Defect-rich boron doped carbon nanotubes as an electrocatalyst for hybrid Li–air batteries

Abstract

Carbon nanotubes (CNTs) have been used as a cathode material for hybrid Li–air batteries (HLABs); however, their catalytic abilities for the oxygen reduction reaction (ORR) and oxygen evolution reaction (OER) are relatively limited. Herein, to improve their catalytic performance, we have synthesized a boron-doped CNT (BC₃NT) with varying boron contents. The Raman spectrum proves that the more defects are due to the higher boron content. And HLABs based on BC₃NT show great cyclability and catalytic performance in air, with an overpotential value of 0.3 V for 200 hours at a current density of 0.05 mA cm⁻². Furthermore, theoretical calculations with density functional theory (DFT) confirm the effects of topological defects of BC₃NT on the ORR and OER. Since boron is doped into the CNT, BC₃NT containing a large amount of defect rings is obtained. And the generation of defects was conducive to the battery reaction. Accordingly, as a positive electrode material for HLABs, BC₃NT is more conducive to the ORR and OER, and the topological defect ring generated by it improved the ORR and OER performance.

---

Introduction

Lithium-ion batteries have brought a significant contribution to our daily lives, but the energy density of Li-ion batteries is close to the theoretical limit. With the development of industrial technology, the demand for energy utilization increases, and therefore it is imperative to develop new energy storage devices with high energy density.^1–6 Littauer et al.⁷ proposed the concept of Li–air batteries (LABs) in 1976. LABs have attracted widespread attention due to their high theoretical energy density, which can reach 11 140 Wh kg⁻¹, almost 10 times that of lithium-ion batteries.^8–12 As the reactant, O₂ is taken directly from the environment in this system, entering from the positive electrode.¹³ It reacts with Li⁺ in the electrolyte, and the cell reaction is: 2Li + O₂ ↔ Li₂O₂. Although the theoretical energy density of LABs is extremely high, many serious problems still have to be solved:¹⁴ for example, lithium dendrites, poor cycling performance, products blocking the porous positive electrode, low actual energy density, low catalytic efficiency, etc.^15–18 The problem of lithium dendrite is common to all lithium batteries; the porous cathode with insoluble material blockage is unique to LABs.¹⁹ To solve the problem, Zhou et al.²⁰ proposed HLABs based on LISICON in 2010, and their cell reaction is: 4Li + O₂ + 2H₂O ↔ 4LiOH. The system comprises a lithium anode, organic-electrolyte, LISICON membrane, water-electrolyte, and porous cathode. Li⁺ can pass through the LISICON membrane preventing the mixing of organic-electrolyte and water-electrolyte.²¹ Compared with LiTFSI-TEGDME organic-electrolyte, LiOH aqueous solutions used in HLABs have higher diffusion coefficients of Li⁺ and O₂. The product LiOH has high solubility. For every 100 g water, 12.8 g LiOH can be dissolved. LiOH is not easy to plug the porous positive electrode, convenient for continuous discharge.^22,23 The electrolytes used in the cathode chamber and anode chamber are LiOH and LITFSI-TEGDME, respectively. The battery is considered a vital candidate to replace Li-ion batteries because of a series of advantages (Scheme 1).

Scheme 1 Structure of HLABs.

