Fabrication of a high-efficiency hydrogen generation Pd/C₃N₅-K,I photocatalyst through synergistic effects of planar and spatial carrier separation

Abstract

Nitrogen-rich graphitic carbon nitride (g-C₃N₅) has garnered significant attention in photocatalytic hydrogen production due to its unique physicochemical properties. However, g-C₃N₅ faces persistent challenges stemming from rapid carrier recombination. In this study, we successfully synthesized a Pd/C₃N₅-K,I photocatalyst through sequential doping of non-metallic heteroatoms and metal nanoparticles. The Pd/C₃N₅-K,I photocatalyst demonstrates a remarkable enhancement in the photocatalytic H₂ evolution rate (2.9 mmol g⁻¹ h⁻¹), approximately 14 times higher than that of pure g-C₃N₅ (0.2 mmol g⁻¹ h⁻¹). Characterization and calculation analyses reveal that iodine doping into the g-C₃N₅ skeleton leads to the formation of planar C–I bonds, facilitating C₃N₅ layer electron–hole pair separation. The subsequent insertion of Pd nanoparticles between the layers leads to electron accumulation on C₃N₅-K,I, while resulting in hole concentration on Pd nanoparticles, thereby facilitating thorough spatial separation of electron–hole pairs. The strategic co-doping of iodine and palladium nanoparticles effectively restrains carrier recombination of g-C₃N₅ by connecting intra- and inter-layer interactions. This study constitutes a novel conceptual framework for CN-based photocatalysts, offering a promising approach to improve their performance in hydrogen production applications.

---

1. Introduction

Hydrogen energy has emerged as a leading alternative to increasingly depleted fossil fuels.^1–3 Since the discovery by Fujishima and Honda in 1972 that TiO₂ can decompose water to produce hydrogen (H₂), extensive research efforts have been dedicated to exploring various photocatalysts, including TiO₂, metal-oxo clusters, and carbon nitride.^4–7 Among these photocatalysts, graphite-phase carbon nitride (g-C₃N₄) has attracted widespread attention due to its remarkable economic efficiency, excellent light stability, and tunable bandgap characteristics.⁸ Wang et al. initially demonstrated the photocatalytic performance of g-C₃N₄ in producing H₂ under visible-light irradiation,⁹ and since then various modifications have been made to enhance its performance. For instance, Jin et al. developed a g-C₃N₄/Ni_xMo_1−xS₂ photocatalyst with anchored Ni–Mo–S nanoparticles, which improved the efficiency of photocatalytic H₂ evolution by regulating the charge separation process.¹⁰ Lai et al. modified g-C₃N₄ by S-doping to improve its light absorption ability and carrier separation efficiency.¹¹ Yang et al. achieved an effective improvement in carrier separation efficiency by forming a Schottky barrier through close contact between Cu nanoparticles and g-C₃N₄.¹² Notably, nitrogen-rich g-C₃N₅, with a narrower bandgap of 2.2 eV, was successfully synthesized by Vinu et al. using 3-amino-1,2,4-triazole as a precursor.¹³ It has shown enhanced vis-light response compared to g-C₃N₄, being activatable under ultraviolet, visible, and near-infrared irradiation.^14–16 The recently synthesized PtCu–C₃N₅,¹⁷ Pt/C₃N₅¹⁸ and CoOOH·CoOx/P-C₃N₅¹⁹ have exhibited remarkable H₂ production efficiency, highlighting the significant potential of g-C₃N₅ in photocatalytic applications. However, despite its potential, the inherent low electrical conductivity and suboptimal electronic structure of g-C₃N₅ impede the transfer of photoinduced charge carriers, limiting its photocatalytic performance.^20–22 Therefore, constructing highly efficient g-C₃N₅-based catalysts to achieve better charge separation efficiency has emerged as a critical focus in the field of photocatalysis.

To address the bottleneck in g-C₃N₅ photocatalysis, doping non-metallic heteroatoms with different electronegativities (e.g., boron, iodine, bromine, and sulfur) has been demonstrated as a vital approach to modulate the electronic states and chemical properties of g-C₃N₅. These dopants can generate trapping sites for electrons or holes, inhibiting the recombination process.^23,24 The charge carrier distribution in heteroatom-doped g-C₃N₅ can generate an electric field and induce electron rearrangement, facilitating charge migration on the CN layers.^25,26 However, while heteroatom doping benefits in-plane charge transfer, it has limited impact on charge transfer between adjacent layers.²⁷ To significantly improve spatial charge transfer and exciton dissociation, incorporating appropriate metal nanoparticles as bridging sites between adjacent CN layers is crucial.^28,29 Metal nanoparticles doped on CN supports enable efficient modification of the electronic structure and form directional charge-transfer channels.^30,31 Extensive research has shown that introducing metal nanoparticles can establish a build-in electric field towards the metal, leading to efficient spatial light-induced charge separation.^32–34 Although these strategies have been explored individually, few researchers have examined the synergistic effect of heteroatoms and metal nanoparticles on planar and spatial charge separation. Resolving both interlayer and in-plane charge transfer in g-C₃N₅ to achieve superior hydrogen production performance in photocatalysis is the primary motivation of this work.

Here, we designed a modified vis-light-catalyst by loading Pd nanoparticles onto g-C₃N₅ doped with potassium and iodine. This Pd/C₃N₅-K,I photocatalyst exhibits excellent hydrogen production properties even with fewer cocatalysts. Characterization results propose a rational chemical structure for the Pd/C₃N₅-K,I photocatalyst, where stable C–I bonds form in the g-C₃N₅ skeleton, and Pd nanoparticles are intercalated into the interlayer of C₃N₅-K,I. By analyzing the calculated electron–hole distribution, we innovatively proposed a synergistic mechanism for distinct charge transfer pathways within Pd/C₃N₅-K,I, thereby studying the photocatalytic hydrogen evolution reaction (HER) in depth. A small amount of I doped in g-C₃N₅ significantly improves planar carrier separation. When Pd nanoparticles are further introduced into the space among adjacent layers of C₃N₅-K,I, significant interlayer electron–hole pair separation is achieved. This mechanism indicates a significant aggregation of electrons on the C₃N₅-K,I layers, with the interlayer palladium nanoparticles exhibiting a hole-rich state. The synergy between planar and spatial carrier separation remarkably boosts charge carrier dynamics, notably enhancing HER activity.

