Poly(catechol) modified Fe3O4 magnetic nanocomposites with continuous high Fenton activity for organic degradation at neutral pH

Fe3O4 magnetic nanoparticles (MNPs) have been widely used as a recyclable catalyst in Fenton reaction for organic degradation. However, the pristine MNPs suffer from the drawbacks of iron leaching in acidic conditions as well as the decreasing catalytic activity of organic degradation at a pH higher than 3.0. To solve the problems, Fe3O4 MNPs were modified by poly(catechol) (Fe3O4/PCC MNPs) using a facile chemical co-precipitation method. The poly(catechol) modification improved both the dispersity and the surface negative charges of Fe3O4/PCC MNPs, which are beneficial to the catalytic activity of MNPs for organic degradation. Moreover, the poly(catechol) modification enhanced the efficiency of Fe(II) regeneration during Fenton reaction due to the acceleration of Fe(III) reduction by the phenolic/quinonoid redox pair. As a result, the Fenton reaction with Fe3O4/PCC MNPs could efficiently degrade organic molecules, exampled by methylene blue (MB), in an expanded pH range between 3.0 and 10.0. In addition, Fe3O4/PCC MNPs could be reused up to 8 cycles for the MB degradation with negligible iron leaching of lower than 1.5 mg L−1. This study demonstrated Fe3O4/PCC MNPs are a promising heterogeneous Fenton catalysts for organic degradation.


