Fe-Doped g-C3N4/Bi2MoO6 Heterostructured Composition with Improved Visible Photocatalytic Activity for Rhodamine B Degradation

The binary heterostructured semiconducting visible light photocatalyst of the iron-doped graphitic carbon nitride/bismuth molybdate (Fe-g-C3N4/Bi2MoO6) composite was prepared by coupling with Fe-doped g-C3N4 and Bi2MoO6 particles. In the present study, a comparison of structural characteristics, optical properties, and photocatalytic degradation efficiency and activity between Fe-doped g-C3N4 particles, Bi2MoO6 particles, and Fe-g-C3N4/Bi2MoO6 composite was investigated. The results of X-ray diffraction (XRD) examination indicate that the hydrothermal Bi2MoO6 particles have a single orthorhombic phase and Fourier transform infrared (FTIR) spectroscopy analysis confirms the formation of Fe-doped g-C3N4. The optical bandgaps of the Fe-doped g-C3N4 and Bi2MoO6 particles are 2.74 and 2.73 eV, respectively, as estimated from the Taut plots obtained from UV-Vis diffuse reflectance spectroscopy (DRS) spectra. This characteristic indicates that the two semiconductor materials are suitable for absorbing visible light. The transmission electron microscopy (TEM) micrograph reveals the formation of the heterojunction Fe-g-C3N4/Bi2MoO6 composite. The results of photocatalytic degradation revealed that the developed Fe-g-C3N4/Bi2MoO6 composite photocatalyst exhibited significantly better photodegradation performance than the other two single semiconductor photocatalysts. This property can be attributed to the heterostructured nanostructure, which could effectively prevent the recombination of photogenerated carriers (electron–hole pairs) and enhance photocatalytic activity. Furthermore, cycling test showed that the Fe-g-C3N4/Bi2MoO6 heterostructured photocatalyst exhibited good reproducibility and stability for organic dye photodegradation.


