Graphene-Co₃O₄ nanocomposite as electrocatalyst with high performance for oxygen evolution reaction.

Graphene-Co₃O₄ composite with a unique sandwich-architecture was successfully synthesized and applied as an efficient electrocatalyst for oxygen evolution reaction. Field emission scanning electron microscopy (FESEM) and transmission electron microscopy (TEM) analyses confirmed that Co₃O₄ nanocrystals were homogeneously distributed on both sides of graphene nanosheets. The obtained composite shows enhanced catalytic activities in both alkaline and neutral electrolytes. The onset potential towards the oxygen evolution reaction is 0.406 V (vs. Ag/AgCl) in 1 M KOH solution, and 0.858 V (vs. Ag/AgCl) in neutral phosphate buffer solution (PBS), respectively. The current density of 10 mA/cm(2) has been achieved at the overpotential of 313 mV in 1 M KOH and 498 mV in PBS. The graphene-Co₃O₄ composite also exhibited an excellent stability in both alkaline and neutral electrolytes. In particular, no obvious current density decay was observed after 10 hours testing in alkaline solution and the morphology of the material was well maintained, which could be ascribed to the synergistic effect of combining Co₃O₄ and graphene.

Solar-driven electrochemical transformation of small molecules, such as water splitting and carbon dioxide reduction, is one of the most promising approaches for producing clean and sustainable energy12. Photoelectrocatalytic or electrocatalytic water splitting is a process including water oxidation and reduction3456. However, water splitting is mainly hindered by the oxygen evolution reaction (OER, 4H+/4e−), which has several steps and requires large overpotential7. Currently, some noble metals (ruthenium and iridium) and their compounds exhibit excellent activity towards OER with low overpotential and high current density89. However, their application are hindered by the scarcity and high cost. Therefore, discovering efficient and inexpensive catalysts is critical to enhance the OER current density and reduce the overpotential.

In the past few years, considerable efforts have been dedicated to utilizing the earth-abundant metal oxides such as cobalt oxide10111213 nickel oxide141516 and manganese oxide1718 materials, as electrocatalysts for OER. Among these catalysts, Co3O4 has attracted extensive attentions owing to its high activity and superior stability12192021. For example, mesoporous Co3O4 has been reported as the OER catalyst with current densities of 27.2 mA/cm2 at 1 V (vs. Ag/AgCl) and the crystalline Co3O4 showed a relatively small Tafel slope (49 mV/decade)1920. In addition, Co3O4 could also be utilized in photocatalytic system. Jiao et al. reported the photocatalytic properties of the mesoporous Co3O4 combined with the [Ru(bpy)3]2+ with a high turnover frequency (TOF) of ~2.2 × 10−3 s−1 per Co atom22. However, the strong causticity of alkaline solution and the intrinsically low conductivity of Co3O4 have impeded the further development of Co3O4 as OER catalyst.

Combining Co3O4 nanoparticles with conductive substrates can efficiently enhance the conductivity of the catalyst and significantly affect its catalytic activity and stability. Carbon materials, such as carbon nanotube (CNT), graphene (G) and mesoporous carbon, have been widely employed as the supporting substrates owing to their high conductivity and large specific surface area23242526. Recently, Dai's group synthesized Co3O4 nanocrystals supported on graphene by hydrothermal reaction and demonstrated high electrocatalytic performance27. Zhao and coworkers reported OER catalysts of Co3O4 nanoparticle/graphene composites, fabricated by the layer-by-layer assembly28. Co3O4/CNT (single-walled or multi-walled) materials were also prepared as high performance catalysts towards oxygen evolution2930.

Herein, we report a simple method to prepare a unique sandwich-architectured catalyst composed of graphene and Co3O4 (G-Co3O4). Ultrafine Co3O4 particles uniformly anchor onto both sides of graphene sheets. The unique sandwich-architecture leads to large amount loading of the active Co3O4 nanocrystals and enhances electron transfer kinetics between the materials. Therefore, the catalytic activity and stability of the catalyst have been substantially promoted. G-Co3O4 composite catalyst shows low overpotentials of 313 mV and 498 mV to achieve the current density of 10 mA/cm2 in the alkaline and neutral conditions, respectively. Furthermore, there is no obvious current density decay after the stability test.

