Flower-Shaped C-Dots/Co3O4{111} Constructed with Dual-Reaction Centers for Enhancement of Fenton-Like Reaction Activity and Peroxymonosulfate Conversion to Sulfate Radical

Novel flower-shaped C-dots/Co3O4{111} with dual-reaction centers were constructed to improve the Fenton-like reaction activity and peroxymonosulfate (PMS) conversion to sulfate radicals. Due to the exposure of a high surface area and Co3O4{111} facets, flower-shaped C-dots/Co3O4{111} could provide more Co(II) for PMS activation than traditional spherical Co3O4{110}. Meanwhile, PMS was preferred for adsorption on Co3O4{111} facets because of a high adsorption energy and thereby facilitated the electron transfer from Co(II) to PMS. More importantly, the Co–O–C linkage between C-dots and Co3O4{111} induced the formation of the dual-reaction center, which promoted the production of reactive organic radicals (R•). PMS could be directly reduced to SO4−• by R• over C-dots. On the other hand, electron transferred from R• to Co via Co–O–C linkage could accelerate the redox of Co(II)/(III), avoiding the invalid decomposition of PMS. Thus, C-dots doped on Co3O4{111} improved the PMS conversion rate to SO4−• over the single active site, resulting in high turnover numbers (TONs). In addition, TPR analysis indicated that the optimal content of C-dots doped on Co3O4{111} is 2.5%. More than 99% of antibiotics and dyes were degraded over C-dots/Co3O4{111} within 10 min. Even after six cycles, C-dots/Co3O4{111} still remained a high catalytic activity.


Introduction
Being attributed to higher oxidative potential (E 0 = 2.5-3.1 V) and longer half-life (t 1/2 = 30-40 µs), sulfate radical (SO 4 − •) could degrade organics more efficiently than OH• (E 0 = 2.80 V, t 1/2 = <1 µs) in neutral solutions. Generally, a Fenton-like reaction was significantly influenced by pH. In neutral/alkaline condition, transition metal ion might form precipitate, resulting in preventing the reaction. On the other hand, SO 4 − • could react with OH − to produce OH• with lower oxidative ability. To enhance the peroxymonosulfate (PMS) conversion to SO 4 − •, various transition metals (Fe, Cu, Mn, and Co) were investigated as activator [1][2][3]. Among them, cobalt oxides [CoO, CoO 2 , CoO(OH), Co 2 O 3 , and Co 3 O 4 ] were considered as the most promising catalysts due to high standard reduction potential (E 0 = +1.92 V vs. NHE) and redox of Co(III)/Co(II) [1]. During the process of a Fenton-like reaction, PMS can not only react with Co(II) as the electron acceptor but also reacts with Co(III) as the electron donor, accompanied by the and the generation of SO4 − • or SO5 − • [4]. However, similar to classic Fenton reaction, the rate-limiting step [Co(III) + HSO5 − → Co(II) + SO5 − • + H + ] and low conversion PMS to SO5 − • seriously restrict the application of cobalt oxides as catalysts [5]. In addition, the active components of catalysts might be converted into low activity substances during the reaction process due to poor stability. For example, cobalt oxides might be corroded and dissolved during the redox process, resulting in deterioration of catalytic performance.
In general, catalysts activity and stability could be influenced by surface morphology and crystal phase. Firstly, the mass transfer between organics and catalyst active sites was affected by the catalysts morphology. Li et al. constructed FexCo3−xO4 with porous nanocages structure, which indicated that such morphology significantly improved the catalytic activity and stability [6]. Secondly, atoms arrangement of catalyst, which led to exposure of transition metal with different chemical value, was depended on the crystal structure [7,8]. As for Co3O4, both Co( Ⅲ ) and Co(Ⅱ) occupied lattice point of octahedron {110} facet, while Co(Ⅱ) entirely occupied the lattice point of tetrahedron {111} facet [8,9]. More importantly, different crystal facets possessed different adsorption energies and electrons transfer properties, which might directly influence the catalytic reaction. Hensen et al synthesized CeO2 with different crystal planes to investigate its effect on CO oxidation, which found that the free energy barrier of Pd/CeO2(100) for the CO catalytic cycle was higher than that of Pd/CeO2(111) [10]. Thus, it is expected that Co3O4 constructed by tetrahedron{111} can not only provide more Co(Ⅱ) for PMS activation, but also possess higher adsorption energies and electrons transfer properties for PMS activation.
On the other hand, although increase of Co(Ⅱ) exposed can obviously improve the catalytic activity, Co(Ⅲ) would be still accumulated in the system because of the rate limitation step [Co(III) + HSO5 − → Co(II) + SO5 − • + H + ] with the reaction time. Since the oxidative potential of SO5 − • is relatively low, the degradation rate of organics by SO5 − • is much lower than by SO4 − •. Thus, the selective PMS conversion to SO4 − • over cobalt oxides still remained to be improved. In general, cation-π interaction could be formed between the cations and aromatic systems, which could be applied in the classic Fenton system for improvement of transition metal redox [11]. Zubir et al. demonstrated that unpaired π electrons of GO could transfer electrons between GO and iron centers via cation-π interactions, which significantly improved the recyclability of Fe3O4 [12]. In addition, the active transition complex [C-H2O2], which was formed between covalent carbon networks and H2O2, can directly transfer electron from the π-system to H2O2 with the generation of OH• [13]. Analogous to g−C3N4, carbon quantum dots (C-dots) are considered as a potential carbon-based co-catalyst due to their graphene structure [14,15]. Wang et al. synthesized the SiO2@C-dots/phosphotungstate catalyst with an inert SiO2 core and a catalytic active shell made up of the prepared amphiphilic phosphotungstate and C-dots, which proved that C-dots were an efficient co-catalyst in the system [16]. Our previous work also demonstrated that an electron-rich Cu center and electron-deficient π-electron conjugated system could be constructed by doping of g−C3N4 in Cu-Al2O3, which provided two electron-transfer routes for OH• generation in the presence of H2O2 [11]. Meanwhile, addition of C-dots in the hydrothermal system could modulate the formation of crystal structure because of the large number of oxygen functional groups on the surface of C-dots, which act as "surfactants". Thus, it is expected to simultaneously enhance the catalytic activity and selective PMS conversion to SO4 − • by doping of C-dots into Co3O4{111}. Notably, Cdots can improve the production of reactive organic radicals (R•) from the reaction of SO4 − • and organics. Such reactive organic radicals play an important role in the enhancement of catalytic activity and selective PMS conversion to SO4 − •. Some of electrons can be transferred from R• to Co(III) via Co-O-C linkage for Co(III) reduction to Co(Ⅱ) without PMS decomposition, and other electrons can be transferred from R• to PMS over C-dots with the generation of SO4 − •.
Based on the above research background and conception, flower-shaped Cdots/Co3O4{111} with high active crystal facets are constructed for improvement of catalytic activity and selective PMS conversion to SO4 − •. Being attributed to {111} facets,  [4]. However, similar to classic Fenton reaction, the rate-limiting step [Co(III) + HSO5 − → Co(II) + SO5 − • + H + ] and low conversion PMS to SO5 − • seriously restrict the application of cobalt oxides as catalysts [5]. In addition, the active components of catalysts might be converted into low activity substances during the reaction process due to poor stability. For example, cobalt oxides might be corroded and dissolved during the redox process, resulting in deterioration of catalytic performance.
In general, catalysts activity and stability could be influenced by surface morphology and crystal phase. Firstly, the mass transfer between organics and catalyst active sites was affected by the catalysts morphology.  with porous nanocages structure, which indicated that such morphology significantly improved the catalytic activity and stability [6]. Secondly, atoms arrangement of catalyst, which led to exposure of transition metal with different chemical value, was depended on the crystal structure [7,8]. As for Co3O4, both Co( Ⅲ ) and Co(Ⅱ) occupied lattice point of octahedron {110} facet, while Co(Ⅱ) entirely occupied the lattice point of tetrahedron {111} facet [8,9]. More importantly, different crystal facets possessed different adsorption energies and electrons transfer properties, which might directly influence the catalytic reaction. Hensen et al synthesized CeO2 with different crystal planes to investigate its effect on CO oxidation, which found that the free energy barrier of Pd/CeO2 (100) for the CO catalytic cycle was higher than that of Pd/CeO2(111) [10]. Thus, it is expected that Co3O4 constructed by tetrahedron{111} can not only provide more Co(Ⅱ) for PMS activation, but also possess higher adsorption energies and electrons transfer properties for PMS activation.