Introduction
Nanosized semiconducting photocatalysts exhibit environmental harmlessness, are stable in photochemical reactions, and are easy and flexible to prepare by various physical and chemical approaches, and are therefore being developed and used for the degradation of organic pollutants and/or contaminants, splitting of water to generate hydrogen and oxygen, and conversion of solar energy to reduce carbon dioxide and produce chemical energy at room temperature or ambient temperature, which are important and interesting topics of scientific research over the past two decades [1][2][3][4][5][6].Small enough optical bandgap semiconductors are being explored to allow efficient absorption overlap with the solar spectrum and to make effective use of solar energy, as visible light contains as much as 43% sunlight, for the development of visible-light-derived semiconducting photocatalysts [7,8].
Metal-free polymeric graphitic carbon nitride (g-C 3 N 4 ), synthesized from earth-abundant elements of carbon and nitride, exhibits specific electrical properties due to special layer structures [9][10][11].It has been widely used for photocatalytic water splitting to produce hydrogen and oxygen, as well as degradation of organic matter in visible light due to its superior optical properties and band structure feature (optical bandgap energy around 2.7 eV) [12,13].The iron (Fe)-doped semiconducting photocatalytic material system is found to have enhanced electron transfer, leading to improved photocatalytic performance [11,14].Bismuth molybdate (Bi 2 MoO 6 ) is one of the most important members in Aurivllius oxide materials.The lattice structure of γ phase Bi 2 MoO 6 consists of [Bi 2 O 2 ] 2+ layers sandwiched between [MoO 4 ] 2− slabs, which favor electron conduction and absorb visible light.Bi 2 MoO 6 has a narrower optical bandgap energy (2.5-2.8 eV) compared to typical wide-bandgap oxide photocatalysts such as TiO 2 and ZnO (3.2-3.4 eV).It is a non-toxic inorganic material system and has excellent thermal and chemical stability, making Bi 2 MoO6 well suited as a reusable visible-light-driven photocatalyst [13,15,16].
However, both Fe-doped g-C 3 N 4 and Bi 2 MoO 6 exhibit rapid recombination of photogenerated electron-hole pairs, resulting in poor quantum yields, which limits and restricts their practical optoelectrical and optoelectrochemical applications [10,13,17].Several approaches have been proposed to inhibit photoinduced carrier recombination, such as composition modification (i.e., impurity doping), textural microstructure optimization, and heterojunction structure design [18][19][20][21].Bandgap engineering based on the heterojunction design of visible-light-driven photocatalysts is a promising and feasible route to increase the effective use of solar energy for high-speed chemical reactions and the treatment of organic pollutants in the environment [10].The development of visible-light-derived photocatalysts based on coupling with graphitic carbon nitride (g-C 3 N 4 ) and oxide semiconductor nanoparticles has attracted considerable attention and interest [12,13,[17][18][19]22].
Both photocatalytic material systems, graphitic carbon nitride (g-C 3 N 4 ) and metal oxide semiconductors, possessed simple fabrication procedures, cost-effectiveness, chemical inertness, thermal and long-term stability, non-toxic and environmentally friendly properties, and good visible light photocatalytic response.The development of visiblelight-derived photocatalysts based on coupling with g-C 3 N 4 nanostructures and oxide semiconductor nanoparticles has attracted considerable attention and interest.Recent studies have reported that composite photocatalysts of graphitic carbon nitride (g-C 3 N 4 ) and bismuth-based oxide semiconductor materials (such as g-C 3 N 4 /Bi 2 O 3 , g-C 3 N 4 /BiVO 4 , g-C 3 N 4 /Bi 2 MoO 6 , and g-C 3 N 4 /Bi 2 WO 6 ) to form type II or Z-scheme heterojunctions can enhance light adsorption capability and effectively inhibit or reduce the recombination of photogenerated electrons and holes, and then exhibit high photocatalytic activity.
In this study, two semiconductor materials with bandgap energies below 3.0 eV, Fedoped g-C 3 N 4 and Bi 2 MoO 6 , were selected as visible-light-derived photocatalysts and coupled to form Fe-g-C 3 N 4 /Bi 2 MoO 6 composite to improve the photocatalytic activity and the photodegradation response performance of RhB in aqueous solution under visible light illumination at room temperature [23].A comparative study of the structural characteristics, optical properties, and photocatalytic degradation efficiency between Fe-doped g-C 3 N 4 particles, Bi 2 MoO 6 particles, and Fe-g-C 3 N 4 /Bi 2 MoO 6 composite was investigated and reported.In addition, a cyclic experiment was carried out to evaluate the repeatability and stability of the developed composite photocatalyst.
Molecules 2024, 29, x FOR PEER REVIEW 3 of 13 with the standard card of JCPDS 21-0102, confirming that the Bi2MoO6 sample has an orthorhombic crystal (patterns (ii) and (iii)) [18].The average crystallite sizes of Fe-doped g-C3N4 and Bi2MoO6 samples were calculated from the main diffraction peaks for the (002) plane of Fe-doped g-C3N4 and the (131) plane of Bi2MoO6 using the Scherrer equation to be 4.3 nm and 38.6 nm (Table 1), respectively.In addition, the diffraction peaks of the Feg-C3N4/Bi2MoO6 composite agree well with the diffraction peaks of the Bi2MoO6 particles.T. Ma et al. presented similar XRD examination results [18].Moreover, the average crystallite size of the Fe-g-C3N4/Bi2MoO6 composite (40.5 nm) is approximately 5% larger than that of the Bi2MoO6 particles (38.6 nm).The morphology and particle size of the three obtained particle samples are characterized by TEM observation (Figure 2).ImageJ software version 2024 was used to estimate the particle size from each corresponding TEM image.These TEM micrographs show that the Fe-doped g-C3N4 sample exhibits an irregular and typical layered platelet morphology with an average particle size of 1860 nm (Figure 2a) and the Bi2MoO6 sample is composed of a roughly rectangular and rod-shaped granular morphology with an average particle size of 328 nm (Figure 2b). Figure 2c is a TEM micrograph of the Fe-g-C3N4/Bi2MoO6 composite showing a significant complex morphology with an average particle size of 1455 nm associated with Bi2MoO6 particles (dark region) and Fe-g-C3N4 (light grey region).It  The morphology and particle size of the three obtained particle samples are characterized by TEM observation (Figure 2).ImageJ software version 2024 was used to estimate the particle size from each corresponding TEM image.These TEM micrographs show that the Fe-doped g-C 3 N 4 sample exhibits an irregular and typical layered platelet morphology with an average particle size of 1860 nm (Figure 2a) and the Bi 2 MoO 6 sample is composed of a roughly rectangular and rod-shaped granular morphology with an average particle size of 328 nm (Figure 2b). Figure 2c is a TEM micrograph of the Fe-g-C 3 N 4 /Bi 2 MoO 6 composite showing a significant complex morphology with an average particle size of 1455 nm associated with Bi 2 MoO 6 particles (dark region) and Fe-g-C 3 N 4 (light grey region).It provides direct evidence to confirm the formation of heterojunction structures in the hybrid material.This feature can benefit from improving photoinduced charge separation and then preventing their rapid recommendation, as well as providing an efficient electron transfer route compared to pure Fe-doped g-C 3 N 4 and Bi 2 MoO 6 particles [10].
