English

• Di-n–butyl phthalate (DBP), a typical kind of phthalate esters, has been considered a priority controlled hazardous contaminant by U.S. EPA. It occurs at low concentrations in aquatic environments and poses considerable risk of endocrine-disrupting effects and metabolic disorders to human beings [1,2]. Moreover, the biodegradation kinetics of DBP are very slow in the natural environment, from dozens of days to more than 20 years [3]. Presently, the typical treatments, including adsorption, membrane separation, biodegradation as well as chemical oxidation, suffer from different drawbacks with regard to their remedy time scale and large energy consumption [4]. Consequently, it is highly demanded to explore a cost-efficient strategy for eliminating trace-level DBP.

  Sulfate radical (SO₄^•−)-based advanced oxidation processes (SR-AOPs) exhibit a favorable standard reduction potential (2.5–3.1 V vs. NHE), wide pH adaptability (2.0–9.0), longer half-life with regard to the hydroxyl radical (30–40 µs vs. 20 ns) and high oxidation selectivity [5]. These make it possible to employ SR-AOPs to decompose a variety of organic substances in sewage. SO₄^•− can be effectively generated by activating peroxymonosulfate (PMS) in conjunction with ultrasound, photolysis, heat, homogenous or heterogeneous catalysts [6]. Cobalt-based catalysts have been investigated, and it has been discovered that they are the most effective activators for cleaving the O—O bond in an asymmetric PMS structure [7]. Cobalt ferrite (CoFe₂O₄) belonging to the spinel-type ferrites family (S.G. Fd-3m) has indicated high activity and magnetic properties with excellent stability in order to degrade persistent pollutants [8]. The coupling effect between Co and Fe species might be more effective as a catalyst in PMS activation than their single component equivalent. Simultaneously, CoFe₂O₄ is characterized to have a high abundance of oxygen vacancies and surface hydroxyl groups, which also help to further activate PMS [9]. Ren et al. evaluated the performance of MFe₂O₄ (M = Co, Cu, Mn, Zn) in PMS activation, and concluded that CoFe₂O₄ showed higher catalytic activity toward PMS for the degradation of DBP, and all catalysts displayed excellent recycling and stability in the repeated batch experiment [10]. Nonetheless, the unprotected CoFe₂O₄ primary nanoparticles are prone to agglomeration, as a result, there are fewer active sites that are accessible and there is less dispersion in the reaction solution, which further reduces the effectiveness of PMS activation and organics decomposition [11]. Therefore, it is necessary to disperse CoFe₂O₄ on supporters for stable and efficient catalytic performance.

  Recently, carbon-based materials were used to supporters for transition metal-based catalysts to fabricate carbon-loaded metal composites. Two-dimensional graphene oxide (GO) is a substance with various excellent properties of unique mechanical strength, high electrical conductivity and large specific surface area [12]. Multiple oxygen-containing functional groups including hydroxyl, carboxyl and epoxy situate at the basal plane and the edges of GO [13]. Metal components can be immobilized by such surface functionalization to create a solid structure and promote the separation and dispersion of spinel ferrite particles [14]. Furthermore, GO consists of the hexagonal ring-based carbon network with both sp²- and sp³-hybridized carbon atoms, the free-flowing unpaired electrons from the sp²-hybridized carbon network promote electron transport between particles, further enhancing the catalytic PMS efficiency [15]. These specific surfaces structure of GO make it stand out as a suitable supporter [16]. Although, GO has been used to couple with some metal oxides such as Fe₂O₃, TiO₂, CoO_x, CoFeNi LDH, and even spinel ferrite [17-20]. However, rarely has the combination of GO and CoFe₂O₄ been reported for heterogeneous catalytic PMS. Particularly, the improving mechanism of GO-CoFe₂O₄ composites on catalytic PMS needs to be investigated, with a focus on the function of GO and the interaction between GO and CoFe₂O₄.

  Herein, a series of GO-CoFe₂O₄ composites were fabricated and introduced to catalyze PMS for DBP removal. The degrading behavior and physicochemical properties were taken into consideration when choosing the optimized synthesis faction of catalyst. The impacts of different operational parameters, such as catalyst dosage, PMS concentration, initial DBP concentration and initial pH on the activity of GO-CoFe₂O₄/PMS system were systematically explored. The utilization efficiency of PMS was optimized by designing the GO-CoFe₂O₄ particles dosing process. Besides, the cycling stability performance of GO-CoFe₂O₄ was also examined. The radical quenching tests and ESR detection were conducted to determine the radicals produced by GO-CoFe₂O₄/PMS. Finally, the underlying enhanced mechanism of GO-CoFe₂O₄-catalyzed PMS was thoroughly clarified.

  The detailed experimental methods were shown in Texts S1-S5 (Supporting information).

  SEM and TEM techniques were utilized to witness the morphology of GO, CoFe₂O₄ and GO-CoFe₂O₄ (Fig. 1). Figs. 1a and d showed the multilayer GO structure with a slightly wrinkly surface. CoFe₂O₄ particles with diameters from 100 nm to 200 nm exhibited a rough and irregular polygonal shape and agglomerated appearance (Figs. 1b and e). For GO-CoFe₂O₄, it showed that the GO morphology was clear, and CoFe₂O₄ particles were well distributed on GO sheets (Figs. 1c and f). Besides, the support on GO sheets clearly suppressed the aggregation of CoFe₂O₄, which was attributed to the strong absorption between the multiple functional groups present on the surface of GO and metal sites [21]. It was beneficial to the heterogeneous PMS activation process.

  Figure 1

  

  Figure 1.  SEM and TEM images of (a, d) GO, (b, e) CoFe₂O₄ and (c, f) 5%GO-CoFe₂O₄.

