Thermal Polymerisation Synthesis of g-C 3 N 4 for Photocatalytic Degradation of Rhodamine B Dye under Natural Sunlight

: The photocatalytic performance of g-C 3 N 4 materials prepared by different precursors for Rhodamine B (RhB) dye degradation was studied. Their crystal structure, morphologies, chemical compositions, functional groups, and optical and photoelectrochemical performances of prepared g-C 3 N 4 were analysed and characterised using X-ray diffraction, scanning electron microscopy, X-ray photoelectron spectroscopy, Fourier transform infrared spectroscopy, UV-Vis diffuse reﬂectance spectra, photoluminescence, and electrochemical workstations. The degradation of RhB dye in the presence of visible light and sunlight was utilised to assess the photocatalytic efﬁciency of the g-C 3 N 4 photocatalyst. The results of the photocatalytic comparison experiment showed that the g-C 3 N 4 photocatalyst prepared with urea as a precursor (UCN) has the best photocatalytic performance, achieving 99.61% removal in 40 min. In addition, the photocatalyst UCN can completely degrade 10 mg/L RhB dye within 20 min under sunlight, demonstrating its potential for practical applications under natural sunlight conditions. After four cycles, the degradation rate remains above 99%, demonstrating excellent stability and reusability. Due to its lower average pore number, larger BET-speciﬁc surface area and volume of pores, UCN provides more activity spaces and facilitates the adsorption of pollutant molecules, thereby enhancing photocatalytic activity. It was established through the active substance trapping studies that the main reactive species involved in the photocatalytic degradation process of RhB dye is • O 2 − . This study showed that g-C 3 N 4 synthesised with urea as the precursor has better photocatalytic performance in the degradation of RhB dye.


