Efficient heterogeneous activation of

Chemical Engineering Journal 345 (2018) 364–374

Contents lists available at ScienceDirect

Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej

Efficient heterogeneous activation of peroxymonosulfate by facilely prepared Co/Fe bimetallic oxides: Kinetics and mechanism Liwei Chena, Xu Zuoa, Liang Zhoua, Yang Huangb, Shengjiong Yangc, Tianming Caia, Dahu Dinga,

T ⁎

a

College of Resources and Environmental Sciences, Nanjing Agricultural University, Nanjing 210095, China College of Materials Science and Engineering, Shenzhen University, Shenzhen 518060, China c School of Environmental and Municipal Engineering, Xi’an University of Architecture and Technology, Xi’an 710055, China b

H I GH L IG H T S

G R A P H I C A L A B S T R A C T

bimetallic oxide catalysts were • Co/Fe fabricated through a facile hydrothermal route.

catalysts were obtained by • Different adjusting the Fe/Co ratio during the preparation process.

Fe O performed much better • Co catalytic activity than CoFe O and 0.4

2.6

4

2

• •

4

CoFe2O4·CoOx. Mixed valence of cobalt and more surface hydroxyl groups were found in Co0.4Fe2.6O4. Degradation pathways of norfloxacin were tentatively elucidated.

A R T I C LE I N FO

A B S T R A C T

Keywords: Sulfate radical Antibiotics Degradation pathway Surface hydroxyl groups Normalized rate constant

The contamination issue caused by the antibiotics is now a big challenge due to its great potential hazardous to the human beings. Herein, three Co/Fe bimetallic oxides were synthesized for the activation of peroxymonosulfate (PMS) to degrade norfloxacin. The elemental analysis results confirmed the stoichiometric ratio of Fe/Co was 1.02, 2.03, and 6.65 (CoFe2O4·CoOx, CoFe2O4, Co0.4Fe2.6O4), respectively. The normalized pseudofirst-order kinetic rate constant (kapp) of Co0.4Fe2.6O4 was 0.18 L/(m2·min), which was over 10 times higher than those of CoFe2O4·CoOx (0.010 L/(m2·min)) and CoFe2O4 (0.012 L/(m2·min)). The degradation process would be further accelerated under weak basic condition (pH = 8.5). The enhanced surface hydroxyl groups were found to be the probable reason for the higher efficiency through the XPS analysis. Moreover, the sulfate radical rather than hydroxyl radical dominated the degradation process in the developed Co0.4Fe2.6O4/PMS system. Overall, the magnetic Co/Fe bimetallic oxide catalyst prepared in this study could be served as a high-quality candidate for the catalytic oxidation of refractory organic pollutants.

⁎

Corresponding author at: 1 Weigang, Xuanwu District, Nanjing 210095, China. E-mail address: [email protected] (D. Ding).

https://doi.org/10.1016/j.cej.2018.03.169 Received 8 January 2018; Received in revised form 6 March 2018; Accepted 29 March 2018 Available online 30 March 2018 1385-8947/ © 2018 Elsevier B.V. All rights reserved.


L. Chen et al.

1. Introduction

Fenton and Fenton-like reactions. The octahedral site in the magnetite structure can accommodate both Fe2+ and Fe3+, allowing them to be reversibly oxidized/reduced. The Fe2+ plays an important role during the catalytic reactions according to the classical Haber-Weiss mechanism [21]. Moreover, the active Fe2+ in the magnetite structure can be substituted by several transition metals with varied redox properties, leading to the variation of the catalytic properties. On the other hand, the Co2+/Oxone system has been reported to be the best combination for the generation of SR when comparing with other systems [16]. Accordingly, the Co/Fe bimetallic oxide produced by the substitution of Fe2+ by Co2+ can be imagined as an efficient heterogeneous catalyst for the activation of PMS. For instance, cobalt ferrite (CF, CoFe2O4) could effectively activate PMS to generate SR [22,23]. However, how to further promote its catalytic activity is still a big challenge. The crystalline structure of the metal oxides plays a vital role during the catalytic activation process. It is reported that the structure of Co3−xFexO4 would transform from normal spinel into inverse spinel with the increases of x, leading to the occurrence of new active sites [24]. Accordingly, it is hypothesized that the catalytic activity of the Fe/Co bimetallic oxides might be significantly influenced by the stoichiometric ratio of Fe/Co. However, very little information is available up until now [25]. Su et al. prepared a series of the CoxFe3−xO4 nanocatalysts for the heterogeneous activation of PMS and found the higher cobalt content in the catalyst was benefit for the removal of Rhodamine B [26]. However, the applied cobalt content was very small (from 1/14 to 1/3). If the activity goes better when further increasing the cobalt content (Co/Fe = 1/2 or 1/1) is obscure. Meanwhile, if the cobalt content influences the properties of the final product is largely unknown. Herein, we reported three Co/Fe bimetallic oxides by regulating the stoichiometric Fe/Co ratios through a facile hydrothermal and post calcination routine. Through inductively coupled plasma analysis, the products were designated as CoFe2O4·CoOx, CoFe2O4, and Co0.4Fe2.6O4, respectively. All the obtained products could readily decompose PMS to generate sulfate radicals and the last one gave the highest degradation rate towards norfloxacin (NOR), a typical fluoroquinolone antibiotic widely used in animal husbandry. More importantly, the solid catalysts could be easily separated from water with external magnetic field according to their ferromagnetic properties. Therefore, the magnetic Co/ Fe bimetallic oxides exhibited very promising performance in environmental remediation.

The pollution issue caused by the abuse of antibiotics has drawn considerable attention all over the world [1]. Not only from the compounds themselves, but also from the antibiotic resistance genes (ARGs), big challenges are posed for the first time [2]. Especially in China, where 210,000 tons of antibiotics are produced annually and near half of them are used for animal husbandry [3]. Therefore, the remediation of residual antibiotics from the contaminated environmental media is of great concern. Catalytic oxidation with the assistance of efficient catalysts is an effective way to eliminate antibiotics pollutants from aquatic environments [4–6]. For example, the generated hydroxyl radical (·OH) performs a high redox potential (E0 = 1.8–2.7 V depending on the pH), making it able to destroy organic compounds without extra addition of chemicals and energy [7,8]. However, some main drawbacks such as the generation of sludge (precipitation of Fe3+) and the low working pH significantly limit its field application. Recently, sulfate radical (SR, SO4●−) is found to be more powerful (E0 = 2.5–3.1 V) and selective (react with organic compounds through electron transfer mechanism) than ·OH [9,10]. The second-order-rate constants of reactions between SR and varieties of organic compounds are generally ranged from 106 to 109 M−1 s−1 [11]. Moreover, the SR is less likely being scavenged by dissolved natural organic matters (NOM) ubiquitously existed in aquatic system due to its selective oxidation [12]. Therefore, the SR based advanced oxidation process (SR-AOP) shows great potential in the environmental remediation [10,13]. SR can be produced from peroxymonosulfate (PMS, HSO5−) or persulfate (PS, S2O82−) through varieties of methods such as pyrolysis [14], photolysis [15], and chemical activation [16,17]. Though the oxidation potential of PMS (1.82 V) is lower than PS (2.01 V) [18], its asymmetric structure appears to make it more readily activated [19,20]. For instance, PMS can be activated homogeneously (with transition metallic ions) [16] or heterogeneously (with solid catalysts) [10]. Among them, the heterogeneous activation seems to be more applicable due to the little secondary pollution and high economical efficiency. Therefore, it is very meaningful to develop efficient catalysts for the heterogeneous activation of PMS. Due to the inverse spinel structure (Scheme 1), magnetite (Fe3O4) possesses several unique features, making it an attractive catalyst in the

