Since the development of mesoporous SiO2 in the 1990s [1, 2], the synthesis and catalytic application of mesoporous SiO2-based materials have attracted much attention in the catalysis community [3, 4, 5, 6, 7]. Mesoporous SiO2 materials have ordered nanopores, large surface area, and high thermal stability, which make them ideal for fabricating supported or other functionalized catalysts for a wide range of useful reactions.
In recent years, mesoporous metal oxides such as TiO2, Co3O4, and Al2O3 have attracted great interest because they have unique functionalities associated with their redox and/or acid properties [8, 9, 10]. Nevertheless, owing to the lower thermal stability and higher density of these metal oxides compared with those of SiO2, the synthesis of high-surface-area mesoporous metal oxides constitutes a significant challenge.
Al2O3 is one of the most useful industrial catalysts and supports [11]. Al2O3-supported metal or metal oxide catalysts have found wide applications in the processing of fossil and biomass fuels, the conversion of chemical stocks into useful chemical products, and the cleaning up of air. A number of advances have been made in the synthesis of mesoporous Al2O3 [12, 13, 14, 15, 16, 17, 18, 19]. In particular, a simple solvent evaporation-induced self-assembly (EISA) strategy has been developed for the synthesis of ordered mesoporous Al2O3 [18]. A similar method was later used for the synthesis of mesoporous MgO-Al2O3 [19, 20], CaO-Al2O3 [19, 21], TiO2-Al2O3 [19, 22, 23], CrOx-Al2O3 [19, 24], NiO-Al2O3 [25], ZrO2-Al2O3 [26, 27], LaOx-Al2O3 [28], and CeO2-Al2O3 [29].
The development of mesoporous MOx-Al2O3 has provided new opportunities for catalysis. First, mesoporous MOx-Al2O3 materials themselves can be used as catalysts. For example, mesoporous CrOx-Al2O3 has shown high catalytic activity and stability for the liquid phase oxidation of cyclohexane [24], and mesoporous Cr2O3-Al2O3 has shown good performance in the dehydrogenation of ethane and propane [30].
Alternatively, mesoporous MOx-Al2O3 materials can be used as precursors for making active catalysts after proper pretreatment. For example, mesoporous CoOx-Al2O3 was reduced in H2 to produce Co/mesoporous Al2O3 active for partial oxidation of methane [31]. Mesoporous CuOx-Al2O3 was transformed into Cu/mesoporous Al2O3 active for direct synthesis of dimethyl ether [32] and ethanol synthesis via dimethyl oxalate hydrogenation [33]. Mesoporous NiO-Al2O3 [34, 35, 36], NiO-MgO- Al2O3 [37, 38], NiO-CaO-Al2O3 [39], NiO-LaOx-Al2O3 [40], and NiO-CeO2-Al2O3 [40, 41, 42] were reduced to form metallic Ni nanoparticles supported on their Al2O3-based mesoporous matrix, and their catalytic performance for the CO2 reforming of methane was tested. Mesoporous NiO-Al2O3 was reduced to form supported Ni catalysts for selective CO methanation [43], partial oxidation of methane to syngas [44], and steam reforming of CH4 [45]. Ni/mesoporous ZrO2-Al2O3 [46] and Ni/mesoporous MgO-Al2O3 [47] prepared by one-pot synthesis of mesoporous mixed oxides followed by reduction in H2 were used for steam reforming of ethanol [46] and methane [47], respectively. The advantage of this preparation method (one-pot synthesis followed by reduction) is that the metallic nanoparticles can be highly dispersed on the internal surfaces of the mesoporous oxide matrix, and their catalytic activity may be further tuned by the presence of another metal oxide (other than Al2O3).
