English

• Hydrogen peroxide (H₂O₂) is a versatile and green oxidant with widespread applications in chemical synthesis, pulp industry, wastewater treatment and energy storage [1-3]. The state-of-the-art process for industrial H₂O₂ production mainly relies on the anthraquinone oxidation method, where complex operations and large scale infrastructures are inevitably required with extensive toxic by-products [4]. Photocatalytic oxygen reduction reaction (ORR) provides a promising alternative route for on-demand H₂O₂ production under ambient conditions by using O₂ as the raw material, sunlight as energy input [4-9]. To achieve high-efficiency H₂O₂ production, the rational design of photocatalysts is the key.

  A plenty of semiconducting materials have been reported for photocatalytic H₂O₂ production, such as metal oxides [10-12], metal sulfides [13-15], carbon nitride [16-18], polymers [19-21], covalent organic frameworks [22-24] and metal-organic frameworks (MOFs) [25-27]. Among them, MOFs, as a type of porous crystalline materials, have gained particular interests due to the appealing features of unique ligand-metal charge transfer transitions, large specific surface area, high porosity, variable composition and structure, and easy functionalization [28-30]. To improve the photocatalytic performance of MOFs, promoting the separation of photogenerated electrons and holes is one of the most important step [31,32]. To this end, loading cocatalysts including oxidation cocatalysts (e.g., PbO₂, CoO_x, MnO_x, PdO_x) for capturing holes and reduction cocatalysts (e.g., Pt, Pd) for accumulating electrons is considered as a promising strategy, which has also been widely applied for MOFs [12,33]. By tuning the types, surfactant coating degree, spatial distribution of cocatalysts [34-40], the charge separation property of MOFs has been greatly enhanced in various photocatalytic reactions including CO₂ reduction [40], water splitting [41], hydrogen reduction [42] and H₂O₂ production [40,41,43]. For crystals such as MOFs, the facet engineering can regulate the surface arrangement, exposure and chemical environment of ligands and metals, which may alter the deposition states of cocatalysts and eventually affect the photocatalytic performance [37,38,44,45]. Thanks to the rapid development of synthetic methodologies [46-49], fabrication of MOFs with well-defined and controlled crystal facets can be achieved, thus providing platforms for exploring the facet effect of MOFs on supporting cocatalysts for photocatalytic reactions, especially for ORR toward H₂O₂ production. Nevertheless, relative studies have been seldomly reported.

  Herein, a representative semiconducting MOF, MIL-125-NH₂ (MIL = Materials Institute Lavoisier, denoted as NM), with controllable facet exposure is adopted to load Pd-based co-catalysts for H₂O₂ production via photocatalytic ORR process (Scheme 1). By altering the synthetic condition, {001} dominated NM-1, {001}/{111} co-exposed NM-2 and {111} dominated NM-3 are fabricated. As supports for Pd loading, the distinct surface chemical environments of {001} and {111} facets induce the facet-selective loading of Pd⁰ and PdO dominated cocatalysts, respectively, resulting in Pd-NM-1 (Scheme 1a), Pd-NM-2 (Scheme 1c) and Pd-NM-3 (Scheme 1b). The photocatalytic results show that the Pd-NM-2 exhibits the highest activity with a H₂O₂ production rate of 128.6 mmol L^-1 g^-1 h^-1, outperforming NM-1 to 3, Pd-NM-1 and Pd-NM-3. The performance enhancement of Pd-NM-2 is ascribed to the loading of spatially separated PdO and Pd⁰ dual cocatalysts, where Pd⁰ on {001} facets as the reduction cocatalyst can efficiently trap the photoexcited electrons and PdO on {111} facets as the oxidation cocatalyst can capture the holes. Consequently, the photogenerated electron-hole pairs can be directionally separated, resulting in improved photocatalytic activity.

  Scheme 1

  

  Scheme 1.  Illustration of the photocatalytic mechanism and band structure of Pd-NM-1, 2 and 3.