Introduction
With the continuous advancement of China's industrialisation and urbanisation, environmental pollution, such as water pollution, has become an increasingly severe problem, a crucial factor restricting economic development and social progress [1]. Many efforts have been made to eliminate contaminants from the water. Recently, photocatalysis technology was investigated for removing organic dyes from water because of its mild reaction conditions, low price, cleanliness, and high efficiency [2,3]. Budnyak et al. [4] synthesised the photoactive lignin/Bi 4 O 5 Br 2 /BiOBr bioinorganic complex site and found that 18.9% of Bi 4 O 5 Br 2 /BiOBr exhibited photocatalytic degradation capabilities towards cationic methylene blue and Rhodamine B (RhB) dyes under 20 W blue light irradiation. Li et al. [5] fabricated a C-BiOCl composite photocatalyst and found that it exhibited superior photocatalytic activity with RhB dye degradation irradiated for only 6 min under a 300 W Xe arc lamp. Zhang et al. [3] synthesised ZSM-5/Bi 4 O 5 Br 2 photocatalysts and found that 1ZSM-5/Bi 4 O 5 Br 2 had an excellent photocatalytic performance. The RhB dye removal rate reached 99.86% within 25 min under 300 W Xe arc lamp irradiation. Yang et al. [6] synthesised a nitrogen-deficient porous g-C 3 N 4 photocatalyst that can completely degrade RhB dye in just 20 min under simulated visible light irradiation.
The photocatalytic process depends on highly active substances, such as superoxide radicals (•O 2 − ), hydroxyl radicals (•OH), and holes (h + ) produced by the photocatalyst after absorbing light energy to decompose target pollutants [7]. Various photocatalysts, including TiO 2 -based [8], Bi-based [9,10], Mo-based [11], Zn-based [12], and Ag-based [13] semiconductors, have been developed and effectively removed RhB dye. Nevertheless, many of the aforementioned photocatalysts might give off precious elemental metals, introducing the risk of additional contamination caused by the release of metals from the photocatalysts. Therefore, metal-free photocatalysts are environmentally friendly alternatives that have prospective widespread usage [14].
Graphite carbon nitride (g-C 3 N 4 ) has generated considerable interest as a metal-free photocatalyst because of its straightforward and affordable synthesis; energy bandwidth for solar power collection; thermal stability; trustworthy chemical inertness; and electrochemical, multipurpose, and nicely organised photocatalytic characteristics [15][16][17]. g-C 3 N 4 has the in-plane tri-s-triazine units, and the sp 2 -hybridised carbon and nitrogen form a πconjugated system, advantageous for modifying materials based on nanosheets in electrical and optical electronics [18,19]. g-C 3 N 4 has good photocatalytic performance because of the narrow band gap and capacity to take in light that is visible, which can effectively absorb and use the light of wavelengths below 460 nm [20]. The g-C 3 N 4 band gap is approximately 2.70 eV, and the conduction band (CB) potential is negative (−1.3 eV), providing a strong reduction ability for photogenerated electrons [21]. The synthesis methods of g-C 3 N 4 primarily include deposition [22,23], template [24], solvothermal [25], and thermal polymerisation [26,27]. Thermal polymerisation is frequently employed because of its affordable, easy availability of raw materials, and simple operation. A few molecules with a high nitrogen content, such as cyanamide [28], dicyandiamide [29], melamine [30], and urea [31], served as a single source of molecular antecedents to fabricate g-C 3 N 4 . Ma et al. [32] prepared g-C 3 N 4 from various raw materials using thermal polycondensation, and the shapes of the products obtained using mono-cyanamide, dicyandiamide, and melamine were spherical, irregular, and layered, respectively. Liu et al. [33] prepared 3D interconnected g-C 3 N 4 via one-step thermal polymerisation using melamine and cyanuric acid.