provides direct evidence to confirm the formation of heterojunction structures in the hybrid material.This feature can benefit from improving photoinduced charge separation and then preventing their rapid recommendation, as well as providing an efficient electron transfer route compared to pure Fe-doped g-C3N4 and Bi2MoO6 particles [10].Measurement of the Fourier transform infrared (FTIR) spectrum is carried out to identify the functional groups and study the formation of compounds.The FTIR spectra of the three samples are shown in Figure 3.The absorption peak at 1639 cm −1 and the absorption band at about 3436 cm −1 are attributed to the stretching of the O-H bond and the deformation vibration of moisture absorption [24].The FTIR spectra of the Fe-doped g-C3N4 particles (spectrum (i)) and the Fe-g-C3N4/Bi2MoO6 composite (spectrum (ii)) show little difference when the Bi2MoO6 and Fe-doped g-C3N4 particles are coupled.Two samples exhibit characteristic infrared absorption bands around 3000-3500 cm −1 and 1200-1700 cm −1 , as well as a sharp absorption peak at 806 cm −1 .The infrared absorption peak represents the typical respiratory mode of triazine units, which is specific for g-C3N4 [25,26].According to previous reports, infrared absorption peaks in the 500-900 cm −1 range (spectrum (iii)) can be assigned to typical vibration modes of Bi2MoO6 lattices, including Bi-O and Mo-O bond stretching and Mo-O-Mo bridge stretching modes [18].Measurement of the Fourier transform infrared (FTIR) spectrum is carried out to identify the functional groups and study the formation of compounds.The FTIR spectra of the three samples are shown in Figure 3.The absorption peak at 1639 cm −1 and the absorption band at about 3436 cm −1 are attributed to the stretching of the O-H bond and the deformation vibration of moisture absorption [24].The FTIR spectra of the Fe-doped g-C 3 N 4 particles (spectrum (i)) and the Fe-g-C 3 N 4 /Bi 2 MoO 6 composite (spectrum (ii)) show little difference when the Bi 2 MoO 6 and Fe-doped g-C 3 N 4 particles are coupled.Two samples exhibit characteristic infrared absorption bands around 3000-3500 cm −1 and 1200-1700 cm −1 , as well as a sharp absorption peak at 806 cm −1 .The infrared absorption peak represents the typical respiratory mode of triazine units, which is specific for g-C 3 N 4 [25,26].According to previous reports, infrared absorption peaks in the 500-900 cm −1 range (spectrum (iii)) can be assigned to typical vibration modes of Bi 2 MoO 6 lattices, including Bi-O and Mo-O bond stretching and Mo-O-Mo bridge stretching modes [18].
provides direct evidence to confirm the formation of heterojunction structures in the brid material.This feature can benefit from improving photoinduced charge separa and then preventing their rapid recommendation, as well as providing an efficient e tron transfer route compared to pure Fe-doped g-C3N4 and Bi2MoO6 particles [10].Measurement of the Fourier transform infrared (FTIR) spectrum is carried ou identify the functional groups and study the formation of compounds.The FTIR spe of the three samples are shown in Figure 3.The absorption peak at 1639 cm −1 and absorption band at about 3436 cm −1 are attributed to the stretching of the O-H bond the deformation vibration of moisture absorption [24].The FTIR spectra of the Fe-do g-C3N4 particles (spectrum (i)) and the Fe-g-C3N4/Bi2MoO6 composite (spectrum (ii)) s little difference when the Bi2MoO6 and Fe-doped g-C3N4 particles are coupled.Two s ples exhibit characteristic infrared absorption bands around 3000-3500 cm −1 and 12 1700 cm −1 , as well as a sharp absorption peak at 806 cm −1 .The infrared absorption p represents the typical respiratory mode of triazine units, which is specific for g-C [25,26].According to previous reports, infrared absorption peaks in the 500-900 range (spectrum (iii)) can be assigned to typical vibration modes of Bi2MoO6 lattices cluding Bi-O and Mo-O bond stretching and Mo-O-Mo bridge stretching modes [18].The photoinduced charge transfer process of photogenerated electron-hole pairs was studied using the fluorescence (FL) spectrum.It is well known that the lower the intensity of the FL emission, the lower the recommended efficiency of the excited electrons and holes.Figure 4 shows room temperature FL spectra of Fe-doped g-C 3 N 4 particles, Bi 2 MoO 6 particles, and Fe-g-C 3 N 4 /Bi 2 MoO 6 composite excited by 350 nm light.Three samples exhibit almost identical peak positions at 470, 483, and 493 nm.This suggests that they have a similar optical bandgap energy.Furthermore, the relative intensity of the peaks is of the order of Bi 2 MoO 6 (spectrum (ii)) > Fe-doped g-C 3 N 4 (spectrum (i)) > Fe-g-C 3 N 4 /Bi 2 MoO 6 (spectrum (iii)).It should be noted that the recombination efficiency of photogenerated carriers for the Fe-g-C 3 N 4 /Bi 2 MoO 6 composite is reliably lower than for Fe-doped g-C 3 N 4 particles.
The photoinduced charge transfer process of photogenerated studied using the fluorescence (FL) spectrum.It is well known tha of the FL emission, the lower the recommended efficiency of th holes.Figure 4 shows room temperature FL spectra of Fe-doped gparticles, and Fe-g-C3N4/Bi2MoO6 composite excited by 350 nm l hibit almost identical peak positions at 470, 483, and 493 nm.This a similar optical bandgap energy.Furthermore, the relative intens order of Bi2MoO6 (spectrum (ii)) > Fe-doped g-C3N4 (spectrum (i (spectrum (iii)).It should be noted that the recombination efficie carriers for the Fe-g-C3N4/Bi2MoO6 composite is reliably lower th particles.Figure 5a-c show the low-temperature nitrogen adsorption/d Fe-doped g-C3N4 particles, Bi2MoO6 particles, and Fe-doped g-C3 respectively.The BET surface areas of the three samples are sum column of Table 1.The Bi2MoO6 particles were found to have a sm m 2 /g) than the Fe-g-C3N4 particles (32.90 m 2 /g).When Fe-g-C3N4 is to form Fe-g-C3N4/Bi2MoO6, the surface area decreases slightly to ured pore volume of three samples is also given in Table 1, which teristic is the same as the surface characteristic.1.The Bi 2 MoO 6 particles were found to have a smaller surface area (7.42 m 2 /g) than the Fe-g-C 3 N 4 particles (32.90 m 2 /g).When Fe-g-C 3 N 4 is coupled with Bi 2 MoO 6 to form Fe-g-C 3 N 4 /Bi 2 MoO 6 , the surface area decreases slightly to 31.33 m 2 /g.The measured pore volume of three samples is also given in Table 1, which shows that its characteristic is the same as the surface characteristic.The recorded UV-Vis DRS spectra of Fe-doped g-C3N4 particles, Bi2MoO6 particles, and Fe-g-C3N4/Bi2MoO6 composite are shown in Figure 6a.In the measured wavelength range, three samples exhibited a similar optical absorption spectrum.These spectra showed that the fundamental absorption edge of the Fe-doped g-C3N4 sample is at 483 nm and that the Bi2MoO6 sample has the absorption edge close to 485 nm.As the two different semiconductor particles are coupled, the Fe-g-C3N4/Bi2MoO6 composite sample has an absorption edge at 487.8 nm, showing a slight red shift.The optical bandgap energy of the prepared semiconductor samples can be calculated using the following Tauc relation [22,23]: where hv is the photon energy, A is a constant, and α and Eg are the absorption coefficient and the optical bandgap energy (at wave vector k equal to zero) of the semiconductors.Tauc plots are based on the relationship between (αhv) 2 and hv to determine the optical bandgap energy for direct transitions.By extrapolating the tangent lines of the drop part to the x-axis (photon energy, eV) in Figure 6b, the optical bandgap energies of Fe-doped g-C3N4, Bi2MoO6, and Fe-g-C3N4/Bi2MoO6 were determined to be 2.74 eV, 2.73 eV, and 2.72 eV, respectively.The edge potentials of the valence band (VB) and conduction band (CB) of a semiconductor at the zero charge point can theoretically be predicted by the following equations [10,27], which are related to the Mullikan electronegativity theory.The recorded UV-Vis DRS spectra of Fe-doped g-C 3 N 4 particles, Bi 2 MoO 6 particles, and Fe-g-C 3 N 4 /Bi 2 MoO 6 composite are shown in Figure 6a.In the measured wavelength range, three samples exhibited a similar optical absorption spectrum.These spectra showed that the fundamental absorption edge of the Fe-doped g-C 3 N 4 sample is at 483 nm and that the Bi 2 MoO 6 sample has the absorption edge close to 485 nm.As the two different semiconductor particles are coupled, the Fe-g-C 3 N 4 /Bi 2 MoO 6 composite sample has an absorption edge at 487.8 nm, showing a slight red shift.The optical bandgap energy of the prepared semiconductor samples can be calculated using the following Tauc relation [22,23]: where hv is the photon energy, A is a constant, and α and E g are the absorption coefficient and the optical bandgap energy (at wave vector k equal to zero) of the semiconductors.Tauc plots are based on the relationship between (αhv) 2 and hv to determine the optical bandgap energy for direct transitions.By extrapolating the tangent lines of the drop part to the x-axis (photon energy, eV) in Figure 6b, the optical bandgap energies of Fe-doped g-C 3 N 4 , Bi 2 MoO 6 , and Fe-g-C 3 N 4 /Bi 2 MoO 6 were determined to be 2.74 eV, 2.73 eV, and 2.72 eV, respectively.The edge potentials of the valence band (VB) and conduction band (CB) of a semiconductor at the zero charge point can theoretically be predicted by the following equations [10,27], which are related to the Mullikan electronegativity theory.
where EVB and ECB are edge potentials of the valence band and conduction band, χ is the absolute electronegativity of the semiconductor (χ = 4.72 eV and 5.55 eV for Fe-g-C3N4 and Bi2MoO6), Ee is the free electron energy on the hydrogen scale, which is about 4.5 eV, and Eg is the optical bandgap energy of the semiconductor.Here, the VB and CB edge potentials for Fedoped g-C3N4 and Bi2MoO6 are determined to be 1.58 eV, −1.