2. Experimental

2.1. Materials

Triethanolamine (TEOA, ≥99.0%), H₂PtCl₆·6H₂O and 3-amino-1,2,4-triazole (3-AT) were supplied by Aladdin. PdCl₂ (AR) was purchased from Changchun Third Party Pharmaceutical Technology Co., Ltd. KI (AR) was obtained from Shanghai Yindian Chemical Co., Ltd. All chemicals were in accordance with the requirement of analytical reagents.

2.2. Catalyst preparation

Bulk g-C₃N₅ was synthesized according to the literature.³⁵ 3 g of the 3-AT powder was calcined to 550 °C by 5 °C min⁻¹ for 3 h and then cooled for 12 h in a muffle furnace.

C₃N₅-K,I was synthesized using a thermal polycondensation synthesis method. In detail, 3-AT (200 mg) and KI (10, 60, 100 and 200 mg, respectively) were dispersed in deionized water (20 mL) under magnetic stirring. Next, the suspension was dried at a temperature of 60 °C for a duration of 12 h under vacuum conditions. Following this step, the sample was heated gradually to reach a temperature of 550 °C with an increment rate of 5 °C min⁻¹ and maintained for 3 h to successfully synthesize C₃N₅-K,I.

As illustrated in Scheme 1, the Pd/C₃N₅-K,I catalyst was synthesized using low-temperature solvothermal treatment. In detail, C₃N₅-K,I (80 mg) was added to a solution of ethylene glycol (13 mL), followed by the addition of PdCl₂ (3 mg mL⁻¹, 0.7, 0.9 and 1.10 mL, respectively) into the above solution with continuous stirring. The obtained suspension was subsequently heated to 130 °C at a heating rate of 2 °C min⁻¹, and maintained for another 4 h. The solution was thoroughly washed multiple times with absolute ethanol and deionized water after cooling down. Finally, the Pd/C₃N₅-K,I catalyst was successfully acquired after a 12 h drying process at 60 °C in a vacuum oven. As a control, Pd/C₃N₅ was synthesis adopting a similar method to the Pd/C₃N₅-K,I catalyst. C₃N₅-K,I was substituted with g-C₃N₅, and the other experimental procedures are identical to the synthesis of Pd/C₃N₅-K,I.

Scheme 1 The synthesis process of Pd/C₃N₅-K,I. Gray, blue, white, pink, purple and dark green represent C, N, H, I, K and Pd.

2.3. Catalyst characterization

The morphologies and structures were observed using a scanning electron microscope (SEM TESCAN MIRA4) and a high-resolution transmission electron microscope (HRTEM, JEOL-JEM 2100). Thickness analysis of the prepared samples was conducted employing an atomic force microscope (AFM Multimode8). The Brunauer–Emmett–Teller (BET) surface areas and pore volumes were determined using a Micrometrics ASAP 2460 automatic surface area and porosity analyser. The X-ray diffraction (XRD) patterns were analysed using a D8 ADVANCE A25 diffractometer. Fourier transform infrared (FT-IR) spectra ranging from 400 to 4000 cm⁻¹ were acquired using an IR Nicolet IS 50 infrared spectrometer. The chemical state of the photocatalyst was investigated using X-ray photoelectron spectroscopy (XPS Thermo Scientific ESCALAB 250Xi). Electronic properties were measured at room temperature using a fluorescence spectrometer (PL F-2700). Time-resolved photoluminescence decay spectra (TRPD, λ_ex = 405 nm, λ_em = 460 nm) were acquired using an FL920 transient fluorescence spectrophotometer. Electrochemical measurements were performed using the CHI-660E electrochemical workstation, employing 0.1 M Na₂SO₄ solution as the electrolyte. A dispersion of 5 mg of the photocatalyst in 2 mL of ethyl alcohol was sonicated, and then 40 μL of the resulting solution was dropwise applied onto an FTO substrate (conductive fluorine-doped tin oxide). The substrate served as the working electrode after natural drying. A saturated calomel electrode and platinum sheet were utilized as the reference and counter electrodes, respectively. Electrochemical impedance spectroscopy (EIS) measurements were performed over the frequency range of 0.01–10⁵ Hz. UV-vis absorbance spectra for the samples were measured using a Hitachi UH4150 spectrophotometer. The content of metal Pd was detected by inductively coupled plasma mass spectrometry (ICP-MS, NexION 350X).

2.4. Photocatalytic hydrogen evolution experiments

A closed gas circulation system (CEL-PAEM-D6) was employed for the analysis of photocatalytic H₂-evolution results. In this experiment, 50 mg of the photocatalyst loaded with H₂PtCl₆ (45 μL,10 g L⁻¹) were added to a 100 mL mixed solution (10 mL TEOA, 90 mL DI water). The reaction mixture was then transferred to a quartz reactor and maintained at a temperature of 6 °C using a condensate pump while being stirred magnetically. For the light source, we utilized an Xe lamp (300 W, Beijing Perfectlight, PLS–SEX300DUV), which was positioned approximately 15 cm away from the liquid surface to provide visible light irradiation. Throughout the course of the reaction lasting for 6 h and employing N₂ as the carrier gas, samples were collected every hour and analysed using gas chromatography (GC-7920) to determine the quantity of produced H₂. Additionally, we evaluated the photocatalytic stability over four consecutive reaction cycles. The apparent quantum yield (AQY) was assessed using an Xe lamp equipped with filters at 400, 420, 450 and 500 nm, while the other procedures remained identical to the H₂ evolution conditions. The calculations were performed in accordance with eqn (1),⁶ where n is the amount of H₂ production, NA is the Avogadro constant, P is the light intensity, S is the irradiation area, t is the illumination time, λ is the irradiation wavelength, h is the Planck constant, and c is the speed of light.

