Regulation of nitrogen reduction reaction catalytic performance by varying the sp/sp² hybrid carbon ratio in graphyne/graphene heterojunction catalysts

Abstract

This study investigates the impact of the sp/sp² hybrid carbon ratio on the nitrogen reduction reaction (NRR) catalytic performance of γ-graphyne and graphene-based heterojunction catalysts. Through density of states (DOS) calculations, crystal orbital Hamilton population (COHP) analysis, and charge density difference plots, it is found that the sp/sp² hybrid carbon ratio significantly influences the electronic properties and NRR activity of the catalysts. The Ti@GY1/Gr catalyst exhibits superior performance, attributed to its sensitivity to changes in the sp/sp² hybrid environment, especially when combined with graphene. An increase in the sp/sp² hybrid carbon ratio leads to a decrease in NRR activity, while also modulating the interaction between Ti and the carbon support. The findings highlight the importance of the sp/sp² hybrid carbon ratio in regulating the electronic properties and catalytic performance of heterojunction catalysts, providing insights for the design of more efficient NRR catalysts.

---

Haozhi Wang

Dr Haozhi Wang received his PhD from the School of Chemical Engineering at Dalian University of Technology in 2020. After a postdoctoral stint at the School of Materials Science and Engineering at Tianjin University, he joined the School of Materials Science and Engineering at Hainan University in 2022. His research focuses on the theoretical design of energy-related catalytic materials and on the development of novel materials through high-throughput screening and machine learning approaches.

---

Introduction

The Haber–Bosch process, the cornerstone of industrial ammonia production, poses significant sustainability challenges due to its intensive energy requirements and substantial CO₂ emissions from fossil fuel dependence.^1–3 In contrast, the electrocatalytic nitrogen reduction reaction (NRR) offers a promising alternative by utilizing renewable electricity and operating under ambient conditions, aligning with global goals to conserve energy and minimize environmental impact. Consequently, the development of high-performance NRR electrocatalysts has become a critical focus in advancing sustainable ammonia production.^4,5

Single-atom catalysts (SACs) are at the forefront of NRR research due to their isolated metal centers, which enhance metal utilization and thereby improve catalytic efficiency.^6,7 Similar advancements have been achieved in various electrochemical reactions using SACs, including the hydrogen evolution reaction, oxygen reduction and evolution reactions, and carbon dioxide reduction reaction.^8–19 Carbon materials are among the most ideal supports for SACs due to their high surface area, pore structure, and excellent electrical conductivity.²⁰ The electronic properties of SACs are strongly influenced by the carbon supports, where the interaction between the metal atoms and the carbon support modulates the charge density of the active site and affects the catalytic activity of SACs.^8,21–25 Graphyne (GY), with its sp and sp² hybrid carbon structures, offers unique characteristics, making it a promising material for SACs in the NRR.^26–30 In our previous work, we found that increasing the ratio of sp/sp² hybrid carbon significantly modulates the electronic structure of Ti@GY, thereby improving the catalytic activity of Ti@GY.³¹ The impact of the sp/sp² ratio is primarily seen in the structural characteristics of GY itself. However, recent studies have revealed the presence of exogenous sp² carbon components in the synthesized GY materials.³² As a result, it is important to further explore how these additional exogenous sp² components influence the electronic structure of GY.

In this work, we investigated the catalytic performance of graphyne/graphene heterojunction (GY/Gr)-supported SACs for NRR applications. Based on our previous work, we found that Ti@GY showed the best NRR activity. Therefore, we designed three Ti-based GY/Gr models (Ti@GY1/Gr, Ti@GY2/Gr and Ti@GY3/Gr) with different sp/sp² hybrid carbon ratios and analyzed their electronic structures and potential-determining step free energy change of the NRR. Our results show that the introduction of exogenous sp² carbon alters the charge aggregation region on the surface of GY. As the sp/sp² hybrid carbon ratio increases, the ability of exogenous sp² carbon to transfer electrons to GY decreases, thereby changing the mechanism by which the sp/sp² hybrid carbon ratio regulates GY. Further analysis revealed that the increased sp/sp² hybrid carbon ratio, along with the introduction of exogenous sp² carbon, weakened the interaction between GY and Ti, reducing the electron transfer from Ti to GY and consequently suppressing the catalyst's catalytic activity. Among the catalysts studied, Ti@GY1/Gr exhibited the best performance for the NRR. The detailed calculation parameters are provided in the ESI.†

