Icosahedron kernel defect in Pt₁Ag_x series of bimetallic nanoclusters enhances photocatalytic hydrogen evolution

Abstract

Developing high-efficiency photocatalysts for photocatalytic hydrogen production and understanding the structure–property relationships is much desired. In this study, a family of Pt₁Ag_x (x = 9, 11, 13 and 14) nanoclusters (NCs), including a new Pt₁Ag₁₁(SR)₅(P(Ph-OMe)₃)₇ NC, were designed and synthesized via ligand engineering (SR = 2,3,5,6-tetrafluorothiophenol, P(Ph-OMe)₃ = tris(4-methylphenyl)phosphine). The positive effect of the kernel structural defect on photocatalytic activity was investigated using the photocatalytic water-splitting reaction as a model, and the mechanistic relationship between the defect structure and catalytic activity was clarified. In this series of Pt₁Ag_x bimetallic NCs, the Pt₁Ag₁₁ NC, which exhibits a distinctive defect-containing icosahedral kernel structure, displayed excellent catalytic performance for photocatalytic hydrogen evolution, with the hydrogen production rate reaching 1780 μmol g⁻¹ h⁻¹. The experimental results revealed that the superior catalytic activity of Pt₁Ag₁₁/g-C₃N₄ may originate from the formation of Z-scheme heterojunction between Pt₁Ag₁₁ and the g-C₃N₄, facilitating efficient electron–hole separation and charge transfer. Furthermore, density-functional theory (DFT) calculations reveal the critical role of the defect-containing icosahedron-kernel on photocatalytic activity, which is favourable for the formation of the most stable nanocomposites and the easy absorption of H* intermediates on the Ag sites in Pt₁Ag₁₁/g-C₃N₄. This paper provides insights into the effect that the defects have on the mechanism of the photocatalytic hydrogen evolution reaction at the atomic level and promotes the rational design of high-efficiency photocatalysts.

---

1 Introduction

Photocatalytic water splitting for hydrogen production has attracted extensive interest because it offers a pollution-free and sustainable route to alleviate the energy crisis and tackle environmental issues.^1–3 However, achieving high conversion efficiency remains a significant challenge due to the complex processes involved in photocatalytic reactions, including photo absorption, charge separation, transport, dissociation, and recombination.^4,5 Therefore, the development of high-efficiency photocatalysts for photocatalytic hydrogen production, and a detailed understanding of structure–property relationships, are desired.⁶

Atomic scale metal nanoclusters (NCs) (1–3 nm in diameter) have shown potential for application in photocatalysis due to their unique atomic stacking modes, optical properties and abundance of catalytic active sites.^7,8 Moreover, their ultrasmall size endows metal NCs with discrete molecule-like electronic energy levels, whereby they can be excited as small-band-gap semiconductors to generate electrons and holes.⁹ To avoid the unfavorable agglomeration of metal NCs caused by high surface energies during the photocatalytic reaction, NC-based composite photocatalysts were fabricated to improve catalytic stability and activity.^10,11 In 2013, Negishi et al. reported glutathione-protected Au₂₅ NCs loaded onto BaLa₄Ti₄O₁₅ as hybrid catalysts for the water-splitting reaction; the hybrid catalysts exhibited photocatalytic activity 2.6 times higher than that of co-catalysts loaded with larger gold nanoparticles (10–30 nm).¹² Subsequently, a wide range of metal NCs has been used in photocatalytic hydrogen evolution.¹³ For instance, Lu et al. synthesized Pt₅(GSH)₁₀ NCs and immobilized them on multi-arm CdS nanorods (NRs). The Pt NCs extracted the photoinduced electrons of the CdS NRs and enhanced charge separation, thereby ensuring the Pt₅–CdS composite catalysts exhibited an improved photocatalytic H₂ production rate of 13.0 mmol g⁻¹ h⁻¹ H₂.¹⁴ Wang et al. synthesized Ag₄₄ NCs and immobilized them on TiO₂. The photocatalytic H₂ production of Ag₄₄–TiO₂ was 7.4 mmol g⁻¹ h⁻¹. This elevated performance of Ag₄₄–TiO₂ was attributed to both the extension of the photoresponse time and the efficient separation and transport of charge carriers.¹⁵ Several methods have been proposed to utilize photogenerated carriers effectively for photocatalysis, such as heteroatom doping or Z-scheme heterojunctions,^10,16–19 however, the efficient charge transfer and separation of NC-based hybrid catalysts in photocatalytic reactions still poses a challenge.²⁰

Previous work has revealed that a NC structure that contains defects can significantly enhance catalytic performance. For example, Nematulloev et al. reported that defect-containing Cu₂₈ NCs have structural vertex defects which cause a distortion in the framework and lower the symmetry, thus exhibiting more efficient selectivity in C–C cross-coupling reactions compared with Cu₂₉ NCs.²¹ In addition, Silalahi et al. reported the discovery of hydride-containing 2-electron palladium/copper superatomic alloys, namely PdHCu₁₁ and PdHCu₁₂. The distinctive defects in PdHCu₁₁ expose the central Pd atom, thus providing an active site to catalyze the reaction, resulting in better hydrogen evolution reaction activity than PdHCu₁₂.²² A few reports have been published on the structural defects of metal NCs for catalysis; however, the effect of defects on the formation of heterojunctions or related mechanisms for enhanced photocatalysis have not yet been clarified for composite photocatalysts.

Herein, Pt₁Ag₁₁(SR)₅(P(Ph-OMe)₃)₇ (Pt₁Ag₁₁ for short, where SR is 2,3,5,6-tetrafluorothiophenol and P(Ph-OMe)₃ is tris(4-methylphenyl)phosphine) with three vertex Ag atom defects within the icosahedral kernel was synthesized to construct NCs that exhibit a defect effect. The X-ray crystal structure of Pt₁Ag₁₁ was identified and categorized as belonging to the Pt₁Ag_x family (x = 9, 11, 13 and 14) (Scheme 1). Considering this, the relationship between the defect-containing icosahedral kernel and the catalytic activity was established using photocatalytic water splitting as a model. Nanocomposite photocatalysts were prepared by loading four Pt₁Ag_x NCs onto two-dimensional graphite-like carbon nitride (g-C₃N₄). g-C₃N₄ has a sheet-like structure with a suitable bandgap, is easily functionalized, and exhibits excellent thermal stability, and chemical corrosion resistance.^23,24 In this series of Pt₁Ag_x bimetallic NCs, the Pt₁Ag₁₁ NC, which exhibits a distinctive defect-containing icosahedral kernel structure displayed excellent catalytic performance for photocatalytic hydrogen evolution, with the hydrogen production rate reaching 1780 μmol g⁻¹ h⁻¹. The superior catalytic activity of Pt₁Ag₁₁/g-C₃N₄ may originate from the formation of a Z-scheme heterojunction between the two materials, facilitating the extension of the photoresponse time whilst facilitating more efficient electron–hole separation and charge transfer. Theoretical calculations further demonstrated that defects within the icosahedral kernel facilitated electron transport and separation. This paper offers a detailed understanding of how defects influence the photocatalytic hydrogen evolution reaction at the atomic level, facilitating the rational design of efficient photocatalysts.

Scheme 1 Structure of the Pt₁Ag_x series of nanoclusters (including the kernel and the whole structure). Color code: Pt-yellow, Ag-turquoise, S-red, P-pink, F-bright green, Cl-green, O-orange, C-grey.

2 Experimental section

2.1 Chemicals

Most reagents were purchased from Sigma-Aldrich and used without further purification, including hexachloroplatinic (IV) acid (H₂PtCl₆·6H₂O, 99.99%, metals basis), potassium tetrachloroplatinate (K₂PtCl₄, 99.9%, metals basis), silver nitrate (AgNO₃), 2,3,5,6-tetrafluorothiophenol (SR, C₆H₂F₄S), pentafluorobenzenethiol (PFBT, C₆HF₅S), 2-chloro-4-fluorobenzenethiol (C₆H₄FClS), sodium borohydride (NaBH₄), tris(4-fluorophenyl)phosphine (P(Ph-F)₃), tris(4-methylphenyl)phosphine (P(Ph-OMe)₃) and triphenylphosphine (PPh₃). 5,5-Dimethyl-1-pyrroline-1-oxide (DMPO), tetraphenyl phosphonium bromide (PPh₄Br, 98%), methylene chloride (CH₂Cl₂, HPLC), methanol (CH₃OH, HPLC) and n-hexane (HeX, HPLC grade) were purchased from Energy Chemical.

