Enhanced degradation of triclosan by cobalt manganese spinel-type oxide activated peroxymonosulfate oxidation process via sulfate radicals and singlet oxygen: Mechanisms and intermediates identification

Journal Pre-proofs Enhanced degradation of triclosan by cobalt manganese spinel-type oxide activated peroxymonosulfate oxidation process via sulfate radicals and singlet oxygen: mechanisms and intermediates identification Zhiping Chen, Sijing Bi, Guangyi Zhao, Yuancai Chen, Yongyou Hu PII: DOI: Reference:

S0048-9697(19)34706-0 https://doi.org/10.1016/j.scitotenv.2019.134715 STOTEN 134715

To appear in:

Science of the Total Environment

Received Date: Revised Date: Accepted Date:

18 July 2019 16 September 2019 27 September 2019

Please cite this article as: Z. Chen, S. Bi, G. Zhao, Y. Chen, Y. Hu, Enhanced degradation of triclosan by cobalt manganese spinel-type oxide activated peroxymonosulfate oxidation process via sulfate radicals and singlet oxygen: mechanisms and intermediates identification, Science of the Total Environment (2019), doi: https://doi.org/10.1016/ j.scitotenv.2019.134715

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2019 Elsevier B.V. All rights reserved.

Enhanced degradation of triclosan by cobalt manganese spinel-type oxide activated peroxymonosulfate oxidation process via sulfate radicals and singlet oxygen: mechanisms and intermediates identification Zhiping Chen, Sijing Bi, Guangyi Zhao, Yuancai Chen*, Yongyou Hu

Ministry of Education Key Laboratory of Pollution Control and Ecological Remediation for Industrial Agglomeration Area, School of Environment and Energy, South China University of Technology, Guangzhou, 510006, China

*Corresponding

author’s E-mail: [email protected]; Phone: +86 13672458060; Address:

College of Environment and Energy, South China University of Technology, Guangzhou, Guangdong, 510006, China

Abstract Spinel is a kind of desirable catalyst to activate peroxymonosulfate (PMS) for chemical oxidation of organic contaminants in wastewater treatment. However, apart from classic sulfate radical based AOPs (SR-AOPs), the generation and oxidative pathways of singlet oxygen (1O2) by Co/Mn spinels have been little explored in PMS catalysis. In this study, spinel-type oxide Co2Mn1O4 was successfully synthesized, and used as highly effective catalyst in PMS activation for heterogeneous degradation of TCS (up to 96.4% within 30 min) at initial 1

pH of 6.8, which was also slightly impacted by coexisting ions. Based on radical scavengers and electron paramagnetic resonance (EPR) experiments, sulfate radicals and singlet oxygen (1O2) were unveiled to be the dominant reactive oxygen species (ROS) in Co2Mn1O4/PMS system. Co2Mn1O4 catalyst exhibited reversible redox properties based on the results of cyclic voltammetry (CV). More importantly, the generation of 1O2 might not only promote the TCS removal rate directly, but also facilitate the metal redox cycle in spinel structure in Co2Mn1O4/PMS system. Finally, degradation pathways of TCS in Co2Mn1O4/PMS system were proposed, which involved the breakage of ether bond and cycloaddition reaction. Keywords: Singlet oxygen; Peroxymonosulfate; Cobalt manganese spinel; Redox metal cycle;

2

1

Introduction Advanced oxidation processes based on sulfate radical (SO4•−) (SR–AOPs) have

received considerable attention as an effective technology to degrade contaminants in wastewater. Sulfate radical can be generated by the activation of peroxymonosulfate (PMS) and peroxydisulfate (PDS). It is a powerful oxidant with high redox potential (2.5 – 3.1 V), long half-life time (t1/2 = 30 – 40 μs), wide pH range, and fast reaction rate for contaminants (106 – 109 M-1 s-1) (Matzek and Carter, 2016; Oh et al., 2016). Recently, nonradical mechanism of PMS activated by organic compounds (Chen et al., 2018; Zhou et al., 2015), metal and carbon-based catalysts (Lee et al., 2015) has been proposed. Singlet oxygenation, electron transfer and surface activated complex have been postulated as three dominant reaction pathways for the nonradical mechanism (Duan et al., 2018). Different from radical reaction, nonradical reaction without radical trapping effect is slightly impacted by water matrix (e.g., halogen anions, natural organic matters). In particular, singlet oxygen (1O2) exhibits high redox capacity (2.2 V) (Zhu et al., 2018) and long lifetime (0.2 – 2 μs) (Salokhiddinov et al., 1981), and could react more selectively with electron-rich compounds (e.g. phenols) (Al-Nu'airat et al., 2018). Therefore, nonradical reaction is more suitable to target the priority pollutants at trace levels in the complicated water matrix, avoiding ineffectively consumption of nonselective radicals (e.g., SO4•−, •OH) (Yun et al, 2017). More importantly, the formation of 1O2 could facilitate the recycling of metal redox and optimize electron localization of metal oxides. Liu et al. (2018) demonstrated that the Bi5+/Bi3+ redox 3

couple of sillenite Bi25FeO40 could promote the generation of 1O2 from lattice oxygen. The multiple roles of 1O2 in the activation of PMS by carbon-based catalysts (e.g. reduced graphene oxide (Duan et al., 2016), graphitic carbon nitride (Gao et al., 2018) and N-doped carbon nanotubes frameworks (Duan et al., 2015)) have been clarified so far, but increasing efforts have been made upon metal-derived activator. It has been reported that noble metal nanoparticles supported on Al2O3 and TiO2 could activate PMS through nonradical reaction and transport electrons to co-adsorbed PMS on the catalyst surface (Ahn et al., 2016). Similar to noble metals, nonradical reaction was also found in traditional metals such as Mn, Fe, Cu (Ahn et al., 2019; Du et al., 2019; Fan et al., 2019; Guo et al., 2019; Li et al., 2019; Nie et al. 2017; Shao et al., 2019). Gao et al. (2019) confirmed perovskite-style oxides LaMO3 (M: Fe, Zn, Mn, Ni) could effectively activate PMS in the degradation of ofloxacin with 1O2 as the main reactive oxygen species (ROSs). Li et al. (2019) developed the perovskite-based nanocomposites L1.15FO as superior Fenton catalyst to catalytic oxidize methyl orange through a faster generation of singlet oxygen. Spinel-type transition metal oxides are of great significance in heterogeneous catalysts in PMS activation, due to their low cost, simple preparation, high stability and strong oxidation activity. In a typical spinel structure (AB2O4), A2+ ions occupy one eighth of the tetrahedral sites, while B3+ ions occupy one-half of the octahedral sites and O2- is closely packed in a face-centered cubic configuration (Reddy and Yun, 2016). The metal sites in crystal lattice can be substituted by different types of cations with close values of crystal field stabilization energies for octahedral and tetrahedral sites, which is the main reason for the 4

abundance defects and non-stoichiometry of spinel structures (Zhao et al., 2017). These characteristics are favorable to mediate metal active sites and redox properties of AB2O4 and facilitate catalytic activity in PMS activation. Among common transition metals (e.g. Co, Cu, Mn, Fe, etc.), Co exhibits the highest activity for PMS activation (Yao et al., 2015). However, the Co/PMS system may discharge a large number of cobalt ions with potential toxicity and carcinogenicity (Hu and Long, 2016). Co-based mixed-metal spinel CoxM2-xO4 (1 ≤ x ≤ 2, M =Zn, Cu, Fe, Mn, etc.) could effectively suppress metal leaching and manipulate the catalytic activity of cobalt (Wei et al., 2015). Furthermore, the redox couples between Co2+/Co3+ and Mn+/M(n+1)+ may further promote the charge transfer (Mathew et al. 2002). The formation of Co-O-M bond also optimizes and modulates the electron localization of cobalt (Zhou et al., 2018). Mn-based spinel possesses some special characteristics compared with other transition metals (e.g. Cu, Fe) due to its unique Mn2+/Mn3+/Mn4+ redox loops involving more electron transfer pathways. It has been demonstrated that Mn-Co based Prussian blue analogues (PBAs) catalyst showed better catalytic activity than Fe-Co-PBAs for activating PMS to decolorize Rhodamine B in water (Lin et al. 2016). Additionally, Mn(III) is labile and susceptible to disproportionation to Mn(II) and Mn(IV), which would greatly enhance oxygen mobility. Cobalt manganese spinel-type oxide (CoxMn2-xO4) was first employed for PMS activation in 2015 (Yao et al., 2015), exhibiting excellent activity due to the synergistic effects of Co and Mn species. Till now, CoxMn2-xO4 has been proven to efficiently degrade Rhodamine B (Yao et al., 2015), carbamazepine (Deng et al., 2017) and sulfanilamide (Li et al., 2018) through 5

the activation of PMS. These researches mainly focused on enhancing the catalytic performance of CoxMn2-xO4 by adjusting the proportion of Co/Mn dosage. Nevertheless, 1O2 also has been found to participate in PMS activation with some spinel-type oxides, such as CuCo2O4 (Wu et al, 2019), but few contributions were made to distinguish the role of singlet oxygen when radical and nonradical reactions coexisted. Thus, given that multi-reaction pathways (radical and nonradical) may be involved in CoxMn2-xO4/PMS system, a systematic understanding of reaction mechanism between cobalt manganese spinel-type oxide and PMS deserves further investigation, in particular for the role of singlet oxygen. In this study, CoxMn2-xO4 was successfully synthesized and utilized to catalyze PMS in the degradation of triclosan (5-chloro-2-(2,4-dichlorophenoxy) phenol, TCS), which was taken as model contamination in our experiments. The degradation of TCS was evaluated under various experiment conditions, such as PMS and catalyst dosage, initial pH and inorganic ions. More importantly, the critical role of singlet oxygen in CoxMn2-xO4/PMS was investigated, and a comprehensive reaction mechanism was proposed. Besides, the degradation pathways of TCS in CoxMn2-xO4/PMS system were also produced based on the intermediates detected by GC-MS. This work is expected to provide worthy reference to nonradical mechanism in spinel-type oxides in AOPs. 2

Materials and methods

2.1 Chemicals and reagents Cobalt nitrate (Co(NO3)2·6H2O, ≥99%) and triclosan (TCS, ≥99%) were purchased from Aladdin

(Shanghai,

China).

