Peroxymonosulfate Activation by Fe@N Co-doped BC for the Degradation of Sulfamethoxazole: The Key Role of Pyrrolic N

1. Introduction

Sulfamethoxazole (SMX), as a common drug of sulfonamides, can efficiently treat gastroenteritis, coccidiosis, diarrhea, and other infections and diseases [1]. Due to high water solubility, and potential carcinogenicity of SMX, it would have negative impact on the ecosystem [2,3]. Since SMX was challenging to be degraded by conventional treatment processes, it was urgent to explore an efficient, low-cost, and simple technology for eliminating residual SMX in the aquatic environment.

Advanced oxidation processes (AOPs) have received huge attention due to their high efficiency and extensive adaptability [4]. Among AOPs, persulfate-based advanced oxidation processes (PS-AOPs) have gradually emerged due to its unique advantages such as higher oxidation potential and wider applicability under the environments [5,6]. As a result, there are several approaches for peroxydisulfate (PDS) and peroxymonosulfate (PMS) activation to produce reactive oxygen species (ROSs), such as electrolysis, ultrasound, ultraviolet irradiation, heat, and transition metals [7]. Among these activation methods, biochar-based catalysts have the advantages of less pollution and superior thermal stability [6].

Biochar (BC) was produced via thermochemical treatment of biomass under oxygen-limited atmosphere [8]. Biomass feedstock, such as citrus peels, maize stalk, sawdust, corncob, wood chips and shrimp shell, is rich and low-cost [9]. Furthermore, BC is featured with abundant oxygen-containing functional groups (e.g., C-OH, -COOH, C=O) and micropore or mesopore structures [10]. The redox cycle of metal ions (e.g., Fe³⁺/Fe²⁺, Cu²⁺/Cu⁺) could contribute to generating ROSs in the catalytic oxidation processes due to consecutive cleavage of the peroxide bond (O-O) [11]. Nitrogen doping could adjust the chemical property of the graphitic structure, increase the electron cloud density of the carbon atoms around them, thus regulate chemical properties of biochar-based catalysts [12].

Herein, Fe, N co-doped biochar (Fe@N co-doped BC) was prepared by carbonization-pyrolysis method and used as a PMS activator for the elimination of SMX. Characterizations and batch experiments were employed to assess the degradation ability of Fe@N co-doped BC. Then, the active species in the Fe@N co-doped BC/PMS system were identified by quenching tests and electron spin resonance (ESR) tests. The activation mechanism of PMS near different N configurations was uncovered by X-ray-photoelectron-spectroscopy (XPS) analysis and density functional theory (DFT) calculation. Additionally, the reusability, universal applicability, and mineralization capacity were evaluated.

2. Results and Discussion

2.1. Characterization

Morphologies of samples were explored using scanning electron microscopy (SEM). The surface of primary BC is smooth and blocky. After the doping of urea, a layer of bright blocky stack appeared on the surface of N-doped BC (Figure 1a,b). A large number of iron oxide particles are gathered on the surface of Fe-doped BC, while the carbon substrate exhibits a regular blocky structure (Figure 1c). The surface of Fe@N co-doped BC is loaded with tiny iron oxide nanoparticles, and no isolated distribution of nanoparticles is observed outside the carbon matrix, indicating that these nanoparticles have been anchored to the carbon layer (Figure 1d).

Besides, the ultrastructure of crystal lattice of Fe@N co-doped BC was studied in the high-resolution transmission electron microscope (HRTEM) and selected area electron diffraction (SAED) pattern. As shown in Figure 2a, iron oxide particles are successfully grown on the surface of BC by carbonization-pyrolysis method, and iron oxide particles can be used as the active site for activating PMS. The (444) crystal faces belonging to Fe₃O₄ with lattice spacing of 1.21 nm are calculated by Fast Fourier Transform (FFT), as depicted in Figure 2b. The Miller concentric rings also exhibit Fe@N co-doped BC crystal phase structure, and the Miller index of (111) and (311) also testifies to the formation of Fe₃O₄ on the surface of BC (Figure 2c). Energy dispersive spectroscopy (EDS) shows the distribution of five elements, C, Fe, Si, N and O, and also proves that iron and nitrogen have been successfully incorporated into the carbon material (Figure 2d).

