In the last decade, the potential environmental risks of antibiotics have raised worldwide attention [1, 2]. Therefore, the elimination of antibiotics in wastewater has become an important issue for human health and environmental protection. Among them, tetracycline antibiotics are the second largest class of antibiotics, which are widely produced and used in human and veterinary medicine [3]. Some tetracycline antibiotics may be strongly carcinogenic or tumorigenic. Various methods, including biodegradation, adsorption, the Fenton reaction, and photochemical technology [4-6], aimed specifically at promoting the elimination of antibiotics. However, these methods have obvious disadvantages, including poor stabilities and high prices of the materials, which have limited their practical applications in industry [7, 8]. Therefore, the development of new technology for antibiotic treatment in wastewater is urgent.
In recent years, visible light photocatalysts have received widespread attention and been extensively used in various fields for environment pollution treatment [9]. Photocatalysis, which involves catalysts with strong oxidation abilities, can enable complete mineralization for the degradation of tetracycline antibiotics [10, 11]. Several new types of photocatalysts have been developed, such as TiO2 [12], CdS [13], WO3 [14, 15], and BiOI [16]. However, the majority of these photocatalytic materials have low visible-light photocatalytic activities and stabilities [17], which are still far from satisfactory. The semiconductor photocatalyst g-C3N4, as a metal-free two-dimensional conjugated semiconductor, is produced from inexpensive raw materials and is harmless and non-toxic. g-C3N4 is physically and chemically stable, and the stability is tunable [18]. It has been recognized as a promising photocatalyst [19, 20]. However, g-C3N4 has an sp2 periodic interlayer melon arrangement hybrid array of characteristic units and weak van der Waals (vdW) attractions, which causes difficulties for electron transfer [21, 22].
During the photocatalytic process, electron-hole recombination is easily realized on the catalyst surface, leading to limited photocatalytic activity [23-25]. To solve this problem, many strategies have been developed, such as inner architecture modification [24, 27, 28] and surface functionalization (surface modification and formation of heterojunctions) [29-33]. Surface modification of g-C3N4 can improve the optical and charge transfer properties. For inner architecture modification, amorphous carbon nitride has a much smaller bandgap than the partially crystalline graphitic carbon nitride. This increases the light absorption ability, which leads to an enhanced photocatalytic efficiency, such as black phosphorus and Pt [34-37]. These surface and amorphous modifications are useful strategies for reducing the recombination rate of the charge carriers and transport across the interfaces. However, studies related to alkaline metal oxide modified g-C3N4 are scant. Thus, the g-C3N4 structure was properly designed and customized in this study.
To reduce the recombination rate of electrons-holes and extend the charge carrier lifetimes, a series of Mg/O co-decorated amorphous carbon nitride samples were prepared by a simple one-step in-situ co-pyrolysis method. The graphitic carbon nitride with an amorphous structure achieved short-range electron transport. This created electron-hole separation and suppressed charge recombination, thereby improving the photocatalytic activity. We proposed that MgO could modify the surface of carbon nitride and new active centers could be formed. The adsorption capacity and O2 activation were promoted to generate radicals and participate in the photocatalytic degradation of the antibiotic tetracycline hydrochloride (TC). In the present study, a simple and effective modification method to increase the photocatalytic activity was developed. The Mg/O co-decorated amorphous carbon nitride had a significant effect on the TC degradation. It is crucial to understand the mechanism of the photocatalytic reaction to promote the development of photocatalytic technology.
Urea, magnesium carbonate tetrahydrate, L-ascorbic acid, and ethylenediamine tetraacetic acid disodium salt were purchased from Kelong Chemical Agents (China). Tetracycline hydrochloride was purchased from Biosharp Deionized Chemical Agents (China). Isopropanol was purchased from Chongqing Horizon Chemical Industry (China). Water was used in all the experiments.
