Energy crises and environmental pollution are two major challenges facing the contemporary world that force people to move away from their dependence on fossil fuels [1-3]. Hydrogen (H₂) is a source of renewable green energy that is an attractive substitute for fossil fuels. To realize the effective use of H₂ energy, it is necessary to develop H₂ production technologies that are clean, cheap, efficient, and large-scale [4, 5]. Photocatalytic H₂ evolution can translate solar energy into chemical energy and has the advantages of high product purity, carbon-free environmental friendliness, simplicity, and low energy consumption, thereby making it a promising H₂ production technology [6-8]. However, photocatalytic H₂ evolution technology is currently limited by the low efficiency of photocatalytic H₂ evolution, poor stability of the photocatalyst, and high cost of H₂ evolution [9, 10]. Therefore, developing affordable, easily prepared photocatalysts for photocatalytic H₂ evolution has attracted much attention. Metal-free semiconductor materials are desirable to realize solar energy conversion without using toxic heavy metals [11].

A representative metal-free semiconductor photocatalyst is graphitic carbon nitride (g-C₃N₄), which has excellent physical and chemical properties, such as high chemical and thermal stability, commendable photoelectric properties, and powerful antioxidant behavior [12, 13]. Importantly, g-C₃N₄ can make use of visible light because of its suitable band gap [14]. Thus, g-C₃N₄ has been widely applied in photocatalytic degradation, air purification, photocatalytic water splitting, and photocatalytic carbon dioxide reduction [15, 16]. However, bulk g-C₃N₄ possesses several disadvantages, such as low specific surface area, a high recombination rate of photogenerated electron-hole pairs, and sluggish reaction kinetics [17-19]. These disadvantages severely limit the photocatalytic activity of bulk g-C₃N₄, so it is necessary to improve its photocatalytic performance by modification strategies such as nanostructure formation, molecular doping, construction of heterojunctions, and elemental doping [20-22]. One approach to improve the photocatalytic efficiency of g-C₃N₄ is to exfoliate bulk g-C₃N₄ into monolayer g-C₃N₄ nanosheets [23]. At present, there are two main exfoliation methods of g-C₃N₄: liquid exfoliation and thermal exfoliation [24]. Two-dimensional (2D) g-C₃N₄ nanosheets possess a larger specific surface area with more active sites, higher separation efficiency of photogenerated electron-hole pairs, and increased electron transport compared with the corresponding properties of bulk g-C₃N₄ [25].

Although 2D g-C₃N₄ nanosheets display greatly improved photocatalytic performance over that of bulk g-C₃N₄, their performance still does not meet the expectations of researchers. Therefore, massive effort has been devoted to optimizing the photocatalytic ability of 2D g-C₃N₄ nanosheets [26]. Doping 2D g-C₃N₄ nanosheets with a plasmonic metal (like Au, Ag, Pd, Pt, Rh, or Ir) could further enhance its ability to absorb visible light through the surface plasmon resonance (SPR) effect, which is one of the most common and effective strategies to enhance photocatalytic activity [27-30]. As many researchers have reported [31-39], when a plasmonic metal is coupled with a semiconductor, hot electrons are generated in the plasmonic metal by overcoming the Schottky barrier under visible-light irradiation. Then, a high density of hot electrons flows into the conduction band (CB) of the semiconductor to trigger the reduction reaction, resulting in increased photocatalytic activity. Previously, our group designed a plasmonic composite consisting of Ag nanoparticles supported on monolayer g-C₃N₄ nanosheets, which verified the role of the SPR effect in increasing photocatalytic activity [40]. However, the photocatalytic performance of the Ag/2D g-C₃N₄ composites was still less than ideal.

The purpose of this work is to greatly enhance the photocatalytic performance of monolayer g-C₃N₄ and steer the flow of charge carriers by dispersing Au plasmonic nanoparticles on the surface of monolayer g-C₃N₄ nanosheets. The H₂ evolution rate of the resulting Au/monolayer g-C₃N₄ composites is evaluated. The photocatalytic mechanism of the composites is investigated by photocurrent tests.