  DownLoad: Full-Size Img PowerPoint

  MIL-125-NH₂ with different facets were prepared by using different titanium sources of titanium tetraisopropanolate (TPOT) or titanium butoxide (TBOT) with or without acetic acid (AA) as a modulator during the solvothermal process (see specific details in Experimental section in Supporting information). MIL-125-NH₂ dominated with {001} facets, {001}/{111} facets and {111} facets are denoted as NM-1, 2 and 3, respectively. Scanning electron microscopy (SEM) images (Figs. S1a and b in Supporting information) show regular nanocake morphology of NM-1 with two {001} facets exposed on the top-bottom surfaces [39]. NM-2 exhibits truncated tetragon-like morphology with exposure of two {001} and eight {111} facets (Figs. S1c and d in Supporting information). With the modulation by AA, the {001} facets disappear with {111} facets exposed on all surfaces of NM-3 with octahedral shape (Figs. S1e and f in Supporting information).

  The crystal structures of NM-1, 2 and 3 were studied by powder X-ray diffraction (XRD). As shown in Fig. S2a (Supporting information), the diffraction patterns of all samples are well matched with the simulated MIL-125-NH₂ [39,50]. From NM-1 to NM-3, the relative intensity ratio of {002}/{222} gradually decreases from 2.78, 1.72 to 1.59, verifying the increased exposure of {111} facet, in accordance with the SEM observations. Fig. S3 displays the N₂ adsorption-desorption curves of NM-1, 2 and 3, all exhibiting type I isotherms with micropore-dominated nature. The Brunauer-Emmett-Teller (BET) specific surface areas and total pore volumes are calculated to be 1114.3, 1046.5, 1174.1 m²/g and 0.67, 0.61, 1.05 cm³/g for NM-1, NM-2 and NM-3, respectively (Table S1 in Supporting information).

  Through a chemical impregnation-reduction process by reacting NM with K₂PdCl₄·2H₂O in methanol solution, three Pd-NM composites were obtained. Taking Pd-NM-2 as a typical example (Fig. 1a), the structure and composition were extensively investigated. SEM images of Pd-NM-2 show well-preserved truncated tetragon-like nanocakes. At higher magnification, nanoparticles (NPs) with brighter contrast are found to be uniformly distributed on the surface of NM-2 (Figs. 1b and c). Similar observation is also presented by the transmission electron microscopy (TEM) image (Fig. 1d). The average diameter of NPs is measured to be ~5–10 nm. The high-resolution TEM (HRTEM) images (Figs. 1e and f) of the NPs located on the top-down (region e in Fig. 1d) and side (region f in Fig. 1d) surfaces show clear lattices with d-spacing distances of 0.226 and 0.262 nm, corresponding to the (111) and (101) planes of Pd⁰ and PdO, respectively [51-53]. This suggests the spatially separated loading of Pd⁰ and PdO NPs on {001} and {111} facets of NM-2, respectively. The high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) image and corresponding elemental mapping images (Fig. 1g) show that the Ti, C, O and N rich nanocake was surrounded by Pd rich NPs.

  Figure 1

  

  Figure 1.  (a) Illustration of the synthetic process of Pd-NM-2. (b, c) SEM, (d) TEM, (e, f) HRTEM, (g) HAADF-STEM and corresponding element mapping images of Pd-NM-2. (h) XRD patterns and (i) FTIR spectra of NM-2 and Pd-NM-2. Scale bars are (b) 1 µm, (c) 250 nm, (d) 300 nm, (e, f) 2 nm, (g) 250 nm, respectively.

