Construction of a Z-scheme PCN-222/CoFe₂O₄ heterojunction for efficient photo-Fenton tetracycline hydrochloride degradation: mechanism and pathways

Abstract

This study successfully constructed Z-scheme heterojunction photocatalysts (P/CF-x) by compositing the metal–organic framework (MOF) PCN-222 with spinel ferrite CoFe₂O₄via a ball-milling method for efficient visible-light-driven photo-Fenton degradation of tetracycline hydrochloride (TCH). The optimized P/CF-20 composite achieved 99.46% degradation of 0.02 g per L TCH within 30 minutes, with an apparent rate constant of 0.1452 min⁻¹ under optimized conditions (0.1 g per L catalyst, pH 6, 250 μL 30 wt% H₂O₂). The enhanced performance originates from the Z-scheme heterojunction, which promotes efficient separation of photogenerated charge carriers: electrons from the conduction band of PCN-222 recombine with holes from the valence band of CoFe₂O₄, preserving highly oxidative h⁺ and reductive e⁻. The interfacial charge transfer mechanism was validated through XPS, PL, TRPL, SPV and EIS analyses. EPR and radical trapping experiments confirmed that h⁺, ·OH, and ·O₂⁻ are the dominant reactive species, with h⁺ playing a primary role. LC-MS analysis revealed two degradation pathways, which involve hydroxylation, demethylation, ring-opening reactions, dehydration, radical-mediated cleavage, and finally mineralization into CO₂, H₂O, and NH₄⁺. Furthermore, P/CF-20 exhibited excellent stability (>96% efficiency after 6 cycles) and magnetic recoverability (saturation magnetization: 13.35 emu g⁻¹). This work provides a novel strategy for designing efficient and recyclable heterogeneous photo-Fenton catalysts, offering potential applications in antibiotic-containing wastewater treatment.

---

1 Introduction

Antibiotics represent a class of chemical compounds capable of interfering with the developmental processes of living cells. Tetracycline hydrochloride (TCH, C₂₂H₂₄N₂O₈·HCl), as a highly efficient, cost-effective broad-spectrum antibacterial agent, has been extensively utilized in medical and livestock industries worldwide. However, wastewater generated during its production and application exhibits persistent environmental challenges due to its recalcitrant biodegradation and inherent toxicity.¹ Furthermore, environmental tetracycline hydrochloride demonstrates bioaccumulation potential through food chains, which may induce microbial antibiotic resistance and potentially cause adverse effects on human organ systems.² Current treatment technologies for antibiotic-containing wastewater primarily focus on physical,³ chemical,^4,5 and biological approaches.⁶ Among existing remediation strategies, advanced oxidation processes (AOPs) have emerged as a particularly effective water treatment technology.^7,8

AOPs encompass various techniques including photocatalysis,⁹ ozonation,⁶ electrochemical oxidation,¹⁰ Fenton,¹¹ and Fenton-like processes,¹² typically generating highly reactive oxygen species (ROS) with strong oxidative capacity to degrade recalcitrant and non-biodegradable organic compounds. Studies have reported that heterogeneous Fenton-like processes predominantly produce superoxide radicals (·O₂⁻),¹³ hydroxyl radicals (·OH),¹⁴ and singlet oxygen (¹O₂),¹⁵ which play crucial roles in antibiotic wastewater treatment.

Spinel ferrites (MFe₂O₄ (M = Mg²⁺, Co²⁺, Ni²⁺, Fe²⁺, Mn²⁺, etc.)) have attracted significant attention as photocatalysts due to their wide availability, excellent stability, and facile recyclability.^16–19 Notably, CoFe₂O₄ magnetic nanoparticles (MNPs) stand out as one of the most catalytically active materials, making them promising candidates for nanocomposite design.^18,20 However, despite the strong magnetism of CoFe₂O₄ MNPs which enables efficient recovery, agglomeration inevitably occurs during Fenton-like reactions under excessive dosage conditions,²¹ resulting in reduced active site exposure, diminished adsorption capacity, and consequently severely compromised photocatalytic efficiency.

Metal–organic frameworks (MOFs) have garnered substantial interest due to their exceptional surface area, structural tunability, and functionalization potential.^22–24 Recent studies have explored hybrid systems integrating MOFs with CoFe₂O₄.^25,26 For instance, Rui et al. utilized MIL-101 to disperse CoFe₂O₄ nanoparticles in a sulfate radical (·SO₄⁻) based photo-Fenton system, achieving an enhancement of TCH degradation efficiency from ∼10% to 83% within 40 minutes.²⁷ This demonstrates that incorporating MOFs to disperse CoFe₂O₄ nanoparticles while improving adsorption–desorption dynamics can effectively enhance active site accessibility and catalytic performance. This addresses the challenge of difficult recovery of MOF materials; nevertheless, significant research opportunities remain in enhancing degradation efficiency through the modification of MOF materials. Specifically, Zr-MOFs (such as PCN-222,²⁸ PCN-223,²⁹ PCN-224,⁵ and UIO-66 (ref. 30)) exhibit exceptional chemical stability over a wide pH range, broad visible-light absorption characteristics, and large pore volume-to-surface area ratios. These structural advantages endow them with superior adsorption capabilities, making them widely employed in constructing heterojunction architectures to enhance charge separation efficiency. Through dispersing CoFe₂O₄ MNPs within Zr-MOF matrices, magnetic MOF composites can be engineered as heterogeneous ternary metal catalysts containing Zr, Co, and Fe active sites. This synergistic integration leverages the complementary advantages of both components enabling catalyst recovery and potentially generating novel catalytic interfaces with enhanced reactivity.^5,31–33

In this study, we synthesized PCN-222/CoFe₂O₄ composite materials via a ball-milling method to construct a magnetically recoverable, high-performance green Z-scheme heterojunction photocatalyst. Experimental characterization confirmed the formation of intimate interfacial contact between PCN-222 and CoFe₂O₄, accompanied by robust electronic interactions at the heterojunction interface. This structural configuration facilitates efficient charge transfer. Under visible light irradiation for 30 minutes, the composite demonstrated exceptional photo-Fenton catalytic performance, achieving 99.46% degradation efficiency of tetracycline hydrochloride. The synergistic integration of these two materials not only mitigated the inherent agglomeration tendency of CoFe₂O₄ nanoparticles but also significantly enhanced the photo-Fenton catalytic activity. Furthermore, systematic investigations were conducted to elucidate the underlying photocatalytic degradation mechanisms and propose potential degradation pathways for tetracycline hydrochloride.

