Roughening of windmill-shaped spinel Co₃O₄ microcrystals grown on a flexible metal substrate by a facile surface treatment to enhance their performance in the oxidation of water

Abstract

High-efficiency and Earth-abundant electrocatalysts for the oxidation of water are required in the production of clean energy from the electrolysis or photolysis of water. Spinel Co₃O₄ microcrystals with a windmill shape were grown on a flexible metal substrate. The microcrystals were then roughened by a surface impregnation treatment. A secondary nanostructure grew out of the blades of the windmills to form a micro/nano hierarchical structure. The as-grown micro/nano Co₃O₄ had an excellent electrochemical performance in the oxidation of water. The onset overpotential of the micro/nano Co₃O₄ electrocatalyst for the oxidation of water was about 0.29 V in alkaline solution and the overpotential of the optimum Co₃O₄ electrocatalyst was 0.41 V at a current density of 10 mA cm⁻². These results suggest that the electrochemical performance is associated with the roughness and active surface of the Co₃O₄ electrodes. The turnover frequency of the optimized Co₃O₄ reached 0.39 s⁻¹ at an overpotential of 0.6 V, about 1.4 times higher than that for the pristine Co₃O₄ microcrystals. The turnover frequency of micro/nano Co₃O₄ is higher than, or comparable to, that previously reported for high-efficiency nanosized Co₃O₄ in alkaline solution. Stability tests indicated that these micro/nano Co₃O₄ electrocatalysts were highly durable towards the oxidation of water, with no structural change and no decrease in noticeable activity after operating for 12 h in oxygen-evolving reactions. This work verifies the contribution of surface roughness and an active surface to the electrochemical oxidation of water by the design of an optimum micro/nano Co₃O₄ electrocatalyst. Our current understanding of the catalytic role of a specific micro/nano structure is also strengthened.

---

Introduction

Our dependence on energy is becoming increasingly intense. Renewable and “clean” energy (e.g. solar, geothermal and nuclear energy) are attracting much attention as a result of their sustainability and “clean” nature. Hydrogen is predicted to become an environmentally friendly and promising form of energy as a result of its high energy density and carbon-free emissions. The electrochemical or light-driven splitting of water molecules is an important method of producing clean energy in the form of hydrogen, with negligible environmental issues.^1–3 To split water, two half-reactions involving the reduction of water to hydrogen and the oxidation of water to oxygen are used. Of the two reactions, the oxidation of water (H₂O/O₂) is a multiple proton-coupled four-electron transfer process, which has a high thermodynamic barrier and slow kinetics.⁴ This inherent bottleneck seriously limits the overall efficiency and rate of water-splitting reactions. High-efficiency catalysts to lower this barrier and speed up the kinetics have received much research interest. Noble metal catalysts (e.g. IrO₂, RuO₂ or noble metal complexes) are some of the most efficient catalysts for the oxidation of water.^5–7 However, the rarity and unsustainability of noble metals have limited their widespread application, despite their good performance in both homogeneous and heterogeneous environments.^8,9

Many attempts have been made to find high-efficiency and Earth-abundant alternatives to the noble metal catalysts, such as Co, Mn, Ni and Fe-based metal oxides or molecular complexes.^10–15 Although molecular or heterogeneous catalysts have shown a high electrocatalytic performance for the oxidation of water in various electrolytes, they usually occur as molecular units or as dispersive particles separated from the electrode materials.^16–18 Tedious post-treatment is required to bind or cast the powders onto the surface of the electrode and peel-off is a common problem. As a result, it is challenging to integrate the catalysts directly onto the electrodes for a practical and engineered product with no significant loss of catalytic activity and durability. Some workers have explored constructions based on a concept of structured catalysts.^19–21

As an efficient, Earth-abundant and easily made electrocatalyst, Co₃O₄ has received much attention over the last few decades, although it is slightly less active than the noble metal oxides for the oxidation of water. A variety of spinel Co₃O₄ or doped cobalt oxide structures have been synthesized.^{18,19,22–25} Koza et al.²³ synthesized crystalline and amorphous Co₃O₄ films using an electrodeposition method, which showed high OER activity in alkaline solution. Han et al.¹⁸ reported the deposition of nanostructured cobalt oxide catalysts from molecular cobaloximes for the oxidation of water in an alkaline buffer solution. Zou et al.²⁵ prepared porous Ni-doped Co₃O₄ nanomaterials via the thermal treatment of graphitic C₃N₄ embedded with metal ions.²⁵ The porous structure and Ni doping led to a superior electrocatalytic activity for the OER. Nanostructuring or roughening the pristine surface of catalysts are the most common ways of increasing the number of surface active sites in catalytic reactions. For example, Li et al.²⁴ reported Ni_xCo_3−xO₄ nanowire arrays for electrocatalytic oxygen evolution, in which the surface was roughened by constructing an Ni-doped mesoporous structure. A greater roughness resulted in a higher catalytic activity. Inspired by these studies, we thought that roughening the surface of pristine Co₃O₄ by a facile means could be a promising way of enhancing its catalytic activity for the oxidation of water.