Porous cathode materials and catalysts play an essential part in LABs. Reasonable porous structure and good catalytic performance are indispensable. CNTs have been extensively used as cathode materials for LABs due to their large specific surface area and porous structure.^24–27 Yi et al.²⁸ studied and compared the ORR, OER, and adsorption mechanism of the discharge product of LABs on CNT electrodes. The results show that the distance between the CNT layers should be at least 3.4 Å to avoid an interaction. Moreover, multi-walled carbon nanotubes (MWCNTs) perform better than single-walled carbon nanotubes (SWCNTs) in inhibiting the initial nucleation of lithium oxide and promoting the initial nucleation of Li₂O₂. A comprehensive analysis of adsorption energy and electronic structure shows that the inner structure of Li₂O₂ is composed of small clusters in addition to the coverage area. Li et al.²⁹ used the sol–gel method to uniformly coat RuO₂ on the surface of CNTs as the positive electrode of the LABs, which showed excellent cycling performance. Lim et al.³⁰ prepared an independent mesh cathode composed of CNT-loaded polymer nanofibers to develop practical LABs with a specific energy of 814 Wh kg⁻¹. Chen et al.³¹ reported the inhibition of side reactions by preparing a MnO₂ composite CNT material as the positive electrode and using salicylic acid as the electrolyte, making the battery cycle up to 300 cycles. Using a metal or metal oxide as the catalyst can improve battery performance. However, it is expensive and difficult to prepare, so it is necessary to find a cathode catalyst material that is cheap and easy to prepare.^32–36 Researchers begin to pay attention to nonmetals as catalysts. Boron atoms and carbon atoms are in the same period, their atomic numbers are adjacent, and their radii are similar. It is straightforward to dope them into CNTs, which attracts widespread attention from researchers.^37–39 Li et al.⁴⁰ applied BC₃ nanosheets to LABs, and their calculations showed that the BC₃ structure helps adsorb oxygen molecules and effectively reduces the overpotential. Boron doping can form BC₃ with a one-dimensional nanotube structure in carbon nanotubes.^41–43 BC₃NTs have the advantages of large specific surface area and good catalytic performance, so they are suitable as cathode materials for LABs.^44–46 BC₃NTs contain C–C and B–C bonds but have no unstable B–B bonds, so these structures are stable.^47–49 Defects are one of the characteristics of BC₃NTs, and these defect structures will have a favorable effect on the mass transfer and electrode reaction of LABs.⁵⁰ Therefore, BC₃NTs comply with the requirements of the positive electrode of LABs with mixed electrolytes and are worthy of further study.

In this paper, BC₃NTs with different B proportions were prepared as the positive electrode materials of HLABs. According to Raman spectrum analysis, increasing the boron doping ratio would increase the number of defects, conducive to the electrode reaction. Tests show that BC₃NTs with defects as a positive electrode can make the HLABs cycle up to 165 cycles and maintain an ultra-low overpotential of 0.3 V in the first 200 hours. DFT was used to further study the effect of topological defect rings of different sizes on the electrode reactions.

Experimental

Synthesis of materials

BC₃NTs were fabricated with B₂O₃ powder (purity 99.9%) and CNTs in an argon glove box. The mass ratio of B₂O₃ powder (purity 99.9%) and CNTs are 2 : 1, 5 : 1, and 8 : 1, respectively. Then, they were placed into a vacuum-sealed quartz tube heating to 1050 °C at a rate of 10 °C min⁻¹, and the products were kept for 10 h at 1050 °C. They were taken out after the resistance furnace was cooled to room temperature as shown in Fig. S1 in the ESI.† The experimental drugs used in this study are shown in Table S1 in the ESI.†

Characterization

A scanning electron microscope (SEM, JEM-7800F.) was used to investigate the morphologies of the samples. The valence states of the samples were analyzed by using X-ray photoelectron spectroscopy (XPS, ESCALAB 250). Raman spectra were collected on a HORIBA XploRA PLUS.

Electrochemical characterization

BC₃NT and PTFE solutions were stirred to form a uniform mixture with a mass ratio of 9 : 1, and an appropriate amount of N-methyl-2-pyrrolidone (NMP) dispersant was added to the mixture and stirred to form a uniform and thick black paste. The black paste was coated on carbon paper, and a positive plate with a diameter of 16 mm was obtained by slicing after vacuum drying at 80 °C for 8 h. HLABs adopt the shape of the mold designed by the research group, as shown in Fig. S2 in the ESI,† which is composed of a die holder, negative current collector, lithium foil, LISICON film (Li_1+X+Y(Al, Y)_x(Ti, Ge)_2−XSi_YP_3−YO₁₂), punch set, positive electrode, and positive current collector. The anode chamber electrolyte shall be 1 mol L⁻¹ LITFSI-TGEDME, and the cathode chamber electrolyte shall be 0.1 mol L⁻¹ LiOH. The roundtrip performance measurement shall be carried out at a current density of 0.05 mA cm⁻² and a limited capacity of 50 mAh g⁻¹, and the initial discharge capacity measurement shall be carried out at a current density of 0.05 mA cm⁻².