Results and discussion

We constructed two-dimensional heterojunctions composed of γ-GY and graphene (Gr) and selected three models with varying sp/sp² hybrid carbon ratios (GY1/Gr, GY2/Gr, and GY3/Gr) in this study (Fig. 1a–c). To evaluate the effects of different sp/sp² hybrid carbon ratios on the electronic behavior of the heterojunctions, we calculated the Bader charge. As shown in Fig. 1d–f, the maps show electron enrichment within the benzene ring of GY at the SAC, which contrasts with the observations for isolated GY, suggesting a change in electronic behavior when GY is part of a heterojunction. Bader charge analysis showed that the increase in the sp/sp² hybrid carbon ratio, following the introduction of exogenous sp² carbon, leads to a decrease in electron accumulation on the benzene ring of GY, thereby reducing the catalytic activity of the GY surface. Additionally, by averaging the Bader charge on GY/Gr to each atom of GY and Gr, the data obtained are shown in Fig. S1.† Among the three GY/Gr models, the Bader charge value of GY1 (0.012e) is more positive compared to that of GY2 (0.006e) and GY3 (0.001e), indicating that GY1 accepts more electrons contributed from Gr. This demonstrates that a relatively low sp/sp² hybrid carbon ratio is conducive to the activation of GY when accepting exogenous sp² hybridization. Further analysis of the electron distribution on the heterojunction surface was conducted using the electron localization function (ELF) (Fig. 1g–i). ELF calculations demonstrate that the electron local density shifts with different sp/sp² hybrid carbon ratios, with electrons concentrating around the benzene ring and becoming more widely distributed near the acetylene bond. This suggests that increasing the sp/sp² hybrid carbon ratio promotes electron transfer and the formation of active sites, akin to the behavior observed in isolated GY. Finally, we examined the interaction between the GY and Gr layers in the heterojunctions by analyzing the charge density difference maps (Fig. 1j–l). The charge density difference maps reveal significant interlayer electron transfer within the heterojunctions, with yellow regions indicating charge density accumulation and blue regions indicating charge depletion. In GY1/Gr and GY2/Gr, electrons remain localized on the surface of GY. In contrast, in GY3/Gr, electrons are no longer localized on the surface of GY, with part of them transferred to the Gr side. This suggests that a higher sp/sp² hybrid carbon ratio alters the charge density distribution between the layers, despite the same layer spacing of 3.5 Å. In conclusion, by constructing a heterojunction with the introduction of exogenous sp² carbon, the surface electrons of GY aggregate near the benzene ring. Increasing the sp/sp² hybrid carbon ratio under these conditions leads to a loss of charge density localized in GY, altering its electron distribution.

Fig. 1 (a–c) Heterojunction models constructed for GY1, GY2, GY3 and graphene, (d–f) analysis of Bader charges projected on GY, where positive (negative) charges indicate the loss (gain) of electrons by C atoms, (g–i) electron local density functions corresponding to the three heterojunction models, and (j–l) differential charge density maps of GY1, GY2, GY3 and graphene.

