The preparation of Fe2O3-ZSM-5 catalysts by metal-organic chemical vapour deposition method for catalytic wet peroxide oxidation of m-cresol

Fe2O3-ZSM-5 catalysts (0.6 wt% Fe load) prepared by metal-organic chemical vapour deposition (MOCVD) method were evaluated in the catalytic wet peroxide oxidation (CWPO) of m-cresol in a batch reactor. The catalysts have a good iron dispersion and small iron crystalline size, and exhibit high stability during reaction. In addition, the kinetics of the reaction were studied and the initial oxidation rate equation was given. Catalysts were first characterized by N2 adsorption–desorption isotherms, scanning electronic microscopy, energy-dispersive spectroscopy, X-ray diffraction and X-ray photoelectron spectroscopy. Results show that extra-framework Fe3+ species (presenting in the form of Fe2O3) are successfully loaded on ZSM-5 supports by MOCVD method. Performances of catalysts were tested and effects of different temperature, stirring rate, catalyst amount on hydrogen peroxide, m-cresol, total organic carbon (TOC) conversion and Fe leaching concentration were studied. Results reveal that catalytic activity increased with higher temperature, faster stirring rate and larger catalyst amount. In all circumstances, m-cresol conversion could reach 99% in 0.5–2.5 h, and the highest TOC removal (80.5%) is obtained after 3 h under conditions of 60°C, 400 r.p.m. and catalyst amount of 2.5 g l−1. The iron-leaching concentrations are less than 1.1 mg l−1 under all conditions. The initial oxidation rate equation −rA0=7.2×1011e−81300/RTCA is obtained for m-cresol degradation with Fe2O3-ZSM-5 catalysts.

The preparation of Fe 2 O 3 -ZSM-5 catalysts (0.6 wt% Fe load) prepared by metalorganic chemical vapour deposition (MOCVD) method were evaluated in the catalytic wet peroxide oxidation (CWPO) of m-cresol in a batch reactor. The catalysts have a good iron dispersion and small iron crystalline size, and exhibit high stability during reaction. In addition, the kinetics of the reaction were studied and the initial oxidation rate equation was given. Catalysts were first characterized by N 2 adsorption-desorption isotherms, scanning electronic microscopy, energy-dispersive spectroscopy, X-ray diffraction and X-ray photoelectron spectroscopy. Results show that extraframework Fe 3+ species (presenting in the form of Fe 2 O 3 ) are successfully loaded on ZSM-5 supports by MOCVD method. Performances of catalysts were tested and effects of different temperature, stirring rate, catalyst amount on hydrogen peroxide, m-cresol, total organic carbon (TOC) conversion and Fe leaching concentration were studied. Results reveal that catalytic activity increased with higher temperature, faster stirring rate and larger catalyst amount. In all circumstances, m-cresol conversion could reach 99% in 0. 5-2.5 h, and the highest TOC removal (80.5%) is obtained after 3 h under conditions of 60°C, 400 r.p.m. and catalyst amount of 2.5 g l −1 . The iron-leaching concentrations are less than 1.1 mg l −1 under all conditions. The initial oxidation rate equation −r A0 = 7.2 × 10 11 e −81300/RT C A is obtained for m-cresol degradation with on the thin film morphology. Chu et al. [40] also prepared a copper containing activated carbon catalyst by MOCVD, which performed a good catalytic activity in the catalytic wet air oxidation of phenol.
Liu et al. [45] studied the degradation of m-cresol in a fixed bed reactor with α-Fe 2 O 3 /γ-Al 2 O 3 catalysts prepared by incipient wetness impregnation method, but the kinetics of the reaction had not been intensively studied. Little literature reports the study of CWPO process of m-cresol. Also, catalysts used for m-cresol degradation prepared by MOCVD have not been reported yet. In this work, ZSM-5 was chosen as the support because of its high hydrothermal stability, fine lipophilicity, strong chemical resistance and good catalytic activity [46]. The aims of this work are to (i) prepare novel Fe 2 O 3 -ZSM-5 catalysts which have a good iron dispersion and small iron crystalline size by MOCVD method, (ii) test the catalytic activity of the prepared catalysts in decomposition of m-cresol by CWPO process (the effects of operating parameters including reaction temperature, stirring rate and amount of catalysts on the CWPO procedure are also discussed) and (iii) study the kinetics of the catalytic reaction and give the initial oxidation rate equation of the m-cresol CWPO reaction.

