English

• Antibiotics have been crucial in treating and preventing infectious diseases in both humans and livestock [1,2]. Oxytetracycline (OTC), a broad-spectrum antibiotic, is widely used in animal farming and aquaculture worldwide [3]. Unfortunately, a large proportion of OTC and its metabolites are released into water environments, disrupting environmental microbial communities and posing serious health risks to humans through the food chain [4]. Traditional wastewater treatment methods are ineffective at removing antibiotics due to their stable chemical structures and poor biodegradability [5-7]. Therefore, it is necessary to develop economical and efficient methods for the elimination of OTC.

  In recent years, peroxymonosulfate (PMS)-based advanced oxidation processes (AOPs) have gained significant attention for their effectiveness in removing contaminants from water [8,9]. PMS can be activated through various methods such as heating, ultraviolet irradiation and transition metal catalysis, resulting in the generation of hydroxyl radicals (^•OH), sulfate radicals (SO₄^•‒) and singlet oxygen (¹O₂) for pollutant degradation [10,11]. Among various transition metal activators, Co-based catalysts facilitate a redox cycle by altering the valence state of Co, leading to the production of highly reactive SO₄^•‒ used in AOPs [9,12,13]. However, traditional Co-based materials only utilize the exposed surfaces at the metal active sites for redox reactions, with most internal atoms remaining uninvolved in catalytic processes, leading to limited atom utilization [14]. To address these limitations, single atom catalysts (SACs) have emerged as a cutting-edge solution for PMS activation in heterogeneous catalysts due to their maximum atom utilization [15,16], effectively addressing the aforementioned limitations. SACs are catalysts with immobilized and monodisperse metal atoms on a solid support, allowing for systematic variations in metal atoms-support interactions and coordination environments [17,18]. Hence, selecting an appropriate supporting substrate is crucial in obtaining SACs with exceptional catalytic performance.

  In order to prevent the metal leaching during the catalytic process, carbon-based non-metallic nanomaterials are commonly chosen as supports for SACs. Graphene or graphene-like carbon materials are typical substrates for SACs due to their exceptional physicochemical properties [19]. However, the migration barriers and binding energies of transition metal adatoms on pristine graphene typically range from 0.2 eV to 0.8 eV and 0.2 eV to 1.5 eV, respectively, indicating that adatoms on the dangling-bond-free graphene sheets are highly mobile. This mobility causes metal atoms supported on pristine graphene to form aggregates at room temperature [20]. To stabilize the metal adatoms on graphene, additional modifications are generally required to create anchoring sites, making the preparation of high-performance SACs cumbersome. Alternatively, graphite nitride carbon (g-C₃N₄), with its abundant lone pair electrons and cavity structure, has garnered widespread attention in the field of SACs [21]. The N/C-coordinating framework in g-C₃N₄, characterized by tri-s-triazine structural units, possesses abundant nitrogen coordinators with lone pair electrons, which are beneficial for anchoring metal atoms on the g-C₃N₄ carrier. This results in Co-N coordination with higher binding energies [22,23]. More importantly, Co-N coordination has been proven to be effective in activating PMS, leading to efficient catalytic performance [18]. In addition, g-C₃N₄ support is a good semiconductor photocatalyst which can generate abundant photo-generated electrons under light irradiation, further favoring PMS activation. Therefore, g-C₃N₄ is a promising substrate for developing highly effective Co-SACs for PMS activation, owing to the synergistic interplay between Co-N coordination and visible light irradiation.

  Recent studies have shown promising results for g-C₃N₄ supported Co-SACs in PMS activation, indicating their potential as highly effective catalysts for organic pollutant degradation [24,25]. However, challenges such as complex synthesis methods, structural agglomeration, low specific surface area, and sluggish charge transfer still limit their practical application. Further research is needed to optimize the synthesis methods and improve the performance of Co-SACs for PMS activation. In contrast, g-C₃N₄ with three-dimensional (3D) interconnected structure provides a high specific surface area for the dispersion of Co atoms and ensures good structural stability, preventing stacking during PMS activation reactions [26,27]. This results in stable and efficient catalytic performance. Additionally, while previous studies have primarily focused on the intrinsic catalytic activity of g-C₃N₄ supported Co-SACs in the absence of light irradiation, the synergistic effect between Co-N coordination and light irradiation has not been adequately addressed. Therefore, the development of a simple and effective method for preparing a 3D porous g-C₃N₄-supported Co-SACs with high catalytic performance under visible light irradiation, and the exploration of the synergistic effect between Co-N coordination and visible light irradiation, is both meaningful and challenging.

  In this study, we employ a facile thermal polymerization method using a melamine-cyanuric acid supramolecular (MCS) precursor to disperse Co atoms on a 3D interconnected g-C₃N₄ (SACo-CN) substrate. Characterization results and density functional theory (DFT) calculations confirm that the Co atoms are chemically bonded to the N atoms of g-C₃N₄. This enhances the visible light absorptivity of the SACo-CN catalyst and facilitates the efficient transfer and separation of charge carriers. Moreover, the SACo-CN catalysts exhibited remarkable catalytic activity and stability in the degradation of OTC due to their 3D interconnected structure and the synergistic interplay between Co-N coordination and visible light irradiation. The dominant active species are identified by trapping experiments and electron spin resonance (ESR) spectroscopy, and possible mineralization pathways were elucidated with liquid chromatography-mass spectrometry (LC-MS). This study not only reveals the catalytic mechanism of Co-SACs but also provides a novel approach for synthesizing efficient metal-SACs for environmental remediation.