XRD spectrum of samples are depicted in Figure 3a. The diffraction peaks of BC appear at about 21.6°, corresponding to the C (111) crystal plane respectively. Interestingly, N-doped BC significantly increases the diffraction peak width at 26.6°, which may be due to the incorporation of urea. It has been reported that the addition of urea can improve the graphitization degree of biochar [12]. Fe@N co-doped BC and Fe-doped BC exhibit distinct diffraction peaks at 30.1°, 35.4°, 43.1°, 53.4°, 56.9°, and 62.5°, corresponding to the crystal plane of (220), (311), (400), (422), (511), and (440), respectively. It was demonstrated that Fe₃O₄ nanoparticles had been successfully loaded onto biochar surfaces (JCPDS 89−2355).

FT-IR spectrum exhibits the surface functional groups (Figure 3b). The peaks at 3446 cm^-1 are attributed to the stretching vibration of -OH groups [14]. The content of -OH groups on the surface of Fe@N co-doped BC is significantly higher than that on Fe-doped BC, N-doped BC and BC. The absorption peaks at 2922 cm^-1 and 2852 cm^-1 are attributed to -CH₂ and -CH₃ groups, respectively [15]. The absorption peaks at 1116 cm^-1 and 1384 cm^-1 are attributed to the vibrations of C-OH and C-C [16]. In addition, the absorption peak at 570 cm^-1 is caused by the tensile vibration of Fe-O, which also indicates that Fe and N co-doping breaks the inert structure of the carbon network and generates a new active center [17].

N₂ adsorption-desorption isotherms (Figure 3c) indicate the specific surface areas (SSAs) of N-doped BC (287.67 m²⋅g^-1) are larger than that of BC (194.86 m²⋅g^-1), Fe@N co-doped BC (269.21 m²⋅g^-1) and Fe-doped BC (216.38 m²⋅g^-1) (Table 1).

As shown in Figure 3d, the appearance of both D-band and G-band of catalysts reveals the co-presence of disordered and crystalline graphite structures [18]. Furthermore, the ratio of I_D/I_G is the key parameter to indicate the defective degree of catalysts [19,20]. The I_D/I_G is 1.07, 0.81, 1.02, and 2.14 for BC, Fe-doped BC, N-doped BC, and Fe@N co-doped BC, respectively. The result demonstrates that Fe@N co-doped BC has obtained abundant defects (vacancy and edge defects) during pyrolysis, which is conducive to catalytic oxidation [10].

2.2. Catalytic Oxidation of SMX

The adsorption efficiency of different catalysts for SMX is shown in Figure 4. Under the conditions of pH = 7.0, dosage of catalyst is 0.4 g/L, and reaction temperature is 25℃, the adsorption rate of BC, Fe-doped BC, N-doped BC, and Fe@N co-doped BC for SMX is 25.8%, 23.2%, 30.0%, and 27.9%, respectively, for 40 mins. It is clear that the adsorption performance of N-doped BC is higher than that of BC, Fe-doped BC and Fe@N co-doped BC, which may be due to the doped urea, which gives a porous structure [21]. It is worth noting that neither Fe-doped BC nor Fe@N co-doped BC has a lower adsorption rate of 30.0% for SMX within 40 mins, which indicates that the iron species loaded on the carbon surface may block part of the adsorption sites, then the adsorption performance of the material is reduced.

Without adding any catalyst, the degradation rate of SMX in sole PMS system is 8.0%, which indicates that the self-decomposition ability of PMS is low (Figure 5a). In the BC/PMS system, the degradation rate of SMX increased to 67.3%, indicating that BC can activate PMS to a certain extent. N and Fe doping alone can partially improve the degradation rate of BC, and the degradation rate of N-doped BC and Fe-doped BC for SMX is 69.0% and 71.0%, respectively. Fe@N co-doped BC has the highest catalytic activity under conditions of pH = 7.0, initial SMX concentration of 10.0 mg/L, Fe@N doped BC of 0.4 g/L, PMS concentration of 0.6 mM, and reaction temperature of 25℃, the degradation rate of SMX can reach 90.2%.

