Co-doped CeO₂/N–C nanorods as a bifunctional oxygen electrocatalyst and its application in rechargeable Zn-air batteries

Ce(NO₃)₃·6H₂O (≥99.95%, Aladdin), C₆H₃(CO₂H)₃ (≥99.00%, Aladdin), C₂H₇NO (≥99.00%, AR, Sinopharm Chemical Reagent Co., Ltd, China), Co(NO₃)₂·6H₂O (≥99.0%, AR, Aladdin), NaBH₄ (≥98% Xuzhou Tianhong Chemical Co, Ltd, China), All the chemicals were used without further purification, and the water used were deionized (DI) water (18.2 MΩ cm⁻²).

Ce-MOF precursor was synthesized according to the literature [37, 38] with slightly modified. Typically, 1 mmol Ce (NO₃)₂·6H₂O was dissolved in a mixture of 25 ml water-ethanol solvent (1:1, V/V) to obtain Solution I. 1 mmol trimesic acid (1, 3, 5-H₃BTC) was also dispersed in the same 25 ml mixture solution to gain Solution II. Subsequently, Solution I was added into Solution II under stirring for 2 h in a water bath at 90 °C. The white precipitation was washed by DI water and ethanol, followed by 80 °C into a vacuum oven and dried for 6 h to collect. The Ce-MOF was heated at 900 °C for 2 h in a tube furnace (2 °C min⁻¹) under Ar flow to get CeO₂@N–C composite.

0.4 g CeO₂@N–C and 10 ml ethanolamine (C₂H₇NO) were evenly dispersed in 25 ml DI water with stirring at 50 °C for 3 h. Then 0.4 g Co(NO₃)₂·6H₂O was added into the solution under stirring for 3 h until 0.4 g NaBH₄ was added. After that, the solution was carefully transferred into a 100 ml Teflon-lined stainless-steel autoclave and heated at 80 °C for 6 h. After cooling to room temperature naturally, the product was washed with DI water until the pH of solvent became neutral and then dried at 60 °C for 6 h in an oven. Finally, the black powder of Co₃O₄-CeO₂@N–C composite was obtained.

The morphology, composition, and structure of as-prepared products were characterized by scanning electron microscope (SEM), (Hitachi S4800), transmission electron microscope (TEM, JEOL 2100F), spherical aberration (Cs) corrected TEM (FEI Titan Themis 60–300), powder x-ray diffraction (XRD, Rigaku Ultima IV) and x-ray photoelectron spectroscopy (XPS, PHI Quantum-2000).

Electrochemical measurements were conducted in a conventional three-electrode system controlled by CHI760E electrochemical station. The rotating disk electrode (RDE, diameter: 4 mm) and rotating-ring disk electrode (RRDE) experiments were performed on the RRDE-3A. Graphite rod and Hg/HgO electrode (1 M KOH solution) were used as counter electrode and reference electrode, respectively. The catalyst ink was prepared by dissolving 5 mg catalyst into 1 ml isopropanol and 5 μl 5% Nafion solution and sonicating for 30 min to form catalyst suspension. Then 30 μl of the ink was dropped on the glassy carbon electrode as the working electrodes. The linear sweep voltammetry (LSV) was performed at a scan rate of 5 mV s⁻¹ in O₂-saturated 0.1 M KOH with rotating from 400 to 2500 rpm for ORR. The electron transfer number (n) and % ${{{\rm{HO}}}_{2}}^{-}$ were determined in combination with the RRDE technique:

$$\begin{eqnarray} &&\% {{\rm{HO}}}_{2}^{-}={\rm{200}}\times \displaystyle \frac{{{I}}_{{R}}/{N}}{{{I}}_{{D}}+{{I}}_{{R}}/{N}},\end{eqnarray} \tag{ 1 }$$

$$\begin{eqnarray}&&n\,=\,4\times \displaystyle \frac{{{I}}_{{D}}}{{{I}}_{{D}}+{{I}}_{{R}}/{N}},\end{eqnarray} \tag{ 2 }$$

where I_D and I_R represent the disk and ring currents, respectively. And N is the collection efficiency of the Pt ring (N = 0.42). All the electrochemical measurements were done at room temperature.

Zinc-air battery (ZAB) tests were performed in a home-made battery. The catalyst ink was drop-casted onto a Teflon-coated carbon cloth (W0S1002) with a catalyst loading amount of 1.2 mg cm⁻² as the gas diffusion electrode (GDE). The achieved GDE was dried under vacuum at 85 °C. A control sample is also prepared by using 0.24 mg Pt/C. For battery tests, the GDE layer was used as the air cathode, a polished Zn foil as anode and 6.0 M KOH with 0.2 M Zn(CH₃COO)₂ as the electrolyte. The discharge/charge curves were recorded in ambient on a Neware battery test instrument (Shenzhen, China).