Introduction
Heterogeneous Fenton technique as a promising technique for advanced oxidation processes (AOPs) has been intensively applied for the removal of organic pollutants because of its intrinsic advantages over classical homogeneous Fenton reactions, including the wide working pH range, no iron sludge pollution, reusable catalysts, and low H 2 O 2 consumption (Chen et al. 2017;Goncalves et al. 2020;Luo et al. 2010). Many solid catalysts have been demonstrated to be effective in heterogeneous Fenton reactions (Li et al. 2018;Li et al. 2017). Fe 3 O 4 magnetic nanoparticles (Fe 3 O 4 MNPs) have received great attention in Fenton reactions (Mondal et al. 2020). The Fe 3 O 4 MNPs possess unique properties, magnetism, which can be effectively separated from the solution using a magnetic field (Mercado et al. 2018a;Magnacca et al. 2014). Especially, Gao et al. demonstrated that Fe 3 O 4 magnetite nanoparticles possessed an intrinsic enzyme mimetic activity similar to that of peroxidases in nature, which were widely used to oxidize organic pollutant in the treatment of wastewater. Further, Fe 3 O 4 MNPs possessed highly effective catalytic activity, which possessed a higher binding affinity for the substrate TMB than HRP and exhibited a 40-fold higher level of activity at the same molar catalyst concentration than that of HRP (Gao et al. 2007). Moreover, the Fe(II) in Fe 3 O 4 MNPs plays an important role as the electron donor to initiate the Fenton reactions. The octahedral sites in the magnetite Responsible Editor: Ricardo Torres-Palma * Chuan Wang wangchuan@gzhu.edu.cn 1 structure can accommodate both Fe(II) and Fe(III) simultaneously. Hence, the reversible transformation of Fe(II)/ Fe(III) can happen within the same structure during Fenton reaction (Wang et al. 2020;Wang et al. 2010), that is to say, Fe(III) transforms to Fe (II) (Cai et al. 2021;Xiang et al. 2021). However, the iron leaching from the Fe 3 O 4 MNPs at acidic pH and the low catalytic activity at neutral and alkaline pH could affect the stability and activity of Fe 3 O 4 MNPs, thus limiting their applications . Previous studies have shown, that, for the degradation of 2,4-dichlorophenol (2,4-DCP) with Fe 3 O 4 MNPs, 9.8 mg L −1 of Fe dissolved into the solution at pH 3.0 after 180 min. Slow degradation of 2,4-DCP was observed at pH values of 3.9, 4.6, and 5.0, although a lower pH caused a shorter induction time and a higher kinetic rate (Xu and Wang 2012). Dimethyl phthalate (DMP) was quickly removed in the acidic solution with pH range from 3.0 to 4.5. However, at pH 8.0, the Fe 3 O 4 MNPs had very little activity as Fenton catalysts (He et al. 2015). The low catalytic activity at high pH values is mainly caused by the oxidation of these fine particles with limited Fe(II) regeneration during heterogeneous Fenton processes. Hu et al. reported that the Fe(II) amount in the total surface Fe atoms of Fe 3 O 4 / MWCNTs for used and fresh catalysts was 13.3% and 31.7%, respectively (Hu et al. 2011). Many efforts have been paid to improving the catalytic performance of Fe 3 O 4 MNPs (Hammouda et al. 2015). Niu et al. introduced humic acid (HA) to coat Fe 3 O 4 magnetic nanoparticles, i.e., Fe 3 O 4 /HA, which exhibited a high catalytic ability for H 2 O 2 decomposition. This was caused by rapid electron transfer among the complexed Fe(II)-HA and Fe(III)-HA, leading to the rapid regeneration of Fe(II) species and the fast production of •OH radicals (Niu et al. 2011).
Organic ligands (citrate, oxalate, ethylenediaminetetraacetic acid, etc.) have been used to modify the surface of iron compounds to control their solubility in Fenton or Fenton-like processes (Bai et al. 2013;Baldi Marchetto et al. 2010;Jho et al. 2012). Among those iron ligands, catechol, could form strong coupling with the iron ions which reduced the Fe 3+ /Fe 2+ redox potential (Niu et al. 2011;Kang and Choi 2009). In addition, catechol can directly reduce Fe(III) to Fe(II), which itself transformed to the corresponding quinones (Melin et al. 2015). Therefore, introducing catechol in Fenton or Fenton-like processes can widen the reaction pH to neutral conditions by preventing iron from precipitation and, at the same time, enhance the electron transfer (Chen et al. 2017;Contreras et al. 2009). However, it also contributes a certain amount of total organic carbon (TOC) that consumes a certain amount of H 2 O 2 in a Fenton system. Finally, catechol will be degraded and exhausted, which will result in the degradation of the catalytic performance (Xiao et al. 2016).
The polymerization of catechol, catalyzed by Fe(III), forms poly(catechol) (Elhabiri et al. 2007;Gulley Stahl et al. 2010;Slikboer et al. 2015). Poly(catechol), containing phenolic/ quinonoid redox-active units in the main chain, is a redoxactive polymer. Moreover, the binding of the bidentate enediol ligands from the catechol group converted the under-coordinated iron on the surface sites back to a bulklike lattice structure with an octahedral geometry for the oxygen-coordinated iron, which consequently ended up with tight binding between ligands and iron oxide (Xu et al. 2004). Poly(catechol) has been exploited as adhesives and coatings on the surfaces of organic and inorganic materials due to its unique thermal, structural properties, and the ability to form strong charge transfer complexes with the metal oxides (Faure et al. 2013;Ye et al. 2011). Therefore, we propose that Fe 3 O 4 MNPs modified with poly(catechol) will greatly increase the catalytic activity of Fe 3 O 4 MNPs without sacrifice the structural stability. The introduction of poly(catechol) in Fe 3 O 4 / PCC MNPs could offer the following advantages: i) preventing nanoparticles from agglomeration and broadening the working pH range of Fenton reactions; ii) avoiding Fe(II) oxidation and iron leaching; iii) accelerating the Fe(III)/Fe(II) conversion with the recyclable organic ligands.
In this work, Fe 3 O 4 /PCC MNPs were prepared by a facile co-precipitation method. The obtained material was tested for heterogeneous Fenton degradation of MB, used as a model organic pollutant. The structure, surface charge, electron transfer ability, and catalytic activity of Fe 3 O 4 /PCC MNPs were investigated. The possible mechanisms for Fe 3 O 4 /PCC MNPs formation and enhanced MB degradation were also proposed.