In addition, the pseudo-first-order kinetic model was used to study the degradation process of SMX in different catalytic systems (Eq. 1) (Figure 5b). The linear regression coefficients (R²) were all over 0.90, indicating that the reaction kinetics of degradation of SMX by each catalyst were consistent with pseudo-first-order kinetics. The reaction rate constant (k_obs) of Fe@N co-doped BC/PMS system was 0.041 min^-1, being about 1.50, 1.40, and 1.60 times that of N-doped BC/PMS, Fe-doped BC/PMS, and BC/PMS systems, respectively (Table S1). The typical pollutants degradation rates in different systems were studied in Table S2.

$$\mathrm{I}\mathrm{n}({\mathrm{C}}_{\mathrm{t}}/{\mathrm{C}}_0)=-\mathrm{k}\mathrm{t}$$

(1)

To explore the catalytic performance of Fe@N co-doped BC in depth, the residual PMS concentration was detected by the ABTs colorimetric method. The decomposition rate of PMS in the Fe@N co-doped BC/PMS system was 98.6% within 40 min (Figure 6). It was suggested that the favorable degradation rate of SMX in the Fe@N co-doped BC /PMS system could be due to the rapid decomposition of PMS.

The effects of Fe@N co-doped BC dosage, PMS concentration, initial pH, and SMX concentration on SMX degradation were investigated (Figure 7). In the tested range, an increase of Fe@N co-doped BC dosage and PMS concentration results in an increase in the SMX degradation rate, as an increase in the Fe@N co-doped BC and PMS dosage is conducive to the rapid generation of ROSs in the Fe@N co-doped BC/PMS system [22]. The initial pH of the solution has a significant effect on the SMX removal efficiency. Under neutral conditions (pH = 7.0), the degradation rate of SMX in the Fe@N co-doped BC/PMS system is as high as 90.2%. Under the condition of strong acidity (pH = 3.0), the degradation rate of SMX decreased from 90.2% to 57.6%. Under strong alkaline conditions (pH = 11.0), the degradation rate of SMX decreased from 90.2% to 65.5%. When the initial concentrations of SMX were 0.5 mg/L, 1.0 mg/L, 5.0 mg/L and 10.0 mg/L, the degradation rates of SMX were 98.8%, 98.6%, 95.7% and 90.2%, respectively. In the test range, the initial concentration of pollutants has a slight effect on the degradation rate of SMX.

Co-existing ions can restrain the SMX removal by interacting with reactive species [23]. Typical anions (Cl⁻, H₂PO₄⁻, HCO₃⁻), and humic acid (HA) restrained the SMX elimination to varying degrees (Figure S1). When 5.0 mM and 500.0 mM Cl⁻ were present in the Fe@N co-doped BC/PMS system, SMX removal rate decreased from 90.2% to 79.0% and 70.0%, respectively, within 40 min. The corresponding k_obs decreased from 0.041 min^-1 to 0.025 min^-1 and 0.021 min^-1, which was mainly due to the formation of active chlorine species [24,25]. Compared with Cl⁻, HCO₃⁻ and H₂PO₄⁻ have a more significant inhibitory effect on SMX degradation due to the fact that HCO₃⁻ reacts with •OH and SO₄^•− to form CO₃^•− which has a lower redox potential [26]. Similarly, H₂PO₄⁻ reacts with SO₄^•− to produce less active free radicals, thereby reducing the degradation rate of SMX. In addition, H₂PO₄⁻ easily binds to the active sites on the surface of Fe@N co-doped BC, preventing the catalyst from producing enough ROSs to degrade SMX [27]. When 20.0 mg/L and 50.0 mg/L HA were present in the reaction system, the degradation rates of SMX decreased from 90.2% to 81.0% and 62.0%, respectively, and the corresponding k_obs decreased from 0.041 min^-1 to 0.033 min^-1 and 0.019 min^-1, respectively, within 40 min. On the one hand, HA can consume •OH and SO₄^•−; on the other hand, HA has a strong π-π superposition effect, which is easy to deposit on the surface of Fe@N co-doped BC, thus inhibiting the interaction between Fe@N co-doped BC, PMS and SMX [28].

Figure S2a exhibited the degradation rates of various pharmaceuticals and personal care products (PPCPs) in Fe@N co-doped BC/PMS system. The degradation efficiencies of tetracycline hydrochloride (TC), 4-chloro-3-methyl phenol (CMP), norfloxacin (NOR), and acetaminophen (ACT) were 95.8%, 93.3%, 91.5%, and 94.2% respectively in 40 min. In general, representative PPCPs could be removed efficiently in Fe@N co-doped BC/PMS system. To study the reusability of Fe@N co-doped BC, 5 consecutive degradation tests were conducted (Figure S2(b)). After 5 cycles, the degradation rate of SMX within 40 min decreased from 90.2% to 63.0%. This could be due to the sedimentation of intermediates on the surface of Fe@N co-doped BC, which occupied the reactive sites for activating PMS. The saturation magnetization value of Fe@N co-doped BC was 17.43 emu/g (Figure S2(c)). Fe@N co-doped BC exhibited a strong magnetic response to an external magnetic field and is easily separated from the SMX solution, facilitating the recycling of the catalyst. Figure S2(d) displayed the continuous leaching of Fe species in the redox process. Fortunately, less than 1.04 mg/L of Fe species were leached in reaction solution.