Windmill-shaped spinel Co₃O₄ microcrystals were therefore grown directly on a metal substrate and subsequently roughened by a facile post-treatment of the surface. Film electrodes consisting of a micro/nano structure were formed and these electrodes showed a roughness-dependent electrochemical OER performance. These findings suggest that the OER activity may be enhanced by roughening pristine Co₃O₄ microcrystals into an optimized micro/nano hierarchical structure. The enhancement in activity was attributed to the increased roughness and active surface of the electrodes, which resulted in more active sites being accessible to the reactants during the OER.

Experimental section

Materials

All chemicals were of analytical-reagent grade and were used as received without further purification. The metal foils (NiCr) used as the substrate were purchased from Beijing Xinxing Brain Technology Co. Ltd. The nominal thickness of the foils was 0.03 mm. Before use, a piece of metal foil was first cleaned ultrasonically in acetone for 5 min to remove any grease and was then washed thoroughly using a concentrated HCl solution (37 wt%) for no more than 3 min to remove the surface oxide layer. The foil was then rinsed three times each with deionized water and anhydrous ethanol.

Synthesis of Co₃O₄ films

Cobalt nitrate [Co(NO₃)₂·6H₂O, 3 mmol], ammonium fluoride (NH₄F, 12 mmol) and urea [CO(NH₂)₂, 15 mmol] were dissolved in 54 mL of distilled water with vigorous stirring to give a clear solution. The solution was then transferred to a Teflon-lined stainless-steel autoclave in which a cleaned foil was placed. The autoclave was sealed and placed in an oven at 100 °C for 6 h. After the reaction had taken place, the autoclave was cooled with water. The foil was rinsed many times with deionized water and absolute ethanol and dried at 70 °C in an oven for 6 h. The Co₃O₄ films were eventually obtained by calcination for 3 h in a muffle furnace at 300 °C. This sample is referred to as Co₃O₄. The product grown on the foil was weighed on a micro-balance. For comparison, Co₃O₄ nanowires were grown on pure Ni foil using the same synthesis conditions. The product is referred to as Co₃O₄-NW.

Preparation of micro/nano Co₃O₄ film electrodes

Micro/nano Co₃O₄ films were obtained by a facile impregnation method. The foil covered with a film of Co₃O₄ was impregnated in an aqueous solution of Co(NO₃)₃·6H₂O in which the mass of the solute was controlled to be 1 wt%, 3 wt% or 5 wt% of the Co₃O₄ previously grown on the foil. The impregnated foils were calcined in a muffle furnace for 1 h at 300 °C. The products were named Co₃O₄-1, Co₃O₄-3 and Co₃O₄-5.

Characterization of the materials

X-ray diffraction measurements. The XRD patterns of the samples were collected using a Shimadzu XRD-6000 diffractometer (40 kV, 30 mA, graphite-filtered CuKα radiation, λ = 0.15418 nm).

Scanning electron microscopy. SEM photographs were taken using a Hitachi S4700 microscope operating at an accelerating voltage of 20 kV and equipped with an energy-dispersive X-ray spectrometer for compositional analyses.

Transmission electron microscopy. TEM photographs were taken using a JEOL JEM-2100 microscope. For the TEM observations, the samples were scraped from the substrate and ultrasonically dispersed in ethanol. A drop of the suspension was then deposited on a carbon-coated Cu grid.

Raman measurements. The Raman spectra were collected at room temperature using a Jobin Yvon Horiba HR800 Raman spectrometer. A laser with a wavelength of 633 nm was used as the excitation source. The spectra were collected at three different spots to ensure the reliability of the results.

Elemental analyses. Elemental analyses to determine the composition were carried out using inductively coupled plasma atomic emission spectroscopy (Shimadzu ICPS-7500).

X-ray photoelectron spectrometry. The XPS spectra of the samples were recorded on a Thermo VG ESCALAB250 X-ray photoelectron spectrometer at a pressure of about 2 × 10⁻⁹ Pa using an Al Kα X-ray as the excitation source (1486.6 eV). The shifts in all the binding energies were calibrated using the C1s core level at 284.6 eV.

Electrochemical measurements

The electrochemical performance of the films in the oxidation of water was studied in a standard three-electrode configuration connected to a CHI660C potentiostat (CH Instrument Co. USA). The films were used as the working electrode. A saturated calomel electrode (SCE) and Pt wire (2 cm) were used as the reference and counter electrodes, respectively. The working electrodes with a projection area of 1 × 1 cm were rinsed with deionized water prior to use. Cyclic voltammetry (CV) and linear sweep voltammetry (LSV) were achieved in an alkaline solution (0.1 M KOH, pH 13) at a scan rate of 1 mV s⁻¹. The amperometric (I–t) curves were recorded at a constant applied potential to judge the stability of the catalysts. After the I–t measurements, the samples were rinsed with deionized water and then blow-dried in a N₂ flow for further use. All the potentials measured were converted to a reversible hydrogen electrode (RHE) according to the formula E_RHE = E_SCE + 0.242 + 0.059 pH (units in volts).⁴

Electrochemical impedance spectroscopy (EIS) measurements were carried out at an applied potential of 0.56 V (vs. SCE) in 0.1 M KOH solution on a CHI660C electrochemical workstation (CH Instrument Co.) with an AC amplitude of 10 mV. The impedance spectra were recorded in the frequency range 100 kHz to 0.01 Hz at room temperature. Before the measurement, KOH solution was bubbled with high-purity N₂ at a flow-rate of 25 mL min⁻¹ for 30 min. Before recording each impedance spectrum, the electrode was first equilibrated at a constant potential for 120 s. The experimental impedance data were fitted to a specific equivalent circuit using ZSimPWin software to derive the resistance of the electrodes.