2 Experimental

2.1. Materials

Ferric chloride hexahydrate (FeCl₃·6H₂O), cobalt dichloride hexahydrate (CoCl₂·6H₂O), ammonium hydroxide (NH₃ H₂O), N,N-dimethylformamide (DMF), trifluoroacetic acid (TFA), zirconyl chloride octahydrate (ZrOCl₂·8H₂O), tetrakis(4-carboxyphenyl)porphyrin (TCPP(H₂)), acetic acid (CH₃COOH), acetonitrile (C₂H₃N), tetracycline hydrochloride (TCH, C₂₂H₂₄N₂O₈·HCl), ethylenediaminetetraacetic acid disodium salt (EDTA-2Na), isopropyl alcohol (C₃H₈O/IPA), silver nitrate (AgNO₃), vitamin E (C₂₉H₅₀O₂), Nafion perfluorinated resin (C₉HF₁₇O₅S, 5 wt% in a mixture of lower aliphatic alcohols and water, contains 45% water), hydrogen peroxide (30 wt% H₂O₂), and 5,5-dimethyl-1-pyrroline-N-oxide (DMPO, a spin-trapping agent for radical detection). All reagents were obtained from commercial sources and were used as received without any further modification.

2.2. Preparation of catalysts

2.2.1. Synthesis of PCN-222. The synthesis of PCN-222 was performed according to a modified literature procedure.³⁴ An amount of 0.2000 g TCPP(H₂) and 0.5500 g ZrOCl₂·8H₂O was dissolved in 50 mL of N,N-dimethylformamide (DMF) under sonication until complete dissolution. The mixture was transferred into a 100 mL Teflon-lined autoclave, followed by the addition of 2.25 mL trifluoroacetic acid (TFA). Subsequently, the reaction was heated under hydrothermal conditions at 120 °C for 20 hours in an oven. Upon cooling to room temperature (r.t.), the precipitate was collected via centrifugation at 8000 rpm for 6 minutes and washed repeatedly with DMF and ethanol. Finally, the product was dried under vacuum at 60 °C for one day to obtain the crystalline PCN-222 framework.

2.2.2. Synthesis of CoFe₂O₄. CoFe₂O₄ nanoparticles were synthesized via a one-step hydrothermal method. Specifically, 4 mmol FeCl₃·6H₂O and 2 mmol CoCl₂·6H₂O were dissolved in 10 mL deionized water under magnetic stirring. NaOH (1 M) was added dropwise to adjust the pH of 10–11. The mixture was transferred into a 25 mL Teflon-lined autoclave and sonicated for 30 minutes to ensure uniform dispersion. Hydrothermal treatment was then conducted at 170 °C for one day. After cooling to r.t., the magnetic CoFe₂O₄ product was separated using a rubidium magnet, followed by repeated washing cycles with deionized water and absolute ethanol to remove impurities. The purified material was dried under vacuum at 60 °C for 12 hours and subsequently ground into powder for subsequent use.

2.2.3. Synthesis of PCN-222/CoFe₂O₄ composites. PCN-222/CoFe₂O₄ composites were prepared via a ball-milling method. Precisely weighed PCN-222 and CoFe₂O₄ powders with varying mass ratios were mixed with 1.5 mm zirconia balls in an MSK-SFM-12M planetary ball mill (Shenzhen Technology Co., Ltd), and the mixture was ball-milled at 3000 rpm for 24 minutes. The resulting composites were designated as P/CF-x, where x (50, 40, 30, 20, 10) denotes the wt% of CoFe₂O₄ in the composite.

2.3. Characterization

The crystalline phase composition of the samples was characterized by X-ray diffraction (XRD, Rigaku Ultima IV, Japan). Morphological features and surface elemental distribution were analyzed using scanning electron microscopy (SEM, TESCAN MIRA LMS, Czech Republic) equipped with energy-dispersive X-ray spectroscopy (EDS), complemented by transmission electron microscopy (TEM, FEI Talos F200X, USA). Fourier-transform infrared spectroscopy (FT-IR, Nicolet 6700, Thermo Fisher, USA) was employed to identify functional groups. The elemental composition and electron transfer behavior were investigated through X-ray photoelectron spectroscopy (XPS, Thermo Scientific K-Alpha, USA). Optical properties were evaluated by ultraviolet-visible diffuse reflectance spectroscopy (UV-vis DRS, Shimadzu UV-3600, Japan) and photoluminescence spectroscopy (PL, Hitachi F-7000, Japan). Time-resolved photoluminescence (TRPL) measurements were performed using an Edinburgh FLS-980 fluorescence spectrometer (UK). The surface photovoltage (SPV) measurements were conducted on a CEL-SPS1000 surface photovoltage testing system. Magnetic properties were determined via vibrating sample magnetometry (VSM, Lake Shore 8604, USA), while pore structure parameters and maximum adsorption capacity were obtained through Brunauer–Emmett–Teller (BET) analysis using an ASAP 2460 system (Micromeritics, USA). A microplate reader (Glomax Multi, Promega Co., USA) was used for luminescent bacteria detection. To probe the formation of ·OH and ·O₂⁻, electron paramagnetic resonance (EPR) signals were recorded on a Bruker EMX Plus spectrometer. In the photocatalytic reaction system, 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) was employed as the spin-trapping agent at a concentration of 0.2 M.

2.4. Photocatalytic measurement

The photo-Fenton catalytic degradation of TCH was evaluated under visible light irradiation using a 300 W xenon lamp (CEL-HXF300, Beijing Zhongjiao Jinyuan Technology Co., China) equipped with a UV cutoff filter (λ > 420 nm). In a typical procedure, 5 mg catalyst was dispersed in TCH aqueous solution (20 mg TCH dissolved in 50 mL H₂O). The suspension was magnetically stirred in the dark for 30 min to establish adsorption–desorption equilibrium. Subsequently, 30 wt% H₂O₂ solution was injected to initiate the reaction. During illumination, 1 mL aliquots were collected at 10 min intervals and immediately filtered through 0.22 μm nylon syringe filters (Millipore) to remove catalyst particles. Quantitative analysis of TCH was performed by high-performance liquid chromatography with the following parameters: C18 reverse-phase column (5 μm particle size, 4.6 mm × 250 mm, Agilent ZORBAX Eclipse Plus) was equipped with a photodiode array (PDA) detector set at a detection wavelength of 280 nm. The mobile phase consisted of acetonitrile/0.1% formic acid aqueous solution (83 : 17, v/v), with a flow rate of 1.0 mL min⁻¹. The injection volume was 20 μL. Degradation intermediates were identified using ultra-performance liquid chromatography coupled with triple quadrupole mass spectrometry (UPLC-HRMS, Thermo Scientific Q Exactive) operated in electrospray ionization (ESI) positive mode. Full-scan mass spectra were acquired over the m/z range 50–600. Chromatographic separation was achieved with the same mobile phase composition as HPLC analysis.