Materials
Commercial ZSM-5 zeolites (Si/Al = 80) were supplied by Nankai University Catalyst Factory. Ferric acetylacetonate was purchased from Aladdin Industrial Corporation. Hydrogen peroxide (H 2 O 2 , 30 wt% aqueous) was purchased from Jiangsu Qiangsheng Chemical Co., Ltd. m-cresol and manganese dioxide were purchased from Shanghai Qiangshun Chemical Reagent Factory. Sodium thiosulfate and potassium iodide were purchased from Tianjin Bodi chemical Co., Ltd. Deionized water was selfmade and used in all experimental processes. All of the chemical reagents used in this study were of analytical grade.

Catalyst preparation
The ZSM-5 zeolite columns were first smashed into powders and were divided into five groups (40,60,80, 100, 120 mesh) by the sieves. Fe 2 O 3 -ZSM-5 catalysts were prepared by MOCVD, and the schematic diagram is shown in figure 1. ZSM-5 and ferric acetylacetonate (Fe(acac) 3 ) powders were chosen as support and iron source precursor [38], respectively. ZSM-5 supports were stored inside a crucible and well mixed with Fe(acac) 3 precursor. The crucible was covered before it was put into a quartz tube. Nitrogen was used as protective gas during the process. The deposition temperature was 310°C with a heating rate of 5°C min −1 [47] and the deposition time was 1.5 h. After deposition, the catalysts were calcined in air at 550°C for 6 h with a heating rate of 1°C min −1 . The iron load of the catalysts was 0.6 wt%.

Catalyst characterization
Different techniques were used to characterize the ZSM-5 and Fe 2 O 3 -ZSM-5 samples. N 2 adsorptiondesorption was performed on an ASAP2020 to get the isotherms, BET-specific surface area and pore volume. Scanning electronic microscopy (SEM) was performed on a Zeiss Merlin FE-SEM to obtain the textural and morphological information of the samples. Energy-dispersive spectroscopy (EDS) and element mapping were applied to analyse the iron dispersion on Fe 2 O 3 -ZSM-5 samples. X-ray diffraction (XRD) was performed on an X'Pert PRO instrument using Cu Kα radiation (40 kV, 40 mA) with 2θ range from 5°to 90°to measure the crystallographic structures of the samples. X-ray photoelectron spectroscopy (XPS) spectra were tested on a Kratos Axis Ultra (DLD) instrument using an Al Kα radiation source operated at 15 kV and 10 mA to obtain the binding energy of O 1s and Fe 2p .

