Novel Co3O4 Nanoparticles/Nitrogen-Doped Carbon Composites with Extraordinary Catalytic Activity for Oxygen Evolution Reaction (OER)

Herein, Co3O4 nanoparticles/nitrogen-doped carbon (Co3O4/NPC) composites with different structures were prepared via a facile method. Structure control was achieved by the rational morphology design of ZIF-67 precursors, which were then pyrolyzed in air to obtain Co3O4/NPC composites. When applied as catalysts for the oxygen evolution reaction (OER), the M-Co3O4/NPC composites derived from the flower-like ZIF-67 showed superior catalytic activities than those derived from the rhombic dodecahedron and hollow spherical ZIF-67. The former M-Co3O4/NPC composite displayed a small over-potential of 0.3 V, low onset potential of 1.41 V, small Tafel slope of 83 mV dec−1, and a desirable stability. (94.7% OER activity was retained after 10 h.) The excellent performance of the flower-like M-Co3O4/NPC composite in the OER was attributed to its favorable structure. Electronic supplementary material The online version of this article (10.1007/s40820-017-0170-4) contains supplementary material, which is available to authorized users.


Introduction
Depletion of fossil fuels and the rapidly growing energy demands have necessitated the development of sustainable energy conversion and storage systems such as metal-air batteries, water splitting devices, and fuel cells [1][2][3][4]. The development of durable, highly efficient, low-cost, and eco-friendly electrocatalysts for the oxygen evolution reaction (OER) is crucial for the commercial application of these renewable energy technologies [5,6]. To date, precious metal-based materials, such as RuO 2 and IrO 2 , have been considered as the most optimal catalysts for OER owing to their lowest over-potentials at practical current densities [7]. However, their commercial applications have been severely impeded because of their poor stability, prohibitive cost, and low selectivity [8].
Recently, significant efforts have been made to explore transition metal-based electrocatalysts for the OER because of their low cost, abundant reserves, environmental benignity, and resistance to corrosion in alkaline solutions [9][10][11][12]. Among them, Co-based catalysts have emerged as promising alternatives for precious metal-based catalysts [13][14][15][16]. The electrocatalytic activity for OER is closely related to the active sites and electronic conductivity of the catalysts. Previous research has demonstrated that active sites can be engineered by modulating the particle size, pore structure [17,18], and the crystallinity [19,20] of Co 3 O 4 . Furthermore, coupling with carbon effectively improves the electronic conductivity of the catalysts [21][22][23]. Nevertheless, carbon itself as a catalyst displays relatively low catalytic OER activity. Recent studies have shown that doping with either nitrogen or transition metals into carbon nanostructure can efficiently promote its catalytic performance [23][24][25][26]. The template method has proven to be an effective protocol for obtaining nitrogendoped Co 3 O 4 /C composites. In this method, various organic hybrids, which contain both the transition metal and nitrogen, are used as precursors such as melamine [27], porphyrin [28], polyaniline [29,30], and salen [31]. However, it is hard to control the size, structure, and morphology of these organic hybrids in an exact manner; therefore, deficiencies and non-uniform distributions of active sites are prevalent, which are also crucial for electrocatalytic activity.
Metal organic frameworks (MOFs) have attracted a significant attention as materials for the preparation of nonprecious metal electrocatalysts because of their inherent advantages such as a controllable porous structure, innate doping with heteroatoms, and an ultrahigh surface area [32,33]. Zeolitic imidazolate frameworks (ZIFs) have proven to be promising as pyrolytic precursors for various porous metal oxides/doped carbon composites [34][35][36]. Via direct pyrolysis, carbon layers with a porous structure can be formed in situ with metal nanoparticles encapsulated homogeneously, and sufficient contacts can be formed between the metal nanoparticles and the carbon matrix. Notably, a highly ordered three-dimensional structure promotes the structural stability of MOFs against pyrolysis, and the remarkable surface-to-volume ratio of MOFs can effectively promote the electrochemical catalytic reactions.
Among the variety of MOF materials available, ZIF-67 is one of the most widely investigated ones because of its high concentration of active cobalt sites as well as a facile synthetic method. Herein we have proposed a facile method to prepare Co 3 O 4 /NPC composites with different morphologies derived from ZIF-67. By slightly modulating the synthetic route of the ZIF-67 precursors, it was possible to control the morphology of the product. Thus, in addition to the typical rhombic dodecahedron morphology, novel flower-like ZIF-67 and hollow spherical ZIF-67 were fabricated. These ZIF-67 precursors were then pyrolyzed to obtain the Co 3  , and polyvinylpyrrolidone ((C 6 H 9 NO) n ) were obtained from Sinopharm Chemical Reagent Co. Ltd. All reagents were used as received without further purification.

