Text S1 of the Supporting Information provides a complete list of reagents.

Preparation of Fe/Mg@BC-Mo (Fig. 1a): The wood powder from crushed wood was carbonised under nitrogen at 700°C for 1 hour, then sieved through a 100-mesh screen, washed with 1M HCl to remove ash, and vacuum-filtered with deionised water until neutral pH. The treated biochar powder was dried at 80°C for 3 hours. A mixed solution of FeSO₄·7H₂O and MgSO₄ was prepared in a molar ratio of 2:1 and stirred with the pre-treated biochar for 1 hour for complete mixing, followed by impregnation for 24 hours. The mixture was then vacuum-filtered, washed to a neutral pH, and dried at 80°C for 3 hours. After grinding with MoS₂ in a 1:1 (wt%) ratio until homogeneous, it was pyrolysed under nitrogen at 500°C for 1 hour to obtain the catalyst. Catalysts Fe@BC, Mg@BC, and Fe/Mg@BC were prepared using the same method.

The reaction temperature was maintained at room temperature (25°C), while magnetic stirring was set at 300 rpm. An initial TCH concentration of 20 mg/L was used, along with 5 mM PMS and a catalyst dosage of 0.5 g/L, with a reaction time of 1 hour, to examine the effects of various catalytic systems on pollutant degradation (Text S2). Furthermore, different amounts of MoS₂, catalyst dosage, and simulated anionic concentrations in natural water bodies were varied to evaluate the degradation effectiveness of the catalyst (Text S3). The Fe/Mg@BC-Mo catalyst was collected through vacuum filtration, washed with ultrapure water, and dried for recycling. All experiments were repeated multiple times, with three parallel groups established for each experiment. The results are presented as averages.

Scanning electron microscopy and energy dispersive X-ray spectroscopy (SEM-EDS, S-48DD, Japan) were employed to examine the morphological features of BC, Fe/Mg@BC, and Fe/Mg@BC-Mo, as well as to analyse the elemental composition on the catalyst surface. Brunauer-Emmett-Teller (BET, MicritracBel, Bel sorp-max, USA) was used to assess the catalysts' pore structure and specific surface area. Fourier transform infrared spectroscopy (FTIR, Thermo Fisher Nicolet iS20) was utilised to identify the chemical bonds within the catalysts. The Raman spectrometer (SENTERRA) was employed to evaluate the structural defects in the carbon materials. X-ray diffraction (XRD, Bruker D8 ADVANCE, Germany) was used to determine the catalysts' crystal structure and phase composition within the 2θ range of 10° to 80°. Liquid chromatography-mass spectrometry (Shimadzu LCMS-8040, Japan) was used for quantitative analysis of TCH concentration during degradation and qualitative analysis of the molecular weight and structure of degradation intermediates(Text S4). X-ray photoelectron spectroscopy (XPS, Thermo Fisher-VG Scientific) was employed to assess changes in surface composition and elemental content of the catalyst before and after use. Electron paramagnetic resonance (EPR) spectroscopy was utilised to measure SO₄^·-, ·OH, and ¹O₂ generated in the reaction system. Linearity sweep voltammetry (LSV) and current-time (I-t) curves, conducted with an electrochemical analyser, elucidated the degradation mechanism of TCH in the Fe/Mg@BC-Mo/PMS system. Inductively coupled plasma atomic emission spectroscopy (ICP-AES) was employed to determine the concentrations of metal ions.

The surface morphology of the catalysts BC, Fe/Mg@BC, and Fe/Mg@BC-Mo was examined using scanning electron microscopy (SEM), as shown in Fig. 1. The surface of the biochar BC (Fig. 1b) appeared relatively flat and smooth with uniform and dense pore structures, which resulted from the release of volatile substances during thermal decomposition, leading to the formation of numerous pores in BC. The biochar Fe/Mg@BC (Fig. 1c) displayed a surface covered with a large amount of light-coloured crystalline structures, consistent with the literature[26]. Based on the XRD results, it was inferred that these crystals were compounds formed by the reaction of Fe and Mg ions with functional groups or other components on the biochar surface. With successful loading of metal ions, the surface voids of the catalyst substrate were filled with metal ions or their resulting compounds. Under SEM, the Fe/Mg@BC-Mo catalyst (Fig. 1d) exhibited an evident surface pore structure, characterised by irregular and uniformly attached small particles. This indicated that the addition of MoS₂ promoted a uniform dispersion of metal oxides and increased microporosity, which could provide more active sites for PMS activation. EDS scanning on Fe/Mg@BC-Mo revealed that elements Fe, Mg, and Mo were evenly distributed across the catalyst surface; notably, the Mg content was higher. Correlating this with subsequent XRD findings, it is likely due to the presence of CaMg(CO₃)₂ in the bio-based substrate wood powder[32]-[33].