On the other hand, although increase of Co(Ⅱ) exposed can obviously improve the catalytic activity, Co(Ⅲ) would be still accumulated in the system because of the rate limitation step [Co(III) + HSO5 − → Co(II) + SO5 − • + H + ] with the reaction time. Since the oxidative potential of SO5 − • is relatively low, the degradation rate of organics by SO5 − • is much lower than by SO4 − •. Thus, the selective PMS conversion to SO4 − • over cobalt oxides still remained to be improved. In general, cation-π interaction could be formed between the cations and aromatic systems, which could be applied in the classic Fenton system for improvement of transition metal redox [11]. Zubir et al. demonstrated that unpaired π electrons of GO could transfer electrons between GO and iron centers via cation-π interactions, which significantly improved the recyclability of Fe3O4 [12]. In addition, the active transition complex [C-H2O2], which was formed between covalent carbon networks and H2O2, can directly transfer electron from the π-system to H2O2 with the generation of OH• [13]. Analogous to g−C3N4, carbon quantum dots (C-dots) are considered as a potential carbon-based co-catalyst due to their graphene structure [14,15]. Wang et al. synthesized the SiO2@C-dots/phosphotungstate catalyst with an inert SiO2 core and a catalytic active shell made up of the prepared amphiphilic phosphotungstate and C-dots, which proved that C-dots were an efficient co-catalyst in the system [16]. Our previous work also demonstrated that an electron-rich Cu center and electron-deficient π-electron conjugated system could be constructed by doping of g−C3N4 in Cu-Al2O3, which provided two electron-transfer routes for OH• generation in the presence of H2O2 [11]. Meanwhile, addition of C-dots in the hydrothermal system could modulate the formation of crystal structure because of the large number of oxygen functional groups on the surface of C-dots, which act as "surfactants". Thus, it is expected to simultaneously enhance the catalytic activity and selective PMS conversion to SO4 − • by doping of C-dots into Co3O4{111}. Notably, Cdots can improve the production of reactive organic radicals (R•) from the reaction of SO4 − • and organics. Such reactive organic radicals play an important role in the enhancement of catalytic activity and selective PMS conversion to SO4 − •. Some of electrons can be transferred from R• to Co(III) via Co-O-C linkage for Co(III) reduction to Co(Ⅱ) without PMS decomposition, and other electrons can be transferred from R• to PMS over C-dots with the generation of SO4 − •. Based on the above research background and conception, flower-shaped Cdots/Co3O4{111} with high active crystal facets are constructed for improvement of catalytic activity and selective PMS conversion to SO4 − •. Being attributed to {111} facets, ) entirely occupied the lattice point of tetrahedron {111} facet [8,9]. More importantly, different crystal facets possessed different adsorption energies and electrons transfer properties, which might directly influence the catalytic reaction. Hensen et al synthesized CeO 2 with different crystal planes to investigate its effect on CO oxidation, which found that the free energy barrier of Pd/CeO 2 (100) for the CO catalytic cycle was higher than that of Pd/CeO 2 (111) [10]. Thus, it is expected that Co 3 O 4 constructed by tetrahedron{111} can not only provide more Co(  [5]. In addition, the active components of catalysts might be converted into low activity substances during the reaction process due to poor stability. For example, cobalt oxides might be corroded and dissolved during the redox process, resulting in deterioration of catalytic performance. In general, catalysts activity and stability could be influenced by surface morphology and crystal phase. Firstly, the mass transfer between organics and catalyst active sites was affected by the catalysts morphology. Li et al. constructed FexCo3−xO4 with porous nanocages structure, which indicated that such morphology significantly improved the catalytic activity and stability [6]. Secondly, atoms arrangement of catalyst, which led to exposure of transition metal with different chemical value, was depended on the crystal structure [7,8]. As for Co3O4, both Co( Ⅲ ) and Co(Ⅱ) occupied lattice point of octahedron {110} facet, while Co(Ⅱ) entirely occupied the lattice point of tetrahedron {111} facet [8,9]. More importantly, different crystal facets possessed different adsorption energies and electrons transfer properties, which might directly influence the catalytic reaction. Hensen et al synthesized CeO2 with different crystal planes to investigate its effect on CO oxidation, which found that the free energy barrier of Pd/CeO2(100) for the CO catalytic cycle was higher than that of Pd/CeO2(111) [10]. Thus, it is expected that Co3O4 constructed by tetrahedron{111} can not only provide more Co(Ⅱ) for PMS activation, but also possess higher adsorption energies and electrons transfer properties for PMS activation.
On the other hand, although increase of Co(Ⅱ) exposed can obviously improve the catalytic activity, Co(Ⅲ) would be still accumulated in the system because of the rate limitation step [Co(III) + HSO5 − → Co(II) + SO5 − • + H + ] with the reaction time. Since the oxidative potential of SO5 − • is relatively low, the degradation rate of organics by SO5 − • is much lower than by SO4 − •. Thus, the selective PMS conversion to SO4 − • over cobalt oxides still remained to be improved. In general, cation-π interaction could be formed between the cations and aromatic systems, which could be applied in the classic Fenton system for improvement of transition metal redox [11]. Zubir et al. demonstrated that unpaired π electrons of GO could transfer electrons between GO and iron centers via cation-π interactions, which significantly improved the recyclability of Fe3O4 [12]. In addition, the active transition complex [C-H2O2], which was formed between covalent carbon networks and H2O2, can directly transfer electron from the π-system to H2O2 with the generation of OH• [13]. Analogous to g−C3N4, carbon quantum dots (C-dots) are considered as a potential carbon-based co-catalyst due to their graphene structure [14,15]. Wang et al. synthesized the SiO2@C-dots/phosphotungstate catalyst with an inert SiO2 core and a catalytic active shell made up of the prepared amphiphilic phosphotungstate and C-dots, which proved that C-dots were an efficient co-catalyst in the system [16]. Our previous work also demonstrated that an electron-rich Cu center and electron-deficient π-electron conjugated system could be constructed by doping of g−C3N4 in Cu-Al2O3, which provided two electron-transfer routes for OH• generation in the presence of H2O2 [11]. Meanwhile, addition of C-dots in the hydrothermal system could modulate the formation of crystal structure because of the large number of oxygen functional groups on the surface of C-dots, which act as "surfactants". Thus, it is expected to simultaneously enhance the catalytic activity and selective PMS conversion to SO4 − • by doping of C-dots into Co3O4{111}. Notably, Cdots can improve the production of reactive organic radicals (R•) from the reaction of SO4 − • and organics. Such reactive organic radicals play an important role in the enhancement of catalytic activity and selective PMS conversion to SO4 − •. Some of electrons can be transferred from R• to Co(III) via Co-O-C linkage for Co(III) reduction to Co(Ⅱ) without PMS decomposition, and other electrons can be transferred from R• to PMS over C-dots with the generation of SO4 − •.
) for PMS activation, but also possess higher adsorption energies and electrons transfer properties for PMS activation.
On the other hand, although increase of Co(  [5]. In addition, the active components of catalysts might be converted into low activity substances during the reaction process due to poor stability. For example, cobalt oxides might be corroded and dissolved during the redox process, resulting in deterioration of catalytic performance. In general, catalysts activity and stability could be influenced by surface morphology and crystal phase. Firstly, the mass transfer between organics and catalyst active sites was affected by the catalysts morphology. Li et al. constructed FexCo3−xO4 with porous nanocages structure, which indicated that such morphology significantly improved the catalytic activity and stability [6]. Secondly, atoms arrangement of catalyst, which led to exposure of transition metal with different chemical value, was depended on the crystal structure [7,8]. As for Co3O4, both Co( Ⅲ ) and Co(Ⅱ) occupied lattice point of octahedron {110} facet, while Co(Ⅱ) entirely occupied the lattice point of tetrahedron {111} facet [8,9]. More importantly, different crystal facets possessed different adsorption energies and electrons transfer properties, which might directly influence the catalytic reaction. Hensen et al synthesized CeO2 with different crystal planes to investigate its effect on CO oxidation, which found that the free energy barrier of Pd/CeO2(100) for the CO catalytic cycle was higher than that of Pd/CeO2(111) [10]. Thus, it is expected that Co3O4 constructed by tetrahedron{111} can not only provide more Co(Ⅱ) for PMS activation, but also possess higher adsorption energies and electrons transfer properties for PMS activation.
