Synthesis of catalysts

The catalyst was synthesized through a co-precipitation method, as described by the chemical reaction in Equation 1. FeSO₄·7H₂O and FeCl₃·6H₂O were dissolved in 100 mL of deionized water at varying Fe(II):Fe(III) molar ratios, with a constant total iron concentration of 0.45 mol L^–1. The mixture was ultrasonically agitated until the iron salts were completely dissolved. The solution was sealed with cling film and stirred magnetically at a constant temperature of 50°C and a fixed rotational speed of 500 rpm for 20 min to homogenize the mixture. Subsequently, 10 mL of concentrated ammonia solution (13 mol L^–1) was added dropwise manually at a controlled rate (∼1 mL min^–1) to initiate precipitation. The reaction mixture was then stirred continuously at 50°C and 500 rpm for 30 min to allow particle maturation. The resulting black precipitate was collected by centrifugation, washed thoroughly with deionized water and dried at 60°C for 8 h in a vacuum oven. The final product (magnetic nanoparticles; MNPs) was obtained by gentle grinding of the dried powder.

(1)\begin{equation}{\text{F}}{{\text{e}}^{2 + }} + 2{\text{F}}{{\text{e}}^{3 + }} + 8{\text{O}}{{\text{H}}^ - } \to {\text{F}}{{\text{e}}_3}{{\text{O}}_4} + 4{{\text{H}}_2}{\text{O}}\end{equation}

Experimental procedure

Degradation experiments were conducted in batch mode. Typically, 50 mL of TC solution (20 mg L^–1) was placed in a conical flask. MNP catalyst (0.1 g L^–1 unless specified otherwise) was added, and the mixture was shaken in the dark for 30 min to establish adsorption–desorption equilibrium. The reaction was initiated by adding H₂O₂ (50 mM unless specified otherwise). Aliquots were withdrawn at predetermined intervals and immediately filtered through a 0.22 μm polyethersulfone (PES) syringe filter to remove the catalyst. The filtered solution was then quenched (if necessary). The residual TC concentration was analysed using high-performance liquid chromatography (HPLC; Dalian Elite EClassical-3200; column: C18, mobile phase: methanol/oxalic acid (0.01 M) = 35:65 (v/v), flow rate: 1.0 mL min^–1, detection wavelength: 355 nm). Degradation kinetics were fitted to the pseudo-first-order model (Equation 2):

(2)\begin{equation} - \ln \left( {\frac{{{C_t}}}{{{{\text{C}}_0}}}} \right) = {{\text{K}}_{{\text{obs}}}}t\end{equation}

where C ₀ and C_t are the TC concentrations (mg L^–1) at initial time and time t (min), respectively, and K_obs is the apparent rate constant (min^–1). After reaction, the catalyst was recovered using a magnet, washed and dried for potential reuse studies.

X-ray diffraction analysis

The phase composition and crystal structure of the samples were determined by X-ray diffraction (XRD; D8 ADVANCE, Bruker, Germany) with Cu-Κα radiation (λ = 0.1541nm) operated at 35 kV and 35 mA. The data were collected in the 10–80°2θ range with a step size of 0.05° and a scanning speed of 5° min^–1. For sample preparation, the MNP powders were fully ground in an agate mortar and sieved to 200 mesh. The powders were then evenly sprinkled into the sample holder cavity, gently compacted with a spatula and finally flattened with a glass slide to form a smooth surface level with the holder. Phase identification was performed using the International Centre for Diffraction Data (ICDD) PDF-4+ database with the aid of MDI Jade 6 software.

Fourier-transform infrared spectrometry analysis

Fourier-transform infrared (FTIR) spectrometry analysis was performed using an FTIR-850 spectrometer (Tianjin Gangdong Sci. & Tech. Co., Ltd, China) operated in transmission mode. The samples were prepared by mixing them with dried KBr powder and pressing them into pellets. Spectra were collected in the wavenumber range of 4000 to 500 cm^–1 with a resolution of 4 cm^–1 and an average of 64 scans.