Catalytic wet peroxide oxidation of m-cresol
CWPO of m-cresol was performed in a glass batch reactor at atmospheric pressure for 3 h to test the catalytic activity of Fe 2 O 3 -ZSM-5 catalysts. During the reaction, the reactor was not sealed, and a condenser was used to prevent the loss of the liquid. The size of the catalysts was 178 µm (80 mesh). The batch reactor was heated by a thermal bath to set the desired temperature. A stirring rake was applied to mix the reactants well and eliminate the influence of external diffusion. The reaction solution contained 1 g l −1 m-cresol and 6 g l −1 H 2 O 2 . The amount of oxidant was adjusted to the stoichiometric rsos.royalsocietypublishing.org R. Soc. open   amount needed for the total m-cresol oxidation according to the following reaction: The required amount of H 2 O 2 was first mixed with m-cresol and pre-heated to a desired temperature, and the catalysts were then added into the mixture. The reaction solution was sampled every 30 min to monitor the main parameters including conversions of m-cresol, H 2 O 2 , total organic carbon (TOC) and concentration of iron leaching. Reaction temperature, stirring rate and amount of catalysts were main variables in experiments and were changed systematically to study their effects on the reaction. H 2 O 2 concentrations were obtained by iodometric titration using 0.02 mol l −1 sodium thiosulfate solution and the expression of H 2 O 2 conversion rate (X H 2 O 2 ) was as follows: where C H 2 O 2 (0) and C H 2 O 2 (t) (mg l −1 ) are concentrations of H 2 O 2 at the initial time t = 0 and sampling time t (hours), respectively. After H 2 O 2 concentrations were determined, 0.1 g manganese dioxide was added into sample solutions to eliminate the residual H 2 O 2 [30] which may affect the measurement of other parameters. m-cresol concentrations were measured by applying high performance liquid chromatography (HPLC) (Agilent 1100) equipped with an Agilent HC-C18(2) column and a UV detector which set a wavelength of 272 nm. Methanol and ultrapure water (v/v = 80:20) were employed as the mobile phase and the flow rate was 1.0 ml min −1 . The conversion rate of m-cresol (X m-cresol ) was calculated as follows: where C m-cresol(0) and C m-cresol(t) (mg l −1 ) are concentrations of m-cresol at the initial time t = 0 and sampling time t (hours), respectively. TOC concentrations of samples were measured by a Liqui TOC instrument. The conversion rate of TOC (X TOC ) was as follows: where C TOC(0) and C TOC(t) (mg l −1 ) are concentrations of m-cresol at the initial time t = 0 and sampling time t (hours), respectively. Iron-leaching concentration in sampling solution was tested by atomic absorption spectrophotometer (AA-6800) to analyse the stability of the catalysts. The specific surface areas (S BET ) are calculated from adsorption branches in the relative pressure range of 0.06-0.3 using the BET (Brunauer-Emmett-Teller) method. The total pore volume (V total ) is calculated by analysing N 2 adsorption-desorption isotherms. The microporous surface area (S micro ) and microporous volume (V micro ) are calculated using t-plot method. It can be seen that N 2 adsorption-desorption isotherm shapes of ZSM-5 support and Fe 2 O 3 -ZSM-5 catalysts are similar to those of type II standard. In the beginning, the volume adsorbed increases continually with the increasing relative pressure, which should be attributed to monolayer-multilayer adsorption [48]. At a high relative pressure, both samples show a hysteresis loop and make the isotherm similar to type IV [49], implying the existence of mesopores which have narrowed size distribution on the samples [50]. The mesoporous pore size distributions are acquired by BJH (Barrett-Joyner-Halenda) method. As can be seen, a narrow peak can be observed between the pore width of 2 and 3 nm, indicating that mesopores between these sizes present a wide distribution in ZSM-5 support and Fe 2 O 3 -ZSM-5 catalyst [45]. The result matches well with the N 2 adsorption-desorption isotherms.

Results and discussion
As listed in table 1, after loading iron, the BET-specific surface area of ZSM-5 decreases from 276.2 to 251.7 m 2 g −1 and the total pore volume also falls from 0.343 to 0.313 cm 3 g −1 . The results can be ascribed to the iron oxides which disperse in the pore canal and block the pores [51]. The microporous specific surface area and microporous volume also decrease by 14.8 m 2 g −1 and 0.007 cm 3 g −1 , respectively, implying that the iron species are successfully loaded into the micropore and result in the loss of the microporous specific surface area and microporous volume.