To clarify the phase structure of the sample, X-ray diffraction (XRD) was used to characterise the sample (Fig. 2a). The XRD pattern of BC showed a large amount of CaMg(CO₃)₂ in the biochar substrate, with sharp diffraction peaks at 2θ = 30.9°, confirming the higher Mg element result obtained from EDS scanning. The XRD patterns of Fe/Mg@BC and Fe/Mg@BC- Mo displayed sharp diffraction peaks at 2θ = 45.2° and 65.1°, indexed as Fe₃C, and a diffraction peak for MgO at 2θ = 36.8°. Diffraction peaks of MgFe₂O₄ appeared at 2θ = 42.7° and 62.0°, indicating the formation of a complex iron-magnesium oxide crystal structure[34]. These results suggest that Fe and Mg were successfully attached to the biochar substrate. However, the intensity of MgFe₂O₄ decreased after Mo doping, implying that the uniform dispersion of MoS₂ hindered metal particle aggregation. Fourier transform infrared spectroscopy (FTIR) was used to further analyse the surface functional groups of the catalyst (Fig. 2b). The absorption peak at 3445 cm^-1 was characteristic of hydroxyl (-OH)[35]. Additionally, the peak at 2379 cm^-1 corresponded to the C = O bond[36], while the stretching vibration peaks of carboxyl (-COOH) at 1624 cm^-1 and 1427 cm^-1 indicated that numerous functional groups appeared on the catalyst surface after thermal treatment and acid washing [37]-[38]. These groups could improve pollutant adsorption and offer more active sites for electron transfer, thereby boosting the reaction rate. The abundant hydroxyl, carboxyl, and other functional groups could form stable coordination bonds with Fe and Mg metal ions, enabling the metal ions to be firmly attached to the biochar surface and enhancing the stability, service life, and reusability of the catalyst. Meanwhile, the coordination interactions between these functional groups and metal ions prevented Fe and Mg ions from aggregating during high-temperature pyrolysis, ensuring their uniform dispersion on the catalyst surface. Characteristic peaks at 625 cm^-1 and 450 cm^-1 are attributed to the stretching vibration of Fe-O-Fe or Mg-O bonds and the bending vibration of M-O-C bonds, further confirming that Fe and Mg metal ions were successfully and stably anchored to the surface of the biochar[26],[39].

Figure 2c displays the Raman spectrum of the catalytic material. According to references [3],[40], porous carbon materials exhibit prominent D and G peaks at 1350 cm^-1 and 1580 cm^-1, which relate to the extent of lattice carbon defects and the graphite carbon content of the material[41]. The ratio of the intensities of these two peaks, known as the peak height or area ratio (I_D/I_G), indicates the level of defects in the carbon material. A higher I_D/I_G value signifies a greater presence of defective carbon and a lower degree of graphitisation. The I_D/I_G value of the biochar after loading iron-magnesium oxides and introducing Mo decreased from 3.47 to 2.34, suggesting that the edge sites and functional groups of the original biochar increased during high-temperature pyrolysis[42], resulting in a higher degree of carbon defects. After metal modification, the metal ions reacted with surface functional groups and formed uniform oxide attachments, altering the distribution of active sites and promoting the graphiteification of carbon. The Brunauer-Emmett-Teller (BET) test was used to analyse the pore characteristics and specific surface area of the catalyst. As shown in Fig. 2d, the N₂ adsorption-desorption curves for BC, Fe@BC, Fe/Mg@BC, and Fe/Mg@BC-Mo all display clear IV-type adsorption-desorption isotherms[3],characterisedd by H4 type narrow slit-like lagging rings. The pore size distribution in the catalytic material is broad, consisting of both micropores and mesopores. The micropore content of the catalyst components loaded with metal ions significantly increases compared with BC. Based on Table S1, the BET specific surface area and pore characteristics of BC, Fe@BC, Fe/Mg@BC, and Fe/Mg@BC-Mo show that the specific surface area rises from 124.20 m²·g⁻¹ to 205.91 m²·g⁻¹, and the total pore volume increases from 0.0520 cm³·g⁻¹ to 0.0942 cm³·g⁻¹. Compared to BC, the specific surface area and total pore volume of Fe@BC, Fe/Mg@BC, and Fe/Mg@BC-Mo nearly double. The specific surface area, micropore specific surface area, and pore volume in the Fe/Mg@BC component all demonstrate the best performance. This may be because the metal ions Fe and Mg act more effectively as "porogens" during the high-temperature pyrolysis process, catalyzing the gasification reaction of the biochar's carbon framework to produce more micropores or mesopores[43]. In the Fe/Mg@BC-Mo sample, the high metal loading causes the metal compounds to penetrate or cover the biochar pores during pyrolysis, leading to blockage. Simultaneously, the high loading of MoS₂ and iron-magnesium oxide crystals reduces the specific surface area[44]. In summary, the BET results suggest that the modified Fe/Mg@BC-Mo possesses a large specific surface area and a rich pore structure, which is beneficial for the adsorption of tetracycline hydrochloride salt and provides more active sites for PMS activation.