  DownLoad: Full-Size Img PowerPoint

  Crystal structures of as-prepared materials were examined by XRD. As seen in Fig. S1 (Supporting information), as a precursor, GO had an intense diffraction peak at around 2θ = 9.5°, regarding the (001) interplanar distance of 0.76 nm, which represented that GO formed a well-ordered layered structure [22]. The pattern of CoFe₂O₄ in Fig. 2a exhibited peaks at 18.2°, 30.3°, 35.6°, 43.2°, 53.5°, 57.4° and 62.7° which were attributed to (111), (220), (311), (400), (422), (511) and (440) planes, respectively. These were in a well-defined crystal structure of spinel-type CoFe₂O₄ (JCPDS No. 22–1086, space group Fd 3m (227)). It was evident that the XRD pattern of GO-CoFe₂O₄ was basically consistent with pure CoFe₂O₄, without the typical diffraction peak of GO. It validated the successful deposition of CoFe₂O₄ between the inter-layers of GO. However, the introduction of GO increased the (311) intensity of CoFe₂O₄ obviously. It was due to the fact that a large number of -OH and -COOH groups on GO combined with H atoms of ferrite to strengthen the spinel structural properties of CoFe₂O₄ [23]. The peak intensity of GO-CoFe₂O₄ became low as the GO content increases to 10%. CoFe₂O₄ crystal growth was constrained by GO layers, and in situ-fabricated CoFe₂O₄ was covered by redundant GO layers [24].

  Figure 2

  

  Figure 2.  (a) XRD patterns, (b) FT-IR spectra, (c, d) N₂ absorption-desorption isotherms of different catalysts, the inserts are the corresponding pore size distributions.

  DownLoad: Full-Size Img PowerPoint

  The wide-scan XPS spectra demonstrated the co-existence of Co, Fe, O, C in 5%GO-CoFe₂O₄ composite (Fig. S2 in Supporting information). The EDS analysis in Fig. S3 (Supporting information) further confirmed the surface elemental composition of the CoFe₂O₄ and GO-CoFe₂O₄ material. The weight percentages of Co, Fe and O in CoFe₂O₄ were 31.48%, 56.75% and 10.35%, respectively (Table S1 in Supporting information). Another aspect, the weight percentages of Co, Fe, O and C in 5%GO-CoFe₂O₄ were 30.36%, 49.72%, 11.29% and 5.48%, respectively. The nominal loading weight percentage of GO and the fraction of C showed a strong correlation. Additionally, the less O component indicated the presence of oxygen defects on as-prepared materials.

  According to Fig. 2b, FT-IR spectra of GO at about 1730, 1626, 1230 and 1031 cm⁻¹ could ascribe to stretching vibration of C=O and C=C, O—H bending vibrations and C—O stretching vibrations, respectively [25]. Indeed, FT-IR spectra of pure CoFe₂O₄ displayed a peak at 564 cm⁻¹, which was attributed to the Co-O and Fe-O vibrations [26]. It was worth noting that in the spectra of GO-CoFe₂O₄, the peaks at 1730 and 1626 cm⁻¹, corresponding to the functional groups of C=O and C=C in GO, blue shifted.

  The Raman spectra of GO, CoFe₂O₄ and 5%GO-CoFe₂O₄ were depicted in Fig. S4 (Supporting information). At 1352 and 1584 cm⁻¹, respectively, two prominent Raman peaks matched with the D and G bands of GO were seen [27]. The degrees of defects and disorders of carbon materials could be clearly reflected by the intensity ratio of D band and G band. Since D and G's band intensities were fundamentally the same, graphitization of GO was to a lesser extent. It might be brought on the large number of groups that are present on its surface and cause considerable sp³ hybridization. The D and G bands of GO were not found in Raman spectra of the composites because of the low concentration of GO in them. Additionally, CoFe₂O₄ and 5%GO-CoFe₂O₄ had obvious characteristic peaks at 323, 498 and 627 cm⁻¹, which were accountable for the stretching vibration of Fe-O on the tetrahedron [28].

  The N₂ adsorption-desorption isotherms of CoFe₂O₄ and 5%GO-CoFe₂O₄ were given in Figs. 2c and d. As observed, both materials possessed the mesoporous structure due to the usual Langmuir-isotherm (type Ⅳ) with a hysteresis loop of H3-type and a centralized pore size distribution within 2–4 nm [29]. The related textural properties were summed up in Table S2 (Supporting information). The specific surface area of CoFe₂O₄ and 5%GO-CoFe₂O₄ was 63.38 and 95.84 m²/g, respectively. This result confirmed our hypothesis that the GO sheets would act as a support and inhibit the aggregation of CoFe₂O₄ (as shown in Figs. 1c and f). Benefiting from the larger surface area, 5%GO-CoFe₂O₄ could expose more active sites and adequately contact with PMS and contaminations.

  The catalytic performances of CoFe₂O₄ and GO-CoFe₂O₄ were determined for DBP removal via PMS activation. As presented in Fig. 3a, the individual PMS produced ~6% DBP elimination, indicating its negligible intrinsic oxidizing capacity. The GO/PMS system achieved 19% of DBP removal efficiency within 30 min, indicated that the PMS activation by GO had tiny contribution to DBP removal (Fig. S5 in Supporting information). However, with the addition of CoFe₂O₄, about 49% of DBP was removed within 30 min. Particularly, a significant increase in the catalytic properties could be effectively accomplished by GO loading, suggesting that the coupling effect between CoFe₂O₄ and GO greatly improved the activation capacity toward PMS. Meanwhile, the impact of GO content on the catalytic function of GO-CoFe₂O₄ was examined. The outcomes uncovered a positive correlation between the GO content and the catalytic performance of GO-CoFe₂O₄, however, the promotion effect was limited. For instance, 90% of DBP removal was obtained over 5%GO-CoFe₂O₄, while in the 10% GO-CoFe₂O₄/PMS system, 82% of DBP degradation was attained. The minimal DBP adsorption activity of all catalysts further demonstrated that the removal of DBP was due to degradation rather than adsorption (Figs. S5 and S6 in Supporting information). Additionally, the pseudo-first-order kinetics (ln(C/C) = −kt) could fit the DBP degradation curves well. The rate constant (k_DBP) in 5%GO-CoFe₂O₄/PMS, which was 0.060 min⁻¹ in Fig. 3b, was four times more than that in CoFe₂O₄/PMS, which was 0.015 min⁻¹. Moreover, 5%GO-CoFe₂O₄ exhibited better mineralization of DBP than CoFe₂O₄ (12% vs. 37% of TOC removal) in activating PMS. The results indicated that the heterogeneous structure instead of the physical mixing of GO and CoFe₂O₄ led to the enhanced catalytic function of GO-CoFe₂O₄.