  DownLoad: Full-Size Img PowerPoint

  Fig. 1h presents the XRD pattern of Pd-NM-2, where both the diffraction peaks of MIL-125-NH₂, Pd and PdO are detected, further verifying the growth of Pd NPs and PdO onto NM-2. The Fourier transform infrared (FTIR) spectrum of Pd-NM-2 is displayed in Fig. 1i with NM-2 as comparison. The typical groups of MIL-125-NH₂ (e.g., amino group at 3200–3700 cm^-1, carboxylate group at 1400–1709 cm^-1, and Ti-O bond at 400–800 cm^-1) are observed in both NM-2 and Pd-NM-2 [54]. In the spectrum of Pd-NM-2, the stretching vibration at 1704 cm^-1 is assigned to the Pd-O bond [33,55], indicating the formation of PdO species, in agreement with the HRTEM observation. For Pd-NM-1 and 3, Pd NPs with similar diameter are also evenly adhered on the surface of NM-1 and 3, respectively (Fig. S4 in Supporting information). In the FTIR spectrum of Pd-NM-3, the peak intensity of Pd-O bond is stronger than that for Pd-NM-1 due to the formation of higher amount of PdO species with the exposure of more {111} facets (Fig. S5 in Supporting information). In contrast, extremely weak Pd-O peak is observed in the spectrum of Pd-NM-1, implying the predominated existence form of Pd⁰. The N₂ adsorption-desorption curves of Pd-NM-1, 2 and 3 are shown in Fig. S6 (Supporting information) and the BET specific surface areas and total pore volumes are measured to be 959.9, 633.9, 858.4 m²/g and 0.56, 0.41, 0.89 cm³/g for Pd-NM-1, 2 and 3, respectively (Table S3 in Supporting information). Furthermore, the Pd/Ti molar ratios of Pd-NM-1, 2 and 3 are determined to be 3.45, 3.42 and 3.42 by inductive coupled plasma emission spectrometer (Table S2 in Supporting information), respectively, demonstrating almost same Pd loading amount in three samples.

  X-ray photoelectron spectroscopy (XPS) measurement was further performed to investigate the surface chemical states of the Pd-NM materials. In the XPS survey spectra, Pd, Ti, N, C and O elements were detected in all samples. The high-resolution Pd 3d spectra of Pd-NM (Fig. 2a) show two intense peaks at 341.0 and 335.7 eV assigned to Pd 3d_5/2 and Pd 3d_3/2 orbitals of Pd⁰ [56,57], with two peaks at 343.3 and 338.0 eV attributed to Pd 3d_5/2 and Pd 3d_3/2 orbitals of Pd²⁺ [58,59]. Even the peak positions were close, the Pd⁰/Pd²⁺ ratio was varied (Fig. 2b) with an order of Pd-NM-1 (3.00) > Pd-NM-2 (1.81) > Pd-NM-3 (0.64), suggesting that the formation of Pd⁰ and PdO was preferred on {001} and {111} facets, respectively, in agreement with the HRTEM and FTIR results.

  Figure 2

  

  Figure 2.  (a) High-resolution XPS spectra of Pd 3d and (b) Pd⁰/Pd²⁺ ratios of Pd-NM-1, −2 and −3 calculated by (a). High-resolution XPS spectra of (c) Ti 2p and (d) O 1s of NM-2 and Pd-NM-2.

  DownLoad: Full-Size Img PowerPoint

  To further explore the electronic interaction between Pd and NM, the Ti 2p and O 1s spectra of Pd-NM-2 as a typical example are displayed in Fig. 2c in comparison with NM-2. The Ti 2p spectrum of NM-2 can be divided into two peaks of Ti 2p_3/2 and Ti 2p_1/2 of Ti⁴⁺ at 464.6 and 458.8 eV, respectively [54]. After Pd deposition, the binding energy of Ti 2p_3/2 peaks presents a positive shift of ≈0.3 eV for Pd-NM-2, indicating the electron transfer from Ti to Pd. In the O 1s spectrum of NM-2 (Fig. 2d), the two peaks at 531.9 and 530.3 eV are ascribed to the carboxylate group and M(Ti/Pd)-O bond, respectively [39]. For Pd-NM-2, the intensity of M-O peak increases with a negative shift of 0.2 eV, which may be due to the formation of PdO species. Similar to M-O, the binding energy of carboxylate group also negatively shifts, resulted by the electron transfer from Pd to carboxylate group with electron-withdrawing property [5,60,61]. Further according to previous reports [39,50], Ti-O clusters and carboxylate group are predominately exposed on the {111} and {001} facets, respectively, possibly inducing the facet-selective loading of PdO and Pd⁰, as evidenced by the FTIR and XPS results. Moreover, Pd⁰ and PdO were recognized as reduction and oxidation cocatalysts for accumulating electrons and capturing holes. Notably, a two-step process is inevitably required for sequentially loading oxidative and reductive co-catalysts onto different facets of semiconducting materials [33]. In our synthesis, the redox dual co-catalysts can be separately deposited onto two different facets of MOFs in a one-step growth process, which would greatly promote the electron-hole separation and improve the photocatalytic activity.