Scanning electronic microscopy and energy-dispersive spectroscopy analysis
Scanning electronic microscopy was used to study the morphology and structure of Fe 2 O 3 -ZSM-5 catalysts, and the image is shown in figure 3. As can be seen, iron species are observed and they are uniformly dispersed on the surface of the catalysts. The figures indicate that the iron species of our prepared Fe 2 O 3 -ZSM-5 catalysts are extra-framework Fe 3+ species (Fe 2 O 3 ), which are also located in the pores and on the surface of the zeolites. The results are essentially in agreement with EDS data exhibited in figures 4 and 5. In figure 4, the existence of Fe, Si, Al and O elements in catalysts is confirmed by the EDS elemental analysis spectrum. In figure 5, the distribution condition of different elements on the surface of catalysts is clearly depicted in the EDS elemental mapping images. It can be seen that Fe species are well dispersed on catalyst surface with a uniform dispersion, indicating that the active components are successfully loaded on the catalysts through MOCVD method.

X-ray diffraction
The XRD patterns of ZSM-5 support and Fe 2 O 3 -ZSM-5 catalyst are shown in figure 6. As can be seen, both samples give the diffraction peaks between 2θ = 7-9°and 2θ = 23-25°, matching well with the standard pattern of ZSM-5 according to the literature [52]. The patterns indicate that the crystal phase of ZSM-5 does not change after loading iron species. However, the intensity of characteristic ZSM-5 peak in Fe 2 O 3 -ZSM-5 catalyst decreases. According to the literature [48,53], the iron species loaded on ZSM-5 could enhance the zeolites' absorption of X-ray, causing the decrease in the intensity of characteristic ZSM-5 peaks. On the other hand, the characteristic peaks of Fe cannot be observed for the Fe 2 O 3 -ZSM-5 samples. Li et al. [54] have reported that the intensity of peak could be significantly affected by the size of the species and its dispersion on the support, while many other works [55][56][57] have proved that weak peaks in XRD patterns could be caused by good dispersion and small crystalline size of the component. According to Yoo [58], iron particles which were loaded on the support by CVD method would be small and would disperse well on the support. Zou et al. [56] also proved that no peak on the XRD pattern could be put down to fine dispersion of metal compounds on ZSM-5 support. Low content of the iron species might be another reason affecting the intensity of the peak. The existence of Fe 2 O 3 on support can be proved by the following XPS analysis.

X-ray photoelectron spectroscopy
The results of XPS are displayed in figure 7. All spectra are calibrated by C 1s peak at 284.6 eV. As can be seen in figure 7a, the peaks of Fe 2 O 3 -ZSM-5 before calcination at 711.0 eV [59] and 726 eV [60] represent the binding energy of Fe 2p3/2 and Fe 2p1/2 in Fe(III), respectively. The satellite peak of Fe 2p3/2 can be observed at 716.7 eV, which is the characteristic peak of Fe(III) in Fe 2 O 3 according to many previous works [60][61][62]. In addition, the peaks centring at 722.6 eV and 709 eV represent the binding energy of Fe 2p3/2 and Fe 2p1/2 in Fe(II), respectively, and the satellite peak of Fe 2p3/2 locates at 714. 6

