English

• Hydrogen energy is hailed as a critical future energy source to achieve carbon neutrality, owing to its intrinsic nature of green energy and high energy density [1–4]. However, conventional hydrogen production relies mainly on fossil energy sources, which are usually accompanied with significant carbon emissions [5–7]. Among the myriad related technologies under development, photocatalytic water splitting is considered as an ideal technology for sustainable and green hydrogen production [8–11]. Since 1972 [12], many photocatalysts have been explored for hydrogen production, including metal-organic frameworks [13–15], metal oxides [16–18], transition metal sulfides [19,20], and conjugated polymers [21,22]. Recently, polymer carbon nitride (PCN), a typical class of conjugated polymer photocatalysts, has garnered significant attention due to its advantages including visible light response, chemical stability, and favorable energy band structures for water splitting [23–27]. However, the conventional method of solid-phase thermal polymerization for synthesis of PCN typically accompany with incomplete polymerization, due to the slow diffusion of reactants and heat, which in turn results in a diminished quantum efficiency for photocatalytic hydrogen production [28–32]. Therefore, it is important to improve the photocatalytic efficiency by rational design the synthesis process and thereby modification in the structure and properties of PCN.

  To enhance the polymerization degree of PCN, various advanced techniques have been adopted, including microwave-assisted heating [33], solvent-thermal synthesis [34], and salt melt synthesis (SMS) [35–39]. Among these methods, SMS has attracted particular attentions, owing to its ability to create a liquid environment at elevated temperatures, which accelerates heat and mass transfer, thereby enhances the extension of the π-conjugated system of PCN [40]. For instance, Bojdys et al. successfully synthesized highly crystalline carbon nitride in the presence of LiCl/KCl molten salt conditions [41]. Up to now, various eutectic salt mixtures, such as LiCl/NaCl [42], LiCl/KCl [43], LiBr/NaBr [38] and NaCl/KCl [25], have been successfully developed for the synthesis of poly(heptazine imide) (PHI), which features a condensed conjugated structure. Interestingly, all the results indicate that the use of molten salts as solvents for the preparation of carbon nitride materials will introduce metal ions into the carbon nitride framework [37,44–46]. These metal cations are principally coordinated with negatively charged nitrogen atoms on the imine bridges [47]. Density functional theory analysis shows that the introduction of metal cations alters the local electronic structure of carbon nitride, resulting in the confinement of the photogenerated electrons to the embedded regions of the metal cations, while the holes are distributed away from the embedded regions [48]. This spatial separation of electrons and holes effectively promotes the separation of photogenerated carriers and improves photocatalytic efficiency. The importance of the metal ions in carbon nitride for carrier separation was further demonstrated by Deng et al., who showed that a decrease in the Na⁺ content in PHI led to a significant decrease in the catalytic activity under prolonged light exposure [49]. Lotsch' group found that the different ions present in the pores of PHI have a significant impact on the catalyst's conductivity, which is a key factor influencing the photocatalytic hydrogen production performance [50]. Therefore, rational control of the cations in PHI through molten salts holds great potential for further enhancement of its photocatalytic performance. Although there are some studies have been investigated to understand the role of cations present in PHI, the relationship between cation and its photocatalytic performance remains unclear. To gain a deep insight into this relationship, ternary molten salt systems, such as LiCl/NaCl/KCl [51], have been considered as ideal models to investigate the "structure-activity" relationship. Ternary molten salts system offers a broad tuning range, allowing for precise compositional adjustments to optimize the polymerization process and thus to enable the enhancement in photocatalytic performance.

  In this study, a series of PHI photocatalysts with varying composition, morphologies and properties were synthesized utilizing an optimized ternary molten salt mixture of LiCl/NaCl/KCl as solvent and template (Scheme 1). The results indicate that PHI with nanorod structure is formed in the presence of ternary molten salts. As a result, the nanorod structure of PHI-LiNaK shortens the migration distance of photo-excited charge carriers, thereby markedly promoting the separation efficiency of electron-hole pairs and enhancing the photocatalytic activity for hydrogen production. After in-situ photodeposition of Pt nanoparticles as hydrogen evolution cocatalysts in an aqueous solution using triethanolamine (TEOA) as a sacrificial agent, the apparent quantum efficiency of optimized PHI-LiNaK for photocatalytic hydrogen evolution reaches up to 52.9% at 420 nm.

  Scheme 1

  

  Scheme 1.  A proposed process for the synthesis of PHI with different morphologies in the presence of different types of molten salts.

