English

• Broad-spectrum antibiotics, particularly second-generation fluoroquinolones, have been monitored worldwide in natural water matrices because of their high demand and extensive application in the treatment of diseases caused by bacterial infections [1]. Ciprofloxacin (CIP) is a typical fluoroquinolone antibiotic, which is frequently detected in the effluents of pharmaceutical manufacturing. The trace-level CIP with high ecotoxicity will severely threaten human health and the safety of the natural ecosystem [2]. Therefore, efficient treatment strategies and technologies are urgently required.

  Considering the deficiencies in the biodegradability and mineralization rate of antibiotics, conventional treatment processes using activated sludge cannot efficiently remove CIP, and transformation products (TPs) may have higher ecotoxicity [3-6]. Photocatalysis is a cost-efficient and environmentally friendly technology that has gained increasing attention for the treatment of antibiotics such as CIP [7]. Compared to other technologies, photocatalysis has specific advantages such as solar light utilization, moderate reaction conditions and low secondary pollution [8]. More importantly, photocatalysis can effectively mineralize various antibiotics to non-toxic final products such as H₂O and CO₂ [9]. Recently, metal-free graphitic carbon nitride (g-C₃N₄), synthesized via a facile thermal polymerization method, has been proposed as a promising photocatalyst owing to its medium band gap (~2.65 eV), high photogenerated carrier transfer efficiency, and good stability. Thus, g-C₃N₄ has been widely applied for photocatalytic degradation of various organic contaminants [10,11]. However, the practical application of g-C₃N₄ is limited by its deficiencies in light absorption and conversion in the visible light range, electron transfer efficiency and recoverability [12]. More importantly, the photoinduced electron-hole pairs generally possess a higher recombination ratio in g-C₃N₄, leading to a low utilization rate of photogenerated carriers [11,12].

  Various strategies have been developed to improve the photocatalytic activity of neat g-C₃N₄, In particular, heterostructure/heterojunction construction has been widely proven as an adequate approach for enhancing its photocatalytic activity [13]. Compared with other heterojunctions (type-Ⅱ, Z-scheme and S-scheme), the type-Ⅰ heterojunction facilitates the efficient separation of electron-hole pairs and controls the hole transfer direction and accumulation, resulting in the construction of a stable photocatalyst and the formation of a built-in electric field [14]. However, few studies reported type-Ⅰ heterojunctions because of the special band energy levels of the two semiconductors. Thus, it is necessary to precisely regulate type-Ⅰ heterojunctions coupled with g-C₃N₄.

  Sodium niobate (NaNbO₃), which has thermoelectric, piezoelectric, and ferroelectric properties [15,16], is a potential candidate for environmental photocatalysis applications. In addition, NaNbO₃ with a high valence band level exhibits strong oxidation ability owing to the generation of rich reactive species, such as photoinduced holes (h⁺) and hydroxyl radicals (^•OH), which are suitable for micropollutant degradation. In addition, incorporation of NaNbO₃, as a new electron donor level, can increase the electrons density and transfer efficiency, thus the introduction of NaNbO₃ with high photocatalytic ability in composite material is a potential method for improving photocatalytic performance [17]. Therefore, modification of g-C₃N₄ with NaNbO₃ to construct type-Ⅰ heterojunctions will boost electron transfer, assuming the positive effect of “1 + 1 > 2” during the organic pollutant elimination process. In addition, the morphology and structure of NaNbO₃ could be controlled by varying the synthesis conditions. Therefore, the in-situ coordinated 3-D cubic NaNbO₃ from Na₂Nb₂O₆·H₂O shows better photocatalytic activities because of the crystal plane heterogeneity, thus facilitating photoexcited electron transfer in the type-Ⅰ heterojunction of g-C₃N₄ [18,19].

  In this study, an NaNbO₃ nanocube-decorated g-C₃N₄ nanohybrid (NbNC/g-C₃N₄) was synthesized and used for the photocatalytic degradation of CIP under simulated solar light. Energy-band structures were designed to achieve efficient electron transfer. The photocatalytic detoxification efficiency of CIP was evaluated. The mechanism of reactive species production was revealed through material characterization and radical identification. In addition, the reaction mechanism of CIP attack by radicals was investigated via integrated experimental analysis and computational calculations.

  All chemicals and reagents used in the experiments were of analytical grade or higher without further purification, as provided in Text S1(Supporting information). NbNC/g-C₃N₄ was synthesized via a two-step route, i.e., Na₂Nb₂O₆·H₂O (precursor of NaNbO₃) and g-C₃N₄ were prepared first, followed by a calcination process to achieve in-situ transformation and decoration of NaNbO₃ onto g-C₃N₄. Detailed synthesis methods are shown in Text S2 (Supporting information). The synthesized materials were characterized by field-emission scanning electron microscopy (SEM), high-resolution transmission electron microscopy (HRTEM), X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), Fourier transform infrared spectroscopy. In addition, UV–vis spectrophotometry, photocurrent, electrochemical impedance spectroscopy (EIS), photoluminescence spectra (PL), linear sweep voltammetry (LSV) and Brunauer-Emmett-Teller (BET) nitrogen adsorption were also analyzed to obtain the diffuse reflectance spectra (DRS), photoelectrochemical properties, surface area, and pore properties of the catalysts. The detailed material characterization methods are described in Text S3 (Supporting information).