Preparation of ZIF-67 Precursors
T-ZIF-67 was synthesized according to a previously reported method [37]. In a typical procedure, solutions of Co(NO 3 ) 2 Á6H 2 O (5.82 g) in methanol (400 mL) (solution A) and 2-methylimidazole (6.48 g) in methanol (400 mL) (solution B) were prepared. Solution B was gradually added into solution A with continuous stirring. After standing for a while, layers were observed and the supernatant was eliminated. The solution was then centrifuged and washed with methanol for 3-5 times to remove the excess Co 2? . T-ZIF-67 was finally acquired as a purple solid after drying at 60°C for 3 h.
The synthetic route to M-ZIF-67 was almost the same as that of T-ZIF-67, except that Co(NO 3 ) 2 Á6H 2 O was replaced by CoSO 4 Á7H 2 O (5.62 g). During the synthesis of H-ZIF-67, 1.00 g PVP was added to the methanol solution of 2-methylimidazole as a morphology modifier, and other steps were the same as that for the synthesis of M-ZIF-67.

Preparation of Co 3 O 4 /NPC Composites
The as-prepared M-ZIF-67, H-ZIF-67, and T-ZIF-67 materials were first ground into powders. They were then individually heated to 550°C in air at a heating rate of 5°C min -1 . After keeping at 550°C for 5 h, the powdered materials were cooled down to room temperature at a cooling rate of 5°C min -1 and black M-Co 3 O 4 /NPC, H-Co 3 O 4 /NPC, and T-Co 3 O 4 /NPC powders were obtained, respectively.

Characterization
Powder X-ray diffraction (PXRD) analysis of the materials was performed on a Bruker-AXS D8 Advance X-ray diffractometer with Cu Ka radiation (k = 0.15406 nm). The morphologies and elemental mappings of the samples were obtained from a Hitachi SU70 field-emission scanning electron microscopy (SEM) instrument at 10 kV and 20 kV. The high-resolution transmission electron microscopy (HRTEM) characterization was carried out on a Tecnai F30 microscope at an accelerating voltage of 300 kV. Elemental analysis was performed on a Vario EL III elemental analyzer. The specific surface area and pore size distribution were determined by the Brunauer-Emmett-Teller (BET) method conducted by the TriStar II 3020 surface area and porosity analyzer. Thermogravimetric analysis (TGA) of the samples was carried out on a SDTQ600 thermoanalyzer in air. X-ray photoelectron spectroscopy (XPS) was performed on a Thermo Scientific ESCALAB 250Xi with Al Ka radiation (hm = 1486.8 eV).

Electrochemical Measurements
Cyclic voltammetry (CV) and linear sweep voltammetry (LSV) measurements were taken on an Autolab PGSTAT302N electrochemical workstation (NOVA 1.9). The evaluation of the catalytic activity for the OER was conducted at room temperature in a conventional threeelectrode system. Co 3 O 4 /NPC composites were used as the working electrode, a platinum foil acted as a counter electrode, and a reversible hydrogen electrode (RHE) was employed as the reference electrode. To prepare the working electrode, 5 mg of the active material was dispersed in a mixture of 0.95 mL ethanol and 0.05 mL 5 wt% Nafion solution with sonication for 60 min. Next, the catalyst (20 lL) was pipetted out and dropped onto a glassy carbon electrode with a diameter of 5 mm. It was then fully dried at room temperature for 12 h before measurements (loading *0.510 mg cm -2 ). Figure S1 shows the PXRD patterns of M-ZIF-67, H-ZIF-67, and T-ZIF-67. Apparently, these three materials exhibit the same XRD pattern with principal diffraction peaks at 7.39°, 10.43°, 12.73°, and 18.07°, which are exactly matched with the simulated ZIF-67 pattern. This suggests that the three ZIFs have the same composition. This result was also supported by their FTIR spectra (Fig. S2). The diffraction peaks of T-ZIF-67 were much higher than those of M-ZIF-67 and H-ZIF-67, implying a higher crystallinity of T-ZIF-67 in comparison with the other two ZIF-67 precursors.