Photocatalytic Degradation Performance of Semiconducting Nano-Photocatalysts
Based on the analysis and discussion, a possible photocatalytic degradation mechanism of the Fe-g-C3N4/Bi2MoO6 composite is proposed, involving the photogenerated electron−hole pairs and photodegradation of organic pollutants.The heterojunction structure of Fe-doped g-C3N4 and Bi2MoO6 is shown in Figure 7, where both semiconducting materials are easily excited and generated electrons and holes after irradiation with visible light.Since both the CB and the VB edge potentials of Fe-doped g-C3N4 are more negative than those of Bi2MoO6, the photogenerated electrons in the CB of Fe-doped g-C3N4 could migrate to the CB of Bi2MoO6 and the photogenerated holes in the VB of Bi2MoO6 could migrate to the VB of Fe-g-C3N4.However, the electrons in the CB of Bi2MoO6 cannot reduce O2 to • O2 − because its CB edge potential (−0.32 eV) is higher than the standard redox potential (E 0 (O2/ • O2 − ) = −0.046eV vs. NHE).Similarly, the holes in the VB of Fe-doped g-C3N4 cannot react with H2O or OH − near the surface of Fe-doped g-C3N4 to form • OH (E 0 (OH − / • OH) = 2.4 eV vs. NHE) due to its lower VB edge potential (1.58 eV).