  Figure 3

  

  Figure 3.  (a) DBP removal and (b) k constants and TOC removal in different PMS oxidation systems. (c) Consecutive use of the catalytic activity of 5%GO-CoFe₂O₄. (d) The concentration of leaching metal ions for 5%GO-CoFe₂O₄. Catalysts: 0.1 g/L, C_0[DBP]: 2 µmol/L, C_[PMS]: 20 µmol/L, initial pH of 7.0, T: 20 ± 2 ℃.

  DownLoad: Full-Size Img PowerPoint

  Successive experiments were conducted to assess the potential for stability in the 5%GO-CoFe₂O₄/PMS catalytic system. As figuring out in Fig. 3c, 5%GO-CoFe₂O₄ continued to exhibit high catalytic activity in the sixth runs. A slight decline in DBP degradation efficiency might cause the unavoidable reduction of catalytic sites during recycling and cleaning. Then, ICP-OES analysis was used to look at the metal leaching capabilities of catalysts. Less than 0.05 and 0.02 mg/L of Co and Fe ions leaked after each cycle, respectively (Fig. 3d). Only around 3% of DBP was lost as a result of leaching, indicating the dominant role of heterogenous catalytic reaction on 5%GO-CoFe₂O₄. Therefore, it demonstrated excellent stability and recyclability for DBP elimination during PMS activation.

  Considering the optimal performance of 5%GO-CoFe₂O₄ and its great potential for practical use, these variable parameters, such as catalyst dosage, initial DBP concentration, PMS concentration, initial pH and temperature were analyzed. The results were displayed in Fig. 4. In Fig. 4a, when the catalyst dosage was raised from 0.02 g/L to 0.2 g/L, DBP degradation was accelerated from 55% to 93%. The apparent improvement in DBP removal might be explained by a greater dosage of 5%GO-CoFe₂O₄, which could have created more surface-active sites that assisted in PMS activation and further led to the creation of more reactive ROS. While the DBP degradation minor decreased as the catalyst dosage was increased further, to 0.4 g/L. This finding might primarily be explained by the fact that extra more catalyst particles would self-bond to form aggregates, resulting in a less effective use of active sites on catalyst surfaces [30].

  Figure 4

  

  Figure 4.  Effect of (a) catalyst dosage, (b) initial DBP concentration, (c) PMS concentration and (d) initial pH on DBP degradation in the 5%GO-CoFe₂O₄/PMS system. Catalysts: 0.1 g/L, C_0[DBP]: 2 µmol/L, C_[PMS]: 20 µmol/L, initial pH of 7.0, T: 20 ± 2 ℃.

  DownLoad: Full-Size Img PowerPoint

  The concentration of contaminants, as an essential component in the treatment of wastewater, had a significant influence on the quantity and duration of PMS that were required. Evidently, raising the initial concentration of DBP slowed down its degradation, as seen in Fig. 4b. As the DBP concentration rose from 1 µmol/L to 4 µmol/L, the DBP elimination in the 5%GO-CoFe₂O₄/PMS system fell from 97% to 70%. Higher initial concentrations caused less efficient DBP breakdown because more DBP molecules competed for the limited number of reactive species. Additionally, more pollutant molecules would bind to the surface of catalysts that prevented PMS from reacting with the redox-active centers of 5%GO-CoFe₂O₄. Therefore, an initial DBP concentration of 2 µmol/L was used in our tests to provide appropriate DBP removal and to mimic the trace level in real situations. Furthermore, PMS was an important source of ROSs (i.e., SO₄^•−, ^•OH, ¹O₂) in the oxidation system, also playing a key role in the 5%GO-CoFe₂O₄/PMS system. DBP removal rate increased from 65% to 98% with PMS concentration ranging from 10 µmol/L to 100 µmol/L, as shown in Fig. 4c. The more PMS participated in the catalytic reaction to form ROSs to enhance the degradation of pollutants.

  For the 5%GO-CoFe₂O₄/PMS system, the effect of initial solution pH on DBP removal was identified. As can be observed from Fig. 4d, DBP elimination significantly improved from 64% to 90% at initial pH values varying from 5.0 to 7.0. The effectiveness of DBP degradation decreased to 75% even though the solution pH climbed further to 9.0. It is widely established that the catalyst surface would become neutral, protonated, or deprotonated, respectively, depending on whether the solution pH was near to, below, or above the pH_PZC [31]. The surface of 5%GO-CoFe₂O₄ became deprotonated and negatively charged at pH > 6.8 (pH_PZC), which made HSO₅⁻ less likely to bind to the surface of catalyst owing to electrostatic repulsion. In addition, SO₄^•− would react with OH⁻ to generate ^•OH with relatively weak oxidative capacity, making the inhibition of DBP removal. Whereas, the decreased removal efficacy of DBP in acidic circumstances was considered to be caused by the creation of an H-bond between H⁺ and the O—O group of HSO₅⁻, which led to the comparatively high stability of the oxidant [32].

  The degradation efficiency of DBP with different amounts of humic acid (HA) or various anions in the 5%GO-CoFe₂O₄/PMS system was investigated (Fig. S7 in Supporting information). Almost no apparent inhibition on DBP removal efficiency could be observed in the presence of HA with the concentrations in the range of 0–10 mg/L. Besides, the common inorganic species (e.g., Cl⁻, HCO₃⁻, and NO₃⁻ ions) with different concentrations also exhibited a very limited effect on DBP degradation. Herein, 5%GO-CoFe₂O₄ could be a promising heterogeneous catalyst in PMS activation for the degradation of organic pollutants [33,34].