2.5. Theoretical calculations

The geometry optimization and excited state were performed using the Gaussian 16 (Revision A03) software package at the B3LYP-D3(BJ)/Pd/Stuttgart/C,N,H,K,I/6-311G* level. There were no imaginary frequencies in the optimized structures. The implicit solvation model (SMD) was employed to describe the solvation effect, with water as the solvent. And the electron–hole analysis was carried out using Multiwfn software.

3. Results and discussion

3.1. Morphology and structure analyses

Fig. 1a–c display SEM images illustrating the microstructures and morphologies of g-C₃N₅, C₃N₅-K,I and Pd/C₃N₅-K,I. Unlike the smooth surface of g-C₃N₅ (Fig. 1a), C₃N₅-K,I shows a flat surface with uniform pore distributions (Fig. 1b), likely due to I doping into the g-C₃N₅ skeleton. This structural change is further supported by BET analysis, which indicates that C₃N₅-K,I has an increased specific area and pore volume compared with g-C₃N₅ (Table S1, ESI†). After Pd metalation (Fig. 1c), distinct stratification is observed, attributed to the insertion of Pd into the C₃N₅-K,I layers. Elemental mapping (Fig. 1d) confirms the existence of all expected elements (C, N, K, I, and Pd) in the Pd/C₃N₅-K,I sample, demonstrating the uniform distribution of Pd and I species throughout the CN structure.

Fig. 1 (a)–(c) SEM images of g-C₃N₅, C₃N₅-K,I, and Pd/C₃N₅-K,I. (d) SEM elemental maps of C, N, K, I, and Pd in Pd/C₃N₅-K,I. (e)–(g) HRTEM images of Pd/C₃N₅-K,I. (h) AFM images of g-C₃N₅, C₃N₅-K,I, and Pd/C₃N₅-K,I.

As shown in Fig. 1e, the HRTEM image indicates that Pd/C₃N₅-K,I has a multilayer staggered stack structure. Additionally, small Pd nanoparticles, with an average size of 5.0 nm, are uniformly attached to C₃N₅-K,I (Fig. 1f). A measured crystal lattice spacing of 0.224 nm is assigned to the (111) lattice plane of Pd (Fig. 1g and Fig. S1a and b, ESI†).³⁶ These results demonstrate that uniform-sized Pd nanoparticles are homogeneously anchored on the unique hierarchical structures of C₃N₅-K,I.

The AFM topographical views of g-C₃N₅, C₃N₅-K,I and Pd/C₃N₅-K,I were analyzed, and the sample thickness was calculated using Gwyddion, open-source software. The average thicknesses of g-C₃N₅, C₃N₅-K,I and Pd/C₃N₅-K,I are approximately 1.5, 1.6, and 5.3 nm, respectively (Fig. 1h and Fig. S2, ESI†). Combined with TEM results, AFM observations reveal that Pd nanoparticles are successfully intercalated with C₃N₅-K,I monolayers, resulting in the formation of a multilayer staggered stack structure.

To elucidate the chemical structures, FT-IR spectra of the samples were recorded (Fig. 2a). Characteristic peaks at 810 and 891 cm⁻¹ are associated with the vibration of tris-s-triazine units.³⁷ The absorption band around 1200–1700 cm⁻¹ is attributed to the stretching modes of C–N and C=N heterocycles.³⁸ A broad band at 3000–3300 cm⁻¹ is observed due to N–H stretching vibrations of uncondensed amino groups or adsorbed water molecules.³⁹ These spectra confirm that C₃N₅-K,I and Pd/C₃N₅-K,I retain the fundamental molecular structure of g-C₃N₅.

Fig. 2 (a) FT-IR spectra of g-C₃N₅, C₃N₅-K,I, and Pd/C₃N₅-K,I. (b) XRD patterns of g-C₃N₅, C₃N₅-K,I, and Pd/C₃N₅-K,I. (c) XPS survey spectra of g-C₃N₅ and Pd/C₃N₅-K,I. (d) The XPS spectra of C 1s regions of g-C₃N₅ and Pd/C₃N₅-K,I. (e) The XPS spectra of N 1s regions of g-C₃N₅ and Pd/C₃N₅-K,I. (f) The XPS spectra of Pd 3d regions of Pd/C₃N₅-K,I.

XRD patterns were employed to investigate the crystal structures (Fig. 2b). All materials exhibit two broad peaks at 13° and 27.3°, corresponding to the (100) and (002) planes of g-C₃N₅, respectively.⁴⁰ These peaks are attributed to the interlayer stacking of triazine and conjugated aromatic units in graphitic materials,⁴¹ corroborating the infrared spectroscopy results. Notably, Pd/C₃N₅-K,I displays four additional characteristic peaks at 40.1, 46.4, 68.2, and 82°, corresponding to the Pd (111), Pd (200), Pd (220), and Pd (311) planes, respectively. This indicates that the aggregated Pd atoms are in a metallic state.^42,43 These findings demonstrate the successful incorporation of Pd nanoparticles into the C₃N₅-K,I layers.