2. Experimental section 2.1. Chemicals Ferrous sulfate (FeSO4·7H2O), cobalt nitrate (Co(NO3)2·6H2O), and sodium hydroxide (NaOH) were obtained from Sinopharm Chemical Reagent Co. Ltd., China. Potassium monopersulfate triple salt (Oxone, KHSO5·KHSO4·K2SO4, ≥47% KHSO5 basis) was used as the source of PMS and purchased from Aladdin, China. Norfloxacin (purity grade 98%) and other reagents used for the degradation studies were purchased from Aladdin, China. All the reagents were of at least analytical grade and were used as received. 2.2. Preparation and characterization of catalysts The solid catalysts were prepared through a facile hydrothermal combined with post calcination route. The well-known CoFe2O4 (molar ratio Co:Fe = 1:2) was firstly synthesized. Then, two typical Co/Fe ratios, 1:1 and 1:6.5 (one is above and the other is lower than 1:2) were selected to investigate the effect of stoichiometric ratio on the catalytic activity. For a typical synthesis procedure, known volume of Co (NO3)2·6H2O and FeSO4·7H2O solutions with different concentrations were mixed and magnetically stirred thoroughly. Then, 100 mL NaOH solution (0.75 M) was added into the mixture dropwisely. After the

Scheme 1. Crystalline structures of the spinel Co/Fe bimetallic oxides. 365


L. Chen et al.

to the next degradation run. The reusability experiment was conducted for successive three cycles. Each experiment was conducted for at least 2 times and the average and standard deviations were reported.

addition of NaOH, the resultant mixture was allowed to react for another 30 min. The whole reaction was conducted at 80 °C and air was purged into the reaction system at a rate of 3 L/min. After the reaction, the precipitate was collected by centrifuge and washed several times by DW to remove the residual chemicals. Finally, the completely dried solid product was heated to 400 °C with a temperate ramp of 5 °C/min for 1 h in air. To make the description more readable, the products were designated as CoFe2O4·CoOx, (molar ratio Co:Fe = 1:1), CoFe2O4 (molar ratio Co:Fe = 1:2), and Co0.4Fe2.6O4 (molar ratio Co:Fe = 1:6.5), respectively. For comparison, single metal oxides were prepared in the same conditions without addition of FeSO4 and Co(NO3)2, respectively. The fabricated catalysts were characterized with transmission electron microscopy (TEM), high resolution transmission electron microscopy (HRTEM), X-ray powder diffraction (XRD), X-ray photoelectron spectroscopy (XPS), magnetic hysteresis (M-H), and surface area and porosity analyzer, respectively. The detailed information was provided in the Supporting Information Text S1. The metal contents in the solid samples were measured by Varian 720 inductively coupled plasma optical emission spectrometer (ICP-OES) after complete digestion. The results indicated the atomic ratio of Co:Fe was 1:1.02 for CoFe2O4·CoOx. While for the products CoFe2O4 and Co0.4Fe2.6O4, the ratio was 1:2.03 and 1:6.65, respectively. The pHzpc of the products were determined by mass titration and the pH value was measured by a PHS-3C pH meter (Leici).

2.4. Analytical methods The concentration of NOR was measured with a high performance liquid chromatograph (HPLC, FL 2200). The mobile phase was prepared with methanol, acetonitrile, and water (400 mL: 50 mL: 550 mL), in which 2 mL trifluoroacetic acid was added. The detection wavelength was 278 nm. The concentrations of leached iron and cobalt ions were detected with Varian 720 ICP-OES. To identify the possible degradation intermediates/products, the filtered aqueous sample was collected along with the degradation process and quenched immediately. Then, the liquid sample was concentrated by the solid phase extraction (SPE) on a SPE workstation using HLB cartridge (WAT106202, Waters Oasis). The HLB cartridge was pre-activated prior to the use with 5 mL methanol and 5 mL Milli-Q water, respectively. The final extracted sample was obtained by eluting with 4 mL methanol (2 mL × twice). The Agilent 1200 series HPLC coupled with an Agilent G6410B triple quadrupole (QQQ) mass spectrometer (Agilent Technologies, USA) was adopted for the analysis. The operation details could be found in our previous work [27] and given in the Supporting Information Text S2. 3. Results and discussion

2.3. Degradation experiments 3.1. Characterization of the Fe/Co bimetallic oxides Batch degradation experiments were carried out in a conical flask (100 mL) placed on a rotary shaker (HZ-9310 KB) at room temperature (25 ± 1 °C) unless otherwise stated. A certain amount (10 mg) of the catalyst was ultrasonically dispersed into the NOR solution (15 µM, 50 mL). After that, a desired volume of PMS stock solution (100 mM) was added in the mixture to initiate the degradation process. At predetermined time intervals, the liquid sample was collected and filtered with 0.22 µm microfiltration membrane immediately. To stop the degradation, the equal volume of methanol was injected into the filtered samples to quench the residual radicals. The quenched samples were subjected for the measurement of NOR in a short time. The inherent pH value of the NOR solution was around 6.0 and was adjusted with 0.1 M H2SO4 or NaOH solution if necessary. Noticeably, the quenching assays were carried out by adding excess amounts of ethanol (EtOH) and tbutanol (TBA) prior to the PMS, respectively. The used catalyst was collected and washed with water to remove the adsorbed organics prior

According to the stoichiometric ratios of Co/Fe, three products with different compositions might be occurred. As reported previously, the Fe2+ at the octahedral site could be partially or totally replaced by transition metals including Co2+ [21]. As shown in Scheme 1, for the product Co0.4Fe2.6O4, part of Fe2+ would be substituted by Co2+ and the exact formula should be [Fe3+]tetrahedral[Fe3+ 2+ 2+ Fe0.6 Co0.4 ]octahedralO4. While for the product CoFe2O4, the Fe2+ would be totally substituted by Co2+ at octahedral site and the exact formula should be [Fe3+]tetrahedral[Fe3+ Co2+]octahedralO4. With the further increase of cobalt content, the Fe2+ at octahedral site would be totally replaced by Co2+ and the excess Co2+ would be crystalized on the surface of the product. Consequently, a mixed composite containing both CoFe2O4 and cobalt oxide might be occurred ([Fe3+]tetrahedral[Fe3+ Co2+]octahedralO4·CoOx). Fig. 1 showed the XRD patterns of the prepared samples. The single

