- Email: [email protected]
Engineering hybrid between nickel oxide and nickel cobaltate to achieve exceptionally high activity for oxygen reduction reaction
Journal of Power Sources 272 (2014) 808e815
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Engineering hybrid between nickel oxide and nickel cobaltate to achieve exceptionally high activity for oxygen reduction reaction Zhentao Cui a, Shuguang Wang a, Yihe Zhang b, *, Minhua Cao a, * a
Key Laboratory of Cluster Science, Ministry of Education of China, Beijing Key Laboratory of PhotoelectroniceElectrophotonic Conversion Materials, Department of Chemistry, Beijing Institute of Technology, Beijing 100081, PR China b Beijing Key Laboratory of Mineral Materials and Utilization of Solid Waste, National Laboratory of Mineral Materials, School of Materials Sciences and Technology, China University of Geosciences, Beijing 100083, PR China
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Porous NiO/NiCo2O4 nanotubes were prepared by coaxial electrospinningassisted method. They exhibit significantly enhanced electrocatalytic activity for ORR. The good ORR performance may be attributed to their unique microstructures.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 18 May 2014 Received in revised form 28 July 2014 Accepted 24 August 2014 Available online 2 September 2014
The porous NiO/NiCo2O4 nanotubes are prepared via a coaxial electrospinning technique followed by an annealing treatment. The resultant NiO/NiCo2O4 hybrid is developed as a highly efficient electrocatalyst, which exhibits significantly enhanced electrocatalytic activity, long-term operation stability, and tolerance to crossover effect compared to NiO nanofibers, NiCo2O4 nanofibers and commercial Pt(20%)/C for oxygen reduction reactions (ORR) in alkaline environment. The excellent electrocatalytic performance may be attributed to the unique microstructures of the porous NiO/NiCo2O4 nanotubes, such as heterogeneous hybrid structure, open porous tubular structure, and the well dispersity of the two components. Moreover, the promising and straightforward coaxial electrospinning proves itself to be an efficient pathway for the preparation of nanomaterials with tubular architectures and it can be used for large-scale production of catalysts in fuel cells. © 2014 Elsevier B.V. All rights reserved.
Keywords: Nickel oxide Nickel cobaltate Nanotubes Hybrid Oxygen reduction reaction Fuel cells
1. Introduction Polymer electrolyte membrane fuel cells (PEMFCs) have been considered as a highly promising power source in future energy systems. However, the common obstacles limiting their broad applications are the high cost and the vulnerability toward * Corresponding authors. E-mail addresses: [email protected] (Y. Zhang), [email protected] (M. Cao). http://dx.doi.org/10.1016/j.jpowsour.2014.08.097 0378-7753/© 2014 Elsevier B.V. All rights reserved.
reaction poisons of the Pt-based catalysts as well as inherently sluggish kinetics of the oxygen reduction (ORR) on the cathode [1e4]. It is therefore of great significance to develop alternative inexpensive catalysts with high activity and wide availability, such as Pt-based alloys [5,6] and common metals or metal oxides [7,8]. Recently, experimental and theoretical studies showed that mixed metal oxides especially cobalt- and manganese-based oxides are an important class of promising electrocatalysts towards the ORR in alkaline solutions [9]. In contrast to their single
Z. Cui et al. / Journal of Power Sources 272 (2014) 808e815
809
component, these oxides as electrocatalysts feature a number of advantages, such as high activity, easy availability, low cost, thermodynamic stability, and low electrical resistance. Recently, Chen et al. synthesized spinel-structured CoxMn3xO4 nanocrystals, which exhibit considerably high catalytic activity towards ORR as a result of their high surface area and abundant defects. More specifically, the CoxMn3xO4 nano-crystals have a specific current density of 43.2 mA mg1 and a transferred electron number of 3.7, both of which are comparable to those of the Pt/C catalyst (43.9 mA mg1, 3.9) [10]. Furthermore, hybridizing two different components is an available route for achieving significantly enhanced electrochemical performance. Yang and coworkers reported that Co3O4/Co2MnO4 nanocomposites exhibit high activity for ORR as a synergistic catalyst, which can be attributed to their large surface area and welldispersed two-phase heterogeneous structure [11]. In fact, the hybrid materials with a heterojunction, i.e. an interface that occurs between two different types of semiconductors, have been shown to give rise to superior performance in catalysis, electrochemistry, and magnetism when compared to each component. For example, Co3O4 nanowire/MnO2 core/shell nanobelt arrays exhibited significantly enhanced capacity and rate capability, which can be ascribed to the synergistic effect between Co3O4 and MnO2 [12]. However, the synthesis procedure for achieving the hybridization of two components mentioned above, still suffers from disadvantages such as complicated production process and no universal applicability. Therefore, developing a simple, yet high yield strategy to fabricate hybrid materials with a control composition is highly desirable but still very challenging. On the other hand, recent research on nanostructured materials has demonstrated that their oxygen reduction performance could be significantly improved by manipulating the microstructures of the materials [13,14]. For example, porous structure of catalysts may decrease the mass transport resistance and allow the access of the reactant species to the active surface sites easier, which is greatly beneficial for the ORR process. This fact has got solid support from the work reported by Srinivasan et al., in which porous NiCo2O4 nanotubes (NCO-NTs), nanofibers (NCO-NFs) and nanobelts (NCO-NBs) were prepared. The BET surface area and total pore volume of the NCO-NTs sample are around 36.9 m2 g1 and 0.22 cm2 g1, respectively, which are much larger than those of NCO-NFs (about 16.1 m2 g1 and 0.06 cm2 g1) and NCO-NBs (about 12.7 m2 g1 and 0.05 cm2 g1). As a consequent, the energy density of NCO-NTs is about 1.6 and 2.0 times those of NCONFs and NCO-NBs, respectively [13]. Herein, we demonstrate the preparation of porous NiO/NiCo2O4 nanotubes via a coaxial electrospinning technique followed an annealing treatment. The resultant NiO/NiCo2O4 hybrid displays a unique porous tubular structure, and that features homogeneous interface/chemical distributions of the two single components (NiO and NiCo2O4) at the nanoscale. The porous NiO/NiCo2O4 nanotubes exhibit enhanced activities as an electrocatalyst towards ORR in alkaline solution compared to NiO nanofibers, NiCo2O4 nanofibers and commercial Pt (20%)/C (20 wt.% platinum on carbon, Johnson Matthey). Although the onset potential for the NiO/NiCo2O4 sample is close to that of the Pt(20%)/C catalyst, its current density is higher than that of the Pt(20%)/C catalyst and that its durability is also superior to that of the Pt(20%)/C catalyst.
purchased from Aldrich and cobaltous acetate [Co(Ac)2$6H2O], nickelous acetate [Ni(Ac)2$6H2O], and N,N-dimethylformamide (DMF) were acquired from Sinopharm Chemical Reagent Co., Ltd.
2. Experimental
2.3. Electrochemical measurements
2.1. Materials and methods
The electrochemical measurement of cyclic voltammetry (CV) and rotating disk electrode (RDE) were carried out in a threeelectrode system in 0.1 M KOH at 25 C using a CHI760E electrochemical analyzer. A glass carbon RDE with a diameter of 5 mm
All chemicals used were analytical grade without further purification. Poly(vinylpyrrolidone) (PVP, K90, Mw ¼ 1,300,000) was
2.1.1. Preparation of porous NiO/NiCo2O4 nanotubes The porous NiO/NiCo2O4 nanotubes were fabricated by coaxial electrospinning followed by annealing treatment. In a typical procedure, a solution containing Co(Ac)2$6H2O, Ni(Ac)2$6H2O and PVP was used as the precursor for the outer channel of the nozzle during the coaxial electrospinning process. This solution was prepared by dissolving Co(Ac)2$6H2O (0.4 g), Ni(Ac)2$6H2O (0.4 g) and PVP powder (1.5 g) in DMF (10 g) by vigorous stirring at room temperature for at least 6 h. In a similar way, the inner solution was prepared by dissolving 15 wt.% of PVP in DMF followed by vigorous stirring for 6 h. The inner solution was loaded into the syringe connected to the inner channel of a dual nozzle, which consists of two stainless-steel tubes with the diameters of 1.3 mm (outer) and 0.3 mm (inner). The outer solution in the other syringe was connected to the outer channel of the same dual-nozzle through a Teflon tube. The feeding rates of outer and inner solution are 0.6 and 0.1 mL h1, respectively. At room temperature a high voltage of 12e15 kV was applied to the dual-nozzle, and the distance between coaxial spinneret tip and collector (aluminum foil) was 15 cm. The as-electrospun nanofibers were peeled off from the collector and then they were transferred into an alumina tube furnace for stabilization. The stabilized sample was annealed at 600 C in air atmosphere for 3 h at a heating rate of 1 C min1 to yield final porous NiO/NiCo2O4 nanotubes. 2.1.2. Preparation of NiCo2O4 nanofibers and NiO nanofibers The NiCo2O4 nanofibers were preparation in a similar way. Briefly, a solution containing stoichiometric Co(Ac)2$6H2O (0.70 g), Ni(Ac)2$6H2O (0.35 g) and PVP (1.5 g) in DMF (10 g) was used as outer precursor solution, and 15 wt.% of PVP in DMF was used as inner precursor solution. The feeding rates of outer and inner solutions are 0.6 and 0.1 mL h1 with a high voltage of 12e15 kV and 15 cm distance. The annealing process is similar to that for the porous NiO/NiCo2O4 nanotubes, but the calcination temperature is 450 C. NiO nanofibers were fabricated by the same way to the above, just using only one metal salt [Ni(Ac)2$6H2O]. 2.2. Characterizations The composition and phase purity of the as-synthesized samples were analyzed by powder X-ray diffraction (XRD) with Cu Ka (l ¼ 1.54178 Å) incident radiation by a Shimadzu XRD-6000 operated at 40 kV voltage and 50 mA current. The size and morphology of the resulting products were studied by a H-8100 transmission electron microscopy (TEM) operating at 200 kV accelerating voltage. The field emission scanning electron microscopy (FE-SEM) of the sample was taken on Hitachi S-4800 SEM unit. X-ray photoelectron spectroscopy (XPS) data were recorded with an ESCALAB 250 electron spectrometer using Al K irradiation. The BrunauereEmmetteTeller (BET) surface area of the as-synthesized samples was measured using a Belsorp-max surface area detecting instrument by N2 physisorption at 77 K. The thermogravimetry (TG) analysis of the sample was carried out with a DTG-60AH instrument with a heating rate of 20 C min1 from 25 C to 600 C in the air.