The monolayer g-C₃N₄ was prepared according to our previous report [1]. Melamine (2 g) was placed in a crucible and then annealed at 550 ℃ for 4 h after heating at a rate of 2 ℃/min in a muffle furnace. The obtained sample was ground into a powder and then annealed again at 550 ℃ under the same conditions, the obtained white sample was monolayer g-C₃N₄.

Monolayer g-C₃N₄ (0.1 g) was dissolved in water (10 mL) under sonication. NH₃·H₂O (2 mL), Na₂SO₃ (0.01 g), and isooctane (10 mL) were added to the mixture. The stirred mixture was heated in an oil bath (95 ℃) for 10 min, and then HAuCl₄·4H₂O (0.0011, 0.0021, or 0.0042 g) was added dropwise. The mixture was stirred at 95 ℃ for 1 h. The sample was collected by centrifugal separation and then washed several times with pure water. The Au/monolayer g-C₃N₄ composites were obtained by freeze-drying. The samples produced using different masses of HAuCl₄·4H₂O of 0.0011, 0.0021, and 0.0042 g are denoted as 0.5%, 1%, and 2% Au/monolayer g-C₃N₄, respectively.

The samples were analyzed by X-ray diffraction (XRD) by Bruker D8 diffractometer with Cu K_α radiation (λ = 1.5418 Å ) in the range of 2θ = 10°−80°. The structural information for samples was measured by Fourier transform infrared spectroscopy (FTIR, Avatar 470, Thermo Nicolet) using the standard KBr disk method. The morphology and structure of the samples were investigated with scanning electron microscope (SEM) and transmission electron microscopy (TEM). The SEM images were taken on a field-emission microscope by a JEOL JSM-7001F. The transmission electron microscopy (TEM) images were collected with a JEOL-JEM-2010 (JEOL, Japan) operated at 200 kV. Elemental compositions were detected by X-ray photoelectron spectroscopy (XPS) analysis which was performed on an ESCALab MKII X-ray photo-electron spectrometer using the Mg K_α radiation. Ultraviolet visible (UV-vis) diffuse reflectance spectrums (DRS) of the samples were measured by using a UV-vis spectrophotometer (Shimadzu UV-2450, Japan) in the range of 200 to 800 nm. BaSO₄ was used as the reflectance standard material. The photoluminescence (PL) spectra of the samples were obtained by a QuantaMaster & TimeMaster Spectrofluorometer with an excitation wavelength at 325 nm. X-band ESR spectra were recorded at ambient temperature on a JES FA200 spectrometer. The settings for the ESR spectrometer were as follows: center field, 336.496 mT; sweep width, 5 mT; microwave frequency, 9.5 GHz; modulation frequency, 100 kHz; power, 0.998 mW.

The photocurrent generated by the samples was tested with an electrochemical analyzer (CHI660B, Chen Hua Instruments, Shanghai, China) using a standard three-electrode configuration including a Pt wire counter electrode, Ag/AgCl reference electrode, and Na₂SO₄ (0.1 mol/L) aqueous solution as the electrolyte. The light source was a 500-W Xe arc lamp. Each sample (5 mg) was dispersed in ethanol (0.5 mL) and ethylene glycol (0.5 mL). Then, 20 μL of each dispersion was drop-cast onto an indium tin oxide (ITO) substrate over a fixed area of 0.5 cm². Each sample-modified ITO electrode was dried under an infrared lamp. A potential of −0.2 V (vs. Ag/AgCl) was used in the photocurrent measurements.

Photocatalytic H₂ production by the samples was tested using an online photocatalytic H₂ production system (LbSolar-3AG, PerfectLight, Beijing). Each sample (0.05 g) was dispersed in an aqueous solution consisting of water (90 mL) and hole sacrificial agent (10 mL). The mixtures were degassed before conducting measurements, which were performed using a 300-W Xe lamp (PLS-SXE 300C (BF), PerfectLight, Beijing) with an optical filter (λ > 400 nm) as the light source. An online gas chromatograph (GC D7900P, thermal conductivity detector, nitrogen carrier gas, 5 Å molecular sieve column, Shanghai Fechcomp) was used to determine the photocatalytic H₂ evolution rate.