To investigate whether the incorporation of Mo can enhance the catalytic cycle regeneration performance, Fig. 4a-c show the results of TCH removal over multiple cycles using the catalytic systems Fe/Mg@BC-Mo/PMS, Fe@BC/PMS, and Fe/Mg@BC/PMS. After two cycles with the Fe@BC/PMS system, the TCH removal rates were 59.00% and 40.42%, respectively. After four cycles with the Fe/Mg@BC/PMS system, TCH degradation rates were 83.15%, 76.35%, 65.35%, and 54.77%. The cycling removal results of the Fe/Mg@BC-Mo/PMS system indicated that, with Mo addition, the catalyst could be reused up to nine times, with TCH removal rates within 60 minutes after the 8th and 9th cycles reaching 95.00% and 87.38%, respectively. In summary, adding Mo helps maintain Fe valence states, improves TCH degradation, and prolongs the catalyst’s lifespan. To assess the catalyst’s stability after multiple uses, the catalyst was examined via SEM and XRD after the 8th cycle. Figure 5a-g show that the catalyst's surface did not develop obvious defects or changes after 8 cycles. Although the MgFe₂O₄ oxide on the surface decreased, Fe and Mo remained evenly and stably attached. The XRD spectrum in Fig. 4h reveals that Mo mainly exists as MoS₂, and during the reaction, MgFe₂O₄ and FeC₃ were consumed.

The concentration of metal ions in the solution after 60 minutes of degradation was measured using an inductively coupled plasma atomic emission spectrometer (ICP-AES). The results[52] indicated that the levels of Fe, Mg, and Mo ions gradually decreased as the number of cycles increased(Text S5). To prevent potential secondary pollution from ion leaching, the doping amount of MoS₂ was optimized. As shown in Fig. 3f, reducing the MoS₂ doping ratio to Fe/Mg@BC:MoS₂ = 8:1 (wt%) resulted in a TCH degradation rate of 93.62% within 2 minutes and 97.91% within 60 minutes. This approach maintained high degradation efficiency and addressed environmental concerns related to ion leaching while also lowering catalyst preparation costs.

To examine the degradation mechanism of TCH by the Fe/Mg@BC-Mo/PMS system, this study combined reactive oxygen species (ROSs) quenching experiments and electron paramagnetic resonance (EPR) analysis to identify the main types of ROS generated in the system. As shown in Fig. 4d, L-histidine (L-His), p-benzoquinone (ρ-BQ), tert-butanol (TBA), and methanol (MeOH) were chosen as quenchers for ¹O₂, O₂^·−, ·OH, and SO₄^·−, respectively. Since MeOH can simultaneously quench ·OH and SO₄^·−, the degradation rate dropped from 98.22% to 63.99% within 2 minutes, demonstrating the most substantial quenching effect. Additionally, after 2 minutes, the degradation rates of each group with L-His, ρ-BQ, and TBA were 76.17%, 83.84%, and 77.39%, respectively. Because TBA reacts preferentially with ·OH to combine with MeOH for quenching, this indicates the presence of both ·OH and SO₄^·− active free radicals in the catalytic system, with the higher degradation rate suggesting that more ·OH is generated. Furthermore, the production of O₂^·− in the system is relatively low and may react with other substances during later stages of degradation, leading to its consumption, while ¹O₂ and ·OH are the main contributors to degradation. Further analysis of ROS produced in the Fe/Mg@BC-Mo/PMS system via EPR, as shown in Fig. 4e-f, used 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) to trap ·OH and SO₄^·−, revealing significant signal peaks and confirming that the Fe/Mg@BC-Mo/PMS system generates abundant ·OH and SO₄^·− active free radicals, consistent with the quenching results. When using 2,2,6,6-tetramethylpiperidine-1-oxide (TEMP) as the trap for ¹O₂, the spectral lines displayed three symmetrical (1:1:1) peaks of equal intensity, indicative of typical nitrogen-oxygen free radical triplet peaks[25], signifying that the Fe/Mg@BC-Mo/PMS system produces abundant non-radical active components ¹O₂, aligning with the quenching findings.

