English

• Effective degradation of aromatic compounds produced in industry has been an impending task due to more and more serious environmental pollution problems. However, nowadays a great challenge still remains in completely decomposing aromatic compounds from the complicated multicomponent wastewater through the traditional one-step chemical or biological treatments [1, 2]. In recent years, the advanced oxidation processes (AOPs), including photochemical oxidation, catalytic wet air oxidation, electrochemical oxidation and Fenton oxidation [3-8], have been widely applied in treating toxic or non-biodegradable contaminants [10-13]. In comparison with the other AOPs, the Fenton oxidation method owns some unique advantages such as easy operation, mild reaction conditions, fast and non-selective destruction to organic pollutants, mainly due to the strong oxidative activity of hydroxyl radicals (•OH) produced from catalytic decomposition of hydrogen peroxide (H₂O₂) [8, 9]. The homogeneous Fenton reaction process is still widely used for the treatment of non-biodegradable organic pollutants in industry, which exhibits high degradation efficiency of organic-containing wastewater in low cost, but suffers the disadvantages of narrow working pH range of 2.0–3.5 and generation of a large amount of iron sludge and wasted acid solution [14], producing the second pollution to water and increasing the treatment cost as well. In order to avoid these drawbacks of the homogeneous reaction system, heterogeneous solid Fenton-type catalysts become the promising alternatives, because the catalytic reaction could be operated in the slurry bed reactors or the conventional continuous flow reactors without generation and separation of soluble iron salts or sludge even hopefully under mild operation conditions [15-17].

  In order to improve the activity and stability of the heterogeneous Fenton-like catalysts, transition metal salts with oxidation valence states, such as iron, copper, ruthenium, cerium and manganese have been widely adopted to be loaded on supports using ion-exchanging, wet impregnation or in situ hydrothermal synthesis methods. Among all the active components, iron salts are generally considered as the preferred chemicals and widely used for heterogeneous Fenton processes according to the economic and environmental concerns [18]. Besides, some porous supports with high surface area and open channels, such as natural layered minerals, porous aluminosilicates including clays [19-22], mesoporous silicas [23-25] and zeolites [26-29], could be utilized to load active iron species. It is well known that zeolites are a kind of microporous materials composed of crystalline aluminosilicates in framework, possessing particular properties such as strong surface acidity, ion-exchanging, shape selectivity, and high thermal and hydrothermal stability, which are widely used as the industrial heterogeneous catalyst at large scale [30]. Due to the rich specific crystalline microporosity and strong surface Brønsted acidity, zeolites can selectively adsorb organic molecules from wastewater and provide the required acidic environment in Fenton reaction as well [31]. Meanwhile, zeolites could supply huge surface for uniform dispersion of Fe₂O₃ with high loading capacity, which could stabilize iron species in catalysts and prevent them leaching into the solution. Therefore, the zeolite-based heterogeneous solid Fenton-like catalysts were announced to perform high activity, good stability and repeatability in degradation of organic contaminants in wastewater [31a].

  In order to enhance the catalytic activity and stability of Fenton reaction, a large number of approaches have been proposed to optimize the structural properties of iron-based catalysts including the supports, Fe³⁺/Fe²⁺ ratio and their dispersions. Zeolite Y (FAU type) and ZSM-5 (MFI type), the well-known and most widely used zeolites in industry, were generally applied as supports to load iron cations or iron oxide particles by the methods of ion-exchanging or wet impregnation, respectively [31a]. The catalyst of Fe₂O₃-loading HY zeolite shows high efficiency and stability in degradation of phenol by photo-Fenton method, however, the acidic solution of pH 3 was needed for the reaction [32]. The Brønsted acidity of catalyst was proven to be one of the important factors to affect the reaction activity positively, and the chemical oxygen demand (COD) removal was found to increase with the increment of Brønsted acidity of the catalyst [31b]. It has been reported that hydrophobic Fe-zeolites with high SiO₂/Al₂O₃ ratios (>200) showed a low ion exchange capacity of only 0.09 wt%, but it was regarded as a promising material for removal of MTBE from water [33]. It has been reported that iron containing particles with the size of 2–4 nm in the zeolite performed much higher catalytic efficiency [34].

  Besides, there have been few reports about the effect of size and dispersity of iron species on different types of zeolite to the degradation of organic contaminants by Fe₂O₃/zeolite composites as the heterogeneous Fenton-like catalysts. In our research here, the Fe₂O₃-loading catalysts with different size and dispersity of Fe₂O₃ species were controllably prepared on the supports of zeolite Y, ZSM-5 and mordenite (MOR type), which were then used as the heterogeneous Fenton-like catalysts for degradation of phenol in neutral pH solution as the mimetic organic wastewater. The Fe₂O₃ nanoparticles in size of 5 nm are found to act as the active centers, which can be highly dispersed on zeolite Y matrix (< 9 wt%). The effects of different reaction parameters on the phenol degradation were also explored in detail, such as concentration of H₂O₂, reaction temperature and initial pH values of phenol solution. The catalyst based on zeolite Y with Fe loading of 9% performed the best phenol degradation efficiency (> 90%) under the mild conditions of low temperature (323 K) and neutral system (pH 7). This excellent catalytic performance is ascribed to the combination of strong surface Brønsted acidity of zeolite Y matrix and octahedral Fe³⁺ in highly-dispersed 5 nm Fe₂O₃ nanoparticles for efficient decomposition of H₂O₂. Since the degradation reaction could be performed under mild conditions of ambient temperature (283–323 K) and a wide pH range (4.0–7.0), the catalyst can be easily recovered and show stable catalytic activity, which has the potential to be developed as an industrial Fenton-like catalyst in degradation of organic contaminants in wastewater.