Fig. 1. Raw and simulated XRD patterns of the prepared samples. (A: simulated XRD patterns of single metal oxides and CoFe2O4 with standard PDF cards, B: the comparison of the raw XRD data of CoFe2O4·CoOx, CoFe2O4, and Co0.4Fe2.6O4). 366


L. Chen et al.

Fig. 2. Typical low and high resolution TEM images of products Co0.4Fe2.6O4 (A and B), CoFe2O4 (C) and CoFe2O4·CoOx (D) and the corresponding EDS spectra (inset).

iron reduced the lattice constant. In addition, for the product CoFe2O4·CoOx, the occurrence of cobalt oxide (i.e. Co3O4) might also contribute to the shift (Fig. 1A). Both CoFe2O4 and Co0.4Fe2.6O4 consisted of irregular spherical nanoparticles with an average diameter of 22.6 and 13.1 nm, respectively (Fig. 2). However, for the CoFe2O4·CoOx containing more cobalt, the morphology clearly changed from nanoparticles to rhombic nanosheets. The interesting phenomenon clearly indicated that the stoichiometric ratio of Fe/Co played a role in the morphological control. The average diameter was close to the mean grain size calculated from the XRD pattern through the Scherrer’s equation (Table 1). In addition, the lattice fringe spacing of the product Co0.4Fe2.6O4 was about 0.25 nm, which could be assigned to the (3 1 1) reflection of CoFe2O4 (Fig. 2B). The corresponding EDS spectra clearly demonstrated the presence of Fe, Co and O in the catalysts and the atomic ratios of Fe/Co were 5.34 and 1.98 for Co0.4Fe2.6O4 and CoFe2O4 respectively, which were close to the ICP results.

metal oxides prepared in this study were identified by simulating with JCPDS No. 43-1003 (spinel Co3O4, space group Fd 3m (2 2 7)) and 391346 (cubic γ-Fe2O3), respectively. The main phase of the product CoFe2O4 was spinel cobalt ferrite because all the diffraction peaks were coincided with the typical pattern of spinel CoFe2O4 (JCPDS No. 221086, space group Fd 3m (2 2 7)). The nine well-defined peaks occurring at 2θ of around 18.3°, 30.2°, 35.6°, 37.2°, 43.2°, 47.3°, 53.8°, 57.2°, and 62.7° were corresponded to the Bragg planes of (1 1 1), (2 2 0), (3 1 1), (2 2 2), (4 0 0), (3 3 1), (4 2 2), (5 1 1), and (4 4 0), respectively. No impurity peak was observed. For the catalyst Co0.4Fe2.6O4, though the main diffraction peaks were retained, the intensities were remarkably reduced when compared with those of CoFe2O4 (Fig. 1B). The decrease of diffraction intensity was probably because the less Co-Fe bond occurred in Co0.4Fe2.6O4. On the other hand, with the increase of the cobalt content, the XRD diffraction peaks shifted to the high degree. This was because the dope of cobalt with smaller atomic radius than Table 1 Physicochemical properties of the prepared samples. Sample

Atomic ratio Fe/Coa

dTEMb (nm)

dXRDc (nm)

HC (oe)

MR (emu/g)

MS (emu/g)

SSABET (m2/g)

Pore volume (cm3/g)

CoFe2O4·CoOx CoFe2O4 Co0.4Fe2.6O4

1.02 2.03 6.65

14.8 22.6 13.1

14.6 22.0 12.3

274.3 896.8 319.9

9.0 27 2.7

32.2 63.4 9.9

26.5 28.8 37.3

0.29348 0.31348 0.15140

a b c

measured by ICP. average particle size calculated from the TEM images by using the Nano Measurer 1.2 software. average grain size calculated from the XRD data by using the MDI Jade 5.0. 367


L. Chen et al.

Fig. 3. M-H curves (A) and N2 adsorption/desorption isotherm and pore size distribution (B) of the prepared catalysts. (inset shows the catalysts CoFe2O4 (left) and Co0.4Fe2.6O4 (right) in the water before and after magnetic separation).

In addition, the BET specific surface area (SSABET) of Co0.4Fe2.6O4 (37.3 m2/g) was larger than those of CoFe2O4 (28.8 m2/g) and CoFe2O4·CoOx (26.5 m2/g), probably indicating the better adsorption capacity. The XPS spectra of the catalysts were recorded and given in Fig. 4. As shown in the survey spectra, Fe 2p, Co 2p, and O 1s peaks were clearly identified, revealing the presence of these three elements in the products. The impurity peak at 497.7 eV was ascribed to the Auger peak of residual Na adsorbed on the surface. The other impurity peak at 170.1 eV was probably due to the residual S.

Magnetic property of the catalyst was of great importance from application point of view. According to the previous study, the introduced Co2+ would form strong Co-Fe bond, which could improve the magnetic intensity of the pristine Fe3O4 [28,29]. The results given in Table 1 confirmed the conclusion. The saturation magnetization (MS) value of CoFe2O4 was 63.4 emu/g, which was much higher than that of Fe3O4 reported elsewhere [23,30]. With the increase or decrease of the cobalt content, the Ms value was significantly decreased (Fig. 3A). Noting that, the Co0.4Fe2.6O4 exhibited small remanence (MR) and coercivity (HC), meaning that strong magnetic signal could be obtained at small applied magnetic fields [28]. As shown in Fig. 3A (inset), both CoFe2O4 and Co0.4Fe2.6O4 composites dispersed in water solution could be easily separated by using a magnet within 5 min. The perfect separation ability made the catalysts quite acceptable for the field application. The N2 adsorption-desorption isotherm and corresponding pore size distribution curves were given in Fig. 3B. All of the samples displayed a typical type IV isotherm with a H3-type hysteresis loop at a relative pressure of 0.4–1.0, indicating the presence of the mesoporous structure [31]. The corresponding pore size distribution curves (inset) confirmed that the composites were mainly mesoporous (2–50 nm). The products CoFe2O4·CoOx and CoFe2O4 showed a broad pore size distributions and large pore volumes, the average pore size was 20.15 and 23.82 nm, respectively. However, the Co0.4Fe2.6O4 possessed a well-defined pore size distribution. The appearance of one sharp peak at about 3.76 nm suggested that the pore size of Co0.4Fe2.6O4 was comparatively uniform.

