The g-C₃N₄ nanosheet was prepared via thermal polycondensation of urea. Fifteen grams urea was placed into a covered ceramic crucible and heated to 500°C for 5 h in air, at the heating rate of 10°C min⁻¹. After the reaction, it cooled down to room temperature naturally, the product was collected and grind to powder. The Ag/g-C₃N₄ composite was synthesized by a liquid-phase reduction method. First 0.5 g of g-C₃N₄ was put in 50 mL of water and ultrasonic treated for 5 min. Second, a certain amount of AgNO₃ (5 mM) aqueous solution was put into the above solution and maintain stirring. Third, a certain amount of NaBH₄ [the molar ratio of n(AgNO₃):n(NaBH₄) = 1:5] dissolved in 30 mL of water, and then put into the above solution, stirring for 1 h. Following, the product was centrifuged and washed with absolute ethanol and distilled water, respectively. Finally, these samples are dried in vacuum oven at 70°C for 5 h. By varying the amount of using AgNO₃, a series of samples with different ratios of Ag to g-C₃N₄ [m(Ag):n(g-C₃N₄) = 1, 2, 3, 4, and 5%] were prepared and labeled as 1%-Ag/g-C₃N₄, 2%-Ag/g-C₃N₄, 3%-Ag/g-C₃N₄, 4%-Ag/g-C₃N₄, and 5%-Ag/g-C₃N₄, respectively.

The crystal phases of products were studied by X-ray diffraction (XRD) (X-ray diffractometer, Cu Kα, λ = 1.54056 Å) (Bruker D5005, Germany). Fourier transform infrared (FT-IR) spectra were conducted using a Nicolet Magna 560 (US) spectrophotometer. X-ray photoelectron spectroscopy (XPS) was measured on a PHIQ 1,600 XPS (US) instrument. The weight percentages of Ag in the Ag/g-C₃N₄ photocatalysts was studied by inductively coupled plasma atomic emission spectrometry (ICP-AES, Shimadzu ICP-7510, Japan). High resolution transmission electron microscopy (HRTEM) was taken by a JEOL JEM-2100F (Japan) electron microscope. UV–vis absorbance spectra were collected on a Shimadzu UV-3100 (Japan) spectrophotometer, using BaSO₄ as reference. The photoluminescence (PL) spectra of g-C₃N₄ and Ag/g-C₃N₄ samples were studied on a Varian Cary Eclipse (US) spectrometer equipped with an excitation wavelength of 325 nm.

The photocatalytic activity of pure g-C₃N₄ and Ag/g-C₃N₄ samples was examined by photodegradation of RhB under the simulated sunlight irradiation, which was obtained from an 500 W Xe lamp. Ten milligram of samples were dispersed in 25 mL of RhB aqueous solution (10 mg/L RhB aqueous solution). Prior to the irradiation, the reaction solution was magnetically stirred in the dark for 30 min to get adsorption-desorption equilibrium for the dyes on photocatalyst surface. During the photocatalytic degradation, 2 mL of the sample was withdrawn from the reaction solution at the time intervals of every 15 min and then centrifuged to remove the particles. Then the concentration of RhB was examined by UV–vis spectrophotometer, at the absorbance wavelength of 553 nm. The photodegradation rate of RhB was calculated by the formula: D = C/C₀ × 100%, where C₀ is the initial concentration of RhB, and C is the concentration of RhB at a time t.

The photoelectrochemical performance was studied on a CHI 660D electrochemical work station with a standard three-electrode system. Put g-C₃N₄ or Ag/g-C₃N₄ on the ITO glass surface as the working electrode. A piece of Pt wire and a calomel electrode were used as the counter electrode and reference electrode, respectively. The electrolyte is 0.1 mol/L Na₂SO₄ aqueous solution. Five milligram photocatalysts were mixed with 1 mL ethanol and then the mixture was coated on 2 × 4 cm ITO glass for use as an electrode. Electrochemical impedance spectroscopy (EIS) Nyquist plots were conducted at an open current potential with an amplitude of 5 mV and the frequency range was from 10⁵ to 1 Hz.