810
Z. Cui et al. / Journal of Power Sources 272 (2014) 808e815
after loading the catalyst was used as the working electrode, a Ag/ AgCl (KCl, 3 M) electrode as the reference electrode, and a Pt foil (1 cm2) as the counter electrode. The glassy carbon disk was rinsed with double distilled water and dried at room temperature before the catalyst layer was applied to the disk. The thin film electrode was prepared as follows: 5 mg of catalyst was added into 1.0 mL of ethanol and 40 mL of 5 wt.% Nafion (Dupont) and dispersed by ultrasonication for approximately 30 min to obtain a homogeneous suspension. Next, 30 mL of the dispersion was uniformly dropped onto a freshly polished glassy carbon electrode (3 mm in diameter) and was dried under ambient conditions. By using the same electrode configuration, GCs and commercial Pt(20%)/C catalyst (20 wt.% platinum on carbon black) with the similar amount were also studied for comparison. The electrochemical experiments were carried out in O2 saturated 0.1 M KOH electrolyte for the ORR. The potential range is cyclically scanned between 0.8 and þ0.2 V at a scan rate of 10 mV s1. Cyclic voltammetry curves were recorded by applying a linear potential scan at a sweep rate of 10 mV s1 between 0.8 and 0.2 V after purging O2 or N2 gas for 30 min. The cycling was repeated until a reproducible cyclic voltammetry curve was obtained before the measurement curves were recorded. RDE measurements were conducted at different rotating speeds from 400 to 2500 rpm. Durability test of the NiO/NiCo2O4 and the commercial Pt(20%)/C (Vulcan, 20 wt.%) was also conducted with O2 continuous flow in 0.1 M KOH. All the experiments were conducted at room temperature (25 ± 1 C). 3. Results and discussion A schematic illustration for the fabrication process of the porous NiO/NiCo2O4 nanotubes is shown in Scheme 1, which gives an overview of a typical coaxial electrospinning setup. Briefly, Co(Ac)2, Ni(Ac)2 and PVP with a weight ratio of 1:1:4 were dissolved in DMF under vigorous stirring to form a viscous liquid. Then the viscous liquid was injected into the outer channel of the nozzle and the PVP dissolved in DMF (PVP:DMF ¼ 3:2, weight) into the inner channel (Scheme 1A). The flow rates of the two different liquids were controlled by two separate syringe pumps. After electrospinning process, the obtained composite nanofibers [Co(Ac)2/Ni(Ac)2/PVP] were annealed at different temperatures (Scheme 1B). Detailed experimental procedure is described in the experimental section. In order to obtain optimal calcination parameters, thermogravimetric analysis and differential thermal analysis (TGA/DTA) were performed on the Co(Ac)2/Ni(Ac)2/PVP nanofibers between room temperature and 600 C in air at a heating rate 10 C min1. As shown in Fig. 1a, the TGA curve can be divided into three stages. The first 19% weight loss (before 200 C) can be attributed to the loss of absorbed moisture and crystal water from metal acetates. The following 29% weight loss between 200 and 340 C can be assigned to the decomposition of metal acetates. When the temperature further increases, a prominent exothermic peak at 380 C in the DTA curve is detected, corresponding to the decomposition of PVP, and therefore there is 43% weight loss in the range of 340e400 C. It is worth noting that no weight loss was found when the temperature is above 400 C, indicating that the metal precursor and PVP polymer can be completely decomposed at the temperature higher than 400 C. According to this TGA result, a temperature of 600 C was chosen for thermally treating the precursor to ensure the formation of porous nanotubes. The effect of the temperature on the microstructure of final produce will be discussed in the later section. To determine the phase structure of the annealed product, powder X-ray diffraction (XRD) measurement was conducted. As shown in Fig. 1b, all diffraction peaks match well with the standard peaks of NiO phase (JCPDF card No. 65-2901) and spinel NiCo2O4
Scheme 1. Schematic illustration of the electrospinning and carbonization steps.
phase (JCPDF card No. 20-0781). This result indicates that the Co(Ac)2/Ni(Ac)2/PVP precursor has been completely transformed into one bi-component phase (NiO/NiCo2O4) at 600 C. Furthermore, the XRD patterns of NiO nanofibers and NiCo2O4 nanofibers are showed in Fig. 1c, d and they both correspond well to respective JCPDF cards. The morphology and structure of the precursor (Co(Ac)2/ Ni(Ac)2/PVP) and its annealed product (NiO/NiCo2O4) were investigated by field-emission scanning electron microscopy (FESEM). As shown in Fig. 2a, the Co(Ac)2/Ni(Ac)2/PVP composite nanofibers have a quite uniform and smooth surface with diameters of 500e600 nm and lengths up to several millimeters. After calcination in air at 600 C for 3 h, it is interesting to observe that the resultant NiO/NiCo2O4 sample displays a tube-like structure with open ends (Fig. 2bed) and that there are many holes on the nanotube wall with an average diameter of about 200 nm. The holes formed on the nanotube surface might be due to outward diffusion of decomposed gases (mainly carbon dioxide) during the heating process. Furthermore, compared to Co(Ac)2/Ni(Ac)2/PVP nanofibers, NiO/NiCo2O4 nanotubes display an evident shrinkage in diameter, which could be attributed to the removal of inner PVP and the decomposition of outer metal precursors. From the highmagnification FESEM image (Fig. 2d) we can clearly seen that these nanotubes with numerous nanopores in the tube walls are composed of interconnected primary nanocrystals with an average diameter of about 30 nm and the nanocrystals spontaneously align
Z. Cui et al. / Journal of Power Sources 272 (2014) 808e815
811
Fig. 1. (a) TGA/DTA curve of as-electrospun Co(Ac)2/Ni(Ac)2/PVP composite nanofibers. XRD patterns of porous NiO/NiCo2O4 nanotubes (b), NiO nanofibers (c) and NiCo2O4 nanofibers (d).
with one another to form a porous tube with an interior cavity and a loose tube wall. This porous network is believed to facilitate the inward and outward gas diffusion and thus it also is expected to have an
enhanced electrochemical performance. On the other hand, a possible formation mechanism of the porous nanotubes is proposed as follows: during heating process, the diffusion of gas produced by the inner PVP decomposition leads to the swelling of
Fig. 2. Typical FESEM images at different magnifications. (a) As-electrospun Co(Ac)2/Ni(Ac)2/PVP composite nanofibers and (bed) Porous NiO/NiCo2O4 nanotubes obtained after calcining the Co(Ac)2/Ni(Ac)2/PVP composite nanofibers at 600 C for 3 h.
812
Z. Cui et al. / Journal of Power Sources 272 (2014) 808e815
composite fibers forming hollow structures, whereas the shrinkage of composite fibers originating from the outer metal acetate decomposition results in the formation of porous tube wall. It should be noted that this kind of porosity on the tube wall in our case closely depends on the thermal treatment conditions. Fig. S1 shows the FESEM images of the Co(Ac)2/Ni(Ac)2/PVP composite nanofibers after different calcination temperatures. If the thermal treatment temperature is 500 C, solid, compact nanofibers with uniform diameters are obtained. When the as-electrospun Co(Ac)2/Ni(Ac)2/PVP nanofibers were calcined at 700 C in air for 3 h, nanotubes were formed, but many nanotubes collapsed (Fig. S1ced). Therefore, the most suitable temperature for the formation of well-defined porous nanotubes was 600 C. In addition, FESEM images of NiO nanofibers and NiCo2O4 nanofibers were shown in Fig. S2. As can be seen, they both have quite smooth surface without any pores over the surface. To gain further insight into the morphology and microstructure of the porous NiO/NiCo2O4 nanotubes, TEM analysis was performed. Fig. 3aec shows representative bright-field TEM images for the NiO/NiCo2O4 nanotubes, which clearly reveal welldefined tubular morphology with an average diameter of approximately 200 nm and wall thicknesses of about 50 nm. The high resolution TEM (HRTEM) images further confirm the high crystallinity of the porous NiO/NiCo2O4 nanotubes (Fig. 3d, e). Furthermore, the measured lattice spacing of 0.47 nm is in good
agreement with the (111) interplanar distance of NiCo2O4 phase, and that of 0.24 nm fits well with the (111) interplanar distance of NiO phase. The selected-area electron diffraction (SAED) pattern indicated the polycrystalline nature of the nanotubes (Fig. 3f), consistent with above HRTEM result. The scanning transmission electron microscopy (STEM) and corresponding elemental mapping images of the NiO/NiCo2O4 nanotubes reveal that Co, Ni and O are homogeneously distributed in each individual nanotube (Fig. 3g). In addition, the energy dispersive X-ray spectroscopy (EDX) spectrum shown in Fig. 3h indicates that the atomic ratio of Ni to Co in the resultant NiO/NiCo2O4 nanotubes is around 1:1.3, indicating that the mass content of NiO in the NiO/NiCo2O4 hybrid is 35%. To further examine the porous properties, N2 adsorption/ desorption isotherm measurements of the porous NiO/NiCo2O4 nanotubes are performed, as depicted in Fig. 3i. The nitrogen sorption isotherms can be classified as type IV according to IUPAC classification and the H3-type hysteresis loop reveals the mesoporous structure nature. The pore size of the mesoporous structure is determined to be around 12 nm by the BarretteJoynereHalenda (BJH) method from de desorption branch (the inset in Fig. 3i). The BET surface area is calculated to be around 28.60 m2 g1. This unique porous structure of the NiO/NiCo2O4 hybrid is expected to be an ideal design for fuel cell catalyst and has huge potential to further improve its electrochemical performance because the
Fig. 3. (aec) Typical TEM images of porous NiO/NiCo2O4 nanotubes at different magnifications. (d, e) HRTEM images recorded on different regions of single nanotube. (f) SAED pattern of NiO/NiCo2O4 nanotubes. (g) STEM and corresponding elemental mapping images, (h) EDS spectrum of the porous NiO/NiCo2O4 nanotubes. (i) Nitrogen adsorption/ desorption isotherms of NiO/NiCo2O4nanotubes and inset is pore size distribution.
