Characterizations of Co2(V4O12)(bpy)2
The single crystal structure (CCDC 2323958) was analyzed to affirm the identical structure of as-prepared Co2(V4O12)(bpy)2 to the previous ones18, in which Co2+ is five-coordinated by two nitrogen atoms from bpy and three oxygen atoms from three different [V4O12]4− clusters. It was observed that six terminal oxygen atoms in [V4O12]4− are coordinated to six different Co2+ atoms via Co-O bonds, and the other two terminal oxygen atoms are uncoordinated (Supplementary Fig. 1 and Supplementary Table 1). The powder X-ray diffraction (PXRD) patterns (Fig. 1a) of as-prepared Co2(V4O12)(bpy)2 matched well with the simulated ones from CCDC 2323958 and 152985, implying high purity of the catalyst used in the successive tests and determinations. The optical microscope image (Fig. 1b) showed that Co2(V4O12)(bpy)2 displayed yellow clump crystal with the micron scale, which was affirmed by scanning electron microscopy (SEM) image (Supplementary Fig. 2). High-resolution transmission electron microscopy (HR-TEM) affirmed the exposed facet (3 3 2) of [V4O12]4− of Co2(V4O12)(bpy)2 (Fig. 1c and Supplementary Fig. 3). The corresponding elemental mapping (Fig. 1d) displayed uniform distribution of characteristic elements like Co, V, N, and O.Fig. 1: Chemical and structural characterizations of Co2(V4O12)(bpy)2.a PXRD patterns of as-fabricated Co2(V4O12)(bpy)2 catalyst. b Optical microscope image and (c) HR-TEM image of Co2(V4O12)(bpy)2. d Elemental mapping images of characteristic elements from Co2(V4O12)(bpy)2. XANES spectra of the (e) Co and (f) V K-edge in Co2(V4O12)(bpy)2 (insert: the absorption edge of Co and V, respectively). EXAFS spectra of the (g) Co and (h) V K-edge in Co2(V4O12)(bpy)2. i WT plots of the Co and V K-edge EXAFS for the Co foil, V foil and Co2(V4O12)(bpy)2.The chemical state and coordination environment of Co and V were investigated using X-ray adsorption near edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) spectra. The Co (Fig. 1e) and V (Fig. 1f) element K-edge XANES combined with k-space and fitted data (Supplementary Fig. 4) illustrated that the valence states of Co and V were near +2 and +5, respectively. The EXAFS spectra (Fig. 1g) indicated that Co atoms from Co2(V4O12)(bpy)2 were coordinated to O and N atoms with the bond length of 1.56 Å (Peak I). In addition, no Co clusters or crystalline particles was presented in Co2(V4O12)(bpy)2 structure due to that no obvious peak of the Co–Co bond at 2.18 Å were observed19, which matched well with the results of crystal structure analyses. The EXAFS spectra of V element (Fig. 1h) demonstrated that V-O bond (Peak II) with the length of 1.28 Å was different from those of V2O3 (1.54 Å) and V2O5 (1.54 Å), illustrating that the interactions between V and O atoms in [V4O12]4− were stronger than those in V2O3 and V2O5. The wavelet transform (WT) of Co2(V4O12)(bpy)2 was performed to further clarify the above-mentioned peaks. As illustrated in Fig. 1i and Supplementary Fig. 5, the peaks at 3.67 Å−1 and 3.81 Å−1 of Co2(V4O12)(bpy)2 sample could be ascribed to Co-O/N and V-O bonds, respectively. The WT plots of Co foil, V foil and the single crystal structure determination results revealed that no Co-Co and V-V bond could be observed in Co2(V4O12)(bpy)2.Catalytic activity and involved oxidation mechanismOFX was selected as target micropollutant to investigate the PMS activation performance of Co2(V4O12)(bpy)2. The PMS activation abilities of different systems were investigated without adjusting pH (pH = 6.38) (Fig. 2a and Supplementary Figs. 6, 7). The Co2(V4O12)(bpy)2 displayed the best PMS activation for OFX degradation with the biggest reaction apparent rate constant (kobs) of 2.72 min−1 (Supplementary Fig. 8a), which was further affirmed by PMS consumption experiments (Supplementary Fig. 8b). Additionally, ca. 48.5% of total organic carbon (TOC) removal could be achieved within 5.0 min