2.3. Mechanism Discussion

2.3.1. Identification of ROSs

The above results indicate that SMX can be degraded rapidly in the Fe@N co-doped BC/PMS system, while previous studies have shown that SMX may be degraded via either a radical pathway (O₂^•−, •OH, or SO₄^•−) or a non-radical pathway (¹O₂ or biochar-PMS* complex) in a heterogeneous catalytic oxidation system [6]. To determine the active species produced in this system and their contribution to SMX degradation, quenching experiments and ESR tests were performed. •OH, •OH/SO₄^•− and O₂^•− were quenched by Tert-butanol (TBA), methanol (MeOH) and p-benzoquinone (p-BQ), respectively, while furfuryl alcohol (FFA) and L-histidine (L-his) were used as trapping agents for ¹O₂ [29,30]. The SMX removal rates decreased from 92.0% to 33.4%, 73.5%, 86.7%, 73.2% and 71.8% within 40 min after adding MeOH (0.5 M), TBA (0.5 M), p-BQ (0.1 M), L-his (0.1 M) and FFA (0.1 M), respectively (Figure 8). These results indicate that •OH and SO₄^•− exist in the system, and SO₄^•− may play a dominant role. Since p-BQ has a slight inhibitory effect on SMX degradation, the system does not produce large amounts of O₂^•−. The contribution of each reactive species was listed in Table S3.

ESR was performed to further prove the presence of ROSs in the Fe@N co-doped BC/PMS system using 2,2,6,6-tetramethyl-4-piperidinol (TEMP) and 5,5-dimethyl-1-pyrroline N-oxide (DMPO) as spin-trapping agents [31,32]. A spectrum with seven main peaks belonging to DMPO-X was observed in the Fe@N co-doped BC/PMS system, indicating that Fe@N co-doped BC activated PMS to produce •OH and SO₄^•− (Figure 9a). The signal intensity in the Fe@N co-doped BC/PMS system was significantly higher than that of the Fe-doped/PMS and N-doped/PMS system, indicating that there is a synergistic effect between Fe and N on the surface of biochar. When Fe@N co-doped BC, PMS and TEMP were added to the solution, a typical triplet signal with the intensity ratio of 1: 1: 1 verified the existence of ¹O₂ (Figure 9b). The self-decomposition of PMS could produce a small amount of ¹O₂ [33]. In summary, •OH, SO₄^•− and ¹O₂ are produced in this system, in which SO₄^•− is the main active species and leads the degradation process of free radical pathway, while ¹O₂ leads the degradation process of non-free radical pathway.

2.3.2. Reaction Mechanism

The above experiments demonstrate the production of •OH, SO₄^•− and ¹O₂ in the Fe@N co-doped BC/PMS system, and on the basis of this, the general catalytic mechanism is proposed. First, •OH and SO₄^•− may be produced by the destruction of the O-O bond of the PMS by free-flowing π electrons on the sp² hybrid carbon (Eqs. (1−3)). As can be seen from the XPS spectrum (Figure 10a), the surface of Fe@N co-doped BC is successfully doped with iron and nitrogen, and the specific atomic percentages of the obtained catalyst are shown in Table S4. Fe@N co-doped BC surface contains C-OH, COOH and C=O groups, which can serve as an active site (Figure 10b). As an electron donor functional group, C-OH is more inclined to activate PMS to produce free radicals, such as •OH and SO₄^•−. As electron withdrawing functional groups, C=O and COOH are more inclined to degrade SMX through non-radical pathway. After the catalytic reaction, Fe (II) content decreased from 43.7% to 40.8%, while Fe (III) content increased from 36.4% to 37.9% (Figure 10c). The doped Fe (II) reacts with HSO₅⁻ to form •OH and SO₄^•− (Eqs. (4−5)), and the generated free radicals can be converted flexibly. Subsequently, Fe (III) reacts with HSO₅⁻ to produce SO₅^•−, and Fe (III) is reduced to Fe (II) and maintains a complete cycle (Eq. 6). These results indicate that Fe₃O₄ plays a leading role in the activation of PMS in the free radical pathway.