  In a typical procedure, Fe-containing zeolites were prepared by wet impregnation in deionized water. In a typical process, 1.4 g Fe (NO₃)₃·9H₂O was dissolved in 20 mL deionized water and 2.0 g dehydrated zeolites were added into the solution. The mixture was stirred at 30 ℃ for 6 h, further dried at 80 ℃ under stirring, and then calcined in air at 600 ℃ for 6 h. All the samples denoted as Fex-y were synthesized with the same procedure described above, here, x represents the support, for example, Y means zeolite Y, M means mordenite, Z means ZSM-5, and y represents the mass content of loading Fe element.

  FeY-9 M was prepared by mechanical mixing method. First, 1.4 g Fe(NO₃)₃·9H₂O and 2.0 g dehydrated zeolite Y were mixed well and grinded for 10 min. Then the mixture was calcined in air at 600 ℃ for 6 h.

  FeY-IE was prepared by ion-exchanging and calcination method. 1.4 g Fe(NO₃)₃·9H₂O and 2.0 g dehydrated zeolites were dispersed into 20 mL deionized water, after stirring for 2 h, the sample was filtered and washed for 5 times, and then calcined in air at 600 ℃ for 6 h.

  Fig. 1 and Fig. S2 (Supporting information) show the powder Xray diffraction patterns (XRD) of parent zeolites and Fe-loading zeolite catalysts prepared by wet impregnation method. Parent zeolite Y, mordenite and ZSM-5 together with the corresponding Fe-loading zeolite catalysts are presented. From Figs. S2A and B, it can be seen that when Fe content is above 6 wt%, the typical diffraction peaks of Fe₂O₃ (2θ = 33° and 35.6°) are clearly observed (JCPDS Card No. 33-0664), which show much stronger and narrower as the Fe content is up to 9 wt%, indicating that the particle sizes of Fe₂O₃ on the supports of mordenite or ZSM-5 increase gradually as the increment of Fe contents based on the aggregation of Fe species. Calculated by Scherrer's equation, the grain size of Fe₂O₃ on mordenite is slightly bigger than that on ZSM-5 with the same content of Fe loading (Table S1 in Supporting information). Whereas, the diffraction peaks of Fe₂O₃ on zeolite Y could not be found when Fe content is lower than 9% (Fig. 1, curves b and c), mainly due to that the formed Fe₂O₃ nanoparticles are too small to be detected by XRD [35, 36], showing that Fe₂O₃ nanoparticles dispersed on zeolite Y are much smaller than that on mordenite and ZSM-5 as supports. When Fe content is increased even to 25 wt% (Fig. 1, curves d and f), the diffraction peaks of Fe₂O₃ on zeolite Y are still very weak, indicating that zeolite Y is the optimum support to highly disperse Fe species on its surface. The determined Fe contents in zeolite catalysts by XRF are quite close to that in the raw mixtures, showing that most of the Fe species are loaded on zeolite supports. The average size of Fe₂O₃ nanoparticles on zeolite Y also increases with the Fe content (Table 1), however, they are much smaller than those loaded in mordenite and ZSM-5 with the same Fe contents (Table S1).

  Figure 1

  

  Figure 1.  XRD patterns of the catalysts with different Fe contents based on zeolite Y as the support: zeolite Y (a), FeY-6 (b), FeY-9 (c), FeY-15 (d), FeY-20 (e) and FeY-25 (f).

  DownLoad: Full-Size Img PowerPoint

  Table 1

  

  Table 1.  Chemical compositions, Fe₂O₃ particle sizes and textural properties of different samples.

  DownLoad: CSV

  Fig. S3c (Supporting information) presents the XRD pattern of the FeY-9 M catalyst, which is prepared by mechanical mixing method with the Fe content of 9 wt%. The typical diffraction peaks of Fe₂O₃ (2θ = 33° and 35.6°) on this sample are obvious and much sharper than the corresponding peaks of the catalyst FeY-9 with the same Fe content (Fig. S3b in Supporting information). Both of the Fe-loading samples retain high crystallinity and good framework of FAU type in comparison with zeolite Y (Fig. S3a in Supporting information). The Fe species in FeY-9 M easily aggregate into bulk Fe₂O₃ nanoparticles in size of about 30 nm, which show very weak interactions with the support.

  UV-vis spectra of the samples with different supports and Fe loadings are shown in Fig. 2. The absorption in the UV region (< 250 nm) could be assigned to O→Fe³⁺ ligand-to-metal chargetransfer (LMCT) transitions, proving the existence of isolated iron ions in tetrahedral or octahedral coordination. The slight absorption ranging from 250 nm to 400 nm is corresponding to that of the octahedral irons in oligomeric Fe clusters. As shown in Fig. 2, each sample has a slight and broad absorption band below 400 nm, indicating the presence of a small part of iron ions in tetrahedral or octahedral coordination. The absorption region above 400 nm may be due to d-d transitions of Fe₂O₃ particles [35, 37, 38]. The slope of this band above 400 nm has been considered to reflect the size of Fe₂O₃ particles, and the sharper the slope is, the larger the Fe₂O₃ nanoparticles are [35]. The slope of the band above 400 nm presents the sequence of FeM-9 (e) ≈ FeZ-9 (d) > FeY-20 (c) > FeY-9 (b) > FeY-6 (a), so the sizes of Fe₂O₃ particles also follow this sequence, which is in good agreement with the XRD patterns and the calculation results by Scherrer's equation (Fig. 1 and Fig. S2, Table 1 and Table S1).