Oxone

(peroxymonosulfate, 6

PMS,

≥99%),

5,5-dimethyl-1-pyrroline N-oxide (DMPO, ≥97%), 4-amino-2,2,6,6-tetramethylpiperidine (TEMP, ≥98%), p-benzoquinone (p-BQ, ≥99%) and tetrahydro furfuryl alcohol (FFA, ≥98%) were purchased from Macklin (Shanghai, China). Other reagents were purchased from Shanghai Chemical Reagent Co., China. Deionized water was used throughout this work, and all chemicals and reagents used in this work were of analytical grade without further purification. 2.2 Synthesis of cobalt manganese spinel nanoparticles Cobalt manganese spinel nanoparticles were prepared through facile solution-based oxidation–precipitation and insertion–crystallization processes according to a previous research (Li et al., 2015). Typically, 4 mL aqueous ammonia (25 wt%) was added to 10 mL 0.2 mol L-1 Co(NO3)2 dropwise and mixed completely. Then, 5 mL 0.2 mol L-1 Mn(NO3)2 solution was added to the mixture and continuously stirred for 2 h. The resulting mixture was evaporated and nitrate was decomposed by heating at 180 °C for 8 h. Finally, the obtained black powder was washed several times with deionized water and methanol to remove the adsorbed chemicals on the surface of the catalysts, and then dried at 80°C for 12 h for further experiments. To make the description more readable, the synthesized catalysts were designated as Co2Mn1O4 (molar ratio Co/Mn = 2:1). According to this procedure, cobalt manganese spinel nanoparticles with different molar ratios (Co/Mn = 8:1, 4:1, 2:1, 1:2, 1:4, 1:8) were synthesized and characterized. The catalyst Co2Mn1O4 (Co/Mn = 2:1) exhibited the highest activity (Fig. S1) and was used for the subsequent experiments. For comparison, pure cobalt oxides (CoO) and manganese metal oxides (MnO) were prepared in the same procedure without the addition of Mn(NO3)2 or 7

Co(NO3)2, respectively. 2.3 Characterization of synthesized catalysts The morphology and chemical composition of catalysts were characterized by scanning electronic microscopy – energy dispersion spectroscopy (SEM-EDS). The powder X-ray diffraction (XRD, D8 ADVANCE, Bruker, USA) spectra of the obtained catalysts were collected with Cu-Kα radiation (λ = 1.5418 Å) at 40 kV and 40 mA operating from the 2θ angle of 10 – 80°. The specific surface area and pore size distribution of catalysts were calculated from the N2 adsorption/desorption isotherms at 77 K (ASAP2020, Micromeritics, USA). Raman scattering spectra were taken between 150 and 750 cm-1 at room temperature (DXR 2xi, Thermo Scientific Fisher, USA). X-ray photoelectron spectroscopy (XPS, ULVCA-PHI X-tool, Physical Electronics, USA) spectra were recorded by using Al-Kα radiation at 15 kV and 51 W, and the binding energies were calibrated with C1s at 284.8 eV. The concentration of leached metals was determined by graphite furnace atomic absorption spectrometry (AAS, Pin AAcle 900T, Perkin Elmer, USA). Cyclic voltammetry (CV) was measured with 0.01 v s-1 rate in 3 M NaOH electrolyte using a CHI-660b electrochemical workstation (Shanghai Chenhua Instruments Co.Ltd., China) with a three-electrochemical cell, including a working electrode of glassy carbon covered with catalysts, a counter electrode of platinum wire and a reference electrode of silver chloride electrode (Ag/AgCl). 2.4 Analysis methods TCS concentration was quantified on high-performance liquid chromatography (HPLC,

8

Agilent 1200) equipped with a PDA 2998 detector and a C-18 column (Agilent Eclipse, 5 μm, 4.6 mm × 250 mm). The mobile phase was acetonitrile/water (75:25, v/v) with a flow rate of 1.0 mL min-1. And the analytical wavelength and column temperature were set at 230 nm and 35 °C, respectively. The pHpzc was calculated through testing zeta potentials of catalyst at different pH with Zetasizer Nano Analyzer (Malvern, UK). The intermediates of TCS were identified by Gas Chromatograph-Mass Spectrometry (GC-MS) on a SHIMADZU GCMS-QP2010 Ultra instrument in EI mode using a full scan range of 20 – 580 m/z. The column temperature was initially set at 50 °C for 2 min, increased to 100 °C at a rate of 6 °C min–1 and then to 200 °C at a rate of 10 °C min–1. Afterward, the temperature was elevated to 280 °C at a rate of 20 °C min–1 followed by 15 min hold. The temperature of injector and ion source was set at 280 °C and 230 °C, respectively. Helium was used as carrier gas at a flow rate of 1 mL min–1. The samples used for GC-MS analysis were prepared as following (Aranami and Readman, 2007). After 30 min reaction, the 200 mL reaction solution was extracted with 20 mL dichloromethane three times. Afterward, the obtained dichloromethane was concentrated to 2 mL on a rotary evaporator and transferred to a GC vial for analysis. The Electron Paramagnetic Resonance (EPR) spectra were measured by Bruker EMX plus instrument (Germany) with DMPO and TEMP as the trapping reagent of SO4•− and 1O2, respectively. 2.5 Experimental procedure Batch experiments were performed in 250 mL conical flask with a magnetic stirrer at room temperature (25 °C). In a typical experiment, the required catalyst was firstly added into 200 mL of 10 mg L-1 TCS solution without pH adjustment. To avoid the effect of ions in buffer 9

solution on the degradation of TCS, the experiment in our work were conducted without adding buffer. Then, the specified concentration of PMS was injected into the obtained mixture to initiate reaction. 1 mL of aliquot was withdrawn at each time interval and filtered immediately by using a 0.22 μm glass fiber filter. Samples were quenched with excess methanol to inhibit further reaction and analyzed at once. In addition, the remained reacted catalyst was collected and washed with deionized water, then dried at 100 °C for further analyses. All experiments were conducted at duplicates, and the average values were used with standard deviations represented by error bars in corresponding graphs. 3

Results and discussion

3.1 Characterization of synthesized catalysts The surface morphology of synthesized catalysts was shown in Fig. 1 (a). CoO was composed of regular cube nanoparticles, while Co2Mn1O4 was spherical nanoparticles with an average diameter about 10 nm when manganese was introduced (Fig. 1 (b)). Whereas, the surface morphology of MnO displayed irregular flaky nanosheets (Fig. 1 (c)). The existence of metal Mn and Co was confirmed by the EDS spectra (Fig. 1 (d)-(f) and Fig. S2), and the measured composition ratios were in good agreement with the theoretic molar ratios (Table S1). The crystallographic structure of the synthesized catalysts was characterized by XRD (Fig. 2(a)), and their unit cell parameters (α0) were calculated (Table S1). The cobalt manganese oxide exhibited the characteristic peaks of CoxMn2-xO4 (JCPDS card No. 23-1237) at around 18.55o (111), 30.54o (220), 35.99o (311), 43.76o (440), 57.91o (511) and 63.62o (440), indicating the successful synthesis of cubic spinel catalyst Co2Mn1O4. Moreover, no other impurity peak 10