X-ray photoelectron spectroscopy analysis

The surface chemical states and bonding configurations were examined via X-ray photoelectron spectroscopy (XPS) on an ESCALAB250Xi spectrometer (Thermo Fisher Scientific, Inc., USA) using a monochromatic Al-Κα X-ray source (1486.6 eV). The analysis was performed with a spot size of 500 μm and a pass energy of 30 eV for high-resolution regional scans. The binding energy scale was calibrated by setting the adventitious carbon C 1s peak to 284.8 eV for charge correction. A Shirley background was applied for baseline subtraction, and the spectra were deconvoluted using the Thermo Avantage software with a Gaussian–Lorentzian (GL (30)) line shape for peak fitting.

Scanning electron microscopy analysis

The morphology of the samples was characterized using field-emission scanning electron microscopy (SEM; Nova Nano SEM, Thermo Fisher Scientific, Inc., USA). The powder samples were directly adhered onto conductive carbon tape. After removing excess powder via gentle air blowing, the samples were sputter-coated with a thin gold layer to enhance conductivity. Imaging was performed at an acceleration voltage of 7 kV and a working distance of 15 mm.

N₂ adsorption/desorption analysis

The specific surface area and pore-size distribution of the prepared samples were studied by using a specific surface area and pore-size analyser (BET, JW-BK200A, JWGB Instruments, China). Prior to measurement, ∼50 mg of each sample was degassed under a dynamic vacuum at 120°C for 6 h to remove moisture and other adsorbed contaminants. The specific surface area was calculated using the Brunauer–Emmett–Teller (BET) method in the relative pressure (P/P ₀) range of 0.05–0.30.

Vibrating sample magnetometer analysis

The magnetic properties of the samples were measured using a vibrating sample magnetometer (VSM; Model 7404, Lake Shore Cryotronics, USA). The powder sample was packed into a non-magnetic gelatin capsule for measurement. The hysteresis loop was recorded at room temperature with an applied magnetic field sweeping from –20 to +20 kOe.

The degradation of TC by the MNP/H₂O₂ system

The synthesis route of MNP is depicted in Fig. 1. Four types of MNPs with Fe²⁺/Fe³⁺ molar ratios of 0.3, 0.5, 0.7 and 1.0 (denoted as MNP-0.3, MNP-0.5, MNP-0.7 and MNP-1.0, respectively) were synthesized to explore the optimal conditions, and the key synthesis parameters are summarized in Table 1. Regarding the adsorption–degradation effect of MNPs with different molar ratios on TC, as shown in Fig. 2, the adsorption–degradation effect of MNP-0.5 on TC is significantly better than those of the other materials. Upon adding H₂O₂, the degradation efficiencies of TC by MNP-0.3, MNP-0.7 and MNP-1.0 reached 96.04%, 88.70% and 86.44% within 120 min, respectively. In contrast, MNP-0.5 achieved complete degradation (100%) of TC within just 90 min. These results demonstrate that both the adsorptive and catalytic performances of MNP-0.5 are superior to those of the other materials. This enhanced activity may be attributed to the Fe²⁺/Fe³⁺ ratio in MNP-0.5 being closest to the theoretical value of 0.5 (Equation 1), which promotes the formation of high-purity Fe₃O₄ with optimal catalytic properties.

Figure 1.

Schematic illustration of the MNP synthesis process.

Figure 2.

Adsorption–catalytic degradation of TC by MNPs. Reaction conditions: [catalyst] = 0.1 g L^–1, [H₂O₂] = 50 mM, [TC] = 20 mg L^–1, pH = 7, T = 298 K.

Table 1.

Key synthesis parameters of MNP with various molar ratios.

The characterization of MNP

XRD was performed to characterize the crystallographic structures of the various MNP catalysts. As shown in Fig. 3a, distinct diffraction peaks were observed at 35.7°2θ, 57.1°2θ and 63.07°2θ, which correspond to the (311), (511) and (440) crystal planes of Fe₃O₄ (PDF #88-0315), respectively, confirming the successful synthesis of Fe₃O₄ (Liu et al., Reference Liu, Gan, Li, Xiong, Xu, Guo and Xue2025b). Furthermore, no detectable impurity peaks were present in the XRD trace, indicating the high purity of the synthesized product. However, the baseline of the trace exhibits slight fluctuations and a curved profile, which may be attributed to the presence of amorphous iron hydroxides (ferric and ferrous hydroxide) formed during the ammonia addition step. These amorphous intermediates may not have fully transformed into well-crystallized Fe₃O₄ (Liu et al., Reference Liu, Gan, Li, Xiong, Xu, Guo and Xue2025b). Further investigation of the surface functional groups of these four materials was conducted using FTIR spectroscopy. As shown in Fig. 3b, the characteristic peak at 572 cm^–1 corresponds to the stretching vibration of Fe–O bonds in Fe₃O, confirming the successful synthesis of Fe₃O₄ (Lu et al., Reference Lu, Jiao, Chen and Li2009). The absence of characteristic peaks for ferrous oxide at 700 cm^–1 indicates no significant impurity phase in the product. The absorption peaks observed at 1631 and 3442 cm^–1 are attributed to the hydroxyl groups’ bending and stretching vibrations from adsorbed water on the surface, respectively (Liu et al., Reference Liu, Wang, Wang and Zhang2013). When comparing Fe₃O₄ synthesized under various conditions, the characteristic peak intensity corresponding to Fe–O stretching vibrations reaches its maximum for MNP-0.5, demonstrating the optimal purity and structural orderliness of Fe₃O₄ at this molar ratio.