The photocatalytic activities of the samples were investigated by their photodegradation of ciprofloxacin (CIP) and methyl orange (MO) under visible-light irradiation. The light source was a 300-W Xe lamp with a 400-nm cutoff filter. Each sample (0.0150 g) was dispersed in CIP or MO aqueous solution (10 mg/L, 50 mL). The reaction used a pump to provide oxygen. To ensure that absorption-desorption equilibrium was reached on the photocatalyst surface, the suspension was stirred for about 30 min in the dark before light exposure. The experiments were conducted under continuous stirring at 30 ℃. Every 30 min, a 3.5 mL aliquot of the suspension was extracted and centrifuged. The supernatant was analyzed using a UV-vis spectrophotometer (UV-2450, Shimadzu, Japan). The contents of pollutants were determined using the maximum absorption wavelengths of CIP of 276 nm and of MO of 463 nm.

Fig. 1(a) shows the photocatalytic performance of the Au/monolayer g-C₃N₄ composites in H₂ evolution from water. The pure monolayer g-C₃N₄ showed low photocatalytic H₂ evolution activity under visible-light irradiation with an H₂ evolution rate of 153 μmol g^‒1 h^‒1. All of the Au/monolayer g-C₃N₄ composites exhibited enhanced photocatalytic activity compared with that of the pure monolayer g-C₃N₄. In particular, the 1% Au/monolayer g-C₃N₄ composite exhibited the highest H₂ evolution rate of 565 μmol g^‒1 h^‒1, which is in line with that of previously reported Au/g-C₃N₄ photocatalysts (Table S1). These results indicated that the introduction of Au plasmonic nanoparticles raised the photocatalytic performance of monolayer g-C₃N₄. The optimal Au content to maximize the H₂ evolution rate was 1%. This result indicates that a moderate amount of Au plasmonic nanoparticles dispersed on the surface of the monolayer g-C₃N₄ nanosheets facilitated charge transport. Although the SPR effect of the 2% Au/monolayer g-C₃N₄ should be enhanced compared with that of the composite with 1% Au, the number of active sites of monolayer g-C₃N₄ will decrease because Au covers more of the surface of monolayer g-C₃N₄. The excess Au nanoparticles do not enhance the migration and separation efficiency of photogenerated electron-hole pairs, and may serve as recombination centers of photogenerated electron-hole pairs, thus lowering the H₂ evolution rate [41-43].

The performance of Au/monolayer g-C₃N₄ and that of previously reported Ag/monolayer g-C₃N₄ are compared in Fig. S1 [40]. The 1% Au/monolayer g-C₃N₄ composite showed a higher H₂ evolution rate (565 μmol g^‒1 h^‒1) than that of 1% Ag/monolayer g-C₃N₄ (173 μmol g^‒1 h^‒1), suggesting that Au/monolayer g-C₃N₄ possessed higher photocatalytic activity than Ag/monolayer g-C₃N₄ at the same doping amount. In addition, the H₂ evolution rate of 1% Au/monolayer g-C₃N₄ was also higher than that of 2% Ag/monolayer g-C₃N₄ (342 μmol g^‒1 h^‒1), which was the optimal doping amount in our previously reported system. Therefore, compared with our previously reported Ag/monolayer g-C₃N₄ composites, Au/monolayer g-C₃N₄ displayed dramatically improved photocatalytic activity at a much lower doping amount.