  DownLoad: Full-Size Img PowerPoint

  Typically, ternary molten salt mixtures (LiCl/NaCl/KCl) and binary molten salts (LiCl/NaCl, LiCl/KCl, and NaCl/KCl) were selected to control the polymerization process of PHI, and the obtained samples were denoted as PHI-LiNaK, PHI-LiNa, PHI-LiK, and PHI-NaK, respectively. More details can be found in Supporting information. The morphology and microstructure of the synthesized samples were investigated using scanning electron microscopy (SEM) and high-resolution transmission electron microscopy (HR-TEM). As shown in Figs. 1a-c, the PHI samples synthesized using binary molten salts exhibit fragmentary and two-dimensional sheet-like structures. Interestingly, PHI-LiNaK (Fig. 1d) displays an elongated nanorod morphology (~20 nm in diameter), which is notably different from the PHI samples synthesized in the presence of binary molten salts. The nanorod morphology is also clearly confirmed in the high-resolution TEM images as shown in Fig. S2 (Supporting information). The formation of nanorods in the presence of ternary LiCl/NaCl/KCl can be attributed to the more stringent molten salt conditions (higher LiCl content and lower NaCl/KCl content), which affect the expansion of the crystal in the two-dimensional plane. The exposed surface of PHI-LiNaK reveals lattice fringes of 1.03 nm (Fig. 1e), corresponding to the in-plane repeating units [52,53]. The limited number of in-plane repeating units further suggests that the in-plane growth of polymer crystals is strongly restricted. Additionally, the lattice spacing of 0.32 nm observed in Fig. 1f corresponds to the interlayer stacking [52,53]. It is also evident to observe from the TEM image that the interlayer stacking in PHI-LiNaK is also restrained.

  Figure 1

  

  Figure 1.  (a-d) SEM image of PHI-LiNa, PHI-LiK, PHI-NaK and PHI-LiNaK. (e, f) HR-TEM image of PHI-LiNaK. Inset: Enlargement of the selected area. (g) Powder XRD patterns, (h) FT-IR spectra of PHI-LiNaK, PHI-LiNa, PHI-LiK and PHI-NaK. (i) Solid-state ¹³C NMR spectrum of PHI-LiNaK.

  DownLoad: Full-Size Img PowerPoint

  The chemical structure of the catalyst was further characterized using X-ray diffraction (XRD), Fourier transform infrared (FT-IR) spectroscopy, nuclear magnetic resonance (NMR), and Raman spectra (Raman). As shown in Fig. 1g, all the PHI samples showed two major diffraction peaks at around 8° and 27°, which corresponded to the stacking patterns of the in-plane repeating units corresponding to the heptazine units and the interlayer stacking mode of graphite-phase lamellar structure (corresponding to the HRTEM results), respectively, which belonged to the typical packing and stacking mode of classical PHI. The crystalline changes are primarily due to the intercalation of ions. Furthermore, the angular shift of the diffraction peak around 27° can be attributed to variations in the hydrated radii of different metal cations (Na⁺ 3.6 Å, K⁺ 3.3 Å), which induce shifts in the diffraction angle. The backbone structure of the catalysts was further investigated by FT-IR (Fig. 1h). The spectra displayed similarities across all four PHI catalysts, indicating that the different molten salts do not significantly affect the fundamental backbone of the heptazine based framework. The sharp peak at 804 cm⁻¹ corresponds to the out-of-plane bending vibration of the heptazine ring, while the peak at 990 cm⁻¹ is attributed to M⁺-NC₂, suggesting the incorporation of metal ions into the carbon nitride framework. The broad absorption at 1100–1700 cm⁻¹ is related to the stretching vibration of the heterocyclic ring of triazine/heptazine units. A distinct absorption peak at 2180 cm⁻¹ represents the unpolymerized terminal cyano group, and its intensity is related to the melting point of the molten salts [54]. Additionally, the broad absorption peak at 3000–3700 cm⁻¹ derives from the N—H of the terminal uncondensed amino group and the O—H vibration of the surface adsorbed water molecule [55].

  Solid-state ¹³C NMR analysis was performed on the PHI-LiNaK sample to confirm the basic chemical structure. The spectrum (Fig. 1i) revealed two distinct resonance signals consistent with previous reports. The resonance signal at 156 ppm corresponds to the characteristic carbon of N₂—C=N (C—N₃) in heptazine units, while the signal at 163 ppm represents the characteristic carbon of N=C—NH_x in heptazine units [52]. Additionally, the Raman spectra (Fig. S3 in Supporting information) also show that the four samples have similar structures. All the above characterizations demonstrate that, despite the significant morphological changes was obtained, there were no notable changes in the chemical structure of the PHI-based catalysts.

