Synthesis of crystalline carbon nitride with molten salt thermal treatment for efficient photocatalytic reduction and removal of U(VI)

It is critical to recover uranium from wastewater containing U(VI) and to ensure the commercial development of nuclear-related energy sources. The extraction and recovery methods of uranyl ions from contaminated areas have become clearer with the rapid development of photocatalytic technology. However, there are several challenges, such as low-charge carrier migration and a lack of active sites. Crystallized carbon nitride (CCN) catalysts at different temperatures were successfully obtained to improve the efficiency of photocatalytic reduction of uranium. The characterization results show that the crystallinity of the material increased and improved the electron–hole separation efficiency. We investigated the effects of catalyst dosage, pH value of a solution, and concentration of U(VI) on the photocatalytic reactions as well as the photoreduction mechanisms. The efficiency of photocatalytic reduction of uranium is enhanced by 2.5 times with good stability in five cycles compared with bulk carbon nitride. The results demonstrate that CCN can effectively remove U(VI) by photocatalytic reduction, which has a wide range of applications prospects in the treatment of uranium-containing wastewater.


Introduction
Uranium is one of the most important and effective ingredients in nuclear reactors, which is crucial in the development of alternative energy sources [1][2][3]. Conversely, uranium contamination poses a potential threat to human health and the environment. The treatment of uranium-containing wastewater is a major issue in environmental protection that has attracted attention globally [4][5][6]. Therefore, the recovery of uranium from wastewater and the safe treatment of uranium-containing wastewater are important to ensure the sustainable development of nuclear energy. Various traditional technologies have been developed for the treatment of uranium-containing wastewater [7,8]. However, traditional processing methods are not fully applicable because of their high cost, complex operation, and low processing efficiency. Because soluble hexavalent uranium can be reduced to insoluble tetravalent uranium for recovery and extraction, it is considered a green and beneficial strategy [9][10][11][12][13].
However, most photocatalysts contain metal components, and the scarcity and metal toxicity of these photocatalysts restrict their large-scale practical applications [14][15][16][17][18][19][20]. Graphitic carbon nitrides (g-C 3 N 4 ) are the most effective candidates in the field of photocatalysis as a metal-free photocatalyst [21][22][23][24]. This is because of their merits, such as good electronic structure, unique optical properties, chemical stability, and environmental friendliness. Furthermore, these excellent properties enable g-C 3 N 4 to be widely used in visible-light photocatalytic applications. However, the bulk of g-C 3 N 4 synthesized from traditional heatinduced polycondensation exhibits a lower crystallinity and moderate photocatalytic activity, which is because of the low degree of polymerization and slow photo-generated electron transfer rate. We improve the photocatalytic performance of traditional g-C 3 N 4 based on these defects. Dai et al. [25] used thermal condensation to prepare the carbon dots modified porous graphitic carbon nitride (CNCD) for photoreduction of U(VI) under visible LED light irradiation. Wang et al. [26] used the supramolecular self-assembly method to prepare 3D g-C 3 N 4 and investigated the photo-immobilization process of U(VI). Furthermore, Wang et al. [27] prepared adjustable mesoporous g-C 3 N 4 using the template method. Its recyclability and high selectivity reflect the potential value of the treatment of uranium from wastewater. Therefore, it is imperative to achieve high photocatalytic activity by regulating the morphology of carbon nitride.
Recently, the preparation of crystalline carbon nitride (CCN) has been an effective strategy to enhance photocatalytic activity [28][29][30][31]. Because highly crystalline g-C 3 N 4 has a significantly tighter layer stacking and fewer defects, which contributes to increased charge carrier migration. Additionally, Li et al. [29] prepared highly crystalline g C3N4 hollow spheres (CCNHS) using the molten salt method and cyanuric acid melamine as a precursor and increased the number of active sites for photocatalytic reactions. Li et al. [32] destroyed the internal structure of carbon nitride by doping heteroatoms to improve the photocatalytic reaction. However, CCN prepared by molten salt thermal treatment still has some shortcomings, such as a low specific surface area and few accessible active sites Synthesis of crystalline carbon nitride with molten salt thermal… because of the relatively low controllable morphology of nanostructures. Furthermore, the preparation of CCN with regular nanostructures by molten salt thermal treatment to improve the specific surface area and the number of active sites, thus enhancing photocatalytic activity, remains a great challenge.
In this study, we used preheated melamine as the precursor to synthesize crystalline CN semiconductors using KCl/LiCl salts as heptazine units. Notably, increasing the crystallinity of g-C 3 N 4 not only improves the light harvesting in the visible light region but also accelerates electron-hole separation efficiency and improves the capability toward solar-driven photocatalysis reduction U(VI). In addition, we systematically analyzed the prepared catalyst. Furthermore, we investigated the effects of the pH of the solution, amount of catalyst, and initial concentration of U(VI) on the photocatalytic reduction of uranium. The CCN-500 exhibited high photocatalytic activity for uranium reduction. Additionally, we also investigated the possible photocatalytic mechanism.