Based on the heterojunction structure, the stable adsorption sites of Ti were first investigated (Fig. S2†). Theoretical calculations revealed that the adsorption energies of Ti on GY1 (−0.43 eV), GY2 (1.19 eV), and GY3 (1.44 eV) were considerably lower than those on the corresponding Gr side (3.44 eV, 3.83 eV, and 3.85 eV), indicating that the interaction between Ti and sp-hybrid carbon is stronger, with Ti preferentially adsorbing on the GY side. Moreover, as the sp/sp² carbon ratio increases, the adsorption strength of Ti weakens, suggesting that less sp/sp² carbon is needed to regulate the adsorption of Ti. To explore the influence of the sp/sp² hybrid carbon ratio on the intrinsic properties of Ti@GY catalysts after the introduction of exogenous sp² carbon, Bader charge analysis and charge density difference mapping were conducted for the three Ti@GY/Gr systems (Fig. 2a–c). The Bader charge results demonstrated a significant charge reduction on Ti for GY1/Gr (−1.39e), GY2/Gr (−1.22e), and GY3/Gr (−1.31e), which is attributed to the introduction of exogenous sp² carbon that led to electron accumulation near the alkyne bonds. In Ti@GY1/Gr, the three alkyne bonds adjacent to Ti were activated due to the electron distribution, which is consistent with the lower adsorption energy of Ti (−0.43 eV). As the sp/sp² hybrid carbon ratio increased, the alkyne bonds near Ti received fewer electrons, leading to reduced activation and higher adsorption energies for Ti in Ti@GY2/Gr (1.19 eV) and Ti@GY3/Gr (1.44 eV), which is consistent with the observed changes in adsorption energy. Charge density difference maps further revealed that Ti@GY1/Gr exhibited a smaller pore size, enabling Ti to form a symmetric charge density distribution with the surrounding alkyne bonds (Fig. 2d–f). In contrast, Ti@GY2/Gr and Ti@GY3/Gr featured larger pore sizes, with the charge density difference concentrated on the two alkyne bonds closer to the benzene ring. These observations suggested that Ti@GY1/Gr has smaller pores, enabling Ti to interact with more alkyne bonds and thus enhancing the adsorption energy. The charge redistribution on the Ti@GY/Gr surface affected its work function, which represented the energy required for electrons to transition from the Fermi level to the vacuum level (Fig. 2g, h and S3a–d†). The calculated work functions for GY1/Gr (5.02 eV), GY2/Gr (5.89 eV), and GY3/Gr (5.10 eV) decreased upon Ti loading, yielding values of 4.98 eV, 4.76 eV, and 4.97 eV for Ti@GY1/Gr, Ti@GY2/Gr, and Ti@GY3/Gr, respectively. This reduction in work function indicated a decrease in surface energy after metal loading, implying weaker electron binding. Additionally, work function calculations revealed variations in interlayer potential energy differences within the heterojunctions. GY1/Gr (8.38 eV) and Ti@GY1/Gr (4.99 eV) exhibited the smallest interlayer potential energy differences, while GY2/Gr (11.00 eV), Ti@GY2/Gr (10.48 eV), GY3/Gr (11.88 eV), and Ti@GY3/Gr (11.87 eV) showed larger differences. Comparing work functions before and after Ti adsorption, it was observed that the potential energy difference (ΔE) between the Gr and GY layers decreased post-Ti adsorption, with the trend slowing as the sp/sp² hybrid carbon ratio increased. For Ti@GY3/Gr, the change in ΔE was minimal (0.01 eV), suggesting that the sp/sp² hybrid carbon ratio effectively modulated the interlayer potential energy difference, thereby fine-tuning the overall electronic structure of the heterojunction. Notably, the reduction in potential energy difference was primarily attributed to an increase in potential energy on the Gr side following Ti loading on the GY side rather than a decrease on the GY side. It indicated that Ti loading enhanced electrostatic interactions between the GY and Gr layers, elevating the potential energy on the Gr side.

Fig. 2 (a–c) Bader charge distribution, (d–f) differential charge density maps, (g and h) work function, (i) density of states and (j) crystal orbital Hamilton populations for Ti@GY1/Gr, Ti@GY2/Gr and Ti@GY3/Gr.

