Preparation of Co-CNK-OH and Its Performance in Fenton-like Photocatalytic Degradation of Tetracycline

: In this study, Co-modiﬁed alkalinized g-C 3 N 4 (named Co-CNK-OH) was prepared for the Fenton-like photocatalytic degradation of tetracycline (TC) via a simple yet effective calcination– impregnation method. In all samples of CNK-OH with different Co 2+ loadings, Co-CNK-OH catalyst with the optimal content (9%) exhibited the highest catalytic activity, with 87.1% tetracycline removal and 50% removal efﬁciency of the total organic carbon (TOC). Mechanism studies revealed that the 9%Co-CNK-OH catalyst had the lower electrical resistance after alkalization treatment and Co 2+ modiﬁcation, leading to a signiﬁcantly accelerated interfacial charge transfer to the electron acceptor as well as effectively separating electrons and holes. The intermediates generated during the TC degradation in the photo-Fenton process were detected by HPLC-MS, which proved that the holes, superoxide radicals, and singlet oxygen are the key reactive species in the Fenton-like photocatalysis. This study provides a new option for the treatment of TC in wastewater.


Introduction
In recent years, an endless variety of antibiotics has been synthesized and manufactured as therapeutic drugs to treat infectious diseases in humans, animals, and even plants around the world [1].Antibiotic-induced sewage pollution is highly concerned as a large quantity of antibiotics has entered water bodies with the widespread application of antibiotics.Tetracycline (TC), which is widely used to treat diseases in livestock animals, is difficult to degrade in nature, resulting in the continuous release and accumulation of TC in the natural water environment [2,3].Therefore, it is urgent to develop more effective methods for degrading tetracycline in wastewater treatment technologies.
Common methods used to degrade TC fall into several categories, such as biological treatment, adsorption and desorption, membrane treatment, and advanced oxidation technologies [4][5][6][7].Among the advanced oxidation technologies, the Fenton process is currently a popular method of water-pollution treatment.However, the classical Fenton process has many limitations on its practical applications, such as a relatively narrow pH range and heavy iron deposition, leading to increased treatment costs and low efficiency in the utilization of H 2 O 2 [8,9].Thus, Fenton-like technologies have increasingly become alternative choices [10][11][12].
Graphite-like carbon nitride (g-C 3 N 4 ), a common two-dimensional (2D) material, has attracted extensive research attention in the field of photocatalysis because of its nontoxicity, low cost, good stability, and visible-light response [13][14][15].Moreover, g-C 3 N 4 can stimulate the generation of hydroxyl radicals (•OH) from H 2 O 2 and is often used as a Fenton-like catalyst [16].However, the rapid recombination of electron-hole pairs in g-C 3 N 4 limits its application on the degradation of organic pollutants [17].To solve this problem, Thi Quyen et al. [18] improved the light absorption behavior by doping S in g-C 3 N 4 , which can effectively reduce the recombination rate of electron-hole pairs in g-C 3 N 4 .Bao et al. [19] prepared Cu-doped g-C 3 N 4 by a simple template-mediated supramolecularself-assembly method, in which the specific surface area of the g-C 3 N 4 was considerably increased while achieving enhanced light absorption and higher separation and transfer efficiency of the photogenerated carriers.The energy bandgap in g-C 3 N 4 can be adjusted by doping with metallic or nonmetallic elements to prevent the recombination of electrons and holes [14,[18][19][20][21][22].Alkalinization treatment is an important surface/interface modulation tool to reduce the energy bandgap by grafting hydroxyl groups, which can inhibit the recombination of photogenerated electrons and holes and improve the Fenton-like catalytic activity, with limited improvement in the photocatalytic performance [23][24][25].The combination of alkalization treatment and metal-ion modification can effectively enhance the performance of photocatalysis.For example, Cong et al. [26] improved the specific surface area and electrochemical activity by embedding Co ions in g-C 3 N 4 , and the produced Co-g-C 3 N 4 showed excellent catalytic performance in the PMS-activated degradation of TC.
To the best of our knowledge, a Co-doped g-C 3 N 4 alkalized catalyst has never been reported.In this work, a novel Co 2+ -doped g-C 3 N 4 alkalized catalyst (Co-CNK-OH) was prepared by a simple calcination-impregnation approach.On the one hand, the basic treatment can transplant numerous hydroxyl groups on the surface of the g-C 3 N 4 to generate active hydroxyl radicals and decrease the energy bandgap of the g-C 3 N 4 .On the other hand, the doping of metal ions can significantly improve the Fenton-like catalytic activity of the g-C 3 N 4 through the variable valence state of the metal ions.Compared with similar products, the Co-CNK-OH catalyst exhibits superior degradation performance for TC under visible-light irradiation.This study provides a new option for the treatment of tetracycline in wastewater.