Results and Discussion
The morphologies of the ZIF-67 precursors and the asprepared Co 3 O 4 /NPC composites were studied by SEM. T-ZIF-67 showed a rhombic dodecahedron morphology with particle sizes of *1 lm, which is the typical morphology of ZIF-67 (Fig. 1a). On the other hand, the morphology of M-ZIF-67 was flower-like with particles of size *1.6 lm (Fig. 1b) and that of H-ZIF-67 was hollow spherical with a diameter of *800 nm and shell thickness *200 nm (Fig. 1g, h). After pyrolysis at 550°C for 5 h under air, all three ZIF-67-derived composites inherited the morphologies of their precursors without either particle agglomeration (Fig. S3) or structural collapse, indicating a high structural stability of the obtained Co 3 O 4 /NPC composites. Specifically, the surfaces of T-ZIF-67 shrunk into a rhombic dodecahedron center with Co 3 O 4 nanoparticles uniformly embedded in the carbon scaffold (Fig. 1b). M-ZIF-67 and T-ZIF-67 underwent similar changes in morphology to yield M-Co 3 O 4 /NPC (Fig. 1e) and H-Co 3 O 4 /NPC (Fig. 1h), respectively. To determine the elemental composition of the composites, elemental mapping analysis was conducted. As shown in Fig. 1c, f, i and Table S1, all three composites were mainly comprised of cobalt and oxygen, with trace amounts of carbon and nitrogen. This implied that the pyrolysis of ZIF-67 yields a nitrogen-doped carbon scaffold encapsulated in situ with Co 3 O 4 nanoparticles. Further detailed investigations were performed by using HRTEM (Fig. 2) To clearly illustrate the process of morphology control, the schematic diagrams of the synthetic procedure are presented in Fig. 3. In the traditional synthetic method of ZIF-67, Co(NO 3 ) 2 Á6H 2 O has been used as the metal source. In this work, we used CoSO 4 Á7H 2 O as the metal source instead. The introduction of SO 4 2? species accelerated the nucleation of ZIF-67, leading to multiple polyhedrons being embedded mutually, until finally flower-like ZIF-67 particles had formed. As for H-ZIF-67, PVP was employed as a template. As shown in Fig. 3, 2-methylimidazole combined with the PVP molecular chain via hydrogen bonds when they were dissolved together in methanol. This interaction between the ligands and the template forced ZIF-67 to grow along the chain, resulting in flake-like ZIF-67, which then piled together to form a hollow sphere.
The To gain an in-depth understanding of the pore structure of the three composites, the nitrogen adsorption-desorption isotherms and the corresponding pore size distribution curves of ZIF-67 precursors and Co 3 O 4 /NPC composites were determined. As shown in Fig. 5a, the nitrogen adsorption-desorption isotherms of M-ZIF-67, H-ZIF-67, and T-ZIF-67 agreed with Langmuir I. In M-ZIF-67, the quantity of adsorbed N 2 increased dramatically at a low relative pressure, indicating abundant micropores in the flower-like particles. Besides, at the tail of the isotherm (high relative pressure), the absorbance increased quickly, suggesting a large amount of mesopores. Similarly, T-ZIF- .703 m 2 g -1 , respectively. Noticeably, the adsorption type changed from Langmuir I to Langmuir III after pyrolysis (Fig. 5c), and the pore sizes became larger and the distribution was more dispersive (Fig. 5d)   The thermal stabilities of the three composites were investigated by TGA under air atmosphere. As shown in Fig. 6, heavy mass losses for M-ZIF-67, H-ZIF-67, and T-ZIF-67 started at 550, 300, and 400°C, respectively. When the temperature increased to 950°C, the weights remained at 44.85%, 11.17%, and 36.96%, respectively. The dramatic weight loss was attributed to the combustion of the carbon species. It is noteworthy that both H-ZIF-67 and T-ZIF-67 went through a slight mass loss before decomposition, while M-ZIF-67 was stable below 500°C. This phenomenon indicated that the thermal stability of M-ZIF-67 was much superior to that of H-ZIF-67 and T-ZIF-67. Figure 7 shows the XPS results of the M-Co 3 O 4 /NPC catalyst. As shown in Fig. 7a, the full XPS spectra provided evidence for the presence of Co, O, and C. For the regional Co 2p spectrum, two major peaks at 780.0 and 795.0 eV were observed, which were correlated to the Co 2p 3/2 and Co 2p 1/2 spin-orbit peaks of Co 3 O 4 , respectively. In addition, two shakeup satellites, which were characteristic of Co 3 O 4 , were clearly observed at 789.9 and 804.3 eV [38]. The high-resolution spectrum of O 1s could be deconvoluted to three subpeaks (Fig. 7d). Peaks cen-  [22,39].