To further analyze the effect of the sp/sp² hybrid carbon ratio on the electronic structure of the Ti active site, the density of states (DOS) was investigated (Fig. 2i). The calculated d-band centers for Ti@GY1/Gr, Ti@GY2/Gr, and Ti@GY3/Gr were 1.50 eV, 0.83 eV, and 0.65 eV, respectively, indicating that as the sp/sp² hybrid carbon ratio increased, the d-band centers of Ti shifted closer to the Fermi level. Aligned with our previous findings on Ti@GY,³¹ demonstrating that the introduction of exogenous sp² carbon did not disrupt the regulatory effect of the hybrid carbon ratio on the d-band center. Notably, Ti@GY3/Gr exhibited d-band centers closest to the Fermi level, with a lower DOS compared to Ti@GY1/Gr and Ti@GY2/Gr. However, the DOS of Ti in Ti@GY2/Gr near the Fermi level was more continuous and delocalized. Furthermore, crystal orbital Hamilton population (COHP) analysis was performed to evaluate the strength of chemical bonds between atoms (Fig. 2j). The COHP results revealed that Ti@GY1/Gr had fewer antibonding orbitals below the Fermi level, while Ti@GY3/Gr exhibited the most antibonding orbitals, suggesting that Ti bonding on GY1/Gr was more stable and that increasing the sp/sp² hybrid carbon ratio weakened the bond strength between Ti and carbon atoms. These findings were consistent with the single-atom binding energies of Ti@GY1/Gr (−0.43 eV), Ti@GY2/Gr (1.19 eV), and Ti@GY3/Gr (1.44 eV), further corroborating the COHP results.

The adsorption of N₂ is a critical initial step in the nitrogen reduction reaction (NRR), as it directly governs the subsequent hydrogenation process. The bonding of N₂ with the catalyst surface was investigated through charge density difference maps, as illustrated in Fig. 3a–c. In Ti@GY1/Gr (Fig. 3a), the map reveals a significant accumulation of electron density around the Ti atom, indicated by the blue regions, suggesting a strong interaction between the Ti atom and the N₂ molecule, which likely enhances the adsorption strength of N₂. In Ti@GY2/Gr (Fig. 3b), the electron density around the Ti atom is less concentrated, with more dispersed blue regions, indicating a weaker interaction between the Ti atom and the N₂ molecule and potentially lower adsorption strength. In Ti@GY3/Gr (Fig. 3c), the differential charge density map shows the most dispersed electron density around the Ti atom, with minimal blue regions, suggesting the weakest interaction between the Ti atom and the N₂ molecule among the three models and likely the lowest adsorption strength. The differential charge density maps clearly demonstrate that a higher electron density around the Ti atom corresponds to a stronger interaction with N₂, leading to enhanced adsorption strength, while a lower electron density around the Ti atom results in a weaker interaction and reduced adsorption strength. It is worth noting that the lower the sp/sp² hybrid carbon ratio, the higher the electron density around Ti, which further confirms the significant impact of the electronic structure on the adsorption behavior. Indeed, Ti@GY3/Gr exhibited a lower N₂ adsorption energy (−0.59 eV) compared to Ti@GY1/Gr (0.13 eV) and Ti@GY2/Gr (−0.17 eV), suggesting that N₂ adsorption was more thermodynamically favorable on Ti@GY3/Gr. Further analysis of the Bader charge revealed that the electron density on the surface of Ti@GY/Gr decreased as the sp/sp² hybrid carbon ratio increased, particularly near the sp/sp² hybrid carbon regions (Fig. 3d–f). This finding implied that, although the bridging between sp/sp² hybrid carbon and Ti in Ti@GY/Gr facilitated electron transfer to N₂, the electron transfer process was progressively suppressed with increasing hybrid carbon ratios, thereby impairing N₂ activation. To further elucidate the influence of the sp/sp² hybrid carbon ratio on the electronic interactions between Ti and N₂, the density of states (DOS) for the N₂ adsorption models was calculated (Fig. 3g and h). Notably, Ti@GY1/Gr exhibited pronounced orbital splitting in its DOS following N₂ adsorption, with a similar phenomenon observed in the p-orbital DOS of N₂, confirming the effective activation of N₂ by Ti@GY1/Gr. In contrast, Ti@GY2/Gr and Ti@GY3/Gr lacked such orbital splitting, indicating that N₂ adsorption and activation were less efficient. Finally, the stability of the Ti–N₂ bonding was evaluated using crystal orbital Hamilton population (COHP) calculations (Fig. 3i). The results revealed a negative correlation between the sp/sp² hybrid carbon ratio and the ICOHP value, suggesting that increasing the sp/sp² hybrid carbon ratio strengthened the bonding between N₂ and Ti. However, this enhanced bonding may not necessarily benefit the catalytic activity of the NRR, as stronger N₂ binding could hinder the activation and reduction processes essential for efficient catalysis.