2.5. Photoelectrochemical measurement

Electrochemical impedance spectroscopy (EIS), Mott–Schottky analysis, and transient photocurrent response (I–t) measurements were performed on an electrochemical workstation (CHI-660E, CH Instruments) using Na₂SO₄ (0.1 M) as the electrolyte. A platinum electrode and an Ag/AgCl electrode served as the counter electrode and reference electrode, respectively. Working electrode preparation: 5 mg of the target material was dissolved in a solution containing 800 μL methanol, 200 μL deionized water, and 100 μL Nafion perfluorinated resin. The mixture was sonicated for 5 min to form a homogeneous ink. 10 μL of the resulting dispersion was drop-casted onto the conductive surface of an FTO-coated glass substrate (1 cm × 1 cm) and dried at 60 °C to obtain a uniformly coated working electrode.

3 Results and discussion

3.1. Morphological and structural analysis

The XRD patterns of PCN-222, CoFe₂O₄, and PCN-222/CoFe₂O₄-20 (P/CF-20) are shown in Fig. 1a. For CoFe₂O₄, the detected peaks at 2θ = 18.29°, 30.09°, 35.44°, 43.06°, 56.98° and 62.59° corresponding to the (111), (220), (311), (400), (511) and (440) crystal planes of the standard card of Fd3m CoFe₂O₄, respectively (PDF No. 00-022-1086), demonstrated the successful synthesis of the spinel structure of CoFe₂O₄. The prepared PCN-222 peak pattern at 2θ = 6.68°, 7.10°, 8.24° and 9.74° was consistent with the simulated PCN-222, indicating that PCN-222 was successfully synthesized. Notably, the composite P/CF-20 retains all characteristic peaks of both CoFe₂O₄ and PCN-222 without peak shifts or broadening, indicating that ball milling preserves the crystallinity of the individual components, the crystal plane of the material did not change, and the crystal form remained intact.

Fig. 1 (a) XRD patterns of P/CF-20, PCN-222, simulated PCN-222 and CoFe₂O₄. (b) FT-IR spectrum of TCPP(H₂), PCN-222, P/CF-20 and CoFe₂O₄. SEM images of (c) PCN-222, (d) CoFe₂O₄ and (e) P/CF-20. (f) Elemental mapping (C, N, Zr, Co, Fe, and O) of P/CF-20.

The FT-IR spectra of TCPP(H₂), PCN-222, CoFe₂O₄ and composites are shown in Fig. 1b. For PCN-222, the peaks at 3420 cm⁻¹ and 3311 cm⁻¹ correspond to O–H/N–H stretching vibrations inherited from TCPP(H₂), while the peaks in the 1400∼1600 cm⁻¹ range arise from aromatic C=C skeletal vibrations. The absorption peak for Zr–O was observed at 655 cm⁻¹,^5,35 and the characteristic peaks at 582 cm⁻¹ in CoFe₂O₄ were attributed to Fe–O.^16,36 Critically, these peaks can also be observed in P/CF-20, indicating that the structure of the material was not damaged by ball milling.

SEM images of PCN-222, CoFe₂O₄ and P/CF-20 composites are illustrated in Fig. 1c–f. PCN-222 exhibits a rod-like morphology with diameters ranging from 500 to 700 nm (Fig. 1c), while CoFe₂O₄ is a nanometer-sized spherical material (Fig. 1d). After ball milling, the spherical CoFe₂O₄ material is attached to the surface of the PCN-222 rod material in the composite material as shown in Fig. 1e. From the TEM image shown in Fig. S1, it can be observed that CoFe₂O₄ is tightly bound to the surface of PCN-222. The lattice fringes with a crystal plane spacing of 0.254 nm obtained from analysis are consistent with the (311) plane of CoFe₂O₄. In addition, EDS elemental mappings of P/CF-20, as shown in Fig. 1f, reveal the homogeneous spatial distribution of Zr, C, N, O, Fe, and Co within the composite. Specifically, Zr and N signals originate from the PCN-222 framework, whereas Fe and Co correspond to the CoFe₂O₄ phase. The thermal stability of PCN-222 was examined by TGA under a N₂ atmosphere (Fig. S2), consistent with prior literature studies.³⁷ The observed mass losses correspond to the desorption of physisorbed water molecules (∼100 °C) and the thermal decomposition of TCPP organic linkers (>350 °C).

The interfacial electronic states and elemental composition of P/CF-20 were investigated by XPS. All binding energy values were calibrated against the C 1s peak at 284.8 eV. As shown in Fig. 2a, the XPS survey spectrum of P/CF-20 confirms the presence of Zr, C, N, O, Co, and Fe. The high-resolution O 1s spectrum is deconvoluted into five peaks at 532.96, 532.01, 530.88, 530.20, and 529.65 eV, as shown in Fig. 2b. These peaks can be sequentially attributed to the –OH, C=O and Zr–O in PCN-222, Fe–O and Co–O in CoFe₂O₄. The coexistence of these oxygen species confirms the successful integration of PCN-222 and CoFe₂O₄. Notably, the ·OH peak exhibits a negative shift compared to pristine PCN-222, suggesting enhanced electron density around oxygen atoms due to interfacial interactions. For pristine CoFe₂O₄, the Fe 2p_3/2 and Fe 2p_1/2 peaks are observed at 711.1 and 723.8 eV, respectively (Fig. 2c). In P/CF-20, these peaks shift to 710.8 eV (Fe 2p_3/2) and 724.2 eV (Fe 2p_1/2),³⁸ indicating a positive binding energy shift. This shift reflects a reduction in electron density around Fe atoms. Deconvolution of the Fe 2p_3/2 peak reveals two components at 710.8 eV (Fe²⁺) and 713.1 eV (Fe³⁺), with the Fe²⁺/Fe³⁺ area ratio significantly increasing compared to pristine CoFe₂O₄. This demonstrates the partial reduction of Fe²⁺ to Fe³⁺ during composite formation, consistent with the observed positive binding energy shift. The Co 2p_3/2 spectra of both CoFe₂O₄ and P/CF-20 are resolved into two peaks corresponding to Co³⁺ and Co²⁺, confirming the coexistence of mixed Co oxidation states. This redox-active behavior facilitates Fe²⁺/Fe³⁺ and Co²⁺/Co³⁺ cycling during catalytic reactions. As shown in Fig. 2e, the Zr 3d_5/2 peak shifts from 182.92 eV in PCN-222 to 183.10 eV in P/CF-20, indicating an increase in binding energy. This shift, combined with the altered Fe and O electronic states, suggests the formation of Zr–O–Fe interfacial bonds. The bridging oxygen atom likely originates from μ-OH groups in PCN-222. Upon bond formation, the electronegative oxygen withdraws electron density from both Zr and Fe, leading to reduced electron cloud density and higher binding energies for these metals.