2.2 Synthesis of Pt₁Ag₁₁(SR)₅(P(Ph-OMe)₃)₇

40 mg of AgNO₃ and H₂PtCl₆·6H₂O (12 mg, 0.0225 mmol) were dissolved into 10 mL of methanol in a 50 mL round-bottomed flask. The solution was stirred vigorously at room temperature for 10 min. The solution immediately turned brown. Subsequently, 40 μL of 2,3,5,6-tetrafluorothiophenol was added into the flask. After 5 min of reaction, 200 mg of tris(4-methylphenyl)phosphine dissolved in 10 mL of CH₂Cl₂ was added under vigorous stirring. The color of the solution became transparent. After half an hour, 3 mL of an aqueous solution of NaBH₄ (80 mg) was added quickly to the reaction mixture under vigorous stirring. The solution color immediately changed from brown to black. The reaction was subsequently carried out for a duration of 9 hours under a N₂ atmosphere at room temperature. The crystals were crystallized from CH₂Cl₂/hexane at room temperature in the dark to afford red block single crystals after 2 weeks.

2.3 Synthesis of Pt₁Ag₁₄(SR)₆(P(Ph-F)₃)₈

40 mg of AgNO₃ and H₂PtCl₆·6H₂O (12 mg, 0.0225 mmol) were dissolved into 10 mL of methanol in a 50 mL round-bottomed flask. The solution was stirred vigorously at room temperature for 10 min. The solution immediately turned brown. Subsequently, 40 μL of 2,3,5,6-tetrafluorothiophenol was added into the flask. After 5 min of reaction, 200 mg of tris(4-fluorophenyl)phosphine dissolved in 10 mL of CH₂Cl₂ was added under vigorous stirring. The color of the solution became transparent. After half an hour, 2 mL of an aqueous solution of NaBH₄ (25 mg) was added quickly to the reaction mixture under vigorous stirring. The solution color immediately changed from brown to black. The reaction was subsequently carried out for a duration of 22 h under a N₂ atmosphere at room temperature. The crystals were crystallized from CH₂Cl₂/hexane at room temperature in the dark to afford orange-red block single crystals after 5 days.

2.4 Synthesis of Pt₁Ag₁₄(SR)₆(PPh₃)₈

40 mg of AgNO₃ and H₂PtCl₆·6H₂O (12 mg, 0.0225 mmol) were dissolved into 10 mL of methanol in a 50 mL round-bottomed flask. The solution was stirred vigorously at room temperature for 10 min. The solution immediately turned brown. Subsequently, 40 μL of 2,3,5,6-tetrafluorothiophenol were added into the flask. After 5 min of reaction, 200 mg of triphenylphosphine dissolved in 10 mL of CH₂Cl₂ was added under vigorous stirring. The color of the solution became transparent. After half an hour, 2 mL of an aqueous solution of NaBH₄ (40 mg) was added quickly to the reaction mixture under vigorous stirring. The solution color immediately changed from brown to black. The reaction was subsequently carried out for a duration of 12 h under a N₂ atmosphere at room temperature. The crystals were crystallized from CH₂Cl₂/hexane at room temperature in the dark to afford yellow rod-like single crystals after 2 weeks.

2.5 Synthesis of the g-C₃N₄ nanosheet

Pristine g-C₃N₄ was synthesized via a thermal polymerization method. 5 g of melamine was calcined at 550 °C for 4 h at a rate of 5 °C min⁻¹ in a muffle furnace. Then, the resulting yellow sample was collected and spread in a crucible, then calcined at 500 °C for 2 h at a heating rate of 5 °C min⁻¹ to obtain the carbon nitride nanosheet.

2.6 Preparation of Pt₁Ag_x/g-C₃N₄ nanocomposites

For the preparation process, 5 mg of PtAg nanoclusters were dissolved in 10 mL of methylene chloride in a round-bottom flask (RBF), which gave a medium to intense color. To the brown-colored solution, 100 mg of g-C₃N₄ nanosheet were added and the mixture was stirred for 12 h. The color of the supernatant became faint as the nanoclusters were successfully embedded onto the g-C₃N₄ nanosheet. The reaction mixture was then centrifuged at 8000 rpm for 10 min. The clear DCM layer was discarded, and the g-C₃N₄ precipitate was dried under vacuum for a few hours. The dry powder was characterized via TEM analyses and used for the catalytic reactions.

2.7 Characterization

The data collection for the single-crystal X-ray diffraction (SC-XRD) analysis of all the nanocluster crystal samples was carried out on a Stoe Stadivari diffractometer under a nitrogen flow, using graphite-monochromatized Cu Kα radiation (λ = 1.54186 Å). Data reductions and absorption corrections were performed using the SAINT and SADABS programs, respectively. The structure was solved by direct methods and refined with full-matrix least squares on F2 using the SHELXTL software package. All non-hydrogen atoms were refined anisotropically, and all the hydrogen atoms were set in geometrically calculated positions and refined isotropically using a riding model. All crystal structures were treated with PLATON SQUEEZE. The diffuse electron densities from these residual solvent molecules were removed. The CCDC number of Pt₁Ag₁₁ is 2380779.

Electrospray ionization mass spectrometry (ESI-MS) measurements were performed on a MicrOTOF-QIII high-resolution mass spectrometer. All UV-vis spectra of the nanoclusters were recorded using an Agilent 8453, and the samples were dissolved in CH₂Cl₂ whose background correction was made using a CH₂Cl₂ blank. Transmission electron microscopy (TEM, JEM-2010) was used to investigate the morphologies and energy-dispersive X-ray spectroscopy (EDS) analyses were performed on a JEOL JEM-2100F FEG TEM operated at 200 kV. Nanocluster powder samples were used for the analysis. X-ray photoelectron spectroscopy (XPS, Thermo-VG Scientific, E = 1486.60 eV, Mg Kα radiation, USA) measurements were performed to detect the elemental composition. The samples were analyzed for photoluminescence spectra using a MicroTime 200 fluorescence spectrophotometer at room temperature. The emission lifetimes were measured with nanoclusters on a HORIBA FluoroMax-4P. The nanocluster was purged with N₂ for 5 min, then saturated with O₂ for 5 min, respectively.

The powder X-ray diffraction (PXRD) measurements of the materials were carried out using a diffractometer operating (Smartlab 9 kW, Cu Kα radiation) at 40 kV and 200 mA, in a 2θ range of 10–80° with a step width of 0.01°. Fourier-transform infrared spectroscopy (FT-IR) was collected on a Thermo Scientific Nicolet iS50R spectrometer. UV-vis diffuse reflectance spectroscopy (DRS) was conducted on a Shimadzu UV-2600i spectrophotometer at room temperature using BaSO₄ as the reference. The electron paramagnetic resonance (EPR) spectra were obtained using a JEOL JES FA200 to detect ˙O₂⁻ and ˙OH radicals using 5,5-dimethyl-1-pyrroline-1-oxide (DMPO) as a spin trap. Inductively coupled plasma-atomic emission spectrometry (ICP-AES) measurements were performed on an Atomscan advantage instrument from Thermo Jarrell Ash Corporation (USA), whereby 2.5 mg of PtAg NC/g-C₃N₄ nanocomposites were dissolved in 1 mL of concentrated nitric acid and 3 mL of deionized water for testing.

2.8 Photoelectrochemical measurements

The electrochemical impedance spectroscopy (EIS), photocurrent–time profiles and Mott–Schottky diagram were recorded on a CHI760E electrochemical workstation with a standard three-electrode system, where photocatalyst-coated fluorine-doped tin oxide (FTO) was used as the working electrode, Pt plate as the counter electrode, and a saturated Ag/AgCl electrode as the reference electrode. A 0.1 M Na₂SO₄ solution was used as the electrolyte. The as-synthesized samples (10 mg) were added into 400 μL of ethanol and a 20 μL Nafion mixed solution, and the working electrodes were prepared by dropping the suspension (200 μL) onto an FTO glass substrate before drying at room temperature. Mott–Schottky plots were measured at 500, 1000, and 1500 Hz, respectively. EIS was recorded with a bias potential of −1.4 V in the dark.