Characterization of As-Prepared Samples
Figure 1a shows the XRD spectra of the as-prepared g-C 3 N 4 , CNK-OH, and Co-CNK-OH heterojunctions with different Co contents.Two typical characteristic peaks for g-C 3 N 4 were found at 13.0 • and 27.4 • , corresponding to the (100) and (200) crystal planes of the g-C 3 N 4 , respectively [27].The (100) plane at 13.0 • disappeared compared with that of the pure g-C 3 N 4 when the g-C 3 N 4 was alkalized with KCl and NH 4 Cl, implying that K + ions were embedded into the g-C 3 N 4 plane and an attraction force between K and N atoms.This was consistent with the results of the K 2p XPS spectra [28].As for the (200) plane, a slight shift in the peak from 27.4 • to 28.1 • for the CNK-OH indicated a smaller interlayer spacing of the (200) plane in the CNK-OH [16].The intensity of the diffraction peak gradually decreased with increasing Co 2+ content, indicating that Co 2+ was successfully doped in the framework of the g-C 3 N 4 and that the Co-doping partly damaged the in-plane structure of the g-C 3 N 4 [29].The XRD diffraction peak of the CNK-OH was basically unchanged, even with Co 2+ ions loading, indicating that Co 2+ ions might have been embedded in the center of the triazine ring unit [30,31].
FT-IR was used to investigate the chemical structures and compositions of the g-C 3 N 4 , CNK-OH, and 9%Co-CNK-OH materials (Figure 1b), and characteristic peaks at 800 cm −1 and between 1200 and 1700 cm −1 were observed for all the samples.The peak at 800 cm −1 was attributed to the typical vibration of the repeating triazine unit.The absorption band at 1200-1700 cm −1 could be assigned to the stretching vibrations of the aromatic CN heterocycles in the g-C 3 N 4 [32].Compared with the peaks for the g-C 3 N 4 , three additional peaks appeared at ~994 cm −1 , 1148 cm −1 , and 2180 cm −1 for the CNK-OH and 9%Co-CNK-OH by the introduction of K + ions.The peaks at 994 cm −1 and 1148 cm −1 corresponded to the vibrational modes of the O-H and C-O or C=O bonds, respectively, verifying that hydroxyl groups were effectively grafted on the surface of the g-C 3 N 4 by this alkalinization modification method [27].The band at 2180 cm −1 belonged to the cyano group (C≡N), which suggested that the introduction of K + ions ruptured the C-N-C bonds in the g-C 3 N 4 to form new C≡N bonds [16].The FT-IR spectrum of the 9%Co-CNK-OH sample was similar to that of the CNK-OH sample, and no evident change was found because of the metal modification.