Fig. 2 XPS spectra of PCN-222, pure CoFe₂O₄ and P/CF-20. (a) Survey scan, (b) O 1s, (c) Fe 2p, (d) Co 2p, (e) Zr 3d, and (f) N 1s.

3.2. Optical and electrochemical properties

A comprehensive analysis of the optical and electrochemical properties was conducted through UV-vis DRS, Mott–Schottky measurements, electrochemical impedance spectroscopy (EIS), time-resolved photoluminescence (TRPL) and surface photovoltage (SPV). As shown in Fig. S3, pristine PCN-222 exhibits strong visible-light absorption, while the P/CF-20 composite demonstrates a redshifted absorption edge toward longer wavelengths after CoFe₂O₄ incorporation, indicating enhanced visible-light harvesting capability. As shown in Fig. S4, the optical bandgap energies of PCN-222 and CoFe₂O₄ were determined as 1.81 eV and 1.39 eV, respectively, using the Tauc plot method based on the equation:³⁹

(ahv)² = A(hv − E_g) (1)

where a represents the absorption coefficient, h denotes the Planck constant, A denotes an energy-independent constant, v denotes the light frequency and E_g denotes the band gap. The flat-band potentials of CoFe₂O₄ and PCN-222 were determined via Mott–Schottky measurements (Fig. 3a and b). CoFe₂O₄ exhibits a positive slope in the Mott–Schottky plot, characteristic of a p-type semiconductor, with a flat-band potential of 1.02 eV vs. Ag/AgCl (0.82 eV vs. NHE). Conversely, PCN-222 displays a negative slope, confirming its n-type semiconductor nature, with a flat-band potential of −0.66 eV vs. Ag/AgCl (−0.46 eV vs. NHE). Generally, for p-type semiconductors, the flat-band potential approximates the valence-band maximum (E_VB), while for n-type semiconductors, it aligns with the conduction-band minimum (E_CB). For CoFe₂O₄, the E_VB value is determined to be 0.82 eV (vs. NHE), while the E_CB value of the PCN-222 is positioned at −0.46 eV (vs. NHE). According to formula E_g = E_VB − E_CB, the E_CB value of CoFe₂O₄ was calculated as −0.57 eV, and the E_VB value of PCN-222 is 1.35 eV.

Fig. 3 (a and b) Mott–Schottky plots of CoFe₂O₄ and PCN-222. (c) The transient photocurrent response. (d) EIS Nyquist plots of P/CF-20, PCN-222, and CoFe₂O₄, (e) TRPL spectra of PCN-222 and P/CF-20, and (f) SPV spectra of P/CF-20, PCN-222, and CoFe₂O₄.

Photocurrent response, EIS Nyquist, and TRPL analyses were employed to assess charge carrier dynamics. As shown in Fig. 3c, P/CF-20 exhibits the highest photocurrent density among all samples, indicating superior charge separation efficiency.⁵ The reduced semicircle diameter in EIS Nyquist plots (Fig. 3d) follows the order: PCN-222 > CoFe₂O₄ > P/CF-20, reflecting the lowest charge transfer resistance in P/CF-20. This facilitates rapid interfacial charge migration and enhances the availability of photogenerated carriers for tetracycline hydrochloride degradation.

In PL spectroscopy, lower fluorescence intensity generally indicates a slower recombination rate of photogenerated electron–hole pairs.⁴⁰ PL spectra acquired under 363 nm excitation (Fig. S5) show reduced emission intensity for P/CF-20 compared to PCN-222, signifying suppressed electron–hole recombination.

Furthermore, TRPL spectra of PCN-222 and P/CF-20 were recorded under 375 nm laser excitation, with emission signals detected at 657 nm and 652 nm, respectively (Fig. 3e). The fluorescence lifetimes were calculated using the amplitude-weighted average formula:⁴¹

t = (B₁τ₁² + B₂τ₂²)/(B₁τ₁ + B₂τ₁) (2)

where τ₁ and τ₂ represent radiative lifetimes, and B₁ and B₂ denote the normalized amplitudes of each decay component, while the τ_avg is presented in Table S1. The average fitted carrier lifetime decreases from 1.00 ns (PCN-222) to 0.84 ns (P/CF-20), indicating enhanced non-radiative transitions.⁴² This reduction is attributed to interfacial electron transfer between PCN-222 and CoFe₂O₄, consistent with the proposed Z-scheme heterojunction mechanism, where charge carriers undergo directional migration across the interface, effectively suppressing radiative recombination.

Concurrently, SPV spectroscopy further corroborates these findings. The significantly enhanced SPV signal intensity of P/CF-20 (Fig. 3f) confirms improved charge transfer efficiency toward the surface and demonstrates superior charge-carrier separation efficiency.⁴³

3.3. Photocatalytic activity

3.3.1. Photocatalytic degradation of TCH. TCH (0.02 g L⁻¹) was selected as the target pollutant to assess the photocatalytic performance of the synthesized catalysts under a catalyst dosage of 0.1 g L⁻¹. A 30-minute dark adsorption phase preceded light irradiation to establish adsorption–desorption equilibrium. Comparative degradation experiments were conducted for PCN-222, CoFe₂O₄, and P/CF-x composites (Fig. 4a). After 50 minutes of illumination, pristine CoFe₂O₄ achieved only 15.37% TCH removal, while PCN-222 reached 92.48% TCH removal with the reaction plateauing thereafter. In contrast, all P/CF-x composites exhibited near-complete degradation (99.99%). Notably, within the first 10 minutes of irradiation, PCN-222, P/CF-50, P/CF-40, P/CF-30, P/CF-20, and P/CF-10 reduced TCH concentrations by 23.56%, 32.46%, 31.75%, 31.94%, 41.27%, and 31.15%, respectively. This significant enhancement in catalytic efficiency is attributed to the formation of a Z-scheme heterojunction, which suppresses intrinsic charge recombination and increases the availability of photogenerated carriers for TCH degradation.

Fig. 4 TCH photo-Fenton degradation (a), pseudo-first-order model (b) and rate constant diagram (c) of PCN-222, CoFe₂O₄ and P/CF-x. TCH photo-Fenton degradation (d), pseudo-first-order model (e) and rate constant diagram (f) with different H₂O₂ quantities of the reaction system. TCH photo-Fenton degradation (g), pseudo-first-order model (h) and rate constant diagram (i) at different pH values of the system.