2.9 Photocatalytic hydrogen evolution

The photocatalytic hydrogen production tests were performed in a Pyrex top-irradiation reaction vessel with a stationary temperature at 25 °C under full-spectrum light, which was connected to a glass closed gas system (Labsolar-6A, Perfect Light). A 300 W Xe lamp was employed to serve as the light source. 50 mg of the as-prepared photocatalyst was ultrasonically dissolved in 10 mL of triethanolamine (TEOA, sacrifice reagent) and 90 mL of H₂O and then loaded into a 370 mL sealed quartz reactor and evacuated using a vacuum pump. The photocatalytic hydrogen performance of the photocatalyst was calculated using gas chromatography (GC-5190, China) equipped with a thermal-conductivity detector (TCD, with Ar as the carrier gas and a 5 Å molecular sieve column). The injection temperature and detection temperature were set to 100 °C during testing, and the column furnace temperature was set to 50 °C. The production of H₂ every 30 min was monitored using a gas chromatograph. The temperature of the whole process was kept at 8 °C using circulating cooling water.

2.10 Apparent quantum efficiency (AQE) measurements

The AQE measurement methods were similar with those of the photocatalytic measurements except for the wavelength of the light. Particularly, different monochromatic light (360 nm, 380 nm, 400 nm, 420 nm, 440 nm) was utilized to assess the quantum efficiency of the photocatalyst. Based on the amounts of evolved H₂, the AQE is calculated via the following formula:
where n (mol) refers to the amount of H₂ molecules, v (mol s⁻¹) represents the rate of H₂ production, N_A (6.022 × 10²³ per mol) represents the Avogadro constant, ℏ (6.626 × 10⁻³⁴ J S) represents the Planck constant, c (3 × 10⁸ m s⁻¹) is the speed of light, S is the irradiation area (cm²), P refers the intensity of the irradiation light (W cm⁻²), t signals the photoreaction time (s), and λ is the wavelength of the monochromatic light (m).

The simplified formula is translated as follows:

2.11 Computational method

The Vienna ab initio simulation package (VASP) was selected to perform our density-functional theory (DFT) calculations. The exchange–correlation functional was treated through the generalized gradient approximation (GGA) using the Perdew–Burke–Ernzerhof (PBE) functional. The cutoff energy was set to 570 eV for all calculations. The Brillouin zone (BZ) was sampled using a 2 × 2 × 1 gamma-centered Monkhorst–Pack grid. The energy and the force standard received on each atom were set to be 10⁻⁶ eV and 0.02 eV Å⁻¹, respectively. To avoid the interaction between adjacent slabs, a vacuum space of 20 Å was used. The visualization for electronic and structural analysis software (VESTA 3) was utilized for visualization and plotting. The VASPKIT code was employed for the postprocessing of the calculated data.

The charge density difference (Δρ) was determined by

Δρ = ρ_*M − ρ_M − ρ_* (1)

where ρ_* and ρ_M represent the charge densities of the g-C₃N₄ monolayer and the clusters (Pt₁Ag₉, Pt₁Ag₁₁, Pt₁Ag₁₃ or Pt₁Ag₁₄) and ρ_*M refers to total charge densities of the cluster-adsorbed g-C₃N₄ systems.

In addition, Bader charges were utilized to detect charge transfer between the substrate and clusters. To evaluate the energetic stability of the g-C₃N₄-loaded clusters, we calculated their binding energy (E_bind), which is defined as:

E_bind = E_*M − E_* − E_M (2)

where E_*M, E_* and E_M are the total energies of the M-adsorbed g-C₃N₄ monolayer, pure g-C₃N₄ monolayer and the clusters M (Pt₁Ag₉, Pt₁Ag₁₁, Pt₁Ag₁₃ and Pt₁Ag₁₄), respectively. The value of E_bind reflects the interaction between the anchored species and the catalyst, and a more negative value indicates a stronger interaction. The free energy difference (ΔG) of the HER process was calculated according to the approach proposed by Norskov et al. The formula can be defined as below:

ΔG = ΔE + ΔE_zpe − TΔS (3)

where ΔE is the adsorption energy, ΔE_zpe and ΔS are the difference in the zero-point energy and entropy between the adsorbed state and the gas phase, respectively. Here, T = 298.15 K.

The HER can be decomposed into two steps, and the reaction equation can be written as:

* + H⁺ + e⁻ → H* (4)

H* + H⁺ + e⁻ → * + H₂ (5)

where * means the adsorbed materials, and H* represent the adsorbed intermediates. For each step of the hydrogen reactions, the free energy difference can be written as:where ΔG₁ and ΔG₂ denote the free Gibbes energy difference of the proton adsorption and H₂ desorption steps, respectively.

3 Results and discussion

The icosahedral kernel structure was chosen as the model for investigating the defect effect because it is the most widely observed structure in NCs. A novel Pt₁Ag₁₁ NC was synthesized, and can be categorized as part of the Pt₁Ag_x NC series along with three previously reported NCs, Pt₁Ag₉,²⁵ Pt₁Ag₁₃ ⁷ and Pt₁Ag₁₄.²⁶ Single-crystal X-ray crystallography (SC-XRD) revealed that Pt₁Ag₁₁ crystalized in the triclinic space group P1̄. The Pt₁Ag₁₁ NC comprised a Pt₁Ag₉ kernel, stabilized by five P(Ph-OMe)₃ ligands, one Ag(SR)₂[P(Ph-OMe)₃] motif and one Ag(SR)₃[P(Ph-OMe)₃] motif (Fig. 1a). The chemical composition of Pt₁Ag₁₁ was definitively confirmed using electrospray ionization mass spectrometry (ESI-MS) in positive ion mode, as shown in the Fig. 1b. The spectrum contained two intense peaks at m/z of 2500.726 Da (calculated = 2500.765 Da) and 2511.792 Da (calculated = 2511.755 Da), corresponding to the characteristic peaks of Pt₁Ag₁₁(C₆HF₄S)₅(C₂₁H₂₁O₃P)₇Ag₂(CH₃OH) and Pt₁Ag₁₁(C₆HF₄S)₅(C₂₁H₂₁O₃P)₇Ag₂(CH₃ONa). Transmission electron microscopy (TEM) images indicated that Pt₁Ag₁₁ has a uniform appearance, with an average size of 1.32 nm (Fig. S1†). The UV-vis spectrum of Pt₁Ag₁₁ in CH₂Cl₂ displays one strong peak centered at 372 nm and three weak shoulder peaks at 425, 455, and 550 nm (Fig. 1c). For comparison, the UV-vis spectra of the Pt₁Ag₉, Pt₁Ag₁₃, and Pt₁Ag₁₄ NCs revealed different absorption peaks (Fig. S2†). Moreover, the Pt₁Ag_x series in the solid state emitted between 600 and 700 nm (Fig. S3†).

Fig. 1 (a) Structural anatomy of a Pt₁Ag₁₁ nanocluster (Pt₁Ag₉ kernel; Pt₁Ag₉@5P; Pt₁Ag₉@Ag₁S₃P₁@Ag₁S₂P₁; Pt₁Ag₁₁(C₆H₂F₄S)₅(C₂₁H₂₁O₃P)₇ overall structure), color labels: yellow, Pt; turquoise, Ag; red, S; pink, P; bright green, F; orange, O; gray, C; white, H. (b) ESI-MS of the Pt₁Ag₁₁ nanoclusters [Pt₁Ag₁₁(C₆HF₄S)₅(C₂₁H₂₁O₃P)₇Ag₂(CH₃OH) (☆) and Pt₁Ag₁₁(C₆HF₄S)₅(C₂₁H₂₁O₃P)₇Ag₂(CH₃ONa) (★)]; (c) UV-vis spectra of the Pt₁Ag₁₁ nanoclusters. The XPS spectra of (d) Pt 4f and (e) Ag 3d for the Pt₁Ag_x nanoclusters (x = 9, 11, 13, 14).