The MgO-CN composites were synthesized as follows. First, 10 g of magnesium carbonate tetrahydrate was dissolved in 30 mL of deionized water and ultrasonically stirred for 30 min. This solution was dried at 80 ℃ for 10 h, after which the precursor in an alumina crucible was heated at 580 ℃ for 4 h (20 ℃/min), and MgO powder was obtained. A certain amount of the prepared MgO powder and 10 g of urea were added to deionized water and ultrasonically stirred for 1 h. The solid composite precursor was dried at 60 ℃ for 10 h and calcined at 550 ℃ for 2 h (10 ℃/min) to yield MgO-CN powder. Composites with different mass ratios of MgO to urea were labeled MgO-CN-0.4, MgO-CN-0.8, MgO-CN-1.2, and MgO-CN-1.5, corresponding to mass ratios of 0.4, 0.8, 1.2, and 1.5 wt%, respectively.
X-ray diffraction (XRD, D8 Advance, Bruker Co., Germany) with Cu Kα radiation was used to characterize the crystal structures and phase compositions of the synthesized samples. Fourier-transform infrared spectroscopy (FT-IR, IR Prestige-21, Shimadzu, Japan) was employed to characterize the chemical bonds present on the composite surface. Scanning electron microscopy (SEM, JSM-7000F, JEOL, Japan) and transmission electron microscopy (TEM, JEM-2100F, JEOL, Japan) were used to examine the morphologies and microstructures of the products. N2 adsorption-desorption analysis (BET-BJH, ASAP 2020, Micromeritics) at –196 ℃ was employed to determine the surface areas and pore size distributions of the prepared samples. UV-vis diffuse reflectance spectroscopy (UV-vis DRS, UV-2550, Shimadzu, Japan), with 100% barium sulfate as a reference, was used to analyze the optical properties and band gaps. The photoluminescence (PL, F-7000, HITACHI, Japan) spectra were obtained to probe the charge-recombination rates of the electron-hole pairs of the samples. Electron spin resonance (ESR) measurements were used to identify the active species in the photodegradation process.
The photocatalytic behavior of the prepared composite samples was assessed by the degradation removal efficiency of TC in aqueous solutions. In a typical run, 100 mg of photocatalyst and 100 mL TC solution (30 mg/L) were stirred with a magnetic bar to achieve adsorption-desorption equilibrium in the dark. After adsorption equilibrium, the sample was exposed to visible light irradiation for a certain time using a 12 W LED lamp (> 400 nm) to initiate the photocatalytic reaction. At time intervals, a certain amount of solution was collected and immediately centrifuged at 5000 rpm, after which these aliquots were analyzed using a UV/Vis spectrophotometer (Shimadzu Inc., Japan). The characteristic absorption band of TC is at 357 nm [38]. The degradation rate (η) of TC is calculated as η= (1–Ct/C0) × 100%, where C0 denotes the initial concentration of the TC solution and Ct denotes the concentration of the TC solution at time t.
Isopropyl alcohol (IPA), ascorbic acid (AA), and ethylenediamine tetraacetic acid disodium salt (EDTA-2Na) were utilized to scavenge ·OH radicals, ·O2– radicals, and holes, respectively. First, 100 mL of TC solution (30 mg/L) and 0.1 g of photocatalyst were stirred with a magnetic bar to achieve the adsorption-desorption equilibrium in the dark. The mixed solutions with different trapping agents were irradiated under visible light to initiate the photocatalytic reactions. The samples used in the ESR experiments were dissolved in a 50 mmol/L 5, 5-dimethyl-1-pyrroline-N-oxide (DMPO) solution tank (aqueous dispersion for DMPO-·OH and methanol dispersion for DMPO-·O2–) before irradiation by visible light.
DFT-D2 calculations were carried out with the "Vienna ab initio simulation package" (VASP5.4) [39]. A plane-wave expansion basis set with a cut-off energy of 450 eV was employed. We used 3×3×1 Monkhorst-pack k-point grid for the structural optimization and to calculate the electron structure. Before the electronic properties were determined, the atoms were relaxed until the residual force was less than 0.03 eV/Å .
The XRD patterns of MgO are shown in Fig. 1(a). The sharp peaks at 36.9°, 42.9°, 62.3°, 74.7°, and 78.6° correspond to the (111), (200), (220), (311), and (222) planes, respectively, and were the characteristic peaks of MgO [40]. The patterns of the CN and MgO-CN samples with different mass ratios of MgO are shown in Fig. 1(b). Pure g-C3N4 has two typical diffraction peaks at 13.1° (in-plane tri-s-triazine units) and 27.6° (interplanar staking of aromatic segments), which are attributed to the (100) and (002) lattice planes [41-43], respectively. The signature diffraction peaks of MgO were not observed because of the low loading mass of the MgO nanoparticles in the composites. With increased loading amounts, the intensities of the CN signals decreased. These results indicated that the aromatic systems of the in-plane periodicity were destroyed. Accordingly, the layers with periodic accumulation were re-disturbed, and amorphous carbon nitride was generated.