The synthetic route of the catalyst is shown in figure 1. First, based on previous literature reports, the Ce-MOF₉₀₀ precursor was slightly modified to prepare a bundled Ce-MOF₉₀₀ (Ce³⁺/Ce⁴⁺) carbon–nitrogen nanorod precursor template with mixed valence. Then the amine group of ethanolamine chelated the metal cobalt ion, and the hydroxyl group at the other end was adsorbed and anchored on the carbon–nitrogen nanorod precursor template. Secondly, cobalt ions were reduced to Co₃O₄ nanoparticles by NaBH₄ through low temperature hydrothermal method, which are highly uniform and dense anchored on the CeO₂ carbon–nitrogen framework, avoiding the agglomeration. Finally, Co₃O₄-CeO₂@N–C catalyst with mixed valence state (Ce³⁺/Ce⁴⁺) is prepared successfully.

Zoom In Zoom Out Reset image size

The crystal structure and surface chemical composition were investigated by XRD and XPS, respectively. Figure 2 is the XRD pattern including the two samples. The CeO₂/NC (i) shows diffraction peaks at 2θ angles of 28.58°, 33.16°, 47.52°, 56.46°, 59.39°, 69.50°, 76.74° and 79.09°, which belong to the (111), (200), (220), (311), (222), (400), (331) and (420) planes of CeO₂ (PDF#89–8436), respectively. After Co doping, Co₃O₄-CeO₂@N–C (ii) exhibits the new diffraction peaks at 19.68° and 38.50°. Both the peaks are corresponding to the (111) and (222) planes of Co₃O₄ (PDF#43–1003), indicating the Co element was successfully doped. Furthermore, the strength of peaks is weak because it is covered by the CeO₂@N–C, which can be further proved by the results of TEM. In addition, after comparison, the peaks width of Co₃O₄-CeO₂@N–C is larger than CeO₂@N–C. Thus, the grain size of Co₃O₄-CeO₂@N–C (∼65 nm), computed by the Debye–Scherrer equation, is smaller than those of CeO₂@N–C (∼80 nm). Smaller particle size is beneficial to increase the specific surface area and further increase the active sites to increase the catalytic performance.

Zoom In Zoom Out Reset image size

To explore the effect of Co doping on the surface chemical composition and valence state of Co₃O₄-CeO₂@N–C and CeO₂@N–C, XPS measurement was also employed. Figure 3(a) shows the XPS survey spectrum of Co₃O₄-CeO₂@N–C. Typical peaks of Ce_3d (924–878 eV), Co_2p (805–770 eV), C_1s (280–295 eV), N_1s (393–410 eV) and O_1s (526–536 eV) is discovered in corresponding energy range, which can illustrate the co-existence of Ce, Co, C, N and O elements. Peak fitting of the elements core-level is all obtained after deduction of Shirley background and is shown in figures 3(b)–(f). In the high-resolution Ce_3d spectrum (figure 3(b)), it can be deconvolved into eight peaks. The peaks are denoted as μ and ν to identify the spin–orbit composition of Ce 3d_5/2 and Ce 3d_3/2, respectively. Two peaks which labeled as μ' and ν' located at 903.16 eV and 885.12 eV are assigned to the Ce³⁺. The remaining six peaks of μ''' (916.82 eV), μ'' (907.62 eV), μ° (901.09 eV), ν''' (898.49 eV), ν'' (888.89 eV) and ν° (882.53 eV) are attributed to the characteristic peaks of Ce⁴⁺ [23, 39, 40]. Therefore, Ce³⁺ and Ce⁴⁺ are co-existence in the Co₃O₄-CeO₂@N–C and the mainly oxidation states were tetravalent [23, 40]. The electronic state of Ce³⁺ is 3d¹⁰4f¹, so there is no doubt about the exist of lone electron (4f¹ orbital). Lone electron is beneficial to enhance the interaction between ceria and the surrounding Co atoms. In the other hand, the percentage contents of Ce ³⁺ (μ' + ν') and Ce ⁴⁺ (μ''' + μ'' + μ° + ν''' + ν'' + ν°) can get by calculate the fitted area that under the corresponding peaks. Co₃O₄-CeO₂@N–C (0.194) has the larger ratio value than CeO₂@N–C (0.188), proving more defects in Co₃O₄-CeO₂@N–C. Moreover, the exist of Ce³⁺ can not only donate charge compensation, but also help to form the unsaturated chemical bonds in the CeO₂ crystal. Both of them are in favor of generating oxygen vacancies and promoting the oxygen adsorption on the catalyst surface subsequently [39, 41].