Although mesoporous Al2O3 has been occasionally used to support metals or metal oxides [48, 49, 50, 51, 52], the employment of mesoporous MOx-Al2O3 materials as supports for the preparation of supported metal or metal oxide catalysts has been rarely reported [53, 54, 55, 56, 57, 58]. For example, Chou and co-workers loaded NiO onto mesoporous MgO-Al2O3 [56] and CaO-Al2O3 [57], and reduced the catalysts in H2. The obtained supported Ni catalysts showed good performance for the CO2 reforming of CH4. Because MOx incorporated into the mesoporous framework may influence the catalytic performance of supported metal or metal oxide catalysts, it would be desirable to synthesize this kind of multi-component catalyst to obtain tunable catalytic performance.
Herein, a series of mesoporous MOx-Al2O3 (M = Mn, Fe, Co, Ni, Cu, Ba) materials were prepared by a one-pot EISA method. The selected MOx additives were chosen because they are either active components or promoters of heterogeneous catalysts useful for many reactions. Rh2O3 nanoparticles were then loaded onto the mesoporous mixed-oxide supports via wet impregnation (Scheme 1). The resulting catalysts were characterized by N2 adsorption-desorption measurements, X-ray diffraction (XRD), transmission electron microscopy (TEM), inductively coupled plasma optical emission spectrometry (ICP-OES), and X-ray photoelectron spectroscopy (XPS), and their catalytic performance for CO oxidation and N2O decomposition was studied.
The triblock copolymer Pluronic P123 (MW = 5800) was purchased from Sigma-Aldrich. Aluminum isopropoxide, 50% Mn(NO3)2 solution, Fe(NO3)3·9H2O, Co(NO3)2·6H2O, Ni(NO3)2· 6H2O, Cu(NO3)2·3H2O, Ba(NO3)2, ethanol, and HNO3 (67 wt%) were purchased from Sinopharm Chemical Reagent.
Mesoporous MOx-Al2O3 materials were prepared according to a previous EISA method [14]. In a typical synthesis, 4.0 g of P123 was dissolved in 80 mL of ethanol at room temperature. Next, 6.4 mL of 67 wt% nitric acid, 36 mmol of aluminum iso-propoxide, and 4 mmol of metal nitrate were added. The mixed solution was stirred for 5 h and then dried at 60 °C for 48 h. The final gel was heated in air in a muffle oven to 700 °C at a ramp rate of 2 °C/min, and calcined at 700 °C for 4 h. The obtained sample was denoted as M-MA, where M represents Mn, Fe, Co, Ni, Cu, or Ba, and MA is an abbreviation for “mesoporous Al2O3”. The actual M content of each sample was determined by ICP-MS. Mesoporous Al2O3 (denoted as MA) was prepared in a similar way, except that no metal nitrate was added.
Rh/MA and Rh/M-MA catalysts were prepared by wet impregnation. A 2 mL Rh(NO3)3 solution (0.01 g/mL based on Rh) was placed in an agate mortar containing 1.98 g M-MA. The mixture was ground until dry under an infrared lamp. The obtained powders were heated in a muffle oven to 500 °C at a ramp rate of 10 °C/min, and calcined at 500 °C for 3 h. The Rh in Rh/MA and Rh/M-MA represents the decomposition product (i.e., Rh2O3) of Rh(NO3)3. The actual Rh content of the seven samples was measured by ICP-MS.
Wide-angle powder XRD measurements were performed on an MSAL XD2 instrument with Cu Kα radiation from 2θ = 10.0° to 80.0°. The scanning rate was 4°/min, and the step length was 0.01°. Small-angle X-ray scattering (SAXS) data were recorded on a NanoStar U SAXS System from 2θ = 0.2° to 2.8°.
N2 adsorption-desorption isotherms were obtained using a Micromeritics Tristar 3000 instrument. Before the measurement, the sample (0.1−0.2 g) was degassed in vacuum at 300 °C for 3 h. The specific surface areas of the samples were calculated from the N2 adsorption-desorption data using the BET method.
ICP-OES analysis was performed on a PerkinElmer OPTIMA 2100 DV optical emission spectrometer. A 50-mg sample was dissolved in a mixture of 9 mL HCl, 3 mL HNO3, 1 mL HClO4, 0.5 mL H2O2, and 3 mL HF reagents, and heated at 150 °C for 2-3 h. After that, 3 mL HCl, 1 mL HNO3, 1 mL HF, and 0.5 mL HClO4 were added again, and the mixture was transferred to an autoclave, heated at 180 °C for 4 h, allowed to cool, and then diluted with distilled water for analysis.