  The PMS decomposition was detected in the 5%GO-CoFe₂O₄/PMS system, its change trend was consistent with the DBP degradation (Fig. 5a). 91% of PMS was rapidly consumed in the first 3 min, indicating that PMS molecules were rapidly decomposed to generate large amounts of ROSs at this stage. PMS could not be detected at the end of the catalytic reaction. The reaction stoichiometric efficiency was further used to quantify the utilization efficiency of PMS for DBP degradation, it was defined as the ratio of the number of moles of DBP oxidized (Δ[DBP]/[DBP]₀) to the number of moles of PMS consumed (Δ[PMS]/[PMS]₀) in the presence of 5%GO-CoFe₂O₄ [35]. As shown in Fig. 5b, PMS utilization increased over reaction time (U_0.5 = 0.63 to U₃₀ = 0.87). Besides, with the increase of initial DBP concentration, PMS utilization showed an increasing trend. It was attributed to the fact that the produced amount of ROSs far exceeded the required amount of ROSs for DBP degradation in the initial stage of the reaction, so more DBP molecules could consume the excess ROSs, thereby improving the utilization of PMS. Unfortunately, increasing the initial DBP concentration improved the utilization of PMS, but decreased DBP removal rate. Additionally, we tried to reduce the catalyst dosage to 0.02 g/L, DBP degradation and PMS decomposition were 53.1% and 70%, respectively (Fig. 5c). PMS utilization remained around 0.7 (Fig. 5d). Apparently, although PMS was well utilized when increasing DBP concentration or decreasing catalyst dosage, DBP degradation was not satisfactory.

  Figure 5

  

  Figure 5.  (a) DBP removal and PMS decomposition in 5%GO-CoFe₂O₄/PMS. (b) Effect of DBP concentration on PMS utilization in 5%GO-CoFe₂O₄/PMS. Effect of (c, d) catalyst dosage and (e, f) catalyst dosing method on DBP removal, PMS decomposition and PMS utilization. Catalysts: 0.1 g/L (a, b, e, f) or 0.02 g/L (c, d), C_0[DBP]: 2 µmol/L (a, c-f), C_[PMS]: 20 µmol/L, initial pH of 7.0, T: 20 ± 2 ℃.

  DownLoad: Full-Size Img PowerPoint

  Considering the practical application, it is necessary to find a reasonable method to improve the PMS utilization efficiency while ensuring DBP degradation. In view of the above, a method, batch delivery of catalyst, was carried out. 0.01 g of catalyst was added to the system at reaction times of 0, 3, and 6 min, and PMS and DBP concentrations were determined. Such a dosing method made PMS always maintain at the stage of rapid decomposition. Therefore, PMS could sustainably generate ROSs to degrade DBP. DBP degradation efficiency reached 80% within 10 min and 90% after 20 min (Fig. 5e). On the other hand, PMS utilization was also maintained at a relatively high value (U₁₀ = 0.87) (Fig. 5f). This dosing method not only saved the amount of catalyst and the reaction time but also achieved the desired degradation efficiency of substance.

  DFT calculation was performed to investigate PMS adsorption characteristics and the cleavage of O—O bond on the (311) plane of CoFe₂O₄ and GO-CoFe₂O₄, respectively. As shown in Fig. 6, the atomic structures of a PMS molecule adsorbing on the surface of CoFe₂O₄ (311) (panel a) and GO-CoFe₂O₄ (100) (panel b) were established, respectively. As summarized in Table S3 (Supporting information), relative to a free PMS molecule without activation, the O—O bond length (l_O—O) of PMS was considerably lengthened after binding to catalysts. This implied an enhanced potential to break O—O bonds. The adsorption energy (E_ads(HSO₅⁻)) of PMS on (311) plane of GO-CoFe₂O₄ was −3.14 eV, which was higher than the corresponding value for CoFe₂O₄ (311) (E_ads(HSO₅⁻) = −2.36 eV), respectively. Moreover, the results of the bard charge analysis revealed that the PMS molecules and GO-CoFe₂O₄ experienced higher charge transfer (Q = 0.16 |e|). The stronger binding affinity and electron transferability indicated that the PMS bound on the GO-CoFe₂O₄ (311) was more active. It was in line with the conclusions of the experiments that GO-CoFe₂O₄ possessed better catalytic performance for PMS activation.

  Figure 6

  

  Figure 6.  The optimized crystal models of PMS adsorption on the (311) surface of (a) CoFe₂O₄ and (b) GO-CoFe₂O₄. The blue, purple, yellow, red, gray and white balls represent Co, Fe, S, O, C and H atoms, respectively.

  DownLoad: Full-Size Img PowerPoint

  In general, the generation efficiency of ROSs during PMS activation was mostly dictated by the catalysts' capacity to transport electrons. In order to highlight the synergy between GO and CoFe₂O₄ in GO-CoFe₂O₄ activating PMS, the electrochemical analysis was conducted to measure the redox potential and charge transfer ability in the of 0.5 mol/L Na₂SO₄ solution mixing with PMS. CV behaviors of CoFe₂O₄ and GO-CoFe₂O₄ were explored with a three-electrode device (Fig. 7a). Clear reduction currents were seen with the CoFe₂O₄ or GO-CoFe₂O₄ electrodes in the presence of PMS, proving that the reduction process took place on catalyst surfaces in tandem with the breakdown of PMS. Furthermore, GO-CoFe₂O₄ demonstrated a greater current density and superior reductive ability to CoFe₂O₄ for coordinating the redox process. It was proposed that GO-CoFe₂O₄ potentially exhibited a more reliable redox circulation than pure CoFe₂O₄, which made electron transfer possible in PMS activation.

  Figure 7

  

  Figure 7.  (a) Cyclic voltammograms and (b) electrochemical impedance spectroscopy of different catalysts.