  The microscopic chemical structure and chemical composition of the catalyst were further characterized by using X-ray photoelectron spectroscopy (XPS, Fig. 2 and Fig. S4 in Supporting information) and inductively coupled plasma optical emission spectroscopy (ICP-OES, Table S1 in Supporting information). All PHI samples contain C, N, and O elements, and the PHI catalysts synthesized using different molten salts contain different contents of metal cations, which can also be revealed from inductively coupled plasma. The C 1s spectra (Fig. 2a) of PHI-LiNaK can be deconvoluted into three peaks, namely, the strong peak at 288.5 eV corresponds to the sp² hybridized carbon with N—C=N in the heterocyclic aromatic ring, 286.8 eV corresponds to the C—NH_x at the end of the nitrogen-containing heterocyclic ring or adsorbed CO₂, and 284.8 eV corresponds to the indeterminate carbon (for charge correction). The high-revolution N 1s spectra (Fig. 2b) inverse convolution has four peaks at 403.9, 401.4, 400.5, and 398.8 eV. In principle, the binding energy of 403.9 eV corresponds to the charge effect in the carbon-nitrogen heterocyclic ring, while the binding energy of 401.4 eV corresponds to the amino group (C—NH_x) deposited at the end of the surface. Besides, the binding energies of 400.5 eV and 398.9 eV correspond respectively to the triple-coordinated N (C—N₃) and double-coordinated N (C—N=C) in the heptazine ring respectively [56–58]. The XPS spectra of C 1s and N 1s of PHI-LiNaK only showed few changes compared to the other three PHI-based samples. The variation in binding energies can be attributed to the different coordination environments of N atoms and metal ions in different catalysts.

  Figure 2

  

  Figure 2.  XPS results of high-resolution spectra of C 1s (a), N 1s (b), Na 1s (c), K 2p (d), Li 1s (e) and Cl 2p (f) of PHI-LiNaK, PHI-LiNa, PHI-LiK and PHI-NaK.

  DownLoad: Full-Size Img PowerPoint

  The Na 1s and K 2p high-resolution XPS spectra (Figs. 2c and d) of the PHI-LiNaK sample exhibit significantly lower intensities compared to the other three samples, indicating differences in the Na⁺ and K⁺ contents within the catalysts. ICP-OES results further confirm the variations in the intensities of these metal ions in PHI-based catalysts synthesized in the presence of different molten salts. Additionally, a small amount of Li⁺ ions and Cl^- ions (Figs. 2e and f) are presented in the PHI catalyst. The presence of Li⁺ ions is known to suppress the growth of heptazine units and occupy coordination sites on the CN heterocycle that would otherwise be available for other alkali metal ions. This could be the reason of the observed variations in Na⁺ and K⁺ contents. The molten salt environment not only influences the ionic composition in the synthesized PHI catalysts but also acts as a structural directing agent to control the crystal growth process. Dong et al.'s research has shown that Na⁺ ions are located in the pores of the CN plane, whereas K⁺ ions reside between the CN layers [59]. In the case of PHI-LiNaK, the lower Na⁺ content restricts in-plane growth, while the presence of K⁺ promotes growth along the interlayer stacking direction as a c-axis structural directing agent [52]. Under the synergistic effects of Na⁺ and K⁺ ions, the PHI-LiNaK catalyst ultimately adopts a nanorod morphology.

  To gain a deep insight on the effect of ions concentration, further characterizations were conducted. Firstly, low-temperature N₂ adsorption-desorption isotherms (Fig. 3a) revealed that the specific surface area of the PHI-based catalysts is positively correlated with the total ion concentration. Subsequently, the photoelectrochemical tests of PHI-based samples were examined. The ultraviolet-visible diffuse reflectance spectra (UV–vis DRS, Fig. 3b) shows that PHI samples synthesized with different molten salts exhibit similar absorption edge at around 460 nm. It should be noted that the conduction band position of PHI-LiNaK was determined to be −0.84 eV (Table S2 in Supporting information), based on the Tauc plots (Fig. S5 in Supporting information) and VB-XPS spectra (Fig. 3c), which is significantly more negative than those of the other three samples. The more negative conduction band position suggests that PHI-LiNaK possesses stronger reductive capability.

  Figure 3

  

  Figure 3.  (a) Low-temperature N₂ adsorption-desorption isotherms, (b) UV–vis light absorption curves, (c) VB-XPS spectra, (d) room-temperature steady-state photoluminescence (PL) spectra, (e) time-resolved fluorescence spectra, and (f) transient photocurrent responses of PHI-LiNaK, PHI-LiNa, PHI-LiK and PHI-NaK.

  DownLoad: Full-Size Img PowerPoint

  Furthermore, the carrier separation and recombination behavior of the PHI catalysts were investigated by using steady-state room-temperature fluorescence spectroscopy and time-resolved fluorescence spectroscopy. As shown in Figs. 3d and e, PHI-LiNaK exhibited the lowest fluorescence emission intensity, indicating that the carrier recombination is effectively suppressed [60]. Additionally, it exhibited the shortest fluorescence lifetime, further suggesting that PHI-LiNaK bears superior charge separation efficiency. Moreover, transient photocurrent measurements (Fig. 3f) show that PHI-LiNaK generates the highest photocurrent under illumination, confirming its superior electron-hole separation ability [61]. EPR measurements were also performed to investigate nitrogen defects in different samples (Fig. S6 in Supporting information). PHI-LiNa exhibited the strongest peak intensity, while PHI-LiK, due to its better crystallinity, showed the lowest peak intensity. The C/N atomic ratio (Table S3 in Supporting information) calculated from the XPS results further confirmed the findings from the EPR spectra.