3.2. Catalytic activity The degradation profiles clearly indicated that the Co0.4Fe2.6O4/ PMS system was highly efficient in oxidizing NOR. Almost complete degradation was achieved within 3 min when the concentrations of NOR and PMS were 15 and 500 µM, respectively. The degradation process apparently followed pseudo-first-order reaction (Fig. S1). Therefore, the pseudo-first-order kinetic (ln(c0/c) = kapp × t) was adopted to mimic the degradation process. Besides, the heterogeneous activation of peroxides was usually controlled by intrinsic reactions on the catalyst surface [32]. Therefore, the as-obtained apparent reaction rate constant kapp was normalized with SSABET per volume

(

kapp SSA × Dose

) to evaluate the catalytic reactivity [33]. The

, L/(m2·min)

corresponding normalized rate constants were given in the inset of Fig. 5. As shown, the normalized kapp of Co0.4Fe2.6O4 was 0.18 L/ (m2·min), which was over 10 times higher than those of CoFe2O4·CoOx (0.010 L/(m2·min)) and CoFe2O4 (0.012 L/(m2·min)). The Co0.4Fe2.6O4 also performed better activity performance than the single metal oxides (Fig. 5). Moreover, the Co0.4Fe2.6O4 performed a higher adsorption capacity than CoFe2O4 and CoFe2O4·CoOx, which was in well agreement with the larger SSABET of Co0.4Fe2.6O4 (Table 1). All the results clearly illustrated that the Co0.4Fe2.6O4/PMS oxidation system was highly effective for the elimination of NOR from aqueous solution. Though the degradation rate was relatively low in the CoFe2O4·CoOx or CoFe2O4 activated PMS systems, both of them could be completed within 30 min (Fig. S2). Correspondingly, the overall apparent reaction rate constants were 0.12 and 0.08 min−1, respectively. 3.3. Effect of solution pH and dose of PMS

Solution pH usually plays a critical role in SR-AOPs because it influences the speciation of organic compounds as well as the generation of

Fig. 4. The XPS survey spectra of the prepared CoFe2O4 and Co0.4Fe2.6O4. 368


L. Chen et al.

Fig. 5. Nofloxacin degradation profiles in different oxidation systems and the normalized pseudo-first order reaction rate constants (inset). ([NOR]: 15 µM, [PMS]: 0.5 mM, [Catalyst]: 0.2 g/L, pH: 6.0, T = 293 K).

Fig. 6. Effects of solution pH (A) and dose of PMS (B) on the degradation of NOR in Co0.4Fe2.6O4/PMS system. ([NOR]: 15 µM, [Catalyst]: 0.2 g/L, T = 293 K, [PMS]: 0.5 mM in (A), pH: 6.0 in (B)).

As shown in Fig. 6B, the degradation process could be significantly accelerated by increasing the dose of PMS. For instance, the complete degradation was even achieved within 2 min when the dose of PMS was 1 mM. In this case, the kapp even reached 2.05 min−1 (Fig. S3), which was extremely high when compared with other oxidation systems [35,40,41]. For example, the kapp of NOR degradation in a thermal-PS system was only 0.0017–0.0575 min−1 [42]. The kapp of SDZ degradation were 0.23 min−1 and 0.20 min−1 in CuFeO2/PMS and CuFe2O4/PMS systems, respectively [33]. Conversely, the degradation process was remarkably slowed down when the dose of PMS was reduced to 0.1 mM. The kapp was only 0.1 min−1, approximately one twentieth of that in the case of 1 mM PMS.

reactive species [34]. Three typical pH values (4.0, 6.0, 8.5) representing acidic, near neutral, and basic conditions were investigated in the present study. As shown in Fig. 6A, the results clearly indicated that the degradation process was inhibited at extremely acidic condition, which was also reported in Co2+/PMS[35] and CoxFe3−xO4/PMS [36] systems. The pseudo-first-order rate constant kapp decreased from 1.34 min−1 to 0.37 min−1 (Fig. S3). On the other hand, the degradation was slightly enhanced at pH of 8.5. The kapp increased from 1.34 min−1 to 2.25 min−1. The best degradation of sulfadiazine (SDZ) was also achieved at pH of 8.5 in a CuFeO2/PMS oxidation system [33]. According to the previous work, the optimal pH for Co2+/PMS decomposition was near neutral (i.e., 4 < pH < 9) [9,37]. The inhibition at acidic condition was probably due to the stabilization effect of H+ on the HSO5− (PMS mainly exist as HSO5− at pH of 4.0–8.5) [38]. In addition, low pH would cause the protonation of piperazine moiety of NOR and decrease its reactivity towards radicals [39]. The high efficiency of Co0.4Fe2.6O4 /PMS oxidation system at neutral and alkaline conditions suggested that this process was suitable for the treatment of surface water.

3.4. Effect of reaction temperature The degradation process was accelerated with increasing the reaction temperature (Fig. 7). The kapp was firstly obtained through the modeling studies (Fig. S4). The activation energy (Ea) was then 369


L. Chen et al.

Fig. 8. Contribution of ·OH and SR throughout the degradation process. ([NOR] = 15 µM, [PMS] = 0.5 mM, [Catalyst] = 0.2 g/L, [EtOH] = [TBA] = 0.5 M, pH: 6.0, T = 293 K).

Fig. 7. Effect of reaction temperature on the degradation of NOR in Co0.4Fe2.6O4/PMS oxidation system. ([NOR] = 15 µM, [PMS] = 0.5 mM, [Catalyst] = 0.2 g/L, pH: 6.0).