Z. Cui et al. / Journal of Power Sources 272 (2014) 808e815
Like other manganese-based spinel compounds, NiCo2O4 also exhibits excellent electrochemical properties [13,14,19]. With its unique microstructures including hybridization and porous nanotube structure, the as-prepared NiO/NiCo2O4 sample is expected to display significantly enhanced electrochemical properties. The electrocatalytic properties of the NiO/NiCo2O4 nanotubes were first evaluated for ORR through conventional three-electrode cyclic voltammetry (CV) in N2 or O2 saturated 0.1 M KOH aqueous solutions. As shown in Fig. 5a, there was no significant peak obtained within the potential range between 0.8 and 0.2 V in the N2saturated solution. On the contrary, when oxygen was introduced, the electrode showed a well-defined cathodic peak appearing at 0.16 V (vs Ag/AgCl) at the same potential range, which can be attributed to electrocatalytic reduction of oxygen by the porous NiO/NiCo2O4 nanotube catalyst. This ORR potential is very close to that of the commercial Pt(20%)/C catalyst (Fig. 5b), and more positive than those of NiCo2O4 nanowires and macroporous NiCo2O4 sheets, [14,19] indicating a pronounced electrocatalytic activity of the NiO/NiCo2O4 sample for ORR. Furthermore, possible crossover effect of the NiO/NiCo2O4 catalyst is also investigated because it is one of important parameters for electrochemical evaluation of electrocatalysts for practical fuel cell applications. Therefore, the CV responses to methanol introduced into the O2-saturated electrolyte were measured for both NiO/NiCo2O4 nanotubes and commercial Pt(20%)/C catalyst. It can be seen from Fig. 5a that after the addition of 3 M methanol, NiO/NiCo2O4 nanotube catalyst retained highly stable current response. In contrast, the CV curve of the Pt(20%)/C electrode shows a strong methanol oxidation peak at around 0.13 V (Fig. 5b). These results indicate that the NiO/ NiCo2O4 nanotube catalyst has a good ability for avoiding crossover effects in the methanol-containing electrolyte. To further prove the role of the NiO/NiCo2O4 nanotubes in the ORR electrochemical process, linear sweep voltammograms (LSVs) on a rotating disk electrode (RDE) were recorded at different rotating speeds from
unique structures are crucial in facilitating the transfer of ions and electrons at the electrode/electrolyte interface. The surface electronic state and the composition of the assynthesized porous NiO/NiCo2O4 nanotubes were analyzed by Xray photoelectron spectroscopy (XPS). As shown in Fig. 4a, the survey XPS spectrum revealed the presence of Ni, Co, O, and C elements. Three typical signals of the O1s, Co2p and Ni2p core levels are detected from the XPS profiles shown in Fig. 4bed. In the Co2p spectra (Fig. 4b), two kinds of Co species (Co2þ and Co3þ) were detected. The binding energies at 779.3 and 794.3 eV can be ascribed to Co3þ, while another two fitting peaks at 781.0 and 795.7 eV can be attributed to Co2þ [15]. In the Ni 2p spectra (Fig. 4c), two kinds of nickel species containing Ni2þ and Ni3þ can also be observed. The fitting peaks at 854.0 and 871.7 eV can be indexed to Ni2þ, while the fitting peaks at 855.9 and 873.8 eV can be indexed to Ni3þ [16]. The satellite peaks at around 861.0 and 879.4 eV are two shake-up type peaks of nickel at the high binding energy side of the Ni2p3/2 and Ni2p1/2 edge [17]. The O 1s spectra (Fig. 4d) show a strong MeOeM peak at 529.48 eV, which is indicative of the lattice oxygen. The other two peaks at 532.83 and 531.15 eV can be attributed to CeO and OeC]O bonds, respectively, corresponding to a number of defect sites with low oxygen coordination in the materials with small particle size [18], which may be due to the multiplicity of physically and chemically adsorbed water at and within the surface [15]. Above result demonstrates that the surface of the as-prepared NiO/NiCo2O4 nanotubes has a composition containing Co2þ, Co3þ, Ni2þ and Ni3þ. Moreover, the calculated Co:Ni atomic ratio is 1.3:1, which is far low than that of stoichiometric NiCo2O4 (2:1), indicating that excessive NiO exists on the surface of the NiCo2O4. This result is well agreement with above XRD result. Here, the redox couples of Co2þ/Co3þ and Ni2þ/ Ni3þcould afford enough active sites for oxygen reduction, which may be one of the important factors contributing to the high electrocatalytic performance of NiO/NiCo2O4 hybrid.
(a) Ni2p
(b)
Ni1s
Intensity (a.u.)
Intensity (a.u.)
Co2p O(Auger) O1s Co(Auger)
C1s
Ni3p Ni3s
200
Ni3+
Co
800
790
780
Binding Energy (eV)
770
M-O-M
Ni2+
2+
Ni
Co3+ 3+
Co2+
Ni 2p3/2
Ni 2p1/2
Co 2p1/2 Co2+
Sat.
(d) Sat.
Intensity (a.u.)
Sat.
400 600 800 1000 1200 810 Binding Energy (eV)
(c)
890
Co 2p3/2
Sat. Ni3+
880 870 860 Binding Energy (eV)
Intensity (a.u.)
0
813
850
O-C=O C=O
540
535 530 525 Binding Energy (eV)
Fig. 4. (a) XPS spectra for the porous NiO/NiCo2O4 nanotubes: (a) Full-scan and high-resolution XPS spectra of (b) Co 2p, (c) Ni 2p, and (d) O 1s.
Z. Cui et al. / Journal of Power Sources 272 (2014) 808e815 2
(a)
O2 saturated N2 saturated O2 saturated in 3M CH3OH
1 0
-0.16 V
-1
10
Current density/mA cm-2
Current density/mA cm-2
814
8 6
-0.8
-0.6
-0.4
-0.2
0.0
0.2
Potential/V vs. Ag/AgCl
(b)
N2 saturated
4 2 0
-2
-2
O2 saturated O2 saturated in 3M CH3OH
-0.8
-0.6
-0.4
-0.2
0.0
0.2
Potential/V vs. Ag/AgCl
Fig. 5. CVs of the porous NiO/NiCo2O4 nanotubes (a)Pt(20%)/C (b) on a glassy-carbon RDE electrode in N2-saturated, O2-saturated 0.1 M KOH, and O2-saturated 0.1 M KOH and 3 M CH3OH solutions. (c) Rotating disk electrode voltammograms of the porous NiO/NiCo2O4 nanotubes in O2-saturated 0.1 M KOH with different electrode rotation rates. (d) The KouteckyeLevich plots derived from the rotating disk electrode measurements. (e) Rotating disk electrode voltammograms of the porous NiO/NiCo2O4 nanotubes, NiO nanofibers, NiCo2O4 nanofibers and Pt(20%)/C in O2-saturated 0.1 M KOH with a scan rate of 10 mV s1 at 1600 rpm. (f) The chronoamperometric curves of the porous NiO/NiCo2O4 nanotubes and Pt(20%)/C in O2-saturated 0.1 M KOH.