Photocatalytic Degradation Performance of Semiconducting Nano-Photocatalysts
Based on the analysis and discussion, a possible photocatalytic degradation mechanism of the Fe-g-C 3 N 4 /Bi 2 MoO 6 composite is proposed, involving the photogenerated electron−hole pairs and photodegradation of organic pollutants.The heterojunction structure of Fe-doped g-C 3 N 4 and Bi 2 MoO 6 is shown in Figure 7, where both semiconducting materials are easily excited and generated electrons and holes after irradiation with visible light.Since both the CB and the VB edge potentials of Fe-doped g-C 3 N 4 are more negative than those of Bi 2 MoO 6 , the photogenerated electrons in the CB of Fe-doped g-C  As shown in the Fe-g-C3N4/Bi2MoO6 heterojunction material system of Figure 7, photogenerated electrons in the CB of Bi2MoO6 can quickly jump to the VB of Fe-doped g-C3N4 to combine the holes in Fe-doped g-C3N4, which feature could lead to the electron accumulation in the CB of Fe-doped g-C3N4 side and the holes in the VB of Bi2MoO6 side.Therefore, the electrons in the CB of Fe-doped g-C3N4 can capture adsorbed O2 on its surface to form • O2 − and the holes in the VB of Bi2MoO6 could oxidize H2O or OH − to form • OH to achieve the photocatalytic degradation reaction [4].That is the main physical mechanism for improving the photocatalytic performance of heterojunction-based composite photocatalysts.
The pH values of the single RhB aqueous solution, RhB aqueous solution with Fe-g-C3N4 nanoparticles, RhB aqueous solution with Bi2MoO6 nanoparticles, and RhB aqueous solution with Fe-g-C3N4/Bi2MoO6 nanocomposite were 5.08, 4.37, 4.93, and 4.50, respectively.We measured the pH value and monitored the visible transmission of mixed aqueous solutions of RhB dye and different developed photocatalysts and found that each mixed aqueous solution continued towards the neutral and clear with the photodegradation process.The photocatalytic degradation performance and activity of the semiconducting particles and the composite was evaluated by determining the photodegradation rate of the RhB aqueous solution under visible light irradiation.
Figure 8a shows the variation of the photocatalytic degradation rate (Ct/C0, where C0 is the initial concentration of dye and Ct is the dye concentration after light irradiation time t) of RhB aqueous solution with visible light irradiation time for the three semiconducting photocatalysts [7].The photodegradation efficiency of photolysis (i.e., without photocatalyst, only the RhB) by exposure to visible light for varying periods of time is negligible due to its relatively stable bonding structure.The dark adsorption capacity for 30 min of the Bi2MoO6 photocatalyst was 1.7%, and the those of the Fe-doped g-C3N4 and Fe-g-C3N4/Bi2MoO6 photocatalysts were about 5%.According to the measured results, the characteristic of the photodegradation efficiency with irradiation time for P25 TiO2 is similar to that of Bi2MoO6, and the RhB dye removal efficiencies of three types of photocatalysts were observed to follow the order Fe-g-C3N4/Bi2MoO6 > Fe-doped g-C3N4 > Bi2MoO6 under identical conditions.It is observed that the photocatalytic degradation efficiency of Figure 8a shows the variation of the photocatalytic degradation rate (C t /C 0 , where C 0 is the initial concentration of dye and C t is the dye concentration after light irradiation time t) of RhB aqueous solution with visible light irradiation time for the three semiconducting photocatalysts [7].The photodegradation efficiency of photolysis (i.e., without photocatalyst, only the RhB) by exposure to visible light for varying periods of time is negligible due to its relatively stable bonding structure.The dark adsorption capacity for 30 min of the Bi 2 MoO 6 photocatalyst was 1.7%, and the those of the Fe-doped g-C 3 N 4 and Fe-g-C 3 N 4 /Bi 2 MoO 6 photocatalysts were about 5%.According to the measured results, the characteristic of the photodegradation efficiency with irradiation time for P25 TiO 2 is similar to that of Bi 2 MoO 6 , and the RhB dye removal efficiencies of three types of photocatalysts were observed to follow the order Fe-g-C 3 N 4 /Bi 2 MoO 6 > Fe-doped g-C 3 N 4 > Bi 2 MoO 6 under identical conditions.It is observed that the photocatalytic degradation efficiency of the Fe-g-C 3 N 4 /Bi 2 MoO 6 composite photocatalyst can reach 95.20% after visible light irradiation for 75 min, and the degradation efficiencies for the P25 TiO 2 , Bi 2 MoO 6 , Fe-doped g-C 3 N 4 , and Fe-g-C 3 N 4 /Bi 2 MoO 6 photocatalysts were 47.57%, 51.42%, 85.40%, and 95.53%, respectively, when under visible light irradiation for 90 min.Among the three semiconducting photocatalysts, the Fe-g-C 3 N 4 /Bi 2 MoO 6 composite exhibits the highest photocatalytic performance towards the RhB aqueous solution because it has a large specific surface area and the heterojunction structure can effectively inhibit electron-hole recombination.
the Fe-g-C3N4/Bi2MoO6 composite photocatalyst can reach 95.20% after visible light irradiation for 75 min, and the degradation efficiencies for the P25 TiO2, Bi2MoO6, Fe-doped g-C3N4, and Fe-g-C3N4/Bi2MoO6 photocatalysts were 47.57%, 51.42%, 85.40%, and 95.53%, respectively, when under visible light irradiation for 90 min.Among the three semiconducting photocatalysts, the Fe-g-C3N4/Bi2MoO6 composite exhibits the highest photocatalytic performance towards the RhB aqueous solution because it has a large specific surface area and the heterojunction structure can effectively inhibit electron-hole recombination.To quantitatively investigate the reaction kinetics of dye photodegradation, the degradation rate data were fitted by the pseudo-first-order approximation in a typical model of −ln (Ct/C0) = kat, where ka is the apparent first-order rate constant (min −1 ) and t is the irradiation time [17].As presented in Figure 8b, the apparent reaction rate constants for the P25 TiO2, Fe-doped g-C3N4, Bi2MoO6, and Fe-g-C3N4/Bi2MoO6 photocatalysts were 0.0058, 0.0199, 0.0076, and 0.0376, respectively (the last column of Table 1).It is worth noting that the reaction rate constant of Fe-g-C3N4/Bi2MoO6 photocatalyst is 4.95 times higher than that of the Bi2MoO6 photocatalyst and 1.9 times higher than that of the Fe-doped g-C3N4 photocatalyst.These results reveal that the photocatalyst has a large specific surface area, which provides more photodegradation reaction sites and leads to a high reaction rate.In addition, the interface between the heterojunction of two semiconductors plays a critical role in the photocatalytic activity.
Since the Fe-g-C3N4/Bi2MoO6 composite photocatalyst exhibited the best photodegradation performance in the present study, we performed a cycling test to reuse it as the photocatalyst for treatment in a similar 10 ppm identical RhB aqueous solution under visible light for three cycles to explore stability and reproducibility.After three cycles, the degradation rate of the developed composite photocatalyst decreased to less than 7% (as shown in Figure 9), which may be due to loss of a small amount of photocatalyst during the cycling experiment.The cycling test showed that it was photostable and a good candidate for practical photoelectrochemical application.To quantitatively investigate the reaction kinetics of dye photodegradation, the degradation rate data were fitted by the pseudo-first-order approximation in a typical model of −ln (C t /C 0 ) = k a t, where k a is the apparent first-order rate constant (min −1 ) and t is the irradiation time [17].As presented in Figure 8b, the apparent reaction rate constants for the P25 TiO 2 , Fe-doped g-C 3 N 4 , Bi 2 MoO 6 , and Fe-g-C 3 N 4 /Bi 2 MoO 6 photocatalysts were 0.0058, 0.0199, 0.0076, and 0.0376, respectively (the last column of Table 1).It is worth noting that the reaction rate constant of Fe-g-C 3 N 4 /Bi 2 MoO 6 photocatalyst is 4.95 times higher than that of the Bi 2 MoO 6 photocatalyst and 1.9 times higher than that of the Fe-doped g-C 3 N 4 photocatalyst.These results reveal that the photocatalyst has a large specific surface area, which provides more photodegradation reaction sites and leads to a high reaction rate.In addition, the interface between the heterojunction of two semiconductors plays a critical role in the photocatalytic activity.
Since the Fe-g-C 3 N 4 /Bi 2 MoO 6 composite photocatalyst exhibited the best photodegradation performance in the present study, we performed a cycling test to reuse it as the photocatalyst for treatment in a similar 10 ppm identical RhB aqueous solution under visible light for three cycles to explore stability and reproducibility.After three cycles, the degradation rate of the developed composite photocatalyst decreased to less than 7% (as shown in Figure 9), which may be due to loss of a small amount of photocatalyst during the cycling experiment.The cycling test showed that it was photostable and a good candidate for practical photoelectrochemical application.