Figure 3.

(a) XRD traces of MNP and (b) FTIR spectra of MNP.

The N₂ adsorption–desorption technique obtained the BET surface areas and pore-size distributions of the MNP catalysts. All samples exhibited similar type-IV curves according to the International Union of Pure and Applied Chemistry (IUPAC) classification (Fig. 4), reflecting their inherent microporous and mesoporous features (Ye et al., Reference Ye, Zeng, Tan, Wu, Liang and Song2020; Huang et al., Reference Huang, Xiao, Yi, Li, Liu and Liu2022). Specifically, MNP-0.5 possessed a larger specific surface area of 68.79 m² g^–1 compared to MNP-0.3 (39.30 m² g^–1), MNP-0.7 (56.89 m² g^–1) and MNP-1.0 (27.98 m² g^–1).

Figure 4.

N₂ adsorption–desorption isotherms of MNPs.

Based on the specific surface area measurements, MNP-0.5 and MNP-0.7 were selected for further investigation of their surface morphology, and the catalysts were characterized by SEM. As shown in Fig. 5a, MNP-0.5 is nanoscale in size and presents a uniformly distributed, spherical, stacked structure but a tendency to form aggregates. This is typical of magnetic nanomaterials synthesized by co-precipitation due to their strong interparticle magnetic dipole interactions and hydrogen bonding facilitated by their surface hydroxyl groups (Fouad et al., Reference Fouad, Zhang, El-Didamony, Yingnan, Mekuria and Shah2019; Rani et al., Reference Rani, Winantari, Amrillah and Setyawan2025). Statistical analysis revealed that the spherical MNP-0.5 particles predominantly ranged between 20 and 30 nm in diameter, with an average size of 23.36 nm (Fig. S1). As shown in Fig. S2, MNP-0.7 exhibited a similar spherical packing morphology. The particle sizes of MNP-0.7 were primarily distributed from 20 to 40 nm, with an average diameter of 29.85 nm (Fig. S3). The successful synthesis of nanoscale Fe₃O₄ was confirmed. Compared with MNP-0.7, the particle size of MNP-0.5 is smaller, which is also conducive to the greater exposure of active sites (Zou et al., Reference Zou, Li, Li and Zhao2025). VSM measurements (Fig. 5b) revealed that MNP-0.5 displays paramagnetic behaviour, characterized by a high saturation magnetization (M_s) of 71.02 emu g^–1 and negligible coercivity and remanence. This strong magnetic response enabled the rapid and efficient separation of the catalyst from the solution within seconds, as illustrated in the inset of Fig. 5b, which captures the catalyst being attracted to an external magnet.

Figure 5.

The characteristics of MNP-0.5: (a) SEM images of MNP-0.5. (b) Room-temperature magnetization curve and the magnetic properties of MNP-0.5; (c) Fe 2p XPS spectrum; and (d) O 1s XPS spectrum.