DFT calculation

Firstly, a CNT model of (8,0) was established, and then some of the C atoms was replaced by B atoms to obtain BC₃NT, as shown in Fig. 1(a). There were 128 atoms in total, including 32 B atoms, with C atoms represented in gray and B atoms in purple. A seven-membered ring topology defect (7MR) was created by rotating one of the bonds in the nanotube by 90°, forming two heptagons with common bonds and two pentagons separated by bonds as shown in Fig. 1(b). Similarly, an eight-membered ring topology defect (8MR) was created and is shown in Fig. 1(c). A density functional theory calculation was performed using the Dmol³ module of Materials Studio 2019 software. The nanotube was placed in the crystal cell of 20 Å × 20 Å × 17.08 Å, and the model was established as shown in Fig. S3 in the ESI.† The bond lengths of B–C and C–C were 1.55 Å and 1.41 Å, respectively.

Fig. 1 Schematic diagram of the nanotube structure: (a) BC₃NT, (b) 7MR, and (c) 8MR.

The energy convergence of the optimization process is shown in Fig. S4 in the ESI.† The spin was not restricted in the calculation. The PBE functional under the Generalized Gradient Approximation (GGA) was used in the electron exchange-correlation function. The Grimme method was used for dispersion correction to accurately describe the van der Waals forces between the reactants and the substrate. The basis set we used was DNP, double numeric quality basis set with polarization function that is implemented by default in the DMol³ package. The total orbital cutoff was 4.9 Å so that the introduced basis group error is small enough to deal with the relativistic effect of metal atoms using the all-electron relativistic theory. Brillouin region integration was optimized by using a special K point method in Bronkhorst-pack form, and the structure was optimized by using K points of 1 × 1 × 7. The criterion of energy convergence in atomic structure optimization was 1.0 × 10⁻⁵ Ha, and the criterion of Hellmann–Feynman atomic force convergence was that the force of each atom was not more than 0.002 Ha Å⁻¹, and the maximum displacement was 0.005 Å.

Under alkaline conditions, the ORR and OER four-electron reaction path is given in eqn S1–S8 in the ESI.† For each basic ORR and OER step, the Gibbs reaction free energy ΔG was defined as the difference between the final state and the initial state free energies, given by the following expression:

ΔG = ΔE + ΔZPE − TΔS + ΔG_U + ΔG_PH (1)

ΔE is the reaction energy of reactants and product molecules adsorbed on the catalyst surface, calculated by DFT. ΔZPE and ΔS are the zero-point energy and entropy changes caused by the reaction. By varying the state ΔG_U = −neU bias effect for each initial, intermediate, and final state, where U is the relative potential, e is the transferred basic charge, and n is the number of transferred proton-electron pairs. ΔG_PH is a correction for H⁺ free energy, depending on whether the reaction being considered is reduction or oxidation:

ΔG_PH = − k_BTln[H⁺] = PH × k_BTln¹⁰ (2)

where k_B is the Boltzmann constant and T is temperature.

The overpotential value is obtained by the difference between the maximum free energy and the equilibrium potential:

η = max [ΔG]/e − 1.23 (3)

where η stands for the overpotential.

η_total = η_ORR + η_OER (4)

η_total stands for the total overpotential, η_ORR stands for the ORR overpotential, and η_OER stands for the OER overpotential.

Results and discussion

Characterization of the material

SEM diagrams before and after the material doping are shown in Fig. 2 and clearly show the surface morphology changes of CNTs before and after the boron doping. As can be observed, the surface of CNTs is relatively smooth, and there is no substance adsorbed on it. Therefore, CNTs are not conducive to the adsorption of oxygen molecules and have fewer catalytic active sites, limiting the ORR and OER on its surface. Fig. 2(b)–(d) show the SEM images of the samples with doping ratios of 2 : 1, 5 : 1, and 8 : 1, respectively. It can be seen that some substances are attached to their surfaces, which become rough. Fig. 2(e) and (f) show the EDS images when the doping ratios is 8 : 1, and the green color represents the doping element B. The adhesion of these active substances on the surface of carbon nanotubes is very uniform, indicating a good doping effect. To better explore the elementary composition and content of the prepared boron-doped carbon nanotube material, X-ray photoelectron spectroscopy (XPS) was used to test BC₃NT within the binding energy range of 0–1350.0 eV to conduct qualitative and semi-quantitative analysis of the material. The full XPS spectrum of BC₃NT is shown in Fig. 3(a). The spectrum of B 1s is shown in Fig. 3(b). Table S2 in the ESI† shows the element content measured by XPS. It can be observed that the three doped materials all show characteristic peaks near 193.4 eV, indicating that the boron element has been successfully doped into the carbon nanotube material. With the increase of the doping ratios of B₂O₃, the peak value becomes larger and larger. The atomic content of B elements is 2.84%, 4.68%, and 5.15%, respectively, and the content of the boron atom doped into the carbon nanotubes increases.