Procedures for Preparing or Synthesizing Three Types of Visible Light Photocatalysts
The iron-doped graphitic carbon nitride (Fe-doped g-C3N4) was synthesized by the simple calcination method.A total of 4.0 g of melamine and 0.02 g iron(iii) nitrate nonahydrate [Fe(NO₃)₃•9 H₂O, Alfa Aesar] were dissolved in 20 mL of distilled water with 1 mL

Procedures for Preparing or Synthesizing Three Types of Visible Light Photocatalysts
The iron-doped graphitic carbon nitride (Fe-doped g-C 3 N 4 ) was synthesized by the simple calcination method.A total of 4.0 g of melamine and 0.02 g iron(iii) nitrate nonahydrate [Fe(NO 3 ) 3 •9 H 2 O, Alfa Aesar] were dissolved in 20 mL of distilled water with 1 mL of hydrochloric acid (HCl, 37%, J.T. Baker) under magnetic stirring at 250 rpm for 30 min at room temperature.The resulting white suspension was dried at 80 • C for 8 h to remove the liquid phase and then heated at 2 • C/min up to 550 • C and held for 4 h in a box furnace without atmosphere protection to obtain the Fe-g-C 3 N 4 product.A typical hydrothermal method was used to synthesize bismuth molybdate (Bi 2 MoO 6 ) nanoparticles.Bismuth(III) nitrate pentahydrate (Bi(NO 3 ) 3 •5H 2 O, Alfa Aesar) was dissolved in 2 mol/L nitric acid (HNO 3 , J.T. Baker) solution and sodium molybdenum oxide dihydrate (Na 2 MoO 4 •2H 2 O, Alfa Aesar) was also dissolved in 2 mol/L sodium hydroxide (NaOH, SHOWA) solution.The two resulting solutions were slowly mixed and stirred at 300 rpm for 30 min, then the pH of the mixed solution was adjusted to 5 to form a clear yellowish-white suspension.The prepared solution was transferred to a Teflon-lined stainless steel autoclave (Parr Instrument Company, model 4744, USA) and the synthesis of the hydrothermal reaction was maintained at 200 • C for 8 h.After natural cooling of the autoclave to room temperature, the Bi-based oxide precipitates were subjected to high-speed centrifugation, washed several times with distilled water and once with ethanol, and finally dried at 60 • C for 24 h to obtain the Bi 2 MoO 6 product.
The heterojunction material coupled with Fe-doped g-C 3 N 4 and Bi 2 MoO 6 was prepared by a facile solution process by stirring and mixing the as-synthesized Fe-doped g-C 3 N 4 nanostructures and the Bi 2 MoO 6 nanoparticles (weight ratio is 3:1) in ethanol at 400 rpm for three days at RT and then drying at 60 for 24 h and grinding to form the composite photocatalysts [30].All chemicals used in this study were analytical-grade reagents.

Characterization of Physical Properties and Measurement of Photocatalytic Performance
The phase structure and crystallinity of these as-prepared compound particles were examined using a Bruker D8 Discover X-ray diffractometer (Bruker, Billerica, MA, USA) with Cu Kα radiation in the 2θ scanning range from 10 • to 70 • .Morphology observation and particle size estimation of three types of particle samples were investigated using a JEOL JEM2100F transmission electron microscope (TEM, Tokyo, Japan).Infrared absorption spectra were recorded using a PerkinElmer Frontier Fourier Transform Infrared spectroscopy (FT-IR, Waltham, MA, USA) in the frequency range 400 to 4000 cm −1 .Fluorescence emission spectra were recorded at room temperature using a Shimadzu RF-5301PC spectrofluorophotometer (Kyoto, Japan) with a xenon lamp as the 350 nm excitation light source.The near UV-Vis absorption spectrum of each sample was recorded on a JASCO V-770 UV-Vis/NIR spectrophotometer (Oklahoma, OK, USA), which was used as a standard UV-Vis diffuse reflectance experiment in the wavelength range of 300−800 nm.The specific surface area value for each prepared particle sample was measured using nitrogen adsorption/desorption isotherms on a Micromeritics ASAP2020 surface area and porosimetry analyzer (Norcross, GA, USA) using the Brunauer-Emmett-Teller (BET) method.
The photocatalytic activities of three as-prepared semiconducting photocatalysts were evaluated by photodegrading the aqueous solution of Rhodamine B (RhB) under visible light irradiation for different times (from 0 to 90 min at 15 min intervals) using a xenon lamp, with a maximum power of 500 W, and a cut-off filter (λ ≥ 400 nm) as light source.The distance between the light source and the Pyrex glass cell was about 60 cm.To investigate the photodegradation efficiency, 100 mg of each photocatalyst was dispersed in an aqueous solution of RhB dye (100 mL, 10 mg/L).Before irradiation for the degradation reaction, each photocatalyst suspension should be kept in the dark for 30 min to ensure that there is sufficient contact and adsorption/desorption equilibrium between the photocatalyst and the organic dye.The mixed solution was then exposed to a fixed power of 300 W of visible light for the desired time.Before and after exposure to visible light, the concentration of the RhB aqueous solution was determined by measuring the absorbance characteristic on a Hitachi U-2900 spectrophotometer (Tokyo, Japan).
In order to identify the influence of semiconductor photocatalysts on degradation reaction activity ability for RhB dye, a blank experiment without photocatalysts was performed as a photolysis.We also selected Degussa (Evonik) P25 titanium dioxide (TiO 2 ) nanoparticles as a reference sample to compare the photodegradation reaction rate and efficiency with the three developed semiconductor photocatalysts.To evaluate the recyclability and stability of the composite photocatalysts, cycling test was carried out.To do this, the suspension was collected for the next measurement by centrifugation at 500 rpm for 5 min at set time intervals.

Figure
Figure 5a-c show the low-temperature nitrogen adsorption/desorption isotherms for Fe-doped g-C 3 N 4 particles, Bi 2 MoO 6 particles, and Fe-doped g-C 3 N 4 /Bi 2 MoO 6 composite, respectively.The BET surface areas of the three samples are summarized in the fourth column of Table1.The Bi 2 MoO 6 particles were found to have a smaller surface area (7.42 m 2 /g) than the Fe-g-C 3 N 4 particles (32.90 m 2 /g).When Fe-g-C 3 N 4 is coupled with Bi 2 MoO 6 to form Fe-g-C 3 N 4 /Bi 2 MoO 6 , the surface area decreases slightly to 31.33 m 2 /g.The measured pore volume of three samples is also given in Table1, which shows that its characteristic is the same as the surface characteristic.
where E VB and E CB are edge potentials of the valence band and conduction band, χ is the absolute electronegativity of the semiconductor (χ = 4.72 eV and 5.55 eV for Fe-g-C 3 N 4 and Bi 2 MoO 6 ), E e is the free electron energy on the hydrogen scale, which is about 4.5 eV, and E g is the optical bandgap energy of the semiconductor.Here, the VB and CB edge potentials for Fe-doped g-C 3 N 4 and Bi 2 MoO 6 are determined to be 1.58 eV, −1.14 eV and 2.41 eV, −0.32 eV, respectively.The flat band potentials of g-C 3 N 4 and Bi 2 MoO 6 semiconductors have been evaluated by the Mott-Schottky curves of S. We et al. and M. Xue et al. to be −1.19 eV and −0.32 eV, respectively[28,29].Such characteristics are close to our calculated results.Molecules 2024, 29, x FOR PEER REVIEW 7 of 13 EVB = χ − Ee − 0.5 Eg, 14 eV and 2.41 eV, −0.32 eV, respectively.The flat band potentials of g-C3N4 and Bi2MoO6 semiconductors have been evaluated by the Mott-Schottky curves of S. We et al. and M. Xue et al. to be −1.19 eV and −0.32 eV, respectively[28,29].Such characteristics are close to our calculated results.
3 N 4 could migrate to the CB of Bi 2 MoO 6 and the photogenerated holes in the VB of Bi 2 MoO 6 could migrate to the VB of Fe-g-C 3 N 4 .However, the electrons in the CB of Bi 2 MoO 6 cannot reduce O 2 to • O 2 − because its CB edge potential (−0.32 eV) is higher than the standard redox potential (E 0 (O 2 / • O 2 − ) = −0.046eV vs. NHE).Similarly, the holes in the VB of Fe-doped g-C 3 N 4 cannot react with H 2 O or OH − near the surface of Fe-doped g-C 3 N 4 to form • OH (E 0 (OH − / • OH) = 2.4 eV vs. NHE) due to its lower VB edge potential (1.58 eV).As shown in the Fe-g-C 3 N 4 /Bi 2 MoO 6 heterojunction material system of Figure 7, photogenerated electrons in the CB of Bi 2 MoO 6 can quickly jump to the VB of Fe-doped g-C 3 N 4 to combine the holes in Fe-doped g-C 3 N 4 , which feature could lead to the electron accumulation in the CB of Fe-doped g-C 3 N 4 side and the holes in the VB of Bi 2 MoO 6 side.Therefore, the electrons in the CB of Fe-doped g-C 3 N 4 can capture adsorbed O 2 on its surface to form • O 2 − and the holes in the VB of Bi 2 MoO 6 could oxidize H 2 O or OH − to form • OH to achieve the photocatalytic degradation reaction [4].That is the main physical mechanism for improving the photocatalytic performance of heterojunction-based composite photocatalysts.

Figure 7 .
Figure 7. Schematic diagram of the possible photogenerated charge carrier, carrier transfer process, and photocatalytic degradation reaction mechanism of a heterostructured Fe-g-C3N4/Bi2MoO6 composite photocatalyst under visible irradiation.

Figure 7 .
Figure 7. Schematic diagram of the possible photogenerated charge carrier, carrier transfer process, and photocatalytic degradation reaction mechanism of a heterostructured Fe-g-C 3 N 4 /Bi 2 MoO 6 composite photocatalyst under visible irradiation.The pH values of the single RhB aqueous solution, RhB aqueous solution with Fe-g-C 3 N 4 nanoparticles, RhB aqueous solution with Bi 2 MoO 6 nanoparticles, and RhB aqueous solution with Fe-g-C 3 N 4 /Bi 2 MoO 6 nanocomposite were 5.08, 4.37, 4.93, and 4.50, respectively.We measured the pH value and monitored the visible transmission of mixed aqueous solutions of RhB dye and different developed photocatalysts and found that each mixed aqueous solution continued towards the neutral and clear with the photodegradation process.The photocatalytic degradation performance and activity of the semiconducting particles and the composite was evaluated by determining the photodegradation rate of the RhB aqueous solution under visible light irradiation.Figure8ashows the variation of the photocatalytic degradation rate (C t /C 0 , where C 0 is the initial concentration of dye and C t is the dye concentration after light irradiation time t) of RhB aqueous solution with visible light irradiation time for the three semiconducting photocatalysts[7].The photodegradation efficiency of photolysis (i.e., without photocatalyst, only the RhB) by exposure to visible light for varying periods of time is negligible due to its relatively stable bonding structure.The dark adsorption capacity for 30 min of the Bi 2 MoO 6 photocatalyst was 1.7%, and the those of the Fe-doped g-C 3 N 4 and Fe-g-C 3 N 4 /Bi 2 MoO 6 photocatalysts were about 5%.According to the measured results, the characteristic of the photodegradation efficiency with irradiation time for P25 TiO 2 is similar to that of Bi 2 MoO 6 , and the RhB dye removal efficiencies of three types of photocatalysts were observed to follow the order Fe-g-C 3 N 4 /Bi 2 MoO 6 > Fe-doped g-C 3 N 4 > Bi 2 MoO 6 under identical conditions.It is observed that the photocatalytic degradation efficiency of the Fe-g-C 3 N 4 /Bi 2 MoO 6 composite photocatalyst can reach 95.20% after visible light irradiation for 75 min, and the degradation efficiencies for the P25 TiO 2 , Bi 2 MoO 6 , Fe-doped g-C 3 N 4 , and Fe-g-C 3 N 4 /Bi 2 MoO 6 photocatalysts were 47.57%, 51.42%, 85.40%, and 95.53%, respectively, when under visible light irradiation for 90 min.Among the three semiconducting photocatalysts, the Fe-g-C 3 N 4 /Bi 2 MoO 6 composite exhibits the highest photocatalytic performance towards the RhB aqueous solution because it has a large specific surface area and the heterojunction structure can effectively inhibit electron-hole recombination.

Figure 8 .
Figure 8.(a) Dynamic photocatalytic degradation and (b) linear transfer −ln (Ct/C0) of the kinetic curves for degradation of the RhB aqueous solution at 10 ppm over three different semiconducting photocatalysts and Degussa P25 TiO2 nanoparticles under visible light irradiation.

Figure 8 .
Figure 8.(a) Dynamic photocatalytic degradation and (b) linear transfer −ln (C t /C 0 ) of the kinetic curves for degradation of the RhB aqueous solution at 10 ppm over three different semiconducting photocatalysts and Degussa P25 TiO 2 nanoparticles under visible light irradiation.

Molecules 2024 , 13 Figure 9 .
Figure 9.Comparison of the photocatalytic degradation performance of RhB aqueous solution at 10 ppm with three cycles for the Fe-g-C3N4/Bi2MoO6 heterostructured photocatalyst.

Figure 9 .
Figure 9.Comparison of the photocatalytic degradation performance of RhB aqueous solution at 10 ppm with three cycles for the Fe-g-C 3 N 4 /Bi 2 MoO 6 heterostructured photocatalyst.

Table 1 .
Comparison of microstructural characteristics,