On the other hand, although increase of Co(Ⅱ) exposed can obviously improve the catalytic activity, Co(Ⅲ) would be still accumulated in the system because of the rate limitation step [Co(III) + HSO5 − → Co(II) + SO5 − • + H + ] with the reaction time. Since the oxidative potential of SO5 − • is relatively low, the degradation rate of organics by SO5 − • is much lower than by SO4 − •. Thus, the selective PMS conversion to SO4 − • over cobalt oxides still remained to be improved. In general, cation-π interaction could be formed between the cations and aromatic systems, which could be applied in the classic Fenton system for improvement of transition metal redox [11]. Zubir et al. demonstrated that unpaired π electrons of GO could transfer electrons between GO and iron centers via cation-π interactions, which significantly improved the recyclability of Fe3O4 [12]. In addition, the active transition complex [C-H2O2], which was formed between covalent carbon networks and H2O2, can directly transfer electron from the π-system to H2O2 with the generation of OH• [13]. Analogous to g−C3N4, carbon quantum dots (C-dots) are considered as a potential carbon-based co-catalyst due to their graphene structure [14,15]. Wang et al. synthesized the SiO2@C-dots/phosphotungstate catalyst with an inert SiO2 core and a catalytic active shell made up of the prepared amphiphilic phosphotungstate and C-dots, which proved that C-dots were an efficient co-catalyst in the system [16]. Our previous work also demonstrated that an electron-rich Cu center and electron-deficient π-electron conjugated system could be constructed by doping of g−C3N4 in Cu-Al2O3, which provided two electron-transfer routes for OH• generation in the presence of H2O2 [11]. Meanwhile, addition of C-dots in the hydrothermal system could modulate the formation of crystal structure because of the large number of oxygen functional groups on the surface of C-dots, which act as "surfactants". Thus, it is expected to simultaneously enhance the catalytic activity and selective PMS conversion to SO4 − • by doping of C-dots into Co3O4{111}. Notably, Cdots can improve the production of reactive organic radicals (R•) from the reaction of SO4 − • and organics. Such reactive organic radicals play an important role in the enhancement of catalytic activity and selective PMS conversion to SO4 − •. Some of electrons can be transferred from R• to Co(III) via Co-O-C linkage for Co(III) reduction to Co(Ⅱ) without ) exposed can obviously improve the catalytic activity, Co(  [5]. In addition, the active components of catalysts might be converted into low activity substances during the reaction process due to poor stability. For example, cobalt oxides might be corroded and dissolved during the redox process, resulting in deterioration of catalytic performance. In general, catalysts activity and stability could be influenced by surface morphology and crystal phase. Firstly, the mass transfer between organics and catalyst active sites was affected by the catalysts morphology. Li et al. constructed FexCo3−xO4 with porous nanocages structure, which indicated that such morphology significantly improved the catalytic activity and stability [6]. Secondly, atoms arrangement of catalyst, which led to exposure of transition metal with different chemical value, was depended on the crystal structure [7,8]. As for Co3O4, both Co( Ⅲ ) and Co(Ⅱ) occupied lattice point of octahedron {110} facet, while Co(Ⅱ) entirely occupied the lattice point of tetrahedron {111} facet [8,9]. More importantly, different crystal facets possessed different adsorption energies and electrons transfer properties, which might directly influence the catalytic reaction. Hensen et al synthesized CeO2 with different crystal planes to investigate its effect on CO oxidation, which found that the free energy barrier of Pd/CeO2(100) for the CO catalytic cycle was higher than that of Pd/CeO2(111) [10]. Thus, it is expected that Co3O4 constructed by tetrahedron{111} can not only provide more Co(Ⅱ) for PMS activation, but also possess higher adsorption energies and electrons transfer properties for PMS activation.
On the other hand, although increase of Co(Ⅱ) exposed can obviously improve the catalytic activity, Co(Ⅲ) would be still accumulated in the system because of the rate limitation step [Co(III) + HSO5 − → Co(II) + SO5 − • + H + ] with the reaction time. Since the oxidative potential of SO5 − • is relatively low, the degradation rate of organics by SO5 − • is much lower than by SO4 − •. Thus, the selective PMS conversion to SO4 − • over cobalt oxides still remained to be improved. In general, cation-π interaction could be formed between the cations and aromatic systems, which could be applied in the classic Fenton system for improvement of transition metal redox [11]. Zubir et al. demonstrated that unpaired π electrons of GO could transfer electrons between GO and iron centers via cation-π interactions, which significantly improved the recyclability of Fe3O4 [12]. In addition, the active transition complex [C-H2O2], which was formed between covalent carbon networks and H2O2, can directly transfer electron from the π-system to H2O2 with the generation of OH• [13]. Analogous to g−C3N4, carbon quantum dots (C-dots) are considered as a potential carbon-based co-catalyst due to their graphene structure [14,15]. Wang et al. synthesized the SiO2@C-dots/phosphotungstate catalyst with an inert SiO2 core and a catalytic active shell made up of the prepared amphiphilic phosphotungstate and C-dots, which proved that C-dots were an efficient co-catalyst in the system [16]. Our previous work also demonstrated that an electron-rich Cu center and electron-deficient π-electron conjugated system could be constructed by doping of g−C3N4 in Cu-Al2O3, which provided two electron-transfer routes for OH• generation in the presence of H2O2 [11]. Meanwhile, addition of C-dots in the hydrothermal system could modulate the formation of crystal structure because of the large number of oxygen functional groups on the surface of C-dots, which act as "surfactants". Thus, it is expected to simultaneously enhance the catalytic activity and selective PMS conversion to SO4 − • by doping of C-dots into Co3O4{111}. Notably, Cdots can improve the production of reactive organic radicals (R•) from the reaction of SO4 − • and organics. Such reactive organic radicals play an important role in the enhancement of catalytic activity and selective PMS conversion to SO4 − •. Some of electrons can be ) would be still accumulated in the system because of the rate limitation step [Co(III) + HSO 5 − → Co(II) + SO 5 − • + H + ] with the reaction time. Since the oxidative potential of SO 5 − • is relatively low, the degradation rate of organics by SO 5 − • is much lower than by SO 4 − •. Thus, the selective PMS conversion to SO 4 − • over cobalt oxides still remained to be improved. In general, cation-π interaction could be formed between the cations and aromatic systems, which could be applied in the classic Fenton system for improvement of transition metal redox [11]. Zubir et al. demonstrated that unpaired π electrons of GO could transfer electrons between GO and iron centers via cationπ interactions, which significantly improved the recyclability of Fe 3 O 4 [12]. In addition, the active transition complex [C-H 2 O 2 ], which was formed between covalent carbon networks and H 2 O 2 , can directly transfer electron from the π-system to H 2 O 2 with the generation of OH• [13]. Analogous to g-C 3 N 4 , carbon quantum dots (C-dots) are considered as a potential carbon-based co-catalyst due to their graphene structure [14,15]. Wang et al. synthesized the SiO 2 @C-dots/phosphotungstate catalyst with an inert SiO 2 core and a catalytic active shell made up of the prepared amphiphilic phosphotungstate and C-dots, which proved that C-dots were an efficient co-catalyst in the system [16]. Our previous work also demonstrated that an electron-rich Cu center and electron-deficient π-electron conjugated system could be constructed by doping of g-C 3 N 4 in Cu-Al 2 O 3 , which provided two electron-transfer routes for OH• generation in the presence of H 2 O 2 [11]. Meanwhile, addition of C-dots in the hydrothermal system could modulate the formation of crystal structure because of the large number of oxygen functional groups on the surface of C-dots, which act as "surfactants". Thus, it is expected to simultaneously enhance the catalytic activity and selective PMS conversion to SO 4 − • by doping of C-dots into Co 3 O 4 {111}. Notably, C-dots can improve the production of reactive organic radicals (R•) from the reaction of SO 4 − • and organics. Such reactive organic radicals play an important role in the enhancement of catalytic activity and selective PMS conversion to SO 4 − •. Some of electrons can be transferred from R• to Co(III) via Co-O-C linkage for Co(III) reduction to Co( 2 of 20 . However, similar to classic Fenton reaction, the tep [Co(III) + HSO5 − → Co(II) + SO5 − • + H + ] and low conversion PMS to SO5 − • rict the application of cobalt oxides as catalysts [5]. In addition, the active f catalysts might be converted into low activity substances during the reacue to poor stability. For example, cobalt oxides might be corroded and disthe redox process, resulting in deterioration of catalytic performance. l, catalysts activity and stability could be influenced by surface morphology ase. Firstly, the mass transfer between organics and catalyst active sites was he catalysts morphology. Li et al. constructed FexCo3−xO4 with porous ucture, which indicated that such morphology significantly improved the ity and stability [6]. Secondly, atoms arrangement of catalyst, which led to ransition metal with different chemical value, was depended on the crystal ]. As for Co3O4, both Co( Ⅲ ) and Co(Ⅱ) occupied lattice point of octaheet, while Co(Ⅱ) entirely occupied the lattice point of tetrahedron {111} facet portantly, different crystal facets possessed different adsorption energies transfer properties, which might directly influence the catalytic reaction. ynthesized CeO2 with different crystal planes to investigate its effect on CO ich found that the free energy barrier of Pd/CeO2(100) for the CO catalytic her than that of Pd/CeO2(111) [10]. Thus, it is expected that Co3O4 constructed n{111} can not only provide more Co(Ⅱ) for PMS activation, but also possess tion energies and electrons transfer properties for PMS activation.