Fig. 2 SEM diagrams: (a) CNT, (b) m_B2O3 : m_CNT = 2 : 1, (c) m_B2O3 : m_CNT = 5 : 1, and (d) m_B2O3 : m_CNT = 8 : 1; ((e) and (f)) EDS images.

Fig. 3 XPS: (a) full spectrum diagram; (b) B 1s characteristic peak.

Fig. 4(a)–(c) represent the B 1s peak fitting diagrams of the samples when B₂O₃ : CNT = 2 : 1, B₂O₃ : CNT = 5 : 1, and B₂O₃ : CNT = 8 : 1. When B₂O₃ : CNT = 2 : 1, five peaks are fitted, located at 188.5 eV, 190.0 eV, 191.3 eV, 193.0 eV, and 193.7 eV, corresponding to B₈C, BC₃, BC₂O, BCO₂, and B₂O₃. When B₂O₃ : CNT = 5 : 1, three peaks are fitted, located at 189.8 eV, 191.3 eV, and 193.7 eV, corresponding to BC₃, BC₂O, and B₂O₃. When B₂O₃ : CNT = 8 : 1, three peaks are fitted, located at 190.0 eV, 191.3 eV, and 193.7 eV, corresponding to BC₃, BC₂O, and B₂O₃, respectively. From the above analysis, it can be seen that there are more impurities generated at 2 : 1, and various carbon oxides are generated, which may be caused by the insufficient quality of B₂O₃ and the inadequate reaction. When the mass of B₂O₃ increases, there exist two kinds of carbon oxides in addition to BC₃ we need at 5 : 1. When the mass of B₂O₃ increases (8 : 1), there is a certain amount of excess B₂O₃. Fig. 4(d)–(f) show the Raman spectra of the samples with different doping ratios to analyze the structural changes of boron-doped carbon nanotubes. The mechanism of defect-rich boron doped carbon nanotubes is that boron doping into CNTs will lead to the graphite's broken hexagonal symmetry, thus inducing an increased D band in the Raman spectrum. As can be seen, the D peak and G peak are obvious at 1360 cm⁻¹ and 1600 cm⁻¹ under different doping ratios. The D peak represents the disordered graphite lattice, and the G peak represents the ordered crystalline graphite carbon. The number of defects of the electrode materials can be evaluated by the numerical ratio (I_D/I_G) between the D peak and G peak. The calculation results show that the I_D/I_G value increases gradually with the increase in content of boron doped into the carbon nanotubes, indicating that the number of surface defects on the carbon nanotubes increases gradually with the increase of boron content.

Fig. 4 B 1s XPS: (a) m_B2O3 : m_CNT = 2 : 1, (b) m_B2O3 : m_CNT = 5 : 1, and (c) m_B2O3 : m_CNT = 8 : 1. Raman spectra: (d) m_B2O3 : m_CNT = 2 : 1, (e) m_B2O3 : m_CNT = 5 : 1, and (f) m_B2O3 : m_CNT = 8 : 1.

Electrochemical performance of the HLABs

The roundtrip performance was measured with a limited capacity of 50 mAh g⁻¹, as shown in Fig. 5. 52 cycles were performed for CNTs as a positive electrode, and 75 cycles were performed for the B-doped CNT at B₂O₃ : CNT = 2 : 1. As the mass ratio of B₂O₃ and CNT is 5 : 1, the battery with BC₃NT cycles 140 times. As the mass ratio of B₂O₃ and CNT is 8 : 1, the battery with BC₃NT cycles 165 times. It can be seen that the number of cycles increases after using the B-doped CNT as a positive material, and the larger the doping ratio, the greater the number of cycles. The number of cycles reached the maximum when the doping ratio was 8 : 1, indicating that the higher the doping ratio of B, the higher the catalytic activity of the material. When pure CNT is used as the positive electrode, the initial overpotential can only maintain 20 cycles. Then, the battery appears to decay due to the poor reversibility. When the doping ratio is 2 : 1, an overpotential of 0.5 V can be maintained for 100 hours; when the doping ratio is 5 : 1, an overpotential of 1 V can be maintained for 260 hours; when the doping ratio is 8 : 1, an ultra-small overpotential of 0.3 V can be maintained for 200 hours. The battery shows excellent reversibility. Therefore, when the doping ratio is 8 : 1, the cycle times of the battery can be improved, reversibility can be improved, the overpotential can be reduced, and the battery runs more smoothly.