Morphology and Structural Properties
To investigate the microstructures of the g-C 3 N 4 , CNK-OH, and 9%Co-CNK-OH materials, SEM and TEM images of the as-prepared samples are shown in Figure 2a-f.
To confirm the elemental distribution, the EDS and elemental mapping images of the Co-CNK-OH are shown (Figure 3a-g).It was obvious that the surface of the g-C 3 N 4 was smooth while the surface of the CNK-OH became rougher with dispersed particles after the alkalization treatment (Figure 2a,b), which may have been due to the influence of the grafted hydroxyl group on the surface of the g-C 3 N 4 [33].Compared with the CNK-OH and g-C 3 N 4 surfaces, smaller solid particles were observed, and the surface of the Co-CNK-OH was much rougher after the Co 2+ modification (Figure 2c).From the microstructure of the CNK-OH, numerous nanosheets on the surface could be clearly seen (Figure 2e).However, the modification with Co 2+ resulted in a more fragmented and uniformly dispersed nanosheet structure, in agreement with the SEM analysis results (Figure 2e,f).As for Co-CNK-OH, it was clear that C, N, O, K, and Co were uniformly distributed on the surface of the Co-CNK-OH material in Figure 3c-g, and the previously mentioned evidence implied the formation of Co-CNK-OH heterojunctions at the microscopic level.
To clarify the detailed chemical bonding and valence states of the constituent elements in Co-CNK-OH, the XPS spectra of the 9%Co-CNK-OH show the peaks of C 1s, N 1s, O 1s, K 2p, and Co 2p (Figure 4a).In the spectrum of C 1s (Figure 4b), the characteristic peaks at binding energy (BE) = 288.43 and 284.79 eV correspond to the N=C=N group and C=C bond of the g-C 3 N 4 , respectively [34].The weak peak at 286.27 eV was due to the C-OH group [31], confirming the grafting of the hydroxyl groups on the surface of the g-C 3 N 4 .Four characteristics peaks at 398.33, 400.90, 403.79, and 404.54 eV were found in the N 1s spectrum (Figure 4c).The characteristic peaks at 398.33, 400.90, and 403.79 eV could be ascribed to the C-N=C, N-(C) 3 , and C-N-H groups in the g-C 3 N 4 , respectively [14,35], while the peak at 404.54 eV was caused by the delocalization of the π-electrons in the g-C 3 N 4 [31].In the O 1s spectrum (Figure 4d), two peaks were fitted, in which the peak at 532.52 eV was assigned to the binding of water molecules, while the peak at 531.11 eV belonged to the OH bond [31,36], which further proved the implantation of the hydroxyl group in the g-C 3 N 4 .The K 2p spectrum depicts two characteristic peaks.The peak at 293.08 eV could be attributed to the N-K bond (Figure 4e), such as that in KN 3 , indicating that K + ions can break the C-N bond in g-C 3 N 4 and insert into the plane of g-C 3 N 4 by the mutual attraction between the K and N atoms.This is crucial to the subsequent grafting of the surface hydroxyl groups, in accordance with the results of the FTIR [16].As for the spectrum of Co 2p (Figure 4f), the characteristic peaks at 780.88 and 796.63 eV were the main peaks of Co 2p 3/2 and Co 2p 1/2 , respectively; BE of Co 2p 1/2 (796.63 eV) was lower than that of metallic cobalt (799 eV) while BE of Co 2p 3/2 (780.88 eV) was higher than that of cobalt oxide (779-780 eV), suggesting that the doped Co 2+ ions combined with the N atoms in the g-C 3 N 4 to form Co-N bonds [37][38][39]; the peaks at 784.34, 787.85, and 802.41 eV are satellite peaks, representing the vibrations of the Co 2+ ions [26, 40,41].From the combined results of the XRD, FT-IR, SEM, TEM, EDS, and XPS, it was concluded that Co 2+ ions were doped in CNK-OH.To confirm the elemental distribution, the EDS and elemental mapping images of the Co-CNK-OH are shown (Figure 3a-g).It was obvious that the surface of the g-C3N4 was smooth while the surface of the CNK-OH became rougher with dispersed particles after the alkalization treatment (Figure 2a, b), which may have been due to the influence of the grafted hydroxyl group on the surface of the g-C3N4 [33].Compared with the CNK-OH and g-C3N4 surfaces, smaller solid particles were observed, and the surface of the Co-CNK-OH was much rougher after the Co 2+ modification (Figure 2c).From the microstructure of the CNK-OH, numerous nanosheets on the surface could be clearly seen (Figure 2e).However, the modification with Co 2+ resulted in a more fragmented and uniformly dispersed nanosheet structure, in agreement with the SEM analysis results (Figure 2e, f).As for Co-CNK-OH, it was clear that C, N, O, K, and Co were uniformly distributed on the surface of the Co-CNK-OH material in Figure 3c-g, and the previously mentioned evidence implied the formation of Co-CNK-OH heterojunctions at the microscopic level.To confirm the elemental distribution, the EDS and elemental mapping images of the Co-CNK-OH are shown (Figure 3a-g).It was obvious that the surface of the g-C3N4 was smooth while the surface of the CNK-OH became rougher with dispersed particles after the alkalization treatment (Figure 2a, b), which may have been due to the influence of the grafted hydroxyl group on the surface of the g-C3N4 [33].Compared with the CNK-OH and g-C3N4 surfaces, smaller solid particles were observed, and the surface of the Co-CNK-OH was much rougher after the Co 2+ modification (Figure 2c).From the microstructure of the CNK-OH, numerous nanosheets on the surface could be clearly seen (Figure 2e).However, the modification with Co 2+ resulted in a more fragmented and uniformly dispersed nanosheet structure, in agreement with the SEM analysis results (Figure 2e, f).As for Co-CNK-OH, it was clear that C, N, O, K, and Co were uniformly distributed on the surface of the Co-CNK-OH material in Figure 3c-g, and the previously mentioned evidence implied the formation of Co-CNK-OH heterojunctions at the microscopic level.main peaks of Co 2p3/2 and Co 2p1/2, respectively; BE of Co 2p1/2 (796.63 eV) was lower than that of metallic cobalt (799 eV) while BE of Co 2p3/2 (780.88 eV) was higher than that of cobalt oxide (779-780 eV), suggesting that the doped Co 2+ ions combined with the N atoms in the g-C3N4 to form Co-N bonds [37][38][39]; the peaks at 784.34, 787.85, and 802.41 eV are satellite peaks, representing the vibrations of the Co 2+ ions [26, 40,41].From the combined results of the XRD, FT-IR, SEM, TEM, EDS, and XPS, it was concluded that Co 2+ ions were doped in CNK-OH.The N 2 adsorption-desorption isotherms and corresponding pore-size-distribution curves of the g-C 3 N 4 , CNK-OH, and 9%Co-CNK-OH were further analyzed (Figure 5), and all the samples belonged to type IV [42].
Catalysts 2023, 13, x FOR PEER REVIEW 6 of 18 The N2 adsorption-desorption isotherms and corresponding pore-size-distribution curves of the g-C3N4, CNK-OH, and 9%Co-CNK-OH were further analyzed (Figure 5), and all the samples belonged to type IV [42].From Table 1, g-C3N4, CNK-OH, and 9%Co-CNK-OH had specific surface areas of 15.279, 5.146, and 8.243 m 2 /g, respectively.The specific surface area of CNK-OH was smaller than that of g-C3N4, which could be attributed to the mixing of KCl with the g-C3N4 samples, whereby K + ions entered into g-C3N4 to inhibit N2 adsorption and partially block the mesopores [24]; however, a slightly increased specific surface area was obtained From Table 1, g-C 3 N 4 , CNK-OH, and 9%Co-CNK-OH had specific surface areas of 15.279, 5.146, and 8.243 m 2 /g, respectively.The specific surface area of CNK-OH was smaller than that of g-C 3 N 4 , which could be attributed to the mixing of KCl with the g-C 3 N 4 samples, whereby K + ions entered into g-C 3 N 4 to inhibit N 2 adsorption and partially block the mesopores [24]; however, a slightly increased specific surface area was obtained after Co 2+ loading.