The occurrence of catalytic reactions necessitates adsorption, and numerous studies have highlighted that optimal adsorption can yield enhanced catalytic efficiency [34,35]. Zhang et al. [36] loaded Cu on an oxide semiconductor surface to improve wastewater remediation, and the results showed that Cu particle deposition can improve the adsorption activity of a single oxide and thus improve the photocatalytic activity. Luo et al. [37] demonstrated that the adsorption process promotes the photodegradation of pollutants by g-C 3 N 4 , while the photodegradation of dyes supported on g-C 3 N 4 regenerates its adsorption capacity many times, indicating that the photocatalytic process also enhances pollutant adsorption on g-C 3 N 4 . Yu et al. [38] confirmed that a modified g-C 3 N 4 /MgZnAl-calcined layered double hydroxide composite (MCN/cLDH) possesses a large S BET -specific surface area, which enhances its adsorption capacity and improves the separation efficiency of photogenerated carriers. Furthermore, MCN/cLDH nanostructures exhibit enhanced synergistic degradation through adsorption-photocatalysis to remove oxytetracycline (95.73%) from artificial seawater. Ni et al. [39] demonstrated that the addition of g-C 3 N 4 to CCN (Co 3 O 4 QDs/3D g-C 3 N 4 ) catalyst led to an increase in tetracycline adsorption, indicating that an optimal amount of g-C 3 N 4 as a Co 3 O 4 QDs carrier can enhance adsorption and expose more active sites, thereby achieving the best synergistic effect between adsorption and oxidative degradation. Liu et al. [40] prepared an efficient three-dimensional (3D) hollow porous C-doped polymeric carbon nitride (CPCN) catalyst, and found that the formed 3D hollow porous structure gave the CPCN catalyst a large specific surface area and numerous exposed active sites, promoted the substrate adsorption reaction and maintained its structural stability, and thus improved its photolytic water hydrogen production and pollutant elimination activity.
In this study, melamine and urea were utilised as precursors for the preparation of g-C 3 N 4 (MCN and UCN, respectively) through thermal polymerisation. The photocatalytic degradation ability and practical application potential of MCN and UCN were assessed by comparing the change in RhB dye concentration under visible light and sunlight irradiation. The crystal structure, morphology, chemical composition, functional groups, optics, photoelectric chemical properties, and stability of g-C 3 N 4 were investigated. To understand how RhB molecules behaved in the reaction procedure, three-dimensional excitation-emission matrix fluorescence spectra (3D EEMs) were used. In comparison to MCN, UCN has better photocatalytic capabilities for degrading RhB dye when exposed to visible light and sunlight. This study serves as a reference for research on the synthesis and properties of g-C 3 N 4 .
2.2. Synthesis of g-C 3 N 4 g-C 3 N 4 was synthesised using the thermal polymerisation process with melamine and urea as the reactive precursors, and 15 g melamine and urea were placed, respectively, into a 50 mL covered crucible in a muff furnace. The heating rate was set to 5 • /min; the target temperature was 550 • C, maintained for 3 h; and the substance was ground for 5 min to obtain g-C 3 N 4 photocatalytic materials, named MCN and UCN.