Characterization of Catalysts
Cetrimonium Bromide (CTMAB) showed significant influence on the surface morphology of Co 3 O 4 during the hydrothermal reaction process. Co 3 O 4 synthesized by hydrothermal reaction without CTMAB are spherical particles with the diameter of 2 µm (Figure 1a). However, due to a decrease of the surface energy, ethanol and cobalt salts could form nano-sheets in the present of CTMAB. Thus, the product of Co 3 O 4 and Cdots/Co 3 O 4 {111}, which are synthesized in the present of CTMAB, shows a flower-shaped structure (diameter = 5 µm) with a nano-sheets thickness of 67 nm (Figure 1b,c). Notably, the flower-shaped structure increased the surface area and pore volume of catalyst, which could be confirmed by the results of the BET surface area (Table 1). The BET surface area of flower-shaped C-dots/Co 3 O 4 {111} was 109.11 m 2 /g, which was much higher than that of spherical Co 3 O 4 {110} (35.45 m 2 /g). A large specific surface area is beneficial to improving the catalytic activity of the catalyst. Such difference between spherical Co 3 O 4 {110} and flower-shaped Co 3 O 4 {111} was due to that surfactants were preferentially adsorbed on crystal faces, which limited the directional growth of nanocrystals and stabilizing the crystal faces. Various ordered aggregates formed by self-assembly of surfactants in solution can be used as microreactors or templates, which could control the morphology of nano-materials [19]. Thus, addition of CTMAB probably played the dominant role in the formation of {111} crystal structure of Co 3 O 4 during the process of hydrothermal reaction. In addition, TEM images (Figure 1f) confirm that the C-dots are successfully doped in the flower-shaped Co 3 O 4 {111} in which the lattice fringes spacing of 0.322 nm is in accordance with the (002) lattice planes of graphitic carbon [20]. EDS images further demonstrate that the C element is homogenously distributed on the catalyst surfaces (Figure 1g), which show that the C-dots are uniformly doped on Co 3 O 4 .  [21]. In general, the Co 3 O 4 crystal is a typical spinel structure (Co 2+ (Co 3+ ) 2 O 2-4 ), which consists of octahedron (CoO 6 ) and tetrahedron (CoO 4 ). Notably, Co(III) species are located in octahedral sites while Co(II) species are in tetrahedral sites (Figure 2b). Since Co(II) is the catalytic site for PMS select conversion to SO 4 − • [22], tetrahedron structure (CoO 4 ) with more Co(II) exposed can provide more catalytic sites for PMS activation. Therefore, in perspective of atomic arrangement, accessible Co(II) with PMS is beneficial for improvement of catalytic activity. To a certain degree, it can be considered that the tetrahedron structure (CoO 4 ) shows higher catalytic activity compared with octahedron structure (CoO 6 ). Furthermore, as shown in TEM (Figure 1d), the 'd' spacing of lattice fringe corresponding to (111) and (220) planes of spherical Co3O4 are respectively 0.464 nm and 0.284 nm, and the inter planar angles between these two planes are 90°, suggesting the {110} crystal structure of the spherical Co3O4 [17]. Different from spherical Co3O4{110}, the 'd' spacing of lattice fringe corresponding to (220) and (442) planes of flower-shaped Co3O4{111} are respectively 0.286 nm and 0.167 nm (Figure 1e), and the inter planar angle between these two planes is 30°, suggesting the {111} crystal structure of flower-shaped Co3O4{111} [18]. Such difference between spherical Co3O4{110} and flower-shaped Co3O4{111} was due to that surfactants were preferentially adsorbed on crystal faces, which limited the directional growth of nanocrystals and stabilizing the crystal faces. Various ordered aggregates formed by self-assembly of surfactants in solution can be used as microreactors or templates, which could control the morphology of nano-materials [19]. Thus, addition of CTMAB probably played the dominant role in the formation of {111} crystal structure of Co3O4 during the process of hydrothermal reaction. In addition, TEM images ( Figure 1f) confirm that the C-dots are successfully doped in the flower-shaped Co3O4{111} in which the lattice fringes spacing of 0.322 nm is in accordance with the (002) lattice planes of graphitic carbon [20]. EDS images further demonstrate that the C element is homogenously distributed on the catalyst surfaces (Figure 1g), which show that the Cdots are uniformly doped on Co3O4.
XRD spectra clearly describe the crystal structure of catalysts synthesized under different hydrothermal reaction conditions (Figure 2a)  . Since there was no impurity peaks in XRD spectra, which could confirm the excellent crystalline of flower-shaped Co3O4{111}, spherical Co3O4{110}, and flower-shaped C-dots/Co3O4{111} [21]. In general, the Co3O4 crystal is a typical spinel structure (Co 2+ (Co 3+ )2O2-4), which consists of octahedron (CoO6) and tetrahedron (CoO4). Notably, Co(III) species are located in octahedral sites while Co(II) species are in tetrahedral sites (Figure 2b). Since Co(II) is the catalytic site for PMS select conversion to SO4 − • [22], tetrahedron structure (CoO4) with more Co(II) exposed can provide more catalytic sites for PMS activation. Therefore, in perspective of atomic arrangement, accessible Co(II) with PMS is beneficial for improvement of catalytic activity. To a certain degree, it can be considered that the tetrahedron structure (CoO4) shows higher catalytic activity compared with octahedron structure (CoO6).           The redox properties of catalysts were characterized by the H 2 -TPR (Figure 4b). The β peak represents the reduction of Co(II) to Co 0 in Co 3 O 4 (Equation (2)), while the α peak represents the reduction of Co(III) to Co(II) (Equation (3)). The α peak and β peak of spherical Co 3 O 4 {110} are 319 • C and 398 • C, respectively. Due to the oxygen vacancy on the surface of flower-shaped Co 3 O 4 {111} and C-dots/Co 3 O 4 {111} (Figure S3), the reduction temperature (Co 2+ → Co 0 , Co 3+ → Co 2+ ) decreased obviously. The surface oxygen vacancy can promote the reduction of the catalyst, which leads to the reduction of the flower-shaped Co 3  {111} with different C-dots amounts are provided to explore the optimal doping of C-dots for PMS activation ( Figure S2). The temperature of Co 3+ reduction to Co 2+ decreases gradually as the C-dots content increase from 0.5% to 2.5%, which indicates that doping of C-dots can improve the reducibility of Co 3 O 4 {111}. However, the temperature of Co 3+ reduction to Co 2+ increases as the C-dots content increased from 2.5% to 3.5%, suggesting that excessive doping of C-dots weaken the reduction ability of Co 3 O 4 . Similar results are also observed in TPR results of Co 2+ reduction to Co 0 . Such adverse effect can be explained by that excess C-dots could cover the surface of the catalyst, resulting in the inhibition of the Co 3 O 4 reduction by H 2 . Thus, C-dots (2.5%)/Co 3 O 4 {111} were selected for further studies.