The crystal phase of as-prepared samples is studied by XRD measurements, and the XRD patterns are shown in Figure 1. The pure g-C₃N₄ nanosheets and Ag/g-C₃N₄ nanocomposites have two dominant peaks at 13.1° and 27.5°, indexed to g-C₃N₄ (JCPDS87-1526) (Yang et al., 2013b). The peak at 27.5° is ascribed to the typical (002) plane with planar distance of 0.33 nm corresponding to interlayer-stacking of aromatic segments. The peak at 13.1° with distance of 0.675 nm is indexed to the (100) plane corresponding to in-plane structural packing (Dong et al., 2011; Liu et al., 2011). Compared with the pure g-C₃N₄ nanosheets, the intensity of the diffraction peak at 27.5° becomes weaker with increasing content of loading Ag NPs. The diffraction peak related to Ag NPs is not found, because of the low Ag loading amount and the high dilution effect of Ag NPs on the g-C₃N₄ surface (Zhou et al., 2014; Fu et al., 2015). As follows, the XPS and EDS data demonstrate the existence of Ag loading on the g-C₃N₄ surface.

The FTIR spectra are similar between the pure g-C₃N₄ nanosheets and Ag/g-C₃N₄ composites with different Ag loading amounts (Figure 2). The peak at 1,639 cm⁻¹ can be ascribed to the stretching vibration of C-N groups, and the peaks at 1,242, 1,327, 1,568 and 1,408 cm⁻¹ can be attributed to the aromatic C-N stretching vibration (Aghdam et al., 2017). The peak at 809 cm⁻¹ corresponds to the breathing mode of triazine units (Sun et al., 2012). The peak at 3171 cm⁻¹ is attributed to the stretching vibration of N-H group (Yang et al., 2013a). All these characteristic FTIR peaks suggest that the overall structure of g-C₃N₄ maintains the original form after Ag NPs loading.

The morphology and microstructure of the pure g-C₃N₄ and 3%-Ag/g-C₃N₄ samples were investigated by TEM measurements. The TEM image of pure g-C₃N₄ shows that it is a two-dimensional nanosheet with some holes in the size range of 10–30 nm (Figure 3A). TEM image of the 3%-Ag/g-C₃N₄ sample (Figure 3B) shows that Ag NPs, observed as black dots, uniformly disperse on g-C₃N₄ surfaces. The size of Ag NPs is from 6 to 20 nm, indicating that these Ag NPs are Ag clusters on the surface of g-C₃N₄. The size distribution of Ag NPs on 3%-Ag/g-C₃N₄ is presented in Figure 3C, which is mainly in the range of 8–18 nm. The energy dispersive X-ray spectrum (EDS) also confirms that Ag NPs exist on the surface of g-C₃N₄ (Figure 3D). Also, it shows that the 3%-Ag/g-C₃N₄ sample is consisted of C, N, and Ag elements, which confirms that Ag NPs successfully adsorbed on the g-C₃N₄ surface.

The surface elemental composition and chemical states of Ag/g-C₃N₄ are studied by XPS, here 3%-Ag/g-C₃N₄ is selected for study (Figure 4). The elements C, N, O, and Ag are clearly observed in the survey spectrum (Figure 4A). The peak located at 531 eV is assigned to O, which may be the water molecules at the sample surface (Hu et al., 2015). Two C 1 s peaks locate at 284.8 eV and 288.3 eV (Figure 4B). The peak located at 284.8 eV is assigned to sp²-hybridized C atoms, and the 288.3 eV peak can be assigned as N-C = N₂ groups (Wu et al., 2015). As shown in Figure 4C, the peaks of N 1 s locate at 398.8, 400.5, and 401.5 eV, which can be ascribed to sp² bonded nitrogen C-N-C groups, sp³ tertiary nitrogen N-(C)₃ and amino functional groups (C-N-H), respectively (Qi et al., 2019). The spectrum of Ag 3d (Figure 4D) shows that the peaks located at 367.4 and 374.0 eV can be assigned as Ag 3d^5/2 and Ag 3d^3/2, respectively (Yang et al., 2016). This confirms that Ag NPs are successfully coasted on g-C₃N₄ surfaces. The peak at 368.1 eV is assigned as Ag(I), which indicates the formation of Ag₂O on the surface of metallic Ag (Tian et al., 2015). The actual content of Ag in the Ag/g-C₃N₄ composites was studied by ICP-AES analysis. The result shows that the weight percentages of Ag in 1%-Ag/g-C₃N₄, 2%-Ag/g-C₃N₄, 3%-Ag/g-C₃N₄, 4%-Ag/g-C₃N₄, and 5%-Ag/g-C₃N₄ were measured to be 0.72, 1.43, 2.09, 2.74, and 3.55, respectively. The measured value by ICP-AES is a little smaller than the theoretical value for the weight percentages of Ag in Ag/g-C₃N₄ composites, but both of the two changing trends are the same.