  The optical properties of the samples were examined by UV–vis diffuse reflectance spectra (DRS). As shown in Fig. S7 (Supporting information), all NM samples exhibit similar light absorption from the UV to visible region, in accordance with reported results [62,63]. The two absorption bands at ~300 and 400 nm correspond to the Ti-O clusters and ligand-based absorption, respectively. After loading Pd NPs, enhanced light absorption especially in the region of 500–800 nm is found in all Pd-NM samples, possibly attributed to the localized surface plasmon resonance absorption from the Pd NPs [64].

  The photocatalytic H₂O₂ production of the samples via 2e-ORR was assessed under simulated sunlight (AM 1.5 G) illumination. As shown in Figs. 3a and b, NM-1 to 3 exhibit relatively poor activity with low H₂O₂ amounts of 0.13, 0.10 and 0.11 mmol/L, respectively. After the deposition of Pd⁰/PdO cocatalysts, Pd-NM-2 delivers significantly enhanced H₂O₂ yield of 0.62 mmol/L, much higher than that of Pd-NM-1 (0.31 mmol/L) and Pd-NM-3 (0.41 mmol/L). Additionally, the H₂O₂ production rate of Pd-NM-2 is calculated to be 128.6 mmol L^-1 g^-1 h^-1, superior to most reported MOF-based photocatalysts and even other non-MOF heterostructures under similar reaction conditions (Fig. 3c and Table S4 in Supporting information), indicating its excellent photocatalytic activity for H₂O₂ production. Except for high activity, Pd-NM-2 also shows good photocatalytic stability. After six cycles, above 83% of H₂O₂ generation amount was retained compared with the initial cycle (Fig. 3d). In the SEM image and XRD pattern (Fig. S8 in Supporting information) of used Pd-NM-2, no obvious change of morphology crystalline structure is found, further suggesting the robust structure during photocatalysis.

  Figure 3

  

  Figure 3.  (a) Time-dependent H₂O₂ photoproduction profiles. (b) H₂O₂ evolution rate. (c) Comparison of H₂O₂ production rate with reported photocatalysts. (d) Cycling test of Pd-NM-2.

  DownLoad: Full-Size Img PowerPoint

  In order to understand the origin of enhanced performance of Pd-NM-2, a series of photo-/electro-chemical measurements were conducted. Firstly, linear sweep voltammetry (LSV) curves were collected to investigate the ORR activity of the samples. In the presence of light and O₂ (Fig. 4a), the current density of Pd-NM-2 is higher than that without O₂ or light, suggesting the occurrence of photocatalytic ORR [65,66]. Moreover, among the samples of Pd-NM-1 to 3, Pd-NM-2 displays the highest current density and most positive onset potential (Fig. 4b), revealing the highest ORR activity [58,67]. The photocarrier separation and transfer manners of the samples were explored by steady-state photoluminescence (PL) spectroscopy, time-resolved photoluminescence (TRPL) spectra and photocurrent response. As shown in Fig. 4c, Pd-NM-2 exhibits the weakest PL emission peak intensity among all samples with lowest electron-hole recombination efficiency. The TRPL spectra (Fig. 4d and Table S5 in Supporting information) display that Pd-NM-2 presents longer average decay lifetime (3.78 ns) than Pd-NM-1(2.64 ns) and Pd-NM-3 (3.19 ns), indicating that charge carrier lived longer in Pd-NM-2. In the photocurrent curves (Fig. 4e), the photocurrent density follows the order of Pd-NM-2 > Pd-NM-3 > Pd-NM-1. The PL, TRPL and photocurrent results indicate the crucial role of Pd⁰/PdO cocatalysts in facilitating the photocarrier migration and charge separation.