Preparation and characterization of catalysts
Fe 3 O 4 /PCC MNPs were prepared by a facile chemical coprecipitation method using iron salts and catechol as precursors . In brief, FeCl 3 ·6H 2 O (10 mmol) and FeSO 4 ·7H 2 O (5 mmol) were dissolved into 75 mL deionized water, before adding 75 mL of catechol aqueous solution (1.5 mM). The mixture was also used as the poly(catechol) precursor. The solution was standing for 30 min before adding into 100 mL of ammonium hydroxide (3.3 M) rapidly. The solution was aged for 120 min under vigorous stirring. The whole synthesis processes were performed in an ambient atmosphere. The black magnetic nanoparticles were separated by an external magnet and were washed with deionized water until pH neutral. The collected nanoparticle powder was dried in a vacuum oven at 50°C for 24 h to obtain Fe 3 O 4 /PCC MNPs. The pristine Fe 3 O 4 MNPs without PCC were synthesized following the same procedure without adding catechol. The poly(catechol)-Fe was obtained following the same procedure without adding ammonium hydroxide. All the products were stored in a desiccator under ambient temperature for further experiments.
The morphology and size distribution of Fe 3 O 4 /PCC and Fe 3 O 4 MNPs were obtained from a transmission electron microscope (TEM, Tecnai™ G2 Spirit, FEI, USA). The phase structures of the Fe 3 O 4 /PCC and Fe 3 O 4 MNPs were determined by X-ray diffraction (XRD, Empyrean, Netherlands) and Raman spectroscopy (Renishaw, inVia Qontor, Germany). To verify the formation of poly(catechol), the surface chemistry of catechol, poly(catechol), Fe 3 O 4 , and Fe 3 O 4 /PCC MNPs was analyzed using Fourier transform infrared (FTIR) spectrometer (Cary 630, Agilent). The FTIR samples were prepared in pressed KBr pellets. The interaction between Fe 3 O 4 and poly(catechol) was analyzed using X-ray photoelectron spectroscopy (XPS, ESCALAB 250, ThermoVG Scientific, USA) with Al Kα (1486.6 eV) as the X-ray source. All XPS spectra were corrected using the C 1s line at 284.6 eV. The thermal stability of Fe 3 O 4 and Fe 3 O 4 /PCC was performed by a thermogravimetric analyzer (TGA, TGA/DSC 1, Mettler-Toledo, Switzerland). The tests were performed at heating rate of 5°C min −1 from room temperature to 1000°C under the nitrogen flow, and the weight retention-temperature curves were recorded. The zeta potentials of the catalyst suspensions at different pH values were determined by an analyzer (Zetasizer, Malvern 3000). The electron transfer ability of the catalysts was examined by cyclic voltammetry measurements using an electrochemical workstation (CHI660, CH Instruments, Chenhua, Shanghai, China) in a cell with a three-electrode configuration. Glassy carbon electrodes deposited with Fe 3 O 4 or Fe 3 O 4 /PCC MNPs respectively were used as working electrodes. A Pt foil and a saturated calomel electrode (SCE) were used as the anode and the reference electrode, respectively. The measurements were carried out in the potential range of 0 to 1.0 V versus Hg/HgO at a scan rate of 10 mVs −1 .

Degradation procedures
The degradation procedures were carried out in a beaker (100 mL) shaken at a speed of 180 rpm. In a typical reaction, MB solution (50 mL) with a certain concentration was prepared with the addition of the specified amount of Fe 3 O 4 /PCC MNPs. The pH of the reaction solution was adjusted to a required value using H 2 SO 4 (1.0 M) or NaOH (1.0 M) solution. The degradation reactions were initiated by adding H 2 O 2 to the suspension once the adsorption equilibrium was achieved. At different time intervals, a suspension sample (0.6 mL) was collected, and the reaction was immediately quenched with the solution of pure methanol (30 μL). The solid samples were separated from the solution using an external magnet. The supernatant liquid was collected for analysis. To test the stability of Fe 3 O 4 /PCC composites, the catalyst was gathered via an external magnet, followed by washing, and drying under vacuum, and then reused in a fresh solution of MB and H 2 O 2 several times. Each experiment was run in triplicate, and average values and standard deviations are presented.