  Figure 2

  

  Figure 2.  Diffuse reflectance UV–vis spectra of the catalysts with different Fe contents: (a) FeY-6, (b) FeY-9, (c) FeY-20, (d) FeM-9 and (e) FeZ-9.

  DownLoad: Full-Size Img PowerPoint

  Fig. 3 and Fig. S4 (Supporting information) show the TEM images of the synthesized catalysts of FeY-9, FeM-9 and FeZ-9, respectively. Bulk Fe₂O₃ nanoparticles with the grain size of 30– 40 nm can be clearly observed in the samples of FeM-9 (Figs. S4a and b) and FeZ-9 (Figs. S4c and d), indicating that the loaded Fe₂O₃ particles are difficult to uniformly disperse on the surface of mordenite or ZSM-5 as supports, mainly owing to that Fe species have a strong aggregation potential to form large particles on the surface during calcination process. The average particle size of Fe₂O₃ on mordenite or ZSM-5 observed by TEM images is quite consistent to the calculations with Scherrer's equation by XRD patterns (Table S1). However, TEM images of FeY-9 in Figs. 3a and b show that all Fe₂O₃ nanoparticles are highly dispersed on the surface of zeolite Y with an average particle size of 5 nm, which are far smaller than those in FeM-9 or FeZ-9. As shown in Fig. S10, no obvious bulk Fe₂O₃ particles appear on the surface of zeolite Y. HRTEM images in Figs. 3c and d clearly display the lattice fringes of crystalline zeolite Y and the spherical Fe₂O₃ nanoparticles ((104) crystal face with d space of 0.27 nm) in size of 5 nm which is marked with yellow circles deposited on the surface, in accordance with the results of XRD pattern (Fig. 1, curve c) and UV–vis curve (Fig. 2, curve b), further proving highly uniform dispersion of Fe₂O₃ nanoparticles on zeolite Y. The high-angular annular dark-field scanning transmission electron microscopy (HAADF-STEM) image and corresponding energy dispersive X-ray (EDX) element mapping (Fig. 3e) of FeY-9 also confirm that the ultrafine Fe₂O₃ nanoparticles with size of 5 nm can be uniformly dispersed on the surface of zeolite Y.

  Figure 3

  

  Figure 3.  TEM (a, b) and HRTEM images of FeY-9 (c, d), the inset in panel d is the SAED patterns. (e) HAADF-STEM image and energy-dispersive X-ray element mapping of Fe, Al, O and Si elements in FeY-9.

  DownLoad: Full-Size Img PowerPoint

  XPS is one of the most efficient methods to resolve the types and valence states of the surface elements in the catalysts. Fe 2p could be split into Fe 2p_3/2 and Fe 2p_1/2 doublets because of spinorbit coupling [39, 40]. According to the different binding energies of iron species in Fe 2p_3/2, Fe³⁺ and Fe²⁺ on the surface of catalysts can be fitted readily in the XPS curves as shown in Fig. S5 in Supporting information. The fitted peaks in Fe 2p_3/2 at 709.3 eV and 710.8 eV for FeY-9 are ascribed to Fe²⁺ and Fe³⁺, respectively (Table S2 in Supporting information) [41]. According to their integral peak areas, the ratio of Fe³⁺/Fe²⁺ is equal to be about 10.1, suggesting that Fe³⁺ is the prevailing iron species in form of Fe₂O₃ on the surface of catalysts, in consistence with the weak broad diffraction peaks of Fe₂O₃ in XRD pattern (Fig. 1, curve c). The XPS curve of FeY-9 A, which is the sample of FeY-9 after Fenton-like reaction for 2 h, is given in Fig. S5b for comparison. Its binding energy of Fe 2p_3/2 doublet slightly shifts to lower positions of 708.2 eV and 709.9 eV, respectively, and the ratio of surface Fe³⁺/Fe²⁺ in this sample decreases to 7.8 (Table S2), indicating that partial Fe³⁺ species as the active centers are reduced to Fe²⁺ ones in combination with the production of hydroxyl radicals from H₂O₂ decomposition through the process of Fenton oxidation reaction [32].

  The N₂ adsorption-desorption isotherms of parent zeolite Y and Fe₂O₃-loading catalyst of FeY-9 show the typical Type Ⅰ curves (Fig. S6 in Supporting information), attributed to the filling of nitrogen in zeolite micropores at the low relative pressure, which are quite similar with that of ZSM-5 and mordenite and their Fe₂O₃-loading catalysts (not shown here). Table 1 lists the data of BET surface area (S_BET), micropore surface area (S_micro), total pore volume (V_total) and micropore volume (V_micro) of the parent HY zeolite and Fe₂O₃-loading catalysts. The S_BET, S_micro and V_micro of zeolite Y are estimated to be 465 m²/g, 366 m²/g and 0.193 cm³/g, respectively, all of which decrease slightly as the Fe content increases from 6% to 25%, suggesting that the formed ultrafine Fe₂O₃ nanoparticles gradually occupy the surface of zeolite Y, so partial micropores are blocked by the nanoparticles. Combined with the results of XRD, TEM and UV–vis, it can be concluded that preserving of most of the micropores is attributed to the uniform and high dispersion of Fe₂O₃ nanoparticles on the surface of zeolite Y, especially in the samples of FeY-6 and FeY-9.