  Figure 4

  

  Figure 4.  (a) LSV curves of Pd-NM-1, 2 and 3 under light and O₂-enriched condition. (b) Pd-NM-2 under light and Ar-enriched condition, dark O₂-enriched condition, and light O₂-enriched condition. (c) PL spectra and (d) TRPL spectra, and (e) photocurrent curves of Pd-NM-1, 2 and 3. (f) ESR signals of DMPO-·O₂ ^- for Pd-NM-1, 2 and 3.

  DownLoad: Full-Size Img PowerPoint

  To gain further insight into the mechanism of photocatalytic H₂O₂ production, the 5, 5-dimethyl-pyrroline-N-oxide (DMPO) spin-trapping electron spin resonance (ESR) technique was applied for detecting the ·O₂^- radicals (Fig. 4f) [20]. Typical peaks assigned to DMPO-·O₂^- are detected in all Pd-NM samples. This implies that the generation of H₂O₂ goes through a two-step single-electron reduction process, where ·O₂^- acts as the key intermediate product (O₂ + e^- → ·O₂^- and ·O₂^- + 2H⁺ + e^- → H₂O₂) [20]. Moreover, Pd-NM-2 possesses the strongest signal intensities among these samples, demonstrating the higher activity.

  Collectively, the components of Pd⁰/PdO are selectively deposited onto the {001} and {111} facets of NM-2 via a one-step growth process, resulting in Pd-NM-2 decorated by spatially separated dual cocatalysts. During the photocatalysis process, Pd⁰ as the reduction cocatalyst can efficiently trap the photoexcited electrons from NM-2, whereas PdO as the oxidation cocatalyst can capture the holes. As a result, the photogenerated charges can be directionally separated [40], efficiently restricting the electron-hole recombination and thus resulting in improved photocatalytic performances than Pd-NM-1 and 3 with single cocatalyst.

  In summary, this work elucidates the facet effect of semiconducting MOFs on supporting cocatalysts for photocatalytic 2e-ORR to H₂O₂ production. Three types of MIL-125-NH₂ with facet exposure of {001}, {001}/{111} and {111} are employed for loading Pd-based co-catalysts. Specifically, the distinct surface chemical environments of {001} and {111} facets of MIL-125-NH₂ results in the facet-dependent loading of Pd⁰ and PdO dominated cocatalysts. Benefiting from the significantly facilitated charge separation by spatially separated dual co-catalysts, the optimized sample of Pd-NM-2 exhibits a H₂O₂ production rate of 128.6 mmol L^-1 g^-1 h^-1, superior to pristine and single cocatalyst modified MIL-125-NH₂ samples.

  Declaration of competing 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.

  CRediT authorship contribution statement

  Peiyang Du: Writing – original draft, Conceptualization. Ling Yuan: Writing – original draft, Data curation, Conceptualization. Tong Bao: Formal analysis. Yamin Xi: Formal analysis. Jiaxin Li: Formal analysis. Yin Bi: Formal analysis. Luli Yin: Formal analysis. Jing Wang: Writing – review & editing, Conceptualization. Chao Liu: Writing – review & editing.

  Acknowledgments

  The authors acknowledge support from the National Natural Science Foundation of China (NSFC, Nos. 51908218 and 21905092), the Youth Elite Sailing Program of Shanghai Institute of Technology (No. 1021GK240006002-A07), Scientific Research Foundation of Shanghai Institute of Technology (No. 10120K226156-A06-YJ2022-62), Shanghai Higher Education Institution Young Teacher Training Funding Program (No. ZZ202312031) and the Key Laboratory of Jiangsu Province for Chemical Pollution Control and Resources Reuse (Nanjing University of Science and Technology).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.110472.