Zoom In Zoom Out Reset image size

Figure 3(c) shows the XPS survey spectrum of Co_2p. The peak at 781.30 eV belongs to Co²⁺, while the peak at 796.81 eV is ascribed to Co³⁺. Generally, Co³⁺ plays an important role in the electrocatalytic performance toward ORR [42]. More details with C_1s is depicted in figure 3(d) and it can be divided into four peaks (C₁–C₄). Two main peaks located at 284.38 eV (C₁) and 286.22eV (C₃) are attributed to C–C/C–N and C–O [43, 44], which are generated from the organic ligands. In addition, there are the other two weak peaks at 285.08 eV (C₂) and 289.38 (C₄), which is responsible for the characteristic of C–O and π–π* [39]. The high resolution XPS spectrum of N_1s shown in figure 3(e) are fitted by three parts at 399.08 eV (N₁), 400.19 eV (N₂) and 404.41 eV (N₃), corresponding to pyrrolic-N, graphitic-N and oxidized-N[43, 45]. Based on previous reports in open literatures, the graphitic-N is generally considered to greatly improve the ORR catalytic performance of N-doped carbon materials, especially the limiting current density on ORR[45, 46].

The O_1s XPS spectrum is shown in figure 3(f), where the peak at 529.61 eV (O₁) stands for the lattice oxygen (O_latt), while the peak at 531.82 eV (O₂) represents the surface active oxygen (O_sur). Surface active oxygen consist of superoxide species (O^2-) and peroxide species (O^-), both of them are activated oxygen species which comes from the oxygen defects [47]. The ratios of O_sur/O_latt of Co₃O₄-CeO₂@N–C₉₀₀ (0.31) are higher than CeO₂@N–C₉₀₀ (0.19), which is meets the catalytic activities of the samples.

The morphology information of the Co₃O₄-CeO₂@N–C were obtained by SEM and TEM. From the low magnification SEM images (figure 4(a)), the Co₃O₄-CeO₂@N–C are composed by bundles of nanorods. Moreover, the high magnification SEM image shown in figure 4(b) indicates that each nanorod with three dimensions is about 2 μm height, 100 nm length and width. Compared with Co₃O₄-CeO₂@N–C, CeO₂@N–C exhibited fewer distributed nanorods (figure S1 (available online at stacks.iop.org/NANO/33/415404/mmedia)). What is more, at higher resolutions, CeO₂ nanoparticles agglomerate on the surface of nanorods, while distribute uniformly after Co modification. Such morphology could inevitably lead to lower specific surface area and less available active sites, affecting their catalytic performance seriously. In TEM image (figure 4(c)), the structure of nanorods bundles of Co₃O₄-CeO₂@N–C can be further confirmed, and some of CeO₂ grain which closed to hexagon also can be observed on the nanorod (figure 4(d)). As shown in high-resolution TEM image of figure 4(e), Co₃O₄ and CeO₂ nanoparticles distribute on the carbon–nitrogen nanorods uniformly, and the close contact between them was confirmed by lattice analysis. As a result, the lattice fringes of 0.32 nm, 0.25 nm and 0.28 nm correspond to the (112) plane of CeO₂ and the (311), (220) plane of Co₃O₄, respectively. Obviously, Co₃O₄ nanoparticles anchored on the surface are expected to cooperate with CeO₂ to promote the ORR/oxygen storage performance. The select area electron diffraction (SAED) pattern (figure 4(f)) shows well-defined diffraction rings, which can be well indexed to the (111), (200), (220), (422), (311) and (422) planes. All the planes are origin from CeO₂ except (422) plane is attributed to the Co₃O₄. In addition, the elements distribution in Co₃O₄-CeO₂@N–C is measured by mapping (figure 4(g)) and EDS (figure 4(h)), which further verified the co-existence of C, N, O, Ce, and Co, and their uniform distribution on the entire carbon support.