[19,47], and CoxFe3−xO4 nanoparticles [26] activated PMS systems. Moreover, the degradation process was nearly entirely inhibited after adding 0.5 M EtOH, indicating the negligible contribution of SO5●− [9,47]. The leached metal ions were characterized to investigate the contributions of homogeneous and heterogeneous activations. As shown in Fig. S6, the leaching amount of Co2+ was around 73.0 µg/L for CoFe2O4, which was relatively low than that of Co3O4 (143.1 µg/L) prepared in this study. The result confirmed that the CoFe2O4 can markedly suppress cobalt leaching due to the strong Fe-Co interactions. However, the leaching amount of Co2+ was enhanced for Co0.4Fe2.6O4 catalyst (521.7 µg/L), which was quite strange. To the best of our knowledgement, the leaching amount of Co2+ for Co0.4Fe2.6O4 should be between those of CoFe2O4 and Co3O4. The more serious leach phenomenon might be related with the low XRD diffraction intensity of the product Co0.4Fe2.6O4 (Fig. 1). A possible reason was that the poor crystallinity led to the dissolution of Co2+. Undoubtedly, the more leached Co2+ would improve the activation process through the homogeneous activation. In order to clarify the role of dissolved Co2+ during the activation process, a homogeneous Co2+ activation test was then conducted. The result (Fig. S7) clearly indicated that there was heterogeneous activation mechanism besides the homogeneous activation. The high-resolution XPS spectra of Fe 2p, Co 2p, and O 1s were then taken for the deconvolution. According to the electrostatic interactions between photoionized Fe 2p core hole and unpaired Fe 3d electrons, spin-orbit coupling and crystal field interactions [48], the nature of Fe 2p peak broadening was very complicated. As shown in Fig. 9A, the centroids of Fe 2p3/2 and Fe 2p1/2 peaks were located at 710.8 and 724.3 eV respectively, agreeing well with the values reported in previous works [48,49]. Besides, two satellite peaks located at 719.5 and 732.8 eV were also observed probably due to the charge transfer and/or shake-up processes [48]. Moreover, the Fe 2p3/2 spectra could be divided into two Gaussian peaks (710.5 and 712.6 eV in CoFe2O4, 710.8 and 712.9 eV in Co0.4Fe2.6O4), revealing the presence of two nonequivalent bonds of Fe ions in the samples [50]. For the CoFe2O4 sample, the doublets could be ascribed to Fe3+ ions in octahedral sites and tetrahedral sites, respectively. However, it was relatively difficult to ascribe the doublets to Fe3+ in different sites because the crystal structure might be distinct in Co0.4Fe2.6O4 sample. Although no remarkable difference was observed in the Fe 2p spectra, the Co 2p spectra was shifted significantly (Fig. 9B). For the CoFe2O4, the peaks at 780.1 and 795.9 eV with the spin-orbit splitting of 15.8 eV were caused by Co 2p3/2 and Co 2p1/2, respectively [51]. The satellite peaks at higher binding energies were typical shake-up peaks.

calculated through the Arrhenius equation by plotting lnk versus 1/T:

Lnk−app = ln A−

Ea RT

(1)

where A and R are the pre-exponential factor and gas constant (8.314 J K−1 mol−1), respectively. From the slope given in the inset of Fig. 7, the Ea value was calculated to be around 5.7 kJ/mol, revealing a typical diffusion-controlled reaction [43]. Meanwhile, the extremely low activation energy might confirm the high catalytic reactivity of Co0.4Fe2.6O4 [33]. 3.5. Stability of the catalyst The stability of the prepared catalyst Co0.4Fe2.6O4 was evaluated by the reusability experiment. An obvious reduction of the catalytic activity was observed probably due to the leaching of the active Co2+ and the passivation of the catalyst (Fig. S5) [27]. However, more information was needed to verify this speculation, which would be discussed in the following section. Noting that, the degradation efficiency could be recovered by extending the degradation duration. In the third run, the degradation efficiency still could achieve over 80% when the reaction time was 60 min, which was also quite acceptable when comparing with other developed catalysts [44]. 3.6. Activation mechanism The underlying activation mechanism was tried to be elucidated by the oxidative species determination combined with catalyst characterizations. The radical quenching study could give valuable information on the roles of oxidative species in the degradation process. Radicals including SR and ·OH are usually occurred in the SR-AOPs [33,45]. To distinguish their contributions respectively, different chemical probes were usually used. For example, EtOH could scavenge ·OH and SR from the oxidation system with comparable high reaction rate constants (k[·OH] = (1.2–2.8) × 109 M−1 s−1 and k[SR] = (1.6–7.7) × 107 M−1 s−1). On the other hand, the alcohols without α-hydrogen like TBA would preferably react with ·OH rather than SR (k[·OH] = (3.8–7.6) × 108 M−1 s−1 and k[SR] = (4–9.1) × 105 M−1 s−1) [16,46]. The inhibition of the degradation process brought by the two quench agents was shown in Fig. 8. The results clearly indicated that the degradation process was a radical-involved process. The primary contributor was SR while ·OH played a minor role throughout the degradation process. The similar conclusion was well documented in Co2+ [9,16], Co3O4 370


L. Chen et al.

Fig. 9. XPS spectra: Fe 2p (A), Co 2p (B), and O 1s spectra of CoFe2O4 (C) and Co0.4Fe2.6O4 (D) catalysts. (inset of B shows the corresponding high-resolution XPS spectra of Co 2p3/2).

On the other hand, the Co 2p3/2 peak shifted from 780.1 to 781.0 eV for the catalyst Co0.4Fe2.6O4, indicating the variation of oxidation state of Co. The Co 2p3/2 peak was further deconvoluted by Gaussian peaks. As shown in the inset of Fig. 9B, the Co 2p3/2 peak of CoFe2O4 was composed of two peaks at 779.6 and 780.2 eV, which were ascribed to the Co2+ in octahedral and tetrahedral sites, respectively. On the other hand, the Co 2p3/2 peak of Co0.4Fe2.6O4 was composed of two peaks with binding energies of 780.6 and 782.7 eV, which were assigned to the Co2+ and Co3+, respectively [23,50]. In the O 1s region (Fig. 9C and D), the peaks located at around 530 eV could be assigned to the lattice oxygen binding with Fe and/or Co. The other O 1s peak located at higher binding energy (∼532.0 eV) could be assigned as lattice hydroxyl [52]. The results clearly illustrated that the lattice hydroxyl was significantly enhanced in Co0.4Fe2.6O4. The enhanced surface hydroxyl groups might be the reason for the higher efficiency of Co0.4Fe2.6O4 because the degradation process depended greatly on the quantity of surface hydroxyl sites [23,53]. Accordingly, based on the above results and discussion, the heterogeneous activation mechanism in Co0.4Fe2.6O4 activated PMS system was presented in Fig. 10. As well documented, the redox reaction between Co2+ and Co3+ played critical role (Eq. (2)) [22,23]. The surface hydroxyl groups formed by the metal ions and H2O bind with PMS through hydrogen bond. The bond would be cracked by the electrons discharged through the redox reactions and finally SR was generated. The SR could attack the organic pollutants directly through electron transfer mechanism or otherwise react with H2O to produce ·OH (Eqs. (3) and (4)) [54,55]. The generated ·OH would further oxidize the NOR

Fig. 10. Heterogeneous activation mechanism of PMS by Co0.4Fe2.6O4.

through hydrogen abstraction or addition mechanism. On the other hand, the lattice oxygen (O2−) donated electrons to keep the balance of the charge on the surface of catalyst (Eq. (5)) [23]. Consequently, the Co3+ was reduced to Co2+ (Eq. (6)). Besides, the resultant Co3+ might be reduced to Co2+ through Eq. (7) because the reduction of Co3+ by 371


L. Chen et al.

Fig. 11. HPLC-UV spectra of the NOR solution during the degradation process (A), the HPLC-DAD spectra and the corresponding total ion chromatogram spectra of the concentrated sample after degradation (B). ([NOR]: 15 µM, [Catalyst]: 0.2 g/L, [PMS]: 0.5 mM, pH: 6.0, T = 293 K).

Fe2+ was thermodynamically favorable [22,30].