Results

Synthesis of G-Co3O4 nanocomposite

The synthesis of the sandwich-architectured composite is illustrated in Fig. 1. Cobalt (II) acetate was added into the xylene and oleic acid mixed solution. The oleic acid in the mixture acted as a capping agent in order to control the particle growth and prevent colloidal particles from aggregation. Then, the as-prepared GO solution was introduced into the above solution under vigorous stirring. CoII cations attached on both sides of the GO nanosheets by electrostatic interaction31. During the following refluxing process and the reducing process by adding NaBH4, Co nanoparticles were loaded on the surface of rGO nanosheets. The final product of G-Co3O4 with unique sandwich-architecture was obtained after calcination.

Figure 1

A schematic illustration for preparing G-Co3O4 nanocomposite with a sandwich-architecture.

The characterization of the G-Co3O4 composite

The morphology of the precursor obtained after filtration and calcination in argon atmosphere was characterized by scanning electron microscopy (SEM, see Supplementary Fig. S1 online), which highly maintained the layer structure of graphene oxide. The morphology of the synthesized G-Co3O4 composite was investigated by SEM. The low magnification image in Fig. 2a clearly shows that the obtained G-Co3O4 composite still displayed layer structure, which is similar to the pristine graphene (see Supplementary Fig. S2a online). The high magnification in the inset of Fig. 2a shows two layers of Co3O4 particles homogeneously distributed. From the layer structure and the homogeneous distribution of the Co3O4 particles, we infer that the Co3O4 particles attached on both sides of the graphene nanosheets, which could form a unique sandwich-architecture. However, it is difficult to clearly find graphene because of the low content (8.8 wt.%, see Supplementary Fig. S3 online). The transmission electron microscopy (TEM) image in Fig. 2b further demonstrates the homogeneous distribution of Co3O4 nanoparticles on graphene substrate. High resolution TEM (HRTEM) in Fig. 2c clearly confirmed the sandwich-architecture. The graphene was definitely imbedded between the parallel layers of Co3O4 particles even after a strong sonication, suggesting a relatively strong interaction between graphene and Co3O4 particles. The graphene nanosheets acted as a binder to link neighboring Co3O4 particles together and also increased the conductivity of the composite. They can further prevent the aggregation of the Co3O4 nanoparticles during thermal treatment. We also synthesized pristine Co3O4 through similar process without adding graphene oxide. The SEM image (see Supplementary Fig. S2b online) shows the particle size of pristine Co3O4 is much bigger than that in G-Co3O4 composite. Furthermore, as shown in Figure 2d, the lattice fringes in HRTEM image and the selected area electron diffraction (SAED) pattern further confirmed the formation of crystalline Co3O4.

Figure 2

Microscope observation of the G-Co3O4 composite.

(a) Low and high magnification SEM images of G-Co3O4. (b) Low magnification TEM image of G-Co3O4. (c) High resolution TEM (HRTEM) image of G-Co3O4, showing a sandwich-like microstructure. (d) HRTEM image of Co3O4 nanocrystals and SAED pattern of G-Co3O4.

The crystal structure of the obtained G-Co3O4 composite and pristine Co3O4 nanoparticles was determined by X-ray diffraction (XRD) method as shown in Fig. 3a. The prominent peaks at 31.2, 37.1, 45.1, 59.7, 65.6° of G-Co3O4 and pristine Co3O4 can be indexed to face-centered cubic phase (Fd3m, JCPDS card No. 76-1802). The broad diffraction peak appeared at around 24.8° in the G-Co3O4 composite is attributed to the (002) direction of graphene. Raman spectra of the G-Co3O4 composite and pristine Co3O4 are shown in Fig. 3b. Two peaks at ca. 470 and 671 cm−1 can be assigned to Co3O4. Moreover, G-Co3O4 also displays another two obvious peaks at 1327 and 1584 cm−1, which can be indexed to the peaks of graphene3233. The XRD and Raman spectra further confirm the successful synthesis of G-Co3O4 composite catalyst.

Figure 3

The characterization of the G-Co3O4 composite.

(a) XRD patterns of Co3O4 and G-Co3O4 composite. (b) Raman spectra of Co3O4 and G-Co3O4 composite catalyst.