The reaction kinetics were analyzed using a pseudo-first-order kinetic model (Fig. 4b), where k represents the apparent rate constant. As shown in Fig. 4c, the k values of P/CF-x initially increased with higher CoFe₂O₄ loading but declined at excessive loadings. This reduction arises from CoFe₂O₄ agglomeration, which impedes light absorption and weakens heterojunction synergy. The optimal P/CF-20 composite exhibited the highest k value of 0.1623 min⁻¹, demonstrating its superior catalytic activity. Consequently, P/CF-20 was selected for subsequent mechanistic and application studies. Nitrogen adsorption–desorption analysis was performed at 77 K to evaluate the specific surface area and permanent porosity (Fig. S6). The P/CF-20 demonstrated a Brunauer–Emmett–Teller (BET) surface area of 1099.8 m² g⁻¹ with a Type IV isotherm characteristic of mesoporous materials, accompanied by a mesoporous structure around 2.5 nm pore size distribution. This configuration facilitates the adsorption of TCH molecules with dimensions of 14.68 × 7.32 × 10.52 Å (Fig. S7) within the pore channels,⁴⁴ thereby enhancing their accessibility to reactive species for subsequent attack and degradation.

3.3.2. Effect of H₂O₂ concentration. H₂O₂ plays a pivotal role in the photo-Fenton system. To investigate its dosage effect, H₂O₂ volumes ranging from 0 to 300 μL were tested (Fig. 4d). The degradation kinetics was fitted using a pseudo-first-order model (Fig. 4e). As shown in Fig. 4f, the apparent rate constant (k) increased progressively from 0.0821 to 0.1454 min⁻¹ as H₂O₂ dosage rose from 0 to 250 μL. However, further increasing H₂O₂ to 300 μL reduced k to 0.1393 min⁻¹. This decline is attributed to radical scavenging effects, where excess H₂O₂ reacts with ·OH via:⁴⁵

This side reaction decreases active oxidative species, thereby diminishing degradation efficiency.

3.3.3. Effect of pH. The photocatalytic degradation efficiency of TCH is significantly influenced by solution pH, which governs both the ionization state of TCH and the surface charge properties of the catalyst. Initial pH values (2, 4, 6, and 8) were adjusted using 1 M HCl or NaOH. As shown in Fig. 4g, the degradation efficiency markedly decreases at pH 2, likely due to partial deactivation of P/CF-20 under strongly acidic conditions, which destabilizes the catalyst structure and reduces active site accessibility. Optimal catalytic performance is achieved under weakly acidic conditions (pH 6), yielding a reaction rate constant of 0.1452 min⁻¹. When pH = 2, excessive H⁺ ions protonate catalyst surface groups, impeding TCH adsorption. Potential dissolution of metal species (e.g., Fe³⁺ and Co²⁺) destabilizes the heterojunction.

3.4. Stability and reusability

The stability and recyclability of photocatalysts are critical for practical applications. Recycling experiments were conducted by recovering P/CF-20 after photocatalytic reactions (Fig. 5a). After six consecutive degradation cycles, the adsorption capacity of P/CF-20 slightly decreased, leading to a minor reduction in photocatalytic rate, likely due to residual contaminants trapped within its pores. Nevertheless, the degradation efficiency for TCH remained consistently high at 96.48% at 50 min with no significant loss. The magnetic recyclability of P/CF-20 was assessed via Vibrating Sample Magnetometry (VSM). As shown in Fig. 5b, the saturation magnetization of P/CF-20 is 13.35 emu g⁻¹, enabling facile magnetic separation from the reaction system. Post-reaction X-ray diffraction (XRD) characterization after four cycles (Fig. 5c) confirmed the structural integrity of P/CF-20. All characteristic diffraction peaks remained unchanged, and no obvious phase transformations were observed, demonstrating its robust crystallinity and chemical stability. These results validate P/CF-20 as an efficient and recyclable heterogeneous photocatalyst. P/CF-20 exhibited remarkable efficacy even at elevated TCH concentrations (Fig. S8). At a catalyst dosage of 0.1 g L⁻¹, degradation efficiencies reached 96.97% for 0.04 g per L TCH and 94.13% for 0.05 g per L TCH within 50 min. The marginal decline at higher pollutant loads may arise from mass transfer limitations or active site saturation.

Fig. 5 (a) Photocatalytic performance of P/CF-20 under visible light irradiation over for consecutive cycles. (b) VSM of CoFe₂O₄ and P/CF-20. XRD pattern (c) obtained before and after the reaction. (d) Effect of inorganic anions and degradation effect in actual water. (e) Radical capture experiment of TCH and the rate constant diagram for the pseudo-first-order model. (f) EPR diagram of P/CF-20 for DMPO-·O₂⁻ under visible light irradiation.

In practical wastewater systems, inorganic anions (e.g., Cl⁻ and SO₄²⁻) commonly coexist with TCH and may adversely affect photocatalytic degradation. To investigate the influence of inorganic anions on catalytic degradation efficiency in real aqueous environments, 4 mM solutions of Cl⁻, SO₄²⁻, PO₄³⁻, CO₃²⁻, HCO₃⁻, or NO₃⁻ were individually introduced into the reaction system. Parallel degradation assessments were performed using seawater and tap water as representative environmental matrices. As shown in Fig. 5d, the addition of Cl⁻ and SO₄²⁻ induced moderate inhibition of degradation efficiency (8.75–9.24% reduction), attributable to radical scavenging reactions in which these anions convert ·OH into less reactive ·Cl₂⁻ and ·SO₄⁻, respectively.^46,47 The introduction of NO₃⁻ induced significant suppression of the degradation efficiency (22% reduction), primarily attributed to its capacity to scavenge surface-bound hydroxyl radicals (·OH) and photogenerated holes (h⁺) at the catalyst interface.⁴⁸

Performance assessments in actual water matrices revealed robust degradation capabilities in both tap water (∼92.53%) and seawater (∼97.26%), demonstrating strong application potential. Notably, seawater matrices show higher degradation efficiency relative to tap water, presumably due to abundant dissolved metal ions providing supplementary Fenton-like active sites for radical generation.

3.5. Photodegradation mechanism of TCH by the P/CF-20 photocatalyst

To elucidate the reactive species involved in the degradation process, radical trapping experiments were performed using AgNO₃ (e⁻ scavenger), EDTA-2Na (h⁺ scavenger), vitamin E (·O₂⁻ scavenger), and isopropanol (IPA, ·OH scavenger) during P/CF-40-mediated photo-Fenton catalysis (Fig. 5e). Pseudo-first-order kinetic fitting (Fig. S9) revealed that TCH degradation efficiencies at 40 min were 99.76%, 90.26%, 99.01%, and 98.71% under the respective scavenging conditions, compared to 97.37% for the control group (k = 0.0801 min⁻¹). The significant inhibition observed with EDTA-2Na (k = 0.0501 min⁻¹) confirms that h⁺ serves as the dominant active species in TCH degradation. Conversely, AgNO₃ accelerated the reaction (k = 0.143 min⁻¹), likely due to enhanced h⁺ availability via electron trapping. The partial activity retention with vitamin E and IPA suggests secondary roles of ·O₂⁻ and ·OH in the degradation process. EPR spectroscopy with DMPO spin-trapping (Fig. 5f) confirmed the generation of ·O₂⁻ radicals under visible light. No signals were detected in the dark, whereas six characteristic peaks of DMPO-·O₂⁻ emerged after 5 min of illumination, verifying the production of ·O₂⁻ by P/CF-x composites.