Fig. 3 (a–c) Differential charge density maps for the N₂ adsorption models, (d–f) Bader charge distribution, (g) d-orbital density of states for Ti, (h) p-orbital density of states for N, and (i) COHP for Ti–N.

Finally, to evaluate the NRR performance of the three Ti@GY/Gr catalysts, four reaction pathways were systematically calculated (Fig. 4a and S4†), and the structures of all reaction intermediates supported on the material were determined (Fig. S5–S10†). Among the three catalysts, Ti@GY2/Gr (0.88 eV) and Ti@GY3/Gr (0.88 eV) exhibited the highest free energy change for the potential-determining step (PDS), while Ti@GY1/Gr (0.32 eV) showed the lowest PDS energy barrier. These results clearly demonstrated that the strength of N₂ adsorption did not directly correlate with NRR catalytic activity. Instead, Ti@GY1/Gr achieved an optimal N₂ adsorption strength, which was attributed to its lower sp/sp² hybrid carbon ratio. This resulted in a reduced interlayer potential energy difference (4.99 eV), thereby enhancing the NRR catalytic performance. In parallel, the hydrogen evolution reaction (HER) competes with the NRR. To assess the selectivity of the three catalysts for the NRR over the HER, the NRR selectivity was calculated. The results revealed that all three heterojunction catalysts exhibited high NRR selectivity compared with the HER (Fig. 4b), indicating that they were less prone to the HER and more likely to favor the NRR pathway. Additionally, to confirm the thermodynamic stability of the catalysts, ab initio molecular dynamics (AIMD) simulations were performed on Ti@GY1/Gr, Ti@GY2/Gr and Ti@GY3/Gr (Fig. S11†). The results showed that the free energy and temperature of Ti@GY1/Gr, Ti@GY2/Gr and Ti@GY3/Gr at 300 K remained within a narrow range throughout the simulation, demonstrating the structural stability of the system over time. The fourth hydrogenation step was identified as the potential-determining step for both Ti@GY1/Gr and Ti@GY2/Gr, whereas for Ti@GY3/Gr, the decisive step shifted to the first hydrogenation step. By correlating the free energy change of the fourth hydrogenation step with the changes in N–N bond lengths for each catalyst, a negative correlation was observed (Fig. 4c). Specifically, as the N–N bond length decreased (or increased), the Gibbs free energy barrier for hydrogenation increased (or decreased). This suggested that catalysts with more activated intermediates and shorter N–N bond lengths exhibited lower free energy change for the decisive hydrogenation step, thereby promoting a more efficient NRR process. To elucidate the relationship between the electronic/geometric features of SACs and their NRR performance, a comprehensive correlation analysis was performed (Fig. 4d and e). For SACs, Fig. 4d reveals several significant correlations. The RDS exhibits a strong negative correlation with the d-band center of Ti (BCT, −0.88), indicating that a higher BCT leads to a lower RDS energy barrier and enhanced catalytic activity. The BCT also shows a strong positive correlation with the single-atom binding energy of Ti (E_b, 0.88), suggesting that a higher BCT corresponds to stronger adsorption of Ti atoms on the catalyst surface. The average bond length of Ti and C (Ti–C) is positively correlated with the ICOHP values for Ti and C (I_c, 0.97), indicating that longer Ti–C bonds enhance the electronic interaction between Ti and C atoms. Additionally, the ICOHP values (I_c) are negatively correlated with the work function of SACs (W_f, −0.89), suggesting that stronger electronic interactions between Ti and C atoms result in a lower work function and potentially improved catalytic activity. Other notable correlations include a negative correlation between BCT and W_f (−0.81), a negative correlation between Ti–C and E_b (−0.81), a negative correlation between Bader charge of Ti (BT) and W_f (−0.75), and a positive correlation between BT and E_b (0.74). For the N₂ adsorption models, Fig. 4e presents a series of significant correlations. The RDS energy barrier is perfectly positively correlated with the adsorption energy of N₂ (E_ads, 1.00), indicating that higher N₂ adsorption energy directly increases the RDS energy barrier, potentially reducing catalytic activity. The E_ads is strongly negatively correlated with the N–N bond length (−0.99), suggesting that stronger N₂ adsorption shortens the N–N bond length, thereby activating N₂. The N–N bond length is positively correlated with the Ti–N bond length (0.95), indicating that as the N–N bond shortens, the Ti–N bond lengthens. The Ti–N bond length is positively correlated with the center of the p-band of the N₂ molecule (BCN, 0.95), suggesting that changes in the Ti–N bond length affect the electronic structure of N₂. Additionally, the E_ads is negatively correlated with the BCN (−0.92), and the N–N bond length is negatively correlated with the Bader charge of the N₂ molecule (BN, −0.98), further highlighting the impact of adsorption energy on N₂ activation. The Ti–N bond length is also negatively correlated with BN (−0.95), indicating that as the Ti–N bond lengthens, the Bader charge of N₂ increases.