Fig. 2 shows the FT-IR spectra of the CN, MgO, and MgO-CN-1.2 samples. The typical IR peaks of g-C3N4 were detected at 810 and 1230–1640 cm–1, which were attributed to the breathing mode of the triazine units and the characteristic stretching modes of aromatic C–N (C(sp2)-N, C(sp2)=N) heterocycles, respectively [44, 45]. The peak in the 3000–3500 cm– 1 region was assigned to the N–H stretching and bending vibrations. The strong and broad peaks of MgO located in the 450–600 cm–1 region were attributed to the stretching vibrations. The absorption peaks at 1643 and 3418 cm–1 were assigned to the bending vibrations of water [46], respectively. The band at 1488 cm–1 was assigned to the carbonate ions on the surface of the sample due to the absorption from carbon dioxide in the atmosphere [47]. The vibration peaks of the MgO-CN-1.2 composite were slightly weaker than those of pure g-C3N4. Moreover, for the modified photocatalyst, the intensities of these peaks decreased, and the N–H stretching at 3160 cm–1 also weakened due to the addition of Mg. Less exposure of the surface amino groups was observed [48, 49]. Note that the peak at 2160 cm–1 represents the N=C=N stretching vibrations [50], indicating that the C-N heterocycles were deformed. This result also implies that the insertion of MgO in the urea polymerization only destroys the periodic interlayer melon arrangement but simultaneously maintains the basic strand atomic structures. As a result, these characteristics can afford unique amorphous arrangements of short-range order and long-range disorder of CN.
The morphologies CN, MgO, and MgO-CN-1.2 were analyzed by SEM and TEM. As shown in Fig. 3(a–c), a stack of porous thin nanosheets was observed in the SEM image of CN. The MgO was composed of bulk and dense thick layers. The MgO-CN sample manifested a flake and sheet-shaped morphology. The TEM images, as shown in Fig. 3(d–e), showed the microstructures of the samples. The CN consisted of layered nanosheets. The lattice fringes of the MgO samples can be clearly observed. The HRTEM image in Fig. 3(e) shows that the lattice of MgO-CN-1.2 was not detected, which further clarified the formation process of the amorphous carbon nitride. This TEM result implies that the introduction of MgO can induce the amorphization of pristine CN, and the difference in the photocatalytic activity was not caused by the morphology.
The surface chemical compositions and bonding configurations of the CN and MgO-CN-1.2 composites were investigated by XPS. In the XPS spectra, the binding energies of Mg and O were identified. The C 1s spectrum (Fig. 4(a)) deconvoluted into peaks at 284.6 and 288.1 eV, corresponding to adventitious hydrocarbons and sp2-bonded carbon in N–C=N groups, respectively. The N 1s spectra of the CN and MgO-CN-1.2 composites were usually fitted into four peaks (Fig. 4(b)). The peaks at 398.4 and 399.9 eV indicated the formation of sp2 hybridized aromatic nitrogen (C=N–C groups) and hybridized tertiary nitrogen (N–(C)3 groups), respectively. The peaks at 401.4 and 404.1 eV were not obvious. This can be also observed in the O 1s spectra (Fig. 4(c)). The peaks at 532.9 eV in the O 1s spectra were attributed from surface adsorbed water. Compared to pure CN, an additional peak at 531.8 eV in the O 1s spectrum of MgO-CN-1.2 appeared, which was ascribed to N-C-O formation during calcination with air. The XPS peaks of the Mg 2p spectra (Fig. 4(d)) of the MgO-CN-1.2 composites were located at 49.8 eV.