To investigate the chemical valence states and elemental compositions, XPS analyses of Pd/C₃N₅-K,I and g-C₃N₅ were conducted. The binding energies were calibrated by referencing the C 1s peak at 284.8 eV. The presence of C, N, K, I, and Pd in the Pd/C₃N₅-K,I sample corroborates the EDS elemental analysis results (Fig. 2c). The high-resolution C 1s spectrum (Fig. 2d) of g-C₃N₅ is deconvoluted into three peaks at 284.8, 286.27, and 288.33 eV, corresponding to C–C bonds,⁴⁴ the C=N bond,⁴⁵ and sp²-hybridized C bonded to nitrogen atoms within triazine rings,⁴⁶ respectively. In contrast, the C 1s spectrum of Pd/C₃N₅-K,I exhibits four individual peaks at 284.8, 285.97, 286.65, and 288.35 eV. The additional peak at 285.97 eV is ascribed to the formation of C–I bonds within the Pd/C₃N₅-K,I structure. Previous research has shown that specific C atoms in g-C₃N₅ with low electron density tend to react with negatively charged I atoms to form C–I covalent bonds.⁴⁷ This is further corroborated by the I 3d XPS spectrum (Fig. S3a, ESI†), where lower C–I peaks appear at 629.95 (I 3d_3/2) and 618.51 eV (I 3d_5/2) compared with those in iodine salts.^48,49 As illustrated in Fig. 2e, the N 1s spectra of both g-C₃N₅ (398.83/400.17/401.41 eV) and Pd/C₃N₅-K,I (398.85/400.30/401.30 eV) can be attributed to C–N=C, N–(C)₃, and C–NH_(x) groups, respectively.⁵⁰ Notably, upon loading Pd nanoparticles onto C₃N₅-K,I, the C and N atoms shift to higher binding energies, which is favorable for electron transfer in Pd/C₃N₅-K,I.⁵¹ The Pd 3d orbital of Pd/C₃N₅-K,I was fitted with two components (Fig. 2f). The peaks at 334.67 (Pd 3d_5/2) and 340.25 eV (Pd 3d_3/2) are assigned to Pd⁰.^52,53 These peaks are accompanied by higher energy satellite peaks at 336.15 and 341.60 eV, respectively. In summary, the XPS results provide strong evidence for the presence of C–I bonds and Pd nanoparticles in the Pd/C₃N₅-K,I structure.

3.2. Optical and photoelectrochemical properties

Fig. 3a presents PL emission spectra of the samples, which are crucial for understanding carrier trapping and recombination processes, as well as elucidating the photocatalytic mechanism. All samples exhibit a characteristic emission peak at near 460 nm, originating from the intrinsic band gap of g-C₃N₅. Notably, the PL intensity of C₃N₅-K,I is quenched compared to that of g-C₃N₅, suggesting that C–I bond formation suppresses charge recombination. Pd/C₃N₅-K,I displays the lowest PL intensity, attributed to the strong interaction between Pd nanoparticles and C₃N₅-K,I, which enhances charge carrier separation.⁵⁴

Fig. 3 (a) PL spectra of samples. (b) TRPL spectra of samples. (c) Photocurrent transient responses of samples. (d) EIS spectra of samples. (e) UV-vis DRS spectra of samples. (f) Band gaps of samples.

TRPD spectra (Fig. 3b) corroborate the PL results, showing an attenuation trend similar to that of the PL spectra. The average fluorescence lifetime (τ_av) was fitted using a triexponential decay function. The τ_av values for g-C₃N₅, C₃N₅-K,I, and Pd/C₃N₅-K,I are 1.13, 1.11, and 1.09 ns, respectively. The shortest PL lifetime decay of Pd/C₃N₅-K,I indicates further enhanced charge carrier separation.

To investigate the charge transfer mechanism in depth, transient photocurrent response experiments were conducted over three on/off cycles under vis-light irradiation. g-C₃N₅ exhibits very low photocurrent densities, indicating a limited electron migration capacity (Fig. 3c). In contrast, Pd/C₃N₅-K,I demonstrates a significantly higher photocurrent response intensity, suggesting more efficient electron–hole pair separation. This enhancement can be ascribed to the rapid electron transfer from Pd nanoparticles to the g-C₃N₅ surface, establishing a Schottky contact.⁵⁵

EIS was employed to further elucidate charge carrier recombination processes. The radius of the semicircle in the Nyquist plot reflects the interfacial charge transfer resistance, which is related to carrier separation and transfer.⁵⁶Fig. 3d shows that Pd/C₃N₅-K,I exhibits the smallest arc radius compared to g-C₃N₅ and C₃N₅-K,I, indicating the lowest resistance among these samples. This suggests that Pd/C₃N₅-K,I possesses superior electroconductivity, with Pd nanoparticles facilitating electron conduction and enhancing electron–hole pair separation efficiency.

The optical absorption characteristics were analyzed using UV-vis DRS, as shown in Fig. 3e. All samples demonstrated vis-light responsiveness, confirming that they are vis-light-driven photocatalysts. The significant red-shift in absorption for C₃N₅-K,I and Pd/C₃N₅-K,I was advantageous for improving photocatalytic reactions.⁵⁷ Notably, Pd/C₃N₅-K,I exhibited the highest vis-light absorption, attributed to the uniform distribution of Pd nanoparticles on the C₃N₅-K,I surface. The band gap energies of the samples were determined by applying the Kubelka–Munk equation to the UV-vis DRS spectra (Fig. 3f).⁵⁸ The band gap (E_g) values of C₃N₅-K,I (1.66 eV) and Pd/C₃N₅-K,I (0.90 eV) were narrower compared to that of pure g-C₃N₅ (1.69 eV), indicating that I and Pd nanoparticles can modulate the electronic structure of g-C₃N₅.⁵⁹

The band diagram of the semiconductors was constructed using Mott–Schottky plots and VB-XPS measurements, which are commonly employed to determine the energy levels of the conduction band (CB) and valence band (VB).⁶⁰ As shown in Fig. 4a–c, all photocatalysts are n-type semiconductors, as indicated by their positive slopes.⁶¹ The CB levels of g-C₃N₅, C₃N₅-K,I and Pd/C₃N₅-K,I are determined to be −0.44, −0.53, and −0.55 V, respectively. The VB-XPS measurements in Fig. 4d confirm the VB values of g-C₃N₅ (1.15 V), C₃N₅-K,I (1.13 V), and Pd/C₃N₅-K,I (0.35 V). Consequently, the energy level structures of the photocatalysts can be derived, as illustrated in Fig. 4e. Notably, the CB potential of Pd/C₃N₅-K,I is more negative than that of g-C₃N₅, indicating that more photogenerated electrons will accumulate in the CB under vis-light irradiation, which is highly beneficial for the HER.⁶²

Fig. 4 Mott–Schottky plots of g-C₃N₅ (a), C₃N₅-K,I (b), and Pd/C₃N₅-K,I (c). (d) Valence band XPS spectra of samples. (e) Estimated bandgaps of samples.