The interfacial charge transfer mechanism in P/CF-20 was further validated through DFT calculations and in situ XPS analysis. A heterojunction model was constructed by coupling the (311) facet of CoFe₂O₄ with PCN-222, followed by structural optimization (Fig. S10). The differential charge density in Fig. 6a visually demonstrates the charge rearrangement at the interface between PCN-222 and CoFe₂O₄. The average charge density plot (Fig. 6b) and integral curve plot (Fig. 6c) further elucidate the direction of electrons transfer along the z-axis plane, revealing directional electron transfer from PCN-222 to CoFe₂O₄ upon heterojunction formation and yielding a net electron gain of 0.237|e| within the interfacial region (Z < 8.339 Å). As shown in Fig. S11, DFT calculations demonstrate the heterojunction's adsorption affinity for H₂O₂, yielding a low adsorption energy of −3.325 eV. This indicates the stable adsorption of H₂O₂ on the heterojunction surface, followed by its cleavage, which facilitates the Fenton reaction. Fig. 6d–f show the in situ irradiation XPS spectra of P/CF-20. Compared to the peaks under dark conditions, the Co 2p and Fe 2p peaks shift to lower binding energies after light irradiation, while the Zr peak shifts to a higher binding energy. This suggests the transfer of photogenerated electrons from PCN-222 to CoFe₂O₄, further supporting the formation of a Z-scheme heterojunction.

Fig. 6 (a) Differential charge density map of the PCN-222/CoFe₂O₄ interface (the blue region indicates electron depletion, the yellow region indicates electron accumulation, and the level of the isosurface is set to 0.005 e Å⁻³). Plot graph of the plane-averaged curve (b) and integral curve (c) of the heterojunction. XPS spectra of (d) Co 2p (e) Fe 2p, and (f) Zr 3d for P/CF-20 in darkness and under light illumination.

As illustrated in Scheme 1, the energy band alignment of PCN-222 and CoFe₂O₄ forms a Z-scheme heterojunction. Under visible light, electrons (e⁻) of PCN-222 are excited from the VB (1.35 eV vs. NHE) to the CB (−0.46 eV vs. NHE), while e⁻ in CoFe₂O₄ migrates from its VB (0.82 eV vs. NHE) to the CB (−0.57 eV vs. NHE). The CB electrons from PCN-222 recombine with VB holes from CoFe₂O₄, thereby preserving highly oxidative holes (h⁺) in the VB of PCN-222 and reductive electrons (e⁻) in the CB of CoFe₂O₄. The CB potential of CoFe₂O₄ is more negative than the O₂/·O₂⁻ redox potential (−0.33 eV vs. NHE),⁴⁹ enabling CB electrons to reduce dissolved O₂ to ·O₂⁻. Additionally, ·OH can be produced via Fenton-like reactions where divalent metal ions (M²⁺, M = Fe and Co) activate H₂O₂:

M²⁺ + H₂O₂ → M³⁺ + ·OH + OH⁻ (M = Fe, Co)

Scheme 1 Possible photocatalytic mechanism of the P/CF-x heterojunction.

The redox potential of Co²⁺/Co³⁺ (1.81 V) was higher than that of Fe²⁺/Fe³⁺ (0.77 V),⁵⁰ which suggested that the reduction of Co³⁺ by Fe²⁺ was feasible. These redox reactions of Co and Fe could promote the continuous activation of H₂O₂ to produce free radicals. The redox cycling of Fe and Co synergistically enhances ·OH production, while the Z-scheme mechanism prolongs charge carrier lifetimes. This dual functionality drives efficient TCH degradation, achieving near-complete mineralization within 40 min.

3.6. Photodegradation intermediate analysis

The intermediate products generated during the photocatalytic degradation of TCH were identified using liquid chromatography-mass spectrometry (LC-MS), and two plausible degradation pathways were proposed based on experimental evidence (Fig. 7a). In Pathway I, intermediate P1 is formed primarily through simultaneous dehydration, hydroxylation of the amide group mediated by ·OH, and loss of amino groups. P1 undergoes sequential transformations, demethylation and keto–enol tautomerization yielding P2 and P3. Reduction of the ketone group in P3 by h⁺, followed by deacetylation, produces P4. Dehydration, oxidation, and ring-opening reactions convert P4 into smaller intermediates P5 and P6. In Pathway II, TCH first loses a water molecule under h⁺ attack to form P7.⁵¹ The N-methyl group of P7 is oxidized by ·O₂⁻, leading to further dehydration and loss of an amino group, generating intermediates P8 and P17.⁵² In the sub-pathway II-1, P8 is oxidized and undergoes ring-opening to form P9. Sequential dealkylation and ring cleavage of P9 produce P11. ·OH abstracts a hydrogen atom from the phenolic moiety of P11, generating a phenoxy radical that undergoes β-scission or hydrogen transfer, resulting in dehydroxylation to form P13.⁵³ Progressive fragmentation of P13 yields monocyclic intermediates P14, P15, and P16. In the Sub-pathway II-2, P17 undergoes ring-opening and dealkylation to form P18. Subsequent demethylation and amino group loss convert P18 into P19. P20 is generated via h⁺-mediated reduction of a ketone group and ·OH attack on unsaturated carbons, followed by ring cleavage to yield small-molecule P21. All intermediates derived from these pathways are ultimately mineralized into CO₂, H₂O, and NH₄⁺, confirming the near-complete degradation of TCH.

Fig. 7 (a) Possible pathways of TCH photodegradation. (b) Toxicity assessment through luminescent bacteria. Acute toxicity changes during the degradation of TCH and its intermediates: green algae (c), daphnid (d) and fish (e).

To assess the potential secondary pollution of the degraded TCH solution to aquatic ecosystems and organisms, acute toxicity induced by the P/CF-20 photocatalyst was evaluated through the bioluminescence inhibition rate of Vibrio fischeri (Fig. 7b). The initial solution exhibited 28.17% bacterial luminescence inhibition, attributable to the EC₅₀ concentration of 104 μM.⁵⁴ At 10 min degradation, inhibition increased to 34.92%, likely due to the generation of toxic intermediates. Ultimately, the inhibition rate decreased to 9.13% in the fully degraded solution.