≡

Co2 +

SO•4−

+

HSO− 5

+ H2 O→

→

≡Co3

HSO−4

+

SO•4−

+

OH−

+ •OH− k < 3.0 × 103s−1

fragmentation patterns (Fig. S9), the possible molecular formulas of these intermediates/products were presented in Table S1. Based on the identified intermediates/products, the degradation pathway was tentatively proposed in Fig. 12. The [NOR]●+ was initially produced due to the attack of SR on the CeH bond on the piperazine ring or quinolone group of NOR. Noting that, the [NOR]●+ was very unstable and would quickly react with H2O to generate the hydroxylated NOR (product 7a) [42,56]. The hydroxylated NOR would be further transformed to product 10 (m/z 368) through addition of –OH. Then, the C]C bond on the quinolone group was break down to generate product 11 (m/z 372). Afterwards, the product 8 (m/z 344) were generated through the losing of C]O bond. The eOH bonds were oxidized to ketone/aldehyde bonds to form the product 7b (m/z 336). Noting that, the formation of C]O bond might indicate the involvement of ·OH in the oxidation system besides SR [56]. Product 1 (m/z 198) was finally generated through the elimination of all the ketone and aldehyde bonds. In addition, the piperazine ring cleavage was also widely reported in the AOP degradation processes towards fluoroquinolone antibiotics [56,57,59]. As shown in the pathway II, product 9 (m/z 350) was formed due to the ring open reaction of piperazine ring. Products 6 (m/z 322), 5 (m/z 294), 4 (m/z 279), 3 (m/z 251), and 2 (m/z 233) were generated by the successive elimination of eC]O, eC]O, eCH2eNH2 (+O), eC]O, and eF. This pathway had been well documented in previous studies [56,57].

(2) (3)

SO•4− + OH− → SO24− + •OH k = (6.5 ± 1.0) × 107M−1s−1 (pH > 8.5) (4)

SO•5−

+ O2 − →

≡SO•4−

+ O2

(5)

2 + + SO• − + H O ≡ Co2 + + HSO− 2 5 → Co 5

(6)

≡ Fe2 + + ≡Co3 + → ≡Fe3 + ≡ Co2 + E0 = 1.04V vs. NHE

(7)

To clear elucidate the reason for the decline of the efficiency during the successive three reaction cycles shown in Fig. S5, the recycled catalyst Co0.4Fe2.6O4 was characterized with XPS. As shown in Fig. S8, the elemental compositions were not significantly changed through the reaction. This was coincided with the low leached Co2+ ions detected in Fig. S6. Moreover, this finding perhaps revealed that the losing of active Co2+ might not be the primary reason for the decline of the efficiency. On the other hand, the surface oxygen species varied remarkably after the catalytic reaction. As clearly illustrated in Fig. S8, the O 1s sub-peak located at binding energy of around 532.0 eV was greatly weakened after the catalytic reaction, indicating the loss of hydroxyl groups. This observation well verified the above deduction that the surface hydroxyl groups were the reason for the high efficiency of Co0.4Fe2.6O4.

4. Conclusions 3.7. Degradation pathway In summary, we reported Co/Fe bimetallic oxide catalysts with different stoichiometric Fe/Co ratios for the heterogeneous activation of PMS to generate sulfate radicals (SR). The developed catalyst with Co/Fe ratio of 1:6.5 (Co0.4Fe2.6O4) performed the highest catalytic efficiency. The target pollutant norfloxacin with initial concentration of 15 µM could be completely degraded within few minutes by using the Co0.4Fe2.6O4 (0.2 g/L)/PMS (0.5 mM) system at room temperature. The pseudo-first-order kinetic rate constant even reached 2 min−1 when the pH was 8.5 or PMS was 1 mM. The reaction was a typical diffusioncontrolled reaction. The quenching study clearly indicated that the SR other than ·OH dominated the degradation process. The XPS results revealed that the high catalytic efficiency of Co0.4Fe2.6O4 might be due to the enhancement of the surface hydroxyl groups. Finally, 12 degradation intermediates/products were identified through HPLC-MS and the possible pathway was tentatively proposed.

The degradation intermediates/products were detected by LC-MS/ MS. The chemical structures were tentatively elucidated either by interpretation of multistage mass (MS2) spectra and/or comparison with literature spectra [27]. The oxidation of NOR in AOPs mainly occurred through two pathways, which were piperazine ring cleavage and the oxidation of quinolone moiety [56,57]. Noting that, the oxidative species played distinct roles in these mechanisms. As well known, the addition/abstraction of hydrogen was attributed to the attack of ·OH. However, the SR mainly attacked the pollutant molecular through electron transfer mechanism [58]. In this case, the primary oxidative species was found to be SR. Therefore, the electron transfer mechanism might play a vital role. As shown in the HPLC and the total ion chromatogram (TIC) spectra (Fig. 11), 12 degradation intermediates/products (besides NOR) were detected. Based on mass spectra and

372


L. Chen et al.

Fig. 12. Proposed degradation pathways of NOR in Co0.4Fe2.6O4/PMS oxidation system. ([NOR]: 15 µM, [Catalyst]: 0.2 g/L, [PMS]: 0.5 mM, pH: 6.0, T = 293 K).

Acknowledgments

References

This work was financially supported by the National Natural Science Foundation of China (No. 51608274) and the Fundamental Research Funds for the Central Universities (KJQN 201749, KYZ 201619).

[1] J.L. Martinez, Environmental pollution by antibiotics and by antibiotic resistance determinants, Environ. Pollut. 157 (2009) 2893–2902. [2] V.K. Sharma, N. Johnson, L. Cizmas, T.J. McDonald, H. Kim, A review of the influence of treatment strategies on antibiotic resistant bacteria and antibiotic resistance genes, Chemosphere 150 (2016) 702–714. [3] H.-C. Su, G.-G. Ying, L.-Y. He, Y.-S. Liu, R.-Q. Zhang, R. Tao, Antibiotic resistance, plasmid-mediated quinolone resistance (PMQR) genes and ampC gene in two typical municipal wastewater treatment plants, Environ. Sci. Processes Impacts 16 (2014) 324–332. [4] L. Tang, J. Wang, G. Zeng, Y. Liu, Y. Deng, Y. Zhou, J. Tang, J. Wang, Z. Guo, Enhanced photocatalytic degradation of norfloxacin in aqueous Bi2WO6 dispersions containing nonionic surfactant under visible light irradiation, J. Hazard. Mater. 306 (2016) 295–304. [5] Y.F. Su, G.B. Wang, D.T.F. Kuo, M.L. Chang, Y.H. Shih, Photoelectrocatalytic degradation of the antibiotic sulfamethoxazole using TiO2/Ti photoanode, Appl.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.cej.2018.03.169.