Optical and Photoelectric Properties
The optical properties of the catalysts were measured by the UV-Vis diffuse reflectance spectra (Figure 6a).The absorption edge of g-C3N4 was located at about 460 nm, with a large bandgap energy of about 2.7 eV. Compared with the g-C3N4 and CNK-OH absorption edges, those of all the Co-CNK-OH catalysts with different Co contents shifted slightly toward the long-wave range, and the intensities of the visible-light absorbance were increased by the Co 2+ modification.The Eg values of the g-C3N4, CNK-OH, and Co-CNK-OH were deduced from the linear extension of the (ahν) 2 versus hv curve to the X-axis.The Eg values of g-C3N4, CNK-OH, 3%Co-CNK-OH, 6%Co-CNK-OH, 9%Co-CNK-OH, 12%Co-CNK-OH, and 15%Co-CNK-OH were 2.71, 2.69, 2.72, 2.71, 2.68, 2.69, and 2.69 eV, respectively (Table 2).PL emission spectroscopy is an effective method to detect the combination of photogenerated electrons and holes.The fluorescence intensity of g-C 3 N 4 was very strong, while the fluorescence intensity of the CNK-OH modified by alkalization was sharply reduced to about 1/8 of that of g-C 3 N 4 (Figure 6b, c).The PL intensity of Co-CNK-OH further decreased compared with that of CNK-OH, indicating that the introduction of Co 2+ ions could effectively promote the separation of electrons and holes.The PL intensity of 9%Co-CNK-OH reached the minimum value and then PL intensity increased instead while further increasing the Co content, suggesting that the optimum dopant was 9%Co-CNK-OH for the most efficient separation of electron-hole pairs.
Mott-Schottky curves were applied to detect the type of semiconductor.As shown in Figure 7, the observed curves of all the samples show significantly positive slopes, indicating that g-C 3 N 4 , CNK-OH, and Co-CNK-OH were all categorized as n-type semiconductors.As listed in Table 2, the E FB and E VB of g-C 3 N 4 , CNK-OH, and Co-CNK-OH were calculated by extending the linear part of C −2 = 0 to obtain E FB , E CB = E FB + 0.197 V (the standard electrode potential of Ag/AgCl), and E VB = E g + E CB [16].
Catalysts 2023, 13, x FOR PEER REVIEW 8 of 18 decreased compared with that of CNK-OH, indicating that the introduction of Co 2+ ions could effectively promote the separation of electrons and holes.The PL intensity of 9%Co-CNK-OH reached the minimum value and then PL intensity increased instead while further increasing the Co content, suggesting that the optimum dopant was 9%Co-CNK-OH for the most efficient separation of electron-hole pairs.Mott-Schottky curves were applied to detect the type of semiconductor.As shown in Figure 7, the observed curves of all the samples show significantly positive slopes, indicating that g-C3N4, CNK-OH, and Co-CNK-OH were all categorized as n-type semiconductors.As listed in Table 2, the EFB and EVB of g-C3N4, CNK-OH, and Co-CNK-OH were calculated by extending the linear part of C −2 = 0 to obtain EFB, ECB = EFB + 0.197 V (the standard electrode potential of Ag/AgCl), and EVB = Eg + ECB [16].Electrochemical impedance spectroscopy (EIS) is commonly used to probe the separation efficiency of electrons and resistance to interfacial charge transfer of semiconductor materials.The EIS curves of g-C3N4, CNK-OH, and Co-CNK-OH are shown in Figure 8.The arc radius of 9%Co-CNK-OH was significantly smaller than those of the other samples, implying that the electrical resistance of 9%Co-CNK-OH was the lowest, in accordance with the best electron-transfer performance and suggesting that Co 2+ modification significantly promote the interfacial charge to the electron acceptor and separate electrons Electrochemical impedance spectroscopy (EIS) is commonly used to probe the separation efficiency of electrons and resistance to interfacial charge transfer of semiconductor materials.The EIS curves of g-C 3 N 4 , CNK-OH, and Co-CNK-OH are shown in Figure 8.The arc radius of 9%Co-CNK-OH was significantly smaller than those of the other samples, implying that the electrical resistance of 9%Co-CNK-OH was the lowest, in accordance with the best electron-transfer performance and suggesting that Co 2+ modification significantly promote the interfacial charge to the electron acceptor and separate electrons and holes effectively [43].