Materials
Melamine, potassium chloride, lithium chloride, uranyl nitrate hexahydrate, methanol, and arsenazo III were purchased from Aladdin and used without further purification. Deionized water was used in all experiments.

Synthesis of CN samples
First, 3.0 g of melamine was heated in a muffle furnace with a heating rate of 1 °C/ min from room temperature to 550 °C, kept at 550 °C for 4 h, and then cooled to room temperature. The resulting products were collected and ground into powder, which was denoted as CN.

Synthesis of CCN-X samples
Here, 3.0 g melamine was heated at 400, 450, 500, and 550 °C for 2 h with a heating rate of 1 °C/min in a muffle furnace in an air atmosphere, respectively. Thereafter, 0.5 g of the preheated sample was grounded with 2.75 g KCl and 2.25 g LiCl in a vacuum glovebox. Furthermore, the mixture was heated to 550 °C for 4 h at a rate of 2 °C/min under an air atmosphere in a muffle furnace. The product was washed several times with boiled deionized water and collected after drying overnight at 60 °C in an oven after cooling to room temperature. These samples were denoted as CCN-400, CCN-450, CCN-500, and CCN-550, respectively. For comparison, 1.0 g melamine was grounded with 5.5 g KCl and 4.5 g LiCl in a vacuum glovebox. The mixture was heated to 550 °C for 4 h at a rate of 2 °C/min under an air atmosphere in a muffle furnace.

Characterization
The morphology of as-prepared catalysts was characterized by scanning electron microscope (SEM, Hitachi, accelerating voltage of 10 kV) and transmission electron microscope (TEM, FEI Talos F200s, accelerating voltage of 200 kV). The X-ray diffraction (XRD, D8 advance, Bruker) instrument equipped with Cu Kα radiation (λ = 1.5406 Å) at steps of 0.02° was used to measure the structures of the samples, respectively. Fourier-transform infrared (FT-IR) spectra were obtained on a Thermo Nicolet instrument, and the spectra acquisition range was 500-4000 cm −1 with a spectral resolution of 4 cm −1 . The elements on the surfaces of the samples were determined using X-ray photoelectron spectroscopy (XPS, Thermo Scientific K α Thermo Scientific) with Al-Kα radiation at 15 kV and 15 mA. The binding energies in XPS analysis were corrected by referring to 284.8 eV for C-1s. The UV-vis diffuse reflectance spectrum of the samples was measured using a spectrophotometer (TU-1901, PERSEE, Chinese) utilizing BaSO 4 as a reference. The N 2 adsorption-desorption isotherms were measured using a TriStar II 3020 physical adsorption instrument, the BET-specific surface area was calculated accordingly at the relative pressure (p/p 0 ) from 0.01 to 1.0, and the total pore size and pore volume were determined at a p/p 0 of 0.99.