Material Characterisation
The crystal structure information of MCN and UCN was recorded using powder X-ray diffraction (XRD) on a PANalytical X'Pert 3 Powder diffraction instrument. The specific functional groups of MCN and UCN were measured using Fourier transform infrared spectroscopy (FTIR) (IS10, Thermo Fisher Scientific, Madison, WI, USA), and the morphology and structure were evaluated using scanning electron microscopy (SEM) (JSM-7900F Plus, Nidec Corporation, Kyoto, Japan). The surface chemical composition and electronic state of MCN and UCN were determined using X-ray photoelectron spectroscopy (XPS) (Thermo Scientific K-Alpha Nexsa, Waltham, MA, USA), and the N 2 adsorption-desorption isotherms and Brunauer-Emmett-Teller (BET) surface area were evaluated using a specific surface area analyser (ASAP 2460). The optical properties were obtained using diffuse reflectance spectra (DRS) via a 3600 UV-Vis spectrophotometer. The photoluminescence (PL) spectra of MCN and UCN were studied using an FLS 980 fluorescence spectrophotometer (Edinburgh Instruments, Livingston, UK) employing an excitation wavelength of 325 nm. Electrochemical impedance (EIS) and photocurrent response tests were conducted on MCN and UCN using an electrochemical workstation (CHI660E, Shanghai Chenhua Instrument Co., Ltd., Shanghai, China). A standard three-electrode system is adopted (sample-coated conductive glass is the working electrode, platinum electrode is the opposite electrode, and Ag/AgCl is the reference electrode). The working electrode is prepared as follows: 10 mg of powder sample is weighed and dispersed in 1 mL of ethanol solution, followed by the addition of 50 µL of Nafion solution. The mixture is then subjected to ultrasonic treatment for 30 min to achieve a homogeneous suspension. Subsequently, 150 µL of the suspension is dropped onto an ITO glass substrate and dried at room temperature for testing. The measurement process was conducted in a 0.1 M Na 2 SO 4 aqueous solution. Raman analysis was performed using a confocal Raman microscope (CRM) (Alpha300R, WITec GmbH, Ulm, Germany) equipped with a TEM single-frequency laser (λ = 532 nm, laser power = 40 mW, WITec GmbH, Germany). Elemental analysis of MCN and UCN was performed using the PerkinElmer 24,000 Series II CHNS/O analyser. The EEM fluorescence spectrum was acquired using a fluorescence spectrophotometer (Aqualog-UV-800).

Photocatalytic Performance Measurement
The photocatalytic performance was assessed by comparing the photocatalytic degradation of RhB dye (dissolved in deionised water) using MCN and UCN. The procedure was as follows: 50 mg of MCN (or UCN) was weighed and placed into a solution in a 250 mL beaker (100 mL, 10 mg/L, 20 mg/L, and 30 mg/L RhB dye). To ensure a balance between adsorption and desorption, each suspension was agitated in the darkness for 60 min and then exposed to a 300 W Xenon lamp (λ ≥ 420 nm) or direct sunlight. Experiments using sunlight illumination were carried out outside between 13:00 pm and 14:00 pm (2023/5; weather: fine; 26-35 • C). During the photocatalytic procedure, 5 mL of suspension was taken at predetermined time intervals, and the photocatalyst powder was separated using a 0.45 µm filter membrane. The RhB dye concentration (λ max = 554 nm) in the suspension at different irradiation time intervals was measured using a 6100A UV-Vis spectrophotometer (Shanghai Yuanxi Instruments Co., Ltd., Shanghai, China), and the Lambert-Beer law was used to calculate the RhB removal efficiency.

Crystal Structure and Functional Group Analysis
The crystal properties and composition of the synthesised photocatalysts were characterised using the XRD technique. The diffraction pattern in Figure 1a displays two distinct peaks at 2θ values of 13.18 • and 27.86 • , which can be attributed to the (100) and (002) planes of g-C 3 N 4 , respectively (JCPDS NO. 01-087-1526) [41]. The (100) crystal plane indicates the periodic arrangement of the structural units in the layer, and the (002) crystal plane corresponds to the interlayer stacking of aromatic segments. The positions of these two diffraction peaks correlate with those previously reported for g-C 3 N 4 [42]. Compared with MCN, UCN's peak intensity and width of the (002) crystal plane is weaker and broader, indicating that urea as a precursor can inhibit the crystal growth of g-C 3 N 4 . The average crystal diameters were calculated using Scherrer's equation: D = Kλ/βcosθ, where D is the average crystal diameter (nm); K is the Scherrer constant (0.89); λ is the wavelength of the X-ray (1.5406 • A); β is the full width at half-maximum intensity and θ corresponds to the angle of maximum peak [3,43], and the calculated average crystallite sizes of MCN and UCN are equal to 3.67 and 1.77 nm, respectively.