Fig. 4 Ti@GY1/Gr, Ti@GY2/Gr, and Ti@GY3/Gr for the NRR: (a) Gibbs free energy plots of the enzymatic pathway, (b) HER and NRR selectivity, (c) free energy change versus changes in N–N bond lengths for the reaction intermediates NHNH₂ to NH₂NH₂, and correlation plots of electronic and geometric structural features in (d) SAC and (e) N₂ adsorption models. (PDS: potential-determining step, BCT: d band center of Ti, E_b: single-atom binding energy of Ti, Ti–C: the average bond length between Ti and C, I_C: ICOHP values for Ti and C, BT: the Bader charge of Ti, W_f: the work function of SACs, E_ads: the adsorption energy of N₂, N–N: bond lengths between nitrogen atoms in N₂, Ti–N: bond lengths between nitrogen atoms in N₂, BN: the Bader charge of the N₂ molecule, and BCN: the centre of the p-band of the N₂ molecule.)

Conclusion

In conclusion, this study systematically explored the impact of the sp/sp² hybrid carbon ratio on the electronic and catalytic properties of Ti@GY/Gr heterojunctions for the NRR. The free energy change for the PDS was found to be 0.32 eV, 0.88 eV, and 0.88 eV for Ti@GY1/Gr, Ti@GY2/Gr, and Ti@GY3/Gr, respectively, with Ti@GY1/Gr exhibiting the lowest energy barrier and thus the highest catalytic activity. The N₂ adsorption energies were −0.59 eV for Ti@GY3/Gr, 0.13 eV for Ti@GY1/Gr, and −0.17 eV for Ti@GY2/Gr, indicating that Ti@GY3/Gr had the strongest N₂ adsorption strength. However, the selectivity for the NRR over the HER was highest for Ti@GY1/Gr, highlighting the importance of balancing adsorption strength and reaction kinetics. The d-band center of Ti was found to shift closer to the Fermi level as the sp/sp² hybrid carbon ratio increased, which influenced the adsorption and activation of N₂. These findings provide valuable insights into the design of efficient NRR catalysts by fine-tuning the electronic and geometric structures through the adjustment of the sp/sp² hybrid carbon ratio. Future work could explore the potential of other transition metals and different heterojunction configurations to further improve the catalytic performance for the NRR.