Fig. 5 Cycle performance: (a) CNT, (b) B₂O₃ : CNT = 2 : 1, (c) B₂O₃ : CNT = 5 : 1, and (d) B₂O₃ : CNT = 8 : 1.

In order to further explore the cause of battery failure, the battery was disassembled and it is found that white substances cover the negative lithium electrode. The white substances pierced LISICON film, resulting in film breakage, and the battery stopped discharging, as shown in Fig. S5 in the ESI.† To further explore the reasons for the breakage of the LISICON film, the products on the surface of the negative lithium film were characterized. Fig. S6 in the ESI† shows the SEM results. As can be seen, a large number of lithium dendrites are generated on its surface, and a large number of lithium dendrites grow through the LISICON film, making the battery fail. The surface of some lithium dendrites is covered with a layer of other substances, and XPS is carried out to study the elemental composition. The full XPS diagram is shown in Fig. S7(a) in the ESI.† The material covering the surface of lithium dendrite contains a large number of O elements and a small amount of Li and C elements. Peak fitting was performed for the O 1s and Li 1s spectra, and the results were obtained as shown in Fig. S7(b) and (c).† It contains LiOH, Li₂CO₃, and Li₂O₂. In conclusion, lithium dendrites grew from the negative electrode and pierced the LAGP film, resulting in the leakage of electrolyte LiOH into the positive electrode chamber, and a small amount of Li₂CO₃ was generated by a reaction with CO₂.

Fig. 6 shows the results of initial discharge capacity with the CNT; the mass proportions of B₂O₃ and CNT are 2 : 1, 5 : 1, and 8 : 1 as positive electrodes, respectively, at a current density of 0.05 mA cm⁻² and an electrolyte concentration of 0.1 mol L⁻¹. The maximum specific capacities are 5000 mAh g⁻¹, 5650 mAh g⁻¹, 7200 mAh g⁻¹ and 8900 mAh g⁻¹. BC₃NT can increase the specific capacity, and with the increase of the doping ratio, the specific capacity also increases because the decomposition rate of LiOH precipitation increases after doping B into the CNT. From the point of view of the discharge voltage platform, when the CNT is used as the positive electrode, the discharge voltage platform is relatively low, which is 2.4 V; when the B-doped CNT is used as the positive electrode, the discharge voltage platform is about 3 V. The reason for the increase in initial discharge capacity is that with the increase of the doping ratio, the material has more and more defects. Compared with common CNTs, the defect position can be used as the active site of the reaction to significantly improve the electrochemical reaction efficiency, which positively affects the battery discharge. The fundamental parameters we obtained through electrochemical measurements, such as the number of cycles, initial discharge capacity, and overpotentials, have significant advantages compared with previous results (Table S3 in the ESI†).

Fig. 6 Diagram of the discharge capacity: (a) CNT, (b) B₂O₃ : CNT = 2 : 1, (c) B₂O₃ : CNT = 5 : 1, and (d) B₂O₃ : CNT = 8 : 1.

DFT calculation

For HLABs, the key to the battery reaction under alkaline conditions is the adsorption of O₂ molecules. At the same time, a perfect CNT is not conducive to O₂ adsorption, so this paper does not calculate its free energy. Fig. 7 shows the calculated state density of O₂ molecules at the adsorption site and the 2p orbital of the B atom and the established four-electron reaction path through *O₂, *OOH, *O, and *OH. The O₂ molecule is adsorbed on the B atom. For perfect BC₃NT and 7MR, the O₂ molecule is completely adsorbed on the B atom, while for 8MR, the O₂ molecule is desorbed, which is more conducive to subsequent reactions. It can be seen from Fig. 7(d) that in the perfect BC₃NT, fewer parts correspond to the density peaks of the B atom and O₂ molecular state, and the coincidence area is small, indicating that it produces relatively weak hybridization. As shown in Fig. 7(e) and (f), the degree of hybridization becomes stronger and stronger with the increase of the defect ring. The B atom in 8MR corresponds well to the state density peak of the adsorbed O₂ molecule, and the coincidence area is large, indicating that it has the strongest hybridization.