Photocatalytic U(VI) reduction and elution experiment
Photocatalysts (25 mg) were added to the UO 2 2+ solution (50 mL 40 mg/L), and 2 mL methanol was added as a hole sacrificial agent. A 300 W Xeon-lamp with a cutoff filter (λ ≥ 420 nm) was used as the light source. The light intensity was 500 mW/cm 2 , optical current was 15 A, and the distance from the light source to the liquid level was 5 cm. Furthermore, the solution was stirred in the air environment for 0.5 h to achieve adsorption/desorption equilibrium. At certain time intervals, 1 mL of the suspension was analyzed to remove the photocatalyst by centrifugation. The photocatalyst was removed by centrifugation and collected after completing the photocatalytic reduction experiment, followed by adding 20 mL of 3 M HNO 3 into the catalyst and oxidizing it for 30 min. Furthermore, 1 mL of the suspension was analyzed to remove the photocatalyst by centrifugation. The Shimadzu UV-2600 UV-vis spectrophotometer at 652 nm wavelength was used to detect the U(VI) concentrations.

Characterization of the CCN-X
The morphological structure of CN and CCN-500 was confirmed by SEM and TEM, as shown in Fig. 1. The SEM image of CN (Fig. 1a) shows the structure of small irregular particles, which was the common morphology of g-C 3 N 4 prepared 1 3 Synthesis of crystalline carbon nitride with molten salt thermal… by conventional thermal polymerization. However, as shown in Fig. 1b, g-C3N4 underwent obvious changes and formed a polycrystalline structure after molten salt thermal treatment. Additionally, as shown in Fig. 1c, d, the obvious crystal structure and lattice fringe can be observed by TEM, suggesting the formation of crystalline g-C 3 N 4 [29]. In addition, nitrogen sorption analysis (Fig. 1e, f and Table 1) indicated that the obtained CN and CCN-500 samples had a BET surface area of 23 and 50 m 2 g −1 and a pore volume of 0.053 and 0.116 cm 3 g −1 , respectively. Furthermore, all samples exhibited type-IV isotherms with H3-type hysteresis loops, indicating mesoporous properties [25,33]. The larger specific surface area not only increased the contact area between CCN-500 and UO 2 2+ but also increased the number of active sites on the surface of CCN-500 samples, enhancing the CCN-500 activity.
The crystal structures of the as-prepared CCN-X and CN were studied by XRD, as shown in Fig. 2a. Many diffraction peaks appeared on the XRD patterns of the CCN-X samples, which correspond to g-C 3 N 4 based on heptazine and triazine [28]. The (100) and (002) peak positions of the CCN-500 samples shifted from 12.6° and 27.3° to 8.1° and 28.1° compared with the CN, corresponding to  the increasing planar repeat and decreasing interlayer distance, respectively. Furthermore, oxygen atoms can combine with the carbon in the heptazine units in the air atmosphere, disrupting the structure of the heptazine units and breaking them down into triazine units.
The internal chemical bonds of synthesized CN and CCN-X were studied by FTIR as shown in Fig. 2b. The peaks at 789, 1700-1200 cm −1 were the characteristic peaks of heptazine derivatives [34], attributable mainly to the stretching vibrations of C-N and C=N in the aromatic carbon nitride heterocyclic structures. The peak located at 2150 cm −1 was attributed to the terminal cyan groups C≡N, and the broad peaks at 3500-3000 cm −1 were attributed to the terminal amino group [35]. The FTIR spectrum of CCN-500 was similar to that of CN, indicating that the carbon nitride prepared by molten salt thermal treatment maintains the chemical core structure.
The elemental valence states and surface chemical state of the sample were further explored by XPS in Fig. 3. According to the XPS survey spectra (Fig. 3a), the peaks at 288 and 398 eV were attributed to the C 1s and N 1s. To thoroughly analyze the chemical bond between C and N atoms, high-resolution spectra of C 1s and N 1s were analyzed, respectively (Fig. 3b, c). As shown in Fig. 3b, C 1s had three peaks at 284.8, 286.6, and 288.2 eV, which corresponded to the C-C bond attributed to the sp 2 hybridized carbon, C-NH x on the edges of sub-units, and the heterocycle of aromatic g-C 3 N 4 sp 2 hybridized N-C=N, respectively [36,37]. As shown in Fig. 3c, CN had three peaks at 398.5, 400.1, and 401.2 eV, corresponding to C-N=C, N-(C) 3, and C-NH x , respectively [23,38]. All the binding energy N 1s spectra of CCN-500 shifted toward the higher binding energy side, suggesting that CCN-500 was more prone to losing electrons and was conducive to the photocatalytic reduction of uranium. Moreover, only one of the O 1s peaks was present at 531.9 eV and 532.6 for CN and CCN-500 (Fig. 5d), respectively, which can be attributed to the surface-absorbed H 2 O species [39]. Additionally, elemental analysis (Table 1)   Synthesis of crystalline carbon nitride with molten salt thermal… structural integrity of the CCN-500 sample. This phenomenon may be caused by crystallization.