was observed in the XRD spectra of the Co2Mn1O4. For comparison, pure metal oxides of cobalt and manganese (CoO and MnO) were also fabricated and exhibited pronouncedly difference from Co2Mn1O4 in crystallinity (Fig. 1(a) and (c)). The average lattice constants (α0) of Co2Mn1O4 fell in the range between CoO and MnO (Table S1), as manganese ions with larger radius were incorporated into the octahedral sites of cobalt ions (ionic radius of Co2+-0.65 Å, Co3+- 0.61 Å, Mn2+-0.67 Å, Mn3+-0.65 Å, Mn4+-0.53 Å). The incorporation would cause the discordance of metal valence states and partial disorder of spinel structure (Li et al., 2009; Naveen and Selladurai, 2014; Sadighi et al., 2017). Raman spectrum, a sensitive method to detect minor amounts of impurity phases and lattice distortions, was conducted to verify the spinel structure of Co2Mn1O4. Generally, five theoretical Raman-active modes (A1g + Eg + 3T2g) can be theoretically perceived in the cubic spinel with the space group Fd-3m (Oh). And the high frequency region above 570 cm−1 could be assigned to vibrational modes involving the motion of oxygen atoms within the MO6 octahedral (Li et al, 2011). As shown in Fig. S3, compared to CoO (678 cm−1) and MnO (640 cm−1), the high bond of Co2Mn1O4 exhibited obviously lower intensity and shifted towards lower wavenumber (590 cm−1), which could be attributed to the partial replacement of Co2+ by lighter Mn2+ (Bouchard and Gambardella, 2011; Chandramohan et al, 2011). The peak centered at around 470 cm-1 could be assigned as Co-O bond in the tetrahedral sites (Bouchard and Gambardella, 2011; Malavasi et al, 2002), and the middle frequency region from 300 to 400 cm−1 was probably related to oxygen modes within MnO6 and/or MnO4 sites. Because only MnO exhibited vibrational bands in this region, whereas CoO exhibited no bands here. The 11

bonds at low frequencies (about 250 cm−1) were due to metal-oxygen bond of octahedral void (MO6)(i.e., Eg and T2g). However, the lower band at 190 cm-1 of CoO catalyst assigned to F2g was not observed in Raman spectra of Co2Mn1O4 (Bijelić et al, 2015). The N2 adsorption−desorption isotherms of synthesized catalysts were all belonged to type IV with a distinct hysteresis loop, which was typical for mesoporous inorganic oxides with good pore connectivity (Fig. 2 (b)). On the pore size distribution curves, the narrow peaks of Co2Mn1O4 and CoO centered at about 5 and 10 nm, respectively, indicating the uniform distribution of mesopores. And the broader peak of MnO displayed less uniformity of particles. The surface area of Co2Mn1O4 was 110.1 m2 g−1, largely exceeding CoO (56.72 m2 g−1) and MnO (81.48 m2 g−1). 3.2 Catalytic performance of Co2Mn1O4 3.2.1

Effective catalytic degradation of TCS by Co2Mn1O4 The optimal dosages of catalyst (0.02 g L−1) and PMS (0.05 g L-1) were first determined

according to TCS degradation and used in all the experiments without pH adjustment (the initial pH of 10 mg L-1 TCS is about 6.8). As shown in Fig. S4, no degradation of TCS was observed (below 5 % TCS in 30 min) in the presence of PMS or Co2Mn1O4 alone, indicating the adsorption of Co2Mn1O4 and the oxidation by PMS itself could be ignored during the whole experiment. The removal efficiency of TCS could reach 38.7 % and 31.3 % within 30 min in CoO/PMS and MnO/PMS systems, respectively, indicating that CoO and MnO could also activate PMS to some extent. Only when 0.05 g L-1 PMS and 0.02 g L-1 Co2Mn1O4 were added together (Co2Mn1O4/PMS system), the removal of TCS reached up to 96.4 % within 30 min, 12

exhibiting superior catalytic performance among metal oxide catalysts (Table S2). Meanwhile, the activated experiment of PMS was also performed with the simply mechanical mixing of CoO and MnO (the same ratio of 2:1 as Co2Mn1O4), and the much lower catalytic activity (29.7 % in 30 min) than Co2Mn1O4 was obtained at the same condition, indicating the special lattice structure of spinel oxides might promote the catalytic oxidation of TCS (Huang et al. 2018). In addition, it was observed that about 0.87 mg L-1 cobalt and 0.49 mg L-1 manganese were leached in Co2Mn1O4/PMS system after reaction. According to the “Emission standard of pollutants for copper, nickel and cobalt industry” of China, total cobalt emission concentration must be less than 1.0mg/L, so the Co leaching in our work meets the standards. To test the durability of catalyst, a continuous utilization experiment of Co2Mn1O4 was conducted. 2 mL 1 g L-1 TCS and 1 mL 10 g L-1 PMS were added simultaneously into mixture at 0, 60 and 120 min. As depicted in Fig. S5, the removal efficiencies of TCS have declined from 100 % (60 min) to 92.0 % (120 min) and 78.8 % (180 min). The slight decrease of TCS removal in Co2Mn1O4/PMS system was possibly ascribed to the loss of active sites on Co2Mn1O4 surface and the accumulation of organic intermediates. To further determine the stability of catalyst, XRD spectra of Co2Mn1O4 before and after reaction (Fig. S6 (a)) were measured and it could be observed that Co2Mn1O4 still maintained the same spinel crystal structure after reaction. Raman active modes of the used catalyst shown in Fig. S6 (b) suggested no obvious shift was observed after reaction. Nevertheless, the intensity of Raman spectra increased and A1g peak slightly moved toward the higher Raman shifts (at around 600 cm-1) compared to fresh catalyst (590 cm-1), which might be attributed to the decrement of lattice oxygen bonded with metal ions. 13

This result is coincided with XPS results below (Section 3.4). 3.2.2

Kinetics As shown in Fig. 3 (a) and (b), all degradations of TCS followed the pseudo-first order

reaction model (R2 ＞ 0.997) with different PMS dosage and catalyst loading. Moreover, the observed degradation rate of TCS was significantly improved from 0.0201 to 0.2189 min-1 with increasing PMS dosage (0.01 – 0.1 g L-1), as the increasing PMS dosage could activate more reaction sites to attack TCS molecules (Fig. S7 (a)). In a low Co2Mn1O4 dosage (0.01 g L-1), only 78.2% of TCS was degraded within 30 min (k = 0.0742 min-1), and as the dosage increased to 0.1g/L, TCS could be completely removed within 30 min (k = 0.5772 min-1) (Fig. S7 (b)). Kinetic was also employed to model the TCS degradation without pH adjustment (initial pH at about 6.8). Because the reaction involves three species (i.e., PMS, TCS and catalyst Co2Mn1O4), the overall reaction rate of TCS degradation (r) can be expressed in Eq. (1), where r, n1, n2 and n3 represent the overall apparent reaction rate constant, and the reaction orders related to dosage of PMS, TCS and Co2Mn1O4 catalyst, respectively. As shown in Fig. 3 (a), a linear correlation between ln(kobs) of TCS and ln([PMS]0) with a slope approximating 1.078. It was also found that the decomposition of PMS was independent of the TCS concentration (Fig. 3 (c)), so the n2 is equal to 0 and the kinetic model could be expressed as Eq. (2). Likewise, the ln(kobs) of PMS decomposition during TCS degradation was linearly correlated to ln([Co2Mn1O4]0) with a slope of 0.511 (insert graph in Fig. 3 (a)). Based on the above experiment results, 1.078 and 0.511 could be assigned to n1 and n3 respectively (Eq. (3)). The similar linear phenomena were also reported in prior studies that CuO activated PDS to degrade 14

2,4-dichlorophenol (2,4-DCP) (Zhang et al., 2014) and CuCo2S4 spinel activated PMS to degrade bisphenol S (Xu et al., 2018). It is clear that the dosages of PMS and Co2Mn1O4 have significant influence on the reaction efficiency, in particular, especially PMS concentration. The overall reaction rate has an order of 1.589 (i.e., n1+ n2 + n3). (1) (2) (3) 3.2.3

Effect of pH The influence of initial pH was presented in Fig. 4. The degradation of TCS displayed

comparable removal rate in the pH range of 5 – 10 and attained the highest degradation rate (0.112 min-1) at the neutral condition (pH = 7). However, extremely acid or alkaline environment has an inhibitory effect on TCS removal. The effect of pH on TCS removal could be explained by properties of catalyst surface, TCS and PMS. As shown in Fig. 4 (b), the pKa value of TCS is 7.9 and thus it mainly existed in its anion form (TCS−) over pH 8 by losing hydrogen (Lindstrom et al., 2002). While the pKa1 and pKa2 of PMS are nearly 0 and 9.4 (Betterton and Hoffmann, 1990), hence PMS mainly existed in mono-anion form (HSO5−) at the pH range of 3 – 9 and dianion form (SO52−) at pH > 9.4 (Fig. 4 (b)). Moreover, the pHpzc of CoMn2O4 was determined to be 7.17. Although the k value tended to increase with the increasing molar fraction of TCS− and SO52−, the electrostatic repulsion between negatively charged catalyst surface and SO52−, TCS- might become more obvious in alkaline solution (Yao et al., 2017). Moreover, the cobalt and manganese hydroxide complexes would be generated 15

over pH 9.0, which would further hinder the catalytic activity of CoMn2O4. When it came to acid condition, the decline of k was quite apparent on account of the collapse of cubic spinel structure for the catalyst CoMn2O4 (Oh et al., 2015). As no buffer was added in this work, the pH of solution decreased obviously after reaction as shown in Fig 4 (b). 3.2.4

