hierarchical NPO@CNTs derived from Ni-MOF [53] and other Ni–Co MOF derived materials [[54], [55], [56], [57]]. The ragone plot correlating E and P of these hybrid devices was shown in Fig. 6f.

4. Conclusions

In summary, Ni–Co MOF was obtained by hydrothermal process, and the M-Ni-Co MOF (M = Zn, Fe, Mn) was obtained by hydrothermal process based on Ni–Co MOF with the best electrochemical performance. M-Ni-Co MOF with different morphology were made by introducing different M ions. The electrochemical performance of Zn–Ni–Co MOF was up to 1135 F/g at 1 A/g current density. In addition, Zn-3//AC hybrid supercapacitors performed a energy density of 58 Wh/kg at a power density of 775 W/kg. This study introduce a method to make MOFs with different morphology to get a better electrochemical performance, which can also be applied to other MOF materials.

CRediT authorship contribution statement

Yige Yan: Writing – original draft, Writing – review & editing, Visualization, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Mohan Huang: Writing – review & editing, Visualization. Yongsheng Wang: Supervision, Resources, Project administration, Funding acquisition. Dawei He: Supervision, Resources, Project administration, Funding acquisition. Jiaqi He: Writing – review & editing, Funding acquisition, Formal analysis.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is/are the supplementary data to this article: