Facile Production of a Fenton-Like Photocatalyst by Two-Step Calcination with a Broad pH Adaptability.

A novel heterogeneous Fenton-like photocatalyst, Fe-doped graphitic carbon nitride (Fe-g-C3N4), was produced by facile two-step calcination method. This Fe-g-C3N4 catalyzed rhodamine B degradation in the presence of H2O2 accompanied with visible light irradiation. transmission electron microscopy(TEM), x-ray diffraction (XRD), FT-IR, x-ray photoelectron spectroscopy (XPS), and photoluminescence fluorescent spectrometer (PL) characterization analysis methods were adopted to evaluate the physicochemical property of samples. It can be observed that the Fe-g-C3N4 exhibited excellent photocatalytic Fenton-like activity at a wide pH range of 3-9, with rhodamine B(RhB) degradation efficiency up to 95.5% after irradiation for 45 min in the presence of 1.0 mM H2O2. Its high activity was ascribed to the formation of Fe-N ligands in the triazine rings that accelerated electron movement driving the Fe(III)/Fe(II) redox cycle, and inhibited photo-generated electron hole re-combinations for continuous generation of reactive oxygen species by reactions between Fe(II) and H2O2. The main active oxygen species were hydroxyl radicals, followed by superoxide radicals and hole electrons. This produced catalyst of Fe-g-C3N4 shows excellent reusability and stability, and can be a promising candidate for decontamination of wastewater.


Introduction
Advanced oxidation process such as the Fenton reaction and photocatalysis has been extensively studied due to their high degradation activity toward the refractory organic pollutants in recent years [1]. The Fenton reaction has been proven to mineralize most of the organic pollutants [2]. However, the conventional homogeneous Fenton process suffers from the narrow pH range and large amounts of iron precipitation sludge [3]. To overcome these shortcomings, researchers worldwide have focused on the heterogeneous Fenton reaction that immobilizes iron species on the carriers such as clays [4], zeolites [5], and photocatalytic compounds. Furthermore, introducing the Fe-N ligands can effectively increase the pH range of the conventional Fenton reaction [6]. Ligands such as phthalocyanines and tetra-amido macrocyclic can form Fe-N ligands. For example, Gupta et al. used the Fe-TAML catalysts to degrade pentachlorophenol at a high pH of 10 [7]. Zhu et al. developed a Fe(II)-phthalocyanine compound for the degradation of antibiotic carbamazepine [8]. The formation of Fe-N ligands not only stabilizes the iron species, but also accelerates the Fe(III)/Fe(II) redox cycle, thus resulting in the ability to work at higher pH values [9].
Graphitic carbon nitride (g-C 3 N 4 ) is a polymer that combines superior photocatalytic properties with chemical stability [10]. It is easy to prepare through the one-step polymerization of cheap precursors of melamine, cyanoguanidine, and urea [11], and it contains structures with six nitrogen lone-pair electrons, which are ideal sites for chemical modification. However, the fast electron-hole recombination

Preparation of the Catalysts
Urea and FeCl 3 •7H 2 O were used for the preparation of the calcination of g-C 3 N 4 and Fe-g-C 3 N 4 .The two-step calcination of g-C 3 N 4 was optimized according to the method described in a previous study with modifications [21]. One-step calcination (product with prefix '1 st ') was performed by heating 5 g urea with or without 0.105 g FeCl 3 •7H 2 O to 550 • C at 5 • C/min for 4 h in a muffle furnace. For two-step calcination ('2 nd '), the calcinated products were allowed to spontaneously cool to room temperature, followed by a 2 h-calcination step in a tube furnace at 520 • C (5 • C/min) in a nitrogen atmosphere (Figure 1a).

Degradation Performance of the Catalysts
The catalytic activity of the obtained catalysts were evaluated by RhB degradation efficiency at room temperature. For photocatalysis or Fenton-like photocatalysis, a 300W Xe lamp (CEAULIGHT, CEL-HXF300E7,Beijing, China) equipped with a 400 nm UV filter was used that emitted visible light (400-780 nm) at 280 mW/cm 2 . RhB was used as the target pollutant at an initial concentration of 10 mg/L (filtered through 0.45-µm acetate membranes prior to use).
For the photocatalysis reaction, as a typical process, 0.02 g of the catalyst (g-C 3 H 4 and Fe-g-C 3 H 4 ) was added into 100 mL RhB solution (10 mg/L). Before irradiation, the suspension was stirred for 40 min in the dark to ensure light-independent adsorption-desorption equilibriums had been reached (presented in Supplementary Materials in Figure S1). Then, a 300 W Xe lamp equipped with a cut off filter (>420 nm) as a light source and placed 12 cm above the surface of the reaction solution was turned on.
For the Fenton-like photocatalysis reaction, under the same condition as the photocatalysis reaction, except that 1.0 mmol/L H 2 O 2 was added into the suspension before the irradiation, the doses of which were determined by our pre-experiments ( Figure S2). A sample without any catalyst control was included as the control. At a specific time interval, 3 mL of the suspension solution was collected and filtered through a 0.22 µm filter for measure by a UV-Vis Spectrophotometer (UV-2600, SOPTOP, China) at 554 nm. When studying the influence of pH, 0.1 mol/L HCl and NaOH solutions were employed to adjust the initial pH value. The influence of pH on the treated solution during the reaction was not addressed, since the fluctuations of pH value were small (±0.2) [22].
The degradation kinetics of RhB in photocatalysis or the Fenton-like photocatalysis process were fitted based on a pseudo-first-order kinetic model (Equation (1)), where C 0 (mg/L) is the concentration of RhB after adsorption in dark (mg/L), Ct is the concentration at time t (min), and k (min −1 ) is the pseudo first-order rate constant.
The stability and reusability of the synthesized catalysts were tested by repeating the Fenton-like photocatalysis experiment. For this, at the end of a photocatalytic experiment, the used Fe-g-C 3 N 4 was filtrated, washed, and dried before being used for the following test.

Characterization
TEM images of the two calcination products were similar, but the material turned from flaxen-collared to brownish upon Fe doping (Figure 1b,c). Element mapping of Fe-g-C 3 N 4 (Figure 1d-g) confirmed the presence of Fe responsible for the color change. The Fe-signal overlapped with that of N, indicating interactions between Fe and N. One-step or two-step calcination of Fe-g-C 3 N 4 resulted in a nanosheet structure similar to that of g-C 3 N 4 visible by SEM ( Figure S2). The S BET of 1 st g-C 3 N 4 , 2 nd g-C 3 N 4 , 1 st Fe-g-C 3 N 4 and 2 nd Fe-g-C 3 N 4 was 56.78, 78.54, 45.61, and 63.52 m 2 /g, respectively, which implies that the two-step calcination can increase pore diameter and the BET specific surface area of both materials (Table S1). The higher surface area and larger pore diameter offer more active sites at the catalyst surface, which potentially enhance catalytic activity.  XRD patterns of materials by one-step and two-step calcination ( Figure 2a) displayed a peak of the (002) plane around 27.5 • , characteristic of a graphite-like structure (JCPDS No. 87-1526). The peak at 13.0 • was attributed to the crystal plane of (100) in-planar ordering of tri-s-triazine units [23], and this peak disappeared following Fe-doping, while the peak intensity for the (002) plane was significantly reduced, suggesting the absence of a rigid layer. There were no extra diffraction peaks observed for other phases such as Fe-related secondary phase or impurity, indicating Fe successfully doped into the g-C 3 N 4 lattice. The FTIR spectra ( Figure 2b) displayed characteristic bands in all samples for stretching vibration of -NH and -NH 2 groups (3000-3400 cm −1 ), C=N, and C-N bonds (1240-1640 cm −1 ) [24], and triazine (805 cm −1 ). Iron peaks were not detected in Fe-g-C 3 N 4 , suggesting that all Fe was doped into the g-C 3 N 4 .

Degradation Performance of the Catalysts
The photocatalysis performance of the catalysts were tested under visible light irradiation. As exhibited in Figure 3a, the degradation efficiency of RhB by photolysis can be ignored in the absence of the catalyst, whereas the photocatalytic degradation efficiency of RhB by 2 nd g-C3N4 was 51.7% within 45 min, which was 13.2% higher than that of 1 st g-C3N4. The photocatalytic activities were poor for both samples doped with Fe (1 st Fe-g-C3N4 and 2 nd Fe-g-C3N4). The sequence of degradation effects from high to low was 2 nd g-C3N4, 1 st g-C3N4, 2 nd Fe-g-C3N4, and 1 st Fe-g-C3N4. From the fitted curve in Figure 3b, we found the degradation process of RhB conformed to the first-order kinetics model, and the calculated rate constant was 0.01446, 0.01077, 0.00436, 0.00776 min −1 , for 2 nd g-C3N4, 1 st g-C3N4, 2 nd Fe-g-C3N4, and 1 st Fe-g-C3N4, respectively. The photocatalysis performance of 2 nd g-C3N4 was the best, which was ascribed to the highest SBET (78.535 m 2 /g) and pore volume (0.383 cc/g) As shown in Figure 2c, the PL emission of all the samples exhibited a primary emission peak around 450 nm, which is consistent with the earlier reports [25][26][27]. We found that two-step calcination could reduce peak intensity. As Fe doping strongly reduced peak intensity, probably due to the transfer of photoelectrons to Fe through chemical bridging, resulting in inhibited recombination between photoelectrons and holes [28], this was of benefit to the photocatalytic activity of the catalyst activity of the catalyst. PL emission spectra are attributed to the band-gap emission in light of its absorption band edge, estimated by the UV-Vis DRS ( Figure S4).
The surface chemical state was analyzed by XPS (Figure 2d), and the atomic content was calculated (Table S2). The proportion of Fe element was increased by two-step calcination compared to the one-step products. Peaks in the high-resolution of the N1s and Fe2p spectra were attributed to molecular components. The N1s spectrum of g-C 3 N 4 in Figure 2e contains four peaks that were attributed to pyridinic N (395.3 and 397.6 eV) [29], C-N (396.6 eV), and C=N-C (401.1 eV), respectively [30]. Fe-doping added peaks at 397.35 and 398.0 eV corresponding to iron nitride (N-Fe) and (Fe-(CN)), respectively [31], accounting for 4.8 and 4.5% of the total N. The Fe2p spectrum in Figure 2g illustrates that surface Fe represents Fe bonded with nitrogen (Fe-N) at 706.7 eV, FeO at 709.3 eV, and Fe 2 O 3 at 724.0 eV, comprising 37.8%, 36.9%, and 6.4%, respectively [32,33]. No metallic Fe was detected on the surface of Fe-g-C 3 N 4 , confirming that all Fe was completely doped into the g-C 3 N 4 lattice and mainly in the form of Fe-N.

Degradation Performance of the Catalysts
The photocatalysis performance of the catalysts were tested under visible light irradiation. As exhibited in Figure 3a, the degradation efficiency of RhB by photolysis can be ignored in the absence of the catalyst, whereas the photocatalytic degradation efficiency of RhB by 2 nd g-C 3 N 4 was 51.7% within 45 min, which was 13.2% higher than that of 1 st g-C 3 N 4 . The photocatalytic activities were poor for both samples doped with Fe (1 st Fe-g-C 3 N 4 and 2 nd Fe-g-C 3 N 4 ). The sequence of degradation effects from high to low was 2 nd g-C 3 N 4 , 1 st g-C 3 N 4 , 2 nd Fe-g-C 3 N 4 , and 1 st Fe-g-C 3 N 4 . From the fitted curve in Figure 3b, we found the degradation process of RhB conformed to the first-order kinetics model, and the calculated rate constant was 0.01446, 0.01077, 0.00436, 0.00776 min −1 , for 2 nd g-C 3 N 4 , 1 st g-C 3 N 4 , 2 nd Fe-g-C 3 N 4 , and 1 st Fe-g-C 3 N 4 , respectively. The photocatalysis performance of 2 nd g-C 3 N 4 was the best, which was ascribed to the highest S BET (78.535 m 2 /g) and pore volume (0.383 cc/g) that had the most active sites. It can be concluded that two-step calcination improved properties in the degradation of dyes.
Nanomaterials 2020, 10, x FOR PEER REVIEW 7 of 12 that had the most active sites. It can be concluded that two-step calcination improved properties in the degradation of dyes. It can be seen from Figure 3c that the degradation efficiency of RhB with the addition of H2O2 only reached about 10% at 45 min. The H2O2 under visible light without a photocatalyst could not improve RhB degradation efficiency, both confirming that H2O2 only or H2O2 with visible light irradiation could marginally degrade RhB [34]. The Fenton-like photocatalysis performance under visible light irradiation in the presence of 1.0 mM H2O2 was investigated. In the presence of H2O2, the degradation efficiency of RhB greatly increased after doped Fe ions (Figure 3c). The 2 nd Fe-g-C3N4 had the highest degradation efficiency, reaching up to 95.5% within 45 min (89% within 15 min). In comparison, the degradation efficiency of RhB by 2 nd Fe-g-C3N4 under photocatalysis, Fenton-like, and Fenton-like photocatalysis within 45 min was 24.2%, 76.4%, and 95.5%, respectively.
In addition, it can be seen that the 2 nd Fe-g-C3N4 and 1 st Fe-g-C3N4 showed higher catalytic activity than that of the 2 nd g-C3N4 and 1 st g-C3N4. The rate constant of the 2 nd Fe-g-C3N4 (0.5424 min −1 ) was the highest among all the materials, which was about 1.46, 1.86, and 2.32 times higher that of the 1 st Fe-g-C3N4 (0.3722 min −1 ), 2 nd g-C3N4 (0.2921 min −1 ), and 1 st g-C3N4 (0.2335 min −1 ). This phenomenon was likely because iron doping can promote the production of active oxygen species [35]. The effect of pH value was tested with the 2 nd Fe-g-C3N4 only, in the presence of H2O2 ( Figure  4). It shows that at the pH values of 3, 5, 7, and 9, the degradation efficiency of RhB was 96.0%, 95.2%, 92.9%, and 96.6%. Although RhB degradation was more rapid at a pH of 3, degradation was still highly efficient at a pH of 5-9. The reason is that Fe doping into the g-C3N4 formed Fe-N bonds, which promoted the migration and circulation of Fe ions and avoided the precipitation of iron salts during the Fenton-like reaction. This represents a major improvement compared to the classic Fenton It can be seen from Figure 3c that (Figure 3c). The 2 nd Fe-g-C 3 N 4 had the highest degradation efficiency, reaching up to 95.5% within 45 min (89% within 15 min). In comparison, the degradation efficiency of RhB by 2 nd Fe-g-C 3 N 4 under photocatalysis, Fenton-like, and Fenton-like photocatalysis within 45 min was 24.2%, 76.4%, and 95.5%, respectively.
In addition, it can be seen that the 2 nd Fe-g-C 3 N 4 and 1 st Fe-g-C 3 N 4 showed higher catalytic activity than that of the 2 nd g-C 3 N 4 and 1 st g-C 3 N 4 . The rate constant of the 2 nd Fe-g-C 3 N 4 (0.5424 min −1 ) was the highest among all the materials, which was about 1.46, 1.86, and 2.32 times higher that of the 1 st Fe-g-C 3 N 4 (0.3722 min −1 ), 2 nd g-C 3 N 4 (0.2921 min −1 ), and 1 st g-C 3 N 4 (0.2335 min −1 ). This phenomenon was likely because iron doping can promote the production of active oxygen species [35].
The effect of pH value was tested with the 2 nd Fe-g-C 3 N 4 only, in the presence of H 2 O 2 ( Figure 4). It shows that at the pH values of 3, 5, 7, and 9, the degradation efficiency of RhB was 96.0%, 95.2%, 92.9%, and 96.6%. Although RhB degradation was more rapid at a pH of 3, degradation was still highly efficient at a pH of 5-9. The reason is that Fe doping into the g-C 3 N 4 formed Fe-N bonds, which promoted the migration and circulation of Fe ions and avoided the precipitation of iron salts during the Fenton-like reaction. This represents a major improvement compared to the classic Fenton reaction, which requires a pH of 2-4. The Fe-N ligand not only keeps the iron stable under different pH values, but also improves the redox properties [36]. reaction, which requires a pH of 2-4. The Fe-N ligand not only keeps the iron stable under different pH values, but also improves the redox properties [36]. We performed recycling tests by recovering the nanoparticles. As shown in Figure 5a, the remove efficiency of RhB by 2 nd Fe-g-C3N4 had no significant reduction after five cycles, with the degradation efficiency of 95.5%, 94.9%, 93.6%, 92.5%, and 90.3%, respectively. Furthermore, the fresh sample and that used after five times had similar XRD characteristic peaks (Figure 5b). The above results indicate the catalytic had an excellent Fenton-like photocatalytic reusability and stability.

Mechanism of Fenton-Like Photocatalytic Reaction
The reactive species were detected under light and dark conditions by EPR, as shown in Figures We performed recycling tests by recovering the nanoparticles. As shown in Figure 5a, the remove efficiency of RhB by 2 nd Fe-g-C 3 N 4 had no significant reduction after five cycles, with the degradation efficiency of 95.5%, 94.9%, 93.6%, 92.5%, and 90.3%, respectively. Furthermore, the fresh sample and that used after five times had similar XRD characteristic peaks (Figure 5b). The above results indicate the catalytic had an excellent Fenton-like photocatalytic reusability and stability.

Mechanism of Fenton-Like Photocatalytic Reaction
The reactive species were detected under light and dark conditions by EPR, as shown in Figures  6a,b. Under visible light irradiation, the specific spectrum clearly appeared, which proved the production of •OH and •O2 − during the catalytic process. In order to further prove the main reactive species, trapping tests were carried out. IPA (1 mmol/L) was used as a scavenger of •OH in the reaction, EDTA-2Na (1 mmol/L) and N2 were used as scavengers of holes and•O2 [37], respectively. Compared with the residual RhB (~4%) for 45 min without any scavenger, the residual RhB was 36.7, 8.1, and 17.2%with IPA, EDTA-2Na, and N2, respectively (Figure 6c). This result indicates that •OH was the primary reactive species, followed by •O2 − and hole electrons.
Based on the above results and discussion, a tentative mechanism for Fe-g-C3N4 dependent Fenton-like photocatalysis is shown in Figure 6d. Fe doped g-C3N4, not only led to a soluble and reactive form of iron (Fe(III) and Fe(II)) at different pH values, but also the advantageously modified the redox properties by ligand-field effects [6]. When the Fe-g-C3N4 was excited by visible light, the

Mechanism of Fenton-Like Photocatalytic Reaction
The reactive species were detected under light and dark conditions by EPR, as shown in Figure 6a,b. Under visible light irradiation, the specific spectrum clearly appeared, which proved the production of •OH and •O 2 − during the catalytic process. In order to further prove the main reactive species, trapping tests were carried out. IPA (1 mmol/L) was used as a scavenger of •OH in the reaction, EDTA-2Na (1 mmol/L) and N 2 were used as scavengers of holes and•O 2 [37], respectively. Compared with the residual RhB (~4%) for 45 min without any scavenger, the residual RhB was 36.7, 8.1, and 17.2%with IPA, EDTA-2Na, and N 2 , respectively ( Figure 6c). This result indicates that •OH was the primary reactive species, followed by •O 2 − and hole electrons.
Nanomaterials 2020, 10, x FOR PEER REVIEW 9 of 12 electrons shifted to the conduction band and the holes left in its valence band (Equation (2)). As the interfacial charge transfer effect, parts of the photo-excited electrons reduced Fe(III) into Fe(II) (Equation (3)) (Fe(III)/Fe(II), 0.77 V, vs. NHE) [38]. Normally, the transformation from Fe(III) to Fe(II) is the decisive step for the whole reaction, much lower than the other steps. Nevertheless, it is facilitated with Fe-N ligands, thus created an excellent catalytic ability of Fe-g-C3N4. Then, Fe(II) reacted with H2O2 to produce •OH (Equation (4)), which oxidized RhB with a high efficiency. Moreover, the photo excited electron could also combine with the dissolved O2 in the solution to form•O2 − (Equation (5)).

Conclusions
In summary, the produced Fe-g-C3N4 by two-step calcination was successfully synthesized. Secondary calcination can increase specific surface area and enhance Fenton-like photocatalysis efficiency. When iron was doped into the g-C3N4 framework, Fe-N bonds were formed, which could effectively reduce the electron-hole recombination rate and broaden the scope of reaction pH in the Based on the above results and discussion, a tentative mechanism for Fe-g-C 3 N 4 dependent Fenton-like photocatalysis is shown in Figure 6d. Fe doped g-C 3 N 4 , not only led to a soluble and reactive form of iron (Fe(III) and Fe(II)) at different pH values, but also the advantageously modified the redox properties by ligand-field effects [6]. When the Fe-g-C 3 N 4 was excited by visible light, the electrons shifted to the conduction band and the holes left in its valence band (Equation (2)). As the interfacial charge transfer effect, parts of the photo-excited electrons reduced Fe(III) into Fe(II) (Equation (3)) (Fe(III)/Fe(II), 0.77 V, vs. NHE) [38]. Normally, the transformation from Fe(III) to Fe(II) is the decisive step for the whole reaction, much lower than the other steps. Nevertheless, it is facilitated with Fe-N ligands, thus created an excellent catalytic ability of Fe-g-C 3 N 4. Then, Fe(II) reacted with H 2 O 2 to produce •OH (Equation (4)), which oxidized RhB with a high efficiency. Moreover, the photo excited electron could also combine with the dissolved O 2 in the solution to form•O 2 − (Equation (5)).

Conclusions
In summary, the produced Fe-g-C 3 N 4 by two-step calcination was successfully synthesized. Secondary calcination can increase specific surface area and enhance Fenton-like photocatalysis efficiency. When iron was doped into the g-C 3 N 4 framework, Fe-N bonds were formed, which could effectively reduce the electron-hole recombination rate and broaden the scope of reaction pH in the range of 3-9. The main active oxygen species were •OH, followed by •O 2 − and hole electrons. This produced catalyst of Fe-g-C 3 N 4 shows excellent reusability and stability, and can be a promising candidate for dye degradation.
Author Contributions: S.J. and H.L. were responsible for the experimental work, Y.Y. and X.L. supervised the laboratory work, Z.Z. led the research. All authors have read and agreed to the published version of the manuscript.
Funding: This work was realized with the financial support of the National Natural Science Foundation of China (51978006).