Cyclic voltammetry (CV) was used to test the charge transfer at the interfacial region of Co 3 O 4 /PMS in a three-electrode system (Figure 4c). Ag/AgCl was used as reference electrode, platinum was used as pair electrodes, and glassy carbon was used as working electrode. The mixture of 0.1 M Na 2 SO 4  3CoO

Catalytic Activity and Stability
The degradation rate of OTC and ENR by PMS over flower-shaped C-dots/Co 3 O 4 {111} are much higher than that over other catalysts (Figure 5a,c). More than 99.3% OTC was degraded in 5 min (PMS = 0.075 mM), while more than 99.7% ENR was degraded in 10 min over the flower-shaped C-dots(2.0 wt%)/Co 3 O 4 {111} (PMS = 0.3 mM). The degradation kinetics by PMS over various catalysts can be well described by pseudo-first-order kinetic model (−lnC/C 0 = kt), where 'C 0 ' is the initial concentration of OTC, 'C' is the concentration of reactant at time 't', and 'k' is the rate constant. The rate constants followed the order of flower-shaped C-dots/Co 3 (Figure 5b,d). Accordingly, the addition of C-dots could obviously improve the catalytic activity of Co 3 O 4 . The optimal content of C-dots doped in flower-shaped Co 3 O 4 {111} was 2%.   The influences of catalyst dosage and oxidant concentration on the catalytic degradation of ENR were tested. The degradation rate of ENR increases from 56.4% to 99.4% within 10 min while the dosage of flower-shaped C-dots/Co3O4{111} increases from 0.01 to 0.05 g·L −1 ( Figure S3a). This was because an increase of the catalysts dosage could supply more active sites for catalytic generation of radicals, which promoted the antibiotics degradation. Similar to the catalyst dosage, the increase of PMS concentration can also improve the degradation of ENR ( Figure S3b). As the PMS concentration increased from 0.05 to 0.3 mM, the removal rate of ENR increased from 74.2% to 99.8% within 10 min. Since the generation rate of the radicals was related to the catalyst dosage, further increase of PMS concentration from 0.3 to 0.7 mM could not obviously improve the removal rate of ENR anymore. Therefore, the optimal concentration of PMS for ENR degradation was 0.3 mM.
Since high temperature is beneficial to PMS activation and generation of radicals, the removal rate of ENR obviously increases as the temperature increases from 25 °C to 55 °C ( Figure S3c). The degradation kinetic constants also followed the order of K55 °C (2.837 min −1 ) > K45 °C (1.355 min −1 ) > K35 °C (0.913 min −1 ) > K25 °C (0.469 min −1 ). More than 99% of ENR could be degraded at 25 °C within 10 min. In consideration of energy consumption, the room temperature (25 °C) was selected for subsequent catalytic degradation experiments. Furthermore, due to changing the surface charge of catalyst, solution pH also influenced the degradation process obviously. The degradation rate of ENR increases as the initial solution pH increases from 3.0 to 6.0 ( Figure S3d). Only 50% of ENR was removed The influences of catalyst dosage and oxidant concentration on the catalytic degradation of ENR were tested. The degradation rate of ENR increases from 56.4% to 99.4% within 10 min while the dosage of flower-shaped C-dots/Co 3 O 4 {111} increases from 0.01 to 0.05 g·L −1 ( Figure S3a). This was because an increase of the catalysts dosage could supply more active sites for catalytic generation of radicals, which promoted the antibiotics degradation. Similar to the catalyst dosage, the increase of PMS concentration can also improve the degradation of ENR ( Figure S3b). As the PMS concentration increased from 0.05 to 0.3 mM, the removal rate of ENR increased from 74.2% to 99.8% within 10 min. Since the generation rate of the radicals was related to the catalyst dosage, further increase of PMS concentration from 0.3 to 0.7 mM could not obviously improve the removal rate of ENR anymore. Therefore, the optimal concentration of PMS for ENR degradation was 0.3 mM.
Since high temperature is beneficial to PMS activation and generation of radicals, the removal rate of ENR obviously increases as the temperature increases from 25 • C to 55 • C ( Figure S3c). The degradation kinetic constants also followed the order of K 55 • C (2.837 min −1 ) > K 45 • C (1.355 min −1 ) > K 35 • C (0.913 min −1 ) > K 25 • C (0.469 min −1 ). More than 99% of ENR could be degraded at 25 • C within 10 min. In consideration of energy consumption, the room temperature (25 • C) was selected for subsequent catalytic degradation experiments. Furthermore, due to changing the surface charge of catalyst, solution pH also influenced the degradation process obviously. The degradation rate of ENR increases as the initial solution pH increases from 3.0 to 6.0 ( Figure S3d). Only 50% of ENR was removed at pH 3.0 within 20 min while ENR was almost completely degraded at pH 6.0. This is because that Co leached from flower-shaped C-dots/Co 3 O 4 {111} in acid conditions was unfavorable to PMS activation. Notably, the concentrations of Co leached from catalysts in solution could meet the limit of Environmental Quality Standards for Surface Water in China (1.0 mg/L) ( Figure S4). However, the degradation rate of ENR decreases as pH increases from 6.0 to 9.0, which indicated that alkaline conditions made adverse effect to catalytic degradation. This was because the large amount of OH − reacted with SO 4 − • to generate OH•, resulting in lower oxidation ability in neutral/basic conditions.
In order to evaluate the reusability of flower-shaped C-dots/Co 3 O 4 {111}, OTC (10 mg/L), and ENR(10 mg/L) solutions are used to further test the catalytic degradation stability of C-dots/Co 3 O 4 {111} in the present work (Figure 6a,b). After six cycling runs, the degradation rate of OTC was still more than 90% within 2 min, and even 100% within 5 min (Figure 6a). Furthermore, the degradation kinetics did not decrease after the sixth cycle runs, showing a perfect catalytic stability of flower-shaped C-dots/Co 3 O 4 {111}. Similar to catalytic degradation of OTC, flower-shaped C-dots/Co 3 O 4 {111} also show high catalytic degradation rate of ENR, Rh B, and MB even after six cycles (Figure 6b-d), which also confirm the excellent stability of flower-shaped C-dots/Co 3 O 4 {111} for PMS activation.
In addition, the surface state of catalyst after six cycling runs is characterized by XPS spectra (Figure 6e). In this work, we propose that Co-O-C bonds conjugated in flower-shaped C-dots/Co 3 O 4 {111} facilitate electron transfer and correlate well with the transformation from R to R•. Such synergistic interaction between C-dots and Co 3 O 4 {111} could accelerate the Co(III)→Co(II) redox cycles. The bridge role played by C-dots could transfer electrons to Co(III) for its quick reduction as organic pollutants are oxidized [12]. XPS spectra of catalysts before and after reaction confirmed this inference. The value of Co(II)/Co(III) of flower-shaped C-dots/Co 3 O 4 {111} did not decreased after catalytic reaction (Figure 6f), which indicated that Co(II) was not oxidized to Co(III). This synergistic effect can not only effectively maintain the number of active sites of catalysts, but also reduce the ineffective decomposition of PMS. As for flower-shaped Co 3 O 4 {111}, the value of Co(II)/Co(III) decreased from 1.36 to 1.16 obviously after cycled uses (Figure 6f)

Density Functional Theory (DFT) Calculation
In general, adsorption of PMS on catalysts was the premise of activation process. According to the density functional theory (DFT), the adsorption energy of PMS on {110} and {111} facets of Co 3 O 4 are calculated by using the Materials Studio 7.0 CASTEP program ( Figure S5). The convergence criteria were as follows: the maximal force on the atoms was 0.03 eV Å −1 , the stress on the atomic nuclei was less than 0.05 GPa, the maximal atomic displacement was 0.001 Å, and the maximal energy change per atom was 1.0 e −5 eV.