XPS analysis was conducted to further determine the elemental composition and valence states of MNP-0.5. The survey spectrum (Fig. S4) exhibits characteristic peaks of Fe 2p, O 1s and C 1s at ∼720, 530 and 284 eV, respectively. The elemental iron content was quantified as 20.42 wt.%. The C 1s signal can be attributed to adventitious carbon, which was used as a reference for binding energy calibration (Fig. S5). As shown in the high-resolution Fe 2p spectrum (Fig. 5c), the peaks located at 710.75 and 712.66 eV correspond to Fe(II) 2p_3/2 and Fe(III) 2p_3/2, respectively (Ouyang et al., Reference Ouyang, Zeng, Zhu, Yao, Yang and Long2023). A satellite peak is also observed at 718.91 eV. Peak-fitting analysis revealed that the relative proportions of Fe(II) and Fe(III) are 39.80% and 60.20%, respectively. The slightly greater Fe(II) content compared to the theoretical value may be attributed to the excessive addition of ammonia during synthesis, which promoted the formation of a stable iron–ammine complex that helped preserve Fe(II) against oxidation, thereby increasing its relative concentration (Li et al., Reference Li, Xiang, Li, Gu, Dong and Wu2023). The high-resolution O 1s spectrum (Fig. 5d) displays three fitted peaks at binding energies of 530.29, 531.59 and 532.16 eV assigned to lattice oxygen, oxygen defects and adsorbed oxygen, respectively. Their relative proportions are 66.85%, 26.89% and 6.26%. A notable amount of oxygen vacancies and defects can be associated with the use of high-concentration ammonia water during synthesis, which could perturb the crystallization process and promote the formation of defective Fe₃O₄ (Zhu et al., Reference Zhu, Zhou, Fei, Cheng, Zhu, Deng and Ma2022).

Optimization and practical assessment of the MNP/H₂O₂ system

As shown in Fig. 6a, when H₂O₂ was added alone, the removal rate of TC reached only 4.84% after 120 min of reaction, indicating that H₂O₂ exhibited negligible degradation efficiency regarding TC. In the MNP-0.5-only system, the removal rate of TC reached 19.89%, which can be attributed to the inherent adsorption capacity of Fe₃O₄. In contrast, when MNP-0.5 and H₂O₂ were introduced, the TC removal rate reached 100% within 90 min, indicating that MNP-0.5 effectively activates H₂O₂ to degrade TC.

Figure 6.

(a) TC degradation experiments of various systems. (b) Dosage of catalyst and (c) K_obs of MNP-0.5 degradation of TC at various dosages. (d) Dosage of H₂O₂ and (e) K_obs of MNP-0.5 degradation of TC at various H₂O₂ dosages. The oxidation activity of MNP-0.5 is show for: (f) various pH value environments; (g) effects of coexisting ions; (h) actual wastewater experiment; (i) the degradation of various types of antibiotics. Reaction conditions: [catalyst] = 0.1 g L^–1, [H₂O₂] = 50 mM, [contaminants] = 20 mg L^–1, pH = 7 (if not adjusted), T = 298 K.

The influence of parameters such as catalyst dosage and H₂O₂ concentration was further evaluated to achieve the best performance of the catalyst. Figure 6b illustrates the combined adsorption and degradation effects of TC under various MNP-0.5 dosages. The adsorption efficiency progressively increased from 14.03% to 26.55% as the catalyst loading rose from 0.05 to 1.0 g L^–1, which is attributable to the enhanced availability of active adsorption sites at higher nanoparticle concentrations. Notably, the catalytic degradation exhibited distinct characteristics across various dosage ranges. While systems containing ≥0.1 g L^–1 MNP-0.5 achieved complete TC removal within 120 min, the 0.05 g L^–1 system demonstrated inferior performance, with only 83.04% removal efficiency. Kinetic analysis (Fig. 6c) revealed the relationship between catalyst loading and degradation rate. K_obs increased significantly from 0.0103 to 0.0510 min^–1 as the dosage increased from 0.05 to 0.25 g L^–1. However, further increases in catalyst loading beyond 0.25 g L^–1 resulted in a paradoxical decrease of K_obs to 0.0490 min^–1, suggesting nanoparticle aggregation-induced site blocking at elevated concentrations (Ren et al., Reference Ren, Wang, Hu, Guo, Gao and Huang2024). This particle agglomeration phenomenon probably reduces both active surface area and catalytic accessibility. The optimal MNP-0.5 dosage was determined as 0.1 g L^–1 for subsequent experiments based on a comprehensive evaluation of reaction kinetics and economic feasibility. This loading achieved efficient TC degradation (100% removal in 90 min) while maintaining favourable reaction kinetics (K_obs = 0.0495 min^–1) without significant nanoparticle aggregation effects.