  To further investigate the advantages of Co-N coordination in g-C₃N₄ supported Co-SACs, we calculated the differential charge densities, Bader charge populations and PMS activation energy barriers for graphene- and g-C₃N₄-supported Co-SACs based on DFT calculations, as shown in Fig. 1 and Table S1 (Supporting information). The yellow and green regions in Figs. 1a and b reveal electron accumulation and depletion, respectively, indicating electron transfer from Co atoms to g-C₃N₄ and graphene substrates. Bader charge population analysis (Table S1) shows that the numbers of electrons lost by Co atoms in graphene- and g-C₃N₄-supported Co-SACs are 0.751 and 0.675 e, respectively. This implies that electron transfer in g-C₃N₄-supported Co-SACs is easier to achieve than in graphene-supported Co-SACs due to the stability of Co-N coordination, which is beneficial to the activation of PMS. This is further confirmed by the lower PMS activation energy barrier for g-C₃N₄-supported Co-SACs (0.56 eV) compared to graphene-supported Co-SACs (0.98 eV), as shown in Figs. 1c and d. Additionally, the abundance of photo-generated electrons in g-C₃N₄-supported Co-SACs under light irradiation favors the activation of PMS. Therefore, the g-C₃N₄ demonstrates promise as an optimal substrate for developing highly effective Co-SACs for PMS activation, owing to the synergistic interplay between Co-N coordination and visible light irradiation.

  Figure 1

  

  Figure 1.  (a, b) Differential charge densities, (c, d) energy profiles of PMS activation reaction for graphene and g-C₃N₄ supported Co-SACs.

  DownLoad: Full-Size Img PowerPoint

  Fig. 2a illustrates the detailed synthetic process for producing SACo-CN catalysts. The process starts with forming an MCS precursor, which features a robust 3D framework due to H-bonding between melamine and cyanuric acid in an aqueous solution (Fig. S1 in Supporting information) [28]. Simultaneously, Co ions are uniformly distributed on the surface of the 3D cross-linking complex precursor or scattered between molecular groups. During the subsequent calcination treatment, thermal polycondensation releases chemical gases such as NH₃, creating abundant porosity within the 3D network structure. This self-templating process of the intricate precursor enables the precise formation of 3D interconnected g-C₃N₄. High-temperature calcination treatment ensures the firm anchoring of Co atoms on the surface of g-C₃N₄, resulting in a uniform dispersion of Co atoms on the 3D interconnected g-C₃N₄ samples.

  Figure 2

  

  Figure 2.  (a) Schematic illustration of the synthesis of SACo-CN catalysts. (b) TEM images, (c) HAADF-STEM images (HRTEM images insert), and (d) elemental mapping images of SACo-CN 1.5. (e) Co K-edge X-ray absorption near-edge structure (XANES) spectra, (f) Fourier transform extended X-ray absorption fine structure (FT-EXAFS) spectra, (g) wavelet transform (WT) plots of the SACo-CN 1.5, Co foil, CoO, and Co₃O₄. (h) FT-EXAFS fitting curves of the SACo-CN 1.5 at Co K-edge. (i) k space fitting curves of the SACo-CN 1.5.

  DownLoad: Full-Size Img PowerPoint

  The crystal phase structures of the catalysts were determined using X-ray diffraction (XRD) analysis. As depicted in Fig. S2 (Supporting information), two diffraction peaks at 12.7° and 27.8° can be attributed to the (100) and (002) planes of g-C₃N₄, respectively [29]. These two peaks are also present in the SACo-CN samples, with no significant peak shifts after the introduction of Co atoms, indicating that the fundamental structure of g-C₃N₄ remains unchanged and that Co elements do not integrate into the lattice of g-C₃N₄. Furthermore, no diffraction peaks corresponding to zero-valent cobalt or cobalt oxide are detected in any samples, suggesting that Co atoms are uniformly dispersed on the g-C₃N₄ matrix [30]. Fig. S3 (Supporting information) presents the Fourier transform infrared (FT-IR) spectra of the catalysts. For g-C₃N₄, the absorption peaks around 2800–3500 cm⁻¹ are attributed to the N–H stretches of terminal amino groups and O–H stretching vibrations of adsorbed H₂O molecules. The signals in the 1200–1700 cm⁻¹ region are associated with the stretching mode of aromatic C–N groups, and the signal around 810 cm⁻¹ is linked to the out-of-plane bending mode of the 3, s-triazine units [31,32]. Notably, the emergence of a new peak around 2168 cm⁻¹ in SACo-CN samples suggests the stretching vibration of -CN, indicating possible coordination bond between cobalt and the CN aromatic ring [33,34].

  The morphologies of g-C₃N₄ and SACo-CN samples were examined using scanning electron microscopy (SEM). Fig. S4 (Supporting information) illustrates the 3D porous structure of g-C₃N₄, prepared through the MCS precursor thermal polymerization method, which displays a loose and interconnected network. Similarly, SEM images in Fig. S5 (Supporting information) reveal that the SACo-CN samples also exhibit a 3D crosslinked network structure akin to pure g-C₃N₄. Energy-dispersive X-ray (EDX) analyses in Fig. S6 (Supporting information) confirm that the introduction of Co atoms does not alter the fundamental morphologies of the g-C₃N₄ samples. The 3D network structure, assembled by cyanuric acid and melamine, endows g-C₃N₄ with abundant pore structures and a substantial specific surface area. Fig. S7 (Supporting information) shows that both samples exhibit gradual adsorption in the middle relative pressure range and significant adsorption at 0.9 < P/P₀ < 1, suggesting the presence of mesopores and macropores [35,36]. These pore structures within the 3D network facilitate mass transfer and diffusion, thereby enhancing the photocatalytic performance [25]. The surface area of SACo-CN 1.5 is 67.202 m²/g, significantly higher than that of g-C₃N₄ (49.897 m²/g), as indicated in Table S2 (Supporting information). The large specific surface areas of these 3D cross-linked structures not only minimize the likelihood of Co atom agglomeration but also ensure a large active area for effective interaction between the catalysts and target pollutants.