A cycle between enhanced emission intensity with N₂ and quenched emission intensity with O₂ was detected for the Pt₁Ag₁₁ NC. Meanwhile, the average PL lifetime of the Pt₁Ag₁₁ NC was 4.28 μs, which was prolonged to 5.91 μs in a N₂ atmosphere, and reduced to 2.16 μs in an O₂ atmosphere. Similar results were observed for the other three NCs, indicating that the Pt₁Ag_x series of NCs (x = 9, 11, 13, and 14) were phosphorescent⁷ (Fig. S4–S6†). In addition, the valence states of the Pt and Ag atoms in the Pt₁Ag_x series of bimetallic NCs were investigated and compared with the results obtained from X-ray photoelectron spectroscopy (XPS). The XPS results revealed the presence of Pt, Ag, P, F, O, C, and S in the Pt₁Ag₁₁ NC (Fig. S7†). As shown in Fig. 1d, the Pt 4f binding energies in Pt₁Ag_x were as follows: Pt₁Ag₁₄ (71.5 eV) > Pt₁Ag₉ (71.4 eV) > Pt₁Ag₁₃ (71.3 eV) > Pt₁Ag₁₁ (71.1 eV). The lower binding energies of the Pt 4f peaks indicate that the Pt atoms in Pt₁Ag₁₁ were more partial to 0 valence (71.0 eV) compared with the other members of the Pt₁Ag_x family of NCs. For the binding energies of Ag, a sequence of Pt₁Ag₁₃ (368.4 eV) ≈ Pt₁Ag₁₁ (368.4 eV) > Pt₁Ag₁₄ (368.2 eV) ≈ Pt₁Ag₉ (368.2 eV) can be established. This illustrates that the valence state of Ag in the Pt₁Ag_x NCs lies between 0 and +1 (Fig. 1e).

In addition to determining the difference in the electronic structure of the Pt₁Ag_x series of NCs, analysis of their geometric structures is crucial. SC-XRD revealed that Pt₁Ag₁₃ crystalized in the triclinic space group P1̄ similar to Pt₁Ag₁₁, whereas Pt₁Ag₁₄ crystalized in the monoclinic space group C2/c; Pt₁Ag₉ was found to have the trigonal space group R3 (Tables S1 and S2†). Comparison of the crystal structures of the Pt₁Ag_x series of NCs indicated that Pt₁Ag₁₄ has a complete icosahedral Pt₁Ag₁₂ kernel, while the remaining three NCs exhibit different icosahedral-kernel defects (Fig. S8†). In Pt₁Ag₁₃, removing the vertex silver atom from the icosahedral kernel caused the loss of one Ag–P bond, whereas the peripheral Ag₁S₃P₁ motif did not change. However, three silver atoms were absent from the icosahedral kernel in Pt₁Ag₁₁, and thus, the Ag₁S₃P₁ motif experienced the loss of an Ag–S bond, leading to the formation of the Ag₁S₂P₁ motif (Fig. 2a). Meanwhile, a new Pt–P bond was formed within the icosahedral kernels in both Pt₁Ag₁₁ and Pt₁Ag₉. Pt₁Ag₁₁ and Pt₁Ag₉ have the same kernel, albeit with only a P ligand included for the latter (Fig. 2b). A detailed comparison of the bond lengths and angles within the kernel structures of the Pt₁Ag₉ and Pt₁Ag₁₁ NCs further suggested that the kernel of Pt₁Ag₁₁ was more distorted (Tables S3 and S4†). In addition, the geometric structures were further compared by analyzing the bond lengths of the four Pt₁Ag_x NCs. The Pt–Ag average distance within the icosahedral kernel in the Pt₁Ag_x NCs was Pt₁Ag₁₄ (2.757 Å) > Pt₁Ag₉ (2.747 Å) > Pt₁Ag₁₃ (2.744 Å) > Pt₁Ag₁₁ (2.743 Å). The Ag–Ag average distance in Pt₁Ag_x was Pt₁Ag₁₄ (2.899 Å) > Pt₁Ag₉ (2.891 Å) > Pt₁Ag₁₃ (2.873 Å) > Pt₁Ag₁₁ (2.861 Å). The above results also indicated that the kernel of the NCs became more compact with the loss of kernel silver atoms. Moreover, the kernel of Pt₁Ag₁₁ was the most compact (Fig. S9–S11†). The geometric defects in the icosahedral kernel of the Pt₁Ag_x NCs were correlated with distinct electronic structures. Specifically, both Pt₁Ag₉ and Pt₁Ag₁₁ exhibited a free electron count of 6e, while Pt₁Ag₁₃ and Pt₁Ag₁₄ displayed counts of 7e and 8e, respectively.

Fig. 2 Structures of the Pt₁Ag_x nanoclusters (x = 9, 11, 13, 14). (a) Framework structure: metallic kernel, ligand, and surface motifs; (b) metallic kernel on the Pt₁Ag_x NCs. Color labels: yellow, Pt; turquoise, Ag; red, S; pink, P; green, Cl. All of the C and H atoms are omitted. The crystal data sources for Pt₁Ag₉, Pt₁Ag₁₃ and Pt₁Ag₁₄ were taken from the literature.^7,25,26

The photocatalytic hydrogen evolution reaction was then used to examine the effect of the structural defects on the catalytic activity. A schematic diagram of the photocatalyst preparation process is shown in Scheme S1.†²⁷ Lamellar g-C₃N₄ nanosheet was first synthesized via secondary calcination and exhibited a larger specific surface area (63.9405 m² g⁻¹) compared with bulk g-C₃N₄ (10.053 m² g⁻¹), as confirmed by BET measurements (Fig. S12†).^28–32 Next, the Pt₁Ag_x NCs with an optimal loading of 5 wt% were anchored onto the g-C₃N₄ nanosheet via impregnation (Fig. S13†). Four types of Pt₁Ag_x/g-C₃N₄ nanocomposites labeled Pt₁Ag₉/g-C₃N₄, Pt₁Ag₁₁/g-C₃N₄, Pt₁Ag₁₃/g-C₃N₄, and Pt₁Ag₁₄/g-C₃N₄ were prepared using this method. Inductively coupled plasma-atomic emission spectroscopy (ICP-AES) was utilized for further analysis, unambiguously identifying the similar Pt and Ag contents of the four types of Pt₁Ag_x/g-C₃N₄ nanocomposite (Table S5†). Moreover, TEM images and energy-dispersive X-ray spectroscopy (EDX) elemental mapping confirmed the presence and homogeneous distribution of the Pt₁Ag_x NCs on the g-C₃N₄ nanosheet (Fig. S14†). The Pt₁Ag_x/g-C₃N₄ nanocomposites were further analyzed using X-ray diffraction (XRD). As shown in Fig. S15a,† the two diffraction peaks at 13.0° and 27.7° that correspond to the (100) and (002) diffraction of the typical graphitic interlayer stacking structure of the pure g-C₃N₄ nanosheet did not change in the nanocomposites. Moreover, as shown in Fig. S15b,† the characteristic Fourier-transform infrared (FTIR) peaks located at approximately 810 cm⁻¹, 1200–1650 cm⁻¹, and 3000–3400 cm⁻¹, corresponding to the vibrations of the s-triazine ring unit, C–N heterocycles, and amino groups (N–H), respectively, remained unchanged in the NCs compared with those of pure g-C₃N₄ nanosheet. These results indicate that the basic backbone of the g-C₃N₄ nanosheet was still well-maintained after the addition of the Pt₁Ag_x NCs. As shown in Fig. S15c,† the optical absorption intensity of Pt₁Ag_x/g-C₃N₄ was enhanced in the full spectrum compared with that of the pure g-C₃N₄ nanosheet. Moreover, according to the Kubelka–Munk function, the corresponding intrinsic bandgap of the g-C₃N₄ nanosheet, Pt₁Ag₁₄/g-C₃N₄, Pt₁Ag₁₃/g-C₃N₄, Pt₁Ag₉/g-C₃N₄, and Pt₁Ag₁₁/g-C₃N₄ were estimated to be 2.76, 2.75, 2.73, 2.71 and 2.69 eV, respectively (Fig. S15d†). The enhanced optical absorption intensity and narrowed bandgap of the as-prepared Pt₁Ag₁₁/g-C₃N₄ contributes to the generation of more photogenerated carriers. The hydrogen evolution performance of the four types of Pt₁Ag_x/g-C₃N₄ (x = 9, 11, 13, and 14) nanocomposites were then obtained from the full solar spectrum.