  DownLoad: Full-Size Img PowerPoint

  Besides, to comprehend the charge transfer kinetics at the catalyst/solution interface, EIS plots of CoFe₂O₄ and GO-CoFe₂O₄ were recorded (Fig. 7b). Regarding to the EIS Nyquist plots, 5%GO-CoFe₂O₄ displayed the smallest semicircle diameter, which denoted the lowest charge transfer resistance. Accelerating the charge transfer during the catalytic decomposition of PMS proved advantageous. CoFe₂O₄ particles were dispersedly supported on GO to achieve the loose structure, and at the same time, the ability of electron transfer about CoFe₂O₄ particles was promoted by GO [36]. Therefore, the advantages of GO loading to CoFe₂O₄, which decreased CoFe₂O₄'s intrinsic charge transfer impedance and helped to improve the charge transport between GO-CoFe₂O₄, PMS, and DBP, were highlighted by the electrochemical results.

  To examine the ROSs generated in the 5%GO-CoFe₂O₄/PMS system, quenching experiments were carried out (Fig. 8a). As a scavenger for SO₄^•− and ^•OH, methanol (MeOH) was utilized, with rate constants of 2.5 × 10⁷ and 9.7 × 10⁸ L mol⁻¹ s⁻¹, respectively [37]. tert–Butyl alcohol (TBA), however, was thought to be the unique radical screening agent to selectively quenching ^•OH (5.6 × 10⁹ L mol⁻¹ s⁻¹) [38]. The addition of 100 mmol/L TBA exhibited an obvious reduction of 42% on DBP degradation and the k_DBP value was decreased from 0.060 min⁻¹ to 0.016 min⁻¹ (Fig. 8b). It denoted that ^•OH existed and help to remove DBP in such an oxidation system. A more noticeable negative impact on the clearance of DBP was seen after overdosing MeOH, the removal rate of DBP fell to 18%. The results had been validated that the results had been validated that the radical oxidation pathway (SO₄^•− and ^•OH) played a leading part in the oxidation of DBP due to almost complete suppression.

  Figure 8

  

  Figure 8.  Effect of different scavengers on (a) DBP degradation and (b) k constants in the 5%GO-CoFe₂O₄/PMS system. Spin-trapping ESR spectra for (c)^•OH/SO₄^•− and (d) ¹O₂ in the different PMS system. Catalysts: 0.1 g/L, C_0[DBP]: 2 µmol/L, C_[PMS]: 20 µmol/L, C_[TBA] = C_[MeOH]: 100 mmol/L, C_[FFA]: 1 mmol/L, C_[SOD]: 50 U/mL, initial pH of 7.0, T: 20 ± 2 ℃.

  DownLoad: Full-Size Img PowerPoint

  Besides, we employed superoxide dismutase (SOD, a typical O₂^•− quenching molecule) to quench the potential O₂^•− [39]. The negligible inhibition in degradation efficiency ruled out the involvement of O₂^•− in such a system. Additionally, after adding furfuryl alcohol (FFA, a classic ¹O₂ scavenger) [40], DBP degradation was incredibly sluggish and 78% of DBP could be broken down. It was hypothesized that ¹O₂ was generated and played a limited role in 5%GO-CoFe₂O₄/PMS. Therefore, these findings verified that the degrading mechanism for DBP during the 5%GO-CoFe₂O₄/PMS process primarily involved SO₄^•−, ^•OH and ¹O₂, with the former two playing a major role.

  To explicitly identify the relevant ROSs functioning in the GO-CoFe₂O₄/PMS catalytic process, in situ ESR studies were also conducted. The lack of distinctive peaks in the DMPO and PMS alone system, as shown in Fig. 8c, proved that no radical could be formed in the absence of catalysts. As anticipated, the CoFe₂O₄/PMS/DMPO system revealed a set of characteristic peaks indexed to DMPO-OH and DMPO-SO₄ adducts, certificating the production of ^•OH and SO₄^•− radicals from the broken of peroxide O—O bond in PMS by CoFe₂O₄ [41]. The three-line peaks of the TEMP-¹O₂ adducts were simultaneously visible in the CoFe₂O₄/PMS/TEMP system with equal intensities, demonstrating the formation of ¹O₂ (Fig. 8d) [42]. Notably, the inclusion of 5%GO-CoFe₂O₄ greatly increased the relative intensity of the ESR signals, indicating the formation of ^•OH, SO₄^•− and ¹O₂ was enhanced by GO substrate in 5%GO-CoFe₂O₄/PMS.

  Moreover, LSV analysis in Fig. S8a (Supporting information) was used to check the electron transfer pathway. The result demonstrated that the current density upon PMS addition followed by DBP addition showed negligible changes, which eliminated the occurrence of electron transfer pathways in the 5%GO-CoFe₂O₄/PMS system. The contribution of high-valence metal-oxo species for DBP degradation in the 5%GO-CoFe₂O₄/PMS system was also excluded by insignificantly adverse effects of DMSO in quenching experiment (Fig. S8b in Supporting information).

  To gain knowledge about the catalytic process of 5%GO-CoFe₂O₄/PMS, the XPS spectra was employed. Significant alterations were seen through comparing the XPS spectra of fresh and used 5%GO-CoFe₂O₄ samples. It was possible to deconvolute Co 2p into two peaks with binding energies (B.E.) of 780.9 and 778.6 eV, corresponding to Co^II and Co^III, respectively [43]. After oxidation, the content of Co^II increased from 46% to 52%, while the proportion of Co^III decreased by 6% (Fig. 9a). The peaks at higher B.E. (712.4 eV) in Fig. 9b were attributed to Fe^III, while lower B.E. (710.2 eV) were associated to Fe^II [44]. 54% and 46% of the total Fe was represented by the relative ratios of Fe^II and Fe^III, respectively. After reaction, the relative ratio of Fe^II dropped slightly, from 54% to 52%, correspondingly, Fe^III changed from 46% to 48%. The findings showed that the redox pairs of Co^II/Co^III and Fe^II/Fe^III played a role in the catalytic cycle, and aided in the conversion of charges to produce ROSs during PMS activation.