Turnover frequency evaluations

The surface concentration of active sites is associated with the redox species. During CV scans, the peak current of the redox species shows a linear dependence on the scan rate. The slope of the line is calculated in terms of the formula: slope = n²F²AΓ₀/4RT, where n is the number of electrons transferred, F is Faraday's constant, A is the area of the electrode, Γ₀ is the surface concentration of active sites (mol cm⁻²), and R and T are the ideal gas constant and the absolute temperature, respectively. The turnover frequency (TOF) was calculated from the formula: TOF = JA/4Fm, where J is the current density at a constant overpotential, A is the area of the electrode, 4 represents the number of moles of electrons consumed to evolve one mole of oxygen, F is Faraday's constant and m is the number of moles of active sites. The TOFs were calculated by assuming that all the cobalt ions in the electrodes were catalytically active.

The amount of oxygen evolved was measured using a NeoFOX oxygen-sensing system (Ocean Optics Inc.) equipped with a FOXY probe inserted into the head-space of a gas-tight cell. The cell was purged with nitrogen for 20 min before measurement. The oxygen fluorescence signal was allowed to stabilize once the nitrogen had stopped. After the residual oxygen signal became constant, the desired potential was applied. Oxygen evolution was monitored at 1.64 V (vs. RHE) for 1 h. The fluorescence data measured was transformed into the amount of O₂ in μmol.

Results and discussion

The Co₃O₄ was grown directly on the metal (NiCr) substrate to construct a flexible film electrode. The post-treatment was achieved by a facile impregnation approach that aimed to increase the surface roughness and construct a micro/nano hierarchical structure. Fig. 1 shows the XRD patterns of the Co₃O₄ and the modified electrodes. All the patterns show clear reflections that can be indexed to the spinel Co₃O₄ phase (JCPDS no. 42-1467). The positions of each reflection among different patterns have negligible shifts. Two intensive reflections at 2θ = 43.9° and 51.2° can be assigned to Ni (111) and (200) reflections (indicated by solid triangles) from the substrate. No other crystalline phase was detected after surface treatment, indicating that the treatment applied to the pristine Co₃O₄ does not change the phase structure (Fig. 1B–D). In comparison, Co₃O₄ was also grown on a pure Ni foil and the XRD pattern could be indexed to just the Co₃O₄ phase (Fig. S1†). The additional reflection at 2θ = 51.2° comes from the Ni foil and the other reflection at 43.9° partially overlaps with the (400) reflection of the spinel phase. The XRD characterization suggests that the metal substrate (either NiCr or Ni) has little effect on the phase composition of the products.

Fig. 1 XRD patterns of Co₃O₄ grown on a metal foil by a surface treatment method. (A) Co₃O₄, (B) Co₃O₄-1, (C) Co₃O₄-3 and (D) Co₃O₄-5 (closed triangle = reflections from Ni).

Raman spectroscopy is a useful tool for revealing the delicate structure associated with the vibration modes of bonds. Fig. 2 shows the Raman spectra of the Co₃O₄ film electrodes. Three intensive vibration bands at about 194, 476 and 681 cm⁻¹ are observed and also two weak bands at 520 and 615 cm⁻¹. These Raman shifts are characteristic of the spinel structure Co₃O₄, consistent with the assignments reported previously.²⁶ Specifically, the band at 681 cm⁻¹ is assigned to the A_1g mode, the band at 476 cm⁻¹ is the E_g mode and the bands at 194, 520 and 615 cm⁻¹ are assigned to the F_2g mode.²⁷ No vibration band related to either the Ni–O nor the Cr–O species was observed. The spectrum of commercially available Co₃O₄ powder is also shown. The A_1g mode appears at 689 cm⁻¹ and the E_g mode at 481 cm⁻¹, with a higher frequency shift of 5–8 cm⁻¹ compared with the micro/nano Co₃O₄ (Co₃O₄-1, Co₃O₄-3, Co₃O₄-5). This finding suggests that the micro/nano hierarchical structure may cause a slight structural distortion in spinel Co₃O₄, leading to a medium shift in the vibration modes towards a lower frequency.²⁶

Fig. 2 Raman spectra of micro/nano Co₃O₄ on a metal foil and the control Co₃O₄ powder (Co₃O₄-P).