  Batch experiments on the photocatalytic performance were performed in a quartz reactor (volume = 250 mL) with a water-cooling system to maintain the reaction temperature at 25 ℃ (Fig. S1 in Supporting information). A Microsolar 300 solar-light simulator (Beijing Perfectlight, Beijing, China) coupled with a Xe lamp (300 W) was used and operated at a light irradiation of 100 ± 0.5 mW/cm² in the AM 1.5 G mode. The typical photocatalytic test was conducted with 10 µmol/L CIP (100 mL) and 0.1 g/L catalysts in the reactor, and the pH was adjusted to 7.0 ± 0.1 by NaOH or HClO₄ (0.1 mol/L). The CIP concentration was detected using a Dionex UltiMate 3000 high-performance liquid chromatograph (HPLC, Thermo Fisher Scientific). The TPs and degradation intermediates were determined using an ultra-HPLC equipped with an electrospray ionization source and a triple quadrupole mass spectrometer (Dionex UltiMate 3000 Series; MS, Thermo Scientific, USA) [20]. The detailed analytical methods are described in Text S4 (Supporting information).

  Density functional theory (DFT) calculations on CIP are performed using Gaussian 16 software (Version C.01) [21]. B3LYP combined with 6–31+G(d, p) is applied for structure optimization [22,23]. Details of the calculation method for the Fukui index are described in Text S5 (Supporting information). To investigate the theoretical ecotoxicity of CIP and its TPs [3]. The Cambridge Sequential Total Energy Package in Materials Studio 2020 was used to calculate the band gap, density of state (DOS), projected DOS (PDOS), work function, electron density, and electron density difference (EDD) of the materials (Text S6 in Supporting information) [24-26]. g-C₃N₄ exhibits a smooth surface and 2-D layered structure (Figs. 1a-g), while pristine Na₂Nb₂O₆·H₂O exhibited as 1-D nanorods (Figs. 1b and e), and the lattice distance of 0.15 nm displayed in the HRTEM image was indexed to the (200) plane of Na₂Nb₂O₆·H₂O (Fig. 1h) [27,28]. After calcination treatment for the in-situ formation and decoration of NaNbO₃ onto g-C₃N₄, the Na₂Nb₂O₆·H₂O 1-D nanowires were transformed to NaNbO₃ 3-D nanocubes via dehydration (Fig. 1f) due to the more thermodynamically stability than NaNbO₃ nanowires [27]. In addition, the surface of NbNC/g-C₃N₄–1 was coarse owing to the niobate decoration (Fig. 1c). The measured lattice distance of 0.39 nm in NbNC/g-C₃N₄–1 was assigned to the (110) plane of orthorhombic NaNbO₃ (JCPDS 01–073–0803) (Fig. 1i), indicating the different crystalline structures for pristine Na₂Nb₂O₆·H₂O and NaNbO₃ phase in NbNC/g-C₃N₄–1. The successful hybridization of the NaNbO₃ nanocubes and g-C₃N₄ can enhance the utilization of solar light, thus facilitating electron transfer to achieve better photocatalytic effectiveness for CIP degradation. In addition, an interface connection between g-C₃N₄ and Na₂Nb₂O₆ phases is clearly observed in Fig. 1i, allowing the electrons transfer between these two components and beneficial to the built-in electric field construction [29].

  Figure 1

  

  Figure 1.  (a-c) Field emission scanning electron microscopy (FESEM), (d-f) TEM and (g-i) HRTEM images of g-C₃N₄, Na₂Nb₂O₆·H₂O and NbNC/g-C₃N₄–1.

  DownLoad: Full-Size Img PowerPoint

  In the XRD patterns (Fig. 2a), the significant peaks of pristine g-C₃N₄ at 13.0° and 27.5° were assigned to the (100) and (002) crystalline planes (JCPDS 01–087–1526), respectively, which can be characterized as the tri-s-triazine system in the interlayer structure and stacking of aromatic units [30]. For pristine Na₂Nb₂O₆·H₂O, the 2θ peaks at 11.4°, 12.7°, 29.1°, 30.4°, and 31.6° corresponded to monoclinic crystal Na₂Nb₂O₆ with a C2/c space group (a = 17.114 Å, b = 5.0527 Å, and c = 16.5587 Å), comprising [NbO₆] octahedra lattice [31]. For pristine NaNbO₃ formed after thermal dehydration from Na₂Nb₂O₆·H₂O, new distinct peaks at 22.8°, 32.5°, 46.5°, 52.4°, and 57.9° appeared, assigned to the (110), (114), (220), (118), and (028) planes of the orthorhombic crystal structure (JCPDS 01–073–0803) (Fig. S2 in Supporting information), respectively, with the Pbcm space group (a = 5.506 Å, b = 5.566 Å, and c = 15.520 Å). An “octahedral tilting” [NbO₆] octahedra corner-sharing network was displayed, indicating that structure fabrication along with phase transformation occurred during dehydration of Na₂Nb₂O₆·H₂O, i.e., the [NbO₆] layers and chemical bonding (Nb–O–Nb and Nb–O–Na) were atomically rearranged to [NbO₆] octahedra via the thermal excitation due to the minimization of Gibbs free energy, resulting in the formation of NaNbO₃ nanocubes (Figs. 1e and f). For all NbNC/g-C₃N₄ nanohybrids, all the Na₂Nb₂O₆·H₂O peaks disappeared while the diffractions of NaNbO₃ emerged, indicating the in-situ crystalline phase transformation via calcination.

  Figure 2

  

  Figure 2.  (a) XRD patterns of NbNC/g-C₃N₄ with different component ratios. The XPS spectra of NaNbO₃, g-C₃N₄ and NbNC/g-C₃N₄–1: (b) The survey, (c) Nb 2p, (d) O 1s, (e) C 1s and (f) N 1s.

  DownLoad: Full-Size Img PowerPoint

  Figs. 2b-f show the XPS spectra of materials. In the high-resolution Nb 3d spectrum (Fig. 2c), the two significant split peaks at 206.1 eV and 208.9 eV for pristine NaNbO₃ are assigned to the Nb 3d_3/2 and 3d_5/2 orbitals, respectively. The spin-orbit separation energy of 2.8 eV corresponded to the [NbO₆] octahedra [32]. There was a positive shift (+0.5 eV) for the two Nb orbitals for NbNC/g-C₃N₄–1 (Table S1 in Supporting information), indicating electron transfer from NaNbO₃ to g-C₃N₄ and the formation of chemical bonds for Nb. In the high-resolution spectra of O 1s (Fig. 2d), the peaks at 529.6, 530.9, and 533.6 eV for pristine NaNbO₃ indexed to the lattice O derived from the [NbO₆] octahedra, Nb–OH, and chemisorbed O/H₂O, respectively. For NbNC/g-C₃N₄–1, the atomic percentage of lattice O increased from 63.07 at% (NaNbO₃) to 70.39 at% (NbNC/g-C₃N₄–1) (Table S1), indicating the formation of Nb–O–C or Nb–O–N, except the main chemical bond Nb–O–Nb in the [NbO₆] octahedra [33]. In the high-resolution spectra of C 1s (Fig. 2e), the peaks at 284.8 eV and 288.3 eV in pristine g-C₃N₄ corresponded to C–C/C=C and N–C=N of the triazine rings, respectively [4]. The N–C=N peak decreased dramatically from 85.16 at% for g-C₃N₄ to 9.82 at% for NbNC/g-C₃N₄–1 (Table S1), indicating new bond formation for C. Moreover, a new peak at ca. 285.9 eV assigned to C–O (accounting for 38.33 at% of C) increased in NbNC/g-C₃N₄–1, which was in good agreement with the formation of Nb–O–C.

  Fig. S3a (Supporting information) shows the photocatalytic effectiveness of various photocatalysts for CIP degradation. The adsorption of CIP was first assessed for all materials (Figs. S3a and S4 in Supporting information). Neat g-C₃N₄ barely adsorbed CIP in 30 min, while neat NaNbO₃ displayed a CIP adsorption efficiency of 39.0%. In addition, the NbNC/g-C₃N₄ composites showed CIP adsorption efficiencies of 10.0%-16.9%, indicating efficient interaction between CIP and NbNC/g-C₃N₄ materials. Then, a pseudo-first order model was used to interpret the kinetic results (Table S2 in Supporting information) [34,35]. 64.1% and 84.5% of CIP was degraded by pristine g-C₃N₄ and NaNbO₃ with degradation kinetic rate constants (k₁) of 0.052 and 0.076 min⁻¹, respectively. Enhanced photocatalytic effectiveness of CIP degradation was observed using all NbNC/g-C₃N₄ nanohybrids. Specifically, NbNC/g-C₃N₄–1 (Na₂Nb₂O₆·H₂O: g-C₃N₄ mass ratio of 1:1) exhibited the highest k₁ value (0.173 min⁻¹), which was 3.3 and 2.3 times that of pristine g-C₃N₄ and NaNbO₃, respectively. Fig. S3b (Supporting information) shows that the optimum NbNC/g-C₃N₄–1 dosage was 0.10 g/L, with a high CIP removal efficiency of 96.8%; further increasing the dosage to 0.20 g/L achieved a slight increase in the CIP removal efficiency to 99.3%.

  After scavenger quenching tests, Fig. 3a shows that the addition of Tiron and potassium iodide (KI) inhibited CIP degradation by 56.5% and 36.3% at 20 min, respectively, indicating that ^•O₂⁻ played the most important role, whereas h⁺ contributed less (Table S3 in Supporting information). However, the addition of tert-butyl alcohol (tBA) slightly retarded the photocatalytic degradation of CIP by 2.6%, suggesting that ^•OH plays a negligible role. The electron spin resonance (ESR) spectra further confirmed the generation of ^•OH and ^•O₂⁻ in the NbNC/g-C₃N₄–1 photocatalytic system under solar irradiation (Figs. 3b and c).

  Figure 3

  

  Figure 3.  (a) Photocatalytic degradation of CIP by NbNC/g-C₃N₄–1 in the presence of various scavengers (Initial CIP concentration = 10 µmol/L, catalyst dosage = 0.1 g/L, temperature = 25 ± 0.2 ℃, solution pH = 7.0 ± 0.1, scavenger dosage = 10 mmol/L). ESR spectra of (b) DMPO-^•OH and (c) DMPO-^•O₂⁻. (d) E_g values calculated from Kubelka-Munk method. (e) XPS-VB spectra and (f) photocurrent spectra for pristine NaNbO₃, g-C₃N₄, and NbNC/g-C₃N₄–1. (g-i) EIS, PL and LSV curves of pristine NaNbO₃, g-C₃N₄ and NbNC/g-C₃N₄–1.

  DownLoad: Full-Size Img PowerPoint

  Based on the UV–vis DRS spectra (Fig. S5 in Supporting information), the measured E_g for NaNbO₃, g-C₃N₄, and NbNC/g-C₃N₄–1 were 4.34, 2.81, and 3.75 eV, respectively (Fig. 3d). The synthesized NaNbO₃ in this study had a markedly wider band gap energy; therefore, the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) were introduced to describe its energy levels instead of the valance band (VB) and conduction band (CB). The light absorption edge of NbNC/g-C₃N₄–1 was red-shifted, with a narrower band gap (3.75 eV) than that of pristine NaNbO₃ (286 nm, 4.34 eV), indicating the enhanced light utilization efficiency of the new material. Although g-C₃N₄ exhibited an absorption edge in the visible light region at 441 nm and an E_g of 2.81 eV, NbNC/g-C₃N₄–1 showed higher photocatalytic activity owing to efficient charge migration and electron-hole separation. In addition, the calculated E_VB and E_CB were 2.07 and −1.68 eV for NbNC/g-C₃N₄–1, respectively (Fig. 3e), and negative E_CB values led to higher electron reactivity on CB, facilitating the production of ^•O₂⁻. The photocurrent spectra clearly show that the photoelectric conversion efficiency of NbNC/g-C₃N₄–1 was better than that of pristine NaNbO₃ and g-C₃N₄ (Fig. 3f). Besides, EIS spectra in Fig. 3g display that NbNC/g-C₃N₄–1 with a smaller arc radius possessed a lower resistance for charge transfer and higher charge transfer efficiency than pristine g-C₃N₄ and NaNbO₃. Then, PL spectra in Fig. 3h also present that NbNC/g-C₃N₄–1 with a lower steady-state PL emission intensity than g-C₃N₄ and NaNbO₃, suggesting NbNC/g-C₃N₄–1 had a lower recombination rate of photoinduced electron-hole pairs than the other two materials. Finally, LSV curves in Fig. 3i show that NbNC/g-C₃N₄–1 had a stronger current density and redox capacity than pristine g-C₃N₄ and NaNbO₃, indicating the electrical conductivity and photocatalytic performance has been obviously enhanced after the heterojunction construction [36].

  To explore the mechanism on enhanced photocatalytic activity of NbNC/g-C₃N₄, the band structure, DOS/PDOS spectra, electron density, work function, and EDD were studied via DFT calculations (Fig. 4). The structural models of the catalysts are presented in Figs. S6a and b (Supporting information). The calculated band gap of pure g-C₃N₄ (Fig. 4a) and NaNbO₃ (Fig. S6c in Supporting information) were 1.281 and 2.373 eV, respectively. The obtained band gap values are lower than the experimental values because of the known limitations of the plain DFT method [36]. Fig. 4d shows that NbNC/g-C₃N₄ possesses a more negative CB than pure g-C₃N₄, indicating that the CB of NbNC/g-C₃N₄ has a high electronic reactivity. In the PDOS spectra for NbNC/g-C₃N₄ (Fig. 4e), the N 2p and O 2p orbitals primarily contributed to the VB, whereas the C 2p and Nb 3d orbitals contributed to the CB. Hence, the electron transfer path in NbNC/g-C₃N₄ followed: N 2p → O 2p → Nb 3d → C 2p based on the PDOS distributions. Meanwhile, Fig. 4e shows that the PDOS of N 2p moves to the left for NbNC/g-C₃N₄ compared to that of pure g-C₃N₄, indicating that the electron density of N 2p increased after the introduction of NaNbO₃ and more electrons were transferred along the above path under light irradiation. Moreover, the d-band center of Nb 3d in NbNC/g-C₃N₄ shifts to 0.40 eV compared with 0.44 eV of NaNbO₃ (Fig. 4e and Fig. S6d in Supporting information), which can be attributed to the compositing with g-C₃N₄. Based on the PDOS of Nb 3d in Fig. 4e, an antibonding state is formed at nearly 2 eV compared to pure NaNbO₃ (Fig. S6d), suggesting that d_x²-y² orbitals with higher coupling ability combine with p orbitals during the bond formation of Nb–O–N/Nb–O–C. Accordingly, the introduction of g-C₃N₄ changes the electronic structure of Nb 3d orbitals, further leading to the alteration of d band center [37]. Therefore, the d-band center of NbNC/g-C₃N₄ was markedly closer to the Fermi level, suggesting that the electrons in Nb 3d were more reactive and could be more effectively transferred to C 2p with higher catalytic activity [38]. The electron density images (Figs. 4c and f) show higher electron densities in the N 2p and C 2p spectra of NbNC/g-C₃N₄ than those of g-C₃N₄, which benefits the photocatalytic reaction. The work function (Fig. 4g) indicates that NaNbO₃ with a low value (4.488 eV) transfers electrons to g-C₃N₄ with a high value (4.689 eV) when the NbNC/g-C₃N₄ heterojunction is constructed. EDD analysis indicates that electrons are more likely to transfer from NaNbO₃ to g-C₃N₄ (Fig. 4h); thus, holes accumulated on NaNbO₃ and electrons gathers on g-C₃N₄ [38,39]. The EDD result clearly shows a 0.37 e^– transfer from NaNbO₃ to g-C₃N₄ based on the Mulliken population analysis of the built-in electric field intensity. Furthermore, because the e^– transfer direction is from NaNbO₃ to g-C₃N₄ based on the EDD result in NaNbO₃/g-C₃N₄ composite, the e^– density is greatly increased in g-C₃N₄, and the built-in electric field is also constructed because of their electrons density difference. Thus, g-C₃N₄ can be likened to a switch with a large amount of e^–, and more e^– transfer occurs along N 2p → O 2p → Nb 3d → C 2p chain under light irradiation, leading to a high e^– utilization rate and activity [40]. Thus, the electron density increased in g-C₃N₄ and the formed built-in electric field significantly boost the separation rate of charge carriers, further promoting photocatalytic activity [39].

  Figure 4

  

  Figure 4.  (a) The band structure, (b) DOS spectra and (c) electron density of pure g-C₃N₄. (d) The band structure, (e) DOS spectra and (f) electron density of NbNC(110)/g-C₃N₄. (g) The work function of g-C₃N₄ and NaNbO₃. (h) EDD of NbNC(110)/g-C₃N₄ (yellow region: electrons depletion; cyan region: electrons accumulation).

  DownLoad: Full-Size Img PowerPoint

  Fig. S7 (Supporting information) shows a schematic of efficient charge migration at the interface of NbNC/g-C₃N₄–1 nanohybrid during photocatalysis. Based on the energy levels of NaNbO₃ and g-C₃N₄, the NbNC/g-C₃N₄–1 structure is a type-Ⅰ heterojunction. Then, under light irradiation, the photogenerated electrons and holes transfer to the same side [14,41]. Specifically, the LUMO potential of NaNbO₃ (−1.93 V) is more negative than the CB of g-C₃N₄ (−0.84 V; Fig. S7), which helps the movement of photogenerated electrons from the LUMO of NaNbO₃ to the CB of g-C₃N₄. In addition, the photogenerated holes cannot move from the VB of g-C₃N₄ (+1.97 V) to the HOMO of NaNbO₃ (+2.41 V) because of the greater negativity of the VB of g-C₃N₄. The transformation of these photogenerated electrons by heterojunctions and the generation of holes in the HOMO and VB can reduce the charge recombination rate and provide more electrons/holes on the surface of the photocatalyst [14,41]. The photocurrent intensity in Fig. 3f and the DFT calculations in Fig. 4 prove the efficient separation of electron-hole pairs compared with neat g-C₃N₄ and NaNbO₃; thus, an internal electric field can also be generated during electron transfer from NaNbO₃ to g-C₃N₄.

  To accurately evaluate the reactive sites of the CIP molecule, DFT calculations of CIP were conducted based on the natural population analysis charge distribution (Fig. 5). Fig. 5b shows the HOMO is placed on the piperazine ring, and the C atom of the benzene ring is bonded with the F atom, which are the most reactive sites for electron loss [42]. Furthermore, the piperazine ring side of CIP^± with a high positive ESP at neutral pH preferred to attach on the negative-charged surface of NbNC/g-C₃N₄–1 (Fig. 5c). Fig. 5d indicates that the most reactive sites of CIP are C5, C8, N10, O17, N19, and N24, with high Fukui indices representing electrophilic attack, i.e., f ^– = 0.0417, 0.1008, 0.0719, 0.0990, 0.0831, and 0.1472, respectively. ^•O₂⁻ is the dominant ROS in this photocatalysis system [43]. Therefore, electrophilic attack for CIP^± oxidation is focused.

  Figure 5

  

  Figure 5.  (a) CIP^± chemical structure. (b) HOMO and LUMO of CIP^±. (c) ESP of CIP^±. (d) NPA charge distribution and Fukui index (f ^–) of CIP^±. (e) Proposed pathways on photocatalytic degradation of CIP by NbNC/g-C₃N₄–1 under solar light. (Initial CIP concentration = 10 µmol/L, catalyst dosage = 0.1 g/L, temperature = 25 ± 0.2 ℃, solution pH = 7.0 ± 0.1). The white, gray, blue, red, and cyan balls denote H, C, N, O and F atoms.

  DownLoad: Full-Size Img PowerPoint

  Table S4 and Fig. S8 (Supporting information) present the TPs of CIP during photocatalytic degradation, and the four CIP degradation pathways are shown in Fig. 5e. Pathway Ⅰ involves the photoinduced substitution of fluorine by the hydroxyl moiety for the formation of TP330 (m/z 330.06). The reaction occurred with the breakage of the C–F bond by attack at the C5 site (f ^– = 0.0417). Pathway Ⅱ involved defluorination and oxidation of the quinolone ring, leading to the formation of TP346 (m/z 346.06). TP330 and TP346 were generated via attack on the C5 (f ^– = 0.0417) and C8 (f ^– = 0.1008) sites. Pathway Ⅲ involved the cleavage of the piperazine moiety via oxidation, which was attributed to the attack on N19 (f ^– = 0.0831) and N24 (f ⁻ = 0.1472) sites with a high Fukui index. TP362 (m/z 336.03) was a key intermediate in this pathway. Pathway Ⅳ involved deep defluorination after dealkylation of the piperazine moiety in Pathway Ⅲ. TP344 (m/z 344.04) was a key intermediate formed upon the ring opening of the piperazine moiety and defluorination of CIP or TP362. Finally, mineralization of these TPs led to formation of inorganic small molecules/ions such as CO₂, H₂O, NH₄⁺ and F⁻ [44].

  Five consecutive CIP degradation tests were performed to evaluate the reusability and stability of NbNC/g-C₃N₄–1 (Fig. S9 in Supporting information), and the CIP removal efficiency just slightly decreased from 97.1% to 90.1%, indicating the good reusability and high stability of NbNC/g-C₃N₄–1. In addition, the physicochemical properties of NbNC/g-C₃N₄–1 after reaction were analyzed by XRD and XPS (Fig. S10 in Supporting information), which also indicate slight changes on crystal phase and composition of the material after photocatalysis [20]. In addition, negligible Na and Nb leached during the photocatalytic reaction based on ICP-OES analysis, indicating good reusability and high stability of NbNC/g-C₃N₄–1.

  In this study, NbNC/g-C₃N₄ nanohybrid was synthesized through in-situ transformation of Na₂Nb₂O₆·H₂O at the interface of g-C₃N₄, which exhibited as type-Ⅰ heterojunction. The optimized NbNC/g-C₃N₄–1 showed high photocatalytic activity for efficient photocatalytic degradation and detoxification of CIP under simulated solar light. In the NbNC/g-C₃N₄ nanohybrid, the electron transfer path was as follows: N 2p → O 2p → Nb 3d → C 2p. Electrons can be effectively transferred along the Nb–O–N/Nb–O–C bonds, which greatly facilitates the migration of photoexcited electrons/holes. Meanwhile, the electron density increased in g-C₃N₄ to form a built-in electric field, because 0.37 e^– transferred from NaNbO₃ to g-C₃N₄ based on Mulliken population analysis, which can significantly boost the separation rate of charge carriers. Upon solar light irradiation, simultaneous degradation and detoxification of CIP by the NbNC/g-C₃N₄ nanohybrid was achieved, implying its considerable application potential in practical 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

  Hui Wang: Writing – original draft, Validation, Resources, Methodology, Investigation, Funding acquisition, Conceptualization. Haodong Ji: Writing – review & editing, Validation, Supervision, Methodology, Funding acquisition, Data curation. Dandan Zhang: Validation, Investigation, Formal analysis, Data curation. Xudong Yang: Validation, Software. Hanchun Chen: Validation, Investigation. Chunqian Jiang: Writing – review & editing, Supervision, Funding acquisition, Conceptualization. Weiliang Sun: Validation, Methodology, Investigation, Data curation. Jun Duan: Validation, Methodology, Data curation. Wen Liu: Writing – review & editing, Validation, Supervision, Project administration, Methodology, Funding acquisition, Data curation, Conceptualization.

  Acknowledgments

  Financial supports from the National Key Research and Development Program of China (Nos. 2021YFA1202500 and 2022YFF1303004), Shenzhen Science and Technology Program (No. JCYJ20220531093205013), the National Natural Science Foundation of China (NSFC) (Nos. 52100069, 52270053 and 52200084), the Beijing Natural Science Foundation (No. 8232035), the Beijing Nova Program (No. 20220484215), the Beijing National Laboratory for Molecular Sciences (No. BNLMS2023011) and Emerging Engineering Interdisciplinary-Young Scholars Project (Peking University), the Fundamental Research Funds for the Central Universities are greatly acknowledged. DFT calculations supported by the High-Performance Computing Platform of Peking University and the National Key Scientific and Technological Infrastructure project “Earth System Numerical Simulation Facility” (EarthLab) are also acknowledged. The work is also supported by the program of “Research on Advanced Treatment Technology of New Pollutants in Domestic Sewage of Residential District”.

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.110200.

  
  
• 1. [1]

     B. Zhou, L. Chen, F. Li, et al., Chin. Chem. Lett. 34 (2023) 108558. doi: 10.1016/j.cclet.2023.108558

  2. [2]

     F. Sun, X. Yang, F. Shao, et al., Chin. Chem. Lett. 34 (2023) 108563. doi: 10.1016/j.cclet.2023.108563

  3. [3]

     C. Dang, F. Sun, H. Jiang, et al., J. Hazard. Mater. 400 (2020) 123225. doi: 10.1016/j.jhazmat.2020.123225

  4. [4]

     W. Liu, Y. Li, F. Liu, et al., Water Res. 151 (2019) 8–19. doi: 10.15302/j-laf-1-010002

  5. [5]

     J. Richard, A. Boergers, C. vom Eyser, K. Bester, J. Tuerk, Int. J. Hyg. Environ. Health. 217 (2014) 506–514. doi: 10.1016/j.ijheh.2013.09.007

  6. [6]

     H. Ji, T. Wang, T. Huang, B. Lai, W. Liu, J. Cleaner Prod. 278 (2021) 123924. doi: 10.1016/j.jclepro.2020.123924

  7. [7]

     S. Li, T. Huang, P. Du, W. Liu, J. Hu, Water Res. 185 (2020) 116286. doi: 10.1016/j.watres.2020.116286

  8. [8]

     J. Xu, P. Liang, X. Shen, et al., Sep. Purif. Technol. 339 (2024) 126734. doi: 10.1016/j.seppur.2024.126734

  9. [9]

     O. Baaloudj, I. Assadi, N. Nasrallah, et al., J. Water Process Eng. 42 (2021) 102089. doi: 10.1016/j.jwpe.2021.102089

  10. [10]

      Y. Liu, L. Chen, X. Liu, et al., Chin. Chem. Lett. 33 (2022) 1385–1389. doi: 10.3390/cryst12101385

  11. [11]

      A. Kumar, P. Raizada, A. Hosseini-Bandegharaei, et al., J. Mater. Chem. A 9 (2021) 111–153. doi: 10.1039/d0ta08384d

  12. [12]

      A. Balakrishnan, E.S. Kunnel, R. Sasidharan, M. Chinthala, A. Kumar, Chem. Eng. J. 475 (2023) 146163. doi: 10.1016/j.cej.2023.146163

  13. [13]

      D. He, Z. Zhang, Y. Xing, et al., Chem. Eng. J. 384 (2020) 123258. doi: 10.1016/j.cej.2019.123258

  14. [14]

      T. Muhmood, M. Xia, W. Lei, F. Wang, Appl. Catal. B: Environ. 238 (2018) 568–575. doi: 10.1016/j.apcatb.2018.07.029

  15. [15]

      Y.F. Xu, M.Z. Yang, B.X. Chen, et al., J. Am. Chem. Soc. 139 (2017) 5660–5663. doi: 10.1021/jacs.7b00489

  16. [16]

      E. Grabowska, Appl. Catal. B: Environ. 186 (2016) 97–126. doi: 10.1016/j.apcatb.2015.12.035

  17. [17]

      X. Liu, P. Du, W. Pan, et al., Appl. Catal. B: Environ. 231 (2018) 11–22. doi: 10.1016/j.apcatb.2018.02.062

  18. [18]

      P. Li, H. Xu, L. Liu, et al., J. Mater. Chem. A 2 (2014) 5606–5609. doi: 10.1039/C4TA00105B

  19. [19]

      P. Li, S. Ouyang, Y. Zhang, T. Kako, J. Ye, J. Mater. Chem. A 1 (2013) 1185–1191. doi: 10.1039/C2TA00260D

  20. [20]

      D. Zhang, J. Qi, H. Ji, et al., Chem. Eng. J. 400 (2020) 125918. doi: 10.1016/j.cej.2020.125918

  21. [21]

      M.J. Frisch, G.W. Trucks, H.B. Schlegel, et al., Gaussian 16 Rev. C.01, 2016.

  22. [22]

      X. Zeng, J. Zhu, G. Zhang, et al., Chem. Eng. J. 468 (2023) 143536. doi: 10.1016/j.cej.2023.143536

  23. [23]

      S. He, E. Wu, M. Shen, ACS ES&T Eng. 3 (2023) 651–660. doi: 10.1021/acsestengg.2c00373

  24. [24]

      M.D. Segall, J.D.L. Philip, M.J. Probert, et al., J. Phy. Condens. Matter 14 (2002) 2717. doi: 10.1088/0953-8984/14/11/301

  25. [25]

      S. Gao, H. Ji, P. Yang, et al., Small 19 (2023) 2206114. doi: 10.1002/smll.202206114

  26. [26]

      S. Gao, R. Wu, M. Sun, et al., Appl. Catal. B: Environ. 324 (2023) 122260. doi: 10.1016/j.apcatb.2022.122260

  27. [27]

      A. Yu, J. Qian, L. Liu, H. Pan, X. Zhou, Appl. Surf. Sci. 258 (2012) 3490–3496. doi: 10.1016/j.apsusc.2011.08.133

  28. [28]

      M. Nyman, A. Tripathi, J.B. Parise, R.S. Maxwell, T.M. Nenoff, J. Am. Chem. Soc. 124 (2002) 1704–1713. doi: 10.1021/ja017081z

  29. [29]

      Y. Yang, S. Zhao, F. Bi, et al., Appl. Catal. B: Environ. 315 (2022) 121550. doi: 10.1016/j.apcatb.2022.121550

  30. [30]

      A. Kumar, P. Raizada, P. Singh, R.V. Saini, A.K. Saini, Chem. Eng. J. 391 (2020) 123496. doi: 10.1016/j.cej.2019.123496

  31. [31]

      J.H. Jung, C.Y. Chen, W.W. Wu, et al., J. Phys. Chem. C 116 (2012) 22261–22265. doi: 10.1021/jp308289r

  32. [32]

      L. Yan, T. Zhang, W. Lei, et al., Catal. Today 224 (2014) 140–146. doi: 10.1016/j.cattod.2013.11.033

  33. [33]

      R. Radha, Y. Ravi Kumar, M. Sakar, K. Rohith Vinod, S. Balakumar, Appl. Catal. B: Environ. 225 (2018) 386–396. doi: 10.1016/j.apcatb.2017.12.004

  34. [34]

      H. Ji, Y. Gong, J. Duan, D. Zhao, W. Liu, Mar. Pollut. Bull. 135 (2018) 427–440. doi: 10.1016/j.marpolbul.2018.07.047

  35. [35]

      H. Ji, Y. Zhu, W. Liu, et al., Colloids Surf. A: Physicochem. Eng. Asp. 561 (2019) 373–380. doi: 10.1016/j.colsurfa.2018.10.048

  36. [36]

      X. Yang, F. Li, W. Liu, et al., Appl. Catal. B: Environ. 324 (2023) 122202. doi: 10.1016/j.apcatb.2022.122202

  37. [37]

      J. Qi, X. Yang, P.Y. Pan, et al., Environ. Sci. Technol. 56 (2022) 5200–5212. doi: 10.1021/acs.est.1c08806

  38. [38]

      H. Shi, G. Chen, C. Zhang, Z. Zou, ACS Catal. 4 (2014) 3637–3643. doi: 10.1021/cs500848f

  39. [39]

      X. Yang, J. Duan, J. Qi, et al., J. Hazard. Mater. 445 (2023) 130576. doi: 10.1016/j.jhazmat.2022.130576

  40. [40]

      H. Zhang, Y. Wang, S. Zuo, et al., J. Am. Chem. Soc. 143 (2021) 2173–2177. doi: 10.1021/jacs.0c08409

  41. [41]

      J. Low, J. Yu, M. Jaroniec, S. Wageh, A.A. Al-Ghamdi, Adv. Mater. 29 (2017) 1601694. doi: 10.1002/adma.201601694

  42. [42]

      J. Duan, L. Chen, H. Ji, et al., Chin. Chem. Lett. 33 (2022) 3172–3176. doi: 10.1016/j.cclet.2021.11.072

  43. [43]

      F. De Vleeschouwer, V. Van Speybroeck, M. Waroquier, P. Geerlings, F. De Proft, Org. Lett. 9 (2007) 2721–2724. doi: 10.1021/ol071038k

  44. [44]

      Y. Wang, X. Sun, J. Ma, et al., Sep. Purif. Technol. 337 (2024) 126392. doi: 10.1016/j.seppur.2024.126392

• 

  Figure 1  (a-c) Field emission scanning electron microscopy (FESEM), (d-f) TEM and (g-i) HRTEM images of g-C₃N₄, Na₂Nb₂O₆·H₂O and NbNC/g-C₃N₄–1.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (a) XRD patterns of NbNC/g-C₃N₄ with different component ratios. The XPS spectra of NaNbO₃, g-C₃N₄ and NbNC/g-C₃N₄–1: (b) The survey, (c) Nb 2p, (d) O 1s, (e) C 1s and (f) N 1s.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) Photocatalytic degradation of CIP by NbNC/g-C₃N₄–1 in the presence of various scavengers (Initial CIP concentration = 10 µmol/L, catalyst dosage = 0.1 g/L, temperature = 25 ± 0.2 ℃, solution pH = 7.0 ± 0.1, scavenger dosage = 10 mmol/L). ESR spectra of (b) DMPO-^•OH and (c) DMPO-^•O₂⁻. (d) E_g values calculated from Kubelka-Munk method. (e) XPS-VB spectra and (f) photocurrent spectra for pristine NaNbO₃, g-C₃N₄, and NbNC/g-C₃N₄–1. (g-i) EIS, PL and LSV curves of pristine NaNbO₃, g-C₃N₄ and NbNC/g-C₃N₄–1.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  (a) The band structure, (b) DOS spectra and (c) electron density of pure g-C₃N₄. (d) The band structure, (e) DOS spectra and (f) electron density of NbNC(110)/g-C₃N₄. (g) The work function of g-C₃N₄ and NaNbO₃. (h) EDD of NbNC(110)/g-C₃N₄ (yellow region: electrons depletion; cyan region: electrons accumulation).

  下载: 全尺寸图片 幻灯片
  

  Figure 5  (a) CIP^± chemical structure. (b) HOMO and LUMO of CIP^±. (c) ESP of CIP^±. (d) NPA charge distribution and Fukui index (f ^–) of CIP^±. (e) Proposed pathways on photocatalytic degradation of CIP by NbNC/g-C₃N₄–1 under solar light. (Initial CIP concentration = 10 µmol/L, catalyst dosage = 0.1 g/L, temperature = 25 ± 0.2 ℃, solution pH = 7.0 ± 0.1). The white, gray, blue, red, and cyan balls denote H, C, N, O and F atoms.

  下载: 全尺寸图片 幻灯片
•