400 to 2500 rpm in 0.1 M KOH electrolyte saturated with O2 (Fig. 5c). It can be clearly seen that the current evidently increases with rotation rate due to the faster oxygen flux to the electrode surface [20,21]. The onset potential was about 0.05 V, very close to that (0.06 V) identified from CV curve (Fig. 5a). The KoutechyeLevich plots (J1 vs u1/2) of the NiO/NiCo2O4 catalyst are obtained from LSVs for different potentials (Fig. 5d) and all plots show fairly good linearity, suggesting first-order reaction kinetics toward dissolved oxygen and similar electron transfer number for the ORR process at different potentials. The transferred electron number (n) per oxygen molecule involved in this ORR process could be determined on the basis of KoutechyeLevich equations (Equation S1) and the KoutechyeLevich plots (J1 vs u1/2). Then, in each case, the n value is calculated to be approximately 3.9 at the potentials ranging from 0.40 to 0.70 V, indicating that the ORR process proceeds via an approximate four-electron pathway [22,23]. To gain further insight into ORR of each sample, LSV measurements were performed on RDE for the porous NiO/NiCo2O4 nanotubes, the NiO nanofibers, NiCo2O4 nanofibers and Pt(20%)/C in O2-saturated 0.1 M KOH with a scan rate of 10 mV s1 at 1600 rpm. From Fig. 5e, it is evident that the NiO/NiCo2O4 sample showed significantly better performance than NiO and NiCo2O4 samples in both onset potential and reaction current density. These results are in agreement with above CV observations (Fig. S3). Furthermore, we can also observe that the onset potential for the NiO/NiCo2O4 sample is close to that of the Pt(20%)/C catalyst, and that its current density is higher than that of the Pt(20%)/C catalyst, suggesting the fact that the ORR catalytic activity of the NiO/ NiCo2O4 catalyst is higher than that of the Pt(20%)/C catalysts. This performance of the NiO/NiCo2O4 sample is also better than those of the reported NiCo2O4-based materials with different morphologies (Table S1). To further assess the ORR reaction mechanism of the NiO/ NiCo2O4 sample, we analyzed its Tafel plots. As shown in Fig. S4, the NiO/NiCo2O4 sample had a Tafel slope of 61.2 mV per decade, slightly smaller than that (68.1 mV) of the Pt(20%)/C per decade, indicating that the ORR kinetics of the NiO/NiCo2O4 sample is similar to that of the Pt(20%)/C. This result further suggests that
the ORR process proceeds via an approximate four-electron pathway. The durability of the NiO/NiCo2O4 and Pt(20%)/C was also assessed through chronoamperometric measurements at 0.2 V in O2-saturated 0.1 M KOH solution for 40,000 s (Fig. 5f). Evidently, a high relative current of 87% can be retained even after 40,000 s for the NiO/NiCo2O4 catalyst. In contrast, the Pt(20%)/C catalyst showed a gradual decrease with a current loss of approximately 30% after 40,000 s. This observation strongly suggests that the durability of the NiO/NiCo2O4 catalyst is superior to that of the Pt(20%)/C catalyst. From the view of a practical application, this feature is particularly important. From the above results, we can see that the electrochemical performance of the as-synthesized NiO/NiCo2O4 hybrid is not only better than those of each component alone, NiO or NiCo2O4, but also comparable to that of the commercial Pt(20%)/C. Combining with reported studies, we believe that this may benefit from its unique microstructures, which can be summarized into following several aspects. First, the heterogeneous hybrid structure of NiO/ NiCo2O4 could ensure fast transport kinetics, which has been confirmed by electrochemical impedance spectroscopy (EIS) measurements performed from 100 KHz to 0.01 Hz. As shown in Fig. 6, the three Nyquist plots all exhibit characteristic of one semicircle in the high frequency range and one inclined line in the low frequency range. According to the Nyquist plots, the charge transfer resistance Rct of the NiO/NiCo2O4 hybrid electrode is determined to be 108.7 U, which is evidently lower than those of the other two electrodes (125.8 U for NiCo2O4 and 152.7 U for NiO). This demonstrates an impressive low charge transfer resistance for the NiO/NiCo2O4 hybrid, which is also common in other hybrids [11]. This feature has been widely confirmed to be mainly responsible for the significant improvement of the performance [11]. Just recently we have successfully constructed mesoporous ZnFe2O4/a-Fe2O3 microoctahedron hybrid, which exhibits superior lithium storage performance [24]. Second, the open porous tubular structure could build short diffusion channels between O2 and electrolyte, which consequentially ensure fast and uniform oxygen and electrolyte distribution inside the electrode. Third, the porous and tubular
Z. Cui et al. / Journal of Power Sources 272 (2014) 808e815
815
tubular architectures and it can be used for large-scale production of catalysts in fuel cells. Acknowledgments This work was supported by the National Natural Science Foundation of China (21471016, and 21271023), the National HighTech Research and Development Program of China (863 Program, 2012AA06A109) and the 111 Project (B07012). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jpowsour.2014.08.097. References Fig. 6. Nyquist plots of the porous NiO/NiCo2O4 nanotubes, NiO nanofibers, and NiCo2O4 nanofibers.
structure could also provide a relatively large electrodeeelectrolyte contact area to ensure high availability of the catalytic active sites. In a word, the unique composition and microstructure of the porous NiO/NiCo2O4 nanotube hybrid are responsible for its excellent electrochemical performance. 4. Conclusion In summary, we have constructed the NiO/NiCo2O4 hybrid with a porous tubular structure by a coaxial electrospinning method followed by a fine annealing treatment. As a potential non-noble metal ORR electro-catalyst for fuel cells, the porous NiO/NiCo2O4 nanotubes exhibit enhanced electrocatalytic activity, long-term operation stability, and tolerance to crossover effect compared to NiO nanofibers, NiCo2O4 nanofibers and commercial Pt(20%)/C via a four-electron pathway in alkaline environment. The onset potential was about 0.05 V and the current density of NiO/NiCo2O4 is about 115% of Pt/C current density. A high relative current of 87% can be retained even after 40,000 s for the NiO/NiCo2O4 catalyst. The unique microstructures of the porous NiO/NiCo2O4 nanotubes with BET surface area of 28.60 m2 g1, such as heterogeneous hybrid structure, open porous tubular structure, and the well dispersity of the two components, are responsible for their significantly improved electrocatalytic activity for ORR. Moreover, the promising and straightforward coaxial electrospinning has proved itself to be an efficient pathway for the preparation of nanomaterials with
[1] M. Winter, R.J. Brodd, Chem. Rev. 104 (2004) 4245e4269. [2] H. Zhang, P.K. Shen, Chem. Rev. 112 (2012) 2780e2832. [3] H.J. Tang, H.J. Yin, J.Y. Wang, N.L. Yang, D. Wang, Z.Y. Tang, Angew. Chem. 125 (2013) 5695e5699. [4] M.K. Debe, Nature 486 (2012) 43e51. [5] V.R. Stamenkovic, B.S. Mun, M. Arenz, K.J.J. Mayrhofer, C.A. Lucas, G. Wang, P.N. Ross, N.M. Markovic, Nat. Mater. 6 (2007) 241e247. [6] C. Wang, D.V. Vliet, K.L. More, N.J. Zaluzec, S. Peng, S.H. Sun, H. Daimon, G.F. Wang, J. Greeley, J. Pearson, A.P. Paulikas, G. Karapetrov, D. Strmcnik, N.M. Markovic, V.R. Stamenkovic, Nano Lett. 11 (2011) 919e926. [7] J.H. Kim, A. Ishihara, S. Mitsushima, N. Kamiya, K.I. Ota, Electrochim. Acta 52 (2007) 2492e2497. [8] V.I. Zaikovskii, K.S. Nagabhushana, V.V. Kriventsov, K.N. Loponov, S.V. Cherepanova, R.I. Kvon, H. Bonnemann, D.I. Kochubey, E.R. Savinova, J. Phys. Chem. B 110 (2006) 6881e6890. [9] G. Du, X. Liu, Y. Zong, T.A. Hor, A. Yuc, Z. Liu, Nanoscale 5 (2013) 4657e4661. [10] F.Y. Cheng, J. Shen, B. Peng, Y.D. Pan, J. Chen, Nat. Chem. 79 (2011) 79e84. [11] D.D. Wang, X. Chen, D.G. Evans, W.S. Yang, Nanoscale 5 (2013) 5312e5315. [12] J.P. Liu, J. Jiang, C.W. Cheng, H.X. Li, J.X. Zhang, H. Gong, H.J. Fan, Adv. Mater. 23 (2011) 2076e2081. [13] L.L. Li, S.J. Peng, Y.L. Cheah, P. Teh, J. Wang, G. Wee, Y. Ko, C. Wong, M. Srinivasan, Chem. Eur. J. 19 (2013) 5892e5898. [14] Y. Xiao, C.G. Hu, L.T. Qu, C.W. Hu, M.H. Cao, Chem. Eur. J. 42 (2013) 14271e14278. [15] J.F. Marco, J.R. Gancedo, M. Gracia, J.L. Gautier, E. Rios, J. Solid State Chem. 153 (2000) 74e81. [16] J.G. Kim, D.L. Pugmire, D. Battaglia, M.A. Langell, Appl. Surf. Sci. 165 (2000) 70e84. [17] J. Xiao, S. Yang, J. Mater. Chem. 22 (2012) 12253e12262. nez, A. Fern s, A.R. Gonz [18] V.M. Jime andez, J.P. Espino alez-Elipe, J. Electron Spectrosc. Relat. Phenom. 71 (1995) 61e71. [19] C. Jin, F.L. Lu, X.C. Cao, Z.R. Yang, R.Z. Yang, J. Mater. Chem. A 1 (2013) 12170e12177. [20] K.P. Gong, F. Du, Z.H. Xia, M. Durstock, L.M. Dai, Science 23 (2009) 760e764. [21] R.L. Liu, D.Q. Wu, X.L. Feng, K. Mullen, Angew. Chem. Int. Ed. 49 (2010) 2565e2569. [22] Y. Li, Y. Zhao, H.H. Cheng, Y. Hu, G.Q. Shi, L.M. Dai, L.T. Qu, J. Am. Chem. Soc. 134 (2012) 15e18. [23] S.Y. Wang, D.S. Yu, L.M. Dai, J. Am. Chem. Soc. 133 (2011) 5182e5185. [24] D. Zhao, Y. Xiao, X. Wang, Q. Gao, M.H. Cao, Nano Energy 7 (2014) 124e133.