Microstructure and Compositions Analysis
SEM was performed on MCN and UCN to study the typical morphology and detailed structure of the as-prepared samples. Figure 2a shows that MCN presents a hierar- The intrinsic chemical structure of the samples can be detected using FTIR spectra. Figure 1b shows that the characteristic peak value at 810 cm −1 corresponds to the out-ofplane bending vibration of the triazine ring [44], and the broader band of approximately 3000 to 3600 cm −1 corresponds to the N-H stretching vibration [45]. The characteristic band at 1675-1200 cm −1 ascribes to the stretching vibration mode of C-N and C=N [46,47]. The tensile vibrations of the bonds in the prepared photocatalyst can be revealed via Raman spectroscopy, as depicted in Figure 1c. The peaks observed at 478 cm -1 , 710 cm −1 , 753 cm −1 , 978 cm −1 , and 1235 cm −1 correspond to the tensile vibrations of C-C and C-N heterocycles [48][49][50].

Microstructure and Compositions Analysis
SEM was performed on MCN and UCN to study the typical morphology and detailed structure of the as-prepared samples. Figure 2a shows that MCN presents a hierarchical structure with layers stacked tightly into blocks. UCN becomes thinner and presents a more dispersed and uniform porous structure ( Figure 2c). The degree of the UCN sample's interlayer stacking structure is low, and the interlayer structure is loose, which can increase the surface area and reactive sites [51]. The results are consistent with those of BET. SEM-EDS was used to analyse the distribution and content percentage of elements on the surface of the photocatalyst. C and N are uniformly distributed on the surface of the photocatalyst (Figure 2b

Microstructure and Compositions Analysis
SEM was performed on MCN and UCN to study the typical morphology and detailed structure of the as-prepared samples. Figure 2a shows that MCN presents a hierarchical structure with layers stacked tightly into blocks. UCN becomes thinner and presents a more dispersed and uniform porous structure (Figure 2c). The degree of the UCN sample's interlayer stacking structure is low, and the interlayer structure is loose, which can increase the surface area and reactive sites [51]. The results are consistent with those of BET. SEM-EDS was used to analyse the distribution and content percentage of elements on the surface of the photocatalyst. C and N are uniformly distributed on the surface of the photocatalyst (Figure 2b    XPS was used to analyse and compare the surface elemental composition of MCN and UCN. The XPS survey spectrum shows that C and N existed in MCN and UCN (Figure 3a). The C 1s spectra are fitted into two peaks at 288.1 and 284.8 eV (Figure 3b). The primary C 1s peak at 288.1 eV is associated with the sp 2 -bonded carbon (N-C=N), and the peak at 284.8 eV corresponds to C-C [31]. The N1s XPS spectrum in Figure 3c displays four peaks at 398.5, 400.4, 401.3, and 404.2 eV [52]. These peaks correspond to C=N-C, ternary N (N(-C) 3 ), N-H bond [7,46], and π-excitations [31], respectively. XPS was used to analyse and compare the surface elemental composition of MCN and UCN. The XPS survey spectrum shows that C and N existed in MCN and UCN (Figure 3a). The C 1s spectra are fitted into two peaks at 288.1 and 284.8 eV (Figure 3b). The primary C 1s peak at 288.1 eV is associated with the sp 2 -bonded carbon (N-C=N), and the peak at 284.8 eV corresponds to C-C [31]. The N1s XPS spectrum in Figure 3c displays four peaks at 398.5, 400.4, 401.3, and 404.2 eV [52]. These peaks correspond to C=N-C, ternary N (N(-C)3), N-H bond [7,46], and π-excitations [31], respectively.