The influence of H₂O₂ concentration (10–100 mM) is presented in Fig. 6d. When the H₂O₂ dosage increased from 10 to 50 mM, the TC removal efficiency improved from 89.12% to 100%, which is attributed to the enhanced generation of ROS at elevated H₂O₂ concentrations. Further increasing the H₂O₂ concentration to 100 mM maintained complete TC removal (100%), indicating that a 50 mM H₂O₂ dosage suffices for complete TC elimination. Figure 6e presents the apparent rate constants for TC removal under various H₂O₂ concentrations. The apparent rate constant increased significantly from 0.0044 min^–1 at 10 mM H₂O₂ to 0.0398 min^–1 at 50 mM H₂O₂. However, increasing H₂O₂ to 100 mM resulted in a decreased K_obs, which is attributed to the scavenging effect of excess H₂O₂ on ·OH and the formation of less reactive hydroperoxyl radicals (HO₂·; Ma et al., Reference Ma, Dong, Zheng, Tang, Zhang and Zhao2025). Thus, 50 mM H₂O₂ was optimal.

The practical potential of MNP-0.5 was systematically evaluated. As shown in Fig. S6, the mineralization capability was confirmed by total organic carbon (TOC) analysis, which revealed a removal efficiency of 34.78% after 120 min in the presence of H₂O₂, a significant enhancement compared to the mere 9.38% removal by adsorption alone. Furthermore, as summarized in Table S1, the degradation efficiency of TC by MNP-0.5 surpasses that of most reported Fe₃O₄-based catalysts. However, recycling experiments (Fig. S7) revealed a limitation in its long-term stability, with the degradation efficiency dropping to 42.91% and 10.57% in the second and third cycles, respectively. This notable deactivation is primarily attributed to the passivation of active sites, probably caused by the formation of a surface oxide layer during the magnetic recovery process (Haji et al., Reference Haji, Sabea, Aljawder, Al-Aradi, Al-Khateeb, Elkanzi and Ahmed2021).

The catalytic activity across a wide pH range (1–13) is shown in Fig. 6f. At pH 1, only 31.5% TC removal was achieved after 120 min, primarily due to the rapid oxidation of Fe²⁺ to Fe³⁺ under strongly acidic conditions, which inhibits the Fenton-like reaction (Liu et al., Reference Liu, Tang, Wang, Zhang and Yuan2024a). In contrast, under alkaline conditions (pH 9–13), TC removal efficiencies decreased to 78.36%, 76.97% and 66.67% at pH 9, 11 and 13, respectively. This decline stems from two synergistic effects: (1) the formation of iron hydroxide precipitates via reaction between Fe ions and hydroxyl groups (OH⁻), which reduces active catalytic sites; and (2) accelerated decomposition of H₂O₂ into H₂O and O₂ under alkaline conditions, suppressing the generation of the hydroxyl radicals (⋅OH) critical for degradation (Liu et al., Reference Liu, Long, Liu, Wang, Zhang, Zhang and Liang2024b). Notably, within the pH range of 3–7, TC underwent near-complete removal, which is primarily attributed to its neutral or zwitterionic state, which facilitates adsorption onto the catalyst surface for enhanced surface-mediated reactions while concurrently maintaining Fe₃O₄ stability, which promotes ROS generation.

The influence of common anions (Cl⁻, F⁻, HCO₃⁻ and H₂PO₄⁻; 10 mM) is shown in Fig. 6g. As the results show, all tested anions exhibited varying inhibitory effects on TC removal. H₂PO₄⁻ demonstrated the most pronounced suppression, reducing the TC removal efficiency to 36.72%. This potent inhibition arises from two mechanisms: (1) the high affinity of H₂PO₄⁻ for active surface sites on MNP-0.5, which blocks catalytic activity; and (2) its scavenging effect on hydroxyl radicals via competitive reactions (Equation 3), thereby diminishing oxidative degradation. In systems containing F⁻ and Cl⁻, TC removal efficiencies decreased to 56.77% and 67.33%, respectively. These halogen ions react with ⋅OH to form less reactive halogenated radicals (Equation 4), weakening the overall oxidative capacity. Similarly, the presence of HCO₃⁻ reduced TC removal to 79.63%, as bicarbonate ions (HCO₃⁻) scavenge ⋅OH to generate carbonate radicals HCO₃∙ (Equation 5), which exhibit lower reactivity regarding TC degradation (Yang et al., Reference Yang, Pignatello, Ma and Mitch2016). These findings collectively highlight that the competitive quenching of reactive radicals by inorganic anions constitutes the primary limiting factor in TC degradation efficiency.