The elemental composition and chemical state of the catalyst surface before and after the Fe/Mg@BC-Mo reaction were analyzed using XPS, as shown in Fig. 6a and d. The proportion of C-OH decreased, while the proportion of C = O increased after the reaction, indicating that the active functional group C-O on the catalyst may be oxidized to C = O to activate PMS[53]-[54]. In the Fe 2p XPS spectrum (Fig. 6b and e), the primary peak of Fe 2p_3/2 appeared at 711.43 eV, along with a characteristic peak at 724.86 eV and distinct satellite peaks at 718.56 eV and 732.11 eV[55]. The binding energy data indicated that the catalyst mainly contained Fe³⁺ in the form of Fe³⁺, with an element proportion of 75.04%. After the reaction, the binding energy values of Fe 2p_3/2 and Fe 2p_1/2 peaks were 711.33 eV and 724.61 eV, respectively, and the proportion of Fe³⁺ increased to 79.03%. This suggests that both Fe²⁺ and Fe³⁺ are present in the catalyst system, and Fe²⁺ is oxidised to Fe³⁺ during PMS activation. Figure 6c and f show the Mo 3d XPS spectra, which indicate that Mo⁴⁺ mainly exists as MoS₂ with a small amount of oxide both before and after the reaction. The proportion of Mo⁴⁺ decreased from 90.64% to 88.67%, while Mo⁶⁺ increased from 9.36% to 11.32%. The changes in Fe³⁺ further confirmed that the incorporation of MoS₂ facilitated the transformation of Fe's valence state[56]. The synergistic effect of Fe²⁺/Fe³⁺ and Mo⁴⁺/Mo⁶⁺ efficiently and quickly enabled the activation and degradation of TCH by PMS[57].

Based on the investigation above, the mechanism by which Fe/Mg@BC-Mo activates PMS is inferred as follows (Fig. 7): The Fe and Mg ions form a complex called magnesium ferrite (MgFe₂O₄) and are evenly and stably attached to the biochar substrate. The Fe²⁺ on the catalyst surface is activated through electron transfer to activate PMS, producing ·OH and SO₄^·−, while Fe²⁺ can also react with O₂ to generate O₂^·− (Equations 7–8). The doping of MoS₂ effectively enhances the cycling of iron valence states, and the Mo⁴⁺ on the catalyst surface can supply electrons to Fe³⁺ to facilitate the reduction and regeneration of Fe²⁺ (Eq. 9). According to reports, MoS₂ can activate PMS to generate reactive oxygen species, thereby degrading pollutants (Equations 10–11)[30]. In the system, a large amount of SO₄^·− can quickly react to form ·OH, and ·OH can react with O₂^·− to produce ¹O₂. Additionally, the oxidised Mo⁶⁺ can also generate some ¹O₂ with O₂^·− (Equations 12–14)[31]. Based on the above analysis, MgFe₂O₄ allows Fe to be stably and evenly attached to the bio-based substrate, preventing the agglomeration phenomenon that may occur in many studies when loading nano Fe⁰ and reducing the risk of metal ion leaching[23]. The synergistic effect of Fe²⁺/Fe³⁺ and Mo⁴⁺/Mo⁶⁺ produces abundant reactive oxygen species, enabling the catalytic system to perform efficiently while enhancing the catalyst's cycle life.