Effect of co-existing anions Inorganic ions (i.e. Ca2+, Mg2+, Cl−, SO42- and HCO3-) are ubiquitous in natural

environment, so the influence of different inorganic ions on TCS catalytic degradation was investigated (Fig. 5 and Fig. S8). Fig. 5 (a) depicted the effect of Cl− on catalytic oxidation of TCS, and it was found that the presence of Cl− ranging from 0 to 5 mM didn’t significantly affect the removal of TCS. In addition, the value k of TCS degradation even increased from 0.106 min−1 to 0.161 min−1 with the addition of 10 mM Cl−, indicating the accelerating effect of Cl− on the removal of TCS. It was attributed to the fact that Cl− would be directly oxidized to Cl2 and HOCl by PMS through two-electron transfer (Eq. (4) and (5)). These producing chlorine species (Cl2 and HOCl) would further facilitate the catalytic reaction. From Fig. 5(b) and (c), the presence of NO3− and SO42− exhibited slight inhibitory on TCS degradation. Likewise, as the most common cations in water matrix, Ca2+ and Mg2+ both exerted little impact on TCS degradation (Fig. 5 (e) and (f)). With 10 mM Ca2+ and Mg2+ adding, the TCS removal efficiencies decreased from 96.4 % to 82.3 % and 84.9 %, and corresponding k value decreased from 0.106 min−1 to 0.048 min−1 and 0.066 min−1 in 30 min. It is owing to the decreased pH from 6.8 to 5.33 and 5.89, respectively. However, the degradation of TCS was reduced dramatically in the presence of HCO3−. With the increase of HCO3− from 0 to 10 mM, the 16

removal efficiencies of TCS sharply decreased from 96.4 % to 33.0 % (Fig. 5(d)). Meanwhile, the solution pH increased from 6.8 to 10.5, indicating pH has a great influence on the catalytic oxidation of TCS in Co2Mn1O4/PMS system. Furthermore, the distinct inhibitory of HCO3− was also resultant from the strong quenching effect of HCO3−/CO32− for SO4•− and 1O2 (the generation of 1O2 was explained in Section 3.3) (Eq.(6)) ( Huie and Clifton, 1990). On the whole, inorganic anions co-exiting in the water have little influence on the catalytic degradation of TCS in Co2Mn1O4/PMS system, except for HCO3−. This result suggested that nonradical reaction was less affected by co-existing anions compared to traditional radicals and provide contribution for TCS degradation in Co2Mn1O4/PMS system. HSO4− + 2Cl− + H+ → Cl2 + SO42− + H2O

(4)

HSO4− + Cl− → HOCl + SO42−

(5)

SO4•− + HCO3− → HCO3• + SO42−

k = (3.11±1.47) × 109 M-1 s-1

(6)

3.3 Identification of reactive oxygen species in Co2Mn1O4/PMS system 3.3.1

Radical quenching study In order to identify ROSs in the Co2Mn1O4/PMS system, radical quenching experiments

were conducted using different scavengers at 1 mM concentration. It was previously reported that SO4•− and HO• were usually considered as reactive radicals in the PMS activation (Buxton et al., 1988; Du et al., 2017). Owing to the different reaction rate with radicals, methanol (MeOH) is usually applied to scavenge both SO4•− and HO• (k(SO4•−,MeOH) = 2.5 × 107 M−1 s−1; k(HO•, MeOH) = 9.7 × 108 M−1 s−1 ), while tert-butyl-alcohol (TBA) is used as a chemical probe for HO• but not for SO4•− (k(HO•, TBA) = 3.8 – 7.6 × 108 M−1 s−1; k(SO4•−,TBA) = 4 – 9.5 × 105 M−1 s−1) 17

(Buxton et al., 1988; Du et al., 2017). Fig. 6 showed that the TCS removal efficiencies decreased from 96.4% to 93.7% and 70.7% after 30 min with the addition of TBA and MeOH, respectively. And the removal of TCS was depressed to 35.0 % as MeOH concentration increased to 100 mmol L-1, while the increasing of TBA has slight influence on the degradation rate of TCS (Fig. S9), indicating that SO4•− mediated ROSs was involved in the degradation of TCS and the influence of •OH (Eq. (7)) was minor. However, the removal efficiencies of TCS (35.0 % − 70.7 %) were still be obtained with the SO4•−scavenger (MeOH), suggesting the participation of other ROSs in TCS degradation. Therefore, we employed furfuryl alcohol (FFA) to further quench the oxidation by 1O2 ( k(FFA, 1O2) = 1.2 × 108 M−1 s−1 (Zhu et al., 2018), and sodium azide (NaN3) to scavenge 1O2 as well as SO4•− (k(NaN3, 1O2) = 3.2 × 107 M−1 s−1 (Yang et al., 2018), k(NaN3, SO4•−) = 2.5 × 109 M−1 s−1 (RE Huie, 1990)). As shown in Fig. 6, the addition of FFA and NaN3 both caused a noticeable decline in the degradation of TCS, which sharply decreased from 96.4% to 68.6 % and 24.98 % after 30 min (Fig. 6 (b)), confirming that 1O2 and SO4•− were the main reactive ROSs involved in the reaction. Some studies reported that 1O2 might be generated from the recombination of superoxide radicals (HO2•/O2•−) (Eq. (8) and (9)) (Yang et al., 2018). In our experiment, p-BQ as the scavenger of HO2•/O2•− had a negligible impact on TCS oxidation, so we deduced that neither HO2• nor O2•− was the main source of 1O2. Lee et al. (2011) reported that 1O2 might be produced from oxygen in the air, thus further experiments were conducted under the N2 and O2 atmosphere respectively (Fig. S10). However, no significant difference was observed under N2 18

or O2 flow, indicating that 1O2 was not produced from the oxygen molecules in the air, which corroborated the findings of previous studies (Li et al., 2019; Liu et al., 2018). The generation mechanism for singlet oxygen would be further discussed in Section 3.4.

3.3.2

SO4•− + H2O → •OH + SO42− + H+

(7)

HO2• + O2•− → 1O2

(8)

2O2•− + 2H+ → 1O2 + H2O2

(9)

ESR The ROSs were further clarified by electron spin resonance (ESR) spectroscopy using

TEMP and DMPO as spin-trapping agents of 1O2 and OH•, SO4•−, respectively. As observed in Fig. 7 (a), both characteristic signals of DMPO-OH (αN = αH = 14.9) and DMPO-SO4 (αN = 13.2 G, αH = 9.6 G, αH = 1.48 G and αH = 0.78 G) appeared in the ESR spectra. According to the literatures (Guo et al., 2019; Kang et al., 2019; Luo et al., 2019a; Yang and Che, 2017; Zhang et al., 2016), the appearance of DMPO-SO4 adducts signal was usually accompanied by the signal of DMPO-OH, and DMPO-SO4 signal was hard to be detected alone in aquatic solution. Besides, the signal intensity of adduct DMPO-OH was much higher than that of adduct DMPO-SO4, possibly due to the fast transformation of DMPO-SO4 to DMPO-OH via nucleophilic substitution (Fig. S11) (Timmins et al., 1999). In addition, the characteristic signals with an intensity ratio of 1:2:1:2:1:2:1 of DMPOX (αN = 7.3, αH = 3.9) appeared (Fig. 7 (a)) due to the oxidation of DMPO by SO4•− with the strong oxidizing ability. More importantly, the intense three-line signals (1:1:1) of TEMP-1O2 adducts (α = 16.9 G) by 1O2 oxidation were 19

also observed in Fig. 7 (b), further confirming the generation of 1O2 (Guo et al., 2019; Luo et al., 2019b; Tian et al., 2017). Moreover, the signals of TEMP-1O2 became stronger with the addition of Co2Mn1O4, indicating that catalyst Co2Mn1O4 could fortify the generation of 1O2. Herein, both the radical quenching and EPR experiment verified that SO4•− and 1O2 were generated and served as the main ROSs in Co2Mn1O4/PMS system. 3.4 Reaction mechanism in the Co2Mn1O4/PMS system To gain further insight into the reaction mechanism in the Co2Mn1O4/PMS system, XPS was performed to record the surface elemental valence and composition of catalysts before and after reactions. Fig. 8 showed that the elements of Co, Mn and O coexisted in the Co2Mn1O4 sample, and the binding energies and relative intensities are summarized in Table S3 based on the deconvolution of Co 2p, Mn 2p and O 1s XPS spectra. The Co 2p spectra were composed of spin-orbit doublets of Co 2p3/2 and Co 2p1/2 with two additional satellite peaks at the binding energy (B.E.) of 780.0 and 794.5 eV, respectively, which could be deconvoluted into Co2+ and Co3+ (Zhang et al., 2018). After catalytic reaction, the intensity ratio of CoII/CoIII increased from 0.821 to 1.448, implying a much higher CoII concentration in used catalyst (Fig. 8 (a)). As for the deconvoluted Mn 2p spectra, there were three oxidation states at 640.9 eV (MnII), 642.2 eV (MnIII) and 643.9 eV (MnIV), respectively. The ratio of MnII decreased from 25.05 % to 19.07 %, while MnIII and MnIV increased from 55.72 % and 19.23 % to 57.79 % and 22.44 %, respectively (Fig. 8 (b)). The average oxidation state of surface Mn increased from 3.1 to 3.3 (Table S3), indicating that catalyst after reaction was enriched with metal ions of higher chemical state. Interestingly, although CoⅡ and MnⅡ 20