in Co2(V4O12)(bpy)2/PMS system (Supplementary Fig. 8c), verifying the better mineralizing capacity of Co2(V4O12)(bpy)2 than Co3O4 and V2O5. The Co2(V4O12)(bpy)2 could also accomplish ultra-high elimination efficiencies toward other organic micropollutants like tetracycline (TC), bisphenol A (BPA), 3,4-dichlorophenol (3,4-DCP), chloroquine phosphate (CQ), and sulfamethoxazole (SMX) (Fig. 2b and Supplementary Fig. 8d), in which the corresponding k-value (min−1 M−1) surpassed almost all reported catalysts for PMS activation (Fig. 2c and Supplementary Table 3). It was affirmed that Co2(V4O12)(bpy)2 displayed good reusability (Fig. 2d). Only slight decline of catalytic activity could be observed after five recycles’ operation, possibly due to that some active sites are occupied by the formed intermediates (Supplementary Fig. 9)20. The leaching Co (0.065 mg L−1) and V (0.084 mg L−1) were far below the environmental quality standards for surface water (GB 3838-2002), indicating the great water stability of catalyst. Also, the observations of PXRD patterns, FTIR spectrum, SEM image and HR-TEM image affirmed the decent stability of Co2(V4O12)(bpy)2 (Supplementary Fig. 10a–d).Fig. 2: Catalytic activity, oxidation mechanism and environmental applicant in Co2(V4O12)(bpy)2/PMS system.a The Fenton-like reaction activity test over different catalysts. b Removal of multiple micropollutants by Co2(V4O12)(bpy)2 (inset: the corresponding ESP of the organic pollutants). c Comparison of the kinetics of multiple micropollutants elimination efficiency by the state-of-the-art catalysts on PMS activation system (The related information was presented in Supplementary Table 3). d Recyclability test of Co2(V4O12)(bpy)2. e The oxidation mechanism of 1O2 for multiple micropollutants (inset: the corresponding HOMO and LUMO of the organic pollutants). f Influences of different simulated wastewater on OFX elimination efficiencies. g Inhibitions of OFX and its degradation intermediates on E. coli growth. h Schematic illustration of pilot reactor. i OFX concentration evolution of different systems in the reactor. Experimental conditions: [Catalyst] = 0.2 g L−1, [PMS] = 0.1 mM, [OFX] = 10.0 mg L−1 (in a, d, f), [TC] = [BPA] = [3,4-DCP] = [CQ] = [SMX] = 5.0 mg L−1 (in b). The error bars in the figures represented the standard deviations from triplicate tests.To identify the primary reactive oxygen species (ROSs) during PMS activation, chemical quenching experiments were conducted (Supplementary Note 5), in which methanol (MeOH), tert-butanol (TBA), benzoquinone (BQ), and furfuryl alcohol (FFA) were selected as quenchers of SO4•−/•OH, •OH, O2•− and 1O2 quencher, respectively20,21,22. The introduction of MeOH and FFA into Co2(V4O12)(bpy)2/PMS system could inhibit the OFX degradation efficiencies from 100% to 68.2% and 45.9%, respectively (Supplementary Fig. 11a). While negligible influence on catalytic OFX degradation efficiencies could be observed when TBA and BQ were added. It was observed that PMS could be directly consumed by FFA (Supplementary Fig. 11b), comparable to the previous studies23. Therefore, L-histidine, as another 1O2 quencher, was adopted to quench 1O2 to further identify its contribution. The results (Supplementary Fig. 11c, d) indicated that both the OFX degradation efficiency and rate were seriously inhibited. Importantly, the OFX catalytic rate in D2O was faster than that in deionized water (Supplementary Fig. 11c, d), further confirming the significant contribution of 1O2 in Co2(V4O12)(bpy)2/PMS system. The above results demonstrated that SO4•− and 1O2 might play primary contribution during the degradation process. Also, electron paramagnetic resonance (EPR) analyses results (Supplementary Fig. 12) showed that strong DMPO-SO4•− and TEMP-1O2 signals could be observed, implying SO4•− and 1O2 were yielded in Co2(V4O12)(bpy)2/PMS system. Besides, 1O2 quantitative analyses via EPR technology (Supplementary Fig. 12b) were performed to avoid overestimating its contribution, revealing that the as-produced 1O2 in Co2(V4O12)(bpy)2/PMS system could survive longer in D2O than in H2O24,25. And the intensity of TEMP-1O2 signal decrease evidently when OFX was added in Co2(V4O12)(bpy)2/PMS/D2O system, confirming the significant role of 1O2 during the OFX degradation process. N2 atmosphere, O2 atmosphere (Supplementary Fig. 13) and aniline quenching experiments (Supplementary Fig. 11a) verified that the generation of 1O2 might be attributed to SO5•− decomposition rather than superoxide radicals (O2•−) transformation26,27. Meanwhile, the results of O2•− detection by fluorescence method (Supplementary Fig. 14) confirmed that O2•− could not be abundantly generated in Co2(V4O12)(bpy)2/PMS system28. PMS-premixing and high-valent metal capture experiments (Supplementary Fig. 15) excluded the contribution of other nonradical pathways like electron transfer process and high-valent Co-intermediate, respectively25. Chemical probe experiments (Supplementary Notes 6, 7)29 affirmed that the produced 1O2 and SO4•− in Co2(V4O12)(bpy)2/PMS system as the major active species could realize the rapid removal toward OFX (Supplementary Fig. 16). However, the TOC removal mainly relied on the oxidation process of surface-bound radicals and the adsorption between intermediates and catalyst (Supplementary Fig. 17 and Supplementary Note 8)30,31,32,33. It had been reported that electron-rich contaminants could be selectively oxidized in non-radical system23,34. The electrostatic potentials (ESP) of some selected organics were calculated, exhibiting that TC, BPA, 3,4-DCP, CQ and SMX with functional groups of electron aggregation displayed higher ESP than benzoic acid (BA) and nitrobenzoic acid (NBA) (Fig. 2b and Supplementary Fig. 18). The difference of 1O2 oxidation ability toward different selected organics was investigated from the view of charge transfer pathway (Fig. 2e). It could be clearly observed that the highest occupied molecular orbital (HOMO) of electron-rich contaminants like TC (−5.25 eV), BPA (−4.84 eV), 3,4-DCP (−5.30 eV), CQ (−4.77 eV) and SMX (−4.96 eV) were slightly lower than the lowest unoccupied molecular orbital (LUMO) of 1O2 (−4.59 eV), indicating that 1O2 could readily withdraw electrons from HOMO of those electron-rich contaminants to further achieve organics degradation35. While the LUMO of 1O2 was much higher than the HOMO of BA (−6.14 eV) and NBA (−6.70 eV), demonstrating that it was difficult to perform the oxidation reaction (Supplementary Fig. 19).The co-existing anions and organic matters in the actual water could react with ROSs to inhibit the catalytic efficiencies36. The results demonstrated that co-existing substances exerted negligible impact on OFX degradation in Co2(V4O12)(bpy)2/PMS system (Supplementary Fig. 20), due to that 1O2 possessed the lower oxidation potential and better selective oxidation ability than radicals. Moreover, the Co2(V4O12)(bpy)2 could efficiently activate PMS to accomplish the OFX degradation in the wide pH range from 3.12 to 8.04 (Supplementary Figs. 21, 22), considering that 1O2 exhibited strong tolerance to solution pH37. Also, Co2(V4O12)(bpy)2/PMS system could achieve complete removal toward OFX in the water samples simulated from actual water solutions (river, lake, tap, purified drinking and secondary settling tank water) within 5.0 min (Fig. 2f and Supplementary Fig. 23). These findings revealed that Co2(V4O12)(bpy)2/PMS system displayed superior anti-interference ability, reflecting the superiority of 1O2 in practical applications for the OFX degradation.Besides UPLC-MS products identification, the Fukui index of OFX was calculated (Supplementary Figs. 24–26) to identify the reactive sites38, which assisted to propose five possible OFX elimination pathways. The toxicity changes of the formed intermediates (Supplementary Fig. 27) showed