As shown in Fig. 3a, when 10 wt% triethanolamine (TEOA) was used as the sacrificial agent, Pt₁Ag₁₁/g-C₃N₄ displayed the highest hydrogen production performance (1780 μmol g⁻¹ h⁻¹), which was 33.3 times that of the pure g-C₃N₄ nanosheet (53.4 μmol g⁻¹ h⁻¹) and 27.9 times that of Pt₁Ag₁₁ (63.8 μmol g⁻¹ h⁻¹), as well as outperforming several other corresponding counterparts (Table S6†). To exclude the reaction of the NCs with the sacrificial agent, we performed a set of light–dark contrast tests, which indicated that no hydrogen was produced in the absence of light (Fig. S16†). Moreover, the hydrogen production rate of Pt₁Ag₁₁/g-C₃N₄ was ∼2.83 times higher than that of Pt₁Ag₉/g-C₃N₄ (628.4 μmol g⁻¹ h⁻¹) and much higher than that of Pt₁Ag₁₃/g-C₃N₄ (448 μmol g⁻¹ h⁻¹) and Pt₁Ag₁₄/g-C₃N₄ (185.2 μmol g⁻¹ h⁻¹) (Fig. 3b and Table S7†), resulting in the following activity trend; Pt₁Ag₁₁/g-C₃N₄ > Pt₁Ag₉/g-C₃N₄ > Pt₁Ag₁₃/g-C₃N₄ > Pt₁Ag₁₄/g-C₃N₄ > g-C₃N₄ nanosheet. The apparent quantum efficiency of Pt₁Ag₁₁/g-C₃N₄ under light irradiation at 420 nm was ∼3.11% (Table S8†). In cycling experiments (Fig. 3c), the hydrogen production rate of Pt₁Ag₁₁/g-C₃N₄ was maintained for three photocatalytic cycles of 9 h. In the fourth cycle, a slight decrease was observed, which may be attributed to shedding of the NCs owing to constant agitation during the cycle reaction. To accurately characterize the changes in the catalysts before and after the reaction, UV-vis diffuse reflection spectroscopy was performed and the characteristic peak position did not change and the intensity decreased slightly, indicating that the overall structure of the catalyst remained unchanged and the structure of Pt₁Ag₁₁ was well maintained (Fig. S17†). In addition, the effect of the ligand on the photocatalytic performance was excluded by obtaining the time-dependent H₂ production profiles for various Pt₁Ag₁₄ NCs with different ligands (Fig. S18†). As shown in Fig. S19,† the photocatalytic hydrogen production rates of Pt₁Ag₁₄-1/g-C₃N₄, Pt₁Ag₁₄-2/g-C₃N₄, and Pt₁Ag₁₄/g-C₃N₄ were 129.7, 149.8, and 185.2 μmol g⁻¹ h⁻¹, respectively. These findings indicated that the influence of the ligand on the photocatalytic performance was minimal.³³

Fig. 3 (a) Photocatalytic H₂ production of Pt₁Ag₁₁, g-C₃N₄ and Pt₁Ag₁₁/g-C₃N₄. (b) Photocatalytic H₂ production of Pt₁Ag_x/g-C₃N₄ (x = 9, 11, 13, and 14). (c) Cycling runs of the Pt₁Ag₁₁/g-C₃N₄ catalyst under light irradiation. (d) Transient photocurrent response and (e) EIS of Pt₁Ag_x/g-C₃N₄ and g-C₃N₄. (f) EPR spectra of ˙OH and ˙O₂⁻ on Pt₁Ag₁₁/g-C₃N₄ under light irradiation or in the dark (x = 9, 11, 13, and 14).

To better understand the superior photocatalytic hydrogen evolution performance of the Pt₁Ag₁₁/g-C₃N₄ catalysts, a series of photo-electrochemistry tests were performed to probe the electron–hole separation and electron transfer behavior of the nanocomposite photocatalyst.^34–39 As shown in Fig. 3d, Pt₁Ag₁₁/g-C₃N₄ exhibited the strongest transient photocurrent response and the response displayed the following trend under light irradiation, Pt₁Ag₁₁/g-C₃N₄ > Pt₁Ag₉/g-C₃N₄ > Pt₁Ag₁₃/g-C₃N₄ > Pt₁Ag₁₄/g-C₃N₄ > g-C₃N₄ nanosheet, suggesting the promoted efficient separation of the electron–hole pairs in the nanocomposite photocatalysts. This was further confirmed by electrochemical impedance spectroscopy (EIS). The Nyquist curve shown in Fig. 3e indicated that the interfacial charge transfer resistance of Pt₁Ag₁₁/g-C₃N₄ was much smaller than those of g-C₃N₄ and the other Pt₁Ag_x/g-C₃N₄ (x = 9, 13, and 14) nanocomposites. Moreover, steady-state photoluminescence (PL) spectra obtained via excitation at 373 nm revealed a broadband centered at 450 nm for the g-C₃N₄ nanosheet and Pt₁Ag_x/g-C₃N₄ nanocomposites with significantly different PL intensities, indicating that the recombination of photogenerated carriers in Pt₁Ag_x/g-C₃N₄ can be efficiently restrained (Fig. S20†). In addition, time-resolved PL spectra (Fig. S21†) were obtained and the fitting of the time-resolved PL spectra determined the PL lifetimes to be 2.089 ns (g-C₃N₄ nanosheet), 2.015 ns (Pt₁Ag₁₄/g-C₃N₄), 1.998 ns (Pt₁Ag₁₃/g-C₃N₄), 1.798 ns (Pt₁Ag₁₁/g-C₃N₄), and 1.619 ns (Pt₁Ag₉/g-C₃N₄). The reduced decay lifetime of Pt₁Ag_x/g-C₃N₄ signified the promotion of efficient separation of electron–hole pairs and an accelerated charge transport process. In summary, the order of charge transfer and separation efficiency of all the photocatalysts closely aligned with the order of the photocatalytic evolution performance.

To understand the photogenerated charge transfer pathways in the Pt₁Ag_x/g-C₃N₄ nanocomposites, the band structure of the g-C₃N₄ nanosheet and Pt₁Ag_x NCs were evaluated. The band gap of 2.76 eV for the g-C₃N₄ nanosheet was determined from the UV-vis absorption spectrum; the conduction band (CB) level was estimated by Mott–Schottky measurements to be −0.31 V vs. NHE (PH = 7, Fig. S22†). Because the metal NCs can be regarded as small band gap semiconductors, the energy band structure of the Pt₁Ag_x NCs was determined in the same way as that of the g-C₃N₄ nanosheet (Fig. S23 and S24†).^10,13 The energy diagram showed that all the lowest unoccupied molecular orbitals (LUMO) of the Pt₁Ag_x NCs were slightly more negative than the CB of g-C₃N₄, suggesting the formation of a type-II or Z-scheme heterojunction (Fig. S25†). Taking Pt₁Ag₁₁/g-C₃N₄ as a model, EPR analysis was performed to distinguish the exact types of heterostructures. To produce ˙O₂⁻ radicals, the potential of the photogenerated electrons in the CB should be more negative than −0.33 V. Meanwhile, the potential of the photogenerated holes in the valence band (VB) should be more positive than 2.40 V to produce ˙OH radicals. Apparently, ˙O₂⁻ radicals and ˙OH radicals could not be produced by the type-II heterojunction due to the high potential of the photogenerated electrons of −0.31 V and the low potential of the photogenerated holes of 0.95 V. However, as shown in Fig. 3f, under light irradiation, the production of ˙O₂⁻ radicals and ˙OH radicals was observed, demonstrating the transfer of the photogenerated electrons from g-C₃N₄ to Pt₁Ag₁₁, forming a Z-scheme heterojunction in Pt₁Ag₁₁/g-C₃N₄.^40,41