The morphology of the Co₃O₄ grown on the foil was observed by SEM. The windmill-like microcrystals were found to grow vertically or were tilted towards the substrate with respect to the pristine Co₃O₄ (Fig. 3A). Detailed observations show a six-bladed windmill structure with a lateral size of about 10–12 μm. This structure has rarely been reported previously for the Co₃O₄ or doped Co₃O₄; nanowires or rods are more usually observed.^20,24 After treatment with an impregnation–calcination procedure, the structure of the microcrystals evolves (Fig. 3B–D). The needle-like branches grow up from the blades of the windmill structure. The density of the needles increases and secondary growth on the primary needle-like structure occurs when the amount of cobalt precursor impregnated increases from 1 to 3% (Fig. 3C). Such a growth mode could roughen the surface of the microcrystals, resulting in more catalytically active sites. However, the needles fuse together and become thicker when the amount of cobalt precursor impregnated is increased to 5% (Fig. 3D), which decreases the roughness and the number of redox-active species. When a pure Ni foil was used as the substrate, only densely grown Co₃O₄ nanowire bundles were observed (Fig. 4), significantly different from the windmill-like structure. Nevertheless, the digital photos of these samples show no visible difference; they all have a black appearance (Fig. S2†). Elemental analyses using EDS measurements suggest that the substrate for Co₃O₄ growth consists of Ni of 73%, Cr of 21% and a small amount of Fe and Mn (Fig. S3 and Table S1†). Therefore the presence of Cr (or Mn, Fe) in the foil could be a key factor in the formation of this structure. It has been reported that Co₃O₄ shows distinct morphologies when grown on different substrates.²⁸ We suggest that the growth mode of Co₃O₄ is affected by the presence of Cr (or Fe, Mn) during hydrothermal growth and subsequent calcination. The preferential adsorption of Cr, Fe or Mn species on the specific facets of Co₃O₄ at the initial stage possibly results in favourable development along the directions of the six branches in the structure. The exact growth mechanism needs to be studied further.

Fig. 3 SEM images of micro/nano Co₃O₄. (A) Co₃O₄, (B) Co₃O₄-1, (C) Co₃O₄-3 and (D) Co₃O₄-5.

Fig. 4 SEM images of Co₃O₄-NW grown on Ni foil.

To further investigate the structure and composition of the samples, elemental mapping was carried out using an EDS instrument attached to a transmission electron microscope. The results show that the micro/nano Co₃O₄ (Co₃O₄-3) consists of the elements cobalt and oxygen (Fig. 5). The amount of additional metal elements (Ni, Cr, Fe, Mn) is negligible, with an atomic ratio less than 0.2% for each element (Table S2†). The effect of the metal substrate on the EDS results cannot be completely avoided as the sample was mechanically scraped from the substrate for TEM characterization. As the micro/nano Co₃O₄ and untreated samples were grown on the same metal substrate, the effect of the substrate (i.e. the composition) on their performance could effectively be neglected. As a result, the difference in the electrochemical performance could be attributed to the structural effect alone, rather than being a compositional effect.

Fig. 5 Elemental mapping of Co₃O₄-3 sample by an EDS instrument attached to the transmission electron microscope.

The surface properties and elemental valences were further studied by XPS. The core level Co2p_3/2 in the Co₃O₄-3 sample was deconvoluted to four fitted peaks (Fig. 6A). The peaks at binding energies (BEs) of 780.0 and 786.0 eV can be assigned to Co³⁺ and its satellite line and the other couple at a higher BE (781.6 and 789.5 eV) can be assigned to Co²⁺ and its satellite line.²⁹ The BE value with a spin–orbit splitting of ca. 15.6 eV (795.6 eV in the Co2p_1/2 region to 780.0 eV in the Co2p_3/2 region) indicates that the Co³⁺ species are predominant on the surface.³⁰ The O1s core level spectra show two fitted peaks O_I and O_II (Fig. 6B). The peak at 530.1 eV (O_I) corresponds to the Co–O species and that at 531.5 eV (O_II) originates from low-coordinated oxygen defects or vacancies.^31,32 No Cr– (or Fe–, Mn–) related species were detected, which is indicative of the absence of such species on the surface of the windmill-like structures. The XPS spectra of Co₃O₄-NW grown on an Ni foil in a control measurement are also shown (Fig. 6C and D). The Co2p_3/2 core level was fitted to four peaks. The BE values corresponding to Co³⁺, Co²⁺ and their satellite lines show little shift compared with those in the Co₃O₄-3 sample. These results indicate that heavy doping of the heteroatoms (Cr, Fe, Mn) to Co₃O₄-3 is not possible as the doping usually causes a BE shift in the core levels.

Fig. 6 XPS core level spectra of Co₃O₄-3 sample. (A) Co2p, (B) O1s and Co₃O₄-NW, (C) Co2p and (D) O1s.

The electrochemical performance of the Co₃O₄ film electrodes towards the oxidation of water was evaluated in a three-electrode cell configuration (0.1 M KOH, pH 13). LSV curves are shown in Fig. 7. The Co₃O₄ electrodes present a lower overpotential and higher current density than the substrate, indicating the catalytic role of Co₃O₄ itself. Compared with the untreated Co₃O₄, the surface treatment further lowers the overpotential for impregnation amounts of 1–3%. The Co₃O₄-3 has an onset overpotential of 0.29 V, lower than that of Co₃O₄ (0.33 V, Table 1). The overpotential of Co₃O₄-3 at a current density of 10 mA cm⁻² is only 0.411 V, a remarkable decrease compared with Co₃O₄ (0.475 V, Table 1). This value is even lower than that of Co₃O₄-NW grown on an Ni foil (0.431 V). Nevertheless, the overpotential of Co₃O₄-5 increases to 0.45 V at the same current density (10 mA cm⁻²). The LSV curves of samples treated with 0.5 and 2% solute (cobalt salt) are shown in Fig. S4.† These results suggest that the sample treated with 3% solute is the optimum. A comparison of the mass activity of the Co₃O₄ samples shows that Co₃O₄-3 is the best catalyst with the highest mass activity (Table S3†).