Fig. 7 Reaction path after optimization: (a) BC₃NT, (b) 7MR, and (c) 8MR. B atom and O₂ molecular state density: (d) BC₃NT, (e) 7MR, and (f) 8MR.

Fig. 8(a)–(f) show the free energy of the ORR and OER processes. Fig. 8(g) and (h) show, respectively, the calculated ORR overpotential and the OER overpotential. An important parameter for evaluating catalytic activity is the limiting reaction barrier, which is determined by the free energy of the rate-determining step (RDS). For the ORR process, when BC₃NT acts as a catalyst, the RDS of the ORR is the reduction of *OOH to *O, and its overpotential is 0.33 V. When 7MR is used as a catalyst, the RDS of the ORR is the reduction of *O to *OH, and its overpotential is 0.15 V. When 8MR is used as a catalyst, the O₂ molecule is dissociated and adsorbed, and its RDS is the reduction of *O₂ to *OOH. This two-site mechanism has an overpotential of 0.17 V. *OH desorption is relatively easy for the three cathode materials, which is convenient for the adsorption of the next O₂ molecule, thereby continuing the reaction. When the BC₃NT catalyst is used for the OER process, the RDS of the OER is the oxidation of *OOH to *O₂. The overpotential value of the battery is 0.6 V. When 7MR is used as a catalyst, its RDS is the oxidation of *O to *OOH, and the overpotential is 0.2 V. When 8MR is used as a catalyst, the RDS of the OER is the oxidation of OH⁻ to *OOH, and the overpotential of the OER is 0.17 V.

Fig. 8 Free energy of the ORR: (a) BC₃NT, (b) 7MR, and (c) 8MR. OER free energy: (d) BC₃NT, (e) 7MR, and (f) 8MR. (g) Overpotential of the ORR. (h) Overpotential of the OER.

When BC₃NT is used as the positive electrode, the free energy value of the intermediate is the most negative, indicating that the chemisorption is relatively strong and the activation energy barrier of the following reaction is relatively high. When BC₃NT, 7MR, and 8MR are used as the positive electrodes, the total overpotential is 0.93 V, 0.35 V, and 0.34 V, respectively. The ORR overpotential and OER overpotential are both increasing. The generation of defects will reduce the overpotential of the battery. The larger the topological defect ring, the smaller the overpotential. The calculated minimum overpotential of 0.34 V is close to the experimental value of 0.30 V, indicating that the calculated results are reliable.

Conclusions

In summary, CNTs doped with different proportions of B were prepared by a high-temperature doping method and were used as a cathode for hybrid Li–air batteries, taking advantage of its catalytic performance. The results of roundtrip performance show that the higher the boron doping ratio is, the more cycles the battery cycle is, and the battery can keep the ultra-low overpotential operation for 200 hours in air. The discharge capacity shows that the larger the B doping ratio is, the larger the battery initial discharge capacity is. In addition, DFT calculation results show that the larger the topological defect ring, the stronger the ability of adsorption towards O₂ molecules and the B atom hybridization. The overpotential of BC₃NT with 7-membered ring and 8-membered ring topological defects is lower than that of pure BC₃NT. The larger the topological defect ring, the smaller the overpotential.

Author contributions

The manuscript was written through the contributions of all authors. All authors have given approval to the final version of the manuscript. Yuyang Wang, Mingfu Yu and Hong Sun contributed equally to the publication. Zhichao Xue and Ying Ma contributed to the writing of data analysis and interpretation. With the help of Ying Ma, Yuyang Wang synthesised and characterised the materials and performed electrocatalytic tests. With the help of Tianyu Zhang, Yuyang Wang performed the DFT calculations. The research was carried out in close collaboration between the research groups.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The authors acknowledge financial support from the National Natural Science Foundation of China (51776131), the Liaoning Revitalization Talents Program (XLYC1802045), and Liaoning province Department of Education fund item (lnqn202007).