  To further confirm the outstanding charge separation efficiency of PHI-LiNaK, Kelvin probe force microscopy (KPFM) (Figs. 4a-d) was used to detect changes in surface potential before and after illumination. Significant changes in surface potential were observed for both samples upon illumination. After illumination, the surface potential of PHI-LiNa decreased by 33.3 mV, while the surface potential of PHI-LiNaK decreased significantly by 59.7 mV. This phenomenon can be attributed to the rapid migration and accumulation of photogenerated electrons on the catalyst surface upon light excitation, leading to a decrease in potential [62,63]. Pt is typically considered an efficient co-catalyst for hydrogen production, as it facilitates the directional migration of charges and reduces the overpotential, promoting the occurrence of the reduction reaction. However, The distribution of Pt can significantly impact the catalytic efficiency [64,65]. As shown in Figs. 4e and g, it is evident that the Pt species loaded onto the PHI-LiNaK surface are more uniformly distributed. High-resolution Pt 4f XPS analysis (Figs. 4f-h and Table S4 in Supporting information) reveals that the proportion of Pt⁰ in PHI-LiNaK is higher. These are all conditions favorable for photocatalytic hydrogen production.

  Figure 4

  

  Figure 4.  (a) KPFM images of PHI-LiNaK in dark and light. (b) The surface potential of PHI-LiNaK before and after irradiation. (c) KPFM images of PHI-LiNa in dark and light. (d) The surface potential of PHI-LiNa before and after irradiation. (e) TEM image and element mapping of PHI-LiNaK. (f) Pt 4f XPS spectra of PHI-LiNaK. (g) TEM image and element mapping of PHI-LiNa. (h) Pt 4f XPS spectra of PHI-LiNa.

  DownLoad: Full-Size Img PowerPoint

  To evaluate the photocatalytic hydrogen production activity of the synthesized samples, all samples were tested under visible light (λ > 420 nm) by in situ photodeposition of platinum as H₂ evolution cocatalysts in aqueous solution with TEOA as a sacrificial agent (Fig. 5a).

  Figure 5

  

  Figure 5.  (a) Photocatalytic H₂ evolution rates of different molten salts. (b) Photocatalytic H₂ evolution rates of various PHI-Li_xNaK. (c) Stability test of H₂ evolution activities of PHI-LiNaK. (d) Wavelength dependence of H₂ evolution rate of PHI-LiNaK (The distance from the LED lamp head to the liquid surface is 23 cm. LED lamps: λ = 380 nm, 507 µW/cm²; λ = 400 nm, 4093 µW/cm²; λ = 420 nm, 1972 µW/cm²; λ = 450 nm, 2286 µW/cm²). (e) Wavelength-dependent AQY of PHI-LiNaK. (f) H₂O₂ production rates of different molten salts.

  DownLoad: Full-Size Img PowerPoint

  Among all the four PHI samples, PHI-LiNaK exhibited best activity, achieving 253 µmol/h, which is significantly higher than those of other three PHI samples synthesized using binary molten salts. This remarkable performance enhancement is attributed to the nanorod morphology of PHI-LiNaK, which reduces the charge carrier transport distance and improves charge separation efficiency.

  To investigate the influence of ion content on catalytic performance, a series of samples were synthesized by varying the mass fraction of LiCl, and their photocatalytic hydrogen production activities were evaluated (Fig. 5b). The results showed that the maximum photocatalytic activity was achieved when the mass fraction of LiCl was 0.7. Initially, SEM characterizations (Fig. S7 in Supporting information) revealed that as the LiCl content increased, the samples gradually transformed from larger two-dimensional nanosheets to three-dimensional nanorods. Next, as shown in Fig. S8 (Supporting information), XRD and FT-IR characterizations confirmed that the PHI-Li_xNaK products synthesized with different LiCl mass ratios consistently exhibited similar chemical structure. These morphological changes, combined with ICP-OES results, validated above conclusion that an increase in Li⁺ ions content occupies more coordination sites, leading to variations in Na⁺ and K⁺ ions contents and thus affecting the polymerization and growth process of the catalyst. As a multilayer catalyst, Na⁺ and K⁺ ions, embedded both in-plane and between layers, play a crucial role in the transport and separation of photogenerated charge carriers. The Na⁺/K⁺ mass ratio in PHI-LiNaK is 0.258. Both in-plane and interlayer ion increase disrupt the balance of synergistic charge carrier transport within and between the layers. UV–vis DRS analysis indicated a gradual blue shift in the light absorption of the samples, which could be attributed to the decreasing in particle size of the catalyst, decreasing visible light absorption. PHI-Li₁NaK showed a distinct n-π* transition, suggesting that an excess of defects in the sample negatively impacted its charge separation efficiency [66]. Further photoluminescence (PL) spectra (Fig. S9 in Supporting information) revealed that the charge separation efficiency was optimal at a Na⁺/K⁺ mass ratio of 0.258 (corresponding to a LiCl mass fraction of 0.7). Deviations from this optimal ratio, either higher or lower, leading to a decrease in charge separation efficiency. This result is consistent with the observed photocatalytic hydrogen production performance.