  Heterogeneous Fenton-like oxidation of phenol was carried out at 323 K and neutral pH of the initial reaction system, which contains 20 mL phenol aqueous solution (1 g/L), the given quantity of 30% H₂O₂ and a certain amount of catalyst. Fig. S7 (Supporting information) illustrates the phenol removal at the given reaction time of 2 h by using parent zeolites and the corresponding Fecontaining catalysts (FeY-9, FeM-9 and FeZ-9). The parent zeolites of Y, mordenite and ZSM-5 can remove about 40%, 10% and 80% phenol from the solution, respectively, which are mostly dependent on the adsorption of the zeolites because of no catalytic active sites of Fe₂O₃ on them and no addition of H₂O₂ in the solution. As a kind of medium pore zeolite, ZSM-5 possesses two-dimensional 10-member-ring (10-MR) pores, whose size is quite close to that of phenol molecules, and relatively hydrophobic surface due to the low surface charge density [42a], leading to the quick and selective adsorption of phenol as high as 80% within 2 h. Whereas, both of zeolite Y and mordenite own 12-MR channels, but their surface is hydrophilic, so the affinity and adsorption rate to H₂O are higher than those to phenol, which result in the adsorption capacity of phenol in these two zeolites as low as 40% and 10%, respectively. As discussed above, since the surface area and pore volume of Fe₂O₃- loading zeolite catalysts decrease, the phenol removal of FeY-9, FeM-9 and FeZ-9 is reduced to 32%, 8% and 71%, respectively, only through the adsorption of phenol because of no H₂O₂ added in the solution. The phenol removal by FeY-9 could achieve 90% with addition of a few amount of H₂O₂, far higher than the removal only by adsorption, indicating that the heterogeneous Fenton-like oxidation reaction happens in the solution.

  Although Fe contents in the three catalysts are about 9%, the sample of FeY-9 performs the best phenol degradation efficiency of 90% under the same conditions, suggesting that the activity of the catalysts probably depends on the support structure and dispersion of iron species on the zeolite surface. As the analysis by XRD, TEM and UV–vis, the Fe₂O₃ nanoparticles in size of about 5 nm are highly dispersed only on the surface of zeolite Y, whereas, large aggregated Fe₂O₃ particles of 30–40 nm in size are formed in the catalysts of FeM-9 and FeZ-9, showing poor dispersion of Fe species on the surface of mordenite and ZSM-5. The small Fe₂O₃ nanoparticles were reported to play the key roles in H₂O₂ decomposition [43]. From the XPS analysis data of FeY-9 (Table S2), the octahedral Fe³⁺ in Fe 2p_3/2 with binding energy of 710.8 eV and Fe 2p_1/2 with binding energy of 724 eV is the dominating component in Fe₂O₃ nanoparticles [41]. In addition, the Fe₂O₃ nanoparticles in small size show lower crystallinity than those in large size, therefore, the Fe₂O₃ nanoparticles in size of 5 nm possess more unsaturated coordination octahedral Fe³⁺ and more edged Fe³⁺ [35, 36], which should be responsible for increasing the activity in decomposition of H₂O₂ to active hydroxyl radicals. However, the catalysts of FeM-9 and FeZ-9 almost have no activity in decomposing H₂O₂, suggesting that Fe₂O₃ particles in size of 30–40 nm perform poor catalytic activity in the decomposition of H₂O₂ and heterogeneous Fenton-like oxidation of phenol as well. Although FeZ-9 exhibits the phenol removal as high as 71% within 2 h, most of the phenol is removed by the Fe-loading ZSM-5 through adsorption, which could be confirmed by FT-IR spectra in Fig. S8b in Supporting information. The absorption bands at 1510 cm^-1 and 1499 cm^-1 are ascribed to the aromatic carbon-carbon stretching vibration of benzene ring, and the absorption at 700 cm^-1 is corresponding to the out-of-plane C–H deformation vibrations in aromatic ring [42b], indicating that some phenol molecules are adsorbed in FeZ-9 A. However, the absorbance bands of phenol are absent in FeY-9 A (Fig. S8a in Supporting information), which confirmed that the adsorbed phenol by the catalyst of FeY-9 was degraded through Fenton-like reaction. Due to the low adsorption capacity to phenol and low reaction activity of bulk Fe₂O₃ particles in size of 30–40 nm, the sample of FeM-9 possesses the lowest phenol removal of 8.1%. The results indicate that octahedral Fe³⁺ species in the Fe₂O₃ nanoparticles in size of 5 nm as the active centers contribute to the decomposition of H₂O₂ and degradation of phenol.

  To further confirm that the highly-dispersed Fe₂O₃ nanoparticles are the active centers, the catalyst of FeY-9 M was prepared by mechanical mixing method with Fe content of 9%, similar with the component of FeY-9. The phenol removal of FeY- 9 M at 323 K for 2 h was only 45% (Fig. 4), far lower than that of FeY- 9. The XRD pattern of FeY-9 M shows the mixed phases of zeolite Y and aggregated Fe₂O₃ particle whose size is about 30 nm (Fig. S3c), and TEM images reveal that the dispersity of Fe₂O₃ in the sample of FeY-9 M was relatively low (Fig. S9 in Supporting information), suggesting that Fe₂O₃ species are difficult to uniformly disperse on the surface of zeolite Y by mechanical mixing method, which are almost inactive in the heterogeneous Fenton-like oxidation. In addition, the catalyst of FeY-IE prepared by ion-exchange was also used to perform the Fenton-like reaction. The phenol removal of FeY-IE at 323 K for 2 h was 52% (Fig. 4), which suggested that the Fe-loading procedure applied by the wet impregnation method can lead to iron ions fixed on ion-exchange positions as well and they can contribute to catalytic activity. However, compared with the catalyst efficiency of FeY-9 which contained the iron species of both iron ions and iron oxide nanoparticles, the iron oxide nanoparticles are mainly responsible for the activity. The contribution of leached iron in the heterogeneous Fenton-like system was also considered, when the catalytic reaction is performed to 50%, then the reaction is stopped and the solid catalyst is removed by centrifugation and the catalytic reaction is then continued with the supernatant solution. The results showed that no further conversion without the solid catalyst is observed (Fig. 4), therefore, we draw a conclusion that octahedral Fe³⁺ of highly-dispersed Fe₂O₃ nanoparticles of 5 nm is the primary active species for heterogeneous Fenton-like oxidation of phenol, which can be obtained by high dispersion of iron species on the surface of zeolite Y through the process of wet impregnation and calcination.