(3)\begin{equation}{{\text{H}}_2}{\text{PO}}_4^ - + \cdot{\text{OH}} \to {{\text{H}}_2}{\text{PO}}_4^\cdot + {\text{O}}{{\text{H}}^ - }\end{equation}

(4)\begin{equation}{\text{C}}{{\text{l}}^ - } + \cdot{\text{OH}} \to {\text{HOC}}{{\text{l}}^{\cdot - }}\end{equation}

(5)\begin{equation}\cdot{\text{OH}} + {\text{HCO}}_3^ - \to {\text{O}}{{\text{H}}^ - } + {\text{HCO}}_3^\cdot\end{equation}

The applicability of the MNPs was tested in copper-containing wastewater (CW), lake water (LW) and tap water (TW; Fig. 6h). Within 120 min, TC removal efficiencies reached 100%, 79.53% and 80.32% in CW, LW and TW, respectively, demonstrating robust tolerance of MNP-0.5 to complex aqueous environments. However, slightly reduced removal efficiencies in LW and TW were observed compared to CW. This attenuation probably stems from two factors: (1) competitive adsorption of coexisting anions and organic contaminants in LW and TW, which may occupy active sites on Fe₃O₄ surfaces; and (2) the consumption of ROS by these impurities during the catalytic process (Wang & Wang, Reference Wang and Wang2021). Despite these interferences, the system demonstrated robust catalytic performance, underscoring its potential for real-world wastewater treatment. As shown in Fig. 6i, TC, TC hydrochloride (TCH), norfloxacin (NOR) and sulfamethoxazole (SMX) were used as target pollutants to investigate the selective removal performance of the MNP-0.5/H₂O₂ system regarding various antibiotics. The MNP-0.5/H₂O₂ system achieved 100% removal of TC within 120 min and an 88.37% removal rate for TCH. However, the system exhibited poor catalytic performance regarding NOR and SMX, with removal rates reaching only 23.45% and 12.33%, respectively. The higher degradation efficiency of TC and TCH stems from their stronger adsorption onto the catalyst, which is a crucial step in the heterogeneous Fenton-like process, where ROSs are primarily generated on the surface, ensuring efficient pollutant–oxidant contact (Wang et al., Reference Wang, Yang, Guo, Wang and Zheng2023; Liu et al., Reference Liu, Wang, Yang, Liu and Liu2025a).

Roles and identification of ROS

To further validate the roles of these reactive species and quantify their contributions to TC degradation, quenching experiments were performed using TBA (⋅OH scavenger), p-BQ (⋅O₂⁻ scavenger) and His (¹O₂ scavenger). As depicted in Figs 7a & S8, adding TBA drastically inhibited TC removal, reducing the removal efficiency from 100% to 46.77%, suggesting that ⋅OH serves as the primary reactive species responsible for TC degradation. In contrast, introducing p-BQ or His resulted in comparatively minor suppression, with TC removal rates decreasing to 88.97% and 81.37%, respectively, indicating that ⋅O₂⁻ and ¹O₂ also participate in the degradation process, but to a lesser extent. Furthermore, quantitative analysis of the contribution rate of ⋅OH was conducted based on scavenging kinetics. As illustrated in Fig. 7b,c, ⋅OH accounted for the majority of the degradation activity, with a relative contribution of 88.92%, underscoring its dominant role in the oxidative degradation of TC.

Figure 7.

Reaction mechanism of MNP-0.5/H₂O₂ systems. (a) TC degradation in the presence of various scavengers and (b) K_obs data; (c) ROS contribution rate; (d) EPR tests; (e) Fe 2p of MNP-0.5 after reaction; (f) the change in the valence state of iron before and after the reaction. Reaction conditions: [catalyst] = 0.1 g L^–1, [H₂O₂] = 50 mM, [contaminants] = 20 mg L^–1, [scavenger] = 10mM, pH = 7, T = 298 K.