BET Surface Area and Pore Size Distribution Analysis
The BET surface area and the distribution of pore sizes for the MCN and UCN were measured using the N2 adsorption-desorption isotherms ( Figure 4). The adsorption-desorption isotherms of the MCN and UCN are classical type IV with a hysteresis loop of type H3, indicating that the catalyst has a mesoporous structure [53]. The BET surface area of UCN is 60.03 m 2 ·g −1 , more than eight times that of MCN. Table 2 shows that UCN exhibits a larger specific surface area, greater pore volume, and smaller average pore size, thereby providing an increased number of active sites that contribute to the adsorption of pollutant molecules [54]. Therefore, the mesoporous structure and large BET surface area

BET Surface Area and Pore Size Distribution Analysis
The BET surface area and the distribution of pore sizes for the MCN and UCN were measured using the N 2 adsorption-desorption isotherms (Figure 4). The adsorptiondesorption isotherms of the MCN and UCN are classical type IV with a hysteresis loop of type H 3 , indicating that the catalyst has a mesoporous structure [53]. The BET surface area of UCN is 60.03 m 2 ·g −1 , more than eight times that of MCN. Table 2 shows that UCN exhibits a larger specific surface area, greater pore volume, and smaller average pore size, thereby providing an increased number of active sites that contribute to the adsorption of pollutant molecules [54]. Therefore, the mesoporous structure and large BET surface area of UCN are expected to enhance the adsorption capacity and photocatalytic degradation efficiency of pollutants.
Water 2023, 15, x FOR PEER REVIEW 7 of UCN are expected to enhance the adsorption capacity and photocatalytic degrada efficiency of pollutants.

UV-Vis Spectroscopy and Photochemical Characteristics
To understand the light absorption capacity of MCN and UCN, UV-vis absorp spectra were analysed, as shown in Figure 5. The sharp innate absorption edge for M and UCN are at around 480 and 460 nm, respectively. The absorption edge's sh primarily due to the porous structure's strong quantum constraint causing the nanosh to become thinner and smaller in the UCN [55]. The Kubelka-Munk formula was use estimation to understand the MCN and UCN energy bands better. This is based on α A (hυ − Eg) n/2 , where a, h, n, A and Eg are the absorption coefficient, the Planck cons light frequency, a constant and band gap energy [56], respectively. The value of n dep on the type of optical transition exhibited by the semiconductor. For g-C3N4 with a d band gap, the value of n is equal to 1 [21,57]. The Eg of MCN and UCN are 2.70 and eV, respectively, similar to the study by Hong et al. [58].

UV-Vis Spectroscopy and Photochemical Characteristics
To understand the light absorption capacity of MCN and UCN, UV-vis absorption spectra were analysed, as shown in Figure 5. The sharp innate absorption edge for MCN and UCN are at around 480 and 460 nm, respectively. The absorption edge's shift is primarily due to the porous structure's strong quantum constraint causing the nanosheets to become thinner and smaller in the UCN [55]. The Kubelka-Munk formula was used for estimation to understand the MCN and UCN energy bands better. This is based on αhυ = A (hυ − Eg) n/2 , where a, h, n, A and Eg are the absorption coefficient, the Planck constant, light frequency, a constant and band gap energy [56], respectively. The value of n depends on the type of optical transition exhibited by the semiconductor. For g-C 3 N 4 with a direct band gap, the value of n is equal to 1 [21,57]. The Eg of MCN and UCN are 2.70 and 2.79 eV, respectively, similar to the study by Hong et al. [58]. Recombining photogenerated carriers within the photocatalyst releases fluorescence energy, and a high fluorescence signal indicates a high photogenerated carrier recombination rate [9,54]. Therefore, PL analysis was used to evaluate the efficiency of photogenerated electron-hole pair separation. Figure 6a shows that MCN and UCN have distinct signal peaks near 480 nm. The UCN fluorescence intensity is significantly weaker than that of MCN, indicating that UCN aids in the separation of photogenerated electron- tion rate [9,54]. Therefore, PL analysis was used to evaluate the efficiency of photogenerated electron-hole pair separation. Figure 6a shows that MCN and UCN have distinct signal peaks near 480 nm. The UCN fluorescence intensity is significantly weaker than that of MCN, indicating that UCN aids in the separation of photogenerated electron-hole pairs and thus improves photocatalytic activity. Furthermore, the photocarrier interface separation and migration of g-C 3 N 4 were measured using the Nyquist plots of EIS. Generally, a smaller radius of the arc on the EIS Nyquist plot indicates more efficient interface charge transfer and lower photogenerated electron-hole pair recombination [59]. In Figure 6b, the EIS Nyquist plot of UCN shows a smaller arc radius than that of MCN, indicating that UCN has a lower interfacial charge transfer resistance and can separate photoinduced carriers more efficiently. The photocurrent responses of MCN and UCN are shown in Figure 6c. The higher the photocurrent density, the better the electron-hole pair separation performance [60]. The photocurrent response of UCN is higher, which further indicates that UCN has a stronger electron-hole pair separation ability. Recombining photogenerated carriers within the photocatalyst releases fluorescence energy, and a high fluorescence signal indicates a high photogenerated carrier recombination rate [9,54]. Therefore, PL analysis was used to evaluate the efficiency of photogenerated electron-hole pair separation. Figure 6a shows that MCN and UCN have distinct signal peaks near 480 nm. The UCN fluorescence intensity is significantly weaker than that of MCN, indicating that UCN aids in the separation of photogenerated electronhole pairs and thus improves photocatalytic activity. Furthermore, the photocarrier interface separation and migration of g-C3N4 were measured using the Nyquist plots of EIS. Generally, a smaller radius of the arc on the EIS Nyquist plot indicates more efficient interface charge transfer and lower photogenerated electron-hole pair recombination [59]. In Figure 6b, the EIS Nyquist plot of UCN shows a smaller arc radius than that of MCN, indicating that UCN has a lower interfacial charge transfer resistance and can separate photoinduced carriers more efficiently. The photocurrent responses of MCN and UCN are shown in Figure 6c. The higher the photocurrent density, the better the electron-hole pair separation performance [60]. The photocurrent response of UCN is higher, which further indicates that UCN has a stronger electron-hole pair separation ability.