The photocatalytic performance of Au/monolayer g-C₃N₄ composites was also evaluated by photodegradation of CIP and MO under visible-light irradiation (Fig. 1(c) and (d)). To rule out the photosensitization process of the Au/monolayer g-C₃N₄ composites, CIP was chosen as a model colorless organic pollutant. CIP is a widely used antibiotic. However, CIP in aquatic environments seriously threatens human health and the safety of ecosystems, so the degradation of CIP has become an important issue. As shown in Fig. 1(c), compared with that of the pure monolayer g-C₃N₄, all of the Au/monolayer g-C₃N₄ composites exhibited enhanced photocatalytic activity for CIP degradation under visible-light irradiation. This clearly indicates that the introduction of Au to monolayer g-C₃N₄ enhances its photocatalytic activity. The 1% Au/monolayer g-C₃N₄ exhibited the highest photocatalytic activity of the composites; its photocatalytic degradation efficiency of CIP was about 61% after irradiation for 3.5 h, which was almost 3.4 times higher than that of pure monolayer g-C₃N₄ (18%). The photocatalytic degradation efficiencies of 0.5% and 2% Au/monolayer g-C₃N₄ were about 41% and 36% after irradiation for 3.5 h, respectively.

Moreover, the photocatalytic degradation of MO was investigated to further evaluate the photocatalytic activity of the Au/monolayer g-C₃N₄ composites; the results are presented in Fig. 1(d). After irradiation for 3.5 h, only about 21% MO was degraded by the pure monolayer g-C₃N₄. The photocatalytic activity increased remarkably when Au nanoparticles were introduced onto the monolayer g-C₃N₄. Again, 1% Au/monolayer g-C₃N₄ possessed the highest photocatalytic activity of the composites; its photocatalytic degradation efficiency of MO was about 98% after irradiation for 3.5 h. In contrast, the photocatalytic degradation efficiencies of 0.5% and 2% Au/monolayer g-C₃N₄ were about 82% and 71% after irradiation for 3.5 h, respectively. The photocatalytic degradation kinetics of the organic pollutants CIP and MO were also determined (Fig. S2).

The results of activity tests presented above indicated that the Au content of the composites had a considerable effect on their photocatalytic performance. The optimal Au content of the Au/monolayer g-C₃N₄ composites was approximately 1%. An appropriate Au content dispersed well on the surface of the monolayer g-C₃N₄ to promote the separation and transfer of photoinduced charge carriers. The photocatalytic H₂ evolution rate decreased when the Au content exceeded 1% because the excess Au nanoparticles might function as a recombination center or cover the active sites of monolayer g-C₃N₄, lowering its adsorption capacity, thereby decreasing the efficiency of charge separation and transfer [42, 43]. Therefore, it was important to optimize the ratio of Au nanoparticles to monolayer g-C₃N₄.

The stability and reusability of a photocatalyst are also relevant to its practical application. Therefore, recycling reactions were conducted for photocatalytic H₂ evolution and the photodegradation of CIP and MO over 1% Au/monolayer g-C₃N₄ under visible-light irradiation; the results are presented in Fig. 1(b) and S3, respectively. After several continuous cycles, no obvious decay of the photocatalytic activity of 1% Au/monolayer g-C₃N₄ was observed, which meant that the composite had high photocatalytic stability. Furthermore, the XRD patterns of 1% Au/monolayer g-C₃N₄ before and after the photocatalytic reactions (Fig. S3(c)) revealed that there was no apparent change in the crystal structure of the sample after several photocatalysis cycles. This indicates that the Au/monolayer g-C₃N₄ composites are very stable during the photodegradation process.

XRD was used to study the crystallinity and purity of the samples. Typical diffraction patterns of monolayer g-C₃N₄ and Au/monolayer g-C₃N₄ composites are shown in Fig. 2(a). Monolayer g-C₃N₄ exhibited a strong (002) peak at around 27.4°, which was attributed to the stacking of the conjugated aromatic system and confirmed that monolayer g-C₃N₄ was indeed g-C₃N₄ [44]. The Au/monolayer g-C₃N₄ hybrid composites also displayed the (002) peak at the same position, suggesting that the introducing of Au had no effect on the crystal structure of monolayer g-C₃N₄. Furthermore, the diffraction peaks observed at 38.2°, 44.3°, 64.6°, and 77.6° in the XRD patterns of the Au/monolayer g-C₃N₄ composites could be assigned to the (111), (200), (220), and (311) planes of metallic Au nanoparticles, respectively [45]. With increasing Au content, the intensity of the diffraction peaks from Au nanoparticles increased gradually. No diffraction peaks from Au were detected for 0.5% Au/monolayer g-C₃N₄, which was attributed to its low content of Au.