  Figure 4

  

  Figure 4.  Phenol removal by using catalysts of parent zeolite Y, FeY-IE, FeY-9 M, FeY-9 and simulation of leached iron respectively. (The catalytic reaction was performed under these conditions: 323 K, initial pH of 7.0, catalyst amount of 0.0375 g/mL, H₂O₂ concentration of 0.14 mol/L, phenol concentration of 1.0 g/L).

  DownLoad: Full-Size Img PowerPoint

  The particle size of the zeolite Y, ZSM-5 and mordenite is about 1 μm, 0.8 μm and 1.5 μm, respectively as shown in Fig. S10, and the results showed that the particle size was not the crucial parameter for highly-dispersed iron loading. As described above, the ZSM-5 zeolite has a strong affinity to phenol due to its hydrophobic surface and matching micropore size with phenol, leading to the fast adsorption of phenol from the solution. Likewise, its hydrophobic surface is difficult to be infiltrated with the aqueous solution containing Fe³⁺ species. Therefore, in order to reduce the surface energy, the loading Fe species perform the aggregation potential to achieve the most stable state, resulting in the formation of large Fe₂O₃ particles on the surface of ZSM-5 during the calcination process. The framework of mordenite contains 12-MR channels (6.5 Å× 7.0 Å) and parallel 8-MR channels (5.7 Å× 2.6 Å). The small side pockets (3.4 Å× 4.8 Å) are formed by the connection of 12-MR channels and 8-MR channels, which greatly limit the loading capacity of fabricated species in mordenite [44]. As the amount of Fe species reach the maximum loading capacity threshold, the iron oxide could block the channels easily and thus aggregate into bulk particles on the surface of mordenite [45]. Zeolite Y possesses large surface area, uniform pore size distribution with a three-dimensional 12-MR channel system, low framework Si/Al ratio, and relatively hydrophilic surface, which is beneficial to deep infiltration by the impregnated aqueous solution containing Fe³⁺. Thereby the iron species can be highly dispersed on its surface, which are proven to act as the active centers and perform high catalytic activity in Fenton oxidation of phenol.

  To further investigate the conditions and mechanism of Fentonlike reaction using Fe-loading zeolite as the catalyst, herein, the effect of H₂O₂ concentration, temperature, and pH value was studied in detail.

  The content of hydrogen peroxide (H₂O₂) in the solution is a significant parameter for heterogeneous Fenton-like oxidation of phenol (Fig. S11 in Supporting information). At low H₂O₂ concentration (0.07 mol/L), the degradation rate is slow due to the lack of generated active hydroxyl radicals from H₂O₂ decomposition, which can immediately react with phenol by deep oxidation. Indeed, in the presence of H₂O₂ of high concentration (0.18 mol/L), as the produced hydroxyl radicals would preferentially react with the excess H₂O₂ to form the unstable intermediate products of HO₂• due to the scavenging effect [31a], which show weak oxidation activity and undesirably compete with phenol degradation catalyzed by the active hydroxyl radicals [46]. Therefore, the adequate concentration of H₂O₂ is found to be about 0.14 mol/L, and the phenol removal can get to the highest value of 90% within 2 h, which drops to 86% when the concentration of H₂O₂ increases to 0.18 mol/L.

  We also examine the effect of temperature on heterogeneous oxidation of phenol as shown in Fig. S12 in Supporting information. The conversion of phenol increases as the temperature increases from 283 K to 323 K at the given reaction time, suggesting that the catalytic activities increase with the reaction temperature. However, the removal of phenol decrease at 343 K, which is mainly due to the decomposition of partial H₂O₂ into water and oxygen at this temperature [12], leading to the actual reduction of hydroxyl radicals as well as the degradation activity to phenol in the solution. When the temperature is increased to 323 K, the phenol removal can reach nearly 85% within 1 h, and 90% of phenol removal can be achieved at 2 h, showing the good activity of FeY-9 and fast degradation of phenol.

  The effect of pH on heterogeneous oxidation of phenol is studied at different pH values from 3.0 to 10.0 under the same reaction conditions (Fig. S13 in Supporting information). The initial pH of the solution is around 7.0, in which some concentrated hydrochloric acid or aqueous ammonia is added to tune the pH values from 3.0 to 10.0. When the pH values of the reaction system equal to 6 or 7, the conversion of phenol is as high as 90%, which means that the weakly acidic or neutral conditions are beneficial for the heterogeneous Fenton-like oxidation of phenol. Cihanoglu has revealed that Brønsted acidity of the catalyst is significant in catalytic activity of phenol degradation [31b], that is to say that the zeolites could provide Brønsted acidity for the heterogeneous Fenton system instead of some adscititious liquid acid to produce hydroxyl radicals efficiently. As reported by literature, absorbance peaks at 1450, 1490 and 1545 cm^-1 correspond to Lewis, Brønsted and Lewis and Brønsted acid sites, respectively [31b]. Because zeolite Y is a kind of common solid acid catalyst with strong surface Brønsted acidity as shown in Fig. S14 in Supporting information, which is enough to supply an acidic environment for the phenol degradation, so the additional acid solution seems unnecessary in our work here, and the Fenton-like oxidation of phenol could be performed in acid-free system. When the reaction system is tuned to the weak alkalinity with pH of 8.0 and 10.0, the phenol in solution is difficult to be degraded deeply with a conversion lower than 55%, indicating that the alkaline conditions impede the activity of iron species in the Fenton-like catalytic reaction [47].