Fig. 7 LSV curves tested in 0.1 M KOH solution with the Co₃O₄ electrodes as electrocatalysts.

Table 1 Electrochemical performance and surface properties of the Co₃O₄ electrodes

Sample	ηonset^a (V)	η10 mA cm^−2 (V)	Rf^b (10^3)	Active surface^c(cm^2)
a Overpotentials (η) calculated using the formula η = ERHE − 1.23 V, where ERHE is the potential referred to the RHE. The onset potential was defined by a tangent method applied to the LSV curves.^24b Roughness factor (Rf) calculated according to the formula Rf = Cdl/(60 × 10^−6 F cm^−2), where Cdl is the double-layer capacitor, i.e. the slope of the linear relationship between the peak current density and the scan rate; 60 × 10^−6 F cm^−2 is a reference value for the surface of oxides.^24c Active surface (A) calculated based on the Randels–Sevcik equation: Ip = 2.687n^3/2υ^1/2D^1/2AC, where n is the number of electrons transferred, υ^1/2 is the square root of the scan rate, D is the diffusion coefficient (1.9 × 10^−5 m^2 s^−1 in 0.1 M KOH), A is the active surface of the electrode and C is the concentration of the active substance.^34 The linear relationship of Ip/υ^1/2 is shown in Fig. S5.†
Co3O4	0.33	0.475	2.91	67
Co3O4-1	0.32	0.447	3.24	76
Co3O4-3	0.29	0.411	3.53	117
Co3O4-5	0.29	0.450	2.98	69
Co3O4-NW	0.30	0.431	3.86	128
Substrate	0.36	0.639	—	—

---

The difference in electrochemical performance may be associated with the structural effect or surface properties of the Co₃O₄. Consequently, the roughness factor (R_f) was calculated in terms of the double-layer capacitance in solution by assuming a standard value of 60 μF cm⁻² for the oxide surface.²⁴ The R_f value gradually increases with the amount of impregnation up to 3% (Table 1), although the R_f (10³) of Co₃O₄-5 shows a decrease down to 2.98, close to the value for pristine Co₃O₄ (2.91). This finding indicates that R_f has an optimum value for Co₃O₄-3, consistent with the electrochemical performance towards the oxidation of water. This clearly shows that R_f plays a critical part in the OEC properties of the electrocatalysts. Enhancing the surface roughness is therefore an efficient way of improving the electrocatalytic performance of a specific material.³³

The active surface (A) of each sample was calculated based on the Randels–Sevcik equation: I_p = 2.687n^3/2υ^1/2D^1/2AC to compare the number of active sites.³⁴ The calculated active surface presents a volcano-like curve, with the highest value of 117 cm² corresponding to Co₃O₄-3, 1.75 times higher than that of untreated Co₃O₄. A higher active surface indicates that more active sites are accessible for catalytic reactions. The analyses of the roughness and active surface are in good agreement with the morphological observations by SEM, where Co₃O₄-3 showed a better developed micro/nano hierarchical structure. Co₃O₄-3 has a similar active surface to Co₃O₄-NW (117 vs. 128 cm²), suggesting that the treatment significantly increases the number of active sites on pristine Co₃O₄ microcrystals.

The Tafel slope could reflect information about the reaction mechanism. The Tafel plots of the Co₃O₄ electrodes are shown in Fig. S6,† where the Tafel slopes were derived by fitting the data in the linear part of the plots. The slopes are 130, 121, 115 and 130 mV dec⁻¹ for Co₃O₄, Co₃O₄-1, Co₃O₄-3 and Co₃O₄-5, respectively. Tafel slopes ranging from about 40 to 120 mV dec⁻¹ have been reported for thermally prepared cobaltites^35–37 and this kinetic parameter is highly dependent on the preparation conditions. The slope of about 120 mV dec⁻¹ obtained in this work for the Co₃O₄ samples suggests that it could follow a mechanism involving a rate-limiting one-electron transfer with no kinetically relevant electrochemical pre-equilibrium.³⁸ The reactions involved are as follows:^38–40

Co₃O₄ + H₂O + OH⁻ ⇄ 3CoOOH + e⁻ (1)

CoOOH + OH⁻ ⇄ CoO₂ + H₂O + e⁻ (2)

Co(IV)_s + OH⁻ ⇄ Co(IV)(OH⁻)_ad (3)

Co(IV)(OH⁻)_ad + OH⁻ → Co(IV)_s + (1/2)O₂ + H₂O + e⁻ (4)

Co(IV)_s and Co(IV)(OH⁻)_ad represent the Co(IV) species on the oxide surface and the adsorbed species, respectively. Oxygen evolution occurs at potentials more positive than the oxidation of Co(III) to Co(IV) (about 1.5 V vs. RHE; Fig. 7), suggesting that Co(IV) is the active catalyst for the OER. The Co₃O₄-3 sample shows a lower Tafel slope, indicative of its improved kinetics for the oxidation of water.