Fig. 2(b) depicts the Fourier transform infrared (FT-IR) spectra of the monolayer g-C₃N₄ and Au/monolayer g-C₃N₄ composites. The FT-IR spectra were similar for the pure monolayer g-C₃N₄ and Au/monolayer g-C₃N₄ composites, which indicates that all the Au/monolayer g-C₃N₄ composites retained the same chemical structure as that of pure monolayer g-C₃N₄. The band at about 808 cm⁻¹ corresponded to the characteristic breathing mode of the s-triazine units [46]. The absorption bands observed at about 1242, 1325, 1415, 1565 and 1640 cm⁻¹ were assigned to the typical stretching vibration modes of CN heterocycles [47]. N–H stretching vibrations were observed between 3600 and 3000 cm⁻¹ [48].

The chemical composition and chemical states of monolayer g-C₃N₄ and Au/monolayer g-C₃N₄ composites were further determined by X-ray photoelectron spectroscopy (XPS). Fig 3(a) shows the survey scan of pure monolayer g-C₃N₄ and Au/monolayer g-C₃N₄ composites; C, N, and O were found in the as-prepared samples. In addition, 1% Au/monolayer g-C₃N₄ contained Au. These results indicated that there were no other impurities in the Au/monolayer g-C₃N₄ composites and Au was successfully doped on monolayer g-C₃N₄. The high-resolution C 1s spectrum in Fig. 3(b) could be fitted with two peaks at about 284.8 and 288.5 eV. The peak at 284.8 eV corresponded to carbon contamination, and that at 288.5 eV was assigned to the typical aromatic C–N=C motif [49]. The high-resolution N 1s spectrum in Fig. 3(c) could be deconvoluted into three peaks at about 398.9, 400.4, and 401.5 eV, which were attributed to nitrogen atoms in the C=N–C, C–N(–C)–C, and C–N–H functional groups, respectively [50]. Based on these XPS analysis, the molecular structure of monolayer g-C₃N₄ remained unchanged after the introduction of a small amount of Au. In addition, the high-resolution Au 4f spectrum in Fig. 3(d) could be deconvoluted into two peaks at about 83.8 and 87.4 eV, which were assigned to Au 4f_7/2 and Au 4f_5/2, respectively. These values were consistent with reported literature [51], further confirming the presence of metallic Au in the Au/monolayer g-C₃N₄ composites.

The morphologies of the monolayer g-C₃N₄ and Au/monolayer g-C₃N₄ were observed by scanning electron microscopy (SEM) and transmission electron microscopy (TEM), as shown in Fig. 4. Fig 4(a) reveals that monolayer g-C₃N₄ was electron transparent and its surface was wrinkled, indicating that monolayer g-C₃N₄ consisted of ultrathin nanosheets with a structure like that of graphene. Fig 4(b) reveals that 1% Au/monolayer g-C₃N₄ was also wrinkled, illustrating that the introduction of Au plasmonic nanoparticles did not damage the ultrathin nanosheet structure of g-C₃N₄. Au plasmonic nanoparticles with a diameter of about 10 nm were dispersed on the surface of the monolayer g-C₃N₄ nanosheets The SEM image in Fig. 4(c) of 1% Au/monolayer g-C₃N₄ further confirmed that the composite maintained its ultrathin nanosheet structure. Chemical element mapping analysis was used to confirm the existence of Au in the composites. As illustrated in Fig. 3(d–f), C, N, and Au were distributed in similar areas, which indicated the existence and uniform distribution of these elements. Importantly, Fig. 3(f) showed that Au was homogeneously distributed, further confirming the presence of Au plasmonic nanoparticles in the Au/monolayer g-C₃N₄ composites.