that PMS activation process over Co2(V4O12)(bpy)2 could degrade OFX into lowly toxic intermediates, even completely mineralize OFX into CO2 and H2O. The growth inhibition influence of OFX and its intermediates on Escherichia coli (E. coli) was determined (Supplementary Note 9), confirming the low toxicity of intermediates. As displayed in Fig. 2g, the diameters of inhibition zones for OFX, the degradation products in 1.0 min and 5.0 min against E. coli were 23.0, 12.0, and 9.0 mm (near blank experiment without OFX), demonstrating that the OFX degradation over Co2(V4O12)(bpy)2 resulted in decreasing toxicity toward E. coli. The plant growth experiment (Supplementary Fig. 28) exhibited that the germination rate and growth status of bean sprouts cultivated with the treated wastewater were comparable to the bean sprouts cultivated with deionized water. The above results indicated that Co2(V4O12)(bpy)2/PMS system displayed satisfactory mineralization ability and detoxification ability.To evaluate the practical application prospects of Co2(V4O12)(bpy)2 as PMS activator, 300.0 mg Co2(V4O12)(bpy)2 powder was fixed onto the graphene cloth (Fig. 2h) to perform the long term treatment to both simulated wastewater containing OFX formulated from pure water and actual secondary sedimentation tank water were performed (Supplementary Fig. 29). Figure 2i exhibited that the OFX elimination efficiency in both pure water and actual secondary sedimentation tank water could maintain >90% up to 56 h and 40 h in Co2(V4O12)(bpy)2/PMS system, while the OFX removal efficiency of the individual PMS system was less than 20%. Additionally, the immobilized Co2(V4O12)(bpy)2 ensured the catalyst stability during the reaction process (Supplementary Fig. 30a, b), in which the leached Co (<0.3 mg L−1) and V (<0.35 mg L−1) concentrations were far below the maximum pollutant level target set by the Chinese National Standard (Co: GB 3838-2002; V: GB 26452-2011). The PXRD pattern (Supplementary Fig. 31) and SEM image (Supplementary Fig. 32) of the used Co2(V4O12)(bpy)2 after the reaction matched well with the fresh sample, demonstrating the stable phase, composition and morphology of Co2(V4O12)(bpy)2. Meanwhile, the TOC removal efficiency could reach about 50% within 48 h in Co2(V4O12)(bpy)2/PMS/pure water system (Supplementary Fig. 30c), implying that OFX could be partly mineralized into CO2 and H2O. These results confirmed that the Co2(V4O12)(bpy)2 catalyst possessed great potential for practical application.Catalytic sites identification and excitation mechanismDFT calculations were utilized to construct Co2(V4O12)(bpy)2 bulk model with spin parallel, which was more stable than spin antiparallel (Supplementary Fig. 33). The Co2(V4O12)(bpy)2 surface model (Supplementary Fig. 34a) was constructed along the (0 0 1) crystal facet, in which Co atoms adjacent the vacuum layer were unsaturated coordination sites. The EPR exhibited an obvious response at g = 2.003 (Supplementary Fig. 34b), affirming the existence of vacancies39. Generally, the electron accumulation area might be more likely to act as the catalytic center40. The results of density of states and partial density of states (PDOS) indicated that electron occupied orbital near Fermi level (EF) were mainly attributed to the contribution of Co 3d, V 3d and O 2p orbitals (Supplementary Fig. 35), in which electronic state of Co 3d displayed obvious split due to the Jahn-Teller effect (Supplementary Fig. 35b). Six possible PMS sorption models (Supplementary Figs. 36, 37) were built to calculate adsorption energy (Eads), considering that both Co and V could be seen as the potential sorption sites to interact with PMS. As shown in Fig. 3a, Co sites from Co2(V4O12)(bpy)2 surface bonded with O1 from PMS, and the hydrogen bond was formed between the terminal hydrogen of PMS and the terminal oxygen of [V4O12]4−. The Eads was calculated to be −1.57 eV, much more negative than