TEM images were obtained using a JEM-2011F transmission electron microscope operated at 200 kV. The catalyst samples were dispersed in ethanol by sonication, and a few drops were dropped onto a carbon-coated copper grid followed by evaporation of the solvent. The size of the Rh2O3 nanoparticles was obtained by measuring 100 particles for each sample using the DigitalMicrograph software.
XPS spectra were recorded on a PerkinElmer PHI 5000 C spectrometer with a Mg Kα X-ray source. The binding energy was calibrated using the C 1s peak at 284.4 eV as a reference, and the deconvolution and fitting of peaks were conducted using the Peak-fit software.
Cu LMN Auger spectra were recorded using an RBD updated PHI-5000C ESCA instrument (PerkinElmer) operated at 14.0 kV with Mg Kα radiation and a detection angle of 54°.
The catalytic activity of the Rh/M-MA and Rh/MA catalysts for CO oxidation and N2O decomposition was measured in a fixed reactor. The catalyst (0.25 g) was loaded into a U-shaped glass tube for the reaction.
For CO oxidation, the reaction gas was composed of 1% CO in air. The flow rate was 50 mL/min. The catalyst was maintained at room temperature for 1 h, and heated to 200 °C at a ramp rate of 0.5 °C/min. The effluent gas was analyzed periodically (every 10 min) using an on-line gas chromatograph (GC; Agilent 7890A, equipped with a TCD detector) capable of separating CO, CO2, N2, and O2, using He as a carrier gas. The CO conversion was calculated as ([CO]in - [CO]out)/[CO]in x100%, where [CO]in is the CO concentration at room temperature (at which no reaction occurs), and [CO]out is the CO concentration at elevated temperature. For comparison, mesoporous MOx- Al2O3 and Al2O3 supports were tested under similar conditions, except that they were heated to 300 °C.
For N2O decomposition, the reaction gas contained 0.5% N2O balanced with He, and the flow rate was 60 mL/min. The catalyst was maintained at room temperature for 1 h, and the reaction temperature was then increased in steps and kept at each temperature step for 0.5 h. The effluent gas was analyzed periodically (every 10 min) using an on-line GC (Agilent 7890A) capable of separating N2O, N2, and O2. The N2O conversion was calculated as ([N2O]in - [N2O]out)/[ N2O]in x 100%.
Fig. 1 shows the wide-angle XRD patterns of the various catalysts. The XRD pattern of Rh/MA displayed three main reflections at 2θ = 37.5°, 45.8°, and 66.9°, which corresponded to the γ-Al2O3 phase (PDF#10-0425). The Rh/M-MA (M = Mn, Fe, Ni, Cu) samples did not exhibit XRD peaks corresponding to crystalline MOx. In contrast, Rh/Co-MA showed additional peaks assignable to CoAl2O4, and Rh/Ba-MA showed some peaks corresponding to BaCO3, probably caused by the reaction of BaO with CO2 generated by the combustion of residual organic species (isopropanol and P123). For all these catalysts, no peaks from Rh2O3 could be detected, indicating that the Rh2O3 particles were highly dispersed on the supports.
Fig. 2 shows the small-angle X-scattering (SAXS) patterns of the samples. Rh/MA exhibited a very intense (100) peak around 2θ = 0.8° and a weak (110) peak around 2θ = 1.0°, indicating its hexagonal mesoporous structure [34]. The observed signals from this sample were very strong, so the raw data were divided by 1000 to enable the SAXS patterns of all the samples to be depicted in one figure. In the cases of the Rh/M-MA (M = Mn, Fe, Ni, Cu) samples, the (100) peak was broader and the (110) peak disappeared completely, which indicated a partial loss of the ordered mesoporous structure. In the cases of Rh/Co-MA and Rh/Ba-MA, the (100) peak was weaker.