To further evaluate the catalytic activity of the micro/nano Co₃O₄, the TOFs were calculated on the basis of the assumption that all Co ions within the structure are catalytically active.⁴¹ It was found that the oxidation peak current of the Co redox species has a linear dependence on the scan rate (Fig. S7†). The line slope is proportional to the surface concentration of the Co species (Γ₀) in terms of the equation: slope = n²F²AΓ₀/4RT. As a result, the TOF values can be calculated from the equation TOF = JA/4Fm, where J is the current density at a constant overpotential, A is the area of the electrode, 4 represents the number of moles of electrons consumed to evolve one mole of oxygen, F is Faraday's constant and m is the number of moles of active sites. TOF values as a function of overpotential are shown in Fig. 8. The TOF values of each sample show a linear increase for overpotentials from 0.35 to 0.65 V, in which the TOFs of Co₃O₄-3 remain higher than the other samples beyond an overpotential of 0.4 V (Fig. 8A). For example, the TOF of Co₃O₄-3 reaches 0.39 s⁻¹ at an overpotential of 0.6 V, about 1.4 times higher than pristine Co₃O₄ (0.28 s⁻¹). However, the TOFs of Co₃O₄-5 are lower than those of pristine Co₃O₄ at overpotentials beyond 0.55 V. This tendency agrees with the results of the electrochemical LSV measurements. It is an indication that Co₃O₄-3 is the optimum micro/nano structure with the highest electrocatalytic activity towards the oxidation of water. The release of O₂ bubbles from the surface of the Co₃O₄-3 electrode was visible with the naked eye at a current density of 10 mA cm⁻² (video 1 in the ESI†). Measurements of the amount of O₂ were made using an O₂ fluorescent probe. The oxygen evolution was monitored in a gas-tight cell at 1.64 V (vs. RHE) for 1 h. Around 73.2 μmol of O₂ were evolved after 1 h of electrolysis under these conditions (Fig. S8†). These results are evidence of the evolution of O₂ from the electrodes.

Fig. 8 (A) TOFs with respect to Co atoms in Co₃O₄ electrocatalysts as a function of the overpotential tested in 0.1 M KOH. (B) Comparison of TOFs of Co₃O₄-3 and Co₃O₄-NW tested in 1 M KOH.

The TOFs of Co₃O₄-3 and Co₃O₄-NW tested in 1 M KOH are compared (Fig. 8B). The former presents higher TOFs than the latter throughout the overpotential range 0.3–0.65 V. At an overpotential of 0.6 V, the TOF of Co₃O₄-3 is 1.26 times higher than that of Co₃O₄-NW (0.54 vs. 0.427 s⁻¹). This suggests that an optimum micro/nano hierarchical structure could be even more active than the nanowire.

It is known that the efficiency of catalysts for the oxidation of water depends on the pH of the solution. Co₃O₄ or cobalt oxide catalysts are commonly used in alkaline solution. Under such alkaline conditions, hydroxyl ions (OH⁻) are the dominant anions in solution and the O₂-evolving reaction is as follows:

4OH⁻(aq.) → O₂(g) + 2H₂O(aq.) + 4e⁻

Consequently, a higher pH (i.e. a larger OH⁻ concentration) promotes the evolution of oxygen. It can be seen that the current density and TOF values for Co₃O₄-3 at the same overpotential (300 mV) remarkably increase when the pH is increased from 13 to 14 (Table 2).

Table 2 Electrochemical performance comparison of Co₃O₄ electrocatalysts with previously reported values

Samples	Current density (mA cm^−2)	η (mV)	pH	TOF (s^−1)	Reference
Co3O4-3	2.71	330	13	0.0312	This work
Co3O4-3	1.58	300	13	0.0182	This work
Co3O4-3	5.02	300	14	0.0452	This work
Co3O4	—	328	14	0.0187	42
Co3O4	—	300	14	0.04	43

---

The TOFs of Co₃O₄-3 were compared with previously reported Co₃O₄ electrocatalysts (Table 2). The optimized catalyst Co₃O₄-3 has a TOF value of 0.0452 s⁻¹ at a 300 mV overpotential (pH 14), 2.4 times higher than the previously reported value for a Co₃O₄ nanocatalyst (0.0187 s⁻¹ at 328 mV, pH 14).⁴² The TOF value shows an increase of 13% compared with that of 0.04 s⁻¹ reported for a high-efficiency Co₃O₄ electrocatalyst under the same conditions.⁴³ The TOF results verify that the structured catalysts are able to retain the high efficiency of their powder counterpart.

The electrochemical impedance spectra can be used to analyze the electron transport and recombination properties.⁴⁴ The electrochemical impedance spectra of micro/nano Co₃O₄ were measured at an applied potential of 0.56 V (vs. SCE) in a 0.1 M KOH solution. Fig. 9 shows the Nyquist plots and the corresponding equivalent circuit. The semicircle geometry corresponds to the Faraday process and shows a Faraday impedance. The equivalent circuit indicates that the EIS profiles are composed of three components, i.e. a solution resistance (R_s), a charge transfer resistance across the electrode/solution interface (R_ct) and a constant phase (CP). The fitted value of R_ct is 5.5 Ω for the Co₃O₄-3 sample, which is the smallest resistance of all the Co₃O₄ electrodes. This indicates that Co₃O₄-3 has better charge transport properties during the oxygen-evolving reactions. The R_ct values reasonably reflect the catalytic activity, i.e. the smaller charge transfer resistance results in a higher OER activity.

Fig. 9 Electrochemical impedance spectra of the Co₃O₄ electrodes measured at a potential of 0.56 V (vs. SCE) in 0.1 M KOH solution.