Furthermore, ECOSAR modeling was employed to simulate the ecotoxicity of TCH and 21 proposed intermediates. According to the toxicity classification scale, for acute toxicity, P18 (m/z = 146) showed elevated toxicity to green algae (Fig. 7c), while P1, P18, P19, and P21 demonstrated increased toxicity to fish (Fig. 7e). These toxic intermediates may account for the initial toxicity surge observed at 10 min degradation during the photo-Fenton process. As shown in Fig. S12, chronic toxicity simulations revealed only P18 (m/z = 146) exhibited a marginal increase in chronic toxicity to daphnids compared to TCH. All other intermediates displayed significantly reduced chronic toxicity. The final degradation products demonstrate negligible environmental risks, confirming the solution's aquatic safety and eco-friendly application potential.

4 Conclusions

In this study, a series of Z-scheme heterojunction photocatalysts (P/CF-x) with varying mass ratios of CoFe₂O₄ were successfully synthesized via a ball-milling method. The optimized P/CF-20 composite exhibited exceptional adsorption capacity and photo-Fenton catalytic activity under visible light irradiation. Compared to pristine PCN-222 and CoFe₂O₄, P/CF-20 demonstrated superior degradation performance, achieving 99.46% removal of tetracycline hydrochloride (TCH, 0.02 g L⁻¹) within 30 minutes under optimized conditions (0.1 g per L catalyst, pH 6, 250 μL 30 wt% H₂O₂), exhibiting accelerated kinetics with an apparent rate constant (k) of 0.1452 min⁻¹, representing a statistically significant 3-fold enhancement over pristine PCN-222 (k = 0.047 min⁻¹). This enhanced activity is primarily attributed to the effective construction of the Z-scheme heterojunction, which facilitates efficient electron transfer from the CB of PCN-222 to the VB of CoFe₂O₄, thereby suppressing charge recombination. Characterization techniques, including PL, TRPL, SPV and photoelectrochemical analyses, consistently corroborated the improved charge separation efficiency and catalytic performance. EPR and radical trapping experiments identified h⁺, ·OH and ·O₂⁻ as the dominant reactive species responsible for TCH degradation. The proposed reaction mechanism, derived from band structure analysis and experimental evidence, highlights the synergistic interplay between Z-scheme charge transfer and radical-mediated oxidation. Furthermore, liquid chromatography-mass spectrometry (LC-MS) analysis revealed two plausible degradation pathways for TCH, involving sequential hydroxylation, dealkylation, ring-opening, and mineralization into CO₂, H₂O, and NH₄⁺, effectively reducing its biotoxicity. This work offers novel insights for designing efficient Z-scheme heterojunction photocatalysts that are magnetically recoverable, aimed at degrading antibiotics via the photo-Fenton process. This approach provides a sustainable strategy for wastewater remediation.

Author contributions

Jiaxin Li: conceptualization, methodology, formal analysis, data curation, writing – original draft, visualization, investigation, validation. Jingchao Liu: supervision, writing – review and editing. Jiawei Li: supervision, writing – review and editing. Yaping Li: supervision, software. Ziyang Zhu: supervision, writing – review and editing. Zirui Wang: visualization, investigation. Yilin Yin: conceptualization, methodology, supervision, writing – review and editing. Zenghe Li: conceptualization, methodology, supervision, writing – review and editing.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the SI.

Supplementary information is available and includes the following: calculation methods, results from TEM, TGA, UV-vis, PL, and BET measurements, as well as other characterization data, and a comparison of the performance reported herein with that of other recently reported photocatalysts. See DOI: https://doi.org/10.1039/d5ta02778k.

Acknowledgements

The authors would like to thank Professor Zenghe Li for providing financial support and guidance.