Table 1 shows the actual contents of Rh and the metal elements in the samples, as determined by ICP-OES. Because Rh2O3 is extremely difficult to dissolve in aqua regia for ICP-OES analysis, the samples were treated under harsh conditions, as described in the Experimental section. The actual Rh contents of Rh/MA and Rh/M-MA (M = Mn, Fe, Co, Ni, Cu, Ba) were determined to be 0.99 wt%, 0.87 wt%, 0.86 wt%, 0.90wt%, 0.97 wt%, 0.92 wt%, and 0.90 wt%, respectively, equivalent to or slightly lower than the theoretical value of 1.0 wt.%. The actual contents of metal elements (Mn, Fe, Co, Ni, Cu, Ba) in the samples were 8.0 wt%, 8.5 wt%, 9.2 wt%, 9.5 wt%, 11.4 wt%, and 16.2 wt%, respectively, slightly lower than the theoretical values of 10.1 wt%, 10.4 wt%, 11.0 wt%, 11.0 wt%, 11.7 wt%, and 19.8 wt%, respectively. Note that the M/Al molar ratio of the starting materials used to synthesize of MOx-Al2O3 was 1/9.
Fig. 3 shows the N2 adsorption-desorption isotherms and pore size distributions of the catalysts. The corresponding textural properties (BET surface area, pore volume, and pore size) are listed in Table 1. The isotherms presented typical type IV curves with H1 shaped hysteresis, characteristic of mesoporous materials. The pore size distributions of Rh/M-MA catalysts were in the range of 2.5-9.5 nm.
The BET surface area and pore volume of the Rh/M-MA (M = Fe, Co, Cu, Ba) catalysts were slightly lower than those of Rh/MA. In contrast, the specific surface areas of Rh/Mn-MA and Rh/Ni-MA were significantly increased. It has been reported that the nickel precursors used in the preparation of ordered mesoporous Ni-V-Al can act as swelling agents and increase the micelle size of triblock copolymers [59]. The manganese precursor may have the same effect in this work.
TEM was used to observe the morphology of the catalysts. Rh/MA exhibited uniform and hexagonal ordered mesopores, as shown in Fig. 4(a). Rh/Ni-MA showed relatively ordered mesopores (Fig. 5(a)), and Rh/Mn-MA (Fig. 4(b)), Rh/Fe-MA (Fig. 4(c)), Rh/Co-MA (Fig. 4(d)), and Rh/Cu-MA (Fig. 5(b)) showed worm-like nanopores. In contrast, Rh/Ba-MA did not have many ordered mesoporous and some BaCO3 aggregates were observed in Fig. 5(c). Fig. 5(d) shows that the lattice spacing of the Rh/Ba-MA sample was 0.379 nm, consistent with the lattice spacing of (111) planes of BaCO3.
Fig. 6 and Fig. 7 show the HRTEM images of the catalysts, highlighting the sizes and distribution of Rh2O3 nanoparticles on the supports. It can be seen that the Rh2O3 nanoparticles were located inside the pores. More importantly, the dispersion of the nanoparticles was very good and no large particles were found on any of the catalysts. The average sizes of the Rh2O3 nanoparticles on the MA and M-MA (M = Mn, Fe, Co, Ni, Cu, Ba) supports were 1.2±0.3, 1.0±0.3, 1.0±0.2, 0.9±0.3, 0.8±0.3, 1.0±0.2, and 1.2±0.3 nm, respectively.
The Rh 3d and metal oxide XPS spectra of the samples are shown in Figs. 8 and 9, respectively. Two evident main peaks at approximately 309-309.3 eV and 313.8-314.1 eV were observed, which can be assigned to Rh3+ [60, 61]. These results confirm that the Rh species in the supports were Rh2O3.