Optical absorption properties
The optical absorption properties and band gap of the samples were also measured by UV-vis DRS. All the samples exhibited a typical semiconductor optical absorption in the visible light range, as shown in Fig. 4a. The band edges of the CCN-X samples exhibited a slight red shift and showed a strong absorption intensity relative to the CN samples under the full spectrum. This is because of the reduced π-π interlayer stacking distance in the CCN-500 frameworks owing to increased crystallization of the CCN-500 samples [23]. In addition, the increased adsorption intensity of CCN-X samples can be attributed to the multiple reflections of incident light in the porous nanocrystal and the existence of nitrogen defects, suggesting that the CCN-X samples have strong light energy utilization under visible light, generating more electron-hole pairs. The band gap energies estimated by the Tauc plots were 2.60, 2.46, 2.56 eV for CN, CCN-0, and CCN-500, respectively [40].
The Mott-Schottky experiments further investigated the band structures in Fig. 4c. The conduction band (CB) of the CN and CCN-X samples was measured by the electrochemical workstation and analyzed by the Mott-Schottky method. The tangent line was made to be the longest straight part of the Mott-Schottky  [41,42], thereby resulting in the VB potentials of 1.82 and 1.98 eV, as shown in Fig. 4d. CCN-500 had a narrower band gap energy, and the VB potential and CB potential of the CCN-500 samples were both improved compared with CN. Therefore, the CCN-500 sample had higher conduction and valence bands, which were more prone to electronic transition in sunlight, favoring the reduction reaction.
The charge carrier separations and recombination rates of the photoexcited carriers were studied by the PL spectra with an excitation wavelength of 375 nm and the time-resolved fluorescence. Luminescence was observed to range between 400-700 nm for the CN and CCN-500 samples, and both demonstrated a maximum at around 470 nm, as shown in Fig. 5a. This may be the n-π * electronic transitions for g-C 3 N 4 of the characteristic band PL phenomenon. The PL intensity for CCN-500 decreased significantly compared to CN, which was attributed to the enhanced crystallinity of the extended conjugated system with suppressed Synthesis of crystalline carbon nitride with molten salt thermal… electron-hole pair recombination. In addition, Fig. 5b shows the time-resolved photoluminescence spectra. The average lifetime of CCN-500 (1.805 ns) was longer than that of CN (1.194 ns) ( Table 2). The longer intrinsic fluorescence lifetime is also expected to improve the transport rate of photocarriers in the CCN-500 samples, suggesting that the CCN-500 can improve the photocatalytic activity of uranium reduction [43,44].  The transient photocurrent responses (i-t) and the electrochemical impedance spectrum (EIS) were further analyzed in the interface electron transfer rate of the as-prepared samples, thus explaining the relationship between the recombination rate of photo-generated electrons-holes and photocurrent. I-t spectra of prepared CN and CCN-500 showed that CCN-500 samples exhibited a higher photocurrent response than the CN samples, which is favorable for the transmission of photoinduced electrons, as shown in Fig. 5c. EIS measurements can also investigate the charge-transfer rate. As shown in Fig. 5d, the results show that both the CN and CCN-X samples performed semicircular Nyquist plots, and the diameter of CCN-500 was significantly decreased, indicating that the improved charge mobility in the electrodes enhanced the photocatalytic activity of uranium reduction.