those of other models, confirming that Co-O1 (Co-PMS*) was the optimal sorption model.Fig. 3: Catalytic sites identification and excitation mechanism exploration.a Optimal PMS adsorption model on Co2(V4O12)(bpy)2 and corresponding adsorption energy. b Electron density difference of optimal PMS adsorption model. c Planar-averaged electron density difference Δρ (Z). d Electron filled situation of Co d orbital and spin state before and after PMS adsorption (e) PDOS of O 2p and Co 3d before and after PMS adsorption. f Change of Mulliken charge of PMS model. XPS spectra of (g) Co 2p (h) O 1 s before and after reaction. i Co K-edge EXAFS spectra, (j) Co K-edge and (k) V K-edge XANES spectra as well as (l) WT plots of Co2(V4O12)(bpy)2 before and after reaction (insert: the absorption edge and white line of Co and V, respectively).Based on optimal model, the corresponding electron density difference was calculated. Figure 3b indicated that PMS could draw electron (0.35 e) from the surface of catalyst, in which the electron accumulation area mainly focused on hydrogen bond rather than peroxo O-O bond (Fig. 3c). To further investigate the source of electron for PMS activation, the Co 4d orbital electron distribution was calculated based on Hund’s rule and Pauli exclusion principle (Fig. 3d). Although Co sites were regarded as the optimal sorption sites toward PMS, weak increase of spin states after PMS adsorption indicated that high-spin complexes with strong oxidation potential were not facilely formed. Therefore, Co sites might not be considered as the only electron transport centers, which was also consistent with the analyses results of PDOS. As shown in Fig. 3e and Supplementary Fig. 38, all the Co 3d, O 2p and V 3d orbitals presented distinct blue shift, indicating that PMS could draw electrons from Co, O, V atoms. Mulliken charge population results (Fig. 3f) indicated that H atom in PMS model after adsorption could obtain more electrons than those of other atoms in original PMS model. The bond length of O-H bond of PMS after adsorption was longer than that of pristine one (Supplementary Fig. 39), while the bond length of peroxo O-O bond nearly remain constant. The above-mentioned results indicated that O-H bond of PMS was easier to be broken than peroxo O-O bond during activation process, due to the pull of hydrogen bond and charge transfer, which was further confirmed by the Gibbs free energy diagrams (Supplementary Fig. 40). To investigate the significant role of hydrogen bond, PMS adsorption model without hydrogen bond was built and calculated. It was found from Supplementary Fig. 41 that the Eads of PMS adsorption model with hydrogen bond was higher than that of PMS adsorption model without hydrogen bond (−1.15 eV), illustrating that PMS adsorption pattern with hydrogen bond was easier to be formed from thermodynamic perspective. To further affirm the role of hydrogen bond interaction, F− (NaF) with biggest electronegativity was introduced into the reaction system to compete with O from [V4O12]4− for forming hydrogen bond. The introduction of F− led to the noticeable inhibition of OFX degradation efficiency and rate (Supplementary Fig. 42), due to that F− could prefer to interact with O-H from PMS to form strong hydrogen bond to decrease the formation of hydrogen bond between PMS and catalyst. It was deemed that Co cations as the PMS sorption sites also played the key role during the Fenton-like reaction. As shown in Supplementary Fig. 43, ethylene diamine tetraacetic acid and phosphate were selected to shield the Co sites to further inhibit the PMS sorption process. As expected, both the OFX elimination performances and PMS consumption decreased, confirming the significant contribution of Co sites. In addition, the PMS activation ability of V sites based on the V-O2 model was also investigated (Supplementary Fig. 37e). As shown in Supplementary Fig. 44a, 0.13 e could be