Electrochemical performance of the OER catalyst

The water oxidation catalytic activities of the as-synthesized nanocomposite were first investigated in alkaline solution (0.1 or 1 M KOH) in a standard three-electrode setup. During the electrochemical test, the working electrode was continuously rotating at 1600 rpm to remove the generated oxygen bubbles. Linear sweep voltammetry (LSV) curves of G-Co3O4 catalysts are shown in Fig. 4a. Ruthenium nanocrystal functionalized carbon black catalysts (30% Ru loaded on carbon black, Ru/C) was also tested for comparison. G-Co3O4 catalyst exhibits significantly higher anodic current and lower onset potential than that of the Ru/C catalysts, in both 0.1 and 1 M KOH solutions (Fig. 4a). In 1 M KOH solution, G-Co3O4 catalyst shows a sharp onset potential at 0.406 V (vs. Ag/AgCl, following the method described by Chen et al.)34, and achieves the current density of j = 10 mA/cm2 at the overpotential of 313 mV, which is much better than that of the Co3O4/SWNTs (593 mV) and mesoporous Co3O4 catalysts (525 mV) in the same alkaline solution, and even comparable to the best performance of G/Co3O4 catalysts (310 mV)202729. In 0.1 M KOH, G-Co3O4 catalyst also exhibits a lower onset catalytic potential (0.446 V vs. Ag/AgCl) and higher current density (achieved a current density of j = 10 mA/cm2 at the overpotential of 359 mV) than those of the Ru/C catalyst.

Figure 4

Electrochemical performance.

(a) Polarization curves of G-Co3O4 and Ru/C catalysts on GC electrodes in 0.1 and 1 M KOH. (b) The Tafel curves of G-Co3O4 in 0.1 and 1 M KOH. (c, d) Chronoamperometric responses (percentage of current retained versus operation time) of G-Co3O4 and Ru/C catalysts in 1 M and 0.1 M KOH electrolytes.

Based on the mass content of the Co3O4 in the composite calculated from the TGA test and assuming that all deposited materials were involved in the electrochemical reaction, the lower limits for turnover frequency (TOF) can be derived from the catalytic current. We calculated a high TOF of 0.45 s−1 referring to per Co atom for G-Co3O4 catalyst at the overpotential of 350 mV in 1 M KOH and 0.194 s−1 at the overpotential of 400 mV in 0.1 M KOH, which are higher than those of previous reports about Co based materials (Co3O4, 0.08 s−1 at overpotential of 507 mV in 1 M KOH and mesoporous Co3O4, 3.16 × 10−3 s−1 at 400 mV in 0.1 M KOH)1020.

Tafel plot is applied to evaluate the efficiency of the catalytic reaction in alkaline solutions (0.1 and 1 M KOH), which is derived from the polarization curves using the Tafel equation η = b log(j/j), where η is the overpotential, b is the Tafel slope, j is the current density, and j is the exchange current density. The G-Co3O4 catalyst exhibits Tafel slopes of b = 56 mV/decade in 1 M KOH and b = 67 mV/decade in 0.1 M KOH (Fig. 4b), which are lower than previously reported G/Co3O4 (67 mV/decade in 1 M KOH) and graphene/NiCo2O4 hybrid paper (156 mV/decade in 0.1 M KOH)2735. The observed Tafel slope value suggests the favorable OER kinetics over G-Co3O4 catalyst and also the good chemical and electronic coupling between the Co3O4 nanoparticles and graphene nanosheets.

The G-Co3O4 catalyst also exhibits good stability in the alkaline solutions, which is another important factor for energy conversion systems. In 1 M KOH solution, the G-Co3O4 electrode shows excellent durability with no obvious activity decay compared with the initial value, while the Ru/C catalyst electrode degrades by 25.8% of the initial value (Fig. 4c). A similar trend is also observed in 0.1 M KOH (Fig. 4d). The result of repeating potential cycling for 1000 cycles also confirmed the good stability of the material. After cycling test, the polarization curve of the G-Co3O4 modified electrode was almost the same as the initial one (see Supplementary Fig. S4 online). Furthermore, no obvious morphology change is observed in the SEM image of G-Co3O4 after long-term stability testing (see Supplementary Fig. S5 online), suggesting that G-Co3O4 catalyst can tolerate long-term corrosion and possess robust mechanical properties.