Data availability

The data supporting this article have been included as part of the ESI.†

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (No. 52231008 and 52301011), the Hainan Provincial Natural Science Foundation of China (No. 524QN226), the Key Research and Development Program of Hainan Province (No. ZDYF2024GXJS006), the Starting Research Fund from the Hainan University (No. KYQD(ZR)23026), the International Science & Technology Cooperation Program of Hainan Province (No. GHYF2023007), and the Opening Project of Key Laboratory of Optoelectronic Chemical Materials and Devices, Ministry of Education, Jianghan University (No. JDGD-202315).

References

1. X. Li, J. You, J. Li, Z. Wang, Y. Zhao, J. Xu, M. Duan and H. Zhang, ChemCatChem, 2024, 16, e202301108.
2. X. Qian, C. Ma, U. B. Shahid, M. Sun, X. Zhang, J. Tian and M. Shao, ACS Catal., 2022, 12, 6385–6393.
3. X. Wang, M. Luo, J. Lan, M. Peng and Y. Tan, Adv. Mater., 2021, 33, 2007733.
4. G. Liu, R. Yao, J. You, L. Liu, B. Yi, Y. Zhao, Y. Li and H. Zhang, Mater. Today Commun., 2024, 39, 109052.
5. S. Wang, Y. Zhou, H. Fang, C. Nie, C. Q. Sun and B. Wang, Chin. Chem. Lett., 2024, 110476.
6. B. Qiao, A. Wang, X. Yang, L. F. Allard, Z. Jiang, Y. Cui, J. Liu, J. Li and T. Zhang, Nat. Chem., 2011, 3, 634–641.
7. W. Zhang, Q. Fu, Q. Luo, L. Sheng and J. Yang, JACS Au, 2021, 1, 2130–2145.
8. Z. Geng, Y. Liu, X. Kong, P. Li, K. Li, Z. Liu, J. Du, M. Shu, R. Si and J. Zeng, Adv. Mater., 2018, 30, 1803498.
9. A. Wang, J. Li and T. Zhang, Nat. Rev. Chem, 2018, 2, 65–81.
10. T. Zhou, Y. Guan, C. He, L. Zhang, X. Sun, Z. Song, Q. Zhang, C. He, X. Jiang, Z. Luo, W. Xing and X. Ren, Carbon Energy, 2024, 6, e477.
11. J. Wang, Z. Li, Y. Wu and Y. Li, Adv. Mater., 2018, 30, 1801649.
12. L. Cao, Q. Luo, W. Liu, Y. Lin, X. Liu, Y. Cao, W. Zhang, Y. Wu, J. Yang, T. Yao and S. Wei, Nat. Catal., 2018, 2, 134–141.
13. B. Guo, Z. Wang, L. Zheng, G. Mo, H. Zhou and D. Luo, Carbon Energy, 2024, 6, e554.
14. Y. Chen, R. Gao, S. Ji, H. Li, K. Tang, P. Jiang, H. Hu, Z. Zhang, H. Hao, Q. Qu, X. Liang, W. Chen, J. Dong, D. Wang and Y. Li, Angew. Chem., Int. Ed., 2021, 60, 3212–3221.
15. H.-Y. Gong, X. Liang, G.-L. Sun, D.-W. Li, X.-J. Zheng, H. Shi, K. Zeng, G.-C. Xu, Y. Li, R.-Z. Yang and C.-Z. Yuan, Rare Met., 2022, 41, 4034–4040.