  
  
• 1. [1]

     A. Santos, R.J. Lewis, G. Malta, et al., Ind. Eng. Chem. Res. 58 (2019) 12623–12631. doi: 10.1021/acs.iecr.9b02211

  2. [2]

     S. Li, K. Rong, X. Wang, et al., Acta Phys. Chim. Sin. 40 (2024) 2403005. doi: 10.3866/pku.whxb202403005

  3. [3]

     Y. Xue, Y. Wang, Z. Pan, et al., Angew. Chem. Int. Ed. 60 (2021) 10469–10480. doi: 10.1002/anie.202011215

  4. [4]

     J.M. Campos-Martin, G. Blanco-Brieva, J.L.G. Fierro, Angew. Chem. Int. Ed. 45 (2006) 6962–6984. doi: 10.1002/anie.200503779

  5. [5]

     Z. Jiang, X. Xu, Y. Ma, et al., Nature 586 (2020) 549–554. doi: 10.1038/s41586-020-2738-2

  6. [6]

     L. Zhou, W. Zhou, M. Yu, et al., J. Liaocheng Univ. (Nat. Sci. Ed. ) 37 (2024) 50–61.

  7. [7]

     C. Zhuang, Y. Chang, W. Li, et al., ACS Nano 18 (2024) 5206–5217. doi: 10.1021/acsnano.4c00217

  8. [8]

     Q. Zhang, M. Zhou, G. Ren, et al., Nat. Commun. 11 (2020) 1731.

  9. [9]

     Z. Tian, C. Han, Y. Zhao, et al., Nat. Commun. 12 (2021) 2039.

  10. [10]

      B. He, Z. Wang, P. Xiao, et al., Adv. Mater. 34 (2022) 2203225.

  11. [11]

      K. Kim, J. Park, H. Kim, et al., ACS Catal. 9 (2019) 9206–9211. doi: 10.1021/acscatal.9b02269

  12. [12]

      O. Tomita, T. Otsubo, M. Higashi, et al., ACS Catal. 6 (2016) 1134–1144. doi: 10.1021/acscatal.5b01850

  13. [13]

      S. Li, K. Dong, M. Cai, et al., eScience (2024) 100208.

  14. [14]

      J. Luo, X. Wei, Y. Qiao, et al., Adv. Mater. 35 (2023) 2210110.

  15. [15]

      A. Wang, H. Liang, F. Chen, et al., Appl. Catal. B 310 (2022) 121336.

  16. [16]

      B. Liu, J. Du, G. Ke, et al., Adv. Funct. Mater. 32 (2022) 2111125.

  17. [17]

      L. Zhou, J. Lei, F. Wang, et al., Appl. Catal. B 288 (2021) 119993.

  18. [18]

      L. Shi, L. Yang, W. Zhou, et al., Small 14 (2018) 1703142.

  19. [19]

      H. Cheng, J. Cheng, L. Wang, et al., Chem. Mater. 34 (2022) 4259–4273. doi: 10.1021/acs.chemmater.2c00936

  20. [20]

      C. Zhao, X. Wang, Y. Yin, et al., Angew. Chem. Int. Ed. 62 (2023) e202218318.

  21. [21]

      L. Yuan, C. Zhang, J. Wang, et al., Nano Res. 14 (2021) 3267–3273. doi: 10.1007/s12274-021-3517-6

  22. [22]

      P. Das, J. Roeser, A. Thomas, Angew. Chem. Int. Ed. 62 (2023) e202304349.

  23. [23]

      M. Kou, Y. Wang, Y. Xu, et al., Angew. Chem. Int. Ed. 61 (2022) e202200413.

  24. [24]

      H. Wang, C. Yang, F. Chen, et al., Angew. Chem. Int. Ed. 61 (2022) e202202328.

  25. [25]

      L. Yuan, Y. Zou, L. Zhao, et al., Appl. Catal. B 318 (2022) 121859.

  26. [26]

      Y. Li, F. Ma, L. Zheng, et al., Mater. Horiz. 8 (2021) 2842–2850. doi: 10.1039/d1mh00869b

  27. [27]

      J. Qiu, D. Dai, J. Yao, Coord. Chem. Rev. 501 (2024) 215597.

  28. [28]