The UV-DRS measurement is used to study the optical adsorption property of the pure g-C₃N₄ nanosheets and Ag/g-C₃N₄ composites (Figure 5). Figure 5A shows that the light absorption edge of the pure g-C₃N₄ is at 440 nm, which agrees with the intrinsic band gap of bulk g-C₃N₄ (Chen et al., 2016). Compared with pure g-C₃N₄ nanosheets, Ag/g-C₃N₄ composites have an additional weak and broad absorption peak around 450–600 nm, which is characteristic of the silver surface plasmon resonance band (Liu et al., 2013). The Ag/g-C₃N₄ composite shows similar light absorption range with that of pure g-C₃N₄, but the visible light adsorption is increased, as shown in Figure 5B. Thus, the samples with the increasing of Ag loading amount change the color from yellow to dark gray.

Photoluminescence measurement is an useful method to analyze the separation efficiency and the life time of photogenerated carriers, as shown in Figure 6. PL spectra of pure g-C₃N₄ and Ag/g-C₃N₄ are taken by the exciting light of 325 nm. A strong broad peak at ~460 nm is observed. Compared with pure g-C₃N₄, the PL intensity of Ag/g-C₃N₄ composites decreases significantly. The weaker peak intensity of PL results in a slower recombination rate of photogenerated carriers (Ong et al., 2014). In Ag/g-C₃N₄ composites, Ag NPs combine with the g-C₃N₄ surface strongly, and effectively reduce the recombination rate of e⁻-h⁺ pairs. This improved separation efficiency of photogenerated carriers leads to increasing e⁻ and h⁺ to join the photocatalytic process of Ag/g-C₃N₄. However, over loading of Ag on g-C₃N₄, such as 5%-Ag/g-C₃N₄, the PL intensity is getting increase again, indicating the increase of the combination of photogenerated carriers. It clearly sees that the Ag/g-C₃N₄ composites with proper loading amounts of Ag have a potential for using as photocatalysts with high activity.

The photocatalytic activity of pure g-C₃N₄ and Ag/g-C₃N₄ for photodecomposition of RhB is tested under the simulated sunlight irradiation. As shown in Figure 7A, compared with that of pure g-C₃N₄, the Ag nanoparticle modified g-C₃N₄ shows an improved photocatalytic performance for decomposition of RhB aqueous. After irradiation for 100 min, the degradation of RhB is about 20% for pure g-C₃N₄ nanosheets and almost 100% for 3%-Ag/g-C₃N₄. Figure 7B shows the apparent reaction rate constant (k) of RhB photodegradation, which shows that the kinetic constant of 3%-Ag/g-C₃N₄ is almost 10 times higher than that of pure g-C₃N₄. When the mass ratio of Ag is in the range of 1–5 wt%, the enhanced photocatalytic activity is observed, due to effective enhanced the separation efficiency of photogenerated e⁻-h⁺ pairs at the Ag/g-C₃N₄ interface and the surface plasmon resonance (SPR) effect of Ag NPs (Duan et al., 2014), which are also supported by PL and UV–vis DRS results. Obviously, 3%-Ag/g-C₃N₄ has the highest photocatalytic activity, which may be due to that the excess Ag NPs will be working as recombination centers, or the active site on g-C₃N₄ surface is blocked (Ge et al., 2011). The photocatalytic activity of as-prepared Ag/g-C₃N₄ composites is compared to the typical photocatalysts including TiO₂ and ZnO. The photocatalytic activity of 3%-Ag/g-C₃N₄ (k = 0.0326 min⁻¹) in this work is higher that TiO₂ (k = 0.0084 min⁻¹) and ZnO (k = 0.0062 min⁻¹) reported previously (Carvalho et al., 2015; Hao et al., 2019). In order to check the reusability of as-prepared Ag/g-C₃N₄ photocatalysts, the recycling test was carried out. As shown in Figure 7C, the photodegradation percentage of RhB is >90% after six cycles, which indicates the as-prepared photocatalysts owns a good stability. A gradually decreased photocatalytic activity is due to the loss of photocatalyst in the recovery process.