3.3. Photocatalytic H₂ evolution performance

As illustrated in Fig. 5a, pure g-C₃N₅ exhibits a modest hydrogen production performance of 53 μmol, which is attributed to the rapid recombination of carriers. Upon modification with KI, the photocatalyst displays a significantly higher H₂-evolution activity (250.3 μmol, Fig. S4, ESI†), indicating that I plays an essential role in the hydrogen production process. With further introduction of Pd nanoparticles into C₃N₅-K,I, a substantial enhancement in H₂-production performance is observed. Specifically, the Pd/C₃N₅-K,I photocatalyst with a Pd loading of 2.25% demonstrates an optimal photocatalytic H₂-generation activity of 753 μmol (the proportion is determined based on the ICP results, Table S2, ESI†). However, the decline in TEOA, which is critical for maintaining the stability of g-C₃N₅,⁶³ along with the agglomeration of Pt, may result in a non-linear increase in H₂ generation over time (Fig. S5, ESI†). As shown in Fig. 5b and c, the H₂ evolution rate and AQY of 2.25% Pd/C₃N₅-K,I reach 2878 μmol g⁻¹ h⁻¹ and 6.07% (400 nm), respectively, which is relatively high compared to those reported for C₃N₅-based photocatalysts (more details in Fig. S6 and Table S3, ESI†). In a control experiment, the hydrogen production performance of Pd/C₃N₅ without I is significantly lower than that of 2.25% Pd/C₃N₅-K,I (Fig. S7, ESI†), demonstrating the important synergistic role of I and Pd nanoparticles in enhancing photocatalytic performance. The negligible reduction in photocatalytic H₂ production of 2.25% Pd/C₃N₅-K,I after four cycles suggests its exceptional stability (Fig. 5d). The effects of four distinct sacrificial agents on H₂ production by g-C₃N₅ are illustrated in Fig. 5e, demonstrating that TEOA is the most effective photocatalytic sacrificial agent among the tested materials. In Fig. 5f, the addition of a small amount of Pt (0.36 wt%) significantly enhances the H₂ production performance for g-C₃N₅ and Pd/C₃N₅-K,I. However, further increasing the Pt content to 1 wt% resulted in only marginal improvements due to the formation of agglomerated Pt clusters. Therefore, 0.36 wt% Pt is considered the most suitable option in this study.

Fig. 5 (a) H₂ generation amounts for different samples. (b) Photocatalytic H₂ evolution rates of g-C₃N₅, C₃N₅-K,I, and Pd/C₃N₅-K,I. (c) Photocatalytic H₂ evolution and AQY (400 nm) of C₃N₅-based catalysts reported in the literature (more details are given in Table S3, ESI†). (d) Photocatalytic stability tests of Pd/C₃N₅-K,I. (e) Different sacrificial agents on H₂ production of g-C₃N₅. (f) H₂ generation amounts for g-C₃N₅ and Pd/C₃N₅-K,I with different Pt contents after 6 hours.

3.4. Photocatalytic reaction mechanism

To elucidate the charge separation process in Pd/C₃N₅-K,I, we conducted geometry optimization calculations for C₃N₅-K,I, Pd₁/C₃N₅-K,I and Pd₃/C₃N₅-K,I, where Pd₁ and Pd₃ represent the single and triple Pd atoms, respectively. Firstly, the electron–hole distributions in the excited states of C₃N₅-K,I, Pd₁/C₃N₅-K,I and Pd₃/C₃N₅-K,I were calculated. The absence of imaginary frequencies in the optimized structures indicated that all configurations are stable. Subsequently, the structural energies of single Pd atoms at various positions were computed. Although the lowest energy was found at a relatively central position, the energy differences among several positions were not significant (Fig. S8, ESI†). As shown in Fig. 6a, the electron–hole pairs in C₃N₅-K,I are significantly separated under the S0–S5, S0–S8 and S0–S9 excited states compared to g-C₃N₅,⁴¹ reflecting the pronounced effect of I atoms on in-plane electron migration in g-C₃N₅. When a single Pd atom is introduced into the interlayer of C₃N₅-K,I (Fig. 6b), it is evident that the Pd atom acts as an electron donor, facilitating the directed concentration of electrons within C₃N₅-K,I and achieving efficient spatial electron–hole pair separation. Consistent results were obtained when triple Pd atoms were introduced into the layers (Fig. 6c). In conclusion, the synergistic action of I and Pd nanoparticles effectively enhances the separation efficiency of both planar and spatial carriers. This calculation result supports the observed high photocatalytic H₂ production performance.

Fig. 6 Electron–hole distribution of (a) C₃N₅-K,I, (b) Pd₁/C₃N₅-K,I, and (c) Pd₃/C₃N₅-K,I in different excited states (Pd₁ and Pd₃ represent the single atom and triple atoms, respectively). Blue and green isosurfaces represent the hole and electron distributions.