Photocatalytic reduction in U(VI) under visible light
Different catalysts were used to remove U(VI) in the same system (m catalyst = 25 mg to investigate the photocatalytic activities of CN and CCN-X, and C U(VI) = 40 mg/L, pH = 5.0), 2 mL methanol was added as a hole sacrificial agent, and the apparent rate constant (k) by the pseudo-first-order reaction (ln(C/C 0 ) = kt) was calculated, as shown in Fig. 6. As shown in Fig. 6a, the removal of U(VI) was negligible in the system with no catalyst, suggesting that the UO 2 2+ self-photolysis can be ignored. However, after a dark reaction in the catalyst system for 30 min, the concentration of the solution remained constant, suggesting that the uranyl ions reached adsorption-desorption equilibrium in the catalyst active site. The CN showed a poor visible-light response under visiblelight radiation. This may be attributed to the traditional thermal polymerization method. Furthermore, the photocatalytic activity of crystalline carbon nitride first increased and then decreased as the pretreatment temperature increased, and the CCN-500 demonstrated an excellent reduction effect in the photocatalytic removal U(VI); approximately 99% U(VI) was removed within 40 min under visible light irradiation. The results show that the efficiency of photocatalytic Synthesis of crystalline carbon nitride with molten salt thermal… reduction of U (VI) can be improved by increasing the crystallinity of the samples. Figure 6b shows the apparent rate constant of photocatalytic removal U(VI). The reaction rate for CCN-500 was 0.132 min −1 , which was 2.5 times higher than that of CN (0.053 min −1 ). Additionally, the photocatalytic elution experiment was investigated to evaluate the practical application significance of the catalyst, as shown in Table 3. The results showed that the elution rates were 66 and 96% for CN and CCN-500, respectively. Furthermore, no U(VI) was detected in the solution after the reaction, and 1.92 mg U(VI) could be detected in the eluate, suggesting that U(VI) was reduced to U(IV) on the catalyst and adsorbed on the catalyst surface.
Dosages of CCN-500 and concentration of U(VI) were tested to investigate the effect of external factors on U(VI) removal: different pH of the U(VI) solution. The removal performance of CCN-500 on U(VI) was studied at different pH values, as shown in Fig. 7a. As a result, the pH increased from 4 to 5, reaching a peak of approximately 100%. However, as the pH increased further, it decreased to 92% at pH = 6. This may be attributed to the adsorption site of U (VI) being more easily captured by H + when the acidity of the solution is strong. Conversely, the dosage of CCN-500 had a certain influence on the photocatalytic performance, as shown in Fig. 7b. The reduction efficiency increased and then gradually decreased after reaching a peak at a dosage of 25 mg with the increased dosage of CCN-500. This may be because the visible light irradiation is being resisted by the excess catalyst, leading to the decreased effect of photocatalytic reduction. Furthermore, the concentration of U(VI) was one of the influences on the photocatalytic performance, as shown in Fig. 7c. The CCN-500 exhibited an upward tendency of photocatalytic reduction performance with the increased U(VI) concentration; the photocatalytic reduction performance reaches the highest until the concentration of U(VI) in 40 mg/L. The above results proved that the pH was 5.0, the concentration of U(VI) was 40 mg/L, and after adding 25 mg CCN-500, 99% of the U(VI) was removed after 40 min of irradiation.
The regeneration and reusability of used photocatalysts for practical applications have great significance. After each photocatalytic cycle, the used CCN-500 was collected, thoroughly washed with 0.5 mol L −1 HNO 3 , and reused for the next photocatalytic reaction. The photocatalytic removal rate of U(VI) decreased from 97.1 to 90.1% after five recycles, as shown in Fig. 7d. The XRD (Fig. 7e) and the FTIR patterns (Fig. 7f) showed the CCN-500 samples after five runs. Furthermore, the used samples exhibited the same structure as the raw samples, suggesting that the photocatalysts have extraordinary stability and reusability. Furthermore, the CCN-500 sample exhibited relatively high photocatalytic reduction U(VI) activity compared with others previously reported in recent years (Table 4).