ions donated electrons to PMS to produce sulfate radicals, the ratio of CoII still increased after the reaction, indicating there might be recycling of metal redox in the Co2Mn1O4 spinel. The O 1s spectra (Fig. 8 (c)) were resolved into three individual peaks at 529.8 eV, 531.0 eV and 531.9 eV, corresponding to the lattice oxygen (Olatt) in the spinel structure, the surface hydroxyl species (OOH) and the adsorbed oxygen (Oads), respectively (Zhuangpeng Huang, 2018). The concentration of oxygen vacancy (V0) was positively correlated with the intensity ratio of OOH to Olatt (OOH/Olatt) with the discrepancy of around 1.1 eV. It was noted that OOH/Olatt increased from 0.367 to 0.573 after reaction (Table S3), confirming that more oxygen vacancies were generated during catalytic reaction and lattice oxygen was involved in the generation of 1O2. Moreover, excess electrons were induced by oxygen vacancies to promote the reduction of metal ions with high valance (Li et al., 2019; Liu et al., 2018). The minor increase of Oads (from 11.31 % to 21.85 %) may be ascribed to the adsorbed oxygen and intermediate products on the surface of catalysts (Li et al., 2019; Huang et al, 2018). To gain further insight into a synergy effect between Co and Mn in the Co2Mn1O4, cyclic voltammetry (CV) was performed to record redox behaviors of as-prepared materials in this work. As seen in the Fig. S13, the cathodic and anodic peaks of CoO at -0.10 V and -0.58 V corresponded to Co2+/Co3+ redox couple, and two oxidative peaks at -0.21 V and -0.40 V for MnO catalyst were assigned as Mn2+/Mn3+/Mn4+ couples. For Co2Mn1O4 catalyst, a weak oxidation shoulder at -0.12 V and a reduction shoulder at -0.41 V were observed. The higher current density and less potential discrepancy between redox couple suggested that the catalyst 21

Co2Mn1O4 combined the advantages of both CoO and MnO, and thus possessed greater capability to accelerate electron transfer. Additionally, their redox peaks became more symmetric compared to CoO, indicating that the interaction of cobalt and manganese could improve the reversibility of cobalt ions. All these would further improve the catalytic activity in PMS activation (Huang et al., 2018). Based on the above results and previous researches (Li et al., 2019; Liu et al., 2018; Yao et al., 2015; Huang et al., 2018), it is reasonable to propose activation mechanism in Co2Mn1O4/PMS system. As shown in Fig. 9, the SO4•− radical-based catalytic process and non-radical 1O2 oxidation process were both involved in Co2Mn1O4/PMS system. For radical reactions, metal ions (M: Co, Mn) on the surface of catalyst served as Lewis acid sites could conduct one electron to activate PMS with the generation of SO4•− (Eq. (10) and (11)) (Deng et al., 2017). Furthermore, the charge transfer between the redox couples of CoⅡ/CoⅢ and MnⅡ/MnⅢ/MnⅣ in the mixed metal oxides may further accelerate the electron transfer of PMS activation based on the CV and XPS analyses (Hu and Long, 2016). With the higher oxidation states (E(CoⅡ/CoⅢ) =-1.82v, E(MnⅢ/MnⅣ)=-0.95v, E(MnⅡ/MnⅢ) =-1.51v), manganese cations in octahedral site is thermodynamically favorable to reduce CoⅢ to CoⅡ (Eq. (12) and (13)) (Ni et al., 2018; Yao et al., 2015). The synergistic effect between the manganese and cobalt ions promoted the generation of SO4•−. On the other hand, non-radical oxidation process by 1O2 also promoted the rate of reaction. As the eg orbital of transition-metal exhibits a strongly spatial overlap with O2p orbital (Zhao et al., 2017; Zhou et al., 2018), the 22

formation of M-O-M (M: Co, Mn) bond would optimize electron localization of metal ions. Moreover, manganese ions with larger ionic radius (Deng et al., 2017; Naveen and Selladurai, 2014; Yao et al., 2015) substituted the cobalt ions in the tetrahedral sites, causing the electronic structure of active center tunable (Naveen and Selladurai, 2014) and generating vacancies on the O sites (Eq. (14)). Meanwhile, some active oxygen *O might further transform to 1O2 (Eq.(15) and the generation of oxygen vacancies facilitated the redox couple of Mn+/M(n+1)+ (M: Co,Mn) in the crystal lattice (Eq.(16)) (Li et al., 2019; Liu et al., 2018; Lu et al., 2018). The blue-shifting of the peaks in Co 2p and Mn 2p spectra (Fig. S12) further demonstrated the weaken metal-oxygen bonds and the generation of oxygen vacancy (Sadighi et al., 2017). The more oxygen vacancies are in favor of interfacial electron transfer, which was also observed in other researches (Lu et al., 2018; Ren et al., 2012; Sadighi et al., 2017; Huang et al., 2018; Gao et al. 2019; Li et al., 2019). Thus, the oxygen defects of catalyst interface might play an important role in the generation of 1O2. ≡Mn+ + HSO5– → ≡M(n+1) + SO4•− + OH–

(10)

≡M(n+1) + HSO5– → ≡Mn+ + SO5•− + H+

(11)

≡MnⅡ + ≡CoⅢ → ≡MnⅢ + ≡CoⅡ

△E = 0.30 V

(12)

≡MnⅢ + ≡CoⅢ → ≡MnⅣ + ≡CoⅡ

△E = 1.66 V

(13)

(14) HSO5– + *O → HSO4– + 1O2 23

(15)

2≡M(n+1)+ + O2− → 2≡M(n+1)+ +

+ O*→ 2≡Mn+ + O*

HSO5– + SO52– → HSO4– + SO42– + 1O2

(16) (17)

3.5 Proposed degradation pathway in the Co2Mn1O4/PMS system A total of five compounds including four intermediates were detected by GC-MS (Fig. S14 and Fig. S15) and summarized in Table S4. In the light of the products identified, plausible degradation pathways of TCS by Co2Mn1O4/PMS oxidation were proposed in Fig.10, involving the breakage of ether bond and cycloaddition reaction. With respect to the first degradation pathway, the ether bond of TCS was first attacked by ROSs to generate 3-chlorophenol and 2,4-dichlorophenol (2,4-DCP). It should be noted that 2-chloro-1,4-benzoquinone could be formed from 3-chlorophenol through cycloaddition reaction. Thus, the detection of 2-chloro-1,4-benzoquinone in this study provides compelling evidence for the occurrence of 1O2 reaction. The primary product 2,4-DCP (retention time 13.09 min) with two isomers (2,6-DCP and 3,5-DCP) has also been detected as the oxidative product of TCS in many studies, such as the oxidative treatments via laccase (Dou et al., 2018), manganese oxides (Zhang and Huang, 2003) and ferrate (Yang et al., 2011), photocatalysis (Aranami and Readman, 2007; Yu et al., 2006) as well as in the biodegradation of TCS (Lee et al., 2012). In the second degradation pathway, singlet oxygen might attack the aromatic ring of TCS to generate endoperoxide, 2-chloro-5-(2,4-dichlorophenoxy)-2,5-cyclohexadiene-1,4-dione, and subsequently converted to 2-chloro-5-(2,4-dichlorophenoxy)-1,4-benzenediol in the solution. Then these endoperoxides were further oxidized to 2,4-DCP, 2-chloro-1,4-benzoquinone and 24

2-chloro-1,4-hydroquinone through the breakage of the ether bond. Finally, with the combined actions of sulfate radicals and singlet oxygen, the intermediate products were further oxidized to small molecule acids (including acetic acid, acrylic acid, maleic acid), which were mineralized to H2O and CO2 ultimately. TCS is classified as a very toxic organic according to the Globally Harmonized System of Classification and Labeling of Chemicals. The EC50 (the concentration for 50% of maximal effect) of TCS for fish, daphnid and green algae were 0.35, 0.39 and 0.765 mg L-1 respectively (Luo et al., 2019). While the acute and chronic toxicity of intermediates detected in this study, i.e., 2,4-DCP, 2-chloro-1,4-benzoquinone and 2-chloro-1,4-hydroquinone , were all lower than that of TCS (Gao et al. 2014; Peng et al. 2019). Thus, we can deduce that TCS could be efficiently removed in Co2Mn1O4/PMS system and its toxicity decreased during the reaction, and no more toxic organic was generated in reaction process. 4

Conclusion In this work, a comprehensive study was performed to investigate PMS activation via

Co2Mn1O4 spinel. Co2Mn1O4 exhibits excellent catalytic activity for PMS activation to degrade TCS in water (96.4 % TCS in 30 min with 0.05 g L-1 PMS and 0.02 g L-1 catalyst dosage) and the kinetic of TCS degradation fitted well with the first-order kinetic. In addition to SO4•−, 1O2 was also unveiled be to the primary reactive oxygen species in Co2Mn1O4/PMS system and the significant role of 1O2 was explored by multiple approaches such as chemical probes, EPR capture, CV and XPS. Based on these results, the comprehensive reaction mechanism was proposed. Heterogeneous metal redox cycle and nonradical reaction coexisted in the 25

Co2Mn1O4/PMS system. The redox couples between CoⅡ/CoⅢ and MnⅡ/MnⅢ/MnⅣ may further promote the electrons transfer and the formation of M-O-M (M: Co, Mn) bond could modulate electronic distribution of metal ions. Simultaneously, the generation of singlet oxygen from lattice oxygen would facilitate the metal redox cycle, which further promoted heterogeneous PMS activation in TCS oxidation. Furthermore, the degradation pathway of TCS involving the breakage of ether bond and cycloaddition reaction was propounded based on the detected intermediates, and 2-chloro-1,4-benzoquinone was first detected in PMS catalytic degradation by GC-MS. This work deepened the understanding of nonradical mechanism in spinel-type oxides activating PMS in AOPs, and further explored the degradation pathways of TCS by SO4•− and 1O2. Acknowledgements The research was financially supported by grants from National Nature Science Foundation of China (41977316, 21677052), Guangdong Water Conservancy Science and Technology Innovation Project (2017-25), Guangdong technological innovation strategy of special funds (key areas of research and development program, grant no.: 2018B020205003), Chinese Central Government for Basic Scientific Research Operations in Commonweal Research Institutes (PM-zx703-201602-044).