16. W. Lai, L. Zhang, W. Hua, S. Indris, Z. Yan, Z. Hu, B. Zhang, Y. Liu, L. Wang, M. Liu, R. Liu, Y. Wang, J. Wang, Z. Hu, H. Liu, S. Chou and S. Dou, Angew. Chem., Int. Ed., 2019, 58, 11868–11873.
17. Y. Ying, K. Fan, X. Luo, J. Qiao and H. Huang, J. Mater. Chem. A, 2021, 9, 16860–16867.
18. H. Yang, Y. Wu, G. Li, Q. Lin, Q. Hu, Q. Zhang, J. Liu and C. He, J. Am. Chem. Soc., 2019, 141, 12717–12723.
19. K. Jiang, S. Siahrostami, T. Zheng, Y. Hu, S. Hwang, E. Stavitski, Y. Peng, J. Dynes, M. Gangisetty, D. Su, K. Attenkofer and H. Wang, Energy Environ. Sci., 2018, 11, 893–903.
20. Y. Wang, X. Ma, Ayeza, C. Wang, Y. Li, C. Yang, Z. Wang, C. Wang, C. Hu and Y. Zhang, New Carbon Mater., 2024, 39, 407–438.
21. P. Zhou, F. Lv, N. Li, Y. Zhang, Z. Mu, Y. Tang, J. Lai, Y. Chao, M. Luo, F. Lin, J. Zhou, D. Su and S. Guo, Nano Energy, 2019, 56, 127–137.
22. X. X. Wang, D. A. Cullen, Y. Pan, S. Hwang, M. Wang, Z. Feng, J. Wang, M. H. Engelhard, H. Zhang, Y. He, Y. Shao, D. Su, K. L. More, J. S. Spendelow and G. Wu, Adv. Mater., 2018, 30, 1706758.
23. Q. Lai, L. Zheng, Y. Liang, J. He, J. Zhao and J. Chen, ACS Catal., 2017, 7, 1655–1663.
24. Y. Li, X. Liu, L. Zheng, J. Shang, X. Wan, R. Hu, X. Guo, S. Hong and J. Shui, J. Mater. Chem. A, 2019, 7, 26147–26153.
25. J. Zhang, Y. Zhao, C. Chen, Y.-C. Huang, C.-L. Dong, C.-J. Chen, R.-S. Liu, C. Wang, K. Yan, Y. Li and G. Wang, J. Am. Chem. Soc., 2019, 141, 20118–20126.
26. X. Zhai, H. Yan, G. Ge, J. Yang, F. Chen, X. Liu, D. Yang, L. Li and J. Zhang, Appl. Surf. Sci., 2020, 506, 144941.
27. Z. Feng, Y. Tang, W. Chen, Y. Li, R. Li, Y. Ma and X. Dai, Phys. Chem. Chem. Phys., 2020, 22, 9216–9224.
28. M. Li, Z. Liu, T. Gao, S. Liu, H. Yu, Z. Wang and H. Sun, Phys. Chem. Chem. Phys., 2023, 25, 4803–4809.
29. M. Li, Q. Fang, X. Zhao, C. Xia, A. Wang, Y. Xie, F. Ma, J. She and Z. Deng, Phys. Chem. Chem. Phys., 2023, 25, 18224–18232.
30. L. Yang, M. Li, D. Holiharimanana, T. Gao, Z. Liu, S. Liu and H. Sun, Mol. Catal., 2023, 551, 113665.
31. Z.-X. Yin, Y.-D. Li, Y.-H. Ye, Y. Liu, M.-F. Li, Z.-J. Yang, X.-R. Zheng, H.-Z. Wang, Y. Wang and Y.-D. Deng, Rare Met., 2024, 43, 5781–5791.
32. X. Gao, Y. Zhu, D. Yi, J. Zhou, S. Zhang, C. Yin, F. Ding, S. Zhang, X. Yi, J. Wang, L. Tong, Y. Han, Z. Liu and J. Zhang, Sci. Adv., 2018, 4, eaat6378.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5ta01226k

‡ Zexiang Yin and Zijun Yang contributed equally to this work.

---