Catalytic Performance and Mechanism of As-Prepared Samples
In order to evaluate the catalytic performance of Co-CNK-OH, the o ments were designed, and the results are shown in Figure S1.From Figure S reaction conditions for the TC degradation were as follows: the amount of c mg L −1 , the concentration of H2O2 was 30 mM, pH = 7, and the initial conce TC was 20 mg L −1 .Different reaction systems were applied to TC degradat in Figure 9a, with visible light alone or with visible light and H2O2, only a s radation of the TC occurred in 40 min.Adding g-C3N4 and CNK-OH led t tion efficiency of TC reaching 25.30 and 47.45%, respectively, indicating tha g-C3N4 promoted the degradation of TC because of the introduction of hyd It was noteworthy that the Co modification enhanced the degradation effic Co-CNK-OH.In detail, the degradation efficiency of TC increased with incr tent (3-9%) and then decreased with further increases in the Co content, w explained by the introduction of defects that act as a complex center for the [44].The highest-efficiency TC degradation with 9%Co-CNK-OH was 87.1 The leaching of Co 2+ ions after catalysis was also investigated by induc plasma mass spectrometry (ICP-MS).The result of the ICP-MS indicated th tration of the free Co 2+ in the solution after the reaction was 0.21 mg/L, w with China's emission standards for copper, cobalt, and nickel indu