373


L. Chen et al.

Catal. B Environ. 186 (2016) 184–192. [6] Z.Y. Bai, Q. Yang, J.L. Wang, Catalytic ozonation of sulfamethazine using Ce0.1Fe0.9OOH as catalyst: Mineralization and catalytic mechanisms, Chem. Eng. J. 300 (2016) 169–176. [7] E. Elmolla, M. Chaudhuri, Optimization of Fenton process for treatment of amoxicillin, ampicillin and cloxacillin antibiotics in aqueous solution, J. Hazard. Mater. 170 (2009) 666–672. [8] M. Pérez-Moya, M. Graells, G. Castells, J. Amigó, E. Ortega, G. Buhigas, L.M. Pérez, H.D. Mansilla, Characterization of the degradation performance of the sulfamethazine antibiotic by photo-Fenton process, Water Res. 44 (2010) 2533–2540. [9] G.P. Anipsitakis, D.D. Dionysiou, Degradation of Organic Contaminants in Water with Sulfate Radicals Generated by the Conjunction of Peroxymonosulfate with Cobalt, Environ. Sci. Technol. 37 (2003) 4790–4797. [10] W.-D. Oh, Z. Dong, T.-T. Lim, Generation of sulfate radical through heterogeneous catalysis for organic contaminants removal: Current development, challenges and prospects, Appl. Catal. B 194 (2016) 169–201. [11] P. Neta, V. Madhavan, H. Zemel, R.W. Fessenden, Rate constants and mechanism of reaction of sulfate radical anion with aromatic compounds, J. Am. Chem. Soc. 99 (1977) 163–164. [12] M. Mahdi Ahmed, S. Barbati, P. Doumenq, S. Chiron, Sulfate radical anion oxidation of diclofenac and sulfamethoxazole for water decontamination, Chem. Eng. J. 197 (2012) 440–447. [13] L.W. Matzek, K.E. Carter, Activated persulfate for organic chemical degradation: a review, Chemosphere 151 (2016) 178–188. [14] Y. Ji, Y. Fan, K. Liu, D. Kong, J. Lu, Thermo activated persulfate oxidation of antibiotic sulfamethoxazole and structurally related compounds, Water Res. 87 (2015) 1–9. [15] M.M. Ahmed, S. Chiron, Solar photo-Fenton like using persulphate for carbamazepine removal from domestic wastewater, Water Res. 48 (2014) 229–236. [16] G.P. Anipsitakis, D.D. Dionysiou, Radical Generation by the Interaction of Transition Metals with Common Oxidants, Environ. Sci. Technol. 38 (2004) 3705–3712. [17] P. Avetta, A. Pensato, M. Minella, M. Malandrino, V. Maurino, C. Minero, K. Hanna, D. Vione, Activation of persulfate by irradiated magnetite: implications for the degradation of phenol under heterogeneous photo-fenton-like conditions, Environ. Sci. Technol. 49 (2015) 1043–1050. [18] K.-C. Huang, R.A. Couttenye, G.E. Hoag, Kinetics of heat-assisted persulfate oxidation of methyl tert-butyl ether (MTBE), Chemosphere 49 (2002) 413–420. [19] L. Chen, T. Cai, W. Sun, X. Zuo, D. Ding, Mesoporous bouquet-like Co3O4 nanostructure for the effective heterogeneous activation of peroxymonosulfate, J. Taiwan Inst. Chem. Eng. (2017). [20] L. Chen, D. Ding, C. Liu, H. Cai, Y. Qu, S. Yang, Y. Gao, T. Cai, Degradation of norfloxacin by CoFe2O4-GO composite coupled with peroxymonosulfate: a comparative study and mechanistic consideration, Chem. Eng. J. 334 (2018) 273–284. [21] R.C.C. Costa, M.F.F. Lelis, L.C.A. Oliveira, J.D. Fabris, J.D. Ardisson, R.R.V.A. Rios, C.N. Silva, R.M. Lago, Novel active heterogeneous Fenton system based on Fe3−xMxO4 (Fe Co, Mn, Ni): The role of M2+ species on the reactivity towards H2O2 reactions, J. Hazard. Mater. 129 (2006) 171–178. [22] P. Hu, M. Long, Cobalt-catalyzed sulfate radical-based advanced oxidation: a review on heterogeneous catalysts and applications, Appl. Catal. B 181 (2016) 103–117. [23] Y. Ren, L. Lin, J. Ma, J. Yang, J. Feng, Z. Fan, Sulfate radicals induced from peroxymonosulfate by magnetic ferrospinel MFe2O4 (M = Co, Cu, Mn, and Zn) as heterogeneous catalysts in the water, Appl. Catal. B 165 (2015) 572–578. [24] S.G. Christoskova, M. Stoyanova, M. Georgieva, Low-temperature iron-modified cobalt oxide system: Part I. Preparation and characterisation, Appl. Catal. A Gen. 208 (2001) 235–242. [25] H. Sun, X. Yang, L. Zhao, T. Xu, J. Lian, One-pot hydrothermal synthesis of octahedral CoFe/CoFe2O4 submicron composite as heterogeneous catalysts with enhanced peroxymonosulfate activity, J. Mater. Chem. A 4 (2016) 9455–9465. [26] S. Su, W. Guo, Y. Leng, C. Yi, Z. Ma, Heterogeneous activation of Oxone by CoxFe3−xO4 nanocatalysts for degradation of rhodamine B, J. Hazard. Mater. 244–245 (2013) 736–742. [27] D. Ding, C. Liu, Y. Ji, Q. Yang, L. Chen, C. Jiang, T. Cai, Mechanism insight of degradation of norfloxacin by magnetite nanoparticles activated persulfate: Identification of radicals and degradation pathway, Chem. Eng. J. 308 (2017) 330–339. [28] J. Hao, W. Yang, Z. Zhang, S. Pan, B. Lu, X. Ke, B. Zhang, J. Tang, Hierarchical flower-like Co3-xFexO4 ferrite hollow spheres: facile synthesis and catalysis in the degradation of methylene blue, Nanoscale 5 (2013) 3078–3082. [29] L. Xu, J. Wang, Fenton-like degradation of 2,4-dichlorophenol using Fe3O4 magnetic nanoparticles, Appl. Catal. B 123–124 (2012) 117–126. [30] J. Deng, Y. Shao, N. Gao, C. Tan, S. Zhou, X. Hu, CoFe2O4 magnetic nanoparticles as a highly active heterogeneous catalyst of oxone for the degradation of diclofenac in water, J. Hazard. Mater. 262 (2013) 836–844. [31] M. Kruk, M. Jaroniec, Gas adsorption characterization of ordered organic-inorganic nanocomposite materials, Chem. Mat. 13 (2001) 3169–3183. [32] W.P. Kwan, B.M. Voelker, Decomposition of hydrogen peroxide and organic

[33]

[34] [35]

[36]

[37]

[38]

[39] [40]

[41]

[42]

[43] [44] [45] [46] [47]