Density Functional Theory (DFT) Calculation
In general, adsorption of PMS on catalysts was the premise of activation process. According to the density functional theory (DFT), the adsorption energy of PMS on {110} and {111} facets of Co3O4 are calculated by using the Materials Studio 7.0 CASTEP program ( Figure S5). The convergence criteria were as follows: the maximal force on the atoms was 0.03 eV Å −1 , the stress on the atomic nuclei was less than 0.05 GPa, the maximal atomic displacement was 0.001 Å, and the maximal energy change per atom was 1.0 e −5 eV. The adsorption energy of PMS (Eabs) on Co3O4 can be calculated by the equation of Eads = EPMS+Cat

The Possible Degradation Pathway of OTC and ENR
Generally, the unsaturated bonds of organic pollutants are easy to be broken in catalytic oxidation process. Based on the Frontier Orbital Theory, the highest occupied molecular orbital (HOMO) acts as the electron donor because of its weak binding force to electrons, while lowest unoccupied molecular orbital (LUMO) acts as electrons acceptor because of its strong binding force to electrons. In the present work, we calculated the bond energy of OTC and ENR by using Gaussian program, which indicates that the overlap of electronic clouds in LUMO and HOMO obviously influence the separation of e − and h + ( Figure S6). The large density of the electron cloud in OTC is mainly in the area of 1C, 9C, 14C, 24N. In the case of ENR, the electronic cloud mainly distributed in 3C, 6C, 19C, 20C. The active sites of pollutants were greatly contributed by the frontier orbitals, and unsaturated bonds strongly influenced the activity of antibiotic molecules. Thus, bonds in these positions of ENR and OTC are broken preferentially. In addition, the electron cloud density of OTC is much greater than that of ENR, which caused a higher degradation rate of OTC compared with ENR.
Furthermore, the intermediates of OTC and ENR degradation are identified by LC-MS/TOF, and iron spectra at different retention time (RT) are presented in ESI Scan (Figures S7 and S8) [20,[26][27][28][29]. Finally, all these substances were thoroughly decomposed into CO 2 [30][31][32][33]. Finally, these substances were decomposed to CO 2 , H 2 O, and other gaseous components. TOC removal over different catalyst is shown in Figure S9, which indicates that more that 60% organics are mineralized over flower-shaped C-dots/Co 3 O 4 {111}. Based on the above results, the possible catalytic degradation pathways of OTC and ENR are proposed (Figure 7).

Proposed Catalytic Mechanism
Generally, sulfate radicals (SO 4 − •) and hydroxyl radicals (OH•) are considered as the most important radicals for degradation of organic pollutants in the PMS activation system [34][35][36]. To identify which radical played the dominant role for the ENR degradation, classic quenching tests were carried out in which ethanol (EtOH) and tert-Butanol (TBA) were used as radical scavenger. In different quenching conditions, the reaction rate constants of sulfate radical and hydroxyl radical are shown in Table 3 [37,38]. As shown in Figure 8a, about 100% of ENR is degraded in 15 min in the absence of scavenger, while the degradation of ENR obviously decreased as the TBA was added in. Notably, compared with TBA, EtOH showed stronger inhibition to catalytic degradation of ENR. Thus, both hydroxyl radical and sulfate radical played the important role in catalytic degradation process.
To further elucidate the catalytic mechanism, DMPO-trapped EPR signals were detected in different aqueous dispersions of the corresponding samples with the addition of PMS (Figure 8). In the absence of catalysts, signals were not observed in the system of DMPO/PMS/ENR dispersion and the blank experiment ( Figure S10). As flower-shaped C-dots/Co 3 O 4 {111} were added, characteristic signals of DMPO-OH and DMPO-SO 4 were still not observed. However, unexpected characteristic signals which were ascribed to the oxidation products of DMPO (DMPOX) with the intensity ratios of 1:2:1:2:1:2:1 were captured ( Figure S10). Since the dosage of oxidant (PMS) was more than pollutants (ENR), the excessive PMS would generate active substance/radicals (OH•, SO 4 − •, H 2 O 2 ) which could oxidize DMPO to DMPOX [34,39]. As the PMS dosage decreased to 1.67 g/L, the characteristic peaks of DMPO-OH adduct (with hyperfine splitting constants of were used as radical scavenger. In different quenching conditions, the reaction rate constants of sulfate radical and hydroxyl radical are shown in Table 3 [37,38]. As shown in Figure 8a, about 100% of ENR is degraded in 15 min in the absence of scavenger, while the degradation of ENR obviously decreased as the TBA was added in. Notably, compared with TBA, EtOH showed stronger inhibition to catalytic degradation of ENR. Thus, both hydroxyl radical and sulfate radical played the important role in catalytic degradation process. Table 3. The reaction rate constants of ethanol (EtOH) and tert-Butanol (TBA) with hydroxyl radical and sulfate radical.

Radical Probo Reaction Rate Constant (M
(1.6-7.7) × 10 7 (1.2-2.8) × 10 9 Tert-Butanol (TBA) (4-9.1) × 10 5 (3.8-7.6) × 10 8 To further elucidate the catalytic mechanism, DMPO-trapped EPR signals were detected in different aqueous dispersions of the corresponding samples with the addition of PMS (Figure 8). In the absence of catalysts, signals were not observed in the system of DMPO/PMS/ENR dispersion and the blank experiment ( Figure S10). As flower-shaped Cdots/Co3O4{111} were added, characteristic signals of DMPO-OH and DMPO-SO4 were still not observed. However, unexpected characteristic signals which were ascribed to the oxidation products of DMPO (DMPOX) with the intensity ratios of 1:2:1:2:1:2:1 were captured ( Figure S10). Since the dosage of oxidant (PMS) was more than pollutants (ENR), the excessive PMS would generate active substance/radicals (OH•, SO4 − •, H2O2) which could oxidize DMPO to DMPOX [34,39]. As the PMS dosage decreased to 1.67 g/L, the characteristic peaks of DMPO-OH adduct (with hyperfine splitting constants of ɑN = ɑH = 14.6 G) and DMPO-SO4 adducts (with hyperfine splitting constants of ɑN = 12.7 G, ɑα H = 10.3 G, ɑβ H = 2.1 G, ɑγ H = 1.1 G) are observed, revealing the simultaneous presence of OH• and SO4 − •. The characteristic peaks of DMPO-SO4 − •and DMPO-OH• were observed in the dispersions of catalysts with their intensities following the order of flower-shape Cdots/Co3O4{111} > flower-shaped Co3O4{111} > spherical Co3O4{110}, which indicated that doping C-dots on Co3O4{111} could effectively improve the generation of sulfate radicals. were used as radical scavenger. In different quenching conditions, the reaction rate constants of sulfate radical and hydroxyl radical are shown in Table 3 [37,38]. As shown in Figure 8a, about 100% of ENR is degraded in 15 min in the absence of scavenger, while the degradation of ENR obviously decreased as the TBA was added in. Notably, compared with TBA, EtOH showed stronger inhibition to catalytic degradation of ENR. Thus, both hydroxyl radical and sulfate radical played the important role in catalytic degradation process. Table 3. The reaction rate constants of ethanol (EtOH) and tert-Butanol (TBA) with hydroxyl radical and sulfate radical.

Radical Probo Reaction Rate Constant (M
(1.6-7.7) × 10 7 (1.2-2.8) × 10 9 Tert-Butanol (TBA) (4-9.1) × 10 5 (3.8-7.6) × 10 8 To further elucidate the catalytic mechanism, DMPO-trapped EPR signals were detected in different aqueous dispersions of the corresponding samples with the addition of PMS (Figure 8). In the absence of catalysts, signals were not observed in the system of DMPO/PMS/ENR dispersion and the blank experiment ( Figure S10). As flower-shaped Cdots/Co3O4{111} were added, characteristic signals of DMPO-OH and DMPO-SO4 were still not observed. However, unexpected characteristic signals which were ascribed to the oxidation products of DMPO (DMPOX) with the intensity ratios of 1:2:1:2:1:2:1 were captured ( Figure S10). Since the dosage of oxidant (PMS) was more than pollutants (ENR), the excessive PMS would generate active substance/radicals (OH•, SO4 − •, H2O2) which could oxidize DMPO to DMPOX [34,39]. As the PMS dosage decreased to 1.67 g/L, the characteristic peaks of DMPO-OH adduct (with hyperfine splitting constants of ɑN = ɑH = 14.6 G) and DMPO-SO4 adducts (with hyperfine splitting constants of ɑN = 12.7 G, ɑα H = 10.3 G, ɑβ H = 2.1 G, ɑγ H = 1.1 G) are observed, revealing the simultaneous presence of OH• and SO4 − •. The characteristic peaks of DMPO-SO4 − •and DMPO-OH• were observed in the dispersions of catalysts with their intensities following the order of flower-shape Cdots/Co3O4{111} > flower-shaped Co3O4{111} > spherical Co3O4{110}, which indicated that doping C-dots on Co3O4{111} could effectively improve the generation of sulfate radicals. were used as radical scavenger. In different quenching conditions, the reaction rate constants of sulfate radical and hydroxyl radical are shown in Table 3 [37,38]. As shown in Figure 8a, about 100% of ENR is degraded in 15 min in the absence of scavenger, while the degradation of ENR obviously decreased as the TBA was added in. Notably, compared with TBA, EtOH showed stronger inhibition to catalytic degradation of ENR. Thus, both hydroxyl radical and sulfate radical played the important role in catalytic degradation process. Table 3. The reaction rate constants of ethanol (EtOH) and tert-Butanol (TBA) with hydroxyl radical and sulfate radical.