Furthermore, EPR spectroscopy was employed to detect ROSs. As shown in Fig. 7d, reactive species were detected using 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) as the spin-trapping agent for ⋅OH and ⋅O₂⁻ and 2,2,6,6-tetramethyl-4-piperidone (TEMP) for ¹O₂. Distinct EPR signals corresponding to DMPO-⋅OH, DMPO-⋅O₂⁻ and TEMP-¹O₂ adducts were observed in the MNP-0.5/H₂O₂ system, confirming the generation of ⋅OH, ⋅O₂⁻ and ¹O₂ during the reaction. These findings are consistent with the results obtained from the quenching experiments.

The residual H₂O₂ concentration was monitored over time via the potassium titanium oxalate method (Fig. S9). Within 120 min, 47.35 mM of H₂O₂ was consumed, corresponding to a utilization efficiency of 94.71%. These results confirm the outstanding catalytic performance of MNP-0.5, which facilitates rapid and efficient generation of ROS. To further investigate the catalyst’s evolution, XPS analysis of the used MNP-0.5 after the reaction was conducted. The high-resolution Fe 2p spectrum revealed a significant shift in the oxidation state of the catalyst after reaction. Compared to the fresh catalyst, the Fe²⁺ content decreased from 39.80% to 30.16%, whereas the Fe³⁺ content increased from 60.20% to 69.84%, as illustrated in Fig. 7e,f. This change can be attributed to the oxidation of surface Fe²⁺ during the activation of H₂O₂ (Equation 6), wherein Fe²⁺ acts as the primary active site. The limited electron transfer from bulk Fe²⁺ to the surface further contributed to the accumulation of Fe³⁺, underscoring the surface-dominated redox mechanism within the MNP-0.5/H₂O₂ system (Dong et al., Reference Dong, Du, Dong, Che, Kong and Cheng2022).

Mechanistic analysis reveals that ⋅OH is predominantly generated via Fe²⁺-activated H₂O₂ decomposition (Equation 6). Subsequent reactions between residual H₂O₂ and ⋅OH yield hydroperoxyl radicals (⋅HO₂; Equation 7). Additionally, Fe³⁺ reacts with H₂O₂ to produce ⋅HO₂ (Equation 8), which decomposes into ⋅O₂⁻. The formation of ¹O₂ is primarily attributed to the dismutation (or self-reaction) of ⋅O₂⁻, as represented in Equation 9 (Xie et al., Reference Xie, He, Pei, Dong, Yang and Xiong2023). Collectively, the degradation of TC in the MNP-0.5/H₂O₂ system involves synergistic contributions from ⋅OH (dominant), ⋅O₂⁻ and ¹O₂. These findings elucidate the multi-radical pathways underpinning the oxidative degradation mechanism.

(6)\begin{equation}{\text{F}}{{\text{e}}^{2 + }} + {{\text{H}}_2}{{\text{O}}_2} \to {\text{F}}{{\text{e}}^{3 + }} + \cdot{\text{OH}}\end{equation}

(7)\begin{equation}\cdot{\text{OH}} + {{\text{H}}_2}{{\text{O}}_2} \to \cdot{\text{H}}{{\text{O}}_2} + {{\text{H}}_2}{\text{O}}\end{equation}

(8)\begin{equation}{\text{F}}{{\text{e}}^{3 + }} + {{\text{H}}_2}{{\text{O}}_2} \to {\text{F}}{{\text{e}}^{2 + }} + \cdot{\text{H}}{{\text{O}}_2} + {{\text{H}}^ + }\end{equation}

(9)\begin{equation}2\cdot{\text{O}}_2^ - + 2{{\text{H}}^ + } \to {}_{ }^1{{\text{O}}_2} + {{\text{H}}_2}{{\text{O}}_2}\end{equation}

Degradation mechanism in the MNP-0.5/H₂O₂ system

Based on the aforementioned results, a degradation mechanism is summarized in Fig. 8. The process begins with the adsorption of TC molecules onto the Fe₃O₄ surface. The adsorbed TC is then rapidly degraded primarily by ·OH, generated from the activation of H₂O₂ by surface-bound Fe²⁺. Concurrently, residual H₂O₂ reacts with ⋅OH to form ⋅O₂⁻ (Equation 7), which subsequently undergoes self-reaction to ¹O₂ (Equation 9). Collectively, these ROS facilitate the ring-opening reactions of TC and its intermediates, ultimately leading to their mineralization into CO₂ and H₂O.

Figure 8.

Mechanistic diagram of MNP-0.5’s adsorption–catalytic degradation of TC.