Catalytic Performance and Mechanism of As-Prepared Samples
In order to evaluate the catalytic performance of Co-CNK-OH, the optimal experiments were designed, and the results are shown in Figure S1.From Figure the optimal reaction conditions for the TC degradation were as follows: the amount of catalyst was 50 mg L −1 , the concentration of H 2 O 2 was 30 mM, pH = 7, and the initial concentration of the TC was 20 mg L −1 .Different reaction systems were applied to TC degradation.As shown in Figure 9a, with visible light alone or with visible light and H 2 O 2 , only a slight selfdegradation of the TC occurred in 40 min.Adding g-C 3 N 4 and CNK-OH led to the degradation efficiency of TC reaching 25.30 and 47.45%, respectively, indicating that the alkalized g-C 3 N 4 promoted the degradation of TC because of the introduction of hydroxyl groups.It was noteworthy that the Co modification enhanced the degradation efficiency of TC by Co-CNK-OH.In detail, the degradation efficiency of TC increased with increasing Co content (3-9%) and then decreased with further increases in the Co content, which could be explained by the introduction of defects that act as a complex center for the charge carriers [44].The highest-efficiency TC degradation with 9%Co-CNK-OH was 87.10% in 40 min.The leaching of Co 2+ ions after catalysis was also investigated by inductively coupled plasma mass spectrometry (ICP-MS).The result of the ICP-MS indicated that the concentration of the free Co 2+ in the solution after the reaction was 0.21 mg/L, which complies with China's emission standards for copper, cobalt, and nickel industrial sources (GB25467-2010, 1 mg/L).To monitor the effect of the mineralization property of the photocatalyst on organic pollutants, the TOC removal efficiency of TC by the 9%Co-CNK-OH catalyst under visible-light irradiation was investigated (Figure 10a).The TOC removal efficiency of the TC by the 9%CO-CNK-OH catalyst was 50% after 60 min.The mineralization property of the 9%Co-CNK-OH catalyst is better than some previous reports (41%) [45].To figure out the mechanism of TC degradation by 9%Co-CNK-OH, TBA, EDTA-2Na, p-BQ, and FFA were used to capture •OH, h + , •O2 − and 1 O2, respectively (Figure 10b).Compared with the efficiency of TC removal (87.10%) by 9%Co-CNK-OH without any trapping-agent addition, the efficiency of TC degradation was reduced by 2.95%, 75.25%, 38.60%, and 17.98% with the addition of TBA, EDTA-2Na, p-BQ, and FFA, respectively, suggesting that h + , •O2 − , and 1 O2 are the main active species for the photocatalytically assisted Fenton-like degradation of TC by Co-CNK-OH.The reusability and stability of the catalyst are critical for practical applications.In order to evaluate the durability of the catalysts, three cycles of repeated experiments were carried out with 9%Co-CNK-OH under similar conditions (Figure 9b).It could be observed that the degradation efficiency of TC decreased from 87.1% to 77.95%, suggesting that the Co-CNK-OH sample had excellent stability.
To monitor the effect of the mineralization property of the photocatalyst on organic pollutants, the TOC removal efficiency of TC by the 9%Co-CNK-OH catalyst under visiblelight irradiation was investigated (Figure 10a).The TOC removal efficiency of the TC by the 9%CO-CNK-OH catalyst was 50% after 60 min.The mineralization property of the 9%Co-CNK-OH catalyst is better than some previous reports (41%) [45].To monitor the effect of the mineralization property of the photocatalyst on organic pollutants, the TOC removal efficiency of TC by the 9%Co-CNK-OH catalyst under visible-light irradiation was investigated (Figure 10a).The TOC removal efficiency of the TC by the 9%CO-CNK-OH catalyst was 50% after 60 min.The mineralization property of the 9%Co-CNK-OH catalyst is better than some previous reports (41%) [45].To figure out the mechanism of TC degradation by 9%Co-CNK-OH, TBA, EDTA-2Na, p-BQ, and FFA were used to capture •OH, h + , •O2 − and 1 O2, respectively (Figure 10b).Compared with the efficiency of TC removal (87.10%) by 9%Co-CNK-OH without any trapping-agent addition, the efficiency of TC degradation was reduced by 2.95%, 75.25%, 38.60%, and 17.98% with the addition of TBA, EDTA-2Na, p-BQ, and FFA, respectively, suggesting that h + , •O2 − , and 1 O2 are the main active species for the photocatalytically assisted Fenton-like degradation of TC by Co-CNK-OH.To figure out the mechanism of TC degradation by 9%Co-CNK-OH, TBA, EDTA-2Na, p-BQ, and FFA were used to capture •OH, h + , •O 2 − and 1 O 2, respectively (Figure 10b).Compared with the efficiency of TC removal (87.10%) by 9%Co-CNK-OH without any trapping-agent addition, the efficiency of TC degradation was reduced by 2.95%, 75.25%, 38.60%, and 17.98% with the addition of TBA, EDTA-2Na, p-BQ, and FFA, respectively, suggesting that h + , •O 2 − , and 1 O 2 are the main active species for the photocatalytically assisted Fenton-like degradation of TC by Co-CNK-OH.
The possible intermediates generated during TC degradation in the photo-Fenton process were investigated by HPLC-MS, as well as the possible TC-degradation pathway by 9%Co-CNK-OH (Figures 11 and 12    In view of this analysis, a possible mechanism for TC degradation by 9%Co-CNK-OH catalyst is proposed (Figure 13).(2) Co 3+ + e − → Co 2+ (5) OH was attributed to the synergistic effect of photocatalysis and Fenton reactions.The possible reactions are listed as follows: 9%Co-CNK-OH + hv → 9%Co-CNK-OH (e − + h + ) e