26

References Ahn Y.Y., Yun E-T., Seo J-W., Lee C., Kim S.H., Kim J.H., Lee J.. 2016. Activation of Peroxymonosulfate by Surface-Loaded Noble Metal Nanoparticles for Oxidative Degradation of Organic Compounds. Environ Sci Technol. 50, 10187-97. https://doi.org/10.1021/acs.est.6b02841 Ahn Y.Y., Bae H., Kim H-I, Kim J.H., Kim S.H., Lee S.G., Lee J.. 2019. Surface-loaded metal nanoparticles for peroxymonosulfate activation: Efficiency and mechanism reconnaissance. Appl. Catal. B Environ. 241, 561-569. https://doi.org/10.1016/j.apcatb.2018.09.056 Al-Nu'airat J., Dlugogorski B.Z., Gao X., Zeinali N., Skut J., Westmoreland P.R., Oluwoye I, Altarawneh M. 2019. Reaction of phenol with singlet oxygen. Phys. Chem. Chem. Phys. 21, 171-183. https://doi.org/10.1039/c8cp04852e Aranami K., Readman J.W.. 2007. Photolytic degradation of triclosan in freshwater and seawater. Chemosphere. 66, 1052-6. https://doi.org/10.1016/j.chemosphere.2006.07.010 Betterton E.A., Hoffmann M.R., 1990. Kinetics and Mechanism of the Oxidation of Aqueous Hydrogen Sulfide by Peroxymonosulfate. Environ. Sci. Technoi. 24, 1819-1824 https://doi.org/10.1021/es00082a005 Bijelić M., Liu X., Sun Q., Djurišić A. B., Xie M., Alan M. C. Ng, Suchomski C., Djerdj I., Skoko Željko, Popović J.. 2015. Long cycle life of CoMn2O4 lithium ion battery anodes with high crystallinity. J. Mater. Chem. A. 3, 14759-14767. https://doi.org/10.1039/C5TA03570H Bouchard M., Gambardella A., 2011. Raman microscopy study of synthetic cobalt blue spinels used in the field of art, J. Raman Spectro. 41. https://doi.org/10.1002/jrs.2645 Buxton G.V., Greenstock C.L., Helman W.P., Ross A.B., 1988. Critical Review of rate constants for reactions of hydrated electrons, hydrogen atoms and hydroxyl radicals (⋅OH/⋅O−) in Aqueous Solution. J. Phys. Chem. Refer. Data. 17, 513-886. https://doi.org/10.1063/1.555805 Chandramohan P., Srinivasan M. P., Velmurugan S., Narasimhan S. V., 2011. Cation distribution and particle size effect on Raman spectrum of CoFe2O4, J. Solid State Chem. 184, 89-96. 27

https://doi.org/10.1016/j.jssc.2010.10.019 Chen J.B., Fang C., Xia W.J., Huang T.Y., Huang Ch.H., 2018. Selective Transformation of β‑Lactam Antibiotics by Peroxymonosulfate: Reaction Kinetics and Nonradical Mechanism, Environ Sci Technol. 52, 1461-1470. https://doi.org/10.1021/acs.est.7b05543 Chen J.B., Zhang L.M., Huang T., Li W.W., Wang Y., Wang Z.M., 2016. Decolorization of azo dye by peroxymonosulfate activated by carbon nanotube: Radical versus non-radical mechanism. J. Hazard. Mater. 320, 571-580. https://dx.doi.org/10.1016/j.jhzmat.2016.07.038 Deng J., Cheng Y., Lu Y., Crittenden J.C., Zhou S., Gao N., Li J., 2017. Mesoporous manganese Cobaltite nanocages as effective and reusable heterogeneous peroxymonosulfate activators for Carbamazepine degradation. Chem. Eng. J. 330, 505-517. https://doi.org/10.1016/j.cej.2017.07.149 Dou R., Wang J., Chen Y., Hu Y., 2018. The transformation of triclosan by laccase: Effect of humic acid on the reaction kinetics, products and pathway. Environ. Pollut. 234, 88-95. https://doi.org/10.1016/j.envpol.2017.10.119 Du X., Zhang Y., Hussain I., Huang S., Huang W., 2017. Insight into reactive oxygen species in persulfate activation with copper oxide: Activated persulfate and trace radicals. Chem. Eng. J. 313, 1023-1032. https://doi.org/10.1016/j.cej.2016.10.138 Du X., Zhang Y., Si F., Yao C., Du M., Hussain I., Kim J., Huang S., Lin Z., Hayat W., 2019. Persulfate non-radical activation by nano-CuO for efficient removal of chlorinated organic compounds: Reduced graphene oxide-assisted and CuO (0 0 1) facet-dependent. Chem. Eng. J. 356, 178-189. https://doi.org/10.1016/j.cej.2018.08.216 Duan X., Ao Z., Sun H., Zhou L., Wang G., Wang S., 2015. Insights into N-doping in single-walled carbon nanotubes for enhanced activation of superoxides: a mechanistic study. Chem. Commun. (Camb) 51, 15249-52. https://doi.org/10.1039/C5CC05101K Duan X., Ao Z., Zhou L., Sun H., Wang G., Wang S., 2016. Occurrence of radical and nonradical pathways from carbocatalysts for aqueous and nonaqueous catalytic oxidation. Appl. Catal. B Environ. 188, 98-105. 28

https://doi.org/10.1016/j.apcatb.2016.01.059 Duan X., Sun H., Shao Z., Wang S., 2018. Nonradical reactions in environmental remediation processes: Uncertainty and challenges. Appl. Catal. B Environ. 224, 973-982. https://doi.org/10.1016/j.apcatb.2017.11.051 Fan J., Qin H., Jiang S., 2019. Mn-doped g-C3N4 composite to activate peroxymonosulfate for acetaminophen degradation: The role of superoxide anion and singlet oxygen. Chem. Eng. J. 359, 723-732. https://doi.org/10.1016/j.cej.2018.11.165 Gao P., Tian X., Nie Y., Yang C., Zhou Z., Wang Y., 2019. Promoted peroxymonosulfate activation into singlet oxygen over perovskite for ofloxacin degradation by controlling the oxygen defect concentration. Chem. Eng. J. 359, 828-839. https://doi.org/10.1016/j.cej.2018.11.184 Gao Y., Zhu Y., Lyu L., Zeng Q., Xing X., Hu C., 2018. Electronic Structure Modulation of Graphitic Carbon Nitride by Oxygen Doping for Enhanced Catalytic Degradation of Organic Pollutants through Peroxymonosulfate Activation. Environ. Sci. Technol. 52, 14371-14380. https://doi.org/10.1021/acs.est.8b05246 Gao Y., Ji Y., Li G., An T., 2014. Mechanism, kinetics and toxicity assessment of OH-initiated transformation of triclosan in aquatic environments. Water Res, 49, 360-370. http://dx.doi.org/10.1016/j.watres.2013.10.027 Guo F., Wang K., Lu J., Chen J., Dong X., Xia D., Zhang A., Wang Q., 2019. Activation of peroxymonosulfate by magnetic carbon supported Prussian blue nanocomposite for the degradation of organic contaminants with singlet oxygen and superoxide radicals. Chemosphere 218, 1071-1081. https://doi.org/10.1016/j.chemosphere.2018.11.197 Huie R.E., Clifton C.L., 1990. Temperature Dependence of the Rate Constants for Reactions of the Sulfate Radical with anions. J. Phys. Chem. 94, 8561-8567. https://doi.org/10.1021/j100386a015 Hu P., Long M., 2016. Cobalt-catalyzed sulfate radical-based advanced oxidation: A review on heterogeneous catalysts and applications. Appl. Catal. B Environ. 181, 103-117. https://doi.org/10.1016/j.apcatb.2015.07.024 Kang J., Wu W., Liu W., Li J., Dong C., 2019. Zero-valent iron (ZVI) Activation of Persulfate (PS) for Degradation of Para-Chloronitrobenzene in Soil. Bull Environ. Contam. Toxicol. 29