transferred from the surface of Co2(V4O12)(bpy)2 to PMS. Supplementary Fig. 44b indicated that the electron accumulation area mainly focused on hydrogen bond rather than peroxo O-O bond in PMS, which was similar to Co-O1 model. Moreover, the dual sites PMS adsorption model was also constructed (Supplementary Fig. 45). The antibonding orbital formed between the adsorption site and PMS was filled with more electrons due to the blue shift of the d-band center41,42, which further led to the deterioration of the PMS adsorption ability (Supplementary Fig. 46a, b). As shown in Supplementary Fig. 47a, the Eads of dual sites PMS adsorption model was lower than that of individual Co or V sites. It was interesting that PMS activation didn’t merely rely on electron transfer from Co metal sites in Co2(V4O12)(bpy)2/PMS system, due to that the electron transfer number (0.37 e) from catalyst to PMS was not altered (Supplementary Fig. 47b). The decrease of Eads signified that the formed active species were more susceptible to desorption, which was more conducive to achieve PMS activation.X-ray photoelectron spectroscopy (XPS), XANES spectra and EXAFS spectra of Co2(V4O12)(bpy)2 before and after reaction combined with DFT calculations were performed to ascertain the electron transfer pathway during the catalytic reaction. Figure 3g exhibited that the percentage of Co3+ increased from 27.7% to 33.6% after the catalytic reaction, indicating that Co2+ might be converted to Co3+ due to electron transfer from Co sites to PMS43. Some electrons were delivered from Co sites to PMS for activation reaction, and other electrons were captured and stored in [V4O12]4− electron reservoir. Meanwhile, the binding energy of Co-N (Supplementary Fig. 48a) and Co-O (Fig. 3h) all shifted to higher energy level than that of pristine catalyst, affirming the electron transfer process from Co sites. Additionally, Supplementary Fig. 48b showed that the red shift of 0.08 eV happened to the binding energy of V 2p of pristine catalyst and after reaction, indicating that [V4O12]4− as the charge receptor could store and transfer electron44. Figure 3h also displayed that the binding energy of O 1 s before and after reaction shifted from 529.65 eV to 529.73 eV, illustrating that O atoms could act as electron donors to transfer electrons to PMS. Co K-edge EXAFS spectra (Fig. 3i) indicated that the peak intensity increased after PMS activation reaction, which could be attributed to the formation of Co-OPMS bond. The white line variation (Fig. 3j) of Co after reaction implied the redistribution of electron density, in which the weak intensity of white line indicated the electronic loss of Co sites. Meanwhile, the enhanced white line intensity of V metal after reaction (Fig. 3k) demonstrated that V draw electrons to cause the increasing charge density45, which matched well with XPS analyses. Furthermore, the WT plots of Co (Supplementary Fig. 49) and V (Fig. 3l) displayed that Co-N/O bond and V-O bond shifted from 3.67 Å−1 and 3.81 Å−1 to 3.92 Å−1 and 3.43 Å−1 after reaction, respectively, indicating that both Co-N/O bond and V-O bond presented dynamic variation during the PMS activation reaction.In situ FTIR under N2 atmosphere (Supplementary Fig. 50a, b) and In situ Raman spectroscopy (Supplementary Fig. 50c, d) were performed to further investigate the intermolecular interaction between PMS and Co2(V4O12)(bpy)2 during the PMS activation45. The FTIR spectra of individual PMS (Supplementary Fig. 51) showed that the peaks at 3676 cm−1 and 1054 cm−1 were assigned to the O-H bond and S-O bond, respectively46. When PMS with a certain concentration was added into the system, the peak of O-H bond (Fig. 4a) exhibited a rapid blue shift from 3676 cm−1 to 3673 cm−1. Then, the peak of O-H bond was kept at 3673 cm−1 after 1.0 min reaction, which might be ascribed to the rapid PMS consumption in Co2(V4O12)(bpy)2/PMS system. The dynamic variation