To gain a molecular level understanding of the effect of the kernel structural defects on photocatalytic activity, density-functional theory (DFT) calculations using the Vienna ab initio simulation package (VASP) were conducted.^42,43 After geometric relaxation, we concluded that the Pt₁Ag_x NCs could maintain a stable structure on the surface of the g-C₃N₄ monolayer (Fig. S26†), which might be attributed to the formation of a Ag–N covalent bond and abundant π–π interactions between Pt₁Ag_x NC and g-C₃N₄. The binding energy (E_bind) of the NCs with the g-C₃N₄ monolayer can also be obtained. The E_bind value of Pt₁Ag₁₁, Pt₁Ag₉, Pt₁Ag₁₃ and Pt₁Ag₁₄ were calculated as −0.95 eV, −0.72 eV, −0.65 eV and −0.68 eV, respectively, where the high E_bind (−0.95 eV) value of Pt₁Ag₁₁ enables it to exhibit the most stable loading structure on the g-C₃N₄ surface compared with the other NCs (Fig. 4a). The differential charge density was analyzed to determine the charge transfer between the NCs and monolayers. As shown in Fig. 4b, the g-C₃N₄ layer accumulated negative charge, while the NCs gathered positive charge, thus forming an internal electric field directing from the NCs to g-C₃N₄. Bader⁴⁴ charge analysis showed that the Pt₁Ag₉ (0.66|e|) and Pt₁Ag₁₁ (0.82|e|) NCs transferred more charge to g-C₃N₄ than was transferred by Pt₁Ag₁₃ (0.48|e|) and Pt₁Ag₁₄ (0.62|e|). This difference indicates a stronger internal electric field in Pt₁Ag₁₁/g-C₃N₄, allowing Pt₁Ag₁₁ to interact more closely with g-C₃N₄. Furthermore, the photocatalytic hydrogen evolution reaction was investigated with ΔG as an indicator of the absorption of H* intermediates on the Ag sites in the Pt₁Ag_x/g-C₃N₄ nanocomposites. As shown in Fig. 4c, Pt₁Ag₁₁/g-C₃N₄ has the lowest Gibbs free energy change (ΔG) (0.23 eV), following the trend of Pt₁Ag₁₁/g-C₃N₄ < Pt₁Ag₉/g-C₃N₄ (0.29 eV) < Pt₁Ag₁₃/g-C₃N₄ (0.33 eV) < Pt₁Ag₁₄/g-C₃N₄ (0.67 eV), which suggests optimal catalytic activity on the surface of the icosahedron with atomic defects. Moreover, the d-band center of the Ag atoms in Pt₁Ag₁₁ are closer to the Fermi level, which suggests a stronger binding interaction between Pt₁Ag₁₁/g-C₃N₄ and the H species (Fig. S27†).

Fig. 4 (a) Binding energies of the Pt₁Ag_x nanoclusters loaded on a g-C₃N₄ monolayer. (b) Differential charge density of nanoclusters loaded on a g-C₃N₄ monolayer, where yellow and cyan areas represent electron accumulation and depletion, respectively. The total amount of charge transfer is given. (c) The calculated Gibbs free energy (ΔG_H*) diagrams for the photocatalytic hydrogen evolution reaction on Pt₁Ag_x/g-C₃N₄. The optimized structures of H* adsorption on Pt₁Ag₉/g-C₃N₄ (blue), Pt₁Ag₁₁/g-C₃N₄ (yellow), Pt₁Ag₁₃/g-C₃N₄ (green), and Pt₁Ag₁₄/g-C₃N₄ (purple). (d) The energy level diagrams and the interface charge transfer route in Pt₁Ag₁₁/g-C₃N₄. An internal electric field (IEF) is formed inside the heterojunction formed by g-C₃N₄ loaded with Pt₁Ag₁₁. E_vacuum, E_f1, E_f2 and E_f3 represent the vacuum level and Fermi levels of g-C₃N₄, Pt₁Ag₁₁ and Pt₁Ag₁₁/g-C₃N₄, respectively. VB and CB represent the top of the valence band and the bottom of the conduction band of g-C₃N₄. HOMO and LUMO refer to the highest occupied molecular orbital and the lowest unoccupied molecular orbital of the electrons in the Pt₁Ag₁₁ cluster, respectively.

Building on this, we explored the mechanism of photocatalytic water splitting for hydrogen production (Fig. 4d). When NCs are loaded onto the surface of g-C₃N₄, free electrons transfer across the interface due to the differing Fermi levels.^45,46 Specifically, electrons from Pt₁Ag₁₁, which has a higher Fermi level, migrate to g-C₃N₄, which has a lower Fermi level, until equilibrium is reached at a common Fermi level. This observation aligns with the differential charge density results shown in Fig. 4b. Under the influence of the Coulomb force, the band edge in g-C₃N₄ at the interface bent downward, and the energy level of Pt₁Ag₁₁ bent upward. When illuminated, electrons in the VB of the g-C₃N₄ within the Pt₁Ag₁₁/g-C₃N₄ heterojunctions absorbs the energy and transits it to the CB, leaving behind holes in the VB. Simultaneously, the HOMO electrons of Pt₁Ag₁₁ gained energy to transition to the LUMO, resulting in a vacant HOMO. Driven by the internal electric field, the photogenerated electrons in g-C₃N₄ were transferred to Pt₁Ag₁₁. The band (and energy level) bending are crucial for the migration of photogenerated carriers under optical excitation.⁴⁷ The downward band bending in g-C₃N₄ facilitates the free flow of electrons while inhibiting hole outflow. Conversely, the upward energy level bending in Pt₁Ag₁₁ suppresses electron de-excitation. Thus, electrons flowing from the CB of g-C₃N₄ recombine with the vacant orbitals in the HOMO of Pt₁Ag₁₁, allowing the electrons in the LUMO of Pt₁Ag₁₁ to participate in the hydrogen evolution reaction. This process enables direct Z-scheme photocatalytic water splitting for hydrogen production.

4 Conclusions

A family of Pt₁Ag_x (x = 9, 11, 13, and 14) NCs, including the new Pt₁Ag₁₁ NC, was designed and synthesized via ligand engineering. Single-crystal X-ray diffraction revealed that the crystal structures of these NCs possessed defect-containing icosahedral kernels where one or three vertical Ag atoms are lost. Using the photocatalytic water-splitting reaction as a model, the positive effect of the kernel structural defect on the photocatalytic activity was elucidated, and the mechanism that links the defect structure and catalytic activity was well established. Importantly, Pt₁Ag₁₁/g-C₃N₄ has the highest hydrogen production rate reaching 1780 μmol g^−1.h⁻¹, that is, ∼2.83, ∼3.97, ∼9.61, and ∼33.3 times higher than that of Pt₁Ag₉/g-C₃N_4, Pt₁Ag₁₃/g-C₃N₄, Pt₁Ag₁₄/g-C₃N₄ and the g-C₃N₄ nanosheets, respectively. DFT calculations further demonstrated the importance of the defect-containing kernel structure for the activation of H₂, electron–hole separation, and the charge transfer efficiency of the Pt₁Ag₁₁/g-C₃N₄ photocatalyst. These improved properties which might be attributed to the formation of Z-scheme heterojunctions, accounting for their very high photocatalytic H₂ production rate. This study lays a strong foundation for the design of highly efficient and precise nanocomposite photocatalysts and provides new insights into the mechanism of photocatalytic hydrogen production at the atomic level.