https://doi.org/10.1007/s00128-018-2511-5 Salokhiddinov K.I., Byteva I.M., Gurinovich G.P., 1981. lifetime of singlet oxygen in various solvents. J. Appl. Spectro. 34, 561–564. https://doi.org/10.1007/BF00613067 Lee D., Zhao F., Rezenom Y.H., Russell D.H., Chu K., 2012. Biodegradation of triclosan by a wastewater microorganism. Water Res. 46, 4226-34. https://doi.org/10.1016/j.watres.2012.05.025 Lee J., Hong S., Mackeyev Y., Lee C., Chung E., Wilson L.J., Kim J-H., Alvarez P.J.J., 2011. Photosensitized oxidation of emerging organic pollutants by tetrakis C(6)(0) aminofullerene-derivatized silica under visible light irradiation. Environ. Sci. Technol. 45, 10598-604. https://doi.org/10.1021/es2029944 Lee H., Lee H-J., Jeong J., Lee J., Park N-B., Lee C., 2015. Activation of persulfate by carbon nanotubes: Oxidation of organic compounds by nonradical mechanism. Chem. Eng. J. 266, 28-33. https://doi.org/10.1016/j.cej.2014.12.065 Li C., Chen C., Lu J., Cui S., Li J., Liu H., Li W., Zhang F., 2018. Metal organic framework-derived CoMn2O4 catalyst for heterogeneous activation of peroxymonosulfate and sulfanilamide degradation. Chem. Eng. J. 337, 101-109. https://doi.org/10.1016/j.cej.2017.12.069 Li C., Han X., Cheng F., Hu Y., Chen C., Chen J., 2015. Phase and composition controllable synthesis of cobalt manganese spinel nanoparticles towards efficient oxygen electrocatalysis. Nat. Commun. 6, 7345. https://doi.org/10.1038/ncomms8345 Li C., Wu J., Peng W., Fang Z., Liu J., 2019. Peroxymonosulfate activation for efficient sulfamethoxazole degradation by Fe3O4/β-FeOOH nanocomposites: Coexistence of radical and non-radical reactions. Chem. Eng. J. 356, 904-914. https://doi.org/10.1016/j.cej.2018.09.064 Li H., Song B., Wang W. J., Chen X. L., 2011. Facile synthesis, thermal, magnetic, Raman characterizations of spinel structure ZnMn2O4. 130, 39-44. https://doi.org/10.1016/j.matchemphys.2011.04.072 Li J., Liang X., Xu S., Hao J., 2009. Catalytic performance of manganese cobalt oxides on methane combustion at low temperature. Appl. Catal. B Environ. 90, 307-312. https://doi.org/10.1016/j.apcatb.2009.03.027 30

Li J., Miao J., Duan X., Dai J., Liu Q., Wang S., Zhou W., Shao Z., 2019. Fine-tuning surface properties of perovskites via nanocompositing with inert oxide toward developing superior catalysts for advanced oxidation. 28, 1804654. https://doi.org/10.1002/adfm.201804654 Lindstrom A., Buerge I.J., Poiger T., Bergqvist P.A., Muller M.D., Buser H.R., 2002. Occurrence and environmental behavior of the bactericide triclosan and its methyl derivative in surface waters and in wastewater. Environ. Sci. Technol. 36, 2322-9. https://doi.org/10.1021/es0114254 Lin K. Y. A., Chen B. J., Chen C. K., Evaluating Prussian blue analogues MII3[MIII(CN)6]2 (MII = Co, Cu, Fe, Mn, Ni; MIII = Co, Fe) as activators for peroxymonosulfate in water, RSC Adv., 2016, 95, 92923-92933. https://doi.org/10.1039/C6RA16011E Liu Y., Guo H., Zhang Y., Tang W., Cheng X., Li W., 2018. Heterogeneous activation of peroxymonosulfate by sillenite Bi25FeO40: Singlet oxygen generation and degradation for aquatic levofloxacin. Chem. Eng. J. 343, 128-137. https://doi.org/10.1016/j.cej.2018.02.125 Lu S., Wang G., Chen S., Yu H., Ye F., Quan X., 2018. Heterogeneous activation of peroxymonosulfate by LaCo1-xCuxO3 perovskites for degradation of organic pollutants. J Hazard. Mater. 353, 401-409. https://doi.org/10.1016/j.jhazmat.2018.04.021 Luo H., Xie Y., Niu J., Xiao Y., Li Y., Wang Y., Zhang Y., Xie T., Cobalt-doped biogenic manganese oxides for enhanced tetracycline degradation by activation of peroxymonosulfate. J. Chem. Technol. Biotechnol. 2019. 94, 752-760. https://doi.org/10.1002/jctb.5820 Luo R., Li M., Wang C., Zhang M., Nasir Khan M.A., Sun X., 2019. Shen J., Han W., Wang L., Li J., Singlet oxygen-dominated non-radical oxidation process for efficient degradation of bisphenol A under high salinity condition. Water Res. 148, 416-424. https://doi.org/10.1016/j.watres.2018.10.087 Luo Z., He Y., Zhi D., Luo L., Sun Y., Khan E., Wang L., Peng Y., Zhou Y., Tsang D.C.W., 2019. Current progress in treatment techniques of triclosan from wastewater: A review. 696, 133990. https://doi.org/10.1016/j.scitotenv.2019.133990 Malavasi L., Galinetto P., Mozzati M. C., Azzonib C. B., Flora G., 2002. Raman spectroscopy of AMn2O4 (A = Mn, Mg and Zn) spinels, Phys. Chem. Chem. Phys. 15. 31

https://doi.org/10.1039/B203520K Matzek L.W., Carter K.E.. 2016. Activated persulfate for organic chemical degradation: A review. Chemosphere 151, 178-88. https://doi.org/10.1016/j.chemosphere.2016.02.055 Mathew, T., Shiju, N. R., Sreekumar, K., Rao, B. S., Gopinath, C. S., 2002. Cu−Co Synergism in Cu1−xCoxFe2O4 Catalysis and XPS Aspects. J. Catal. 210, 405−417. https://doi.org/10.1006/jcat.2002.3712 Naveen AN, Selladurai S. 2014. Investigation on physiochemical properties of Mn substituted spinel cobalt oxide for supercapacitor applications. Electrochimica Acta 125, 404-414. https://doi.org/10.1016/j.electacta.2014.01.161 Ni Q., Ma J., Fan C., Kong Y., Peng M., Komarneni S., 2018. Spinel-type cobalt-manganese oxide catalyst for degradation of Orange II using a novel heterogeneous photo-chemical catalysis system. Ceramics International 44, 19474-19480. https://doi.org/10.1016/j.ceramint.2018.07.184 Nie G., Huang J., Hu Y., Ding Y., Han X., Tang H., 2017. Heterogeneous catalytic activation of peroxymonosulfate for efficient degradation of organic pollutants by magnetic Cu0/Fe3O4 submicron composites, Chinese J. Catalysis 38, 227-239. https://doi.org/10.1016/S1872-2067(16)62566-4 Oh W-D., Dong Z., Hu Z-T., Lim T-T. 2015. A novel quasi-cubic CuFe2O4–Fe2O3 catalyst prepared at low temperature for enhanced oxidation of bisphenol A via peroxymonosulfate activation. J. Mater. Chem. A 3, 22208-22217. https://doi.org/10.1039/C5TA06563A Oh W-D., Dong Z., Lim T-T., 2016. Generation of sulfate radical through heterogeneous catalysis for organic contaminants removal: Current development, challenges and prospects. Appl. Catal. B Environ. 194, 169-201. https://doi.org/10.1016/j.apcatb.2016.04.003 Peng J., Zhang Y., Zhang C., Miao D., Li J., Liu H., Wang L., Gao S., 2019. Removal of triclosan in a Fenton-like system mediated by graphene oxide: Reaction kinetics and ecotoxicity evaluation. Sci. Total Environ. 673, 726-733. https://doi.org/10.1016/j.scitotenv.2019.03.354 Reddy D.H.K., Yun Y-S., 2016. Spinel ferrite magnetic adsorbents: Alternative future materials for water purification? Coordin. Chem. Reviews 315, 90-111. https://doi.org/10.1016/j.ccr.2016.01.012 32