To determine the optimum pyrolysis temperature for OER, the flower-like ZIF-67 was pyrolyzed at different temperatures (350, 450, 550, and 650°C). The electrochemical activities of M-350, M-450, M-550, and M-650 for OER were tested in O 2 -saturated 1.0 M KOH solution. The over-potential at a current density of 10 mA cm -2 is an important metric related to solar fuel synthesis. As shown in the LSV curves (Fig. 8a), M-350, M-450, and M-550 showed comparative catalytic activity, and the overpotentials at a current density of 10 mA cm -2 were 290, 310, and 302 mV, respectively. M-650 displayed a relatively poor catalytic activity with a high over-potential (*370 mV). However, the Tafel slopes revealed the opposite tendency. Tafel plots were established based on the LSV curves (Fig. 8b). The Tafel slope b is a parameter that describes the kinetics of the electrocatalyst for OER, which is determined by the Tafel equation: where g refers to the over-potential, b represents the Tafel slope, and the current density is indicated by J. A smaller value of b implies a faster increase in the rate of the OER as applied to an increase in the potential. The Tafel slope values for M-550 and M-650 were 83 and 79 mA dec -1 , much smaller than those of M-350 (*121 mA dec -1 ) and M-450 (*105 mA dec -1 ). In order to explain these results, the composition and structure analysis was performed by powder XRD. As shown in Fig. S4, the intensity of the diffraction peaks of Co 3 O 4 increased with the pyrolysis temperature, indicating a highly disordered structure of M-350. As the TGA results (Fig. 6) revealed that there was no obvious weight loss from the M-ZIF-67 sample at 350°C, it was reasonable to conclude that M-350 contained a high percentage of carbon. While a highly disordered structure efficiently improved the catalytic activity, the kinetics were compromised by the high carbon content.
Remarkably, M-550 performed well in both metrics. Therefore, the optimum pyrolysis temperature was chosen as 550°C. Therefore, M-ZIF-67, H-ZIF-67, and T-ZIF-67 were pyrolyzed at 550°C under air. As shown in the LSV curves ( Fig. 9a)  The reason for better electrocatalytic performance of M-Co 3 O 4 /NPC over the other two composites was attributed to its favorable structure (Fig. 11). Firstly, the M-Co 3 O 4 /NPC composite derived from the flower-like ZIF-67 was comprised of the nitrogen-doped carbon scaffold with uniformly attached Co 3 O 4 nanoparticles. The unique carbon network provided channels for the electrolyte, allowing intimate contact between the electrode and the electrolyte, hence promoting interfacial charge transfer. Besides, good electrical conductivity of the carbon scaffold likely also facilitated the electron transport. Thirdly, a highly disordered structure implied the presence of more active sites, which were key to the improvement in OER activity. Furthermore, the flower-like carbon matrix showed high structural stability, which could firmly support the Co 3 O 4 nanoparticles and thus improved the stability of the catalyst.

Conclusion
In summary, a facile method for the preparation of Co 3 O 4 / NPC composites with different morphologies has been proposed, in which Co 3 O 4 nanoparticles were uniformly embedded in a nitrogen-doped carbon scaffold. By slightly modulating the synthetic route of the ZIF-67 precursors, it   was possible to achieve control of their morphology. This facile method provided a new means to prepare MOFderived electrocatalysts for the OER. Among the three Co 3 O 4 /NPC composites, the M-Co 3 O 4 /NPC derived from the flower-like ZIF-67 displayed superior electrocatalytic activity. The excellent performance of the M-Co 3 O 4 /NPC composite was attributed to its favorable structure. Firstly, the unique carbon network allowed an intimate contact area between the electrode and the electrolyte, thus promoting interfacial charge transfer. Secondly, the highly disordered structure resulted in more active sites, which were determinant to the electrocatalytic activity for OER. Lastly, the flower-like carbon matrix assumed high structural stability, which firmly supported the Co 3 O 4 nanoparticles, thus improving the stability of the catalyst.