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Engineering hybrid between nickel oxide and nickel cobaltate to achieve exceptionally high activity for oxygen reduction reaction Zhentao Cui a, Shuguang Wang a, Yihe Zhang b, *, Minhua Cao a, * a
Key Laboratory of Cluster Science, Ministry of Education of China, Beijing Key Laboratory of PhotoelectroniceElectrophotonic Conversion Materials, Department of Chemistry, Beijing Institute of Technology, Beijing 100081, PR China b Beijing Key Laboratory of Mineral Materials and Utilization of Solid Waste, National Laboratory of Mineral Materials, School of Materials Sciences and Technology, China University of Geosciences, Beijing 100083, PR China
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Porous NiO/NiCo2O4 nanotubes were prepared by coaxial electrospinningassisted method. They exhibit significantly enhanced electrocatalytic activity for ORR. The good ORR performance may be attributed to their unique microstructures.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 18 May 2014 Received in revised form 28 July 2014 Accepted 24 August 2014 Available online 2 September 2014
The porous NiO/NiCo2O4 nanotubes are prepared via a coaxial electrospinning technique followed by an annealing treatment. The resultant NiO/NiCo2O4 hybrid is developed as a highly efficient electrocatalyst, which exhibits significantly enhanced electrocatalytic activity, long-term operation stability, and tolerance to crossover effect compared to NiO nanofibers, NiCo2O4 nanofibers and commercial Pt(20%)/C for oxygen reduction reactions (ORR) in alkaline environment. The excellent electrocatalytic performance may be attributed to the unique microstructures of the porous NiO/NiCo2O4 nanotubes, such as heterogeneous hybrid structure, open porous tubular structure, and the well dispersity of the two components. Moreover, the promising and straightforward coaxial electrospinning proves itself to be an efficient pathway for the preparation of nanomaterials with tubular architectures and it can be used for large-scale production of catalysts in fuel cells. © 2014 Elsevier B.V. All rights reserved.
Keywords: Nickel oxide Nickel cobaltate Nanotubes Hybrid Oxygen reduction reaction Fuel cells
1. Introduction Polymer electrolyte membrane fuel cells (PEMFCs) have been considered as a highly promising power source in future energy systems. However, the common obstacles limiting their broad applications are the high cost and the vulnerability toward * Corresponding authors. E-mail addresses: [email protected] (Y. Zhang), [email protected] (M. Cao). http://dx.doi.org/10.1016/j.jpowsour.2014.08.097 0378-7753/© 2014 Elsevier B.V. All rights reserved.
reaction poisons of the Pt-based catalysts as well as inherently sluggish kinetics of the oxygen reduction (ORR) on the cathode [1e4]. It is therefore of great significance to develop alternative inexpensive catalysts with high activity and wide availability, such as Pt-based alloys [5,6] and common metals or metal oxides [7,8]. Recently, experimental and theoretical studies showed that mixed metal oxides especially cobalt- and manganese-based oxides are an important class of promising electrocatalysts towards the ORR in alkaline solutions [9]. In contrast to their single
Z. Cui et al. / Journal of Power Sources 272 (2014) 808e815
809
component, these oxides as electrocatalysts feature a number of advantages, such as high activity, easy availability, low cost, thermodynamic stability, and low electrical resistance. Recently, Chen et al. synthesized spinel-structured CoxMn3xO4 nanocrystals, which exhibit considerably high catalytic activity towards ORR as a result of their high surface area and abundant defects. More specifically, the CoxMn3xO4 nano-crystals have a specific current density of 43.2 mA mg1 and a transferred electron number of 3.7, both of which are comparable to those of the Pt/C catalyst (43.9 mA mg1, 3.9) [10]. Furthermore, hybridizing two different components is an available route for achieving significantly enhanced electrochemical performance. Yang and coworkers reported that Co3O4/Co2MnO4 nanocomposites exhibit high activity for ORR as a synergistic catalyst, which can be attributed to their large surface area and welldispersed two-phase heterogeneous structure [11]. In fact, the hybrid materials with a heterojunction, i.e. an interface that occurs between two different types of semiconductors, have been shown to give rise to superior performance in catalysis, electrochemistry, and magnetism when compared to each component. For example, Co3O4 nanowire/MnO2 core/shell nanobelt arrays exhibited significantly enhanced capacity and rate capability, which can be ascribed to the synergistic effect between Co3O4 and MnO2 [12]. However, the synthesis procedure for achieving the hybridization of two components mentioned above, still suffers from disadvantages such as complicated production process and no universal applicability. Therefore, developing a simple, yet high yield strategy to fabricate hybrid materials with a control composition is highly desirable but still very challenging. On the other hand, recent research on nanostructured materials has demonstrated that their oxygen reduction performance could be significantly improved by manipulating the microstructures of the materials [13,14]. For example, porous structure of catalysts may decrease the mass transport resistance and allow the access of the reactant species to the active surface sites easier, which is greatly beneficial for the ORR process. This fact has got solid support from the work reported by Srinivasan et al., in which porous NiCo2O4 nanotubes (NCO-NTs), nanofibers (NCO-NFs) and nanobelts (NCO-NBs) were prepared. The BET surface area and total pore volume of the NCO-NTs sample are around 36.9 m2 g1 and 0.22 cm2 g1, respectively, which are much larger than those of NCO-NFs (about 16.1 m2 g1 and 0.06 cm2 g1) and NCO-NBs (about 12.7 m2 g1 and 0.05 cm2 g1). As a consequent, the energy density of NCO-NTs is about 1.6 and 2.0 times those of NCONFs and NCO-NBs, respectively [13]. Herein, we demonstrate the preparation of porous NiO/NiCo2O4 nanotubes via a coaxial electrospinning technique followed an annealing treatment. The resultant NiO/NiCo2O4 hybrid displays a unique porous tubular structure, and that features homogeneous interface/chemical distributions of the two single components (NiO and NiCo2O4) at the nanoscale. The porous NiO/NiCo2O4 nanotubes exhibit enhanced activities as an electrocatalyst towards ORR in alkaline solution compared to NiO nanofibers, NiCo2O4 nanofibers and commercial Pt (20%)/C (20 wt.% platinum on carbon, Johnson Matthey). Although the onset potential for the NiO/NiCo2O4 sample is close to that of the Pt(20%)/C catalyst, its current density is higher than that of the Pt(20%)/C catalyst and that its durability is also superior to that of the Pt(20%)/C catalyst.
purchased from Aldrich and cobaltous acetate [Co(Ac)2$6H2O], nickelous acetate [Ni(Ac)2$6H2O], and N,N-dimethylformamide (DMF) were acquired from Sinopharm Chemical Reagent Co., Ltd.
2. Experimental
2.3. Electrochemical measurements
2.1. Materials and methods
The electrochemical measurement of cyclic voltammetry (CV) and rotating disk electrode (RDE) were carried out in a threeelectrode system in 0.1 M KOH at 25 C using a CHI760E electrochemical analyzer. A glass carbon RDE with a diameter of 5 mm
All chemicals used were analytical grade without further purification. Poly(vinylpyrrolidone) (PVP, K90, Mw ¼ 1,300,000) was
2.1.1. Preparation of porous NiO/NiCo2O4 nanotubes The porous NiO/NiCo2O4 nanotubes were fabricated by coaxial electrospinning followed by annealing treatment. In a typical procedure, a solution containing Co(Ac)2$6H2O, Ni(Ac)2$6H2O and PVP was used as the precursor for the outer channel of the nozzle during the coaxial electrospinning process. This solution was prepared by dissolving Co(Ac)2$6H2O (0.4 g), Ni(Ac)2$6H2O (0.4 g) and PVP powder (1.5 g) in DMF (10 g) by vigorous stirring at room temperature for at least 6 h. In a similar way, the inner solution was prepared by dissolving 15 wt.% of PVP in DMF followed by vigorous stirring for 6 h. The inner solution was loaded into the syringe connected to the inner channel of a dual nozzle, which consists of two stainless-steel tubes with the diameters of 1.3 mm (outer) and 0.3 mm (inner). The outer solution in the other syringe was connected to the outer channel of the same dual-nozzle through a Teflon tube. The feeding rates of outer and inner solution are 0.6 and 0.1 mL h1, respectively. At room temperature a high voltage of 12e15 kV was applied to the dual-nozzle, and the distance between coaxial spinneret tip and collector (aluminum foil) was 15 cm. The as-electrospun nanofibers were peeled off from the collector and then they were transferred into an alumina tube furnace for stabilization. The stabilized sample was annealed at 600 C in air atmosphere for 3 h at a heating rate of 1 C min1 to yield final porous NiO/NiCo2O4 nanotubes. 2.1.2. Preparation of NiCo2O4 nanofibers and NiO nanofibers The NiCo2O4 nanofibers were preparation in a similar way. Briefly, a solution containing stoichiometric Co(Ac)2$6H2O (0.70 g), Ni(Ac)2$6H2O (0.35 g) and PVP (1.5 g) in DMF (10 g) was used as outer precursor solution, and 15 wt.% of PVP in DMF was used as inner precursor solution. The feeding rates of outer and inner solutions are 0.6 and 0.1 mL h1 with a high voltage of 12e15 kV and 15 cm distance. The annealing process is similar to that for the porous NiO/NiCo2O4 nanotubes, but the calcination temperature is 450 C. NiO nanofibers were fabricated by the same way to the above, just using only one metal salt [Ni(Ac)2$6H2O]. 2.2. Characterizations The composition and phase purity of the as-synthesized samples were analyzed by powder X-ray diffraction (XRD) with Cu Ka (l ¼ 1.54178 Å) incident radiation by a Shimadzu XRD-6000 operated at 40 kV voltage and 50 mA current. The size and morphology of the resulting products were studied by a H-8100 transmission electron microscopy (TEM) operating at 200 kV accelerating voltage. The field emission scanning electron microscopy (FE-SEM) of the sample was taken on Hitachi S-4800 SEM unit. X-ray photoelectron spectroscopy (XPS) data were recorded with an ESCALAB 250 electron spectrometer using Al K irradiation. The BrunauereEmmetteTeller (BET) surface area of the as-synthesized samples was measured using a Belsorp-max surface area detecting instrument by N2 physisorption at 77 K. The thermogravimetry (TG) analysis of the sample was carried out with a DTG-60AH instrument with a heating rate of 20 C min1 from 25 C to 600 C in the air.