References

1. B. D. Petersen, T. C. B. Pereira, S. Altenhofen, D. D. Nabinger, P. M. d. A. Ferreira, M. R. Bogo and C. D. Bonan, Comp. Biochem. Physiol., Part C: Toxicol. Pharmacol., 2021, 242, 108936.
2. J. Scaria, K. V. Anupama and P. V. Nidheesh, Sci. Total Environ., 2021, 771, 145291.
3. Z. Zhang, Y. Chen, Z. Wang, C. Hu, D. Ma, W. Chen and T. Ao, Appl. Surf. Sci., 2021, 542, 148662.
4. H. Yin, Y. Cao, T. Fan, P. Li, S.-m. Chen and X. Liu, J. Taiwan Inst. Chem. Eng., 2021, 119, 70–79.
5. J.-C. Lu, X.-L. Bai, Q.-Y. Zhao, B.-B. Lu and Y. Fu, J. Cleaner Prod., 2024, 456, 142364.
6. M. Grzegorczyk, S. Pogorzelski and P. Rochowski, J. Environ. Chem. Eng., 2022, 10, 107983.
7. L. Yang, Y. Jiao, X. Xu, Y. Pan, C. Su, X. Duan, H. Sun, S. Liu, S. Wang and Z. Shao, ACS Sustain. Chem. Eng., 2022, 10, 1899–1909.
8. H. Zheng, Y. Hou, S. Li, J. Ma, J. Nan and T. Li, Chin. Chem. Lett., 2022, 33, 5013–5022.
9. H. Wang, X. Yuan, Y. Wu, G. Zeng, H. Dong, X. Chen, L. Leng, Z. Wu and L. Peng, Appl. Catal., B, 2016, 186, 19–29.
10. C. Qian, R. Wang, H. Shen, J. Xia, D. Cui, K. Sun, H. Liu, C. Guo, F. Yu, J. Li and W. Bao, Small, 2023, 19, e2304045.
11. A. R. Bracamontes-Ruelas, Y. Reyes-Vidal, J. R. Irigoyen-Campuzano and L. Reynoso-Cuevas, Catalysts, 2023, 13, 748.
12. X. Shen, Z. Zhu, X. Liu, R. Qiao, H. Zhang and Y. Qiu, J. Environ. Chem. Eng., 2024, 12, 113299.
13. L. R. Rad, M. Irani and M. Anbia, J. Environ. Chem. Eng., 2024, 12, 113302.
14. C. Qi, S. Han, J. Lin, J. Cheng, K. Du, Y. Hu and Y. Chen, Catalysts, 2022, 12, 671.
15. Y. Y. Liu, L. J. Chen, X. Zhao and X. P. Yan, Chemistry, 2021, 27, 10151–10159.
16. S. Kumar, M. Kumar and A. Singh, Contemp. Phys., 2022, 62, 144–164.
17. A. R. Tanna, K. M. Sosa and H. H. Joshi, Mater. Res. Express, 2017, 4, 115010.
18. S. O. Aisida, T. C. Chibueze, M. H. Alnasir, O. E. Oyewande, A. T. Raji, C. E. Ekuma, I. Ahmad, T. k. Zhao, M. Maaza and F. I. Ezema, Solid State Sci., 2023, 136, 107107.
19. J. Yan, B. Chai, Y. Liu, G. Fan and G. Song, Appl. Surf. Sci., 2023, 607, 155088.
20. G. He, Y. Wen, C. Ma, X. Li, L. Gao and Z. Sun, Int. J. Hydrogen Energy, 2021, 46, 5369–5377.
21. J. Rodríguez-Chueca, E. Barahona-García, V. Blanco-Gutiérrez, L. Isidoro-García and A. J. Dos santos-García, Chem. Eng. J., 2020, 398, 125606.
22. C. Du, Z. Zhang, G. Yu, H. Wu, H. Chen, L. Zhou, Y. Zhang, Y. Su, S. Tan, L. Yang, J. Song and S. Wang, Chemosphere, 2021, 272, 129501.
23. S. Dawood, S. Shaji, G. Pathiraja, Y. Mo and H. Rathnayake, Phys. Chem. Chem. Phys., 2021, 23, 21677–21689.
24. K. Yu, Y.-R. Lee, J. Y. Seo, K.-Y. Baek, Y.-M. Chung and W.-S. Ahn, Microporous Mesoporous Mater., 2021, 316, 110985.
25. M. Aghayi-Anaraki and V. Safarifard, Eur. J. Inorg. Chem., 2020, 2020, 1916–1937.
26. D.-T. Tran, T.-T.-U. Bui, A.-T. Phan and T.-B. Pham, Environ. Technol. Innovation, 2024, 33, 103502.
27. G. Rui, Z. Hu, H. Wu and J. Ma, J. Porous Mater., 2025, 32, 1333–1349,  DOI:10.1007/s10934-025-01774-1.
28. H. Zhang, Q. Meng, H. Li, G. Wu, K. Li, J. Xu, L. Wang, J. Wu, X. Meng and H. Hou, CrystEngComm, 2022, 24, 7611–7619.
29. S. Chen, Q. Tang, Y. Zeng, Y. Yang, T. Zhu, H. Wang, L. Guo, L. Li and Z. Qian, Mikrochim. Acta, 2023, 190, 70.
30. M. Kim, M. Al Mamunur Rashid, Y. J. Choe, S. H. Kim, J.-H. Lee, K. Jeong and J. Kim, J. Mater. Chem. A, 2023, 11, 9436–9454.
31. W. J. He, H. P. Jia, Z. P. Li, C. Q. Miao, R. H. Lu, S. B. Zhang and Z. Q. Zhang, J. Environ. Chem. Eng., 2022, 10, 108698.
32. R. Bhosale, S. Bhosale, R. Sankannavar, V. Chavan, C. Jambhale, H. Kim and S. Kolekar, ACS Appl. Nano Mater., 2024, 7, 4078–4091.
33. P. Liu, P. Chen, Z. Xing, Z. Li, H. Liu, Y. Wang, Y. Yang, Y. Wang and W. Zhou, J. Mater. Chem. A, 2024, 12, 12155–12162.
34. D.-Y. Zheng, E.-X. Chen, C.-R. Ye and X.-C. Huang, J. Mater. Chem. A, 2019, 7, 22084–22091.
35. S. Zhong, H. Huang, X. Wang, X. Yu and X. Shentu, Lwt, 2024, 198, 115996.
36. K. Manna and S. K. Srivastava, ACS Sustain. Chem. Eng., 2017, 5, 10710–10721.
37. S. Karamzadeh, E. Sanchooli, A. R. Oveisi, S. Daliran and R. Luque, Appl. Catal., B, 2022, 303, 120815.
38. X. Zhao, X. Zhang, Z. Xue, W. Chen, Z. Zhou and T. Mu, J. Mater. Chem. A, 2019, 7, 27417–27422.
39. J. Zhuang, L. Wang, Y. Hu, Y. Wang, L. Lin, J. Xiao, Y. Chen, X. Li, H. Liu, Q. Zhang and Q. An, J. Mater. Chem. A, 2024, 12, 23789–23800.
40. H. Lin, J. Sun, M. He and G. Wang, J. Mater. Chem. A, 2025, 13, 10855–10862,  10.1039/d5ta00062a.
41. Y. Zhang, Z. He, J. Xiong, S. Zhan, F. Liu, M. Su, D. Wang, Y. Huang, Q. Liao, J. Lu, Z. Zhang, C. Yuan, J. Wang, Q. Dai and J. Zhang, J. Mater. Chem. A, 2024, 12, 10369–10384.
42. X. Li, D. Wu, J. Yang, Q. Zhu, J. Wang, Y. Peng, T. Xie and H. Chen, Sol. Energy, 2021, 220, 440–449.
43. Z. Yuan, Y. Cao, Y. Meng, G. Pan, Y. Zheng, Z. Ni and S. Xia, J. Hazard. Mater., 2023, 458, 131895.
44. S. Cong, Y. Zhou, C. Luo, C. Wang, J. Wang, Z. Wang and X. Liu, Angew. Chem., Int. Ed., 2024, 63, e202319894.
45. C. Liu, H. Dai, C. Tan, Q. Pan, F. Hu and X. Peng, Appl. Catal., B, 2022, 310, 121326.
46. T. Wang, C. Zhao, L. Meng, Y. Li, D. Wang and C.-C. Wang, Appl. Catal., B, 2023, 334, 122832.
47. Z. Zhu, L. Wang, W. Zhang, C. Hou, C. Wang and J. Zhao, Chem. Eng. J., 2024, 481, 148313.
48. M. Omar, S. Omar, A. Eltaweil, G. M. El-Subruiti and N. F. Attia, Surf. Interfaces, 2023, 43, 103606.
49. D. Pan, Y. Lu, A. M. Idris, Z. Chen, L. Xu, J. Wang, G. Jiang, Z. Chen and Z. Li, J. Mater. Chem. A, 2024, 12, 10461–10471.
50. H. Wang, X. Zhang, X. Wu, Y. Kang, Y. Huang, Z. Zhao, Z. Li, T. Zhou, J. Wang, W. Zhu and C. Pan, J. Mater. Chem. A, 2025, 13, 9952–9962,  10.1039/d4ta06982j.
51. T. Zhang, X. Guo, H. Pei, F. Zha, X. Tang, H. Tian, R. Guo, N. Liu and Z. Mo, Colloids Surf., A, 2023, 659, 130822.
52. C. Hu, J. He, J. Liang, T. Lin and Q. Liu, Environ. Res., 2023, 231, 116313.
53. L. Pratali Maffei, M. Pelucchi, T. Faravelli and C. Cavallotti, React. Chem. Eng., 2020, 5, 452–472.
54. Q. Zhang, J. Xu, X. Ma, J. Xu, Z. Yun, Q. Zuo and L. Wang, J. Water Process Eng., 2021, 44, 102364.

---