The G-Co3O4 composite has shown high catalytic activity and good stability towards OER in alkaline solution. On the other hand, the enhanced catalytic activities in neutral solution are more desirable owing to the benign nature and weak causticity. The catalytic activity of Co3O4 towards OER is sensitive to a low pH value, especially for the neutral solution according to the previous investigation36. Therefore, it is necessary to evaluate the electrochemical performance of G-Co3O4 catalyst in neutral solution. The G-Co3O4 electrode with the same mass loading was tested in 0.1 M phosphate buffer solution (PBS) at pH 7. The LSV curve of G-Co3O4 catalyst and the corresponding Tafel plot obtained in the neutral solution are presented in Fig. 5a. In the neutral condition, G-Co3O4 catalyst shows an onset potential at 0.858 V (vs. Ag/AgCl) and achieves a current density of j = 10 mA/cm2 at the overpotential of 498 mV, which exhibit better performance than those of the previous report (810 mV)29. Moreover, G-Co3O4 catalyst exhibits a little smaller Tafel plot of 98 mV/decade compared to others (104 mV/decade or 110 mV/decade)2937. The decreased Tafel slope value can be ascribed to the synergistic coupling and the fast charge transport of the materials. Furthermore, the stability of G-Co3O4 is assessed by repeated potential cycling for 1000 cycles, as shown in Fig. 5b. Only a slight decay of the activity (<10%) was observed referred to the polarization curves after the long-term test.

Figure 5

Electrochemical performance of G-Co3O4 composite OER catalyst.

(a) LSV and Tafel slope of G-Co3O4 in neutral solution, (b) OER stability test of G-Co3O4, the initial and 1000th polarization curves in neutral solution.

Discussion

The excellent electrochemical performance of G-Co3O4 catalyst suggests its promising application towards OER both in alkaline and neutral solution. The high activity and excellent durability of G-Co3O4 towards OER are mainly attributed to the unique sandwich-architecture of the materials. A series of controlled experiments, shown in Fig. 6, have demonstrated that the obtained G-Co3O4 catalyst displayed better OER activity than those of the simple physical mixture of Co3O4 and graphene (G+Co3O4), pure Co3O4 and graphene in 1 M and 0.1 M KOH conditions. The activity order is G-Co3O4 > G+Co3O4 > pure Co3O4 > graphene. The strong interaction of Co3O4 nanoparticles with graphene nanosheets and the unique sandwich-architecture have strong effect on the catalytic property of the composite. Simple physical mixing cannot create effective interfacial contacts between the Co3O4 nanoparticles and graphene nanosheets. This was also confirmed by the electrochemical impedance spectroscopy (see Supplementary Fig. S6 online). The Nyquist plots showed that the OER charge resistance of the G-Co3O4 catalyst is the smallest among the other samples (G+Co3O4, Co3O4). This further illustrates that the strong chemical coupling and good interaction between the Co3O4 and graphene can significantly improve the electron transport and reaction kinetics during the OER. In order to further investigate the high performance of G-Co3O4, the effective surface areas were estimated, which was determined by electrochemical capacitance measurements from static cyclic voltammetry (see Supplementary Fig. S7 online)3839. The capacitance for G-Co3O4 composite is 12.4 mF/cm2, which is much higher than G+Co3O4 (2.8 mF/cm2). Therefore, the high performance could be also associated with the high electroactive surface area.

Figure 6

Electrochemical performance.

Polarization curves of G-Co3O4, G+Co3O4, the pristine Co3O4 and graphene in (a) 1 M and (b) 0.1 M KOH solutions.

The observed enhanced activity of the G-Co3O4 catalyst indicates that the sandwich-architecture is favorable for the oxygen evolution, in which the graphene nanosheets provide large specific surface area and good conductivity. This specific structure also leads to large amount loading and much small particle size of the Co3O4 anchored on both sides of the graphene sheets, which provides much more electroactive surface area for the oxygen evolution. In addition, the unique and intimate contact between the graphene and Co3O4 afforded by one-step facile mechanism contributes to the superior catalytic activity.

In conclusion, sandwich-architectured G-Co3O4 composite was successfully synthesized exhibiting good interaction between Co3O4 nanoparticles and graphene nanosheets. G-Co3O4 catalyst exhibits enhanced catalytic performances with high catalytic activities, good stability and favorable reaction kinetics in both the alkaline and neutral solutions, as catalyst for water oxidation. The enhanced catalytic performance demonstrates large electroactive surface area, fast electron transfer rate and superior electrical and chemical coupling of the composite. The unique morphology and excellent electrochemical performance of G-Co3O4 catalyst render this material a promising noble metal free catalyst towards the oxygen evolution reaction.