It has been reported that carbon-based materials can also activate PMS to degrade organic micropollutants through non-free radical pathways, and the non-free radical mechanism may be related to nitrogen doping [34,35]. N 1s can be divided into three peaks, including pyrrolic N, pyridinic N and graphitic N (Figure 10d). The content of graphitic N and pyridinic N increased from 17.6% and 24.5% to 42.4% and 30.5% respectively, while the content of pyrrolic N decreased from 57.9% to 27.1%. Therefore, pyrrolic N is considered to be the main active species in the redox process.

$$\mathsf{\pi }-\mathrm{e}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{n}\mathrm{s}+\mathrm{H}\mathrm{S}{\mathrm{O}}_5^{-}\to \mathrm{S}{\mathrm{O}}_4^{·-}+\mathrm{O}{\mathrm{H}}^{-}$$

(1)

$$\mathsf{\pi }-\mathrm{e}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{n}\mathrm{s}+\mathrm{H}\mathrm{S}{\mathrm{O}}_5^{-}\to \cdot \mathrm{O}\mathrm{H}+\mathrm{S}{\mathrm{O}}_4^{2-}$$

(2)

$$\mathrm{S}{\mathrm{O}}_4^{·-}+{\mathrm{H}}_2\mathrm{O}\to {\mathrm{H}}^{+}+\mathrm{S}{\mathrm{O}}_4^{2-}+\cdot \mathrm{O}\mathrm{H}$$

(3)

$$\mathrm{F}{\mathrm{e}}^{2+}+\mathrm{H}\mathrm{S}{\mathrm{O}}_5^{-}\to \mathrm{F}{\mathrm{e}}^{3+}+\mathrm{S}{\mathrm{O}}_4^{·-}+\mathrm{O}{\mathrm{H}}^{-}$$

(4)

$$\mathrm{F}{\mathrm{e}}^{2+}+\mathrm{H}\mathrm{S}{\mathrm{O}}_5^{-}\to \mathrm{F}{\mathrm{e}}^{3+}+\mathrm{S}{\mathrm{O}}_4^{2-}+\cdot \mathrm{O}\mathrm{H}$$

(5)

$$\mathrm{F}{\mathrm{e}}^{3+}+\mathrm{H}\mathrm{S}{\mathrm{O}}_5^{-}\to \mathrm{F}{\mathrm{e}}^{2+}+\mathrm{S}{\mathrm{O}}_5^{·-}+{\mathrm{H}}^{+}$$

(6)

The PMS activation performances of three N configurations, pyrrolic N, pyridinic N and graphitic N, were compared, as shown in Figure 11. From the top view, it can be seen that PMS can be stably adsorbed on top of pyrrolic N, pyridinic N and graphitic N. For the reactant structure, the distance (D) between PMS and pyrrolic N is the shortest (2.986 Å). The transition state energy barrier of pyrrolic N/PMS, pyridinic N/PMS and graphitic N/PMS was calculated by B3LYP/6-31G (d,p) method as 23.54 kcal/mol, 27.81 kcal/mol and 26.45 kcal/mol, respectively. The above results showed that pyrrolic N could significantly reduce the activation energy required for the reaction, and the catalytic activity of this active site was the highest.

To investigate the charge transfer rate on the surface of carbon-based materials, electrochemical tests were carried out. Electrochemical impedance spectroscopy (EIS) shows that Fe@N co-doped BC has a smaller semicircle diameter and therefore lower electron migration resistance than that of BC, Fe-doped BC, and N-doped BC (Figure 12). The hydrophilicity of carbon-based materials was further investigated by measuring the contact angle [21] (Figure 13). The contact angle of Fe@N co-doped BC is 4.7°, which is lower than that of BC (14.8°), N-doped BC (9.0°), and Fe-doped BC (9.3°), which means that Fe@N co-doped BC has higher hydrophilicity and is easier to obtain excellent degradation properties.