Fig. 5(a) shows the BET surface area measurement of the samples. All samples were classified into a type-Ⅳ curve with H3 hysteresis loops at relative pressures ranging from 0.5 and 1.0, corresponding to mesoporous structures. Fig. 5(b) shows the relative pore-size distributions [51]. For comparison, the specific surface areas and total pore volumes of the CN, MgO, and MgO-CN-X samples are listed in Table 1. The specific surface area of CN was 129.13 m2/g after the coupling with MgO. The surface areas of the MgO-CN-X samples gradually decreased, as the MgO could plug the pore structures, leading to low surface areas and slight decreases of the pore sizes. Based on the photocatalytic activities for TC degradation, the specific surface area and porous structure were not the critical factors for improving the photocatalytic activity.
The optical properties of CN and MgO-CN-1.2 were investigated by UV-vis diffuse reflectance spectroscopy. As shown in Fig. 6(a), an adsorption band up to 225 nm appeared for MgO. The MgO-CN-1.2 exhibited intensified absorption over the entire range of wavelengths due to Mg and O co-decoration. The Eg values of the MgO-CN-1.2 and pristine CN were obtained from the (αhυ)1/2 vs. hυ plots shown in Fig. 6(b), which indicate that the band gaps of pristine CN, MgO, and MgO-CN-1.2 were 2.75, 5.39, and 2.48 eV, respectively. The conduction band (CB) and valence band (VB) positions of CN were –1.15 and 1.60 eV, respectively, whereas those of MgO-CN-1.2 were –1.01 and 1.47 eV. The VB and CB edges of MgO-CN exhibited down-shifts compared to those of pristine CN. The calculated densities of states (DOSs) are shown in Fig. 6(c), which were used to estimate the band energy. Compared with pristine CN, the band gap of MgO-CN-1.2 narrowed significantly and shifted to a lower energy level, which agreed with the UV-vis results. The reduced band gap of MgO-CN-1.2 can be ascribed to the MgO doping. The narrowed band gap reduced the energy required to excite electrons. Fig. 6(d) shows the PL spectra of CN and MgO-CN-1.2. MgO-CN-1.2 had lower fluorescence intensity, which is beneficial for separation efficiency. Mg/O co-decorated amorphous carbon nitride can achieve an optimized band structure and enhance the efficiency of electron-hole separation.
DFT was further employed to determine the electronic structures to analyze the effect of Mg/O co-decorated amorphous carbon nitride on the electron delivery. Fig. 7(a) and 7(b) show the electronic structure of the Mg/O co-decorated g-C3N4. The charge difference distribution of the MgO-CN calculated using DFT is shown in Fig. 7(c). We found that the O atom received electrons from adjacent C, N, and Mg atoms, which was attributed to the differences in electronegativities between the C, N, Mg, and O elements. Inducing an O atom into the CN could aggregate electrons and provide active sites to promote the activation of O2 and H2O molecules on the catalyst surface to generate reactive species. The electron localization function (ELF, Fig. 7(d)) implies that Mg/O co-decoration can provide electronic localization, which could have specific transport orbitals (C→O←Mg) for intralayer electron transfer.
Fig. 8(a) shows the photo-degradation efficiency of TC with the CN and MgO-CN composites. The CN exhibited a lower activity than the composites after 2.5 h of irradiation with visible light. The degradation of TC treated with MgO-CN-1.2 reached about 82%, which was higher than that of the pure g-C3N4 (23.5%). The fits to the kinetics of the TC degradation by g-C3N4 and MgO-CN-X (where X stands for the MgO percentage) are shown in Fig. 8(b). MgO-CN-1.2 exhibited the highest photodegradation efficiency (0.01018 min–1), which was five times higher than that of the CN (0.00205 min–1). The photocatalytic activity results showed that after the introduction of MgO, the photocatalysis efficiency of the composites was greatly enhanced. Fig. 8(c) shows the UV-vis spectra of TC using the MgO-CN-1.2 composite photocatalyst for the specific time intervals. The decrease in the absorbance characteristic peak at 357 nm with the treatment time indicated that the TC was degraded under visible light irradiation for 2.5 h. Based on the UV-vis analysis of the TC aqueous solutions, it can be concluded that the MgO-CN-1.2 can effectively degrade TC via photocatalysis. Meanwhile, the stability of the photocatalyst is one of the essential aspects for advanced photocatalytic performance. We examined the stability of the CN and MgO-CN-1.2 composite photocatalyst with five cycles. The result shows that the MgO-CN-1.2 exhibited a slight decrease of the photocatalytic activity (Fig. 8(d)). The slight decrease in the degradation rate was possibly due to the partial loss of the photocatalyst that occurred during recovery.