Photocatalytic mechanism exploration
The catalyst after the photocatalytic reaction was characterized to explore the photocatalytic mechanism in the process. The SEM image of CCN-500 after photocatalytic reaction and the mapping map of the corresponding elements (C, N, O, U) are shown in Fig. 8a. A huge number of uranium elements were detected on the CCN-500 surface, indicating that the reduction reaction primarily occurs on the catalyst. Furthermore, the XPS spectra of CCN-500 after the photocatalytic reaction are shown in Fig. 8b. The signal can be quantitatively divided into two components according to the high-resolution spectra of U 4f. U(VI) and U(IV) resulted in peaks  at 381.34, 380.40 eV and 392.14, 391.15 eV, which were attributed to U 4f 7/2 and U 4f 5/2 of U(VI) and U(IV), respectively. This indicates that the uranium hexavalent was converted to uranium tetravalent in the solution at the photocatalytic reaction. Furthermore, 52.58% U(IV) and 47.42% U(VI) were detected using XPS analyses, which could be a result of reoxidation of UO 2 by h + or oxyhydrogen radicals.
Furthermore, PBQ (benzoquinone as the scavenger of ·O 2 − ) and IPA (isopropanol as the radical scavenger of ·OH) as the sacrificial agents were added to the reaction system to further investigate the active species in the process of photocatalytic reduction U(VI), and EDTA-2Na (ethylenediaminetetraacetic acid disodium salt the scavenger of h + ) [26,49,50]. As shown in Fig. 8c, the EDTA-2Na and PBQ exhibited a resisting effect on the photocatalytic reduction of uranium during the conversion of methanol to other sacrificial agents, suggesting that h + and ·O 2 − were the main active substances in the photocatalytic reduction process. Additionally, electron spin resonance also observed a strong signal of DMPO-·O 2 − after 5 min visible light irradiation, which confirmed that ·O 2 − was the key role of the photocatalytic system, as shown in Fig. 8d.
The aforementioned results suggested a possible schematic mechanism of U(VI) photo-immobilization. Furthermore, the electrons move to the conduction band and the holes move to the valence band in the catalyst surface under visible-light irradiation. The photo-induced electrons on the CCN-500 reduced the adsorbed O 2 on its surface to superoxide radicals (·O 2 − ), UO 2 2+ with e-and ·O 2 − formed the UO 2 subsided on the catalyst surface, respectively. Moreover, methanol became a sacrificial agent combined when with the holes, which improved the migration of the photogenerated electron-hole pairs effectively. Therefore, the significantly improved photocatalytic activity of the crystalline CCN-500 can be attributed to the synergistic effects of its electronic structure and morphology. Additionally, a mechanism was assumed for the photoreduction of U(VI) by CCN-500 with methanol that can be described by the following equation, and the corresponding schematic illustration is shown in Fig. 9.

Conclusions
In summary, we successfully synthesized CCN-X using molten salt heat treatment to improve the photocatalytic reduction ability of U(VI) to CN. CCN-X with proper preheating temperature has a higher photocatalytic activity for reducing U(VI) than CN under visible light irradiation. CCN-500 exhibits excellent photocatalytic reduction removal of U(VI) rate (0.132 min −1 ) under visible light irradiation in the experiment of photocatalytic reduction of U(VI), which is 2.5 times higher than that of bulk CN. The significant improvement in CCN-500 performance is mainly because of the improvement of crystallinity, the enhancement of photo-generated charge carrier mobility, and the increase in surface area. This study demonstrated the important role of crystallinity on carbon nitrides, showed the potential of using crystalline carbon nitride for photocatalytic reduction and removal of U(VI)and provided an efficient strategy to construct visible-lightresponsive photocatalysts for applicable treatment of uranium from wastewater.