To study the stability of the micro/nano Co₃O₄ electrodes, the I–t curves were measured at a constant applied bias (1.58 V vs. RHE) in alkaline solution (0.1 M KOH). The optimum Co₃O₄-3 electrocatalyst reached an initial current density of 3.4 mA cm⁻², which remained almost constant over a 12 h OER test with no noticeable decrease (Fig. 10A). In comparison, the untreated Co₃O₄ had a current density of 1.4 mA cm⁻² at the beginning. After a 1 h OER test the value slightly decreased to about 1.3 mA cm⁻² (a decrease of 7%), showing a comparable stability to Co₃O₄-3, although with a lower current density. It is understandable that the Co₃O₄ displays a high OER stability, as spinel-type Co₃O₄ has been reported to be a stable electrocatalyst towards the oxidation of water under alkaline conditions.^23,43 SEM observations revealed that the micro/nano hierarchical structure of Co₃O₄-3 showed no change, even after a 12 h OER test (Fig. 10B). The SEM images of other Co₃O₄ samples after a 12 h OER test are shown in Fig. S9† and show little morphological change. XRD characterization also confirmed that the Co₃O₄-3 kept its spinel structure after the OER test, consistent with that before the reactions (Fig. S10†). In addition, the LSV curve for Co₃O₄-3 after a 12 h OER operation (0.1 M KOH, 1.58 V vs. RHE) overlaps that of unused Co₃O₄-3 (Fig. S11†), suggesting excellent recyclability for the oxygen-evolving reactions. Consequently, the micro/nano Co₃O₄ grown directly on the metal substrate could be used as highly efficient, durable and engineered electrocatalysts and are promising for use in the production of clean fuels from the electrolysis of water.

Fig. 10 (A) Current–time curves in the presence of Co₃O₄ electrocatalysts at an applied potential of 1.58 V vs. RHE, 0.1 M KOH. The inset shows the 12 h OER stability of Co₃O₄-3. (B) SEM image of Co₃O₄-3 after 12 h OER test.

Conclusions

Unique windmill-like Co₃O₄ microcrystals were grown directly on a flexible substrate by a hydrothermal–calcination procedure. Post-treatment of these Co₃O₄ microcrystals results in secondary growth on the Co₃O₄ microcrystals, forming a micro/nano hierarchical structure. Structural studies have indicated that the roughness and active surface of pristine Co₃O₄ reaches an optimum value using such a treatment, leading to an enhanced performance in the electrochemical oxidation of water. The optimized micro/nano Co₃O₄ electrodes had an onset overpotential of 0.29 V in a 0.1 M KOH solution. The overpotential is only 0.41 V at a current density of 10 mA cm⁻². The TOF comparison suggests that micro/nano Co₃O₄ possesses a higher activity than previously reported Co₃O₄ nanowires and Co₃O₄ nanoparticles. In addition, the high stability towards the oxidation of water in alkaline solution was confirmed and is indicative of the durable and recyclable properties of these electrodes. This work provides a facile way to roughen the surface of electrocatalysts grown on a substrate to enhance the OER activity. Our findings give a further insight into the catalytic behavior of micro/nano structured catalysts and the mechanism of enhancing this activity.

Acknowledgements

This work was supported by the 973 Program (Grant no. 2011CBA00506), the National Natural Science Foundation of China (Grant no. 21376020) and the Program for Changjiang Scholars and Innovative Research Team in University (Grant no. IRT1205).