The optical properties of a material greatly affect its photocatalytic activity, so the optical properties of the pure monolayer g-C₃N₄ and Au/monolayer g-C₃N₄ composites were investigated by UV-vis diffuse reflectance spectroscopy. As shown in Fig. 5, compared with the light absorption of pure monolayer g-C₃N₄, all of the Au/monolayer g-C₃N₄ composites exhibited enhanced light absorption with a slight red shift. The UV-vis absorption spectra of the Au/monolayer g-C₃N₄ composites displayed a broad absorption peak in the range of about 500-650 nm, which was ascribed to the SPR of Au nanoparticles [51]. The absorption intensity and SPR effect obviously increased with Au content, which was related to the color of the as-prepared samples (sample color changed from white to light grey with increasing Au content). The SPR absorption improved the visible-light absorption intensity and range of the composites, thus enhancing their utilization efficiency of solar energy, which might contribute to enhanced their photocatalytic activity.

The separation, transfer, and recombination of photogenerated electron-hole pairs greatly influence the photocatalytic properties of a material, To elucidate the possible photocatalytic mechanism of the samples, photoluminescence (PL), photocurrent, and electrochemical impedance spectroscopy (EIS) measurements were performed. Fig. 6(a) displays the PL spectra of pure monolayer g-C₃N₄ and Au/monolayer g-C₃N₄ composites. The emission peaks for all samples were at the same position centered around 450 nm, suggesting that introducing Au plasmonic nanoparticles did not affect the position of the emission peak of g-C₃N₄. The PL intensity of the Au/monolayer g-C₃N₄ composites was weaker than that of monolayer g-C₃N₄, indicating that the introduction of Au greatly lowered the recombination probability of photogenerated electron-hole pairs [52]. The PL results showed that introducing Au was conducive to charge separation in the Au/monolayer g-C₃N₄ composites, which then bring about higher photoactivity.

Next, the photocurrents of pure monolayer g-C₃N₄ and Au/monolayer g-C₃N₄ composites were measured to further analyze their separation and transfer of photogenerated electron-hole pairs; the results are displayed in Fig. 6(b). The photocurrents remained stable after five intermittent on-off irradiation cycles and all the composites exhibited a higher photocurrent response than that of monolayer g-C₃N₄. As the Au content increased from 0.5% to 1%, the photocurrent responses gradually rose, which means that the photogenerated electron-hole pairs in the Au/monolayer g-C₃N₄ composites were separated more effectively. The photocurrent response of the composites then gradually decreased as Au content increased to 2%, indicating that the excess Au nanoparticles might become new recombination centers, resulting in the decline of photocatalytic activity. These results are consistent with the photocatalytic activity of the composites, explaining why 1% Au/monolayer g-C₃N₄ possessed the best photocatalytic performance of the composites.

To better understand the role of Au nanoparticles in the photocatalytic activity of the composites, EIS analysis was performed. The Nyquist plots of pure monolayer g-C₃N₄ and Au/monolayer g-C₃N₄ composites are shown in Fig. 6(c). A smaller arc radius for a photocatalyst indicates more efficient charge separation and transfer [53]. All the Au/monolayer g-C₃N₄ composites exhibited a smaller arc radius than that of pure monolayer g-C₃N₄, in good agreement with the photocatalytic activities of these materials. These results illustrate that introducing Au to g-C₃N₄ facilitated charge separation and transfer. The arc radius of the Nyquist plot of 1% Au/monolayer g-C₃N₄ was the smallest of the samples, indicating that 1% Au/monolayer g-C₃N₄ had the fastest charge transfer rate, which was reflected by it exhibiting the highest photocatalytic activity of the materials.