The spatial distribution of the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) is closely related to the separation of photoexcited electrons and holes, significantly influencing photocatalytic efficiency. Due to the severe overlap of HOMO and LUMO distributions in g-C₃N₅ (Fig. 7a), photoexcited charges tend to transfer to the same locations, severely hindering electron–hole separation and resulting in poor photocatalytic performance.⁶⁴ However, the introduction of KI leads to a clear redistribution of the HOMO and LUMO (Fig. 7b), facilitating efficient charge transfer. The electron density of the LUMO for Pd₃/C₃N₅-K,I is concentrated in the C₃N₅-K,I plane, while in the HOMO, the electron density is primarily distributed around the Pd atoms (Fig. 7c). This confirms that electrons are transferred to the HOMO via Pd nanoparticles.⁶⁵ Furthermore, the HOMO–LUMO energy (HLG) of g-C₃N₅ significantly decreases with the introduction of I and Pd, indicating an increase in the electrical conductance of C₃N₅-K,I and Pd/C₃N₅-K,I.⁶⁶ Compared to g-C₃N₅, Pd/C₃N₅-K,I exhibits enhanced electron–hole separation and a narrowed HLG, while Pd maintains electron-supplying characteristics, making it highly promising for photocatalytic applications.

Fig. 7 Electronic structure of the optimized HOMO and LOMO of (a) g-C₃N₅, (b) C₃N₅-K,I and (c) Pd₃/C₃N₅-K,I (Pd₃ represent triple atoms). Blue and green isosurfaces represent the hole and electron distributions.

In this study, we propose a plausible model for the Pd/C₃N₅-K,I photocatalyst under vis-light illumination to elucidate the reaction mechanism, as depicted in Scheme 2. Taking advantage of the synergistic effect of I and Pd nanoparticles, photogenerated electrons with sufficient energy are excited from the VB to the CB, leaving holes in the VB. Meanwhile, Pd nanoparticles and C₃N₅-K,I form an ohmic contact with an electric field directed towards Pd, facilitating electron migration from Pd to C₃N₅-K,I. Consequently, a substantial number of electrons from the Pd nanoparticles are effectively injected into C₃N₅-K,I, leading to a high concentration of electrons accumulating around C₃N₅-K,I, facilitated by the synergistic action of I. This enables the H⁺ (from H₂O) adsorbed on the Pd/C₃N₅-K,I photocatalyst to readily obtain electrons from C₃N₅-K,I, thereby generating H₂.

Scheme 2 The probable photocatalytic HER mechanism of the Pd/C₃N₅-K,I photocatalyst.

4. Conclusions

In this study, to solve the problem of rapid carrier recombination of g-C₃N₅, we successfully synthesized a novel photocatalyst comprising Pd nanoparticles embedded within K,I-doped g-C₃N₅. This novel conceptual framework exhibits two distinct charge transfer pathways both within the layers and across the interlayers. Our comprehensive characterization and computational analyses demonstrate that the strategic formation of C–I bonds enhances the photocatalytic activity through increased in-plane carrier mobility. Crucially, the incorporation of Pd nanoparticles facilitates the injection of electrons into the C₃N₅-K,I layer; meanwhile, Pd nanoparticles exhibit a hole-rich state, thereby enhancing the spatial separation of photoinduced charge carriers. The synergistic effect of these modifications optimizes electron–hole pair separation efficiency and enables the Pd/C₃N₅-K,I photocatalyst to achieve remarkably improved photocatalytic H₂ generation performance. This significant improvement underscores the efficacy of our approach in addressing the limitations of conventional CN-based photocatalysts. Furthermore, this approach paves the way for efficient photocatalytic systems with applications in hydrogen production as well as sustainable energy conversion and environmental remediation.

Author contributions

Yanan Gao: responsible for designing the study, developing the methodology, conducting investigations, performing formal analysis, and writing and preparing the original draft. Jingjing Wang: involved in calculations and formal analysis. Jingxuan Yang and Yajie Wang: contributed to investigations and formal analysis. Wenjuan Tian: engaged in calculations, formal analysis, and supervision of the project. Bin Liu: oversaw the project, provided guidance on writing and editing, managed the project, and secured funding.

Data availability

The authors confirm that the data supporting the findings of this study are available within the article and its ESI.†

Conflicts of interest

The authors declare no conflicts of interest.

Acknowledgements

We gratefully acknowledge the Fundamental Research Program of Shanxi Province (Grant No. 202303021221058).