Notes and references

1. J. M. Cameron, C. Holc, A. J. Kibler, C. L. Peake, D. A. Walsh, G. N. Newton and L. R. Johnson, Chem. Soc. Rev., 2021, 50, 5863–5883.
2. J.-H. Kang, J. Lee, J.-W. Jung, J. Park, T. Jang, H.-S. Kim, J.-S. Nam, H. Lim, K. R. Yoon, W.-H. Ryu, I.-D. Kim and H. R. Byon, ACS Nano, 2020, 14, 14549–14578.
3. T. Liu, X.-Y. Yang and X.-B. Zhang, Adv. Mater. Technol., 2020, 5, 2000476.
4. L. Ye, Y. Hong, M. Liao, B. Wang, D. Wei, H. Peng, L. Ye, Y. Hong, M. Liao, B. Wang, D. Wei and H. Peng, Energy Storage Mater., 2020, 28, 364–374.
5. J. Lai, Y. Xing, N. Chen, L. Li, F. Wu and R. Chen, Angew. Chem., Int. Ed., 2020, 59, 2974–2997.
6. Y. Zhou, C.-H. Wang, W. Lu and L. Dai, Adv. Mater., 2020, 32, 1902779.
7. E. L. Littauer and K. C. Tsai, J. Electrochem. Soc., 1976, 123, 771–776.
8. Y. Xing, N. Chen, M. Luo, Y. Sun, Y. Yang, J. Qian, L. Li, S. Guo, R. Chen and F. Wu, Energy Storage Mater., 2020, 24, 707–713.
9. M. Asadi, B. Sayahpour, P. Abbasi, A. T. Ngo, K. Karis, J. R. Jokisaari, C. Liu, B. Narayanan, M. Gerard, P. Yasaei, X. Hu, A. Mukherjee, K. C. Lau, R. S. Assary, F. Khalili-Araghi, R. F. Klie, L. A. Curtiss and A. Salehi-Khojin, Nature, 2018, 555, 502–506.
10. C.-S. Yang, K.-N. Gao, X.-P. Zhang, Z. Sun and T. Zhang, Rare Met., 2018, 37, 459–472.
11. Q. Liu, Z. Chang, Z. Li and X. Zhang, Small Methods, 2018, 2, 1700231.
12. Y. Jiang, J. Cheng, L. Zou, X. Li, Y. Huang, L. Jia, B. Chi, J. Pu and J. Li, ChemCatChem, 2017, 9, 4231–4237.
13. Q. Lv, Z. Zhu, S. Zhao, L. Wang, Q. Zhao, F. Li, L. A. Archer and J. Chen, J. Am. Chem. Soc., 2021, 143, 1941–1947.
14. K. Chen, D.-Y. Yang, G. Huang and X.-B. Zhang, Acc. Chem. Res., 2021, 54, 632–641.
15. L. Li, Z. Chang and X.-B. Zhang, Adv. Sustainable Syst., 2017, 1, 1700036.
16. P. Tan, B. Chen, H. Xu, H. Zhang, W. Cai, M. Ni, M. Liu and Z. Shao, Energy Environ. Sci., 2017, 10, 2056–2080.
17. D. Aurbach, B. D. McCloskey, L. F. Nazar and P. G. Bruce, Nat. Energy, 2016, 1, 16128.
18. J. B. Goodenough and Y. Kim, J. Power Sources, 2011, 196, 6688–6694.
19. X. Zhu, Y. Yan, W. Wan, Y. Wang, Y. Wu, X. He and Z. Lu, Mater. Lett., 2018, 215, 71–74.
20. P. He, Y. Wang and H. Zhou, Electrochem. Commun., 2010, 12, 1686–1689.
21. M. M. Thackeray, M. K. Y. Chan, L. Trahey, S. Kirklin and C. Wolverton, J. Phys. Chem. Lett., 2013, 4, 3607–3611.
22. N. Ahmad, L. Zhou, M. Faheem, M. K. Tufail, L. Yang, R. Chen, Y. Zhou and W. Yang, ACS Appl. Mater. Interfaces, 2020, 12, 21548–21558.
23. F. Bai, K. Kakimoto, X. Shang, D. Mori, S. Taminato, M. Matsumoto, Y. Takeda, O. Yamamoto, H. Izumi, H. Minami and N. Imanishi, Front. Energy Res., 2020, 8, 187.
24. X. Liu, C. Culhane, W. Li and S. Zou, ACS Omega, 2020, 5, 24367–24378.