of O-H peak could be attributed to: (i) O-H bond from PMS presented strong interaction with catalyst surface; and (ii) O-H bond in PMS was broken due to the interaction, which led to the formation of the new and steady O-H bond between H from PMS and terminal O from [V4O12]4− on the catalyst surface. XPS analyses also indicated (Fig. 3h) that the content of hydroxyl group on the catalyst increased after Fenton-like reaction, which might be attributed to the interaction between terminal O from catalyst and terminal H from PMS. Meanwhile, the slight shift of S-O peak (Fig. 4a) could be attributed to the adsorption effect of metal center toward PMS, which was comparable to DFT calculations. More significantly, the characteristic stretching of O-O at ~ 896 cm−1 (Fig. 4a) appeared under the conditions of N2 atmosphere and the absence of dissolved oxygen47,48, further verifying the generation of 1O2 in Co2(V4O12)(bpy)2/PMS system due to that the interference of O2•− on O-O bond was eliminated. Additionally, In situ Raman with laser excitation at 532 nm was adopted to characterize the change of PMS during the reaction. The characteristic vibration of HSO5− gradually shifted to higher wavenumbers (Fig. 4b), affirming the electron transfer process from catalyst to PMS45. Also, the new peak at about 837 cm−1 disappeared after Co2(V4O12)(bpy)2 was added into the PMS solution20, indicating that the high-spin complexes (catalyst−PMS*) with strong oxidation ability was not formed. Moreover, In situ Raman experiments were performed to facilitate to observe the dynamic variation of metal centers, in which the Raman peaks at 476.8 cm−1, 619 cm−1 and 962.3 cm−1 could be ascribed to the Co-O, Co-N and V-O stretching frequencies (Fig. 4c), respectively49. Compared with the dynamic structure of Co-O and Co-N bond, the slight dynamic change of V-O bond could be neglected (Fig. 4d) due to the stable structural characteristics of [V4O12]4− cluster50, matching well with HR-TEM image. As shown in Supplementary Fig. 52, the crystal lattice of [V4O12]4− of the used Co2(V4O12)(bpy)2 was almost identical to that in the fresh one. On the contrary, noticeable stretching vibration of Co-O and Co-N bonds indicated the mutual affinity between Co sites and PMS because the metal centers will induce structural response to adsorption of PMS. Interestingly, above conclusions matched well with the analysis results of DFT calculations. The included angle of O1-Co1-O2 changed from 140.2° to 126.5° when PMS was adsorbed on the Co sites, while the included angle of O3-V1-O4 only shifted from 107.9° to 107.4° after PMS adsorption (Supplementary Fig. 53). In conclusion, beside the electrons transferred from Co sites to PMS, In situ spectral results combined with DFT calculations jointly indicated that the O-H of PMS preferred to be broken due to the pull of hydrogen bond and electron transfer between [V4O12]4− and PMS, implying the formation of SO5•− might further produce 1O2 and SO4•− for OFX degradation (Fig. 4e).Fig. 4: In situ investigation of PMS activation process and electrochemical tests.a In situ FTIR spectra of Co2(V4O12)(bpy)2/PMS system under N2 atmosphere. b In situ Raman spectra of Co2(V4O12)(bpy)2, PMS, and their reaction. c The shift of different peaks in the In situ Raman spectrum. d The shift of the characteristic peaks of the In situ Raman spectra during the different reaction time. e Stepwise analyses of PMS activation mechanism during Fenton-like reaction. f Influences of different guest molecule activators on OFX degradation efficiencies. Experimental conditions: [Catalyst] = 0.2 g L−1, [PMS] = 0.1 mM, [PDS] = [H2O2] = 3.5 mM, [OFX] = 10.0 mg L−1 (in f). g ECSA measurements based on the results of CV curves of Co2(V4O12)(bpy)2, Co3O4 and V2O5. h Open-circuit potential curves and (i) Amperometric i-t curve in different system using Co2(V4O12)(bpy)2 as working electrode. The error bars in the figures