  The transmission electron microscopy (TEM) was employed to reveal the microstructure of SACo-CN 1.5 catalyst. The TEM image in Fig. 2b shows that a 3D cross-linked and curled structure, consistent with the findings from the SEM images. To further determine the dispersion state of Co elements in the SACo-CN 1.5 catalyst, spherical aberration-corrected high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) was utilized. The HAADF-STEM image in Fig. 2c shows many single-atom sized bright spots anchoring onto the g-C₃N₄ carrier, along with some tiny atomic clusters. The energy-dispersive X-ray spectroscopy (EDS) mapping images in Fig. 2d indicate the uniform distribution of Co elements on the g-C₃N₄ framework.

  The surface chemical elements of the SACo-CN 1.5 catalyst were investigated by X-ray Photoelectron Spectroscopy (XPS). According to the survey spectrum in Fig. S8a (Supporting information), the dominant eleFigments in the SACo-CN 1.5 sample are C and N, with a small amount of O detected, which may be due to the absorption of H₂O or CO₂ on the material surface [37,38]. In the high-resolution C 1s spectrum of SACo-CN 1.5 (Fig. S8b in Supporting information), the two fitted peaks at 284.9 eV and 288.1 eV correspond to C–C/C═C and N═C–N bonds, respectively [39]. Fig. S8c (Supporting information) illustrates the high-resolution N 1s spectrum, with peaks centered at 398.6 eV, 399.8 eV, 401.2 eV and 404.4 eV assigned to the C═N–C, N-(C)₃, C-NH_x and π-excitations of g-C₃N₄, respectively [40]. It is noteworthy that the peak positions of N-(C)₃ and C-NH_x shift significantly to lower binding energies after the introduction of Co atoms, suggesting that the Co atoms may be immobilized by these N atoms, thereby altering the electron distribution of g-C₃N₄ [41]. The high-resolution Co 2p spectrum shown in Fig. S8d (Supporting information) reveals fitted four peaks at 781.1 eV, 796.6 eV, 787.2 eV and 801.9 eV corresponded to the Co 2p_3/2, Co 2p_1/2 states and two satellite peaks, respectively. This indicates that the Co species in SACo-CN 1.5 are in the form of Co²⁺ rather than metallic Co⁰, indexed to the Co-N species [42].

  The X-ray absorption fine structure (XAFS) analysis was employed to investigate the local coordination environments of Co atoms in the SACo-CN 1.5 catalyst. As displayed in Fig. 2e, the near-edge absorption energy of SACo-CN 1.5 closely resembles that of CoO, indicating that Co atoms primarily exist in a + 2 valence state. In the extended X-ray absorption fine structure (EXAFS) spectrum (Fig. 2f) of SACo-CN 1.5 sample, the first shell peak at approximately 1.61 Å is attributed to the Co-N coordination. Notably, no Co-Co or Co-O bonds are observed, further confirming the atomic dispersion of Co in the SACo-CN 1.5 catalyst. Additionally, the wavelet transform (WT) plots of SACo-CN 1.5 distinguish from Co oil, CoO, and Co₃O₄ references, providing further evidence that the Co sites are atomically dispersed on the g-C₃N₄ substrate and coordinated through Co-N chemical bonds (Fig. 2g). According to the results of quantitative EXAFS fitting analysis (Fig. 2h and Table S3 in Supporting information), the Co sites are coordinated with N atoms, and the coordination number for Co is 4.0, thus forming a Co-N4 structure in the SACo-CN 1.5 sample. The k-space fitting curves of the SACo-CN 1.5 are presented in Fig. 2i, and the good fitting results demonstrate the accuracy and reliability of the conclusions provided above.

  The excitation of photogenerated electrons plays a pivotal role for catalytic performance. To demonstrate the solar light absorption capability of the catalyst, UV–vis diffuse reflectance spectra were used. As shown in Fig. 3a, the SACo-CN catalysts exhibit stronger absorption than pure g-C₃N₄ across the entire measurement range. The absorption edges of the composites continuously shift to longer wavelengths with increasing Co atoms, indicating that the absorptivity of the SACo-CN catalysts are enhanced by the supported Co atoms [43]. Furthermore, Fig. S9 (Supporting information) demonstrates that the calculated bandgap of SACo-CN 1.5 (2.41 eV) is significantly lower than that of g-C₃N₄ (2.67 eV), which may be attributed to the formation of Co-CN-x hybrids with narrow bandgap energies [44]. To further elucidate the changes in the band structure caused by the introduction of Co atoms, Mott-Schottky (MS) measurement was conducted to determine the Fermi energy level (E_f) of the samples [45]. As illustrated in Fig. S10 (Supporting information), the E_f values for g-C₃N₄ and SACo-CN 1.5 are −0.93 and −0.85 V (vs. NHE), respectively. Additionally, the XPS valence band (VB) spectra are shown in Fig. S11 (Supporting information), and the VB potentials of g-C₃N₄ and SACo-CN 1.5 are 2.63 and 2.38 V (vs. NHE), respectively. Therefore, the VB energies of g-C₃N₄ and SACo-CN 1.5 are calculated as 1.71 and 1.53 V (vs. NHE). Combined with the bandgaps obtained above, the conduction band (CB) energies of the g-C₃N₄ and SACo-CN 1.5 are determined as −0.98 and −0.88 V (vs. NHE). Fig. 3b presents a plot of the energy band diagrams for the two catalysts, leading to the conclusion that loading a significant amount of Co atoms greatly enhances the utilization of solar energy.

  Figure 3

  

  Figure 3.  (a) The diffuse reflectance spectra, (b) band structure diagram, (c) electrochemical impedance spectroscopy, (d) photocurrent, (e) steady-state and (f) time-resolved PL decay curves of g-C₃N₄ and SACo-CN samples. LUMO and HOMO distributions of (g) g-C₃N₄, (h) SACo-CN 1.5. Activation pathways of PMS on (i) g-C₃N₄ and (j) SACo-CN 1.5.

  DownLoad: Full-Size Img PowerPoint

  The investigation of photogenerated carrier separation and transfer ability in g-C₃N₄ and SACo-CN catalysts was conducted using electrochemical impedance spectroscopy (EIS) and transient photocurrent response. In the EIS spectra, the semicircle is attributed to charge transfer resistance (R_ct) and constant phase element (CPE). The R_ct is primarily associated with the resistance at the interface between the photocatalyst and electrolyte. The corresponding equivalent circuit in Fig. 3c illustrates this relationship, with the inset highlighting the differences in EIS semicircle diameters between SACo-CN catalysts and g-C₃N₄. This disparity indicates that incorporation Co atoms enhance carrier migration, leading to accelerated transport properties [46]. The smallest EIS diameter is observed for the SACo-CN 1.5 catalyst, implying enhanced catalytic performance. However, an excess of Co atoms, as seen in the case of SACo-CN 2.0, leads to increased charge transfer resistance, possibly due to the agglomeration of excess Co atoms, which increases internal migration resistance [47]. Fig. 3d shows that the SACo-CN 1.5 catalyst displays a higher photocurrent intensity than pure g-C₃N₄, demonstrating efficient photoelectron-hole separation efficiency and low charge recombination rate [48]. In Fig. 3e, the photoluminescence (PL) intensities of SACo-CN samples are remarkably reduced with the introduction of atomically dispersed Co, indicating enhanced charge carrier separation efficiency. Additionally, the time-resolved photoluminescence (TRPL) decay experiment was conducted, as presented in Fig. 3f and Table S4 (Supporting information). The calculated average fluorescence decay lifetime of SACo-CN 1.5 (3.84 ns) is shorter than that of g-C₃N₄ (5.26 ns), revealing the properties of effective charge carrier dissociation and migration in SACo-CN 1.5 catalyst [49]. These photoelectrochemical properties indicate that loading of Co atoms significantly promotes the generation of photogenerated electrons, accelerates charge transfer, and reduces the recombination of photoelectron-hole pairs. Consequently, multiple photogenerated electrons and holes excited by visible light can be transferred to the conduction and valence band of g-C₃N₄, which then react with PMS to produce reactive oxygen radicals and used for OTC degradation.

  DFT calculations were conducted to investigate the impact of Co atom loading on the electron distribution and charge carrier mobility of the SACo-CN 1.5 sample, with the results depicted in Figs. 3g and h. For g-C₃N₄, the highest occupied molecular orbital (HOMO) is composed of the 2p orbitals of the N atoms, while the lowest unoccupied molecular orbital (LUMO) is primarily dominated by 2p states of the C atoms. In Fig. 3g, the HOMO and LUMO densities exhibit a regular distribution. However, for SACo-CN 1.5 sample (Fig. 3h), the formation of Co-N bonds disrupts this even distribution, indicating that the loading of Co atoms affects the electronic structure. Charge delocalization generally enhances electron mobility [50], thus the electronic transitions and conduction of the SACo-CN 1.5 catalyst are significantly accelerated after the loading with Co atoms [51].

  The transition states and energy profiles for PMS activation by the g-C₃N₄ and SACo-CN 1.5 catalyst are shown in Figs. 3i and j. Comparison with experimental data reveals that the PMS activation energy barrier for SACo-CN 1.5 (0.56 eV) is noticeably lower than that of g-C₃N₄ (1.12 eV), implying that the PMS activation is kinetically more facile. In addition, the product energies are situated below that of the reactant at −2.06 eV, demonstrating that the activation reaction is thermodynamically exothermic and occurs spontaneously [52]. Furthermore, the calculated adsorption energies listed in Table S5 (Supporting information) suggest that PMS adsorption on SACo-CN 1.5 is more favorable than on g-C₃N₄. Consequently, PMS molecules are more easily activated by the SACo-CN 1.5 catalyst, thereby promoting the degradation of pollutants.

  During the degradation experiments, OTC was used as a target contaminant to evaluate the catalytic performance of the synthesized samples under visible light irradiation. As shown in Figs. S12 and S13 (Supporting information), the capacity for OTC adsorption on various catalysts is almost negligible in the dark. Moreover, under visible light alone, there is no substantial reduction in the presence of OTC. In Fig. 4a and Fig. S14 (Supporting information), the SACo-CN catalysts demonstrate superior catalytic performance compared to g-C₃N₄ under visible light irradiation. The k value of SACo-CN 1.5 is 4.76 times that of g-C₃N₄ (0.0315 min⁻¹), indicating that the introduction of Co atoms (Co-N) is an effective strategy to improve photocatalyst activity. Besides, compared with dark conditions (Fig. S15), the higher OTC removal efficiency of SACo-CN catalysts under visible light emphasizes the crucial role of visible light in further enhancing catalytic degradation. Thus, the highest OTC degradation efficiency (98.8% in 30 min) can be achieved through the synergistic effect of Co-N coordination and visible light irradiation, surpassing many previously reported g-C₃N₄-based catalysts as highlighted in Table S6 (Supporting information). However, when more Co atoms are introduced into g-C₃N₄, the catalytic degradation efficiency decreases, which may be attributed to the accumulation of excessive Co atoms that hinder the photoelectron excitation of the catalyst [44].

  Figure 4

  

  Figure 4.  Degradation efficiency of OTC under different (a) catalysts, (b) PMS concentrations, (c) SACo-CN 1.5 dosages, (d) OTC concentrations, (e) inions, (f) solution pH, and (g) real water bodies. (h) Cycle experiments for OTC degradation. (i) Catalytic degradation of common pollutants by SACo-CN 1.5 catalyst. Experimental conditions unless otherwise specified: [PMS] = 0.5 mmol/L, [catalyst] = 0.4 g/L, [OTC]₀ = 15 mg/L, visible light irradiation.

  DownLoad: Full-Size Img PowerPoint

  As shown in Fig. 4b, increasing the PMS concentration from 0.1 to 0.5 mmol/L significantly raises the degradation efficiency from 75.9% to 98.8% within 30 min. However, boosting the PMS concentration to 1.0 mmol/L only marginally improves catalytic performance, possibly due to excessive free radical self-quenching or dose limitation of the SACo-CN 1.5 catalyst hindering further PMS activation. A similar pattern is observed when exploring the impact of SACo-CN 1.5 dose on OTC degradation (Fig. 4c). Increasing the dose from 0.2 g/L to 0.4 g/L leads to a sharp rise in the elimination rate, but further elevating it to 0.5 g/L shows no change in removal rate due to PMS dose constraints. Notably, the degradation rate reaches 92.3% at 25 mg/L OTC (Fig. 4d), indicating that the SACo-CN 1.5/PMS/vis degradation system maintains excellent catalytic performance even at higher concentrations of organic pollutants.

  To explore the impact of common anions (Cl⁻, H₂PO₄⁻, SO₄²⁻ and NO₃⁻) on the catalytic degradation of antibiotics, 5 mmol of these anions were introduced (Fig. 4e and Table S7 in Supporting information). The results, analyzed by the Box-Behnken experimental methodology and depicted in the 3D surface plots in Fig. S16, indicate that the catalytic system shows strong resistance to anionic interference. The influence of solution pH on the elimination of OTC by the SACo-CN 1.5/PMS system was explored, and the results are shown in Fig. 4f. At pH 3, the OTC degradation rate is slightly inhibited due to the stabilization effect of H⁺ on HSO₅⁻ [53]. Conversely, alkaline conditions enhance the decomposition and activation of PMS, significantly improving the catalytic degradation of OTC. Overall, the SACo-CN 1.5 catalyst maintains excellent OTC degradation performance across a wide pH range of 3–11. Three potential pathways for the mineralization of OTC were determined through a combination of HPLC-MS technology and DFT calculations (Figs. S17, S18 and Tables S8-S10 in Supporting information). Furthermore, according to the results of Toxicity Estimation Software Tool (T.E.S.T.), the potential toxicities of chemical intermediates are reduced during the photocatalytic process (Fig. S19 and Table S11 in Supporting information).

  In Fig. 4g, the degradation of OTC remains largely unaffected in real water environments, such as lake water, river water, and tap water, underscoring the potential applicability of the SACo-CN 1.5/PMS/vis degradation system in practical water settings. Cyclic experiments depicted in Fig. 4h reveal the exceptional durability of the SACo-CN 1.5/PMS degradation system, with the removal rate of OTC still reaching 90% after five cycles. The SEM image of the SACo-CN 1.5 catalyst after five cycles, shown in Fig. S20 (Supporting information), reveals that the catalyst has retained its 3D cross-linked porous structure, suggesting that the catalyst exhibits remarkable reusability throughout the catalytic degradation process. Moreover, with morphology regulation and the loading of the Co atoms, the catalytic degradation efficiencies for metronidazole (MNZ), ofloxacin (OFX), tetracycline (TC), and methyl orange (MO) achieve 95.1%, 87.8%, 97.5%, and 99.8%, respectively (Fig. 4i). This highlights the capability of the SACo-CN 1.5/PMS/vis system to degrade various organic pollutants.

  To ascertain the contributions of reactive species to the catalytic degradation process, capture experiments were conducted. As depicted in Fig. S21a (Supporting information), EtOH (0.2 mol/L), isopropyl alcohol (IPA, 0.2 mol/L), benzoquinone (BQ, 5.0 mmol/L), L-histidine (10.0 mmol/L), and ethylenediaminetetraacetic acid disodium (EDTA-2Na 2 mmol/L) were used as capture agents for SO₄^•‒, ^•OH, O₂^•−, ¹O₂, and h⁺, respectively [54,55]. It was observed that both EtOH and IPA show slight inhibitory effects on OTC degradation, suggesting that SO₄^•‒ and ^•OH are not the primary active species in the catalytic process. In contrast, evident deceleration is noted in the presence of EDTA-2Na, L-histidine or BQ, indicating the involvement of O₂^•−, ¹O₂ and h⁺ in the degradation reaction. The dominant role of h⁺ was evidenced by the strongest inhibitory effect with EDTA-2Na. Furthermore, the presence of the involved species was confirmed through ESR spectroscopy. As shown in Figs. S21b-d (Supporting information), characteristic peaks of these species were not observed in the dark. In contrast, the typical signals of ^•OH, SO₄^•‒ (Fig. S21b), ¹O₂ (Fig. S21c) and O₂^•− (Fig. S21d) were detected under light, verifying the generation of reactive species during OTC degradation under visible light irradiation.

  Based on the experimental results and theoretical calculations, the SACo-CN 1.5/PMS/vis system exhibits efficient catalytic degradation of multiple organic contaminants. The related mechanisms can be summarized in three aspects: (1) The 3D cross-linked structures endow the catalyst with a large specific surface area (67.20 m²/g) for the loading of Co atoms, increasing the possibility of more exposed reaction sites. (2) The incorporation of Co atoms enhances the absorption ability of SACo-CN 1.5 catalysts for visible light, facilitates the separation and transfer of photogenerated electrons, and inhibits the recombination of photoelectron-hole pairs. (3) The synergistic effect between Co-N coordination and visible light irradiation facilitates electron transfer and conduction, improves PMS adsorption on the catalyst, and lowers the energy barrier for PMS activation. This results in highly effective PMS activation and exceptional catalytic performance in the elimination of contaminants. The proposed mechanism of OTC degradation is illustrated in Fig. 5. The incorporation of Co atoms enhances the absorption ability of photocatalysts and modifies the band structure, facilitating the excitation and generation of electrons and holes under visible light irradiation (Eq. 1). The CB position of SACo-CN 1.5 is at −0.88 V vs. NHE, which is more negative than the redox potential of O₂/O₂^•− (−0.33 V vs. NHE). Thus, the e⁻ on CB react with O₂ to produce O₂^•− radicals (Eq. 2), providing the raw material for the reactive species. Subsequently, the h⁺ on the VB react with O₂^•− to generate ¹O₂ radicals (Eq. 3) [56]. In addition, the Co atoms play an important role in the decomposition of PMS through the Co²⁺/Co³⁺ cycle and the production of SO₄^•‒ (Eqs. 4 and 5) [57]. SO₄^•‒ can directly react with H₂O in solution to form ^•OH (Eq. 6). Finally, all these radicals contribute to the mineralization of OTC (Eq. 7).

  $ \mathrm{SACo}-\mathrm{CN} 1.5+h v \rightarrow \mathrm{e}^{-}+\mathrm{h}^{+} $

  (1)

  $ \mathrm{e}^{-}+\mathrm{O}_2 \rightarrow \mathrm{O}_2^{\bullet-} $

  (2)

  $ \mathrm{h}^{+}+\mathrm{O}_2^{\bullet-} \rightarrow{ }^1 \mathrm{O}_2 $

  (3)

  $ \mathrm{Co}^{2+}+\mathrm{HSO}_5^{-} \rightarrow \mathrm{Co}^{3+}+\mathrm{SO}_4^{\bullet-}+\mathrm{OH}^{-} $

  (4)

  $ \mathrm{Co}^{3+}+\mathrm{e}^{-} \rightarrow \mathrm{Co}^{2+} $

  (5)

  $ \mathrm{SO}_4^{\bullet-}+\mathrm{H}_2 \mathrm{O} \rightarrow \mathrm{SO}_4^{2-}+{ }^{\bullet} \mathrm{OH}+\mathrm{H}^{+} $

  (6)

  $ { }^{\bullet} \mathrm{OH} / \mathrm{SO}_4^{\bullet-} / \mathrm{O}_2^{\bullet-} /{ }^1 \mathrm{O}_2 / \mathrm{h}^{+}+\mathrm{OTC} \rightarrow \text { products }+\mathrm{H}_2 \mathrm{O}+\mathrm{CO}_2 $

  (7)

  Figure 5

  

  Figure 5.  Schematic for possible mechanism of OTC degradation by SACo-CN 1.5/PMS/vis system.

  DownLoad: Full-Size Img PowerPoint

  In summary, 3D interconnected g-C₃N₄ catalysts with dispersed Co atoms have been successfully synthesized via the pyrolysis of melamine-cyanuric acid supramolecules. The atomically dispersed Co atoms significantly enhance the electronic transitions and conductivity of the catalyst while reducing the energy barrier for PMS activation. The SACo-CN/PMS/vis system demonstrated remarkable resistance to external interference, exceptional cycling durability, and high removal rates for common organic pollutants. This is attributed to the 3D interconnected structure and the synergistic interplay between Co-N coordination and visible light irradiation. The degradation rates of MNZ, OFX, TC, and MO are 95.1%, 87.8%, 97.5%, and 99.8% within 30 min, respectively. During the catalytic reaction, h⁺ are identified as predominant active species for the degradation of OTC. Additionally, the biotoxicitiy of degradation intermediates of OTC can significantly reduced after treatment with the SACo-CN/PMS/vis system, ensuring the biological safety of the products. This work provides a green and efficient approach for eliminating organic pollutants and demonstrates the feasibility of SACs-PMS/vis system for future application in wastewater treatment.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  CRediT authorship contribution statement

  Tianjun Ni: Writing – original draft, Supervision, Investigation. Hui Zhang: Writing – original draft, Methodology, Supervision. Liping Zhou: Writing – original draft, Supervision, Software. Roujie Ma: Writing – original draft, Software, Investigation. Yanyu Wang: Writing – original draft, Methodology. Zhijun Yang: Writing – review & editing, Data curation. Dan Luo: Writing – review & editing, Data curation. Nithima Khaorapapong: Writing – review & editing, Data curation. Xingtao Xu: Writing – review & editing, Software, Investigation, Supervision. Yusuke Yamauchi: Writing – review & editing, Investigation, Software, Supervision. Dong Liu: Writing – original draft, Supervision, Software, Methodology.

  Acknowledgments

  The authors acknowledge financial support from the National Natural Science Foundation of China (Nos. 22276159, J2224005), the Key research project plan for higher education institutions of Henan province (No. 24ZX009), the Development Program for Key Young Teachers in Colleges and Universities of Henan Province (No. 2020GGJS146), and the Starting Research Fund of Xinxiang Medical University (No. XYBSKYZZ201911). We express our gratitude for the English editing software, such as Wiley Editing Services and ChatGPT, for refining the language and checking grammatical errors in our manuscript. Additionally, this research project was partially supported by the office of the Ministry of Higher Education, Science, Research and Innovation, Thailand under the Reinventing University 2024 Visitting Prefessor Program, and the Queensland node of the NCRIS-enabled Australian National Fabrication Facility (ANFF).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.110659.

  
  
• 1. [1]

     J.S. Wang, X.-H. Yi, X. Xu, et al., Chem. Eng. J. 431 (2022) 133213. doi: 10.1016/j.cej.2021.133213

  2. [2]

     J. Guo, L. Wang, X. Wei, et al., J. Hazard. Mater. 415 (2021) 125591. doi: 10.1016/j.jhazmat.2021.125591

  3. [3]

     W. Qi, J. Long, C. Feng, et al., Water Res. 160 (2019) 361–370. doi: 10.1016/j.watres.2019.05.058

  4. [4]

     S. Nazari, E. Asgari, A. Sheikhmohammadi, et al., J. Environ. Chem. Eng. 11 (2023) 110393. doi: 10.1016/j.jece.2023.110393

  5. [5]

     Y. Tang, J. Kang, M. Wang, et al., J. Environ. Chem. Eng. 9 (2021) 105864. doi: 10.1016/j.jece.2021.105864

  6. [6]

     Y. Liu, L. Tian, Z. Wang, et al., J. Environ. Chem. Eng. 11 (2023) 110723. doi: 10.1016/j.jece.2023.110723

  7. [7]

     S. Wang, J. Wang, Sep. Purif. Technol. 277 (2021) 119457. doi: 10.1016/j.seppur.2021.119457

  8. [8]

     J. Guo, Y. Wang, Y. Shang, et al., PNAS 121 (2024) e2313387121. doi: 10.1073/pnas.2313387121

  9. [9]

     C. Chen, M. Yan, Y. Li, et al., Appl. Catal. B: Environ. 340 (2024) 123218. doi: 10.1016/j.apcatb.2023.123218

  10. [10]

      Y. Gao, Y. Zhou, S.Y. Pang, et al., Water Res. 193 (2021) 116856. doi: 10.1016/j.watres.2021.116856

  11. [11]

      S. Liu, Y. Hu, H. Xu, et al., Appl. Catal. B: Environ. 335 (2023) 122882. doi: 10.1016/j.apcatb.2023.122882

  12. [12]

      C. Jin, M. Wang, Z. Li, et al., Chem. Eng. J. 398 (2020) 125569. doi: 10.1016/j.cej.2020.125569

  13. [13]

      Q. Wang, Z. Xu, Y. Cao, et al., Chem. Eng. J. 427 (2022) 131953. doi: 10.1016/j.cej.2021.131953

  14. [14]

      Y. Shang, Y. Kan, X. Xu, Chin. Chem. Lett. 34 (2023) 108278. doi: 10.1016/j.cclet.2023.108278

  15. [15]

      M. Yang, Z. Hou, X. Zhang, et al., Environ. Sci. Technol. 56 (2022) 11635–11645. doi: 10.1021/acs.est.2c01261

  16. [16]

      J. Guo, B. Gao, Q. Li, et al., Adv. Mater. 36 (2024) 2403965. doi: 10.1002/adma.202403965

  17. [17]

      X. Wei, S. Song, W. Cai, et al., ACS Nano 18 (2024) 4308–4319. doi: 10.1021/acsnano.3c09819

  18. [18]

      X. Wu, S. Liu, Y. Li, et al., Angew. Chem. Int. Ed. 62 (2023) e202305639. doi: 10.1002/anie.202305639

  19. [19]

      J. Li, W. Xia, J. Tang, et al., J. Am. Chem. Soc. 144 (2022) 9280–9291. doi: 10.1021/jacs.2c00719

  20. [20]

      H. Wang, Q. Wang, Y. Cheng, et al., Nano Lett. 12(1) (2012) 141–144. doi: 10.1021/nl2031629

  21. [21]

      Y. Shen, M.J. Martín de Vidales, G. Gorni, et al., Chem. Eng. J. 444 (2022) 136610. doi: 10.1016/j.cej.2022.136610

  22. [22]

      M. Xie, J. Tang, L. Kong, et al., Chem. Eng. J. 360 (2019) 1213–1222. doi: 10.1016/j.cej.2018.10.130

  23. [23]

      S. Singh, P. Kaur, Abhivyakti, et al., Appl. Surf. Sci. 580 (2022) 152254. doi: 10.1016/j.apsusc.2021.152254

  24. [24]

      J. Li, S. Zhao, S.Z. Yang, et al., J. Mater. Chem. A 9 (2021) 3029–3035. doi: 10.1039/d0ta11503g

  25. [25]

      Y. Xu, H. Tang, L. Liu, et al., Chem. Eng. J. 450 (2022) 138180. doi: 10.1016/j.cej.2022.138180

  26. [26]

      X. Zhang, J. Li, J. Li, et al., Chem. Eng. J. 385 (2020) 123394. doi: 10.1016/j.cej.2019.123394

  27. [27]

      C. Li, J. Yin, B. Cao, et al., Ceram. Int. 49 (2023) 6213–6221. doi: 10.1016/j.ceramint.2022.10.081

  28. [28]

      Y. Sheng, Z. Wei, H. Miao, et al., Chem. Eng. J. 370 (2019) 287–294. doi: 10.1016/j.cej.2019.03.197

  29. [29]

      T. Ni, H. Zhang, Z. Yang, et al., J. Colloid. Interf. Sci. 625 (2022) 466–478. doi: 10.1016/j.jcis.2022.06.057

  30. [30]

      Z. Zhao, W. Zhang, W. Liu, et al., Chem. Eng. J. 407 (2021) 127167. doi: 10.1016/j.cej.2020.127167

  31. [31]

      D. Liu, C. Li, C. Zhao, et al., Chem. Eng. J. 438 (2022) 135623. doi: 10.1016/j.cej.2022.135623

  32. [32]

      C. Zhou, P. Xu, C. Lai, et al., Chem. Eng. J. 359 (2019) 186–196. doi: 10.1016/j.cej.2018.11.140

  33. [33]

      M. Li, F. Zhang, M. Kuang, et al., Nano-Micro Lett. 15 (2023) 238. doi: 10.1007/s40820-023-01188-1

  34. [34]

      Y. Zou, J. Li, J. Tan, et al., Chem. Eng. J. 471 (2023) 144531. doi: 10.1016/j.cej.2023.144531

  35. [35]

      M. Jourshabani, Z. Shariatinia, A. Badiei, J. Mater. Sci. Technol. 34 (2018) 1511–1525.

  36. [36]

      Y. Liang, W. Xu, J. Fang, et al., Appl. Catal. B: Environ. 295 (2021) 120279. doi: 10.1016/j.apcatb.2021.120279

  37. [37]