Analytical methods
The MB concentration was measured by UV-vis spectroscopy at the maximum absorption wavelength of MB (660 nm) (Anelise L. et al. 2012;Banerjee S. et al. 2018;Li K.Y. et al. 2017). The MB mineralization was evaluated by a total organic carbon (TOC) analyzer (MultiN/C3100TOC/TN, Analyticjena, AG ). The total carbon content on the catalyst was measured by the elemental analyzer (Vario EL cube). The total leached iron was measured using the orthophenantroline complexometric method (λ = 510 nm). The formation of •O 2 − was determined by a fluorescence method using 4-chloro-7nitrobenz-2-oxa-1,3-diazole (NBD-Cl) as a radical scavenger. The concentration of the reaction product between •O 2 − and NBD-Cl was measured with a fluorescence spectrometer (Hitachi F-4600, Hitachi, Japan) at the emission of 550 nm with the excitation of 470 nm. •OH radicals were identified by electron spin resonance spectroscopy (ESR, ESP 300E, Bruker), operating at the center field strength of 3514 G with 9.85 GHz microwave frequency. For electron spin resonance (ESR) assay, samples (0.5 mL ) were taken instantaneously after 20 min into the heterogeneous Fenton reaction. The samples were mixed with 50 μL of DMPO (500 mM) to form the DMPO-OH adduct. The ESR spectra of the reaction filtrate were recorded.  (Kong et al. 2002), which indicated the polymerization of catechol, and catechol retained only one hydroxyl group during the polymerization processes, being consistent with the reports in the literature (Dubey et al. 1998a, b). The peak at 1280 cm −1 from the Fe 3 O 4 /PCC MNPs and poly(catechol) is attributed to the asymmetrical stretching of C-O-C arom and C-OH Dubey et al. (1998a, b). The peaks observed at 1400 cm −1 , 1460 cm −1 , can be assigned to the stretch of C-C, C=C in the aromatic ring (Aktas et al. 2003), which demonstrates that organic ligand indeed existed in the magnetic nanocomposites.   (Mercado et al. 2018a;Mercado et al. 2018b;Mercado et al. 2014;Mercado and Weiss 2018). The Fe 3 O 4 MNPs in the composites are highly pure Fe 3 O 4 phase with an inverse spinel structure since no XRD peaks from other magnetite were detected (Hu et al. 2011). Moreover, the structural characteristics of Fe 3 O 4 , Fe 2 O 3 , and Fe 3 O 4 /PCC MNPs were characterized by Raman spectroscopy (Fig. S1a). As shown as Fig. S1a, the Raman spectroscopy of Fe 3 O 4 and Fe 2 O 3 were very similar, and it was possible to distinguish hematite domains might be present. Nevertheless, the reason might be the oxidation of Fe(II) to Fe(III) during the Raman test, according with the published study that in situ laser oxidation might take place (Jubb and Allen 2010). From the hysteresis loop of Fe 3 O 4 and Fe 3 O 4 / PCC MNPs (Fig. S1b), there were no coercive force and residual magnetism in Fe 3 O 4 and Fe 3 O 4 /PCC MNPs, indicating the magnetic separation characteristic of Fe 3 O 4 /PCC MNPs (Mercado et al. 2018a;Magnacca et al. 2014).

Results and discussion
To characterize the morphology and size distribution of Fe 3 O 4 /PCC and Fe 3 O 4 MNPs, TEM images were obtained from a transmission electron microscope. The TEM and HRTEM images of Fe 3 O 4 and Fe 3 O 4 /PCC MNPs were illustrated in Fig. 1c, e, d, and f, respectively. As shown as Fig. 1c, d, both Fe 3 O 4 and Fe 3 O 4 /PCC MNPs were in a quasi-spherical shape, and the average particle size of Fe 3 O 4 and Fe 3 O 4 /PCC MNPs was 8.06 nm and 6.32 nm, respectively. According to the formation processes of Fe 3 O 4 /PCC MNPs, there were a large number of phenolic hydroxyl groups distributed around Fe 3 O 4 nanoparticles due to the modification of poly(catechol), and the ionization of phenolic hydroxyl groups made the composites   Fig. 1e, f, the measured d-spacings equal to 0.251 nm and 0.209 nm were assigned to the lattice spacing of the (311) and (400) plane of Fe 3 O 4 and Fe 3 O 4 / PCC, respectively, which was accordance with the results of XRD patterns. Figure 2a shows the XPS spectra of the MNPs. The dominant peaks at the binding energies of~285, 530, and 711 eV are ascribed to the C 1s, O 1s, and Fe 2p, respectively. Figure 2b shows the high resolution Fe 2p spectra of Fe 3 O 4 and Fe 3 O 4 /PCC MNPs. The peaks from the Fe 2p1/2 and Fe 2p3/2 are located at 710.8 and 724.5 eV, respectively. The results are consistent with the literature data for magnetite, which confirmed the successful formation of Fe 3 O 4 (He et al. 2010). Moreover, 32.6% of the total surface iron atoms were in the Fe(II) state in Fe 3 O 4 /PCC MNPs (Fig. S2a). This is much higher than that in the pristine Fe 3 O 4 MNPs (25%, Fig. S2b), which was due to the complexation and reduction of irons by the poly (catechol) (Zubir et al. 2014). The deconvolution of the C 1s spectra (Fig. 2d)