represented the standard deviations from triplicate tests.Based on the above discussions, three kinds of POM-MOFs and three kinds of Co based MOFs were adopted to investigate the universality of the proposed mechanism (Supplementary Fig. 54). The results of quenching and chemical probe experiments (Supplementary Fig. 55–57) indicated that 1O2 played the major contribution for OFX degradation in POM-MOFs/PMS system, which could be attributed to the electron directional transfer effect of polyvanadates. While the Co sites were regarded as the electron donators in Co based MOFs for PMS activation, leading to that the radicals and nonradicals jointly participated the oxidation reaction for OFX elimination. Additionally, the activations of PDS and hydrogen peroxide (H2O2) were also performed to explore the significant role of host-guest interactions on other Fenton-like reactions. As expected, the results (Fig. 4f) indicated that the catalytic activity of H2O2 activation was better than that of PDS activation, in which the catalytic degradation rate of H2O2 activation was about 3.58 times higher than that of PDS activation (Supplementary Fig. 58a). H2O2 adsorption model (Supplementary Fig. 58b) and PDS adsorption model (Supplementary Fig. 58c) displayed that the terminal H atom of H2O2 could also generate interactions with terminal O atom of [V4O12]4− cluster, while PDS adsorption could only rely on Co sites. The calculated adsorption energy indicated that PMS adsorption model displayed highest sorption energy (Eads = −1.57 eV), followed by H2O2 adsorption model (Eads = −0.55 eV) and PDS adsorption model (Eads = −0.24 eV). Therefore, the directly electron transfer from [V4O12]4− could achieve the rapid H2O2 activation to produce active species for OFX degradation. In contrast, PDS activation process could only rely on the electron transfer from Co sites, which displayed poor catalytic OFX degradation performances. Electrochemical characterizations also confirmed that Co2(V4O12)(bpy)2/PMS system possessed better electron transfer ability than those of Co2(V4O12)(bpy)2/H2O2 system and Co2(V4O12)(bpy)2/PMS system (Supplementary Fig. 58d, e). In addition, electrochemical experiments were performed to test the real surface area of various catalysts. Supplementary Fig. 59 showed that cyclic voltammetry (CV) curves of Co2(V4O12)(bpy)2, Co3O4 and V2O5 after 20 electrochemical activation cycles, which could calculate the double-layer capacitance (Cdl) of Co2(V4O12)(bpy)2, Co3O4 and V2O5 based on the current density and different scan rates51. Figure 4g displayed that the Cdl (5.26 mF cm−2) of Co2(V4O12)(bpy)2 was higher than those of Co3O4 (2.44 mF cm−2) and V2O5 (4.88 mF cm−2). The higher Cdl value indicated the stronger electrochemical active surface area52. The open-circuit potential of Co2(V4O12)(bpy)2 rose when PMS was added into the electrolyte (Fig. 4h and Supplementary Fig. 60a, b), and the potential gradually elevated with the addition of PMS and OFX. Amperometric i − t curves (Fig. 4i and Supplementary Fig. 60c, d) displayed that the addition of PMS and OFX could promote current intensity, indicating that OFX as the electron-rich pollutant could boost the electron transfer process between host and guest molecules. This result was also in agreement with the linear sweep voltammetry (LSV) and PMS consumption experiments. The LSV curves (Supplementary Fig. 61a) showed that the current was very small and almost negligible in the individual Co2(V4O12)(bpy)2 system. Compared to the LSV curves in Co2(V4O12)(bpy)2/PMS system, the current increased significantly after adding OFX, indicating the degradation of OFX53. Additionally, the results of PMS consumption experiments (Supplementary Fig. 61b) indicated that the introduction of OFX could accelerate PMS decomposition.
Multi-channel electron transfer induced by polyvanadate in metal-organic framework for boosted peroxymonosulfate activation