  Figure 9

  

  Figure 9.  XPS spectra of 5%GO-CoFe₂O₄: (a) Co 2p, (b) Fe 2p, (c) O 1s and (d) C 1s.

  DownLoad: Full-Size Img PowerPoint

  The O 1s spectra (Fig. 9c) was divided into three distinct peaks, which corresponded to H₂O at ~532.9 eV, adsorbed oxygen species (O_ads) at ~531.1 eV and lattice oxygen species (O_latt) at ~529.7 eV, respectively [45]. The quantity of O_latt was found to have decreased by 7% following the catalytic process, demonstrating that lattice oxygen was used during the catalytic reaction that went along with the redox of metal ions. Accordingly, the percentage of O_ads rose from 37% to 50%, showing that certain O_ads might change into O_latt by obtaining electrons from the system in conjunction with the oxidation of Co^II and Fe^II. Additionally, the role of dissolved O₂ was also be confirmed. As shown in Fig. S9 (Supporting information), the DBP degradation under N₂ condition almost equaled the one observed without gas introduction. Furthermore, there was also no visible discrepancy on ESR signal of TEMP-¹O₂ under air and N₂ atmosphere. Therefore, it indicated the dissolved O₂ in the water did not contribute significantly to the oxidation of DBP and the generation of ¹O₂ in this system.

  Furthermore, C—C/C=C, C—O, C=O, and O—C=O bonds were identified as the fitting peaks at 284.5, 286.2, 288.2, and 289.5 eV, respectively (Fig. 9d) [46]. After catalysis, it was discovered that the relative ratio of C=O in 5%GO-CoFe₂O₄ dropped from 14% to 9%. The study came to the conclusion that C=O took part in the catalytic reaction, and that the newly created SO₄^•− and ^•OH were also capable of reducing the amount of C=O in 5%GO-CoFe₂O₄ through the radical oxidation process.

  Based upon the aforementioned results, possible PMS activation pathways using GO-CoFe₂O₄ were postulated and schematically displayed in Fig. 10. DBP degradation was facilitated by both radical and non-radical routes in the GO-CoFe₂O₄/PMS system. HSO₅⁻ and Co^II/Fe^II on the surface of catalyst reacted to form SO₄^•−, which then oxidized the Co^II/Fe^II to Co^III/Fe^III (Eq. 1). Simultaneously, Co^III/Fe^III reacted with HSO₅⁻ to produce SO₅^•− (Eq. 2), Additionally, H₂O or OH⁻ could partially react with SO₄^•− to generate ^•OH in part (Eq. 3). ¹O₂ was created via the interaction between H₂O and SO₅^•− or the compound of SO₅^•− (Eqs. 4 and 5). During the GO-CoFe₂O₄/PMS process, the redox cycles of Fe^III/Fe^II and Co^III/Co^II improved the transport of electrons and the production of various ROSs. In the meantime, the plenty of oxygen-containing groups on the surface of GO, such as C=O groups, were electron-rich and hence had a significant potential to coordinate a redox process [47,48]. Additionally, free-flowing electrons from sp² hybridized carbon network with unpaired electrons promoted the formation of ROSs [49]. Finally, the DBP molecules were quickly broken down by the producing ROSs into minor organic intermediates or even CO₂ and H₂O (Eq. 6).

  (1)

  (2)

  (3)

  (4)

  (5)

  (6)

  Figure 10

  

  Figure 10.  Scheme of proposed mechanism of PMS activation on GO-CoFe₂O₄.

  DownLoad: Full-Size Img PowerPoint

  In conclusion, a series of GO-CoFe₂O₄ catalysts with various GO loading levels were prepared. They showed enhanced catalytic activity compared to pure CoFe₂O₄ in activating PMS for DBP degradation. The excellent removal effectiveness of DBP was achieved under the optimal conditions of 0.1 g/L 5%GO-CoFe₂O₄, 20 µmol/L PMS, 2 µmol/L of DBP at an initial pH of 7.0, where the attained degradation of DBP was 90%. 5%GO-CoFe₂O₄ demonstrated a significant increase in catalytic function with the k_DBP value of 0.060 min⁻¹, which was almost three times greater than that of CoFe₂O₄ (k_DBP = 0.015 min⁻¹). Moreover, recyclability tests revealed that GO-CoFe₂O₄ was stable and reusable. The PMS utilization efficiency was significantly improved by the batch dosing of catalysts. Electrons cycling between Co^II/Co^III and Fe^II/Fe^III led to the cleavage of PMS to produce SO₄^•− and ^•OH and ¹O₂, which attacked DBP molecules to degrade into tiny molecular species and then further mineralized into CO₂ and H₂O. The synergistic function between CoFe₂O₄ and GO led to effective PMS activation and increased SO₄^•− and ^•OH generation.

  Declaration of competing interest

  The authors declare no conflict of interest.

  Acknowledgments

  We appreciate the financial support of the National Natural Science Foundation of China (Nos. 52200010, 52000050), Postdoctoral Science Foundation of China (Nos. 2022M710954, 2020M670913), Open Project of State Key Laboratory of Urban Water Resource and Environment (Harbin Institute of Technology) (Nos. HC202240, 2021TS22).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2023.108397.

  
  
• 1. [1]

     K. Fent, A.A. Weston, Aquat. Toxicol. 76 (2006) 122–159. doi: 10.1016/j.aquatox.2005.09.009

  2. [2]

     O. Bajt, G. Mailhot, M. Bolte, Appl. Catal. B: Environ. 33 (2001) 239–248. doi: 10.1016/S0926-3373(01)00179-5

  3. [3]

     B. Wang, H.X. Wang, W. Zhou, et al., Environ. Sci. Technol. 49 (2015) 1120–1129. doi: 10.1021/es504455a

  4. [4]

     J. Lee, U.V. Gunten, J.H. Kim, Environ. Sci. Technol. 54 (2020) 3064–3081. doi: 10.1021/acs.est.9b07082

  5. [5]

     W.D. Oh, Z. Dong, T.T. Lim, Appl. Catal. B: Environ. 194 (2016) 169–201. doi: 10.1016/j.apcatb.2016.04.003

  6. [6]

     J.L. Wang, S.Z. Wang, Chem. Eng. J. 334 (2018) 1502–1517. doi: 10.1016/j.cej.2017.11.059

  7. [7]

     P.D. Hu, M.C. Long, Appl. Catal. B: Environ. 181 (2016) 103–117. doi: 10.1016/j.apcatb.2015.07.024

  8. [8]

     J. Deng, Y. Shao, N. Gao, et al., J. Hazard. Mater. 262 (2013) 836–844. doi: 10.1016/j.jhazmat.2013.09.049

  9. [9]

     K. Zhang, D. Sun, C. Ma, et al., Chemosphere 241 (2020) 125021. doi: 10.1016/j.chemosphere.2019.125021

  10. [10]

      Y.M. Ren, L.Q. Lin, J. Ma, et al., Appl. Catal. B: Environ. 165 (2015) 572–578. doi: 10.1016/j.apcatb.2014.10.051

  11. [11]

      L.J. Xu, W. Chu, L. Gan, Chem. Eng. J. 263 (2015) 435–443. doi: 10.1016/j.cej.2014.11.065

  12. [12]

      A.T. Smith, A.M. LaChance, S. Zeng, B. Liu, L. Sun, Nano Mater. Sci. 1 (2019) 31–47. doi: 10.1016/j.nanoms.2019.02.004

  13. [13]

      J.P. Zhou, H.L. Liu, J. Luo, et al., ACS Appl. Mater. Interfaces 8 (2016) 18140–18149. doi: 10.1021/acsami.6b05895

  14. [14]

      Z.Q. Cai, A.D. Dwivedi, W.N. Lee, et al., Environ. Sci. -Nano 5 (2018) 27–47. doi: 10.1039/C7EN00644F

  15. [15]

      L. Tang, X. Meng, D. Deng, X. Bao, Adv. Mater. 31 (2019) 1901996. doi: 10.1002/adma.201901996

  16. [16]

      W. Choi, I. Lahiri, R. Seelaboyina, Y.S. Kang, Crit. Rev. Solid State 35 (2010) 52–71. doi: 10.1080/10408430903505036

  17. [17]

      X. Zhu, Y. Zhu, S. Murali, M.D. Stoller, R.S. Ruoff, ACS Nano 5 (2011) 3333–3338. doi: 10.1021/nn200493r

  18. [18]

      C. Chen, W. Cai, M. Long, et al., ACS Nano 4 (2010) 6425–6432. doi: 10.1021/nn102130m

  19. [19]

      X.R. You, C.Y. Huang, W. Huang, et al., Environ. Sci. : Nano 7 (2020) 554–570. doi: 10.1039/C9EN01163C

  20. [20]

      J. Deng, L.W. Xiao, S.J. Yuan, et al., Sep. Purif. Technol. 255 (2021) 117685. doi: 10.1016/j.seppur.2020.117685

  21. [21]

      R. Tabit, O. Amadine, Y. Essamlali, et al., RSC Adv. 8 (2018) 1351–1360. doi: 10.1039/C7RA09949E

  22. [22]

      W.S. Hung, S.M. Chang, R.L.G. Lecaros, et al., Carbon 117 (2017) 112–119. doi: 10.1016/j.carbon.2017.02.088

  23. [23]

      N. Meng, R.C.E. Priestley, Y. Zhang, H.T. Wang, X.W. Zhang, J. Membrane Sci. 501 (2016) 169–178. doi: 10.1016/j.memsci.2015.12.004

  24. [24]

      D.R. Yang, J. Feng, L.L. Jiang, et al., Adv. Funct. Mater. 25 (2015) 7080–7087. doi: 10.1002/adfm.201502970

  25. [25]

      N. Li, Z.F. Geng, M.H. Cao, et al., Carbon 54 (2013) 124–132. doi: 10.1016/j.carbon.2012.11.009

  26. [26]

      M. Seredych, T.J. Bandosz, J. Mater. Chem. 22 (2012) 23525–23533. doi: 10.1039/c2jm34294d

  27. [27]

      M. Kim, C. Lee, J. Jang, Adv. Funct. Mater. 24 (2014) 2489–2499. doi: 10.1002/adfm.201303282

  28. [28]

      K. Ashwini, A.D. Mashkoor, S. Poorva, V. Dinesh, AIP Conf. Proc. 1148 (2014) 1148–1150.

  29. [29]

      T. Lin, L. Yu, M. Sun, et al., Chem. Eng. J. 286 (2016) 114–121. doi: 10.1016/j.cej.2015.09.024

  30. [30]

      Y. Zhao, H.Z. An, G.J. Dong, et al., App. Surf. Sci. 505 (2020) 144476. doi: 10.1016/j.apsusc.2019.144476

  31. [31]

      Y. Zhao, S. Wang, T. Wei, Y.M. Ren, T.Z. Luan, J. Environ. Chem. Eng. 10 (2022) 107241. doi: 10.1016/j.jece.2022.107241

  32. [32]

      J. Liu, Z.W. Zhao, P.H. Shao, F.Y. Cui, Chem. Eng. J. 262 (2015) 854–861. doi: 10.1016/j.cej.2014.10.043

  33. [33]

      Z.L. Wu, Z.K. Xiong, R. Liu, et al., J. Hazard. Mater. 427 (2022) 128204. doi: 10.1016/j.jhazmat.2021.128204

  34. [34]

      R.H. Zhang, M.L. Chen, Z.K. Xiong, Y. Guo, B. Lai, Chin. Chem. Lett. 33 (2022) 948–952. doi: 10.1016/j.cclet.2021.07.029

  35. [35]

      Z. Wang, R.T. Bush, L.A. Sullivan, C.C. Chen, J.S. Liu, Environ. Sci. Technol. 48 (2014) 3978–3985. doi: 10.1021/es405143u

  36. [36]