Radical Probo
(1.6-7.7) × 10 7 (1.2-2.8) × 10 9 Tert-Butanol (TBA) (4-9.1) × 10 5 (3.8-7.6) × 10 8 To further elucidate the catalytic mechanism, DMPO-trapped EPR signals were detected in different aqueous dispersions of the corresponding samples with the addition of PMS (Figure 8). In the absence of catalysts, signals were not observed in the system of DMPO/PMS/ENR dispersion and the blank experiment ( Figure S10). As flower-shaped Cdots/Co3O4{111} were added, characteristic signals of DMPO-OH and DMPO-SO4 were still not observed. However, unexpected characteristic signals which were ascribed to the oxidation products of DMPO (DMPOX) with the intensity ratios of 1:2:1:2:1:2:1 were captured ( Figure S10). Since the dosage of oxidant (PMS) was more than pollutants (ENR), the excessive PMS would generate active substance/radicals (OH•, SO4 − •, H2O2) which could oxidize DMPO to DMPOX [34,39]. were used as radical scavenger. In different quenching conditions, the reaction rate constants of sulfate radical and hydroxyl radical are shown in Table 3 [37,38]. As shown in Figure 8a, about 100% of ENR is degraded in 15 min in the absence of scavenger, while the degradation of ENR obviously decreased as the TBA was added in. Notably, compared with TBA, EtOH showed stronger inhibition to catalytic degradation of ENR. Thus, both hydroxyl radical and sulfate radical played the important role in catalytic degradation process. To further elucidate the catalytic mechanism, DMPO-trapped EPR signals were detected in different aqueous dispersions of the corresponding samples with the addition of PMS (Figure 8). In the absence of catalysts, signals were not observed in the system of DMPO/PMS/ENR dispersion and the blank experiment ( Figure S10). As flower-shaped Cdots/Co3O4{111} were added, characteristic signals of DMPO-OH and DMPO-SO4 were still not observed. However, unexpected characteristic signals which were ascribed to the oxidation products of DMPO (DMPOX) with the intensity ratios of 1:2:1:2:1:2:1 were captured ( Figure S10). Since the dosage of oxidant (PMS) was more than pollutants (ENR), the excessive PMS would generate active substance/radicals (OH•, SO4 − •, H2O2) which could oxidize DMPO to DMPOX [34,39].  Table 3 [37,38]. As shown in egraded in 15 min in the absence of scavenger, while decreased as the TBA was added in. Notably, comnger inhibition to catalytic degradation of ENR. Thus, adical played the important role in catalytic degradaethanol (EtOH) and tert-Butanol (TBA) with hydroxyl radi- 7 (1.2-2.8) × 10 9 (4-9.1) × 10 5 (3.8-7.6) × 10 8 tic mechanism, DMPO-trapped EPR signals were deons of the corresponding samples with the addition of catalysts, signals were not observed in the system of e blank experiment ( Figure S10). As flower-shaped Ccteristic signals of DMPO-OH and DMPO-SO4 were ected characteristic signals which were ascribed to the OX) with the intensity ratios of 1:2:1:2:1:2:1 were cape of oxidant (PMS) was more than pollutants (ENR), active substance/radicals (OH•, SO4 − •, H2O2) which 34,39]. As the PMS dosage decreased to 1.67 g/L, the adduct (with hyperfine splitting constants of ɑN = ɑH = ith hyperfine splitting constants of ɑN = 12.7 G, ɑα H = are observed, revealing the simultaneous presence of eaks of DMPO-SO4 − •and DMPO-OH• were observed their intensities following the order of flower-shape C-o3O4{111} > spherical Co3O4{110}, which indicated that effectively improve the generation of sulfate radicals. scavenger. In different quenching conditions, the reaction rate concal and hydroxyl radical are shown in Table 3 [37,38]. As shown in of ENR is degraded in 15 min in the absence of scavenger, while NR obviously decreased as the TBA was added in. Notably, com-H showed stronger inhibition to catalytic degradation of ENR. Thus, l and sulfate radical played the important role in catalytic degradate constants of ethanol (EtOH) and tert-Butanol (TBA) with hydroxyl radi- ate the catalytic mechanism, DMPO-trapped EPR signals were deeous dispersions of the corresponding samples with the addition of e absence of catalysts, signals were not observed in the system of persion and the blank experiment ( Figure S10). As flower-shaped Cadded, characteristic signals of DMPO-OH and DMPO-SO4 were wever, unexpected characteristic signals which were ascribed to the f DMPO (DMPOX) with the intensity ratios of 1:2:1:2:1:2:1 were capince the dosage of oxidant (PMS) was more than pollutants (ENR), ould generate active substance/radicals (OH•, SO4 − •, H2O2) which to DMPOX [34,39]. As the PMS dosage decreased to 1.67 g/L, the f DMPO-OH adduct (with hyperfine splitting constants of ɑN = ɑH = O4 adducts (with hyperfine splitting constants of ɑN = 12.7 G, ɑα H = ɑγ H = 1.1 G) are observed, revealing the simultaneous presence of haracteristic peaks of DMPO-SO4 − •and DMPO-OH• were observed atalysts with their intensities following the order of flower-shape Cwer-shaped Co3O4{111} > spherical Co3O4{110}, which indicated that 3O4{111} could effectively improve the generation of sulfate radicals.