[48] [49]

[50]

[51]

[52]

[53]

[54] [55]

[56]

[57]

[58] [59]

374

compounds in the presence of dissolved iron and ferrihydrite, Environ. Sci. Technol. 36 (2002) 1467–1476. Y. Feng, D. Wu, Y. Deng, T. Zhang, K. Shih, Sulfate radical-mediated degradation of sulfadiazine by CuFeO2 rhombohedral crystal-catalyzed peroxymonosulfate: synergistic effects and mechanisms, Environ. Sci. Technol. 50 (2016) 3119–3127. C. Liang, Z.-S. Wang, C.J. Bruell, Influence of pH on persulfate oxidation of TCE at ambient temperatures, Chemosphere 66 (2007) 106–113. Y. Ji, C. Dong, D. Kong, J. Lu, New insights into atrazine degradation by cobalt catalyzed peroxymonosulfate oxidation: kinetics, reaction products and transformation mechanisms, J. Hazard. Mater. 285 (2015) 491–500. S. Su, W. Guo, Y. Leng, C. Yi, Z. Ma, Heterogeneous activation of Oxone by CoxFe3−xO4 nanocatalysts for degradation of rhodamine B, J. Hazard. Mater. 244 (2013) 736–742. K. Chan, W. Chu, Degradation of atrazine by cobalt-mediated activation of peroxymonosulfate: different cobalt counteranions in homogenous process and cobalt oxide catalysts in photolytic heterogeneous process, Water Res. 43 (2009) 2513–2521. T. Zhang, H. Zhu, J.-P. Croué, Production of sulfate radical from peroxymonosulfate induced by a magnetically separable CuFe2O4 spinel in water: efficiency, stability, and mechanism, Environ. Sci. Technol. 47 (2013) 2784–2791. D.L. Ross, C.M. Riley, Aqueous solubilities of some variously substituted quinolone antimicrobials, Int. J. Pharm. 63 (1990) 237–250. M.J. Chen, W. Chu, H2O2 assisted degradation of antibiotic norfloxacin over simulated solar light mediated Bi2WO6: kinetics and reaction pathway, Chem. Eng. J. 296 (2016) 310–318. Y. Ding, L. Zhu, A. Huang, X. Zhao, X. Zhang, H. Tang, A heterogeneous Co3O4Bi2O3 composite catalyst for oxidative degradation of organic pollutants in the presence of peroxymonosulfate, Catal. Sci. Technol. 2 (2012) 1977–1984. H. Guo, N. Gao, Y. Yang, Y. Zhang, Kinetics and transformation pathways on oxidation of fluoroquinolones with thermally activated persulfate, Chem. Eng. J. 292 (2016) 82–91. W. Stumm, J.J. Morgan, Aquatic chemistry: chemical equilibria and rates in natural waters, Cram101 Textbook Outlines to Accompany 179 (1996) A277. A. Ghauch, G. Ayoub, S. Naim, Degradation of sulfamethoxazole by persulfate assisted micrometric Fe0 in aqueous solution, Chem. Eng. J. 228 (2013) 1168–1181. G.P. Anipsitakis, E. Stathatos, D.D. Dionysiou, Heterogeneous activation of oxone using Co3O4, J. Phys. Chem. B 109 (2005) 13052–13055. A.B. Ross, P. Neta, Rate constants for Reactions of Inorganic Radicals in Aqueous Solution, National Bureau of Standards, 1979. T. Zeng, X. Zhang, S. Wang, H. Niu, Y. Cai, Spatial confinement of a Co3O4 catalyst in hollow metal-organic frameworks as a nanoreactor for improved degradation of organic pollutants, Environ. Sci. Technol. 49 (2015) 2350–2357. J. Baltrusaitis, D.M. Cwiertny, V.H. Grassian, Adsorption of sulfur dioxide on hematite and goethite particle surfaces, PCCP 9 (2007) 5542–5554. Y. Huang, D. Ding, M. Zhu, W. Meng, Y. Huang, F. Geng, J. Li, J. Lin, C. Tang, Z. Lei, Z. Zhang, C. Zhi, Facile synthesis of α-Fe2O3 nanodisk with superior photocatalytic performance and mechanism insight, Sci. Technol. Adv. Mater. 16 (2015) 014801. Z. Zhou, Y. Zhang, Z. Wang, W. Wei, W. Tang, J. Shi, R. Xiong, Electronic structure studies of the spinel CoFe2O4 by X-ray photoelectron spectroscopy, Appl. Surf. Sci. 254 (2008) 6972–6975. M. Zong, Y. Huang, H. Wu, Y. Zhao, Q. Wang, X. Sun, One-pot hydrothermal synthesis of RGO/CoFe2O4 composite and its excellent microwave absorption properties, Mater. Lett. 114 (2014) 52–55. X. Zhou, J. Lan, G. Liu, K. Deng, Y. Yang, G. Nie, J. Yu, L. Zhi, Facet-Mediated Photodegradation of Organic Dye over Hematite Architectures by Visible Light, Angew. Chem. 124 (2012) 182–186. Q. Yang, H. Choi, S.R. Al-Abed, D.D. Dionysiou, Iron–cobalt mixed oxide nanocatalysts: Heterogeneous peroxymonosulfate activation, cobalt leaching, and ferromagnetic properties for environmental applications, Appl. Catal. B 88 (2009) 462–469. D.E. Pennington, A. Haim, Stoichiometry and mechanism of the chromium (II)peroxydisulfate reaction, J. Am. Chem. Soc. 90 (1968) 3700–3704. E. Hayon, A. Treinin, J. Wilf, Electronic spectra, photochemistry, and autoxidation mechanism of the sulfite-bisulfite-pyrosulfite systems. SO2-, SO3-, SO4-, and SO5radicals, J. Am. Chem. Soc. 94 (1972) 47–57. T. An, H. Yang, W. Song, G. Li, H. Luo, W.J. Cooper, Mechanistic considerations for the advanced oxidation treatment of fluoroquinolone pharmaceutical compounds using tio2 heterogeneous catalysis, J. Phys. Chem. A 114 (2010) 2569–2575. M. Chen, W. Chu, Photocatalytic degradation and decomposition mechanism of fluoroquinolones norfloxacin over bismuth tungstate: experiment and mathematic model, Appl. Catal. B 168–169 (2015) 175–182. F. Minisci, A. Citterio, C. Giordano, Electron-transfer processes: peroxydisulfate, a useful and versatile reagent in organic chemistry, Acc. Chem. Res. 16 (1983) 27–32. Y. Ji, C. Ferronato, A. Salvador, X. Yang, J.-M. Chovelon, Degradation of ciprofloxacin and sulfamethoxazole by ferrous-activated persulfate: Implications for remediation of groundwater contaminated by antibiotics, Sci. Total Environ. 472 (2014) 800–808.