      S. Wang, M.Y. Zhang, J. Feng, et al., Chem. Eng. J. 430 (2022) 133175. doi: 10.1016/j.cej.2021.133175

  37. [37]

      Z. Liu, H. Ding, C. Zhao, et al., Water Res. 159 (2019) 111–121. doi: 10.1016/j.watres.2019.04.052

  38. [38]

      Y. Zhao, H.Z. An, J. Feng, Y.M. Ren, J. Ma, Environ. Sci. Technol. 53 (2019) 4500–4510. doi: 10.1021/acs.est.9b00658

  39. [39]

      Z.C. Yang, J.S. Qian, A.Q. Yu, B.C. Pan, Proc. Natl. Acad. Sci. U. S. A. 116 (2019) 6659–6664. doi: 10.1073/pnas.1819382116

  40. [40]

      E.T. Yun, J.H. Lee, J. Kim, H.D. Park, J. Lee, Environ. Sci. Technol. 52 (2018) 7032–7042. doi: 10.1021/acs.est.8b00959

  41. [41]

      J.J. You, C.Y. Zhang, Z.L. Wu, et al., Chem. Eng. J. 415 (2021) 128890. doi: 10.1016/j.cej.2021.128890

  42. [42]

      J.E. Yang, Y. Lin, H.H. Peng, et al., Appl. Catal. B: Environ. 268 (2020) 118549. doi: 10.1016/j.apcatb.2019.118549

  43. [43]

      L.P. Wu, B. Li, Y. Li, et al., ACS Catal. 11 (2021) 5532–5543. doi: 10.1021/acscatal.1c00701

  44. [44]

      A.H. Mady, M.L. Baynosa, D. Tuma, J.J. Shim, Appl. Catal. B: Environ. 244 (2019) 946–956. doi: 10.1016/j.apcatb.2018.11.086

  45. [45]

      X.B. Hu, B.Z. Liu, Y.H. Deng, et al., Appl. Catal. B: Environ. 107 (2011) 274–283. doi: 10.1016/j.apcatb.2011.07.025

  46. [46]

      X.T. Li, J. Wang, X.G. Duan, et al., ACS Catal. 11 (2021) 4848–4861. doi: 10.1021/acscatal.0c05089

  47. [47]

      C. Liu, S. Liu, L. Liu, et al., Chem. Eng. J. 379 (2020) 122274. doi: 10.1016/j.cej.2019.122274

  48. [48]

      L.Y. Wu, Z.Y. Li, P.T. Cheng, et al., Water Res. 223 (2022) 119013. doi: 10.1016/j.watres.2022.119013

  49. [49]

      J. Ye, J.D. Dai, D.Y. Yang, C.X. Li, Y.S. Yan, J. Environ. Chem. Eng. 9 (2021) 106076. doi: 10.1016/j.jece.2021.106076

• 

  Figure 1  SEM and TEM images of (a, d) GO, (b, e) CoFe₂O₄ and (c, f) 5%GO-CoFe₂O₄.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (a) XRD patterns, (b) FT-IR spectra, (c, d) N₂ absorption-desorption isotherms of different catalysts, the inserts are the corresponding pore size distributions.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) DBP removal and (b) k constants and TOC removal in different PMS oxidation systems. (c) Consecutive use of the catalytic activity of 5%GO-CoFe₂O₄. (d) The concentration of leaching metal ions for 5%GO-CoFe₂O₄. Catalysts: 0.1 g/L, C_0[DBP]: 2 µmol/L, C_[PMS]: 20 µmol/L, initial pH of 7.0, T: 20 ± 2 ℃.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Effect of (a) catalyst dosage, (b) initial DBP concentration, (c) PMS concentration and (d) initial pH on DBP degradation in the 5%GO-CoFe₂O₄/PMS system. Catalysts: 0.1 g/L, C_0[DBP]: 2 µmol/L, C_[PMS]: 20 µmol/L, initial pH of 7.0, T: 20 ± 2 ℃.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  (a) DBP removal and PMS decomposition in 5%GO-CoFe₂O₄/PMS. (b) Effect of DBP concentration on PMS utilization in 5%GO-CoFe₂O₄/PMS. Effect of (c, d) catalyst dosage and (e, f) catalyst dosing method on DBP removal, PMS decomposition and PMS utilization. Catalysts: 0.1 g/L (a, b, e, f) or 0.02 g/L (c, d), C_0[DBP]: 2 µmol/L (a, c-f), C_[PMS]: 20 µmol/L, initial pH of 7.0, T: 20 ± 2 ℃.

  下载: 全尺寸图片 幻灯片
  

  Figure 6  The optimized crystal models of PMS adsorption on the (311) surface of (a) CoFe₂O₄ and (b) GO-CoFe₂O₄. The blue, purple, yellow, red, gray and white balls represent Co, Fe, S, O, C and H atoms, respectively.

  下载: 全尺寸图片 幻灯片
  

  Figure 7  (a) Cyclic voltammograms and (b) electrochemical impedance spectroscopy of different catalysts.

  下载: 全尺寸图片 幻灯片
  

  Figure 8  Effect of different scavengers on (a) DBP degradation and (b) k constants in the 5%GO-CoFe₂O₄/PMS system. Spin-trapping ESR spectra for (c)^•OH/SO₄^•− and (d) ¹O₂ in the different PMS system. Catalysts: 0.1 g/L, C_0[DBP]: 2 µmol/L, C_[PMS]: 20 µmol/L, C_[TBA] = C_[MeOH]: 100 mmol/L, C_[FFA]: 1 mmol/L, C_[SOD]: 50 U/mL, initial pH of 7.0, T: 20 ± 2 ℃.

  下载: 全尺寸图片 幻灯片
  

  Figure 9  XPS spectra of 5%GO-CoFe₂O₄: (a) Co 2p, (b) Fe 2p, (c) O 1s and (d) C 1s.

  下载: 全尺寸图片 幻灯片
  

  Figure 10  Scheme of proposed mechanism of PMS activation on GO-CoFe₂O₄.

  下载: 全尺寸图片 幻灯片
•