  Subsequently, we evaluated the photochemical stability of PHI-LiNaK (Fig. 5c). After 45 h of illumination, there was no significant change in the final photocatalytic hydrogen production activity compared to the initial activity. After the reaction, we performed XRD, FT-IR, and XPS tests on the post-reaction catalyst (Figs. S10 and S11 in Supporting information). From the comparison of the test results, no significant structural changes were observed in PHI-LiNaK before and after the reaction. Therefore, it can be concluded that PHI-LiNaK exhibits good photochemical stability. The relationship between apparent quantum efficiency and optical absorbance across different wavelength regions is illustrated in Figs. 5d and e, showing a strong correlation at wavelengths of 380, 420, and 450 nm. Notably, the apparent quantum efficiency of PHI-LiNaK reaches up to 52.9% at 420 nm.

  The effects of synthesis temperature (Figs. S12-S15 in Supporting information), holding time (Figs. S16-S19 in Supporting information), and the use of different precursors (Figs. S20 and S21 in Supporting information) on the photocatalytic performance were further investigated. The photocatalytic hydrogen production results revealed that the catalyst exhibited the highest hydrogen production activity when PHI was synthesized using melamine as the precursor at a calcination temperature of 550 ℃ and a holding time of 6 h. According to the hydrogen production results (Fig. S22 in Supporting information), a Pt loading of 3 wt% exhibited the optimal performance.

  Additionally, the photocatalytic performance of the samples for hydrogen peroxide synthesis was tested (Fig. 5f). The trend in hydrogen peroxide synthesis performance aligns well with that of photocatalytic hydrogen production. Under λ > 420 nm illumination, PHI-LiNaK exhibited a hydrogen peroxide synthesis rate of 8.86 mmol L⁻¹ h⁻¹, significantly higher than the other three PHI samples. Finally, we conducted TGA-DSC analysis (Fig. S23 in Supporting information) to investigate the polymerization process, with detailed analyses provided in Supporting information.

  Based on the comprehensive characterization results, it can be inferred that PHI-LiNaK exhibits a high catalytic activity mechanism (Fig. 6). Under visible light irradiation, the nanorod-shaped PHI-LiNaK shows shorter charge transfer distances compared to other nanosheet structures, leading to a greater accumulation of charges on the surface over a given period. Additionally, the Pt nanoparticles loaded on the surface are smaller and more uniformly distributed. Previous studies have reported that Pt nanoparticles act as the active sites for the hydrogen evolution reaction. Therefore, the more uniform and well-dispersed loading of Pt nanoparticles will further enhance the photocatalytic hydrogen evolution performance.

  Figure 6

  

  Figure 6.  Proposed photocatalytic hydrogen evolution mechanism of PHI-LiNaK and PHI-LiNa.

  DownLoad: Full-Size Img PowerPoint

  In summary, by optimizing the composition of molten salts, PHI-LiNaK photocatalyst with excellent hydrogen production performance, was successfully synthesized. The presence of Li⁺ occupies certain coordination sites, reduces the incorporation of other alkali metal ions, while changes in ion concentration led to the formation of a nanorod morphology. This nanorod morphology reduced the migration distance of charge carriers, thereby enhancing charge separation efficiency. Furthermore, the deposition of Pt as a co-catalyst for hydrogen production resulted in optimized PHI-LiNaK, which exhibited an excellent apparent quantum efficiency of 52.9% under visible light (λ > 420 nm). This study highlights the importance of rational control the polymerization process of PHI and optimization in the photocatalytic activity, providing new insights into the synthesis of high-performance photo-catalysts for solar hydrogen production.

  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

  Xiao Liu: Writing – original draft, Software, Methodology, Formal analysis, Data curation. Hangqi Liu: Methodology, Formal analysis, Data curation. Qian Wang: Software, Methodology, Investigation, Formal analysis, Data curation. Dandan Zheng: Writing – original draft, Supervision, Formal analysis, Data curation. Sibo Wang: Supervision, Formal analysis, Data curation. Masakazu Anpo: Writing – review & editing, Supervision, Formal analysis, Data curation. Guigang Zhang: Writing – review & editing, Writing – original draft, Supervision, Project administration, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.

  Acknowledgments

  The present work is financially supported by the National Key R & D Program of China (No. 2021YFA1502100), the National Natural Science Foundation of China (Nos. 22472029, 22172029, 22311540011), the Natural Science Foundation of the Fujian Province (No. 2024J010014), and 111 Project (No. D16008).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2025.111621.