Effect of reaction temperature
The CWPO reaction of m-cresol solution was carried out in a batch reactor at different temperatures (40°C, 60°C and 80°C). Other conditions were set as follows: catalysts 2.5 g l −1 , stirring rate of 400 r.p.m. The performances of catalysts are shown in figure 8. In figure 8a, H     increasing temperature will promote the decomposition of H 2 O 2 into hydroxyl radical [64] and accelerate the decomposition of H 2 O 2 to H 2 O and O 2 . Figure 8b reveals that the conversion rate of m-cresol also increases as the reaction temperature increases. When the reaction temperature is 40°C, it takes 2.5 h for m-cresol conversion to reach 99%. When the temperature is increased to 60°C, it takes 1 h for 99% conversion. When the temperature is further increased to 80°C, it takes only 0.5 h. It has been reported that the degradation rate of organic pollutants in CWPO reaction can be enhanced by rising temperature [65], and this could be due to the production of more hydroxyl radicals which play an important part in the oxidation of m-cresol. We have known that the iron species of our Fe 2 O 3 -ZSM-5 catalysts are extra-framework Fe 3+ species which present in the form of Fe 2 O 3 . According to Yan et al.'s [66] research, extra-framework Fe 3+ species perform higher activity in organic degradation than the framework Fe 3+ , and this can be the reason of high m-cresol conversion in this work. The result is consistent with H 2 O 2 conversion, implying that Fe 2 O 3 -ZSM-5 catalysts are active for m-cresol oxidation. A similar trend can also be observed in TOC conversion from figure 8c. It can be seen that elimination of TOC will be promoted as the temperature increases. For reaction temperatures at 40, 60 and 80°C, TOC conversion reaches 27.9%, 62.6% and 80.5%, respectively, after 3 h reaction. The residual TOC could still be detected while the m-cresol has been totally removed, indicating that m-cresol is degraded into some intermediate products and some hard degradable low weight molecular organic acids which produced the remainder TOC [67]. As for the stability of catalysts, iron-leaching concentration is shown in figure 8d. It can be observed that though the iron-leaching concentration gradually grows as the reaction temperature increases, the iron loss is low. When the temperature was 80°C, after 3 h, the iron concentration was 0.933 mg l −1 , and only 6.2% of iron leached out from Fe 2 O 3 -ZSM-5. The loss of iron ions can be ascribed to the formation of the acid by-product during the reaction [68]. In addition, stirring for a long time could also wash out some of the active components. Even so, catalysts prepared by MOCVD still could significantly reduce the iron leaching, because the diffusion of the metal substance in vapour phase can help the metal components go further into the pore structure and finally be fixed in the pores, making the catalysts more stable [40].

Effect of stirring rate
Different stirring rate can affect the mechanical shearing force of catalysts, which will influence the loss of active component and the destruction of catalyst morphology. Stirring rate can also impact the mass transfer and external diffusion process, which play an important part in reaction rate, product conversion and selectivity. In this experiment, the stirring rate is adjusted to 100, 200 and 400 r.p.m. under same conditions (namely, catalysts 2.5 g l −1 , temperature at 60°C) to study its impact on the reaction. The results are shown in figure 9. As figure 9a reveals, when the stirring rate increases from 100 to 400 r.p.m., H 2 O 2 conversion is raised from 84% to 93%, implying that higher stirring rate will accelerate the decomposition of H 2 O 2 . As for m-cresol conversion depicted in figure 9b, when the stirring rate is 100 r.p.m., the conversion of m-cresol reaches only 83.0% after 0.5 h, and it takes 1.5 h for the conversion to reach 99%. When the stirring rate is raised to 200 and 400 r.p.m., only 1 h is needed to reach 99% conversion. The contact and mass transfer among reagents in the system are enhanced with the increasing stirring rate, thus promoting m-cresol degradation. By comparing figure 9a and b, the trend of m-cresol conversion is not synchronous with H 2 O 2 conversion when the stirring rate increases, so it can be inferred that speeding up the stirring rate will not only promote the degradation of m-cresol but also boost the decomposition of H 2 O 2 into oxygen and water. Figure 9c indicates that the variation trend of TOC conversion (52.7%, 59.0%, 62.6% at 100, 200, 400 r.p.m., respectively, after 3 h) is consistent with m-cresol conversion. When the stirring rate increases, TOC conversion simultaneous rises. At high stirring rate, external diffusion has been eliminated, so the degradation process of m-cresol is controlled by surface reaction on the catalysts and the reaction will not be significantly expedited as the stirring rate is further added. From figure 9d, we can see that though iron-leaching concentration ascends slightly with increasing stirring rate, the maximum concentration is only 0.164 mg l −1 , and 1.1% of the iron is