      W. Liang, P. Wied, F. Carraro, et al., Chem. Rev. 121 (2021) 1077–1129. doi: 10.1021/acs.chemrev.0c01029

  29. [29]

      B. He, Q. Zhang, Z. Pan, et al., Chem. Rev. 122 (2022) 10087–10125. doi: 10.1021/acs.chemrev.1c00978

  30. [30]

      N.F. Suremann, B.D. McCarthy, W. Gschwind, et al., Chem. Rev. 123 (2023) 6545–6611. doi: 10.1021/acs.chemrev.2c00587

  31. [31]

      J. Xiao, Q. Han, Y. Xie, et al., Environ. Sci. Technol. 51 (2017) 13380–13387. doi: 10.1021/acs.est.7b04215

  32. [32]

      Q. Wang, T. Hisatomi, Q. Jia, et al., Nat. Mater. 15 (2016) 611–615. doi: 10.1038/nmat4589

  33. [33]

      L. Bai, F. Ye, L. Li, et al., Small 13 (2017) 1701607.

  34. [34]

      R. Li, H. Han, F. Zhang, et al., Energy Environ. Sci. 7 (2014) 1369–1376.

  35. [35]

      M. Li, S. Yu, H. Huang, et al., Angew. Chem. Int. Ed. 58 (2019) 9517–9521. doi: 10.1002/anie.201904921

  36. [36]

      L. Mu, Y. Zhao, A. Li, et al., Energy Environ. Sci. 9 (2016) 2463–2469.

  37. [37]

      R. Li, F. Zhang, D. Wang, et al., Nat. Commun. 4 (2013) 1432.

  38. [38]

      J. Pan, G. Liu, G.Q.M. Lu, et al., Angew. Chem. Int. Ed. 50 (2011) 2133–2137. doi: 10.1002/anie.201006057

  39. [39]

      X.M. Cheng, X.Y. Dao, S.Q. Wang, et al., ACS Catal. 11 (2020) 650–658.

  40. [40]

      C. Zhou, S. Wang, Z. Zhao, et al., Adv. Funct. Mater. 28 (2018) 1801214.

  41. [41]

      J. Zhang, T. Bai, H. Huang, et al., Adv. Mater. 32 (2020) 2004747.

  42. [42]

      A. Meng, J. Zhang, D. Xu, et al., Appl. Catal. B 198 (2016) 286–294.

  43. [43]

      T. Liu, Z. Pan, J.J.M. Vequizo, et al., Nat. Commun. 13 (2022) 1034.

  44. [44]

      X. Xie, Y. Li, Z.Q. Liu, et al., Nature 458 (2009) 746–749. doi: 10.1038/nature07877

  45. [45]

      W. Jiang, C. Ni, L. Zhang, et al., Angew. Chem. Int. Ed. 61 (2022) e202302575.

  46. [46]

      C.Y. Liu, X.R. Chen, H.X. Chen, et al., J. Am. Chem. Soc. 142 (2020) 6690–6697. doi: 10.1021/jacs.0c00368

  47. [47]

      F.L. Li, Q. Shao, X. Huang, et al., Angew. Chem. Int. Ed. 57 (2018) 1888–1892. doi: 10.1002/anie.201711376

  48. [48]

      L. Jiao, Y. Wang, H.L. Jiang, et al., Adv. Mater. 30 (2018) 1703663.

  49. [49]

      M.S. Yao, J.W. Xiu, Q.Q. Huang, et al., Angew. Chem. Int. Ed. 131 (2019) 15057–15061. doi: 10.1002/ange.201907772

  50. [50]

      F. Guo, J.H. Guo, P. Wang, et al., Chem. Sci. 10 (2019) 4834–4838. doi: 10.1039/c8sc05060k

  51. [51]

      J. Ge, D. He, W. Chen, et al., J. Am. Chem. Soc. 138 (2016) 13850–13853. doi: 10.1021/jacs.6b09246

  52. [52]

      B.T. Meshesha, N. Barrabés, K. Föttinger, et al., Appl. Catal. B 117-118 (2012) 236–245.