810
Z. Cui et al. / Journal of Power Sources 272 (2014) 808e815
after loading the catalyst was used as the working electrode, a Ag/ AgCl (KCl, 3 M) electrode as the reference electrode, and a Pt foil (1 cm2) as the counter electrode. The glassy carbon disk was rinsed with double distilled water and dried at room temperature before the catalyst layer was applied to the disk. The thin film electrode was prepared as follows: 5 mg of catalyst was added into 1.0 mL of ethanol and 40 mL of 5 wt.% Nafion (Dupont) and dispersed by ultrasonication for approximately 30 min to obtain a homogeneous suspension. Next, 30 mL of the dispersion was uniformly dropped onto a freshly polished glassy carbon electrode (3 mm in diameter) and was dried under ambient conditions. By using the same electrode configuration, GCs and commercial Pt(20%)/C catalyst (20 wt.% platinum on carbon black) with the similar amount were also studied for comparison. The electrochemical experiments were carried out in O2 saturated 0.1 M KOH electrolyte for the ORR. The potential range is cyclically scanned between 0.8 and þ0.2 V at a scan rate of 10 mV s1. Cyclic voltammetry curves were recorded by applying a linear potential scan at a sweep rate of 10 mV s1 between 0.8 and 0.2 V after purging O2 or N2 gas for 30 min. The cycling was repeated until a reproducible cyclic voltammetry curve was obtained before the measurement curves were recorded. RDE measurements were conducted at different rotating speeds from 400 to 2500 rpm. Durability test of the NiO/NiCo2O4 and the commercial Pt(20%)/C (Vulcan, 20 wt.%) was also conducted with O2 continuous flow in 0.1 M KOH. All the experiments were conducted at room temperature (25 ± 1 C). 3. Results and discussion A schematic illustration for the fabrication process of the porous NiO/NiCo2O4 nanotubes is shown in Scheme 1, which gives an overview of a typical coaxial electrospinning setup. Briefly, Co(Ac)2, Ni(Ac)2 and PVP with a weight ratio of 1:1:4 were dissolved in DMF under vigorous stirring to form a viscous liquid. Then the viscous liquid was injected into the outer channel of the nozzle and the PVP dissolved in DMF (PVP:DMF ¼ 3:2, weight) into the inner channel (Scheme 1A). The flow rates of the two different liquids were controlled by two separate syringe pumps. After electrospinning process, the obtained composite nanofibers [Co(Ac)2/Ni(Ac)2/PVP] were annealed at different temperatures (Scheme 1B). Detailed experimental procedure is described in the experimental section. In order to obtain optimal calcination parameters, thermogravimetric analysis and differential thermal analysis (TGA/DTA) were performed on the Co(Ac)2/Ni(Ac)2/PVP nanofibers between room temperature and 600 C in air at a heating rate 10 C min1. As shown in Fig. 1a, the TGA curve can be divided into three stages. The first 19% weight loss (before 200 C) can be attributed to the loss of absorbed moisture and crystal water from metal acetates. The following 29% weight loss between 200 and 340 C can be assigned to the decomposition of metal acetates. When the temperature further increases, a prominent exothermic peak at 380 C in the DTA curve is detected, corresponding to the decomposition of PVP, and therefore there is 43% weight loss in the range of 340e400 C. It is worth noting that no weight loss was found when the temperature is above 400 C, indicating that the metal precursor and PVP polymer can be completely decomposed at the temperature higher than 400 C. According to this TGA result, a temperature of 600 C was chosen for thermally treating the precursor to ensure the formation of porous nanotubes. The effect of the temperature on the microstructure of final produce will be discussed in the later section. To determine the phase structure of the annealed product, powder X-ray diffraction (XRD) measurement was conducted. As shown in Fig. 1b, all diffraction peaks match well with the standard peaks of NiO phase (JCPDF card No. 65-2901) and spinel NiCo2O4
Scheme 1. Schematic illustration of the electrospinning and carbonization steps.
phase (JCPDF card No. 20-0781). This result indicates that the Co(Ac)2/Ni(Ac)2/PVP precursor has been completely transformed into one bi-component phase (NiO/NiCo2O4) at 600 C. Furthermore, the XRD patterns of NiO nanofibers and NiCo2O4 nanofibers are showed in Fig. 1c, d and they both correspond well to respective JCPDF cards. The morphology and structure of the precursor (Co(Ac)2/ Ni(Ac)2/PVP) and its annealed product (NiO/NiCo2O4) were investigated by field-emission scanning electron microscopy (FESEM). As shown in Fig. 2a, the Co(Ac)2/Ni(Ac)2/PVP composite nanofibers have a quite uniform and smooth surface with diameters of 500e600 nm and lengths up to several millimeters. After calcination in air at 600 C for 3 h, it is interesting to observe that the resultant NiO/NiCo2O4 sample displays a tube-like structure with open ends (Fig. 2bed) and that there are many holes on the nanotube wall with an average diameter of about 200 nm. The holes formed on the nanotube surface might be due to outward diffusion of decomposed gases (mainly carbon dioxide) during the heating process. Furthermore, compared to Co(Ac)2/Ni(Ac)2/PVP nanofibers, NiO/NiCo2O4 nanotubes display an evident shrinkage in diameter, which could be attributed to the removal of inner PVP and the decomposition of outer metal precursors. From the highmagnification FESEM image (Fig. 2d) we can clearly seen that these nanotubes with numerous nanopores in the tube walls are composed of interconnected primary nanocrystals with an average diameter of about 30 nm and the nanocrystals spontaneously align
Z. Cui et al. / Journal of Power Sources 272 (2014) 808e815
811
Fig. 1. (a) TGA/DTA curve of as-electrospun Co(Ac)2/Ni(Ac)2/PVP composite nanofibers. XRD patterns of porous NiO/NiCo2O4 nanotubes (b), NiO nanofibers (c) and NiCo2O4 nanofibers (d).
with one another to form a porous tube with an interior cavity and a loose tube wall. This porous network is believed to facilitate the inward and outward gas diffusion and thus it also is expected to have an
enhanced electrochemical performance. On the other hand, a possible formation mechanism of the porous nanotubes is proposed as follows: during heating process, the diffusion of gas produced by the inner PVP decomposition leads to the swelling of
Fig. 2. Typical FESEM images at different magnifications. (a) As-electrospun Co(Ac)2/Ni(Ac)2/PVP composite nanofibers and (bed) Porous NiO/NiCo2O4 nanotubes obtained after calcining the Co(Ac)2/Ni(Ac)2/PVP composite nanofibers at 600 C for 3 h.
812
Z. Cui et al. / Journal of Power Sources 272 (2014) 808e815
composite fibers forming hollow structures, whereas the shrinkage of composite fibers originating from the outer metal acetate decomposition results in the formation of porous tube wall. It should be noted that this kind of porosity on the tube wall in our case closely depends on the thermal treatment conditions. Fig. S1 shows the FESEM images of the Co(Ac)2/Ni(Ac)2/PVP composite nanofibers after different calcination temperatures. If the thermal treatment temperature is 500 C, solid, compact nanofibers with uniform diameters are obtained. When the as-electrospun Co(Ac)2/Ni(Ac)2/PVP nanofibers were calcined at 700 C in air for 3 h, nanotubes were formed, but many nanotubes collapsed (Fig. S1ced). Therefore, the most suitable temperature for the formation of well-defined porous nanotubes was 600 C. In addition, FESEM images of NiO nanofibers and NiCo2O4 nanofibers were shown in Fig. S2. As can be seen, they both have quite smooth surface without any pores over the surface. To gain further insight into the morphology and microstructure of the porous NiO/NiCo2O4 nanotubes, TEM analysis was performed. Fig. 3aec shows representative bright-field TEM images for the NiO/NiCo2O4 nanotubes, which clearly reveal welldefined tubular morphology with an average diameter of approximately 200 nm and wall thicknesses of about 50 nm. The high resolution TEM (HRTEM) images further confirm the high crystallinity of the porous NiO/NiCo2O4 nanotubes (Fig. 3d, e). Furthermore, the measured lattice spacing of 0.47 nm is in good
agreement with the (111) interplanar distance of NiCo2O4 phase, and that of 0.24 nm fits well with the (111) interplanar distance of NiO phase. The selected-area electron diffraction (SAED) pattern indicated the polycrystalline nature of the nanotubes (Fig. 3f), consistent with above HRTEM result. The scanning transmission electron microscopy (STEM) and corresponding elemental mapping images of the NiO/NiCo2O4 nanotubes reveal that Co, Ni and O are homogeneously distributed in each individual nanotube (Fig. 3g). In addition, the energy dispersive X-ray spectroscopy (EDX) spectrum shown in Fig. 3h indicates that the atomic ratio of Ni to Co in the resultant NiO/NiCo2O4 nanotubes is around 1:1.3, indicating that the mass content of NiO in the NiO/NiCo2O4 hybrid is 35%. To further examine the porous properties, N2 adsorption/ desorption isotherm measurements of the porous NiO/NiCo2O4 nanotubes are performed, as depicted in Fig. 3i. The nitrogen sorption isotherms can be classified as type IV according to IUPAC classification and the H3-type hysteresis loop reveals the mesoporous structure nature. The pore size of the mesoporous structure is determined to be around 12 nm by the BarretteJoynereHalenda (BJH) method from de desorption branch (the inset in Fig. 3i). The BET surface area is calculated to be around 28.60 m2 g1. This unique porous structure of the NiO/NiCo2O4 hybrid is expected to be an ideal design for fuel cell catalyst and has huge potential to further improve its electrochemical performance because the
Fig. 3. (aec) Typical TEM images of porous NiO/NiCo2O4 nanotubes at different magnifications. (d, e) HRTEM images recorded on different regions of single nanotube. (f) SAED pattern of NiO/NiCo2O4 nanotubes. (g) STEM and corresponding elemental mapping images, (h) EDS spectrum of the porous NiO/NiCo2O4 nanotubes. (i) Nitrogen adsorption/ desorption isotherms of NiO/NiCo2O4nanotubes and inset is pore size distribution.