Effect of amount of catalysts
To investigate the effect of catalyst amount on m-cresol degradation, different catalyst dosages (1, 2.5 and 4 g l −1 ) are added into the reactor under the following conditions: temperature at 60°C, stirring rate of 400 r.p.m. The results are shown in figure 10. As we can see in figure 10a, after 3 h reaction, conversion of H 2 O 2 grows significantly from 67.9% to 99% as the catalyst amount increases from 1 to 4 g l −1 . This phenomenon can be explained by Lewis's collision theory. The occurrence of a reaction needs to meet the requirement: two activated molecules collide in a particular direction (namely, effective collision). When more catalysts are added into the system, the number of active sites increases, so the effective collision probability between activated H 2 O 2 molecules and active sites increases, and H 2 O 2 decomposition is promoted. At the same time, formation of hydroxyl radicals is boosted and benefits the degradation of m-cresol. The result is corresponding to figure 10b. When the catalyst amount is 1 g l −1 , conversion of m-cresol reaches 81.5% after 0.5 h. When the amount is adjusted to 2.5 g l −1 , the figure obviously raises to 93.5%. When the dosage is 4 g l −1 , 99% removal of m-cresol is observed. All three conditions get 99% removal after 1 h. As for TOC conversion portrayed in figure 10c, TOC has obvious higher conversion at condition of 4 g l −1 catalysts than at 2.5 and 1 g l −1 . In the very beginning, conversion rates of 2.5 and 1 g l −1 are nearly the same. After 1.5 h, reaction at 2.5 g l −1 starts to exhibit higher TOC conversion than 1 g l −1 . The results confirm that the increase of the catalyst dosages will take effect in m-cresol degradation. Iron-leaching concentration during reaction is depicted in figure 10d, and the highest value is 1.092 mg l −1 . According to this figure, only 7.3 wt% of active component leaches out from the catalysts, and it is lower than catalysts prepared by incipient wetness impregnation [69]. According to Liu et al. [45], the leached iron ions have little effect on the TOC removal when the concentration is low. The result indicated that the catalysts prepared by MOCVD method exhibited excellent ability to degrade m-cresol with fine stability.

Kinetics of catalytic wet oxidation peroxide of m-cresol with Fe 2 O 3 -ZSM-5
To investigate the kinetics of CWPO of m-cresol, effects of stirring rate and particle size of catalysts on the initial rate of m-cresol conversion were studied first in order to eliminate the external diffusion and internal diffusion, respectively. The results are shown in figures 11 and 12. As can be seen, when the stirring rate is increased to more than 400 r.p.m., the initial rate of m-cresol degradation no longer rises. Similarly, when the particle size of catalysts gets to 80 meshes, the initial rate stops rising. The results mean that the external and internal diffusion have been eliminated at the stirring rate of 400 r.p.m. and particle size of 80 meshes.   If the CWPO of m-cresol over Fe 2 O 3 -ZSM-5 with excess H 2 O 2 is a first-order reaction, it should follow the following reaction rate equation [70]: where C A is m-cresol concentration (mg l −1 ) at time t (min), k 0 (min −1 ) is the pre-exponential factor, E a (J mol −1 ) is the activation energy and T (Kelvin) is the reaction temperature. The equation could be changed into the following form: ln 2) The first-order rate coefficient can be obtained by plotting ln(C A0 /C A )versus time and the slope is the value of coefficient. The plots acquired at different temperatures from 313 to 343 K are shown in figure 13 (when the m-cresol conversion reached 100%, ln C A0 /C A makes no sense, so these points are not depicted in the figure), and the values of the first-order rate coefficients are depicted in table 2. As can be seen, the experimental data are consistent with the rule of the first-order reaction, so it reveals that the catalytic reaction follows first-order kinetics.
According to Arrhenius equation:  It can be changed into another form: Arrhenius plot of ln k versus 1/T is depicted in figure 14. Based on the slope of the plot and the above equation, the activation energy is 81.3 kJ mol −1 and k 0 is 7.2 × 10 11 min −1 . Hence, the initial oxidation rate could be expressed as follows: − r A0 = 7.2 × 10 11 e −81300/RT C A .
(3.5) Data accessibility. This article does not contain any additional data.