      X. Yu, S.F. Ng, L.K. Putri, et al., Small 17 (2021) 2006851. doi: 10.1002/smll.202006851

  38. [38]

      Y. Chen, Y. Qu, X. Zhou, et al., ACS Appl. Mater. Interf. 12 (2020) 41527–41537. doi: 10.1021/acsami.0c11578

  39. [39]

      Z. Jin, Y. Zhang, D. Liu, et al., Sep. Purif. Technol. 277 (2021) 119224. doi: 10.1016/j.seppur.2021.119224

  40. [40]

      T. Xu, D. Wang, L. Dong, et al., Appl. Catal. B: Environ. 244 (2019) 96–106. doi: 10.1016/j.apcatb.2018.11.049

  41. [41]

      H. Li, C. Shan, B. Pan, Environ. Sci. Technol. 52 (2018) 2197–2205. doi: 10.1021/acs.est.7b05563

  42. [42]

      Y. Wu, Z. Chen, W.C. Cheong, et al., Chem. Sci. 10 (2019) 5345–5352. doi: 10.1039/c9sc00475k

  43. [43]

      D. Liu, C. Li, T. Ni, et al., Appl. Surf. Sci. 555 (2021) 149677. doi: 10.1016/j.apsusc.2021.149677

  44. [44]

      W. Wei, J. Luo, S. Liu, et al., J. Solid State Chem. 312 (2022) 123160. doi: 10.1016/j.jssc.2022.123160

  45. [45]

      Y. Liu, Y. Liu, Y. Xu, et al., J. Colloid. Interf. Sci. 608 (2022) 2088–2099. doi: 10.1016/j.jcis.2021.10.146

  46. [46]

      B. Liu, X. Liu, J. Liu, et al., Appl. Catal. B: Environ. 226 (2018) 234–241. doi: 10.3390/s18010234

  47. [47]

      Y. Fu, T. Huang, B. Jia, et al., Appl. Catal. B: Environ. 202 (2017) 430–437. doi: 10.1016/j.apcatb.2016.09.051

  48. [48]

      Z. Yang, H. Zhang, L. Zhou, et al., J. Alloy. Compd. 949 (2023) 169879. doi: 10.1016/j.jallcom.2023.169879

  49. [49]

      M. Han, C. Kang, Z. Qu, et al., Appl. Catal. B: Environ. 305 (2022) 121064. doi: 10.1016/j.apcatb.2022.121064

  50. [50]

      Q. Yang, J. An, Z. Xu, et al., Colloid. Surf. A 614 (2021) 126161. doi: 10.1016/j.colsurfa.2021.126161

  51. [51]

      S. Qureshi, M. Asif, H. Sajid, et al., Mat. Sci. Semicon. Proc. 148 (2022) 106753. doi: 10.1016/j.mssp.2022.106753

  52. [52]

      J. Tan, Z. Li, J. Li, et al., J. Hazard. Mater. 423 (2022) 127048. doi: 10.1016/j.jhazmat.2021.127048

  53. [53]

      H. Li, Q. Gao, G. Wang, et al., Chem. Eng. J. 426 (2021) 131292. doi: 10.1016/j.cej.2021.131292

  54. [54]

      D. Liu, C. Li, J. Ge, et al., Appl. Surf. Sci. 579 (2022) 152180. doi: 10.1016/j.apsusc.2021.152180

  55. [55]

      M. Xu, H. Zhou, Z. Wu, et al., J. Hazard. Mater. 399 (2020) 123103. doi: 10.1016/j.jhazmat.2020.123103

  56. [56]

      G. Liu, T. Zhang, T. Wang, et al., Appl. Catal. B: Environ. 296 (2021) 120370. doi: 10.1016/j.apcatb.2021.120370

  57. [57]

      G. Zhao, W. Li, H. Zhang, et al., Chem. Eng. J. 430 (2022) 132937. doi: 10.1016/j.cej.2021.132937

• 

  Figure 1  (a, b) Differential charge densities, (c, d) energy profiles of PMS activation reaction for graphene and g-C₃N₄ supported Co-SACs.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (a) Schematic illustration of the synthesis of SACo-CN catalysts. (b) TEM images, (c) HAADF-STEM images (HRTEM images insert), and (d) elemental mapping images of SACo-CN 1.5. (e) Co K-edge X-ray absorption near-edge structure (XANES) spectra, (f) Fourier transform extended X-ray absorption fine structure (FT-EXAFS) spectra, (g) wavelet transform (WT) plots of the SACo-CN 1.5, Co foil, CoO, and Co₃O₄. (h) FT-EXAFS fitting curves of the SACo-CN 1.5 at Co K-edge. (i) k space fitting curves of the SACo-CN 1.5.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) The diffuse reflectance spectra, (b) band structure diagram, (c) electrochemical impedance spectroscopy, (d) photocurrent, (e) steady-state and (f) time-resolved PL decay curves of g-C₃N₄ and SACo-CN samples. LUMO and HOMO distributions of (g) g-C₃N₄, (h) SACo-CN 1.5. Activation pathways of PMS on (i) g-C₃N₄ and (j) SACo-CN 1.5.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Degradation efficiency of OTC under different (a) catalysts, (b) PMS concentrations, (c) SACo-CN 1.5 dosages, (d) OTC concentrations, (e) inions, (f) solution pH, and (g) real water bodies. (h) Cycle experiments for OTC degradation. (i) Catalytic degradation of common pollutants by SACo-CN 1.5 catalyst. Experimental conditions unless otherwise specified: [PMS] = 0.5 mmol/L, [catalyst] = 0.4 g/L, [OTC]₀ = 15 mg/L, visible light irradiation.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  Schematic for possible mechanism of OTC degradation by SACo-CN 1.5/PMS/vis system.

  下载: 全尺寸图片 幻灯片
•