  Successive experiments are conducted in order to investigate the stability of catalyst, which are used for the reaction in next cycle after separation from the solution and washing. As seen from Fig. S15 in Supporting information, the catalytic activity gradually decreases during eight consecutive runs under the same reaction conditions, and about 78% of degradation efficiency to phenol is reserved in the 8^th catalytic reaction. The loss of catalytic activity for phenol degradation can be attributed to the slow deactivation of catalysts caused by several factors, including the poisoning of catalytic active sites, the reduction of surface area of catalyst and the slow leaching of iron species during the consecutive runs [17, 48]. Noticeably, the catalytic activity can recover to the original level when it was deeply regenerated under calcination in air at 500 ℃, showing the excellent stability and repeatability as the heterogeneous Fenton-like catalyst.

  The possible catalytic pathway of phenol degradation is shown in Fig. 5. At first, phenol could be adsorbed by the zeolite, either on the surface or in the micropores, which is proven by the blank adsorption experiments of the parent zeolite Y and the catalyst of FeY-9. The process of oxidation/redox cycle of Fe²⁺/Fe³⁺ synchronously happens on the catalyst surface, induced by the reaction of Fe species (Fe²⁺/Fe³⁺) with hydrogen peroxide to generate hydroxyl radicals. On one hand, the nanosized Fe₂O₃ particles can be highly dispersed on zeolite surface, providing enough active sites and large contact surface with hydrogen peroxide. Since the Fe₂O₃ species are deposited on zeolite surface through high-temperature sintering, a strong interaction exists between the Fe₂O₃ nanoparticles and the support. Consequently, the dispersion, geometry and electronic properties of Fe₂O₃ particles could be changed, which can affect the performance of the catalyst, because that the high activity of ultrafine Fe₂O₃ nanoparticles dispersed on zeolite surface is ascribed to the interaction between Fe species and "α- oxygen" sites [35, 36]. On the other hand, the Brønsted acidity of zeolite provides acidic environments around the zeolite surface due to the existence of protons to balance the framework electronegativity, which can greatly promote the generation rate of hydroxyl radicals. More Fe²⁺ species are produced via the reduction of Fe³⁺ species on the catalyst surface, accompanied by the production of active hydroxyl radicals. The octahedral Fe³⁺ in the ultrafine highly-dispersed Fe₂O₃ nanoparticles (5 nm) is the primary active Fe species, and the surface Brønsted acid provides the acidic environment for heterogeneous Fenton-like oxidation of phenol in acid-free system under mild conditions.

  Figure 5

  

  Figure 5.  Catalytic pathway of degradation of phenol on the Fe₂O₃-loading zeolite Y.

  DownLoad: Full-Size Img PowerPoint

  The Fe-containing Fenton-like catalysts are prepared by wet impregnation method, based on the zeolites of Y, mordenite and ZSM-5 with the same Si/Al ratio as supports. The characterizations of XRD, XPS, N₂ adsorption, TEM and UV–vis prove that ultrafine Fe₂O₃ nanoparticles in size of 5 nm can be highly dispersed only on zeolite Y matrix due to much better wettability of its hydrophilic surface. Whereas, the aggregated bulk particles in size of 30–40 nm are formed on mordenite and ZSM-5 due to the relatively hydrophobic surface and high surface energy. The Fe₂O₃-loading zeolite samples are used as the catalysts for phenol degradation in the heterogeneous Fenton-like reaction, and the effects of hydrogen peroxide concentration, reaction temperature and pH of phenol solution are screened in detail. The catalysts supported on mordenite and ZSM-5 perform low degradation efficiency of phenol owing to the low activity of large Fe₂O₃ particles in size of 30–40 nm, otherwise, the adsorption is the primary form of phenol removal. More than 90% of phenol could be deeply degraded by the catalysts of Fe₂O₃-containing zeolite Y within 2 h, which can be performed under the mild conditions of 283–323 K and a wide pH range of 4.0-7.0. The catalyst based on zeolite Y with Fe loading of about 9% exhibits the best degradation efficiency of phenol in acidfree solution. Its high catalytic activity is attributed to the bifunctional properties of strong surface Brønsted acidity and high activity to decompose H₂O₂ into hydroxyl radicals by octahedral Fe³⁺ in the ultrafine highly-dispersed Fe₂O₃ nanoparticles 5 nm in size. Since the phenol degradation is performed in neutral solution under mild conditions, the catalysts can be easily recovered and show stable catalytic activity and repeatability, which own the potential to be used as catalysts for deep degradation of industrial wastewater.

  Acknowledgments

  This work was sponsored by Shanghai Pujiang Program, China (No. 16PJ1401100), and the Shanghai Committee of Science and Technology, China (No.15ZR1402000), Key Basic Research Program of Science and Technology Commission of Shanghai Municipality (No. 17JC1400100), the NSF of China (No. 21673048), National Youth Top Talent Support Program of National High-Level Personnel of Special Support Program (Youth Top-notch Talent Support Program), the State Key Laboratory of Transducer Technology of China (No. SKT1503). The authors extend their appreciation to the International Scientific Partnership Program ISPP at King Saud University for funding this research work through ISPP# 0094.