Z. Cui et al. / Journal of Power Sources 272 (2014) 808e815
Like other manganese-based spinel compounds, NiCo2O4 also exhibits excellent electrochemical properties [13,14,19]. With its unique microstructures including hybridization and porous nanotube structure, the as-prepared NiO/NiCo2O4 sample is expected to display significantly enhanced electrochemical properties. The electrocatalytic properties of the NiO/NiCo2O4 nanotubes were first evaluated for ORR through conventional three-electrode cyclic voltammetry (CV) in N2 or O2 saturated 0.1 M KOH aqueous solutions. As shown in Fig. 5a, there was no significant peak obtained within the potential range between 0.8 and 0.2 V in the N2saturated solution. On the contrary, when oxygen was introduced, the electrode showed a well-defined cathodic peak appearing at 0.16 V (vs Ag/AgCl) at the same potential range, which can be attributed to electrocatalytic reduction of oxygen by the porous NiO/NiCo2O4 nanotube catalyst. This ORR potential is very close to that of the commercial Pt(20%)/C catalyst (Fig. 5b), and more positive than those of NiCo2O4 nanowires and macroporous NiCo2O4 sheets, [14,19] indicating a pronounced electrocatalytic activity of the NiO/NiCo2O4 sample for ORR. Furthermore, possible crossover effect of the NiO/NiCo2O4 catalyst is also investigated because it is one of important parameters for electrochemical evaluation of electrocatalysts for practical fuel cell applications. Therefore, the CV responses to methanol introduced into the O2-saturated electrolyte were measured for both NiO/NiCo2O4 nanotubes and commercial Pt(20%)/C catalyst. It can be seen from Fig. 5a that after the addition of 3 M methanol, NiO/NiCo2O4 nanotube catalyst retained highly stable current response. In contrast, the CV curve of the Pt(20%)/C electrode shows a strong methanol oxidation peak at around 0.13 V (Fig. 5b). These results indicate that the NiO/ NiCo2O4 nanotube catalyst has a good ability for avoiding crossover effects in the methanol-containing electrolyte. To further prove the role of the NiO/NiCo2O4 nanotubes in the ORR electrochemical process, linear sweep voltammograms (LSVs) on a rotating disk electrode (RDE) were recorded at different rotating speeds from
unique structures are crucial in facilitating the transfer of ions and electrons at the electrode/electrolyte interface. The surface electronic state and the composition of the assynthesized porous NiO/NiCo2O4 nanotubes were analyzed by Xray photoelectron spectroscopy (XPS). As shown in Fig. 4a, the survey XPS spectrum revealed the presence of Ni, Co, O, and C elements. Three typical signals of the O1s, Co2p and Ni2p core levels are detected from the XPS profiles shown in Fig. 4bed. In the Co2p spectra (Fig. 4b), two kinds of Co species (Co2þ and Co3þ) were detected. The binding energies at 779.3 and 794.3 eV can be ascribed to Co3þ, while another two fitting peaks at 781.0 and 795.7 eV can be attributed to Co2þ [15]. In the Ni 2p spectra (Fig. 4c), two kinds of nickel species containing Ni2þ and Ni3þ can also be observed. The fitting peaks at 854.0 and 871.7 eV can be indexed to Ni2þ, while the fitting peaks at 855.9 and 873.8 eV can be indexed to Ni3þ [16]. The satellite peaks at around 861.0 and 879.4 eV are two shake-up type peaks of nickel at the high binding energy side of the Ni2p3/2 and Ni2p1/2 edge [17]. The O 1s spectra (Fig. 4d) show a strong MeOeM peak at 529.48 eV, which is indicative of the lattice oxygen. The other two peaks at 532.83 and 531.15 eV can be attributed to CeO and OeC]O bonds, respectively, corresponding to a number of defect sites with low oxygen coordination in the materials with small particle size [18], which may be due to the multiplicity of physically and chemically adsorbed water at and within the surface [15]. Above result demonstrates that the surface of the as-prepared NiO/NiCo2O4 nanotubes has a composition containing Co2þ, Co3þ, Ni2þ and Ni3þ. Moreover, the calculated Co:Ni atomic ratio is 1.3:1, which is far low than that of stoichiometric NiCo2O4 (2:1), indicating that excessive NiO exists on the surface of the NiCo2O4. This result is well agreement with above XRD result. Here, the redox couples of Co2þ/Co3þ and Ni2þ/ Ni3þcould afford enough active sites for oxygen reduction, which may be one of the important factors contributing to the high electrocatalytic performance of NiO/NiCo2O4 hybrid.
(a) Ni2p
(b)
Ni1s
Intensity (a.u.)
Intensity (a.u.)
Co2p O(Auger) O1s Co(Auger)
C1s
Ni3p Ni3s
200
Ni3+
Co
800
790
780
Binding Energy (eV)
770
M-O-M
Ni2+
2+
Ni
Co3+ 3+
Co2+
Ni 2p3/2
Ni 2p1/2
Co 2p1/2 Co2+
Sat.
(d) Sat.
Intensity (a.u.)
Sat.
400 600 800 1000 1200 810 Binding Energy (eV)
(c)
890
Co 2p3/2
Sat. Ni3+
880 870 860 Binding Energy (eV)
Intensity (a.u.)
0
813
850
O-C=O C=O
540
535 530 525 Binding Energy (eV)
Fig. 4. (a) XPS spectra for the porous NiO/NiCo2O4 nanotubes: (a) Full-scan and high-resolution XPS spectra of (b) Co 2p, (c) Ni 2p, and (d) O 1s.
Z. Cui et al. / Journal of Power Sources 272 (2014) 808e815 2
(a)
O2 saturated N2 saturated O2 saturated in 3M CH3OH
1 0
-0.16 V
-1
10
Current density/mA cm-2
Current density/mA cm-2
814
8 6
-0.8
-0.6
-0.4
-0.2
0.0
0.2
Potential/V vs. Ag/AgCl
(b)
N2 saturated
4 2 0
-2
-2
O2 saturated O2 saturated in 3M CH3OH
-0.8
-0.6
-0.4
-0.2
0.0
0.2
Potential/V vs. Ag/AgCl
Fig. 5. CVs of the porous NiO/NiCo2O4 nanotubes (a)Pt(20%)/C (b) on a glassy-carbon RDE electrode in N2-saturated, O2-saturated 0.1 M KOH, and O2-saturated 0.1 M KOH and 3 M CH3OH solutions. (c) Rotating disk electrode voltammograms of the porous NiO/NiCo2O4 nanotubes in O2-saturated 0.1 M KOH with different electrode rotation rates. (d) The KouteckyeLevich plots derived from the rotating disk electrode measurements. (e) Rotating disk electrode voltammograms of the porous NiO/NiCo2O4 nanotubes, NiO nanofibers, NiCo2O4 nanofibers and Pt(20%)/C in O2-saturated 0.1 M KOH with a scan rate of 10 mV s1 at 1600 rpm. (f) The chronoamperometric curves of the porous NiO/NiCo2O4 nanotubes and Pt(20%)/C in O2-saturated 0.1 M KOH.