Photocatalytic Degradation Performance
The degradation of RhB dye was used to gauge both the adsorption and photocatalytic abilities of MCN and UCN.

Photocatalytic Degradation Performance
The degradation of RhB dye was used to gauge both the adsorption and photocatalytic abilities of MCN and UCN. Figure 7 demonstrates that UCN exhibits superior removal efficiency in treating RhB dye at different concentrations under visible light or sunlight, with equilibrium being achieved within 20 min. After 40 min of visible light exposure (Figure 7a), the removal efficiency of UCN towards 30 mg/L RhB dye reaches up to 99.61%, which is higher by 20.31% compared to MCN. MCN has the best effect on the treatment of RhB dye with a concentration of 10 mg/L under visible light, and the removal rate is 98.58% after 40 min of illumination, but the removal efficiency is still lower than that of UCN (99.36%). Additionally, the applicability of the prepared photocatalyst under natural sunlight was also investigated. As shown in Figure 7c, the photocatalyst UCN outperforms the laboratory-simulated visible light catalytic system by completely degrading 10 mg/L of RhB dye in 20 min. This is mostly because that natural sunlight emits not only visible light but also ultraviolet radiation, which can potentially enhance the generation of charge carriers by the catalyst and facilitate the degradation of RhB dye [61]. In order to visualise the photocatalytic degradation of RhB dyes via UCN and MCN, a degradation diagram excluding the adsorption effect was made. As shown in Figure 7b,d, under simulated visible light and real sunlight irradiation, the effect of UCN is still better after deducting the adsorption effect, which indicates the adsorption-photocatalytic synergy. In conclusion, the utilisation of urea yields superior results in the preparation of g-C 3 N 4 . Figure 8 shows the UV-Vis absorption spectra for the photodegradation of RhB dye in the UCN photocatalyst. The maximum absorption peak of RhB dye at 554 nm (λ max ) gradually decreased with the photodegradation process. After 5 min of visible light irradiation, the UCN photocatalyst decomposed approximately 96.15% of the RhB dye. The primary absorption peak's position was significantly reduced, indicating that UCN has excellent photocatalytic performance.
visible light but also ultraviolet radiation, which can potentially enhance the gener of charge carriers by the catalyst and facilitate the degradation of RhB dye [61]. In o to visualise the photocatalytic degradation of RhB dyes via UCN and MCN, a degrad diagram excluding the adsorption effect was made. As shown in Figure 7b,d, under ulated visible light and real sunlight irradiation, the effect of UCN is still better afte ducting the adsorption effect, which indicates the adsorption-photocatalytic synerg conclusion, the utilisation of urea yields superior results in the preparation of g-C3N   In addition, 3D EEMs technology was employed to elucidate the alterations of RhB dye after photocatalysis. Before visible light illumination, the characteristic fluorescence signal appeared at Ex/Em = 460-590/540-700 nm and 200-425/550-670 nm. After being stirred in darkness for 30 min (Figure 9a), the position and shape of the fluorescence peak remained relatively stable, but its intensity slightly decreased, indicating that the RhB molecule was only adsorbed on the catalyst without decomposition. After 40 min, the fluorescence intensity began to decrease. This indicates that the RhB has broken down into pieces and molecular fragments [62]. This is consistent with the results shown in Figure 8.  In addition, 3D EEMs technology was employed to elucidate the alterations of RhB dye after photocatalysis. Before visible light illumination, the characteristic fluorescence signal appeared at Ex/Em = 460-590/540-700 nm and 200-425/550-670 nm. After being stirred in darkness for 30 min (Figure 9a), the position and shape of the fluorescence peak remained relatively stable, but its intensity slightly decreased, indicating that the RhB molecule was only adsorbed on the catalyst without decomposition. After 40 min, the fluorescence intensity began to decrease. This indicates that the RhB has broken down into pieces and molecular fragments [62]. This is consistent with the results shown in Figure 8. signal appeared at Ex/Em = 460-590/540-700 nm and 200-425/550-670 nm. After being stirred in darkness for 30 min (Figure 9a), the position and shape of the fluorescence peak remained relatively stable, but its intensity slightly decreased, indicating that the RhB molecule was only adsorbed on the catalyst without decomposition. After 40 min, the fluorescence intensity began to decrease. This indicates that the RhB has broken down into pieces and molecular fragments [62]. This is consistent with the results shown in Figure 8. The stability and recycling properties of the UCN were verified by carrying out four cycle tests for the photocatalytic degradation of RhB dye. After each photocatalytic degradation, the powder material was extracted and recycled, cleaned three times with deionised water, and dried for the next cycle experiment. Compared with the first time (99.61%), the removal rate shows a slight change (Figure 10a). In four cycles, the treatment rates were 99.58%, 99.49%, 99.36%, and 99.35%. The results indicate that UCN exhibits exceptional stability and reusability. The photostability of UCN was further confirmed by   The stability and recycling properties of the UCN were verified by carrying out four cycle tests for the photocatalytic degradation of RhB dye. After each photocatalytic degradation, the powder material was extracted and recycled, cleaned three times with deionised water, and dried for the next cycle experiment. Compared with the first time (99.61%), the removal rate shows a slight change (Figure 10a). In four cycles, the treatment rates were 99.58%, 99.49%, 99.36%, and 99.35%. The results indicate that UCN exhibits exceptional stability and reusability. The photostability of UCN was further confirmed by measuring the XRD patterns of UCN after four recycling tests on RhB (Figure 10b). The XRD results show that there was no change in the crystal structure of UCN and that all the characteristic peaks are present.  Furthermore, the photocatalytic property of g-C3N4 was compared with other materials explored by other researchers in removing RhB dye, and the results are demonstrated in Table 3. Although the g-C3N4 studied in this work is relatively basic, its degradation rate of RhB dye exceeds that of other composite photocatalysts. The results demonstrate that g-C3N4 exhibits superior photocatalytic performance compared with most materials over a specific time, thereby establishing a robust foundation for further research.  Furthermore, the photocatalytic property of g-C 3 N 4 was compared with other materials explored by other researchers in removing RhB dye, and the results are demonstrated in Table 3. Although the g-C 3 N 4 studied in this work is relatively basic, its degradation rate of RhB dye exceeds that of other composite photocatalysts. The results demonstrate that g-C 3 N 4 exhibits superior photocatalytic performance compared with most materials over a specific time, thereby establishing a robust foundation for further research.