  53. [53]

      T.J. Wang, F.M. Li, H. Huang, et al., Adv. Funct. Mater. 30 (2020) 2000534.

  54. [54]

      L. Yuan, C. Zhang, Y. Zou, et al., Adv. Funct. Mater. 33 (2023) 2214627.

  55. [55]

      Z. Sorinezami, H. Mansouri-Torshizi, M. Aminzadeh, J. Biomol. Struct. Dyn. 37 (2019) 4238–4250. doi: 10.1080/07391102.2018.1546619

  56. [56]

      Y. Tian, W. Li, C. Zhao, et al., Appl. Catal. B 213 (2017) 136–146. doi: 10.1145/3083187.3083198

  57. [57]

      C. Zhang, C. Xie, Y. Gao, et al., Angew. Chem. Int. Ed. 61 (2022) e202204108.

  58. [58]

      W. Li, X.S. Chu, F. Wang, et al., Appl. Catal. B 288 (2021) 120034.

  59. [59]

      M. Guo, Q. Meng, W. Chen, et al., Angew. Chem. Int. Ed. 62 (2023) e202305212.

  60. [60]

      Q. Wu, J. Cao, X. Wang, et al., Nat. Commun. 12 (2021) 483.

  61. [61]

      J.Y. Bai, L.J. Wang, Y.J. Zhang, et al., Appl. Catal. B 266 (2020) 118590.

  62. [62]

      X.M. Cheng, X.Y. Dao, S.Q. Wang, et al., ACS Catal. 11 (2021) 650–658. doi: 10.1021/acscatal.0c04426

  63. [63]

      X.M. Cheng, P. Wang, S.Q. Wang, et al., ACS Appl. Mater. Interfaces 14 (2022) 32350–32359. doi: 10.1021/acsami.2c05037

  64. [64]

      C. Zhou, S. Wang, Z. Zhao, et al., Adv. Funct. Mater. 28 (2018) 1801214.

  65. [65]

      H. Li, Y. Sun, B. Cai, et al., Appl. Catal. B 170-171 (2015) 206–214.

  66. [66]

      B. Xu, Y. Li, Y. Gao, et al., Appl. Catal. B 246 (2019) 140–148.

  67. [67]

      J.H. Lee, H. Cho, S.O. Park, Appl. Catal. B 284 (2021) 119690.

• 

  Scheme 1  Illustration of the photocatalytic mechanism and band structure of Pd-NM-1, 2 and 3.

  下载: 全尺寸图片 幻灯片
  

  Figure 1  (a) Illustration of the synthetic process of Pd-NM-2. (b, c) SEM, (d) TEM, (e, f) HRTEM, (g) HAADF-STEM and corresponding element mapping images of Pd-NM-2. (h) XRD patterns and (i) FTIR spectra of NM-2 and Pd-NM-2. Scale bars are (b) 1 µm, (c) 250 nm, (d) 300 nm, (e, f) 2 nm, (g) 250 nm, respectively.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (a) High-resolution XPS spectra of Pd 3d and (b) Pd⁰/Pd²⁺ ratios of Pd-NM-1, −2 and −3 calculated by (a). High-resolution XPS spectra of (c) Ti 2p and (d) O 1s of NM-2 and Pd-NM-2.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) Time-dependent H₂O₂ photoproduction profiles. (b) H₂O₂ evolution rate. (c) Comparison of H₂O₂ production rate with reported photocatalysts. (d) Cycling test of Pd-NM-2.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  (a) LSV curves of Pd-NM-1, 2 and 3 under light and O₂-enriched condition. (b) Pd-NM-2 under light and Ar-enriched condition, dark O₂-enriched condition, and light O₂-enriched condition. (c) PL spectra and (d) TRPL spectra, and (e) photocurrent curves of Pd-NM-1, 2 and 3. (f) ESR signals of DMPO-·O₂ ^- for Pd-NM-1, 2 and 3.

  下载: 全尺寸图片 幻灯片
•