Proposed Catalytic Mechanism
Generally, sulfate radicals (SO4 − •) and hydroxyl radicals (OH•) are considered as the most important radicals for degradation of organic pollutants in the PMS activation system [34][35][36]. To identify which radical played the dominant role for the ENR degradation, classic quenching tests were carried out in which ethanol (EtOH) and tert-Butanol (TBA) In addition, in spherical Co 3 O 4 {110} and flower-shaped Co 3 O 4 {111} aqueous suspensions, PMS was not only reduced to SO 4 − • (Equation (7)) but also oxidized to large amounts of SO 5 − • (Equation (9)) due to the decomposition of PMS following the classic Fenton reaction mechanism. Thus, with the oxidation from Co(II) to Co(III), the value of Co(II)/Co(III) [ (6) and (7)), R• could transfer the electron to Co (III) for its quick reduction via Co-O-C linkage, resulting in a stable value of Co(II)/Co(III) [flower-shaped C-dots/Co 3 O 4 {111}] before and after catalytic reaction (Figure 6f). Thus, such different mechanism of PMS activation over Co 3 O 4 and flower-shaped Co 3 O 4 {111} led to higher selective PMS conversion to SO 4 − • over flower-shaped C-dots/Co 3 O 4 {111}.  To further elucidate the catalytic mechanism, DMPO-trapped EPR signals were detected in different aqueous dispersions of the corresponding samples with the addition of PMS (Figure 8). In the absence of catalysts, signals were not observed in the system of DMPO/PMS/ENR dispersion and the blank experiment ( Figure S10). As flower-shaped Cdots/Co3O4{111} were added, characteristic signals of DMPO-OH and DMPO-SO4 were still not observed. However, unexpected characteristic signals which were ascribed to the oxidation products of DMPO (DMPOX) with the intensity ratios of 1:2:1:2:1:2:1 were captured ( Figure S10). Since the dosage of oxidant (PMS) was more than pollutants (ENR), the excessive PMS would generate active substance/radicals (OH•, SO4 − •, H2O2) which could oxidize DMPO to DMPOX [34,39]. As the PMS dosage decreased to 1.67 g/L, the characteristic peaks of DMPO-OH adduct (with hyperfine splitting constants of ɑN = ɑH = 14.6 G) and DMPO-SO4 adducts (with hyperfine splitting constants of ɑN = 12.7 G, ɑα H = 10.3 G, ɑβ H = 2.1 G, ɑγ H = 1.1 G) are observed, revealing the simultaneous presence of OH• and SO4 − •. The characteristic peaks of DMPO-SO4 − •and DMPO-OH• were observed in the dispersions of catalysts with their intensities following the order of flower-shape Cdots/Co3O4{111} > flower-shaped Co3O4{111} > spherical Co3O4{110}, which indicated that doping C-dots on Co3O4{111} could effectively improve the generation of sulfate radicals. In addition, in spherical Co3O4{110} and flower-shaped Co3O4{111} aqueous suspensions, PMS was not only reduced to SO4 − • (Equation (7)) but also oxidized to large amounts of SO5 − • (Equation (9)) due to the decomposition of PMS following the classic Fenton reaction mechanism. Thus, with the oxidation from Co(II) to Co(III), the value of In order to further illustrate the co-catalyst effect of C-dots on Co 3 O 4 {111}, PMS activation rate per second on a single active site was calculated. In classic Fenton reaction system, the concentration of OH• in the aqueous dispersion with H 2 O 2 could be directly measured by using the terephthalic acid (TPA) probe method. Accordingly, the turnover frequencies (TOF) of Fenton catalyst could be obtained from the conversion number of H 2 O 2 into OH• per second on a single active site [40]. However, TOF cannot be calculated in PMS activation system because it is difficult to accurately quantify the SO 4 − •. Thus, the turnover numbers (TONs) are selected to evaluate the catalytic efficiency for PMS activation (Equation (3)) [41]. TONs = n (converted reactants)/n (catalyst active sites). ( The number of active sites of catalysts can be calculated by the following method [42]: The specific surface area and Co 2+ /Co 3+ of the catalyst can be obtained by BET and XPS spectra, respectively. Moreover, the unit cell area can be calculated by the length of Co 3 O 4 cell [42]. Then, the PMS concentration was obtained by low-concentration iodide methods (ESI). As expected, C-dots/Co 3 O 4 {111} show higher TONs than Co 3 O 4 {111} ( Figure S11). Such improvement can be explained as follows: Besides of PMS reduction to SO 4 − • by Co(II), graphene structure of C-dots could facilitate the production of R• (R•, the product of SO 4 − •/OH• attacking to organics) (Equation (10)). Such reactive organic radicals promote the PMS conversion to SO 4 − • (Equation (11)). Electrons were not only transferred from R• to Co(III) via the Co-O-C linkage for Co(III) reduction to Co(II), but also transferred from R• to PMS over C-dots with the generation of SO 4 − • (Equation (12)). Thus, the PMS conversion rate is obviously increased after C-dots doped on Co 3 O 4 {111}.
Based on the above results, the possible catalytic mechanism of flower-shape Cdots/Co 3 O 4 {111} is proposed (Figure 9). Both Co 3 O 4 crystal structure and doped Cdots played an important role in PMS activation. On the one hand, flower-shaped Cdots/Co 3 O 4 {111} exposed more Co(II) for PMS activation with generation of SO 4 − •. On the other hand, graphene structure of C-dots facilitated the production of reactive organic radicals (R•) (Equation (10)). Some of electrons are transferred from R• to Co(III) via the Co-O-C linkage with the reduction of Co(III) to Co(II) (Equation (11)), which decrease the 'invalid decomposition' of PMS (Equation (9)). Other electrons are transferred from R• to PMS over C-dots with the generation of SO 4 − • (Equation (12)) (R* is the intermediate product from the reaction of R• and Co(III)/HSO 5 − ). Thus, there were three electron transfer routes in the catalytic process: the first route was from Co(II) to PMS with the generation of SO 4 − •, the second route was from R• to Co(III) via the Co-O-C linkage with the quick reduction of Co(III) to Co(II), and the third route was from R• to PMS via C-dots with the generation of SO 4 − •. Consequently, high catalytic activity and selective PMS conversion to SO 4 − • was achieved.

Catalytic Performance
Four aromatic pollutants (OTC, ENR, MB, and Rh B) were selected to evaluate the catalysts activity. Preliminary experiments for ENR degradation indicated that the optimal doses of the catalyst and PMS were 0.05 g/L and 12.5 mM, respectively ( Figure 6). Thus, these two doses were used in all catalytic degradation experiments unless otherwise specified. Typically, 200 mL aqueous solutions with certain pollutant and 0.01 g catalyst were placed in 250 mL beaker flasks. To ensure the adsorption-desorption equilibrium, the suspensions were magnetically stirred for 30 min before the catalytic reaction. Then, 12.5 mM PMS was added in the suspensions with magnetic stirring (120 rpm). At given time intervals, 2 mL aliquots were sampled and followed by adding methanol into the sample to quench the radicals for termination of the catalytic reaction, and subsequently filtrated through a Millipore filter (pore size 0.45 µm) prior to the analysis. In addition, to test the catalytic stability, the catalyst was recovered by filtration, washed with deionized water, dried, and reused in the following cycle. The degradation rate of pollutants is calculated. Concentrations of OTC and ENR were measured using Agilent 1200 HPLC (Agilent Technologies, Palo Alto, CA, USA) equipped with a quaternary pump and a UV detector. Chromatographic analysis was performed by a reversed-phase C-18 column (4.6 × 150 mm, 5 µm particle diameter), and the specific liquid conditions are listed in Table  S2. The degradation intermediates of OTC and ENR were detected by HPLC coupled with time-of-flight mass spectrometry detection (HPLC-TOFMS, Agilent 1290 Infinity LC/6460 QQQ MS) in positive polarity. Rh B and MB concentrations in solutions were analyzed by the UV−1800 UV-vis spectrometer (Shimadzu, Japan).

Conclusions
Novel flower-shaped C-dots/Co 3 O 4 {111} were constructed for the improvement of the catalytic activity and selective peroxymonosulfate (PMS) conversion to sulfate radicals. Being attributed to the exposure of {111} facets, C-dots/Co 3 O 4 {111} could provide plenty of Co(II) for PMS activation. In addition, the Co 3 O 4 {111} facet not only possessed high adsorption energy of PMS on catalyst, but also facilitated the electron transfer from Co(II) to PMS. On the other hand, the graphene structure of C-dots facilitated the production of reactive organic radicals (R•), which provided the electrons for Co(III) and PMS reduction. Thus, C-dots doped on Co 3 O 4 {111} could obviously enhance the TONs. The catalytic degradation tests indicated that antibiotics and dyes could be efficiently degraded over flower-shaped C-dots/Co 3 O 4 {111}. Even after six cycling runs, flower-shaped C-dots/Co 3 O 4 {111} still remained a high catalytic activity. Additionally, according to degradation intermediates identified by LC-MS/TOF and electron cloud density calculated by Gaussian program calculation, the possible degradation pathways of OTC/ENR were proposed. EPR spectra and radical capture experiments further demonstrated that both OH• and SO 4 − • played a dominant role in catalytic degradation processes. Thus, the present work provided a simple and efficient way for PMS activation, which could be applied in the treatment of wastewater.   Figure S6. Frontier electron densities of OTC and ENR calculated by Gaussian. Figure S7. Iron spectra at different retention time (RT) of OTC catalysis sample. Figure S8. Iron spectra at different retention time (RT) of ENR catalysis sample. Figure S9. TOC removal under different reaction conditions. Figure S10. EPR spectra in various conditions. Center field: 34,800 G; microwave frequency: 9.849 GHz; modulation frequency: 100 kHz; and power: 20.17 mW. Figure S11. The turnover numbers(TONs) of different catalysts prepared (Firstly, 10 mM potassium iodide (KI) solution was prepared to dilute the PMS sample 50 times. Then the solution was shaken for 5 minutes and detected with UV-vis spectrometer at λ = 352 nm). Figure S12. Schematic of synthesize process of flower-shape C-dots/Co 3 O 4 {111}. Table S1. The pseudo-first-order kinetic equations, rate constants (K) and regression coefficients (R 2 ) of degradation of OTC/ENR over different catalysts. Table S2.
Operating conditions for HPLC about OTC and ENR.