Ren Y., Dong Q., Feng J., Ma J., Wen Q., Zhang M., 2012. Magnetic porous ferrospinel NiFe2O4: A novel ozonation catalyst with strong catalytic property for degradation of di-n-butyl phthalate and convenient separation from water. J. Colloid Interface Sci. 382, 90-6. https://doi.org/10.1016/j.jcis.2012.05.053 Sadighi Z., Huang J., Qin L., Yao S., Cui J., Kim J-K., 2017. Positive role of oxygen vacancy in electrochemical performance of CoMn2O4 cathodes for Li-O2 batteries. J. Power Sour. 365, 134-147. https://doi.org/10.1016/j.jpowsour.2017.08.081 Shao P., Tian J., Duan X., Yang Y., Shi W., Luo X., Cui F., Luo S., Wang S., 2019. Cobalt silicate hydroxide nanosheets in hierarchical hollow architecture with maximized cobalt active site for catalytic oxidation. Chem. Eng. J. 359, 79-87. https://doi.org/10.1016/j.cej.2018.11.121 Tian X., Gao P., Nie Y., Yang C., Zhou Z., Li Y., Wang W., 2017. A novel singlet oxygen involved peroxymonosulfate activation mechanism for degradation of ofloxacin and phenol in water. Chem. Commun. 53, 6589-6592. https://doi.org/10.1039/C7CC02820B Timmins G.S., Liu K., Bechara E.J., Kotake Y., Swartz H.M., 1999. Trapping of free radicals with direct in vivo EPR detection: a comparison of 5,5-dimethyl-1-pyrroline-N-oxide and 5-diethoxyphosphoryl-5-methyl-1-pyrroline-N-oxide as spin traps for HO* and SO4*. Free Radic. Biol. Med. 27, 329-33. https://doi.org/10.1016/S0891-5849(99)00049-0 Wei J., Feng Y., Liu Y., Ding Y., 2015. MxCo3−xO4 (M = Co, Mn, Fe) porous nanocages derived from metal–organic frameworks as efficient water oxidation catalysts. J. Mater. Chem. A 3, 22300-22310. https://doi.org/10.1039/C5TA06411B Xu H., Wang D., Ma J., Zhang T., Lu X., Chen Z., 2018. A superior active and stable spinel sulfide for catalytic peroxymonosulfate oxidation of bisphenol S. Appl. Catal. B Environ. 238, 557-567. https://doi.org/10.1016/j.apcatb.2018.07.058 Yang B., Ying G., Zhao J., Zhang L., Fang Y., Nghiem L.D., 2011.Oxidation of triclosan by ferrate: reaction kinetics, products identification and toxicity evaluation. J. Hazard. Mater. 186, 227-35. https://doi.org/10.1016/j.jhazmat.2010.10.106 33

Yang S., Che D., 2017. Degradation of aquatic sulfadiazine by Fe0/persulfate: kinetics, mechanisms, and degradation pathway. RSC Advances 7, 42233-42241. https://doi.org/10.1039/C7RA07920F Yang S., Wu P., Liu J., Chen M., Ahmed Z., Zhu N., 2018. Efficient removal of bisphenol A by superoxide radical and singlet oxygen generated from peroxymonosulfate activated with Fe0-montmorillonite. Chem. Eng. J. 350, 484-495. https://doi.org/10.1016/j.cej.2018.04.175 Yao J., Zeng X., Wang Z., 2017. Enhanced degradation performance of sulfisoxazole using peroxymonosulfate activated by copper-cobalt oxides in aqueous solution: Kinetic study and products identification. Chem. Eng. J. 330, 345-354. https://doi.org/10.1016/j.cej.2017.07.155 Yao Y., Cai Y., Wu G., Wei F., Li X., Chen H., Wang S., 2015. Sulfate radicals induced from peroxymonosulfate by cobalt manganese oxides (Co(x)Mn(3-x)O4) for Fenton-Like reaction in water. J. Hazard. Mater. 296, 128-137. https://doi.org/10.1016/j.jhazmat.2015.04.014 Yu J., Kwong T.Y., Luo Q., Cai Z., 2006. Photocatalytic oxidation of triclosan. Chemosphere 65, 390-9. https://doi.org/10.1016/j.chemosphere.2006.02.011 Yun E.T.,, Yoo H-Y., Bae H., Kim H-I., Lee J., 2017. Exploring the Role of Persulfate in the Activation Process: Radical Precursor Versus Electron Acceptor. Environ. Sci. Technol. 51, 10090-10099. https://doi.org/10.1021/acs.est.7b02519 Wu L., Zhang Q., Hong J., Dong Z., Wang J., 2019. Degradation of bisphenol A by persulfate activation via oxygen vacancy-rich CoFe2O4–x. Chemosphere. 221, 412-422. https://doi.org/10.1016/j.chemosphere.2019.01.049 Zhang H., Huang C., 2003. Oxidative transformation of triclosan and chlorophene by manganese oxides. Environ. Sci. Technol. 37, 2421-30. https://doi.org/10.1021/es026190q Zhang J., Chen M., Zhu L., 2016. Activation of persulfate by Co3O4 nanoparticles for orange G degradation. RSC Advances 6, 758-768. https://doi.org/10.1039/C5RA22457H Zhang J., Zhou W., Yang L., Chen Y., Hu Y., 2018. Co-N-doped MoO2 modified carbon felt cathode for removal of EDTA-Ni in electro-Fenton process. Environ. Sci. Pollut. Res. Int. 34

25, 22754-22765. https://doi.org/10.1007/s11356-018-2373-8 Zhang T., Chen Y., Wang Y., Le Roux J., Yang Y., Croue J.P., 2014. Efficient peroxydisulfate activation process not relying on sulfate radical generation for water pollutant degradation. Environ. Sci. Technol. 48, 5868-75. https://doi.org/10.1021/es501218f Zhao Q., Yan Z., Chen C., Chen J., 2017. Spinels: controlled preparation, oxygen reduction/evolution reaction application, and beyond. Chem. Rev. 117, 10121-10211. https://doi.org/10.1021/acs.chemrev.7b00051 Zhou Y., Sun S., Song J., Xi S., Chen B., Du Y., Adrian C.F., Cheng F., Wang X., Zhang H., Xu Z., 2018. Enlarged Co-O covalency in octahedral sites leading to highly efficient spinel oxides for oxygen evolution reaction. Adv Mater. 30, e1802912. https://doi.org/10.1002/adma.201802912 Zhou Y., Jiang J., Gao Y., Ma J., Pang S., Li J., Lu X., Yuan L., 2015. Activation of peroxymonosulfate by benzoquinone: a novel nonradical oxidation process. Environ Sci Technol. 49, 12941-12950. https://doi.org/10.1021/acs.est.5b03595 Zhu S., Li X., Kang J., Duan X., Wang S., 2019. Persulfate activation on crystallographic manganese oxides: Mechanism of singlet oxygen evolution for nonradical selective degradation of aqueous contaminants. Environ Sci Technol. 53, 307-315. https://doi.org/10.1021/acs.est.8b04669 Huang Z., Chen Z., Chen Y., Hu Y., 2018. Synergistic effects in iron-copper bimetal doped mesoporous γ-Al2O3 for Fenton-like oxidation of 4-chlorophenol Structure, composition, electrochemical behaviors and catalytic performance. Chemosphere. 203, 442-449. https://doi.org/10.1016/j.chemosphere.2018.04.001

35

Fig. 1 SEM-EDS spectra of synthesized catalysts: (a)(d) CoO, (b)(e) Co2Mn1O4 and (c)(f) MnO

Fig. 2 (a) Powder XRD patterns, (b) Nitrogen adsorption and desorption isotherms and corresponding

pore

size

distribution

(inset)

36

of

synthesized

catalysts

Fig. 3 Influence of (a) PMS dose ([Co2Mn1O4] = 0.02 g L−1, [TCS] 0 = 10 mg L−1), and (b) Co2Mn1O4 catalyst dose ([PMS]0 = 0.05 g L−1, [TCS] 0 = 10 mg L−1) and TCS concentration ([Co2Mn1O4] = 0.02 g L−1, [PMS]0 = 0.05 g L−1) on PMS decomposition.

37

Fig. 4 (a) Influence of initial pH on TCS degradation rate in Co2Mn1O4/PMS process; ([Co2Mn1O4] = 0.02 g L−1, [PMS]0 = 0.05 g L−1, [TCS] 0 = 10 mg L−1); (b) Species distribution of PMS and TCS at different pH values.

38

Fig. 5 TCS removal efficiency and rate constant k of Co2Mn1O4/PMS system with co-existing anions at different concentrations, (a) Cl−, (b) NO3−, (c) SO42−, (d) HCO3−, (e) Ca2+ and (f) Mg2+ ([Co2Mn1O4] = 0.02 g L−1, [PMS]0 = 0.05 g L−1, [TCS] 0 = 10 mg L−1).

39

Fig. 6 (a) The quenching effect of different scavengers on TCS degradation in Co2Mn1O4/PMS system and (b) the corresponding rate kinetics ([Co2Mn1O4] = 0.02 g L−1, [PMS]0 = 0.05 g L−1, [TCS] 0 = 10 mg L−1).

Fig. 7 (a) DMPO and (b) TEMP spin-trapping EPR spectra of PMS activation in different systems. ([Co2Mn1O4] = 0.02 g L−1, [PMS]0 = 0.05 g L−1, [DMPO] = 10mM, [TEMP] = 10mM)

40

Fig. 8 (a) General XPS spectra and deconvoluted peaks of (b) Co 2p, (c) Mn 2p, and (d) O 1s for Co2Mn1O4 before and after reaction

41

Fig. 9 Proposed activation mechanism of PMS on cobalt manganese spinel

Fig.10 Pathways of TCS oxidation in the Co2Mn1O4/PMS system

Graphic abstract

42

Highlights 

Activation of PMS using Co2Mn1O4 exhibits a superior effect on degradation of TCS.



Both 1O2 and SO4•− were identified as the main reactive oxygen species.



1O



Degradation pathways for TCS in Co2Mn1O4/PMS were investigated through GC-MS.

2

involved in generation of oxygen vacancy promoted metal redox cycle.

43