Methods

Materials synthesis

Graphene oxide (GO) was prepared by the oxidation and exfoliation of the graphite according to the previously reported procedure4041. Sandwich-like G-Co3O4 was synthesized via an oleic acid assisted method followed by thermal treatment42. In a typical synthesis process, 0.68 g oleic acid was mixed with 40 mL dry xylene under vigorous magnetic stirring. Then, 0.5 g cobalt acetate tetrahydrate (Co(AC)2, Aldrich) was introduced into the mixture and sonicated for 1 min. After that, the mixture was mixed with 40 ml GO solution (2 mg ml−1) by vigorous stirring and reacted in a pre-heated oil bath under refluxing for 4 h at 105°C. Then, 10 ml (5 mg/ml) NaBH4 was added to the reaction mixture and kept stirring for another 10 min. The precipitation was collected by filtration and annealed at 300°C for 3 h in argon atmosphere. The final product was obtained after annealing at 400°C for another 2 h in air. The pristine Co3O4 was synthesized by similar method without the addition of GO solution. The comparision sample of ruthenium nanocrystals supported on carbon black (Ru/C) was synthesized by using hydrophilic and hydrophobic tri-block copolymer F127 as a soft template, followed by low temperature heat treatment43.

Structural characterization

The morphology of the obtained materials was characterized by field emission scanning electron microscopy (FESEM, Zeiss Supra 55VP) and transmission electron microscopy (TEM, Model JEM-2011, JEOL). X-ray diffraction (XRD) patterns were collected on Siemens D5000 diffractometer using Cu Kα radiation with a scanning step of 0.02° per second. Raman spectra were recorded with an inVia Renishaw Raman spectrometer system (HR Micro Raman spectrometer, Horiba JOBIN YVON US/HR800 UV) equipped with a 632.8 nm wavelength laser. Thermal gravimetric analysis (TGA) of G-Co3O4 composite was performed using a TGA/differential thermal analysis (DTA) analyzer (TA Instruments, SDT 2960 module, New Castle, DE, USA) at a heating rate of 10°C min−1 in air from room temperature to 700°C.

Electrochemical measurements

Electrocatalytic activity measurements were carried out on an electrochemical workstation (CHI 660E) in a three-electrode glass cell system. A glass carbon (GC) electrode coated with as-prepared materials was used as the working electrode. A platinum wire was used as the counter electrode (CE). The potential was recorded using an Ag/AgCl (1 M KCl) reference electrode, which was converted to the reversible hydrogen electrode (RHE) according to Nernst equation ERHE = EAg/AgCl + 0.059 × pH + 0.2224. The working electrode was prepared as follows: 4 mg sample was dispersed in 1 ml of 1:1 v/v water/isopropanol by ultrasonication. Then 80 μl Nafion (5 wt%) was added to the solution to obtain a homogeneous ink. The catalyst ink (10 μl) was loaded onto the GC electrode with a diameter of 5 mm. LSV was conducted in KOH solution (0.1 M, 1 M) and phosphate buffer solution (PBS, 0.1 M) at a scan rate of 2 mV s−1. The polarization curves were all corrected by 95% iR compensation and our typical electrochemical cell had Ru = ~8 Ω in 1 M KOH, Ru = ~20 Ω in 0.1 M KOH and Ru = ~26 Ω in 0.1 M PBS. TOFs were calculated according to the equation of TOF = nO2/nCo = (Q/4F)/nCo (F is the faraday constant, 96485 C/mol). Stability was carried out for 10 h at 0.50 V (vs. Ag/AgCl) in 1 M KOH and 0.60 V (vs. Ag/AgCl) in 0.1 M KOH solution. Electrical impedance spectroscopy (EIS) was recorded under the following condition: ac voltage amplitude 5 mV, frequency ranges from 106 to 0.1 Hz, and open circuit. Cyclic voltammograms (CV) of G-Co3O4 and G+Co3O4 measured in 1 M KOH solution in a potential window (−0.16 V to −0.08 V (vs. Ag/AgCl)) without faradaic processes with scan rates of 10, 30, 50, 70, 90 mV/s.

Author Contributions

G.W. designed the experiments, co-wrote the manuscript. Y.Z. performed the major part of experiments, organized the data, and wrote the manuscript. D.S. contributed to the TEM experiments and discussion. S.C., B.S., X.H., K.S., Y.Y. and H.L. involved in discussion. All authors read and approved the final manuscript.