25. K.-X. Wang, Q.-C. Zhu and J.-S. Chen, Small, 2018, 14, 1800078.
26. Z. Chang, J. Xu and X. Zhang, Adv. Energy Mater., 2017, 7, 1700875.
27. A. Nomura, K. Ito and Y. Kubo, Sci. Rep., 2017, 7, 45596.
28. X. Yi, X. Liu, R. Dou, Z. Wen and W. Zhou, ACS Appl. Energy Mater., 2021, 4, 2148–2157.
29. Z. Jian, P. Liu, F. Li, P. He, X. Guo, M. Chen and H. Zhou, Angew. Chem., Int. Ed., 2014, 53, 442–446.
30. A. C. Lim, H. J. Kwon, H. C. Lee, D. J. Lee, H. Lee, H. J. Kim, D. Im and J. G. Seo, Chem. Eng. J., 2020, 385, 123841.
31. X. Hu, J. Wang, T. Li, J. Wang, D. H. Gregory and J. Chen, Nano Lett., 2017, 17, 2073–2078.
32. A. O. Boev, S. S. Fedotov, K. J. Stevenson and D. A. Aksyonov, Comput. Mater. Sci., 2021, 197, 110592.
33. X. Cao, Z. Chen, N. Wang, Z. Han, X. Zheng and R. Yang, Electrochim. Acta, 2021, 383, 138369.
34. L. Zou, S. Guo, Z. Wang, M. Sun, F. Yu, B. Chi and J. Pu, Electrochim. Acta, 2021, 381, 138219.
35. H. Xia, Q. Xie, Y. Tian, Q. Chen, M. Wen, J. Zhang, Y. Wang, Y. Tang and S. Zhang, Nano Energy, 2021, 84, 105877.
36. High-Index Faceted Porous Co3O4 Nanosheets with Oxygen Vacancies for Highly Efficient Water Oxidation | ACS Applied Materials & Interfaces, https://pubs.acs.org/doi/abs/10.1021/acsami.7b18208, (accessed 4 September 2021).
37. I. Mukhopadhyay, N. Hoshino, S. Kawasaki, F. Okino, W. K. Hsu and H. Touhara, J. Electrochem. Soc., 2002, 149, A39.
38. N. E. Benti, G. A. Tiruye and Y. S. Mekonnen, RSC Adv., 2020, 10, 21387–21398.
39. Z. Feng, Y. Ma, Y. Li, R. Li, Y. Tang and X. Dai, Phys. Chem. Chem. Phys., 2020, 22, 1493–1501.
40. F. Zheng, H. Dong, Y. Ji and Y. Li, J. Power Sources, 2019, 436, 226845.
41. B. He, J. Shen, Z. Lu and D. Ma, Appl. Surf. Sci., 2020, 527, 146828.
42. W. He, P. Xue, H. Du, L. Xu, M. Pang, X. Gao, J. Yu, Z. Zhang and T. Huang, Int. J. Hydrogen Energy, 2017, 42, 4123–4132.
43. H. R. Jiang, T. S. Zhao, L. Shi, P. Tan and L. An, J. Phys. Chem. C, 2016, 120, 6612–6618.
44. L. Gao, M. Zhu, Z. Zhang and G. Cui, Electrochim. Acta, 2019, 327, 134980.
45. F.-Q. Li, Y. Zhang and S.-L. Zhang, Nanomaterials, 2021, 11(6), 1395.
46. B. C. Lourencao, R. A. Pinheiro, T. A. Silva, E. J. Corat and O. Fatibello-Filho, Diamond Relat. Mater., 2017, 74, 182–190.
47. C. Shu, R. Huang, J. Wang and D. Su, RSC Adv., 2015, 5, 103019–103022.
48. F. Wu, Y. Xing, L. Li, J. Qian, W. Qu, J. Wen, D. Miller, Y. Ye, R. Chen, K. Amine and J. Lu, ACS Appl. Mater. Interfaces, 2016, 8, 23635–23645.
49. K.-X. Zhang, H. Su, H.-H. Wang, J.-J. Zhang, S.-Y. Zhao, W. Lei, X. Wei, X.-H. Li and J.-S. Chen, Adv. Sci., 2018, 5, 1800062.
50. L. Yang, S. Jiang, Y. Zhao, L. Zhu and Z. Hu, Angew. Chem., Int. Ed., 2011, 50, 7132–7135.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/d1cy01832a

---