References

1. W. Li, L. Lu, C. Cheng, N. Ren, S.-T. Yang and M. Liu, Bioresour. Technol., 2022, 364, 128069.
2. V. Nikolaou, G. Charalambidis, K. Ladomenou, E. Nikoloudakis, C. Drivas, I. Vamvasakis, S. Panagiotakis, G. Landrou, E. Agapaki, C. Stangel, C. Henkel, J. Joseph, G. Armatas, M. Vasilopoulou, S. Kennou, D. M. Guldi and A. G. Coutsolelos, ChemSusChem, 2021, 14, 961–970.
3. N.-N. Vu, S. Kaliaguine and T.-O. Do, ACS Sustainable Chem. Eng., 2019, 8, 853–863.
4. A. Fujishima and K. Honda, Nature, 1972, 238, 37–38.
5. Q. Lan, S. Jin, B. Yang, Q. Zhao, C. Si, H. Xie and Z. Zhang, Trans. Tianjin Univ., 2022, 28, 214–225.
6. M. Ren, J. Meng, Y. Yang, X. Zhang, G. Yang, L. Qin and Y. Guo, Appl. Catal., B, 2024, 345, 123680.
7. Y. Zou, S. Li, D. Zheng, J. Feng, S. Wang, Y. Hou and G. Zhang, Sci. China Mater., 2024, 67, 2215–2223.
8. A. M. Sadanandan, M. Fawaz, N. P. Dharmarajan, M. Huš, G. Singh, C. I. Sathish, B. Likozar, Z. Li, A. M. Ruban, C.-H. Jeon, J.-H. Yang, P. Kumar and A. Vinu, Appl. Catal., B, 2024, 124701.
9. X. Wang, K. Maeda, A. Thomas, K. Takanabe, G. Xin, J. M. Carlsson, K. Domen and M. Antonietti, Nat. Mater., 2009, 8, 76–80.
10. H. Yang, Z. Jin, H. Hu, Y. Bi and G. Lu, Appl. Surf. Sci., 2018, 427, 587–597.
11. Y. Quan, R. Li, X. Li, R. Chen, Y. H. Ng, J. Huang, J. Hu and Y. Lai, Small, 2024, 2406576.
12. P. Song, D. Wang, B. Wang and P. Yang, J. Ind. Eng. Chem., 2024, 138, 472–480.
13. G. P. Mane, S. N. Talapaneni, K. S. Lakhi, H. Ilbeygi, U. Ravon, K. Al-Bahily, T. Mori, D.-H. Park and A. Vinu, Angew. Chem., Int. Ed., 2017, 56, 8481–8485.
14. C. Fu, T. Wu, G. Sun, G. Yin, C. Wang, G. Ran and Q. Song, Appl. Catal., B, 2023, 323, 122196.
15. H. Li, R. Li, G. Liu, M. Zhai and J. Yu, Adv. Mater., 2023, 36.
16. C. Ma, Z. Yu, J. Wei, C. Tan, X. Yang, T. Wang, G. Yu, C. Zhang and X. Li, Appl. Catal., B, 2022, 319, 121951.
17. Q. Liu, X. Du, A. Zhou, J. Chen, X. Wang, R. Wang, M. Cheng, J. Hu, T. Wei, Y. Cui, F. Chen, W. Li, W.-L. Dai and B. Liu, J. Colloid Interface Sci., 2025, 678, 114–124.
18. E. Cui, Y. Lu, Z. Li, J. Sang, Z. Wang, M. Xie, X. Yang, J. Cao and Y. Zhang, Appl. Catal., B, 2024, 347, 123806.
19. S. He, G. Wang, Y. Liu, L. Luo, T. Jiang, T. Fan, X. Zhu, Y. Ding, J. Jing and S. Guan, Environ. Sci.: Nano, 2024, 11, 3202–3213.
20. H. S. Gujral, G. Singh, J. H. Yang, C. I. Sathish, J. Yi, A. Karakoti, M. Fawaz, K. Ramadass, A. A. H. Al-Muhtaseb, X. Yu, M. B. H. Breese and A. Vinu, Carbon, 2022, 195, 9–18.
21. H. Yin, C. Yuan, H. Lv, K. Zhang, X. Chen and Y. Zhang, Sep. Purif. Technol., 2023, 308, 122815.
22. J. Zhang, Z. Li, B. Liu, M. Chen, Y. Zhou and M. Zhu, Appl. Catal., B, 2023, 328, 122522.
23. X. Gao, Y. Zhou, Z. Cheng, Y. Tan, S. Liu and Z. Shen, Int. J. Hydrogen Energy, 2019, 44, 27421–27428.
24. H. Wang, Y. Ma, S. Tang, Y. Liang, D. Zhang, X. Jin, Q. Wang, W. Sun, L. Zheng and W. Li, Carbon, 2024, 218, 118723.
25. Y. Liu, M. Tayyab, W. Pei, L. Zhou, J. Lei, L. Wang, Y. Liu and J. Zhang, Small, 2023, 19, 2208117.
26. H. Long, P. Wang, X. Wang, F. Chen and H. Yu, Appl. Surf. Sci., 2022, 604, 154457.
27. S. Cao, H. Li, T. Tong, H.-C. Chen, A. Yu, J. Yu and H. M. Chen, Adv. Funct. Mater., 2018, 28, 1802169.
28. G. Dukovic, M. G. Merkle, J. H. Nelson, S. M. Hughes and A. P. Alivisatos, Adv. Mater., 2008, 20, 4306–4311.
29. C. Merschjann, S. Tschierlei, T. Tyborski, K. Kailasam, S. Orthmann, D. Hollmann, T. Schedel-Niedrig, A. Thomas and S. Lochbrunner, Adv. Mater., 2015, 27, 7993–7999.
30. M. Luna, M. Barawi, S. Gómez-Moñivas, J. Colchero, M. Rodríguez-Peña, S. Yang, X. Zhao, Y.-H. Lu, R. Chintala, P. Reñones, V. Altoe, L. Martínez, Y. Huttel, S. Kawasaki, A. Weber-Bargioni, V. A. de la Peña ÓShea, P. Yang, P. D. Ashby and M. Salmeron, ACS Appl. Mater. Interfaces, 2021, 13, 50531–50538.
31. Y. Peng, B. Lu, N. Wang, J. E. Lu, C. Li, Y. Ping and S. Chen, ACS Appl. Mater. Interfaces, 2019, 11, 24707–24714.
32. Y. Ben-Shahar, F. Scotognella, I. Kriegel, L. Moretti, G. Cerullo, E. Rabani and U. Banin, Nat. Commun., 2016, 7, 10413.