Abbreviations

Pt1Ag11	Pt1Ag11(SR)5(P(Ph-OMe)3)7, (SR is 2,3,5,6-tetrafluorothiophenol and P(Ph-OMe)3 is tris(4-methylphenyl)-phosphine)
Pt1Ag9	Pt1Ag9(P(Ph-F)3)7Cl3, (P(Ph-F)3 is tris(4-fluorophenyl)phosphine)
Pt1Ag13	Pt1Ag13(PFBT)6(PPh3)7, (PFBT is pentafluorobenzenethiol and PPh3 is triphenylphosphine)
Pt1Ag14	Pt1Ag14(SR)6(PPh3)8, (SR is 2-chloro-4-fluorobenzenethiol and PPh3 is triphenylphosphine)
Pt1Ag14-1	Pt1Ag14(SR)6(P(Ph-F)3)8, (SR is 2,3,5,6-tetrafluorothiophenol and P(Ph-F)3 is tris(4-fluorophenyl)phosphine)
Pt1Ag14-2	Pt1Ag14(SR)6(PPh3)8, (SR is 2,3,5,6-tetrafluorothiophenol and PPh3 is triphenylphosphine)

Data availability

CCDC 2380779 contain the supplementary crystallographic data for this paper. These data can be obtained free of charge via https://www.ccdc.cam.ac.uk/data_request/cif, or by emailing E-mail: data_request@ccdc.cam.ac.uk, or by contacting The Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax: +44 1223 336033.

Author contributions

D. Hu, S. Sun and M. Zhu conceived and designed the project. D. Tan and T. Ding carried out the synthesis and catalytic experiments. S. Jin and D. Tan carried out crystallography and mass spectrometry. K. Shen and C. Xu conducted theoretical calculations. All authors contributed to the discussion of the results and the preparation of the manuscript.

Conflicts of interest

The authors declare no conflict of interest.

Acknowledgements

We acknowledge the financial support provided by the National Natural Science Foundation of China (22371003, 21871001, 22103001) and the Natural Science Fund of the Education Department of Anhui Province (2023AH050108).