Preparation of CNK-OH
The steps for preparing alkalinized g-C 3 N 4 (CNK-OH) were the same as those followed by Li.First, 9 g of melamine, 45 g of KCl, and 0.6 g of NH 4 Cl were mixed and ground thoroughly in a mortar for 30 min.Then, the mixture was transferred to a ceramic crucible, heated to 550 • C at 2.3 • C/min in a muffle furnace, kept there for 4 h, and cooled naturally in the air.The obtained solid was ground into powder, dissolved in 200 mL of deionized water, stirred for 6 h, washed by vacuum filtration, dried in an oven at 80 • C for 12 h, and ground again to obtain the as-prepared CNK-OH.Under the same conditions, pure g-C 3 N 4 was prepared with 4.5 g of melamine but without KCl or NH 4 Cl.

Fabrication of Co-CNK-OH Composites
Co-CNK-OH composites were prepared by an impregnation method [16].First, 0.48 g of CNK-OH was dispersed in 100 mL of deionized water (denoted as solution A); a certain amount of CoCl 2 •6H 2 O was dissolved in 100 mL of deionized water (denoted as solution B).Then, solution B was added dropwise to solution A while stirred slowly, then stirred for 36 h at room temperature.The obtained suspension was sonicated for 3 h, washed three times each with deionized water and ethanol, and dried in an oven at 80 • C for 12 h to obtain the as-prepared Co-CNK-OH sample.The mass ratios of CoCl 2 •6H 2 O and CNK-OH were designed between 3 and 15%, and the Co-CNK-OH samples loaded with different Co contents are expressed as weight percentages of Co-CNK-OH for simplicity.Accordingly, 3%Co-CNK-OH, 6%Co-CNK-OH, 9%Co-CNK-OH, 12%Co-CNK-OH, and 15%Co-CNK-OH samples were prepared.

Degradation Experiments
First, a certain amount of the as-prepared catalyst was dispersed in 100 mL of a 20 mg/L TC solution and stirred magnetically for 30 min under dark conditions to establish adsorption-desorption equilibrium.A certain amount of H 2 O 2 was added, while a 300 W Xe lamp (λ ≥ 420 nm) was turned on at the same time, and about 3 mL of the suspension was taken out every 10 min, filtered through a 0.22 µm filter tip to remove the catalyst, and the absorbance was measured at 357 nm by spectrophotometry to evaluate the efficiencies of the TC degradation by the different catalysts.Every run was measured 3 times, and the relative error was less than 5%.