Zoom In Zoom Out Reset image size

Figure 5(a) shows the RRDE curves of CeO₂@N–C, Co₃O₄-CeO₂@N–C and commercial Pt/C (20 wt%) toward ORR at 1600 rpm. Evidently, the ORR activity of CeO₂@N–C is very low due to the most negative RRDE curve. After the introduction of Co₃O₄, the Co₃O₄-CeO₂@N–C catalyst shows more positive E₀ (0.80 V), half-wave potential (0.75 V) and the Pt-like limiting current density (6.0 mA cm⁻²), indicating that the Co₃O₄-CeO₂@N–C catalyst possesses more active sites and the catalytic performance toward ORR is enhanced. Through the RRDE measurement, we also calculated the yield of H₂O₂ and the electron transfer number (n) to obtain selectivity and electron transfer pathway during ORR, as shown in figure 5(a) and (c). Apparently, the disk current density corresponding to ORR is significantly larger than the ring current density arisen from the H₂O₂ oxidation, implying the predominant product of OH^‒. The H₂O₂ yield observed on Co₃O₄-CeO₂@N–C remains below 12% in the potential range of 0.25–0.6 V and the n is approximately 4, indicating an ideal four-electron transfer pathway proceeded on Co₃O₄-CeO₂@N–C. Particularly, the Co₃O₄-CeO₂@N–C catalyst also showed excellent electrocatalytic activity for OER (figure 5(d)). To obtain the current density of 10 mA cm⁻², the potential required by Co₃O₄-CeO₂@N–C and Pt/C was 1.65V and 1.82V, respectively, that is, Co₃O₄-CeO₂@N–C shows better OER performance. What is more, according to previous literatures [48, 49], a smaller potential gap between ORR and OER (ΔE = E_J=10-E_1/2) indicates better catalytic activity. Compared with Pt/C (ΔE = 0.98 V), Co₃O₄-CeO₂@N–C has a smaller ΔE of 0.92 V, indicating that Co₃O₄-CeO₂@N–C has better ORR and OER bifunctional catalytic activity. The Tafel slope corresponding to Co₃O₄-CeO₂@N–C is 81 mV dec⁻¹ (figure 5(e)), which is smaller than that of Pt/C (121 mV dec⁻¹) and CeO₂@N–C (141 mV dec⁻¹), illustrating that Co₃O₄-CeO₂@N–C is conducive to generating better dynamic processes. Moreover, the long-term stability of the Co₃O₄-CeO₂@N–C was investigated under 0.7 V in O₂-saturated 0.1 M KOH solution (figure 5(f)). After 40 000 s of continuous test, the current density of Pt/C and CeO₂@N–C only retained 66.1% and 42.3%, respectively. While those of Co₃O₄-CeO₂@N–C could reach 81.1%. This result confirms that Co₃O₄-CeO₂@N–C has superior ORR stability than those of Pt/C and CeO₂@N–C catalysts.

Zoom In Zoom Out Reset image size

To further evaluate the practical application potential of the catalyst in Zn-air batteries, Teflon carbon cloth coated with CeO₂@N–C, Co₃O₄-CeO₂@N–C and commercial Pt/C (20 wt%) catalysts were used as the air cathode to assemble ZABs. As shown in figure 6(a), the open-circuit voltage of the ZAB assembled by the Co₃O₄-CeO₂@N–C is 1.46 V, which is higher than that with CeO₂@N–C cathode and only 50 mV smaller than that with Pt/C, suggesting its good catalytic performance. The maximum power density for ZAB assembled by the Co₃O₄-CeO₂@N–C reaches about 110 mW cm⁻² at around 180 mA cm^–2 (figure 6(a)), which outperforms CeO₂@N–C cathode (90 mW cm^–2) and up to 90% of Pt/C. In addition, at the constant discharge current density of 10 mA cm⁻², ZAB assembled by the Co₃O₄-CeO₂@N–C has the highest specific capacity of 780 mAh g⁻¹ (based on the mass of consumed Zn, figure 6(b)). This is superior to CeO₂@N–C (447 mAh g⁻¹) and commercial Pt/C (770 mAh g⁻¹). Moreover, the rechargeability of the air electrode at a current density of 5 mA cm⁻² was evaluated (10 min per cycle and 200 cycles) and the results are shown in figure 6(c). Co₃O₄-CeO₂@N–C shows excellent cycle durability and has the lowest initial charge/discharge voltage gap (0.89 V) in the initial dozens of cycles than those of CeO₂@N–C (1.42 V) and Pt/C (1.1 V). Besides, for Co₃O₄-CeO₂@N–C, the voltage gap of ZAB does not increase significantly within 40 h, which proved that it had high round-trip efficiency, while the gap of ZAB with Pt/C catalyst increases obviously. Finally, by testing at different discharge rates, the rate performance of Co₃O₄-CeO₂@N–C based ZABs was evaluated (figure 6(d)). Under different current densities (5 –100 mA cm⁻²), the discharge potential of ZABs remained basically stable. Especially after high-rate discharge at high current density, the discharge potential could recover at 5 mA cm⁻², which proved the remarkable reversibility for Co₃O₄-CeO₂@N–C. All the above results indicate that the ZABs with Co₃O₄-CeO₂@N–C catalysts exhibit a superior charge–discharge capacity and stability due to their excellent bifunctional ORR/OER electrocatalytic activity and stability, which outperforms most of other recently reported low-cost bifunctional oxygen electrocatalysts (table 1).

Zoom In Zoom Out Reset image size