In Fig. 9, the Mn 2p and Fe 2p XPS peaks are attributed to Mn4+ (MnO2) [62, 63] and Fe3+ (Fe2O3) [64]. Rh/Co-MA exhibited two features assignable to Co2+, a main peak at 781.3 eV and a Co2+ satellite located at 785.8 eV [65, 66]. The Co2+ species in Rh/Co-MA may have been CoAl2O4. This was not only confirmed by the XPS and XRD data, but also confirmed by the blue color of the sample because CoAl2O4 is blue [67]. Rh/Ni-MA exhibited a Ni 2p3/2 main peak at 855.8 eV with a satellite peak at around 862.0 eV, characteristic of NiAl2O4 [39, 59]. Note that the binding energy of pure NiO should be 853.6 eV [36, 37]. Therefore, the Ni species in Rh/Ni-MA was NiAl2O3 [39, 59]. Rh/Cu-MA showed two Cu 2p3/2 peaks at 934.1 and 932.3 eV, with a satellite near 938.5-946.6 eV. The peak at 932.3 eV can be assigned to Cu+ [68], and the peak at 934.1 eV corresponds to Cu2+ [68] in CuO. The ratio of Cu+/(Cu++Cu2+) was 11.45%. Because the Cu 2p3/2 signals of Cu+ and Cu0 can hardly be differentiated, a Cu LMM Auger spectrum was recorded. The broad peak was deconvoluted into a main peak at 917.2 eV (Cu2+) and a minor peak at 913.4 eV (Cu+) [69]. Finally, the peaks observed in the Ba 3d XPS spectrum of Rh/Ba-MA were attributed to Ba2+ (BaO/BaCO3) [70], which means that Ba2+ existed in the form of BaO and/or BaCO3 in the sample.
The elimination of CO is useful in pollution control, H2 fuel cells, and many other applications. Fig. 10 presents the CO oxidation reactivity over the Rh/MA and Rh/M-MA catalysts. Rh/Mn-MA showed the best activity among the samples, and 100% CO conversion was obtained at 145 °C. The catalytic activity (evaluated from T50 values) of the catalysts followed the sequence Rh/Mn-MA (122 °C) > Rh/Fe-MA (130 °C) ≈ Rh/Cu-MA (131 °C) > Rh/Co-MA (136 °C) > Rh/Ni-MA (156 °C) > Rh/MA (161 °C) > Rh/Ba-MA (171 °C). It should be mentioned that the supported Rh catalysts were tested in CO oxidation up to 200 °C, and all of them reached 100% CO conversion below 200 °C. In our control experiments, the MOx-Al2O3 supports were tested in CO oxidation up to 300 °C. As shown in Fig. 11, the activity followed the sequence Cu-MA > Mn-MA > Fe-MA > Co-MA > Ba-MA > MA > Ni-MA. Most of the supports were not very active below 200 °C.
These results mean that although the M-MA supports may show some activity for CO oxidation at or above 200 °C, when Rh2O3 nanoparticles are loaded onto these supports, the main active species for CO oxidation below 200 °C is Rh2O3 but not M-MA. Thus, the role of the MOx (M = Mn, Fe, Co, Ni, Cu, not including Ba) additive is to promote the Rh2O3 species. The reason for this promotional effect is not clear at this stage, but structural effects can be excluded because the Rh/M-MA samples were less ordered than Rh/MA. The promotional effect was also not caused by a difference in the size of the Rh2O3 nanoparticles on different supports, because the average particle sizes were similar. Further research is still needed to understand the observed promotional effect of the MOx additives.
N2O has been identified as a greenhouse gas, and therefore its elimination is very important for the environment. Fig. 12 shows the results of N2O conversion over the Rh/MA and Rh/M-MA catalysts as a function of reaction temperature. The catalytic activity (evaluated from their T50 values) of the Rh/M-MA catalysts followed the sequence Rh/Co-MA (283 °C) > Rh/Ni-MA (287 °C) ≈ Rh/Fe-MA (290 °C) ≈ Rh/Ba-MA (292 °C) > Rh/MA (301 °C) > Rh/Cu-MA (314 °C) > Rh/Mn-MA (321 °C). In our control experiments, the corresponding M-MA supports showed no activity for N2O decomposition (data not shown), thus highlighting the importance of the supported Rh2O3 nanoparticles in this reaction.