  Appendix A. Supplementary data

  Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.cclet.2018.06.026.

• 1. [1]

     S. Esplugas, J. Giménez, S. Contreras, E. Pascual, M. Rodrı'Guez, Water Res. 36(2002) 1034-1042. doi: 10.1016/S0043-1354(01)00301-3

  2. [2]

     P.R. Gogate, A.B. Pandit, Adv. Environ. Res. 8(2004) 553-597. doi: 10.1016/S1093-0191(03)00031-5

  3. [3]

     E.R. Bandala, M.A. Peláez, D.D. Dionysiou, et al., J. Photoch. Photobiol. A 186(2007) 357-363. doi: 10.1016/j.jphotochem.2006.09.005

  4. [4]

     N.K. Daud, B.H. Hameed, J. Hazard. Mater. 176(2010) 938-944. doi: 10.1016/j.jhazmat.2009.11.130

  5. [5]

     V. Kavitha, K. Palanivelu, Water Res. 39(2005) 3062-3072. doi: 10.1016/j.watres.2005.05.011

  6. [6]

     O. Primo, A. Rueda, M.J. Rivero, I. Ortiz, Ind. Eng. Chem. Res. 47(2008) 946-952. doi: 10.1021/ie071111a

  7. [7]

     M.S. Yalfani, S. Contreras, F. Medina, J. Sueiras, Appl. Catal. B:Environ. 89(2009) 519-526. doi: 10.1016/j.apcatb.2009.01.007

  8. [8]

     E. Brillas, I. Sire's, M.A. Oturan, Chem. Rev. 109(2009) 6570-6631. doi: 10.1021/cr900136g

  9. [9]

     F. Pardo, J.M. Rosas, A. Santos, A. Romero, Environ. Sci. Pollut. R 21(2014) 12198-12207. doi: 10.1007/s11356-014-2997-2

  10. [10]

      P.R. Gogate, A.B. Pandit, Adv. Environ. Res. 8(2004) 501-551. doi: 10.1016/S1093-0191(03)00032-7

  11. [11]

      Y. Lee, L. Kovalova, C.S. Mcardell, U.V. Gunten, Water Res. 64(2014) 134-148. doi: 10.1016/j.watres.2014.06.027

  12. [12]

      S. Malato, P. Fernández-Ibáñez, M.I. Maldonado, J. Blanco, W. Gernjak, Catal. Today 147(2009) 1-59. doi: 10.1016/j.cattod.2009.06.018

  13. [13]

      J. Schneider, M. Matsuoka, M. Takeuchi, et al., Chem. Rev. 114(2014) 9919-9986. doi: 10.1021/cr5001892

  14. [14]

      Y. Kang, K.Y. Hwang, Water Res. 34(2000) 2786-2790. doi: 10.1016/S0043-1354(99)00388-7

  15. [15]

      K. Fajerwerg, H. Debellefontaine, Appl. Catal. B:Environ. 10(1996) L229-L235. doi: 10.1016/S0926-3373(96)00041-0

  16. [16]

      E.V. Kuznetsova, E.N. Savinov, L.A. Vostrikova, V.N. Parmon, Appl. Catal. B:Environ. 51(2004) 165-170. doi: 10.1016/j.apcatb.2004.03.002

  17. [17]

      K.A. Sashkina, V.S. Labko, N.A. Rudina, et al., J. Catal. 299(2013) 44-52. doi: 10.1016/j.jcat.2012.11.028

  18. [18]

      M. Wang, G. Fang, P.D. Liu, et al., Appl. Catal. B:Environ. 188(2016) 113-122. doi: 10.1016/j.apcatb.2016.01.071

  19. [19]

      J. Bujdák, Appl. Clay Sci. 34(2006) 58-73. doi: 10.1016/j.clay.2006.02.011

  20. [20]

      V. Rives, M.A. Ulibarri, Coord. Chem. Rev. 181(1999) 61-120. doi: 10.1016/S0010-8545(98)00216-1

  21. [21]

      E.M. Serwicka, K. Bahranowski, Catal. Today 90(2004) 85-92. doi: 10.1016/j.cattod.2004.04.012

  22. [22]

      A. Vaccari, Catal. Today 41(1998) 53-71. doi: 10.1016/S0920-5861(98)00038-8

  23. [23]

      Y. Han, N. Li, L. Zhao, et al., J. Phys. Chem. B 107(2003) 7551-7556. doi: 10.1021/jp026899j

  24. [24]

      F. Zhang, Y. Yan, H. Yang, et al., J. Phys. Chem. B 109(2005) 8723-8732. doi: 10.1021/jp044632+

  25. [25]

      D. Zhao, J. Feng, Q. Huo, et al., Science 279(1998) 548-552. doi: 10.1126/science.279.5350.548

  26. [26]

      M. Neamtu, C. Catrinescu, A. Kettrup, Appl. Catal. B:Environ. 51(2004) 149-157. doi: 10.1016/j.apcatb.2004.01.020

  27. [27]

      O.A. Makhotkina, E.V. Kuznetsova, S.V. Preis, Appl. Catal. B:Environ. 68(2006) 85-91. doi: 10.1016/j.apcatb.2006.07.008

  28. [28]

      G. Ovejero, J.L. Sotelo, F. Martínez, J.A. Melero, L. Gordo, Ind. Eng. Chem. Res. 40(2001) 3921-3928. doi: 10.1021/ie000896g

  29. [29]

      K.M. Valkaj, A. Katovic, S. Zrnčević, J. Hazard. Mater. 144(2007) 663-667. doi: 10.1016/j.jhazmat.2007.01.099

  30. [30]

      X. Meng, F.S. Xiao, Chem. Rev. 114(2014) 1521-1543. doi: 10.1021/cr4001513

  31. [31]

      (a) S. Navalon, M. Alvaro, H. Garcia, Appl. Catal. B: Environ. 99 (2010) 1-26;
      (b) A. Cihanoğlu, G. Gündüz, M. Dükkancı, Appl. Catal. B: Environ. 165 (2015) 687-699.