400 to 2500 rpm in 0.1 M KOH electrolyte saturated with O2 (Fig. 5c). It can be clearly seen that the current evidently increases with rotation rate due to the faster oxygen flux to the electrode surface [20,21]. The onset potential was about 0.05 V, very close to that (0.06 V) identified from CV curve (Fig. 5a). The KoutechyeLevich plots (J1 vs u1/2) of the NiO/NiCo2O4 catalyst are obtained from LSVs for different potentials (Fig. 5d) and all plots show fairly good linearity, suggesting first-order reaction kinetics toward dissolved oxygen and similar electron transfer number for the ORR process at different potentials. The transferred electron number (n) per oxygen molecule involved in this ORR process could be determined on the basis of KoutechyeLevich equations (Equation S1) and the KoutechyeLevich plots (J1 vs u1/2). Then, in each case, the n value is calculated to be approximately 3.9 at the potentials ranging from 0.40 to 0.70 V, indicating that the ORR process proceeds via an approximate four-electron pathway [22,23]. To gain further insight into ORR of each sample, LSV measurements were performed on RDE for the porous NiO/NiCo2O4 nanotubes, the NiO nanofibers, NiCo2O4 nanofibers and Pt(20%)/C in O2-saturated 0.1 M KOH with a scan rate of 10 mV s1 at 1600 rpm. From Fig. 5e, it is evident that the NiO/NiCo2O4 sample showed significantly better performance than NiO and NiCo2O4 samples in both onset potential and reaction current density. These results are in agreement with above CV observations (Fig. S3). Furthermore, we can also observe that the onset potential for the NiO/NiCo2O4 sample is close to that of the Pt(20%)/C catalyst, and that its current density is higher than that of the Pt(20%)/C catalyst, suggesting the fact that the ORR catalytic activity of the NiO/ NiCo2O4 catalyst is higher than that of the Pt(20%)/C catalysts. This performance of the NiO/NiCo2O4 sample is also better than those of the reported NiCo2O4-based materials with different morphologies (Table S1). To further assess the ORR reaction mechanism of the NiO/ NiCo2O4 sample, we analyzed its Tafel plots. As shown in Fig. S4, the NiO/NiCo2O4 sample had a Tafel slope of 61.2 mV per decade, slightly smaller than that (68.1 mV) of the Pt(20%)/C per decade, indicating that the ORR kinetics of the NiO/NiCo2O4 sample is similar to that of the Pt(20%)/C. This result further suggests that
the ORR process proceeds via an approximate four-electron pathway. The durability of the NiO/NiCo2O4 and Pt(20%)/C was also assessed through chronoamperometric measurements at 0.2 V in O2-saturated 0.1 M KOH solution for 40,000 s (Fig. 5f). Evidently, a high relative current of 87% can be retained even after 40,000 s for the NiO/NiCo2O4 catalyst. In contrast, the Pt(20%)/C catalyst showed a gradual decrease with a current loss of approximately 30% after 40,000 s. This observation strongly suggests that the durability of the NiO/NiCo2O4 catalyst is superior to that of the Pt(20%)/C catalyst. From the view of a practical application, this feature is particularly important. From the above results, we can see that the electrochemical performance of the as-synthesized NiO/NiCo2O4 hybrid is not only better than those of each component alone, NiO or NiCo2O4, but also comparable to that of the commercial Pt(20%)/C. Combining with reported studies, we believe that this may benefit from its unique microstructures, which can be summarized into following several aspects. First, the heterogeneous hybrid structure of NiO/ NiCo2O4 could ensure fast transport kinetics, which has been confirmed by electrochemical impedance spectroscopy (EIS) measurements performed from 100 KHz to 0.01 Hz. As shown in Fig. 6, the three Nyquist plots all exhibit characteristic of one semicircle in the high frequency range and one inclined line in the low frequency range. According to the Nyquist plots, the charge transfer resistance Rct of the NiO/NiCo2O4 hybrid electrode is determined to be 108.7 U, which is evidently lower than those of the other two electrodes (125.8 U for NiCo2O4 and 152.7 U for NiO). This demonstrates an impressive low charge transfer resistance for the NiO/NiCo2O4 hybrid, which is also common in other hybrids [11]. This feature has been widely confirmed to be mainly responsible for the significant improvement of the performance [11]. Just recently we have successfully constructed mesoporous ZnFe2O4/a-Fe2O3 microoctahedron hybrid, which exhibits superior lithium storage performance [24]. Second, the open porous tubular structure could build short diffusion channels between O2 and electrolyte, which consequentially ensure fast and uniform oxygen and electrolyte distribution inside the electrode. Third, the porous and tubular
Z. Cui et al. / Journal of Power Sources 272 (2014) 808e815
815
tubular architectures and it can be used for large-scale production of catalysts in fuel cells. Acknowledgments This work was supported by the National Natural Science Foundation of China (21471016, and 21271023), the National HighTech Research and Development Program of China (863 Program, 2012AA06A109) and the 111 Project (B07012). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jpowsour.2014.08.097. References Fig. 6. Nyquist plots of the porous NiO/NiCo2O4 nanotubes, NiO nanofibers, and NiCo2O4 nanofibers.
structure could also provide a relatively large electrodeeelectrolyte contact area to ensure high availability of the catalytic active sites. In a word, the unique composition and microstructure of the porous NiO/NiCo2O4 nanotube hybrid are responsible for its excellent electrochemical performance. 4. Conclusion In summary, we have constructed the NiO/NiCo2O4 hybrid with a porous tubular structure by a coaxial electrospinning method followed by a fine annealing treatment. As a potential non-noble metal ORR electro-catalyst for fuel cells, the porous NiO/NiCo2O4 nanotubes exhibit enhanced electrocatalytic activity, long-term operation stability, and tolerance to crossover effect compared to NiO nanofibers, NiCo2O4 nanofibers and commercial Pt(20%)/C via a four-electron pathway in alkaline environment. The onset potential was about 0.05 V and the current density of NiO/NiCo2O4 is about 115% of Pt/C current density. A high relative current of 87% can be retained even after 40,000 s for the NiO/NiCo2O4 catalyst. The unique microstructures of the porous NiO/NiCo2O4 nanotubes with BET surface area of 28.60 m2 g1, such as heterogeneous hybrid structure, open porous tubular structure, and the well dispersity of the two components, are responsible for their significantly improved electrocatalytic activity for ORR. Moreover, the promising and straightforward coaxial electrospinning has proved itself to be an efficient pathway for the preparation of nanomaterials with
[1] M. Winter, R.J. Brodd, Chem. Rev. 104 (2004) 4245e4269. [2] H. Zhang, P.K. Shen, Chem. Rev. 112 (2012) 2780e2832. [3] H.J. Tang, H.J. Yin, J.Y. Wang, N.L. Yang, D. Wang, Z.Y. Tang, Angew. Chem. 125 (2013) 5695e5699. [4] M.K. Debe, Nature 486 (2012) 43e51. [5] V.R. Stamenkovic, B.S. Mun, M. Arenz, K.J.J. Mayrhofer, C.A. Lucas, G. Wang, P.N. Ross, N.M. Markovic, Nat. Mater. 6 (2007) 241e247. [6] C. Wang, D.V. Vliet, K.L. More, N.J. Zaluzec, S. Peng, S.H. Sun, H. Daimon, G.F. Wang, J. Greeley, J. Pearson, A.P. Paulikas, G. Karapetrov, D. Strmcnik, N.M. Markovic, V.R. Stamenkovic, Nano Lett. 11 (2011) 919e926. [7] J.H. Kim, A. Ishihara, S. Mitsushima, N. Kamiya, K.I. Ota, Electrochim. Acta 52 (2007) 2492e2497. [8] V.I. Zaikovskii, K.S. Nagabhushana, V.V. Kriventsov, K.N. Loponov, S.V. Cherepanova, R.I. Kvon, H. Bonnemann, D.I. Kochubey, E.R. Savinova, J. Phys. Chem. B 110 (2006) 6881e6890. [9] G. Du, X. Liu, Y. Zong, T.A. Hor, A. Yuc, Z. Liu, Nanoscale 5 (2013) 4657e4661. [10] F.Y. Cheng, J. Shen, B. Peng, Y.D. Pan, J. Chen, Nat. Chem. 79 (2011) 79e84. [11] D.D. Wang, X. Chen, D.G. Evans, W.S. Yang, Nanoscale 5 (2013) 5312e5315. [12] J.P. Liu, J. Jiang, C.W. Cheng, H.X. Li, J.X. Zhang, H. Gong, H.J. Fan, Adv. Mater. 23 (2011) 2076e2081. [13] L.L. Li, S.J. Peng, Y.L. Cheah, P. Teh, J. Wang, G. Wee, Y. Ko, C. Wong, M. Srinivasan, Chem. Eur. J. 19 (2013) 5892e5898. [14] Y. Xiao, C.G. Hu, L.T. Qu, C.W. Hu, M.H. Cao, Chem. Eur. J. 42 (2013) 14271e14278. [15] J.F. Marco, J.R. Gancedo, M. Gracia, J.L. Gautier, E. Rios, J. Solid State Chem. 153 (2000) 74e81. [16] J.G. Kim, D.L. Pugmire, D. Battaglia, M.A. Langell, Appl. Surf. Sci. 165 (2000) 70e84. [17] J. Xiao, S. Yang, J. Mater. Chem. 22 (2012) 12253e12262. nez, A. Fern s, A.R. Gonz [18] V.M. Jime andez, J.P. Espino alez-Elipe, J. Electron Spectrosc. Relat. Phenom. 71 (1995) 61e71. [19] C. Jin, F.L. Lu, X.C. Cao, Z.R. Yang, R.Z. Yang, J. Mater. Chem. A 1 (2013) 12170e12177. [20] K.P. Gong, F. Du, Z.H. Xia, M. Durstock, L.M. Dai, Science 23 (2009) 760e764. [21] R.L. Liu, D.Q. Wu, X.L. Feng, K. Mullen, Angew. Chem. Int. Ed. 49 (2010) 2565e2569. [22] Y. Li, Y. Zhao, H.H. Cheng, Y. Hu, G.Q. Shi, L.M. Dai, L.T. Qu, J. Am. Chem. Soc. 134 (2012) 15e18. [23] S.Y. Wang, D.S. Yu, L.M. Dai, J. Am. Chem. Soc. 133 (2011) 5182e5185. [24] D. Zhao, Y. Xiao, X. Wang, Q. Gao, M.H. Cao, Nano Energy 7 (2014) 124e133.