The reactive species (·OH, ·O2–) and holes (h+) are responsible for the photocatalytic oxidation. To investigate the effect of these reactive species on the photodegrading of TC, radical trapping experiments were conducted. As shown in Fig. 9, the photocatalytic activity of MgO-CN-1.2 indicated that the addition of IPA, AA, or EDTA-2Na inhibited the photocatalysis to different degrees. Moreover, the addition of AA (·O2 scavengers) has the most significant influence on decreasing the photocatalytic degradation efficiency of TC. These results suggest that ·OH, ·O2, and h+ play important roles in the photodegradation process of TC.
According to the degradation studies, the MgO-CN composites exhibited greater photoactivities than g-C3N4. The radical trapping experiment results suggested that ·OH, ·O2, and h+ enhanced the photocatalytic process of TC degradation. To further determine the amounts of reactive species present for photocatalytic TC degradation, the reactive species were identified by the ESR spin-trap. As shown in Fig. 10(a), MgO-CN-1.2 exhibited much stronger ·O2− and ·OH signals than pure g-C3N4 while reacting under visible light. The results were associated with the unique electronic structure of the Mg/O co-decorated amorphous carbon nitride. Furthermore, the improvement was beneficial for electron excitation, which prolongs the time of electron-hole recombination. Moreover, the red shift of the adsorption edge broadened the field of light adsorption, which enhanced the utilization of light. Thus, an improvement of the TC degradation was obtained. The separation of the electron-hole pairs had a significant impact on the photocatalytic reaction. The photogenerated electrons and holes were trapped by oxygen and surface hydroxyl groups on the surface of the photocatalyst to form ·O2– and ·OH.
The mechanism of the photocatalytic TC degradation by the MgO-CN catalysts under visible light is illustrated in Fig. 11. In agreement with the related reports [51-53], we found that the VB holes (1.47 eV) from MgO-CN were unable to oxidize the OH–/H2O to ·OH radicals. However, ·OH was detected, as shown in Fig. 10(b). Thus, ·OH can form through O2 reduction at the CB (–1.01 eV) via the path ·O2– → H2O2 → ·OH [54, 55]. The photocatalytic degradation mechanism is as follows:
We reported that the Mg/O co-decorated amorphous carbon nitride exhibited a high photocatalytic activity for TC degradation. Based on detailed experimental data and theoretical calculations, the mechanism of the photocatalytic activity enhancement was proposed. The Mg and O co-decoration turned the periodic interlayer melon arrangement of g-C3N4 into an amorphous phase, increasing the light absorption ability. The electron transport distance decreased, which was favorable for the separation of electron-hole pairs. The Mg and O co-decoration created a unique structure that could generate localized electrons (O atoms caused electron aggregation). Moreover, the reactant activation capacity was enhanced via the C→O←Mg route to create favorable conditions for the O2 molecule activation on the catalyst surface. The reactive species were produced via the path O2 → ·O2→ H2O2 → ·OH. The ESR spectra and DFT calculations revealed the mechanism of the photocatalytic degradation of TC by the Mg/O co-decorated amorphous carbon nitride. The present work can provide new ideas and understanding of catalyst modification for large-scale environmental and remediation applications.
This work was financially supported by the Specialized Innovation of Social and People's Livelihood in Chongqing (cstc2016shmszx20012), the Converting Outstanding Achievements of University-Funded Projects of Chongqing (KJZH17122), the National Natural Science Foundation of China (5160080705), the Key Laboratory Open Project from Chongqing Technology and Business University (1556036), the Scientific and Technological Research Program of Chongqing Municipal Education Commission (KJ130704), and the Innovative Research Project from Chongqing Technology and Business University (yjscxx2016-060-34).
2019,
Vol. 40 Issue (5): 776-785 DOI: S1872-2067(19)63300-0 
PDF
2019,
Vol. 40 Issue (5): 776-785 DOI: S1872-2067(19)63300-0 PDF