The charge separation efficiency is studied by using the photoelectrochemical measurements. Figure 8A shows that the photocurrent response of pure g-C₃N₄, 1%-Ag/g-C₃N₄, 3%-Ag/g-C₃N₄, and 5%-Ag/g-C₃N₄ samples under the simulated sunlight irradiation, which shows stable reproducible photocurrent responses over five on-off cycles. The photocurrent starts when the light was turned on, the photocurrent is close to zero when the light was turned off. The photocurrent density (0.50 μA/cm²) of 3%-Ag/g-C₃N₄ is bigger than the pure g-C₃N₄ (0.07 μA/cm²) and the 5%-Ag/g-C₃N₄ (0.38 μA/cm²). The stronger photocurrent is due to the higher separation efficiency of the photogenerated e⁻-h⁺ pairs of Ag/g-C₃N₄, which is consistent with its higher activity on photocatalytic decomposition of organic dyes.

In order to study the charge separation efficiency, the electrochemical impedance spectroscopy (EIS) Nynquist plots are used, and the EIS Nyquist plot of pure g-C₃N₄, 1%-Ag/g-C₃N₄, 3%-Ag/g-C₃N₄, and 5%-Ag/g-C₃N₄ sample is as shown in Figure 8B. Usually, the smaller the radius of EIS Nyquist plots is, the higher the separation efficiency of charge carriers is (Li et al., 2015). The radius on the EIS Nynquist plot of C₃N₄ is getting smaller after Ag modification, the order is 3%-Ag/g-C₃N₄ < 5%-Ag/g-C₃N₄ < 1%-Ag/g-C₃N₄ < g-C₃N₄, indicating that Ag modification indeed reduces the recombination of charge carries and increase the separation efficiency of photogenerated e⁻-h⁺ pairs, 3%-Ag/g-C₃N₄ is the best, which agrees well with PL spectra.

Figure 9 shows the photocatalytic mechanism of Ag/g-C₃N₄ composites during RhB decomposition under sunlight irradiation. Ag nanoparticle modification enhances the photocatalytic performance of g-C₃N₄ due to the synergistic effect of two aspects, one is the SPR effect of metal Ag, another is the decrease of the recombination rate of photogenerated e⁻-h⁺ pairs (Ingram et al., 2011). When Ag/g-C₃N₄ is irradiated by the simulated sunlight irradiation, the e⁻-h⁺ pairs are separated, e⁻ is excited to CB of g-C₃N₄, h⁺ remains at VB of g-C₃N₄. Then e⁻ transfers to Ag NPs due to the high Schottky barrier of Ag, finally, transfers to the photocatalyst surface to join the reduction reaction. The generated e⁻ from two routes one from the plasmon excited Ag NPs and the other from the photoexcited g-C₃N₄ nanosheets. These e⁻ react with O₂ to generate O^2−•, and O^2−• radicals can discompose of RhB molecules to CO₂ and H₂O. Thus it is concluded that the adsorbed silver NPs have two functions, one is as the electron pool and the other is capture of the photoinduced electrons.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.