  
  
• 1. [1]

     Q. Wang, K. Domen, Chem. Rev. 120 (2019) 919–985. doi: 10.3390/ijms20040919

  2. [2]

     Q. Wang, X.C. Wang, F.X. Zhang, Next Energy 1 (2023) 100016. doi: 10.1016/j.nxener.2023.100016

  3. [3]

     Y.O. Wang, H. Suzuki, J.J. Xie, et al., Chem. Rev. 118 (2018) 5201–5241. doi: 10.1021/acs.chemrev.7b00286

  4. [4]

     F. Dawood, M. Anda, G.M. Shafiullah, Int. J. Hydrogen Energy 45 (2020) 3847–3869. doi: 10.1016/j.ijhydene.2019.12.059

  5. [5]

     B. Zhang, W.J. Luo, L.Y. Pan, et al., Inorg. Chem. Front. 12 (2025) 161–170. doi: 10.1039/d4qi02452d

  6. [6]

     A. Odenweller, F. Ueckerdt, G.F. Nemet, M. Jensterle, G. Luderer, Nat. Energy 7 (2022) 854–865. doi: 10.1038/s41560-022-01097-4

  7. [7]

     Y.H. Zhang, Z.C. Yang, D.D. Zheng, et al., Int. J. Hydrogen Energy 69 (2024) 372–380. doi: 10.1016/j.ijhydene.2024.05.056

  8. [8]

     H. Nishiyama, T. Yamada, M. Nakabayashi, et al., Nature 598 (2021) 304–307. doi: 10.1038/s41586-021-03907-3

  9. [9]

     X.P. Tao, Y. Zhao, S.Y. Wang, C. Li, R.G. Li, Chem. Soc. Rev. 51 (2022) 3561–3608. doi: 10.1039/d1cs01182k

  10. [10]

      J.F. Du, Y.L. Shen, F. Yang, et al., Inorg. Chem. Front. 10 (2023) 832–840. doi: 10.1039/d2qi01665f

  11. [11]

      G.R. Zou, Q. Wang, G. Ye, et al., Adv. Funct. Mater. (2024), https://doi.org/10.1002/adfm.202420899. doi: 10.1002/adfm.202420899

  12. [12]

      A. Fujishima, K. Honda, Nature 238 (1972) 37–38. doi: 10.1038/238037a0

  13. [13]

      E.N. Musa, A.K. Yadav, K.T. Smith, et al., Angew. Chem. Int. Ed. 63 (2024) e202405681.

  14. [14]

      F. Guo, M. Yang, R.X. Li, et al., ACS Catal. 12 (2022) 9486–9493. doi: 10.1021/acscatal.2c02789

  15. [15]

      K. Sun, Y. Huang, F.S. Sun, et al., Nat. Chem. 16 (2024) 1638–1646. doi: 10.1038/s41557-024-01599-6

  16. [16]

      W.B. Yu, H.R. Zhang, H.W. Zhang, et al., Adv. Funct. Mater. 34 (2024) 2410816. doi: 10.1002/adfm.202410816

  17. [17]

      Y. Miseki, H. Kusama, H. Sugihara, K. Sayama, J. Phys. Chem. Lett. 1 (2010) 1196–1200. doi: 10.1021/jz100233w

  18. [18]

      R.T. Chen, Z.F. Ren, Y. Liang, et al., Nature 610 (2022) 296–301. doi: 10.1038/s41586-022-05183-1

  19. [19]

      W. Li, X.S. Chu, F. Wang, et al., Appl. Catal. B 304 (2022) 121000. doi: 10.1016/j.apcatb.2021.121000

  20. [20]

      X. Wang, B.Y. Liu, S.Q. Ma, et al., Nat. Commun. 15 (2024) 2600. doi: 10.1038/s41467-024-47022-z

  21. [21]

      H. Wang, H. Wang, Z.W. Wang, et al., Chem. Soc. Rev. 49 (2020) 4135–4165. doi: 10.1039/d0cs00278j

  22. [22]

      Z.A. Lan, M. Wu, Z.P. Fang, et al., Angew. Chem. Int. Ed. 61 (2022) e202201482. doi: 10.1002/anie.202201482

  23. [23]

      F.K. Kessler, Y. Zheng, D. Schwarz, et al., Nat. Rev. Mater. 2 (2017) 17030. doi: 10.1038/natrevmats.2017.30

  24. [24]

      X.C. Wang, K. Maeda, A. Thomas, et al., Nat. Mater. 8 (2009) 76–80. doi: 10.1038/nmat2317

  25. [25]

      M.L. Chang, Z.M. Pan, D.D. Zheng, et al., ChemSusChem 16 (2023) e202202255. doi: 10.1002/cssc.202202255

  26. [26]

      D.D. Zheng, Q. Wang, Z.M. Pan, et al., Sci. China Mater. 67 (2024) 1900–1906. doi: 10.1007/s40843-023-2752-9

  27. [27]

      H.J. Song, L.Y. Luo, S.Y. Wang, G. Zhang, B.J. Jiang, Chin. Chem. Lett. 35 (2024) 109347. doi: 10.1016/j.cclet.2023.109347

  28. [28]

      X.J. Chen, R. Shi, Q. Chen, et al., Nano Energy 59 (2019) 644–650. doi: 10.1016/j.nanoen.2019.03.010

  29. [29]

      J. Zhang, X.C. Liang, C. Zhang, et al., Angew. Chem. Int. Ed. 61 (2022) e202210849. doi: 10.1002/anie.202210849

  30. [30]