2.4. Degradation Pathways of SMX

The intermediate products of SMX degradation were investigated by UPLC-TOF/MS (Figure S3, Figure S4 and Table S5). In pathway I, the C atom on the isoxazole ring of SMX undergoes electrophilic addition reaction to produce the product I (m/z = 288) [1]. Subsequently, the S-N bond of SMX breaks and product I is transformed into product II (m/z = 133) and product III (m/z = 190). Under ROSs’s attack, the N-O bond on the isoxazole ring breaks to produce product IV (m/z = 117), which is eventually mineralized into CO₂ and H₂O. At the same time, a ring-opening reaction may occur on the isoxazole ring of product I to form product VI (m/z = 246), and a subsequent dehydrogenation reaction to form intermediate product VII (m/z = 216). In pathway III, SMX undergoes electrophilic substitution to produce a product IX (m/z = 270). Subsequently, the S-N bond of SMX breaks and product IX is transformed into product X (m/z = 190) and product XI (m/z = 99). Under ROSs’s attack, product XI was further oxidized to product XII and product XIII. In summary, the degradation mechanism of SMX in the Fe@N co-doped BC/PMS system is shown in Figure 14.

3. Materials and Methods

3.1. Preparation of Catalysts

Slow pyrolysis, as a typical carbonization method, was commonly used for the conversion of biomass wastes into biochar with high carbon and low volatile matter contents [13]. In this work, Fe@N co-doped BC was fabricated using carbonization-pyrolysis strategy [12]. Rice husk was purchased from a biomass pyrolysis power plant in China. Firstly, biomass was washed, dried in a vacuum drying oven at 100℃ for 12 h, ground, and then passed through a 100-mesh sieve to acquire thin powders for further use. 5.56 g FeSO₄⋅7H₂O, 3.00 g urea and 2.50 g ascorbic acid was dissolved into 100 mL deionized water, after that 3.00 g powders were added into the solution and adequately shaken in thermostatic shaking bath. The mixture was dried in a vacuum drying oven for 24 h, and calcined at a constant calcination temperature (800℃) under N₂ atmosphere (5℃/min. of heating rate) for 2 h. The catalyst was washed with deionized water for several times, and the resulting composites were denoted as Fe@N co-doped BC. For comparison, Fe-doped BC, N-doped BC and BC were synthesized in the same method.

3.2. Reaction Procedures

Batch experiments were carried out in 50 mL vials containing 50 mL of SMX (10 mg/L) to evaluate the catalytic ability of Fe@N co-doped BC on PMS activation. The range of Fe@N co-doped BC dosing was 0.2−0.8 g/L, the concentration of PMS was 0.2−0.8 mM, the pH was 3.0−11.0, SMX concentration was 0.5−10.0 mg/L, and the reaction temperature was set to 25℃. All experiments were carried out at 150 rpm, and 1.0 mL reaction solution was taken periodically at intervals and filtered through a 0.22 μm filter. The excess reaction was terminated by adding 1 mL Na₂S₂O₃ (0.2 M). 0.1 M NaOH or H₂SO₄ was added into the reaction solution to adjust the initial pH values before reaction. All experiments were performed in triplicate.

4. Conclusions

In summary, Fe@N co-doped BC was prepared by carbonization-pyrolysis method, and the catalyst obtained abundant catalytic centers (Fe₃O₄, pyrrolic N) and defect sites (I_D/I_G = 0.97). Under optimal conditions, the degradation rate of SMX in the Fe@N co-doped BC/PMS system is 90.2%. The degradation process conforms to pseudo-first-order kinetics with a k_obs of 0.041 min^-1. In addition, various typical PPCPs could be efficiently degraded in the Fe@N co-doped BC/PMS system within 40 mins. The quenching experiment and ESR test results show that •OH, SO₄^•− and ¹O₂ were produced in the Fe@N co-doped BC/PMS system, in which SO₄^•− is the main active species. SO₄^•− can be produced in the redox cycle between Fe (III) and Fe (II), and Fe₃O₄ acts as an important catalytic center to break the inert carbon network structure and promote the formation of ROSs in the reaction system. Before and after the reaction, the content of pyrrolic N decreased from 57.9% to 27.1%. As the main active site, pyrrolic N showed stronger interaction with PMS and significantly reduced the activation energy required for the reaction (∆G = 23.54 kcal/mol). Therefore, this work first solves the problem of low catalytic activity caused by the lack of catalytic center of carbon materials. In addition, the strong magnetic Fe@N co-doped BC has good recyclability, which improves the practical application and long-term operational potential of the system.