  32. [32]

      M. Noorjahan, V.D. Kumari, M. Subrahmanyam, L. Panda, Appl. Catal. B:Environ. 57(2005) 291-298. doi: 10.1016/j.apcatb.2004.11.006

  33. [33]

      R. Gonzalez-Olmos, F.D. Kopinke, K. Mackenzie, A. Georgi, Environ. Sci. Technol. 47(2013) 2353-2360. doi: 10.1021/es303885y

  34. [34]

      K.A. Sashkina, E.V. Parkhomchuk, N.A. Rudina, V.N. Parmon, Microporous Mesoporous Mater. 189(2014) 181-188. doi: 10.1016/j.micromeso.2013.11.033

  35. [35]

      G. Fierro, G. Moretti, G. Ferraris, G.B. Andreozzi, Appl. Catal. B:Environ. 102(2011) 215-223. doi: 10.1016/j.apcatb.2010.12.001

  36. [36]

      G. Moretti, G. Fierro, G. Ferraris, G.B. Andreozzi, V. Naticchioni, J. Catal. 318(2014) 1-13. doi: 10.1016/j.jcat.2014.07.005

  37. [37]

      K. Jíša, J. Nováková, M. Schwarze, et al., J. Catal. 262(2009) 27-34. doi: 10.1016/j.jcat.2008.11.025

  38. [38]

      J. Wang, G. Mai, G. Liu, S. Yu, J. Microbiol. Biotechn. 23(2013) 251-259. doi: 10.4014/jmb

  39. [39]

      P. Xie, L. Chen, Z. Ma, et al., Microporous Mesoporous Mater. 200(2014) 52-60. doi: 10.1016/j.micromeso.2014.08.020

  40. [40]

      Y. Ying, S. Jiang, H. Zhang, X. Zhang, Chem. Eng. J. 259(2015) 243-251. doi: 10.1016/j.cej.2014.08.018

  41. [41]

      P. Tan, J. Catal. 338(2016) 21-29. doi: 10.1016/j.jcat.2016.01.027

  42. [42]

      (a) M. Takeuchi, T. Kimura, M. Hidaka, D. Rakhmawaty, M. Anpo, J. Catal. 246 (2007) 235-240;
      (b) R. Lapuente, F. Cases, P. Garce's, E. Morallo'n, J.L. Va'zquez, J. Electroanal. Chem. 451 (1998) 163-171.

  43. [43]

      X. Li, X. Liu, L. Xu, et al., Appl. Catal. B:Environ. 165(2015) 79-86. doi: 10.1016/j.apcatb.2014.09.071

  44. [44]

      H.V. Le, S. Parishan, A. Sagaltchik, et al., ACS Catal. 7(2017) 1403-1412. doi: 10.1021/acscatal.6b02372

  45. [45]

      C. Wang, B. Zhao, Y. Xie, Chin. J. Catal. 24(2003) 475-482. http://en.cnki.com.cn/Article_en/CJFDTOTAL-CHUA200306016.htm

  46. [46]

      M.B. Kasiri, H. Aleboyeh, A. Aleboyeh, Appl. Catal. B:Environ. 84(2008) 9-15. doi: 10.1016/j.apcatb.2008.02.024

  47. [47]

      D. Wu, Y. Chen, Z. Zhang, et al., Chem. Eng. J. 294(2016) 49-57. doi: 10.1016/j.cej.2016.02.097

  48. [48]

      S. Zhang, X. Zhao, H. Niu, Y. Shi, Y. Cai, J. Hazard. Mater. 167(2009) 560-566. doi: 10.1016/j.jhazmat.2009.01.024

• 

  Figure 1  XRD patterns of the catalysts with different Fe contents based on zeolite Y as the support: zeolite Y (a), FeY-6 (b), FeY-9 (c), FeY-15 (d), FeY-20 (e) and FeY-25 (f).

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Diffuse reflectance UV–vis spectra of the catalysts with different Fe contents: (a) FeY-6, (b) FeY-9, (c) FeY-20, (d) FeM-9 and (e) FeZ-9.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  TEM (a, b) and HRTEM images of FeY-9 (c, d), the inset in panel d is the SAED patterns. (e) HAADF-STEM image and energy-dispersive X-ray element mapping of Fe, Al, O and Si elements in FeY-9.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Phenol removal by using catalysts of parent zeolite Y, FeY-IE, FeY-9 M, FeY-9 and simulation of leached iron respectively. (The catalytic reaction was performed under these conditions: 323 K, initial pH of 7.0, catalyst amount of 0.0375 g/mL, H₂O₂ concentration of 0.14 mol/L, phenol concentration of 1.0 g/L).

  下载: 全尺寸图片 幻灯片
  

  Figure 5  Catalytic pathway of degradation of phenol on the Fe₂O₃-loading zeolite Y.

  下载: 全尺寸图片 幻灯片

  Table 1.  Chemical compositions, Fe₂O₃ particle sizes and textural properties of different samples.

  下载: 导出CSV
•