      J. Zhang, G. Ye, C. Zhang, et al., ChemSusChem 15 (2022) e202201616. doi: 10.1002/cssc.202201616

  31. [31]

      H.T. Wang, L.L. Yu, J.Z. Jiang, Arramel, J. Zou, Acta Phys. Chim. Sin. 40 (2024) 2305047. doi: 10.3866/PKU.WHXB202305047

  32. [32]

      W.G. Zeng, Y.C. Dong, X.Y. Ye, et al., Chin. Chem. Lett. 35 (2024) 109252. doi: 10.1016/j.cclet.2023.109252

  33. [33]

      D.D. Wu, S.N. Hu, H.Y. Xue, et al., J. Mater. Chem. A 7 (2019) 20223–20228. doi: 10.1039/c9ta05135j

  34. [34]

      N.P. Dharmarajan, D. Vidyasagar, J.H. Yang, et al., Adv. Mater. 36 (2023) 2306895.

  35. [35]

      Q. Wang, S.Y. Li, D.D. Zheng, et al., ACS Appl. Energy Mater. 7 (2024) 6090–6095. doi: 10.1021/acsaem.4c01512

  36. [36]

      B.Y. Zhai, H.G. Li, G.Y. Gao, et al., Adv. Funct. Mater. 32 (2022) 2207375. doi: 10.1002/adfm.202207375

  37. [37]

      G.G. Zhang, L.H. Lin, G.S. Li, et al., Angew. Chem. Int. Ed. 57 (2018) 9372–9376. doi: 10.1002/anie.201804702

  38. [38]

      Y.X. Jin, D.D. Zheng, Z.P. Fang, et al., Interdiscip. Mater. 3 (2024) 389–399. doi: 10.1002/idm2.12159

  39. [39]

      Y.M. Zou, S.Y. Li, D.D. Zheng, et al., Sci. China Chem. 67 (2024) 2215–2223. doi: 10.1007/s11426-024-2076-3

  40. [40]

      Z.P. Chen, A. Savateev, S. Pronkin, et al., Adv. Mater. 29 (2017) 1700555. doi: 10.1002/adma.201700555

  41. [41]

      M.J. Bojdys, J.O. Müller, M. Antonietti, A. Thomas, Chem. Eur. J. 14 (2008) 8177–8182. doi: 10.1002/chem.200800190

  42. [42]

      M. Zhou, L.B. Zeng, R. Li, et al., Appl. Catal. B 317 (2022) 121719. doi: 10.1016/j.apcatb.2022.121719

  43. [43]

      C.M. Pelicano, J.X. Li, M. Cabrero-Antonino, et al., J. Mater. Chem. A 12 (2024) 475–482. doi: 10.1039/d3ta05701a

  44. [44]

      Y.X. Jin, C. Zhang, G.G. Zhang, Chin. J. Struct. Chem. 43 (2024) 100414.

  45. [45]

      W. Zhong, D. Zheng, Y.X. Ou, A.Y. Meng, Y.R. Su, Acta Phys. Chim. Sin. 40 (2024) 2406005. doi: 10.3866/pku.whxb202406005

  46. [46]

      Y. Pan, W.P. Si, Y.H. Li, et al., Chin. Chem. Lett. 35 (2024) 109877. doi: 10.1016/j.cclet.2024.109877

  47. [47]

      S.P. Liu, V. Diez-Cabanes, D. Fan, et al., ACS Catal. 14 (2024) 2562–2571. doi: 10.1021/acscatal.3c05961

  48. [48]

      H.L. Gao, S.C. Yan, J.J. Wang, et al., Phys. Chem. Chem. Phys. 15 (2013) 18077. doi: 10.1039/c3cp53774a

  49. [49]

      Q.H. Deng, H.P. Li, W.X. Hu, W.G. Hou, Angew. Chem. Int. Ed. 62 (2023) e202314213. doi: 10.1002/anie.202314213

  50. [50]

      J. Kröger, F. Podjaski, G. Savasci, et al., Adv. Mater. 34 (2022) 2107061. doi: 10.1002/adma.202107061

  51. [51]

      Q. Wang, G.G. Zhang, W.D. Xing, et al., Angew. Chem. Int. Ed. 62 (2023) e202307930. doi: 10.1002/anie.202307930

  52. [52]

      H. Schlomberg, J. Kröger, G. Savasci, et al., Chem. Mater. 31 (2019) 7478–7486. doi: 10.1021/acs.chemmater.9b02199

  53. [53]

      A. Thomas, A. Fischer, F. Goettmann, et al., J. Mater. Chem. 18 (2008) 4893. doi: 10.1039/b800274f

  54. [54]

      Z.L. Zhao, Z. Shu, J. Zhou, et al., J. Mater. Chem. A 10 (2022) 17668–17679. doi: 10.1039/d2ta04213d

  55. [55]