TC Mineralization Experiment
The sampling procedure for TC mineralization was based on that of the catalytic experiment.A total organic carbon analyzer was used to monitor the filtrate, and the TOC removal efficiency of TC was calculated with the following equation: TOC removal efficiency = TOC 0 − TOC t TOC 0 × 100% (9) where TOC 0 is the initial TOC value of the TC solution, and TOC t indicates the TOC value measured at any time t.

Conclusions
In this study, Co-CNK-OH photocatalysts were prepared by a calcination-impregnation method, and the 9%Co-CNK-OH showed the best degradation efficiency (87.1% TC degradation in 40 min and 50% TOC removal in an hour).The doping of Co 2+ expanded the absorption range of visible light and accelerated the transfer of the photogenerated electrons at the interface of the catalysts.The Co (III)/Co (II) redox cycle not only accelerated the consumption of the photogenerated electrons and then promoted the separation of the photogenerated electrons and holes but also accelerated the continuous generation of 1 O 2 , •O 2 − , and h + in the system.The addition of H 2 O 2 and the Co 2+ /Co 3+ redox cycle pro-

Figure 6 .
Figure 6.UV-vis diffuse reflectance patterns of g-C 3 N 4 , CNK-OH and Co-CNK-OH (outer figure), plots of (αhv) 2 versus hv of g-C 3 N 4 , CNK-OH and Co-CNK-OH (sub-figure) (a); PL emission spectra of g-C 3 N 4 , CNK-OH, and 9%Co-CNK-OH (b); and PL emission spectra of X%Co-CNK-OH (c).The absorption edge of g-C 3 N 4 was located at about 460 nm, with a large bandgap energy of about 2.7 eV. Compared with the g-C 3 N 4 and CNK-OH absorption edges, those of all the Co-CNK-OH catalysts with different Co contents shifted slightly toward the long-wave range, and the intensities of the visible-light absorbance were increased by the Co 2+ modification.The E g values of the g-C 3 N 4 , CNK-OH, and Co-CNK-OH were deduced from the linear extension of the (ahν) 2 versus hv curve to the X-axis.The E g values of g-

Figure 10 .
Figure 10.TOC changes in TC by 9%Co-CNK-OH under visible light (a); photo-Fenton degradation efficiencies of TC by 9%Co-CNK-OH with different scavengers under visible light (b).

Figure 10 .
Figure 10.TOC changes in TC by 9%Co-CNK-OH under visible light (a); photo-Fenton degradation efficiencies of TC by 9%Co-CNK-OH with different scavengers under visible light (b).

Figure 10 .
Figure 10.TOC changes in TC by 9%Co-CNK-OH under visible light (a); photo-Fenton degradation efficiencies of TC by 9%Co-CNK-OH with different scavengers under visible light (b).

Figure 11 .
Figure 11.Mass spectra of tetracycline and its corresponding intermediates during photo-Fenton degradation of TC.

Figure 11 .
Figure 11.Mass spectra of tetracycline and its corresponding intermediates during photo-Fenton degradation of TC.
First, numerous photogenerated electrons and holes in the conduction and valence bands were generated under visible-light irradiation and an internal electric field for the 9%Co-CNK-OH samples and •O 2 − was formed.The addition of H 2 O 2 promoted the formation of both 1 O 2 and •O 2 − .The Co (III)/Co (II) redox cycle accelerated the consumption of the photogenerated electrons and then promoted the separation of the photogenerated electrons and holes while keeping the continuous generation of 1 O 2 , •O 2 − , and h + in the system [50,51].Therefore, TC degradation by 9%Co-CNK-OH was attributed to the synergistic effect of photocatalysis and Fenton reactions.The possible reactions are listed as follows: 9%Co-CNK-OH + hv → 9%Co-CNK-OH (e − + h + ) (1) e − + O 2 → •O 2 −

Figure 13 .
Figure 13.Possible photocatalytic mechanism of TC degradation by Co-CNK-OH in Fenton-like photocatalytic systems.

Table 1 .
Specific surface areas and pore-size parameters of g-C 3 N 4 , CNK-OH and 9%Co-CNK-OH.

Table 2 .
E g , E FB , E CB, and E VB of g-C 3 N 4 , CNK-OH and Co-CNK-OH.