Photocatalytic Mechanism
Trapping tests were conducted to study the various species implicated in photodegradation and the primarily responsible species. The primary active species of UCN photocatalytic degradation were detected by adding IPA, BQ, and SO to RhB dye as scavengers of •OH, •O 2 − , and h + [43]. Figure 11 shows that the degree of removal of RhB dye (41.56%) declined significantly with the addition of BQ, while the scavenging effect of IPA (99.58%) and SO (98.53%) was weaker than that of BQ. Therefore, in the reaction process of UCN degradation of RhB dye, •O 2 − serves as the primary reactive species. The valence band (VB) potentials of a semiconductor can be determined using the empirical equation EVB = X − E e + 0.5Eg, where X, E e , and Eg represent the semiconductor's electronegativity (the X value for g-C3N4 is 4.72 eV [69]), the energy of free electrons on the hydrogen scale (4.5 eV), and the semiconductor's bandgap, respectively [70]. Moreover, the CB potentials can be calculated using ECB = EVB + Eg [71]. In the DRS analysis results (Figure 5b), the Eg of UCN and MCN are 2.79 and 2.70 eV, respectively. Therefore, the CB and VB potentials of UCN are calculated to be −1.17 eV (ECB) and 1.62 eV (EVB), while the CB and VB potentials of MCN are -1.13 eV (ECB) and 1.57 eV (EVB), respectively. According to these results, Figure 12 illustrates the plausible photocatalytic mechanism involved in the process of RhB photocatalytic degradation. UCN and MCN tend to produce e − and h + under visible light radiation. Since the CB potential of UCN and MCN is lower than  The valence band (VB) potentials of a semiconductor can be determined using the empirical equation E VB = X − E e + 0.5E g , where X, E e , and E g represent the semiconductor's electronegativity (the X value for g-C 3 N 4 is 4.72 eV [69]), the energy of free electrons on the hydrogen scale (4.5 eV), and the semiconductor's bandgap, respectively [70]. Moreover, the CB potentials can be calculated using E CB = E VB + E g [71]. In the DRS analysis results (Figure 5b), the E g of UCN and MCN are 2.79 and 2.70 eV, respectively. Therefore, the CB and VB potentials of UCN are calculated to be −1.17 eV (E CB ) and 1.62 eV (E VB ), while the CB and VB potentials of MCN are −1.13 eV (E CB ) and 1.57 eV (E VB ), respectively. According to these results, Figure 12 illustrates the plausible photocatalytic mechanism involved in the process of RhB photocatalytic degradation. UCN and MCN tend to produce e − and h + under visible light radiation. Since the CB potential of UCN and MCN is lower than  (Figure 5b), the Eg of UCN and MCN are 2.79 and 2.70 eV, respectively. Therefore, the CB and VB potentials of UCN are calculated to be −1.17 eV (ECB) and 1.62 eV (EVB), while the CB and VB potentials of MCN are -1.13 eV (ECB) and 1.57 eV (EVB), respectively. According to these results, Figure 12 illustrates the plausible photocatalytic mechanism involved in the process of RhB photocatalytic degradation. UCN and MCN tend to produce e − and h + under visible light radiation. Since the CB potential of UCN and MCN is lower than E θ (O2/•O2 − ) = −0.33 eV, the e − in the CB of UCN and MCN can capture the adsorbed O2 on the catalyst surface and reduce it to give •O2 − . Since the EVB of UCN and MCN are more negative than E θ (•OH/H2O) = 2.27 eV [71], •OH cannot be generated on the VB of UCN and MCN. The results correlate with the trapping experiments.

Conclusions
MCN and UCN were prepared via thermal polymerisation using melamine and urea as precursors. Compared to MCN, UCN exhibits superior photocatalytic degradation performance towards RhB dye. Under the irradiation of visible light, the removal rate of RhB dye at 10 mg/L within 20 min reaches 99.36%, while complete degradation can be achieved under sunlight. The photocatalytic efficiency of UCN is enhanced by the layered structures and a large specific surface area, which enrich the reaction sites, improve the transfer of photoinduced charges, and strengthen the absorbance of visible light. The prepared UCN exhibits exceptional stability, even after undergoing four cycles of photocatalytic reaction. Moreover, the trapping experiments established that •O 2 − was the primary reactive species in the photocatalytic degradation of RhB dye under visible light radiation. Therefore, the as-synthesised UCN are promising photocatalysts for environmental applications such as the treatment of industrial wastewater under natural sunlight. This study highlights the potential of UCN as a photocatalyst for RhB dye removal in wastewater treatment and lays a foundation for further exploration of g-C 3 N 4 .