Fe²⁺ + HSO₅⁻ → Fe³⁺ +SO₄^·− + ·OH (7)

Fe²⁺ + O₂ → Fe³⁺ + O₂^·− (8)

Fe³⁺ + e⁻ → Fe²⁺ (9)

≡ Mo(IV)S₂ + HSO₅⁻ → ≡ Mo(V)S₂ + SO₄²⁻ + ·OH (10)

≡ Mo(IV)S₂ + HSO₅⁻ → ≡ Mo(V)S₂ + SO₄^·− + OH⁻ (11)

SO₄^·− + OH⁻ → SO₄²⁻ + ·OH (12)

·OH + O₂^·− → ¹O₂ + OH⁻ (13)

Mo⁶⁺ + O₂^·− → ¹O₂ + Mo⁴⁺ (14)

To further clarify the catalytic performance benefits of the Fe/Mg@BC-Mo system, electrochemical tests[58] were used to compare the linear sweep voltammogram (LSV) curves (Fig. 8a) and current-time (I-t) curves (Fig. 8b) of various catalytic materials, including Fe@BC, Fe/Mg@BC, and Fe/Mg@BC-Mo. The electrochemical results showed that the Fe/Mg@BC-Mo catalyst had the strongest current response at different potentials, and the bimetallic loading outperformed the single-metal Fe loading. This suggests that doping with Mo yields the highest electrocatalytic activity, and the rapid increase in charge transfer rate can enhance PMS activation for TCH degradation. Furthermore, the results of ongoing changes over time for various catalytic components under a constant voltage show that, compared to the Fe@BC and Fe/Mg@BC systems, the low current density in the Fe/Mg@BC-Mo system may be due to a lack of active sites on the catalyst surface. The initial current of the Fe/Mg@BC-Mo system is the highest, indicating that the catalyst has excellent reaction kinetics, which is related to the fact that the catalyst surface has more active sites for rapid mass transfer[59]. The current in the Fe/Mg@BC-Mo/PMS system slightly decreases during the first two minutes. Then it remains nearly constant in the plateau region, maintaining a high and stable steady-state current. Generally, the stability and decay rate of the current reflect the electrochemical stability and durability of the material. This further highlights the advantages of the Fe/Mg@BC-Mo catalytic system in degrading TCH under various conditions.

The experiment utilised a high-performance liquid chromatography-mass spectrometry (HPLC-MS) system to identify the intermediate products of TCH degradation within the catalytic system within 60 minutes and propose possible degradation pathways[60]. The mass spectra of TC and its degradation intermediates are shown in Fig. S1. Figure 9 shows four potential degradation pathways based on the mass spectrometry results. Since H⁺ and Cl^- are present in the aqueous solution in ionic form, the complete hydrolysis of tetracycline hydrochloride TCH produces tetracycline TC. Therefore, the mass-to-charge ratio (m/z) of TC before degradation was P0 = 445[61]. During the degradation process, 13 intermediate products were detected, with m/z values of 477, 434, 412, 408, 392, 350, 304, 301, 279, 192, 163, 135, 115. In pathway I, TC first undergoes demethylation, deamidation, and breaks the C = C double bond to undergo hydroxy addition and transform into P1 (m/z = 392). Then, it undergoes deamidation and ring-opening reactions to generate P2 (m/z = 350) and further dehydroxylation to form P3 (m/z = 304). In pathway II, TC first undergoes deamidation and dehydroxylation to produce P4 (m/z = 412), then experiences a series of demethylation, deamidation, -CHO removal, and ring-opening reactions to form P5 (m/z = 301). In degradation pathway III, one -CH₃ group is removed, and a hydroxyl addition reaction occurs to produce P6 (m/z = 434). Then, another methyl group is removed, and deamidation takes place to convert it into P7 (m/z = 408). This is followed by deamidation, dihydroxylation, and ring-opening reactions to form P8 (m/z = 279). Further removal of the ortho-phenolic hydroxyl group leads to P9 (m/z = 192), which is ultimately broken down into CO₂ and H₂O[62] in pathways I and II, along with intermediate products P10 (m/z = 163), P11 (m/z = 135), and P12 (m/z = 115) that undergo further decomposition. Additionally, in pathway IV, TC is attacked by ·OH, leading to hydroxy addition and hydrogen removal reactions that produce -OH and -C = O-, converting P0 into P13 (m/z = 477). This explains the slight increase in the mass-to-charge ratio of some intermediate products during the degradation process.