Comparison of catalytic activity
The catalytic activity of both Fe 3 O 4 and Fe 3 O 4 /PCC MNPs (1.0 g L −1 ) was evaluated based on the degradation of MB (0.1 mM) with H 2 O 2 (40.0 mM) at an initial pH of 6.0 (Fig. 4a) In the Fe 3 O 4 /PCC MNPs catalyzed MB degradation process, MB was first enriched at the vicinity of the catalyst surface through the electrostatic adsorption, which was due to the facts that the poly(catechol) modification increased the surface negative charges of Fe 3 O 4 /PCC MNPs (Fig. 3a) and improved the dispersity of the nanoparticles (Fig. 1d). The adsorbed MB could be much easier to be oxidized by the hydroxyl radicals generated on the surface of Fe 3 O 4 /PCC MNPs since the organic pollutant adsorbed on the surface of catalysts is much more reactive (Xue et al. 2009;Noorjahan et al. 2005;Gu et al. 2013). The adsorption of MB on the Fe 3 O 4 /PCC MNPs was quantitatively studied and shown in Figs. S3 and S4 (in supporting information: S1.1 and S1.2). In contrast, methyl orange (MO), an anion dye, was neither adsorbed on the Fe 3 O 4 /PCC MNPs nor degraded after the addition of H 2 O 2 shown in Fig. S5 (in supporting information: S2). This confirmed that the degradation of MB was initiated by the adsorption of MB on Fe 3 O 4 /PCC MNPs. A similar effect was found by Gu et al. that a magnetic porous carbon derived from the sludge showed a highly active property as an efficient heterogeneous catalyst to adsorb and degrade naphthalene dye (1,2,4-Acid) in an aqueous solution through a Fentonlike reaction (Gu et al. 2013).
In addition, poly(catechol), coordinated with Fe(III) on the surface of Fe 3 O 4 /PCC MNPs, could facilitate the Fe(II)/Fe(III) conversion by the phenolic/quinonoid redox pair during the Fenton reaction (Feng et al. 2016). To better evaluate the catalytic activity of the catalyst, cyclic voltammetry measurements were adopted, which were shown in Fig. 4b. The cathodic and anodic peaks can be attributed to the reduction of Fe(III)/Fe(II) and the oxidation of Fe(II)/Fe(III), respectively. In comparison with Fe 3 O 4 MNPs, both the cathodic and the anodic peaks of Fe 3 O 4 /PCC MNPs shifted towards smaller potentials, which indicated that the Fe(III)/Fe(II) redox reactions became more easily with the presence of PCC. The peak potential difference between the cathodic and anodic peaks of Fe 3 O 4 /PCC MNPs (92 mV) was also smaller than that of Fe 3 O 4 MNPs (280 mV), which confirmed the acceleration of the Fe(III)/Fe(II) redox cycling with the introduction of poly(catechol). The fast Fe(II) regeneration can greatly promote the degradation of MB in the heterogeneous Fenton system (Ma et al. 2015).  Based on the first-order kinetics, the degradation rate constant, k, slightly decreased with the increase of pH. This was quite different from some Fe 3 O 4 -based degradation system, in which the removal efficiency decreased sharply as the solution pH increased (Zhang et al. 2014a, b;Wang et al. 2012). On one hand, MB was more easily enriched at the vicinity of the catalyst surface with the increase of pH maintaining fast oxidation. On the other hand, poly(catechol) enhanced the Fe(II) regeneration in the Fenton reaction with Fe 3 O 4 /PCC MNPs due to the acceleration of the Fe(II)/ Fe(III) recycling (Fig. 4b). Hence, the Fe 3 O 4 /PCC MNPs became less sensitive to the solution pH.