To confirm that the reason for the enhanced photocatalytic performance of the composites was hot electron injection rather than formation of a Schottky junction, the photocurrents of monolayer g-C₃N₄ and 1% Au/monolayer g-C₃N₄ were measured under UV irradiation (λ = 365 nm). As shown in Fig. 7(a), the photocurrent intensity of 1% Au/monolayer g-C₃N₄ was nearly the same as that of monolayer g-C₃N₄ under UV light, which meant that the photogenerated electrons in the CB of monolayer g-C₃N₄ could not be transferred to the surface of Au plasmonic nanoparticles. That is, a Schottky junction did not play a part in the charge flow process. Therefore, it could be speculated that hot electron injection increased the efficiency of the whole photocatalytic process. As presented in Fig. 7(b), when the Au/monolayer g-C₃N₄ was exposed to visible light, a high density of hot electrons was generated in the Au plasmonic nanoparticles through the SPR effect. The hot electrons then flowed into the CB of monolayer g-C₃N₄ to participate in the photocatalytic reaction, resulting in increased photocatalytic activity.

Based on the above analysis results, a possible photocatalytic H₂ evolution mechanism of Au/monolayer g-C₃N₄ system was proposed, as presented in Fig. 7(b). Under visible-light irradiation, Au plasmonic nanoparticles and monolayer g-C₃N₄ were both excited to separate electron-hole pairs. The hot electrons then transferred to the CB of monolayer g-C₃N₄ because of the SPR effect. Finally, the hot electrons could unite with H⁺ to produce H₂, and the holes could react with triethanolamine to produce the oxidation product.

The possible photocatalytic mechanism of photocatalytic degradation by the Au/monolayer g-C₃N₄ composites is now discussed further. To reveal the main active oxygen species in the photocatalytic process, electron spin resonance (ESR) and trapping experiments were performed using 1% Au/monolayer g-C₃N₄. Fig. 8(a) shows the ESR spectra of superoxide radicals (O₂•⁻). There was no ESR signal in the dark and the obvious characteristic peaks of O₂•⁻ were observed visible-light irradiation. These results illustrate that O₂•⁻ could only be generated when visible light reached the surface of the as-prepared sample. In contrast, no signals from hydroxyl radicals (•OH) were detected (Fig. 8(b)), which signified that •OH was not the main active species; instead, O₂•⁻ played an important role in this photocatalytic reaction. Moreover, according to the results of trapping experiments in Fig. 8(c) and (d), generated holes also affected the photocatalytic degradation process. In the trapping experiments, tert-butanol was used as an •OH scavenger and triethanolamine was used as a hole scavenger. Fig 8(c) reveals that the photodegradation efficiency of CIP over 1% Au/monolayer g-C₃N₄ was strongly inhibited upon the addition of triethanolamine. Conversely, the photocatalytic degradation efficiency only slightly changed upon the addition of tert-butanol. Therefore, holes were another main active species and •OH were not. The same results were found for MO photodegradation (Fig. 8(d)), which indicates that the photocatalytic mechanism might be the same for photodegradation of MO and CIP. The ESR and trapping experiments revealed that O₂•⁻ and holes were the main species affecting the photocatalytic activity of this system. The proposed photocatalytic degradation mechanism based on these results is presented in Fig. 8(e). Under visible-light irradiation, Au plasmonic nanoparticles and monolayer g-C₃N₄ were both excited to separate electron-hole pairs. The hot electrons were transferred to the CB of monolayer g-C₃N₄ because of the SPR effect. The hot electrons could unite with O₂ to produce O₂•⁻, then O₂•⁻ reacted with pollutants. Meanwhile, the holes also could react with pollutants.

An Au/monolayer g-C₃N₄ plasmonic photocatalyst was successfully synthesized by an oil-bath method. The Au/monolayer g-C₃N₄ composites exhibited enhanced photocatalytic activity compared with that of monolayer g-C₃N₄ and 1% Au/monolayer g-C₃N₄ possessed the highest photocatalytic performance. The enhanced photocatalytic activity was chiefly attributed to the SPR effect of the Au nanoparticles, which promoted ultrafast hot electron transfer from Au plasmonic nanoparticles into the CB of monolayer g-C₃N₄, inducing higher transfer and separation efficiencies of the photogenerated electron-hole pairs.