Notes and references

1. Y. Shang, H. Fan, X. Che and W. Wang, Promoted photocatalytic hydrogen evolution via double-electron migration in Ag@g-C₃N₄ heterojunction, Int. J. Hydrogen Energy, 2023, 48, 17370–17382.
2. Y. Han, C. Wang, R. Zhao, J. Han and L. Wang, Ni-doped CdSe/ZnSnO₃ double-shell nanocubes heterojunction for efficient photocatalytic hydrogen evolution, Fuel, 2023, 353, 129247.
3. P. Zhou, I. A. Navid, Y. Ma, Y. Xiao, P. Wang, Z. Ye, B. Zhou, K. Sun and Z. Mi, Solar-to-hydrogen efficiency of more than 9% in photocatalytic water splitting, Nature, 2023, 613, 66–70.
4. A. Shu, C. Qin, M. Li, L. Zhao, Z. Shangguan, Z. Shu, X. Yuan, M. Zhu, Y. Wu and H. Wang, Electric effects reinforce charge carrier behaviour for photocatalysis, Energy Environ. Sci., 2024, 17, 4907.
5. C. Xia, H. Wang, J. K. Kim and J. Wang, Rational Design of Metal Oxide-Based Heterostructure for Efficient Photocatalytic and Photoelectrochemical Systems, Adv. Funct. Mater., 2021, 31, 2008247.
6. A. Meng, W. Tian, H. Yang, X. Wang, X. Wang and Z. Li, Molybdenum sulfide-modified metal-free graphitic carbon nitride/black phosphorus photocatalyst synthesized via high-energy ball-milling for efficient hydrogen evolution and hexavalent chromium reduction, J. Hazard. Mater., 2021, 413, 2008247.
7. C. Fang, C. Xu, W. Zhang, M. Zhou, D. Tan, L. Qian, D. Hu, S. Jin and M. Zhu, Dual-quartet phosphorescent emission in the open-shell M₁Ag₁₃ (M = Pt, Pd) nanoclusters, Nat. Commun., 2024, 15, 5962.
8. Y. Du, H. Sheng, D. Astruc and M. Zhu, Atomically Precise Noble Metal Nanoclusters as Efficient Catalysts: A Bridge between Structure and Properties, Chem. Rev., 2020, 120, 526–622.
9. Y. Wang, X.-H. Liu, R. Wang, B. Cula, Z.-N. Chen, Q. Chen, N. Koch and N. Pinna, Secondary Phosphine Oxide Functionalized Gold Clusters and Their Application in Photoelectrocatalytic Hydrogenation Reactions, J. Am. Chem. Soc., 2021, 143, 9595–9600.
10. H. Wang, X. Zhang, W. Zhang, M. Zhou and H.-L. Jiang, Heteroatom-Doped Ag₂₅ Nanoclusters Encapsulated in Metal-Organic Frameworks for Photocatalytic Hydrogen Production, Angew. Chem., Int. Ed., 2024, e202401443.
11. C. Wang, P. Lv, D. Xue, Y. Cai, X. Yan, L. Xu, J. Fang and Y. Yang, Zero-Dimensional/Two-Dimensional Au₂₅(Cys)₁₈ Nanoclusters/g-C₃N₄ Nanosheets Composites for Enhanced Photocatalytic Hydrogen Production under Visible Light, ACS Sustain. Chem. Eng., 2018, 6, 8447–8457.
12. Y. Negishi, M. Mizuno, M. Hirayama, M. Omatoi, T. Takayama, A. Iwasea and A. Kudo, Enhanced photocatalytic water splitting by BaLa₄Ti₄O₁₅ loaded with ∼1 nm gold nanoclusters using glutathione-protected Au₂₅ clusters, Nanoscale, 2013, 5, 7188.
13. Y. Du, C. Li, Y. Dai, H. Yin and M. Zhu, Recent progress in atomically precise metal nanoclusters for photocatalytic application, Nanoscale Horiz., 2024, 9, 1262.
14. X. Lu, A. Tong, D. Luo, F. Jiang, J. Wei, Y. Huang, Z. Jiang, Z. Lu and Y. Ni, Confining single Pt atoms from Pt clusters on multi-armed CdS for enhanced photocatalytic hydrogen evolution, J. Mater. Chem. A, 2022, 10, 4594.
15. Y. Wang, X.-H. Liu, Q. Wang, M. Quick, S. A. Kovalenko, Q.-Y. Chen, N. Koch and N. Pinna, Insights into Charge Transfer at an Atomically Precise Nanocluster/Semiconductor Interface, Angew. Chem., Int. Ed., 2020, 59, 7748–7754.
16. X. Huang, K. Yin, S. Zhang, T. Wu, Y. Yuan, X. Wang, Y. Jia, Z. Xiao, J. Gu and D. Wang, Interfacial chemical bond-modulated Z-scheme Cs₂AgBiBr₆/WO₃ enables stable and highly efficient photocatalysis, Appl. Surf. Sci., 2023, 637, 157877.
17. J. Wang, X. Li, Y. You, X. Yang, Y. Wang and Q. Li, Interfacial coupling induced direct Z-scheme water splitting in metal-free photocatalyst: C₃N/g-C₃N₄ heterojunctions, Nanotechnology, 2018, 29, 365401.
18. Q. Zhu, H. Shen, C. Han, L. Huang, Y. Zhou, Y. Du, X. Kang and M. Zhu, Rationally construction of atomic-precise interfacial charge transfer channel and strong build-in electric field in nanocluster-based Z-scheme heterojunctions with enhanced photocatalytic hydrogen production, Nano Res., 2024, 17, 5002–5010.
19. Y. Du, J. Li, X. Ma and Q. Guo, Visible light driven mesoporous ag/Ag₂CrO₄/g-C₃N₄ degrades multiple organic pollutants efficiently: synthesis, mechanism, and degradation pathway, Vacuum, 2024, 226, 113323.
20. S. O. Lee, S. K. Lakhera and K. Yong, Strategies to Enhance Interfacial Spatial Charge Separation for High-Efficiency Photocatalytic Overall Water-Splitting: A Review, Adv. Energ. Sust. Res., 2003, 4, 2300130.
21. S. Nematulloev, A. Sagadevan, B. Alamer, A. Shkurenko, R. Huang, J. Yin, C. Dong, P. Yuan, K. E. Yorov, A. A. Karluk, W. J. Mir, B. E. Hasanov, M. N. Hedhili, N. M. Halappa, M. Eddaoudi, O. F. Mohammed, M. Rueping and O. M. Bakr, Atomically Precise Defective Copper Nanocluster Catalysts for Highly Selective C-C Cross-Coupling Reactions, Angew. Chem., Int. Ed., 2023, 62, e202303572.
22. R. P. B. Silalahi, Y. Jo, J.-H. Liao, T.-H. Chiu, E. Park, W. Choi, H. Liang, S. Kahlal, J.-Y. Saillard, D. Lee and C. W. Liu, Hydride-containing 2-Electron Pd/Cu Superatoms as Catalysts for Efficient Electrochemical Hydrogen Evolution, Angew. Chem., Int. Ed., 2023, 62, e202301272.
23. D. Zhu and Q. Zhou, Nitrogen doped g-C₃N₄ with the extremely narrow band gap for excellent photocatalytic activities under visible light, Appl. Catal., B, 2021, 281, 119474.
24. Z. Yu, Y. Li, A. Torres-Pinto, A. P. LaGrow, V. M. Diaconescu, L. Simonelli, M. J. Sampaio, O. Bondarchuk, I. Amorim, A. Araujo, A. M. T. Silva, C. G. Silva, J. L. Faria and L. Liu, Single-atom Ir and Ru anchored on graphitic carbon nitride for efficient and stable electrocatalytic/photocatalytic hydrogen evolution, Appl. Catal., B, 2022, 310, 121318.
25. W. Sun, S. Jin, W. Du, X. Kang, A. Chen, S. Wang, H. Sheng and M. Zhu, Total Structure Determination of the Pt₁Ag₉[P(Ph-F)₃]₇Cl₃ Nanocluster, Eur. J. Inorg. Chem., 2020, 590–594.
26. X. Lin, K. Sun, X. Fu, X. Ren, Y. Yang, C. Liu and J. Huang, Correlating Kernel−Shell Structures with Optical Properties of Pt₁Ag₂₄ and Pt₁Ag₁₄ Nanoclusters, J. Phys. Chem. C, 2021, 125, 2194–2201.
27. A. K. Das, S. Mukherjee, S. S. R, A. S. Nair, S. Bhandary, D. Chopra, D. Sanyal, B. Pathak and S. Mandal, Defects Engineering on Ceria and C−C Coupling Reactions Using [Au₁₁(PPh₃)₇I₃] Nanocluster: A Combined Experimental and Theoretical Study, ACS Nano, 2020, 14, 16681–16688.
28. P. Niu, L. Zhang, G. Liu and H.-M. Cheng, Graphene-Like Carbon Nitride Nanosheets for Improved Photocatalytic Activities, Adv. Funct. Mater., 2012, 22, 4763–4770.
29. W. Yan, L. Yan and C. Jing, Impact of doped metals on urea-derived g-C₃N₄ for photocatalytic degradation of antibiotics: structure, photoactivity and degradation mechanisms, Appl. Catal., B, 2019, 244, 475–485.
30. X. Zhang, F. Wu, G. Li, L. Wang, J. Huang, A. Song, A. Meng and Z. Li, Mechanistic insight into the synergy between platinum cluster and indium particle dual cocatalysts for enhanced photocatalytic water splitting, J. Colloid Interface Sci., 2024, 670, 774–784.
31. C. Huang, Y. Wen, J. Ma, D. Dong, Y. Shen, S. Liu, H. Ma and Y. Zhang, Unraveling fundamental active units in carbon nitride for photocatalytic oxidation reactions, Nat. Commun., 2021, 12, 320.
32. X. Zhang, F. Wu, G. Li, L. Wang, J. Huang, A. Song, A. Meng and Z. Li, Construction of intramolecular donor-acceptor type carbon nitride for photocatalytic hydrogen production, J. Colloid Interface Sci., 2024, 655, 439–450.
33. A. Ma, Y. Ren, Y. Zuo, J. Wang, S. Huang, X. Ma and S. Wang, Ligand-controlled exposure of active sites on the Pd₁Ag₁₄ nanocluster surface to boost electrocatalytic CO₂ reduction, Chem. Commun., 2024, 60, 3162–3165.
34. M. Xu, D. Li, K. Sun, L. Jiao, C. Xie, C. Ding and H.-L. Jiang, Interfacial Microenvironment Modulation Boosting Electron Transfer between Metal Nanoparticles and MOFs for Enhanced Photocatalysis, Angew. Chem., Int. Ed., 2021, 60, 16372–16376.
35. M. Xu, X. Ruan, D. Meng, G. Fang, D. Jiao, S. Zhao, Z. Liu, Z. Jiang, K. Ba, T. Xie, W. Zhang, J. Leng, S. Jin, S. K. Ravi and X. Cui, Modulation of Sulfur Vacancies in ZnIn₂S₄/MXene Schottky Heterojunction Photocatalyst Promotes Hydrogen Evolution, Adv. Funct. Mater., 2024, 34, 2402330.
36. X. Yang, Y. Zhang, J. Deng, X. Huo, Y. Wang and R. Jia, Fabrication of Porous Hydrophilic CN/PANI Heterojunction Film for High-Efficiency Photocatalytic H₂ Evolution, Catalysts, 2023, 13(1), 139.
37. L. Mao, B. Zhai, J. Shi, X. Kang, B. Lu, Y. Liu, C. Cheng, H. Jin, E. Lichtfouse and L. Guo, Supercritical CH₃OH-Triggered Isotype Heterojunction and Groups in g-C₃N₄ for Enhanced Photocatalytic H₂ Evolution, ACS Nano, 2024, 18, 13939–13949.
38. L. Wang, Y. Li, Y. Ai, E. Fan, F. Zhang, W. Zhang, G. Shao and P. Zhang, Tracking Heterogeneous Interface Charge Reverse Separation in SrTiO₃/NiO/NiS Nanofibers with In Situ Irradiation XPS, Adv. Funct. Mater., 2023, 33, 2306466.
39. F. Xing, C. Wang, S. Liu, S. Jin, H. Jin and J. Li, Interfacial Chemical Bond Engineering in a Direct Z-Scheme g-C₃N₄/MoS₂ Heterojunction, ACS Appl. Mater. Interfaces, 2023, 15, 11731–11740.
40. H. Wang, Z. Chen, Y. Shang, C. Lv, X. Zhang, F. Li, Q. Huang, X. Liu, W. Liu, L. Zhao, L. Ye, H. Xie and X. Jin, Boosting Carrier Separation on a BiOBr/Bi₄O₅Br₂ Direct Z-Scheme Heterojunction for Superior Photocatalytic Nitrogen Fixation, ACS Catal., 2024, 14, 5779–5787.
41. H. Shi, C. Li, L. Wang, W. Wang, J. Bian and X. Meng, Photocatalytic reduction of nitrate pollutants by novel Z-scheme ZnSe/BiVO₄ heterostructures with high N₂ selectivity, Sep. Purif. Technol., 2022, 300, 121854.
42. G. Kresse and J. Hafner, Ab initio molecular dynamics for liquid metals, Phys. Rev. B:Condens. Matter Mater. Phys., 1993, 47, 558–561.
43. G. Kresse and J. Furthmuller, Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set, Phys. Rev. B:Condens. Matter Mater. Phys., 1996, 54, 11169–11186.
44. W. Tang, E. Sanville and G. Henkelman, A grid-based Bader analysis algorithm without lattice bias, J. Phys. Condens. Matter, 2009, 21, 084204.
45. L. Xu, M. Yang, S. J. Wang and Y. P. Feng, Electronic and optical properties of the monolayer group-IV monochalcogenides M X (M = Ge, Sn; X = S, Se, Te), Phys. Rev. B, 2017, 95, 235434.
46. S. Grimme, J. Antony, S. Ehrlich and H. Krieg, A consistent and accurate ab initio parametrization of density functional dispersion correction (DFT-D) for the 94 elements H-Pu, J. Chem. Phys., 2010, 132, 154104.
47. X. Jia, J. Wang, Y. Lu, J. Sun, Y. Li, Y. Wang and J. Zhang, Designing SnS/MoS₂ van der Waals heterojunction for direct Z-scheme photocatalytic overall water-splitting by DFT investigation, Phys. Chem. Chem. Phys., 2022, 24, 21321–21330.

---

Footnotes

† Electronic supplementary information (ESI) available. CCDC 2380779. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d5sc01735a

‡ D. Tan, T. Ding and K. Shen contributed equally to this work.

---