Effect of initial pH on the catalytic activity of Fe 3 O 4 /PCC MNPs
The amount of iron leaching from the catalyst was as low as 1.02 mg L −1 , 0.62 mg L −1 , 0.37 mg L −1 , and 0.49 mg L −1 , at the pH of 3.0, 6.0, 9.0, and 10.0, respectively. The possible reason for the negligible iron leaching was that the phenolic hydroxyl groups at the surface of the Fe 3 O 4 /PCC MNPs were hydrolyzed to form a thin layer of H + , which could prevent the diffusion of H + into the Fe 3 O 4 . Thus, Fe 3 O 4 /PCC MNPs were stable even at acidic conditions.

Mineralization of MB and •OH radical detection
Under our reaction conditions, MB concentration was decreased monotonically in the Fenton-like degradation process within 120 min (see in supporting information: S3, Fig. S6). To evaluate the mineralization of MB, the total carbon concentration (TOC) was measured. The total organic content was increased by 0.24 mg after the first circle of Fenton-like degradation with Fe 3 O 4 /PCC MNPs. This value was increased to 1.57 mg after eight reaction cycles. In each reaction cycle, 0.1 mM 50 mL of MB (0.995 mg carbon) was added to the system. The net TOC removal was obtained by calculating the TOC difference coming from both the solid catalyst and liquid solution before and after the degradation reactions. Hence, during the first reaction cycle, 48.3% of TOC was mineralized using 1.0 g L −1 Fe 3 O 4 /PCC MNPs and 40 mM H 2 O 2 at pH 6.0 and 30°C within 120 min. After eight cycles, TOC removal efficiency was maintained at 40.7% (Table 1).
To identify the reactive oxygen species (ROS) responsible for the mineralization of MB, the concentrations of •OH and •O 2 − were detected by ESR and fluorescence method, respectively, during the MB degradation (Ma et al. 2015;Zhang et al. 2014a, b). DMPO and NBD-Cl were used as the probes for the determination of •OH and •O 2 − respectively. As shown in Fig. 6a  of the typical DMPO-•OH adduct with the intensity ratio of 1:2:2:1. There was no significant decrease in the ESR signal after the sixth repeated run of MB degradation with Fe 3 O 4 / PCC MNPs (Fe 3 O 4 /PCC-6th MNPs). However, the intensity of •OH in the Fe 3 O 4 /PCC-H 2 O 2 system was much higher than that of the Fe 3 O 4 -H 2 O 2 system (Fig. 6b). The results indicated that the addition of poly(catechol) in Fe 3 O 4 /PCC MNPs enhanced the transformation of H 2 O 2 into •OH radicals. It confirmed that •OH radicals were the main active species involved in the process of MB degradation (Wang et al. 2020;Wang et al. 2010), which is also responsible for the stable performance of the Fe 3 O 4 /PCC MNP catalyst.

Stability of Fe 3 O 4 /PCC MNPs
Stability is an important property for an effective catalyst. To evaluate the stability of the catalyst, eight reaction cycles for MB removal with Fe 3 O 4 /PCC MNPs (1.0 g L −1 ) were continuously performed using 40.0 mM H 2 O 2 at pH 6.0 and 30°C. Each cycle lasted 120 min. After each reaction, Fe 3 O 4 / PCC MNPs were separated from the reaction solution with a powerful magnet and rinsed with distilled water three times before reused in the next Fenton processes. During the eight times recycling, MB was almost completely removed with negligible iron leaching (< 1.5 mg L −1 ) (Fig. 7). This is consistent with the fact that there was no significant decrease in the concentration of •OH radicals produced after six cycles of running (Fig. 6b), which was due to the enhanced Fe(II) regeneration. The results demonstrated the good stability and high catalytic activity of the Fe 3 O 4 /PCC MNPs.
Since the reduction of Fe(III) to Fe(II) is crucial for the continuous generation of •OH, the amount of surface Fe(II) on the catalyst after the reaction was investigated using XPS analysis shown in Fig. 8a. The survey spectrum from the used Fe 3 O 4 /PCC MNPs exhibited peaks at the binding energies of 285, 530, and 711 eV, which were ascribed to C 1s, O 1s, and   Figure 8b shows the details of the Fe 2p peaks (Fe 2p 1/2 and Fe 2p 3/2 ) of used Fe 3 O 4 /PCC MNPs. The Fe 2p 3/2 peak was deconvoluted into three components at 713.42 eV, 711.30 eV, and 710.03 eV, which were assigned to the Fe(III) tct , Fe(II) oct , and Fe(II) oct , respectively. About 34.8% of the total surface iron atoms were in the Fe(II) state in Fe 3 O 4 /PCC MNPs after MB degradation, close to the fresh catalyst (32.6%). This was different from some studies which observed that the amount of Fe(II) declined gradually during the Fenton reaction (Hu et al. 2011). The stable concentration of Fe(II) is associated with the presence of poly(catechol) which accelerated the reduction of Fe(II) driven by the phenolic/quinonoid conversion. Moreover, the chemical stability of poly(catechol) makes itself resilient against degradation due to the steric hindrance. Hence, it could contribute to the repeated effective catalytic performance of Fe 3 O 4 /PCC MNPs at neutral pH. The stable catalytic efficiency, negligible iron species leaching, as well as convenient recycling of Fe 3 O 4 /PCC MNPs, made it an attractive catalyst for the Fenton reactions.