Notes and references

1. D. G. Nocera, Acc. Chem. Res., 2012, 45, 767–776.
2. N. S. Lewis and D. G. Nocera, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 15729–15735.
3. L. Meda and L. Abbondanza, Rev. Adv. Sci. Eng., 2013, 2, 200–207.
4. M. W. Kanan and D. G. Nocera, Science, 2008, 321, 1072–1075.
5. J. J. Concepcion, J. W. Jurss, M. K. Brennaman, P. G. Hoertz, A. O. T. Patrocinio, N. Y. Murakami Iha, J. L. Templeton and T. J. Meyer, Acc. Chem. Res., 2009, 42, 1954–1965.
6. Y. V. Geletii, Z. Q. Huang, Y. Hou, D. G. Musaev, T. Q. Lian and C. L. Hill, J. Am. Chem. Soc., 2009, 131, 7522–7523.
7. L. Wang, L. Duan, B. Stewart, M. Pu, J. Liu, T. Privalov and L. Sun, J. Am. Chem. Soc., 2012, 134, 18868–18880.
8. T. Reier, M. Oezaslan and P. Strasser, ACS Catal., 2012, 2, 1765–1772.
9. Y. Lee, J. Suntivich, K. J. May, E. E. Perry and Y. Shao-Horn, J. Phys. Chem. Lett., 2012, 3, 399–404.
10. M. W. Kanan, Y. Surendranath and D. G. Nocera, Chem. Soc. Rev., 2009, 38, 109–114.
11. Q. Yin, J. M. Tan, C. Besson, Y. V. Geletii, D. G. Musaev, A. E. Kuznetsov, Z. Luo, K. I. Hardcastle and C. L. Hill, Science, 2010, 328, 342.
12. V. Artero, M. Chavarot-Kerlidou and M. Fontecave, Angew. Chem., Int. Ed., 2011, 50, 7238–7266.
13. A. Singh and L. Spiccia, Coord. Chem. Rev., 2013, 257, 2607–2622.
14. A. Singh, S. L. Y. Chang, R. K. Hocking, U. Bachcde and L. Spiccia, Energy Environ. Sci., 2013, 6, 579–586.
15. P. Du and R. Eisenberg, Energy Environ. Sci., 2012, 5, 6012–6021.
16. J. B. Gerken, J. G. McAlpin, J. Y. C. Chen, M. L. Rigsby, W. H. Casey, R. David Britt and S. S. Stahl, J. Am. Chem. Soc., 2011, 133, 14431–14442.
17. M. Gong, Y. G. Li, H. L. Wang, Y. Y. Liang, J. Z. Wu, J. G. Zhou, J. Wang, T. Regier, F. Wei and H. J. Dai, J. Am. Chem. Soc., 2013, 135, 8452–8455.
18. A. Han, H. Wu, Z. Sun, H. Jia and P. Du, Phys. Chem. Chem. Phys., 2013, 15, 12534–12538.
19. S. Chen and S.-Z. Qiao, ACS Nano, 2013, 7, 10190–10196.
20. X. Liu, Z. Chang, L. Luo, T. Xu, X. Lei, J. Liu and X. Sun, Chem. Mater., 2014, 26, 1889–1895.
21. Z. Lu, W. Xu, W. Zhu, Q. Yang, X. Lei, J. Liu, Y. Li, X. Sun and X. Duan, Chem. Commun., 2014, 50, 6479–6482.
22. M. Hamdani, R. N. Singh and P. Chartier, Int. J. Electrochem. Sci., 2010, 5, 556–577.
23. J. A. Koza, Z. He, A. S. Miller and J. A. Switzer, Chem. Mater., 2012, 24, 3567–3573.
24. Y. Li, P. Hasin and Y. Wu, Adv. Mater., 2010, 22, 1926–1929.
25. X. Zou, J. Su, R. Silva, A. Goswami, B. R. Sathemand and T. Asefa, Chem. Commun., 2013, 49, 7522–7524.
26. J. Yang, H. W. Liu and W. N. Martens, et al., J. Phys. Chem. C, 2010, 114, 111–119.
27. V. Khadzhiev, M. Iliev and I. Vergilov, J. Phys. C: Solid State Phys., 1988, 21, L199.
28. J. Jiang, J. P. Liu, X. T. Huang, Y. Y. Li, R. M. Ding, X. X. Ji, Y. Y. Hu, Q. B. Chi and Z. H. Zhu, Cryst. Growth Des., 2010, 10, 70–75.
29. X. Xiang, L. Zhang, H. I. Hima, F. Li and D. G. Evans, Appl. Clay Sci., 2009, 42, 405–409.
30. J. C. Dupin, D. Gonbeau, I. Martin-Litas, P. Vinatier and A. Levasseur, J. Electron Spectrosc. Relat. Phenom., 2001, 120, 55–65.
31. V. M. Jiménez, A. Fernández and J. P. Espinós, et al., J. Electron Spectrosc. Relat. Phenom., 1995, 71, 61–71.
32. J. K. Chang, C. M. Wub and I. W. Sun, J. Mater. Chem., 2010, 20, 3729–3735.
33. M. S. Faber, R. Dziedzic, M. A. Lukowski, N. S. Kaiser, Q. Ding and S. Jin, J. Am. Chem. Soc., 2014, 136, 10053–10061.
34. A. J. Bard and L. R. Faulkner, Electrochemical Methods Fundamentals and Applications, John Wiley & Sons, 2nd edn, 2003.
35. R. N. Singh, J. F. Koenig, G. Poillerat and P. Chartier, J. Electrochem. Soc., 1990, 137, 1408.
36. X. Nikolov, R. Darkaoui, E. Zhecheva, R. Stoyanova, N. Dimitrov and T. Vitanov, J. Electroanal. Chem., 1997, 429, 157.
37. R. N. Singh, M. Hamdani, J. F. Koenig, G. Poillerat, J. L. Gautier and P. Chartier, J. Appl. Electrochem., 1990, 20, 442.
38. J. B. Gerken, J. G. McAlpin, J. Y. C. Chen, M. L. Rigsby, W. H. Casey, R. D. Britt and S. S. Stahl, J. Am. Chem. Soc., 2011, 133, 14431–14442.
39. E. B. Castro and C. A. Gervasi, Int. J. Hydrogen Energy, 2000, 25, 1163–1170.
40. E. B. Castro, C. A. Gervasi and J. R. Vilche, J. Appl. Electrochem., 1998, 28, 835–841.
41. Y. Surendranath, M. W. Kanan and D. G. Nocera, J. Am. Chem. Soc., 2010, 132, 16501–16509.
42. A. J. Esswein, M. J. McMurdo and P. N. Ross, et al., J. Phys. Chem. C, 2009, 113, 15068–15072.
43. N. H. Chou, P. N. Ross and A. T. Bell, et al., ChemSusChem, 2011, 4, 1566–1569.
44. N. Wang, D. Wang, M. Li, J. Shi and C. Li, Nanoscale, 2014, 6, 2061–2066.

---

Footnote

† Electronic supplementary information (ESI) available: XRD pattern, EDX spectra, elemental analysis, additional electrochemical measurements, and stability studies. See DOI: 10.1039/c4ra07082h

---