33. Y. Peng, B. Lu, F. Wu, F. Zhang, J. E. Lu, X. Kang, Y. Ping and S. Chen, J. Appl. Chem. Sci., 2018, 140, 15290–15299.
34. F. Yan, Y. Wu, L. Jiang, X. Xue, J. Lv, L. Lin, Y. Yu, J. Zhang, F. Yang and Y. Qiu, ChemSusChem, 2020, 13, 876–881.
35. J. Yang, Y. Gao, X. Li, Y. Wang, P. Guan and B. Liu, Mater. Sci. Semicond. Process., 2024, 175, 108267.
36. W.-W. Liu, J. Pan and R.-F. Peng, Rare Met., 2021, 40, 3554–3560.
37. W. Liu, X. Liu, S. Xin, Y. Wang, S. Huo, W. Fu, Q. Zhao, M. Gao and H. Xie, Appl. Energy, 2024, 358, 122552.
38. I. Y. Kim, S. Kim, X. Jin, S. Premkumar, G. Chandra, N. S. Lee, G. P. Mane, S. J. Hwang, S. Umapathy and A. Vinu, Angew. Chem., Int. Ed., 2018, 57, 17135–17140.
39. X. Guan, M. Fawaz, R. Sarkar, C.-H. Lin, Z. Li, Z. Lei, P. D. Nithinraj, P. Kumar, X. Zhang, J.-H. Yang, L. Hu, T. Wu, S. Chakraborty, J. Yi and A. Vinu, J. Mater. Chem. A, 2023, 11, 12837–12845.
40. C. Peng, L. Han, J. Huang, S. Wang, X. Zhang and H. Chen, Chin. J. Catal., 2022, 43, 410–420.
41. H. Che, J. Wang, X. Gao, J. Chen, P. Wang, B. Liu and Y. Ao, J. Colloid Interface Sci., 2022, 627, 739–748.
42. C. Han, W. Yi, Z. Li, C. Dong, H. Zhao and M. Liu, Electrochim. Acta, 2023, 447, 142083.
43. S. Lu, X. Li, B. Yu, J. Ding, Y. Zhong and H. Zhang, Ecotoxicol. Environ. Saf., 2022, 246, 114148.
44. J. Shen, C. Luo, S. Qiao, Y. Chen, Y. Tang, J. Xu, K. Fu, D. Yuan, H. Tang, H. Zhang and C. Liu, ACS Catal., 2023, 13, 6280–6288.
45. J. Xue, Y.-N. Jing, L.-L. Li, X.-L. Yin, Z.-F. Xu, J. Li and Y.-L. Wang, Colloids Surf., A, 2024, 685, 133158.
46. M. Zhai, Y. Zhang, J. Xu, C. Wang and L. Wang, Appl. Surf. Sci., 2024, 652, 159378.
47. S. Ma, D. Yang, B. Li, Y. Guan, M. Wu, J. Wu, Y. Guo, L. Sheng, L. Liu and T. Yao, J. Colloid Interface Sci., 2024, 664, 960–971.
48. Y. Guo, Q. Liu, Z. Li, Z. Zhang and X. Fang, Appl. Catal., B, 2018, 221, 362–370.
49. S. Ma, J. Zhao, Q. Gao, C. Song, H. Xiao, F. Li and G. Li, Angew. Chem., Int. Ed., 2023, 62, e202315564.
50. S. Hu, P. Qiao, X. Liang, G. Ba, X. Zu, H. Hu, J. Ye and D. Wang, Appl. Catal., B, 2024, 346, 123737.
51. D. Ramírez-Ortega, D. Guerrero-Araque, P. Acevedo-Peña, E. Reguera, H. A. Calderon and R. Zanella, Int. J. Hydrogen Energy, 2021, 46, 34333–34343.
52. R. Chen, S. Chen, L. Wang and D. Wang, Adv. Mater., 2023, 36.
53. C. Wan, L. Zhou, L. Sun, L. Xu, D.-G. Cheng, F. Chen, X. Zhan and Y. Yang, Chem. Eng. J., 2020, 396, 125229.
54. N. Ramesh Reddy, U. Bharagav, M. M. Kumari, K. K. Cheralathan, P. K. Ojha, M. V. Shankar and S. W. Joo, Ceram. Int., 2021, 47, 10216–10225.
55. Y. Hu, S. Zhang, Z. Zhang, H. Zhou, B. Li, Z. Sun, X. Hu, W. Yang, X. Li, Y. Wang, S. Liu, D. Wang, J. Lin, W. Chen and S. Wang, Adv. Mater., 2023, 35.
56. B. Su, M. Zheng, W. Lin, X. F. Lu, D. Luan, S. Wang and X. W. Lou, Adv. Energy Mater., 2023, 13, 2203290.
57. X. Yu, H. Su, J. Zou, Q. Liu, L. Wang and H. Tang, Chin. J. Catal., 2022, 43, 421–432.
58. D. Zheng, Q. Wang, Z. Pan, S. Wang, Y. Hou and G. Zhang, Sci. China Mater., 2024, 67, 1900–1906.
59. J. Lu, Z. Chen, Y. Shen, H. Yuan, X. Sun, J. Hou, F. Guo, C. Li and W. Shi, J. Colloid Interface Sci., 2024, 670, 428–438.
60. J. Li, Z. Huang, C. Wang, L. Tian, X. Yang, R. Zhou, M. N. Ghazzal and Z.-Q. Liu, Appl. Catal., B, 2024, 340, 123181.
61. H. Xu, Z. Wang, S. Feng, X. Liu, X. Gong and J. Hua, Int. J. Hydrogen Energy, 2023, 48, 8071–8081.
62. X. Guan, X. Zhang, Z. Li, S. Deshpande, M. Fawaz, N. P. Dharmarajan, C.-H. Lin, Z. Lei, L. Hu, J.-K. Huang, P. Kumar, Z. Sun, S. Chakraborty and A. Vinu, Chem. Mater., 2024, 36, 4511–4520.
63. M. Wang, S. Shen, L. Li, Z. Tang and J. Yang, J. Mater. Sci., 2017, 52, 5155–5164.
64. B. Zhao, J. Xu, D. Gao, F. Chen, X. Wang, T. Liu, X. Wu and H. Yu, Appl. Catal., B, 2024, 355, 124215.
65. M. Younas, M. Yar, H. AlMohamadi, T. Mahmood, K. Ayub, A. L. Khan, M. Yasin and M. A. Gilani, Int. J. Hydrogen Energy, 2024, 51, 758–773.
66. M. F. Fellah, Int. J. Hydrogen Energy, 2019, 44, 27010–27021.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4tc03914a

---