Possible mechanism for activating H 2 O 2 using Fe 3 O 4 /PCC MNPs
Since •OH radicals were the main active species involved in the degradation of MB, it is essential to figure out whether the •OH radicals generated by the homogeneous catalytic reaction is also contributed. The homogeneous Fenton reactions were performed in 3.0 mg L −1 FeSO 4 ·6H 2 O solution. We chose this concentration since it is higher than the maximal amount of total iron leached from the Fe 3 O 4 /PCC MNP catalyst after the oxidation cycles. Only less than 10% of MB degradation was achieved in the homogeneous Fenton process with 40 mM H 2 O 2 at pH 6.0 and 30°C within 120 min, which indicated the importance of the solid state Fe 3 O 4 /PCC in the heterogeneous decomposition of MB (Dong et al. 1995;Rahhal and Richter 1988).
The mechanism for the MB degradation with activated H 2 O 2 in the heterogeneous Fenton reaction on Fe 3 O 4 /PCC MNPs is proposed in Scheme 2. Firstly, MB was adsorbed on Fe 3 O 4 /PCC MNPs through electrostatic adsorption (Scheme 2a). Then, H 2 O 2 was activated on the surface of Fe 3 O 4 /PCC MNPs to produce •OH radicals for MB degradation (Scheme 2b). The adsorption of MB was beneficial to its degradation as discussed above. The presence of quinones in PCC also improves the degradation of organic compounds in the Fenton processes, due to their role as an electron shuttle (Kang and Choi 2009;Fang et al. 2013). The poly(catechol) in Fe 3 O 4 /PCC MNPs introduced the phenolic/quinonoid redox cycle, which in turn accelerated the Fe(III)/Fe(II) cycle (Chen et al. 2017;Leng et al. 2013). This enhanced Fe(II) regeneration during the Fenton reaction (Fig. 4b) helped to maintain the high catalytic activity of Fe 3 O 4 /PCC MNPs in a wide pH range. The improved Fe(II) regeneration by poly(catechol) was verified by the XPS results (Fig. 8b). could also be quickly reduced to Fe(II) by poly(catechol) and semiquinone radicals (SQ) (Reactions 5 and 6), which were essential for the generation of •OH radicals through Reaction 1 (Leng et al. 2013). As the cycles increased, the quinone analogues were accumulated continuously in the system, forming more Fe(II) and •OH radicals. Hence, the MB degradation was accelerated. Author contribution NH conducted the experiment and experimental analysis, contributed to the discussion of the study, and wrote the original draft of the manuscript. CW supervised the research, conducted the experimental analysis, funding acquisition, and reviewed and edited the manuscript. SW and JX contributed to the discussion of the study and wrote the original draft of the manuscript. Funding The present research is supported by the Natural Science Foundation of China (No. 51678554, No. 51978181).
Data availability All data generated or analyzed during this study are included in this published article and its supplementary information files.

Declarations
Ethics approval Not applicable.
Consent to participate Not applicable.
Consent for publication Not applicable.

Competing interests
The authors declare no competing interests.