      X.C. Li, X.X. Chen, Y.X. Fang, et al., Chem. Sci. 13 (2022) 7541–7551. doi: 10.1039/d2sc02043b

  56. [56]

      K. Schwinghammer, M.B. Mesch, V. Duppel, et al., J. Am. Chem. Soc. 136 (2014) 1730–1733. doi: 10.1021/ja411321s

  57. [57]

      D. Dontsova, S. Pronkin, M. Wehle, et al., Chem. Mater. 27 (2015) 5170–5179. doi: 10.1021/acs.chemmater.5b00812

  58. [58]

      J.M. Wu, K.Y. Li, B. Zhou, et al., Angew. Chem. Int. Ed. (2024), https://doi.org/10.1002/anie.202421263. doi: 10.1002/anie.202421263

  59. [59]

      T. Xiong, W.L. Cen, Y.X. Zhang, F. Dong, ACS Catal. 6 (2016) 2462–2472. doi: 10.1021/acscatal.5b02922

  60. [60]

      G.G. Zhang, A. Savateev, Y.B. Zhao, L.N. Li, M. Antonietti, J. Mater. Chem. A 5 (2017) 12723–12728. doi: 10.1039/C7TA03777E

  61. [61]

      T. T. Zhao, Q. Zhou, Y.Q. Lv, et al., Angew. Chem. Int. Ed. 59 (2020) 1139–1143. doi: 10.1002/anie.201911822

  62. [62]

      P.L. Wang, S.Y. Fan, X.Y. Li, et al., Nano Energy 95 (2022) 107045. doi: 10.1016/j.nanoen.2022.107045

  63. [63]

      A. Talebian-Kiakalaieh, M.J. Guo, E.M. Hashem, et al., Adv. Energy Mater. 13 (2023) 2301594. doi: 10.1002/aenm.202301594

  64. [64]

      S.W. Cao, F. Tao, Y. Tang, Y.T. Li, J.G. Yu, Chem. Soc. Rev. 45 (2016) 4747–4765. doi: 10.1039/C6CS00094K

  65. [65]

      T. Shmila, S. Mondal, S. Barzilai, et al., Small 19 (2023) 2303602. doi: 10.1002/smll.202303602

  66. [66]

      G.G. Zhang, G.S. Li, Z.A. Lan, et al., Angew. Chem. Int. Ed. 56 (2017) 13445–13449. doi: 10.1002/anie.201706870

• 

  Scheme 1  A proposed process for the synthesis of PHI with different morphologies in the presence of different types of molten salts.

  下载: 全尺寸图片 幻灯片
  

  Figure 1  (a-d) SEM image of PHI-LiNa, PHI-LiK, PHI-NaK and PHI-LiNaK. (e, f) HR-TEM image of PHI-LiNaK. Inset: Enlargement of the selected area. (g) Powder XRD patterns, (h) FT-IR spectra of PHI-LiNaK, PHI-LiNa, PHI-LiK and PHI-NaK. (i) Solid-state ¹³C NMR spectrum of PHI-LiNaK.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  XPS results of high-resolution spectra of C 1s (a), N 1s (b), Na 1s (c), K 2p (d), Li 1s (e) and Cl 2p (f) of PHI-LiNaK, PHI-LiNa, PHI-LiK and PHI-NaK.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) Low-temperature N₂ adsorption-desorption isotherms, (b) UV–vis light absorption curves, (c) VB-XPS spectra, (d) room-temperature steady-state photoluminescence (PL) spectra, (e) time-resolved fluorescence spectra, and (f) transient photocurrent responses of PHI-LiNaK, PHI-LiNa, PHI-LiK and PHI-NaK.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  (a) KPFM images of PHI-LiNaK in dark and light. (b) The surface potential of PHI-LiNaK before and after irradiation. (c) KPFM images of PHI-LiNa in dark and light. (d) The surface potential of PHI-LiNa before and after irradiation. (e) TEM image and element mapping of PHI-LiNaK. (f) Pt 4f XPS spectra of PHI-LiNaK. (g) TEM image and element mapping of PHI-LiNa. (h) Pt 4f XPS spectra of PHI-LiNa.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  (a) Photocatalytic H₂ evolution rates of different molten salts. (b) Photocatalytic H₂ evolution rates of various PHI-Li_xNaK. (c) Stability test of H₂ evolution activities of PHI-LiNaK. (d) Wavelength dependence of H₂ evolution rate of PHI-LiNaK (The distance from the LED lamp head to the liquid surface is 23 cm. LED lamps: λ = 380 nm, 507 µW/cm²; λ = 400 nm, 4093 µW/cm²; λ = 420 nm, 1972 µW/cm²; λ = 450 nm, 2286 µW/cm²). (e) Wavelength-dependent AQY of PHI-LiNaK. (f) H₂O₂ production rates of different molten salts.

  下载: 全尺寸图片 幻灯片
  

  Figure 6  Proposed photocatalytic hydrogen evolution mechanism of PHI-LiNaK and PHI-LiNa.

  下载: 全尺寸图片 幻灯片
•