The Rh/M-MA (M = Fe, Co, Ni, Ba) catalysts were more active for the N2O decomposition, than Rh/MA, while Rh/Cu-MA and Rh/Mn-MA were less active. However, Rh/Cu-MA and Rh/Mn-MA were more active than Rh/MA for CO oxidation, i.e., the activity trend observed in the N2O decomposition experiments was roughly the reverse of the trend seen in CO oxidation. This may be because N2O decomposition requires the desorption of O2 to complete the catalytic cycle, whereas CO oxidation requires the adsorption of O2 [71].
Fig. 13 and Fig. 14 show the stability of selected catalysts as a function of reaction on stream. Different reaction temperatures were chosen to make sure that the conversions of CO or N2O on the catalysts were high but not 100%. As shown in Fig. 13, Rh/MA, Rh/Mn-MA, and Rh/Cu-MA attained significant CO conversions at 165, 135, and 145 °C, respectively. The initial CO conversions were similar to those observe from the conversion curves in Fig. 10. However, the catalysts became deactivated over time, especially during the initial 10 h on stream. The catalytic activity of each of the samples seemed to be stabilized during the later stage of the reaction, meaning that in practical applications, complete CO conversion can still be achieved if more catalyst is put into the reactor or the reaction temperature is higher. As shown in Fig. 14, Rh/MA, Rh/Co-MA, and Rh/Ni-MA showed significant N2O conversions at 300, 275, and 300 °C, respectively. The initial N2O conversions were similar to those observed in the conversion curves in Fig. 12. The N2O conversions then increased over time and became stable during the later stage of the reaction.
To put the current work in perspective, the activity of the most active Rh/Co-MA obtained in this study was compared with those of other Rh catalysts for CO oxidation and N2O decomposition reported in the literature. Because the catalyst amount, Rh content, CO (N2O) concentration, and flow rate used differ between studies, specific rates over the different catalysts at the same temperature were calculated. As shown in Table 2, Rh/Mn-MA showed a specific rate of 597 mmol gRh-1 h-1 in CO oxidation at 140 °C, higher than those of Rh/MA (97 mmol gRh-1 h-1 in this study and Rh/Al2O3 (0 mmol gRh-1 h-1) reported in the literature [72]. Rh/Mn-MA was also more active than our previous Rh/AlPO4 (228 mmol gRh-1 h-1) and Rh/ZnPO4 (177 mmol gRh-1 h-1) catalysts. However, it is difficult to compare Rh/Mn-MA with Rh/CeO2 (> 107 mmol gRh-1 h-1) [73] or Rh/hydroxyapatite (> 179 mmol gRh-1 h-1) because the latter two catalysts showed 100% CO conversion below 140 °C.
As shown in Table 3, Rh/Co-MA exhibited a specific rate of 346 mmol gRh-1 h-1 in N2O decomposition at 325 °C, making it apparently more active than Rh/MA (279 mmol gRh-1 h-1), Rh/Al2O3 (192 mmol gRh-1 h-1) [74], Rh/Fe-SBA-15 (0 mmol gRh-1 h-1) [75], Rh/Cu-SBA-15 (8 mmol gRh-1 h-1) [75], and Rh/KIT-6 (0 mmol gRh-1 h-1) [76], but less active than Rh/CeO2 (> 1071 mmol gRh-1 h-1) [77]. Rh/Co-MA was more active than Rh/LaPO4 (25 mmol gRh-1 h-1), but it is difficult to compare Rh/Co-MA with Rh/hydroxyapatite (> 277 mmol gRh-1 h-1) because the latter catalyst showed 100% N2O conversion below 325 °C [78].
A series of mesoporous MOx-Al2O3 (M = Mn, Fe, Co, Ni, Cu, Ba) materials were prepared via a one-pot synthetic method. The obtained materials were used as supports for the loading of Rh2O3. The oxides of Mn, Fe, Ni, and Cu were highly dispersed in the mesoporous matrix, whereas those of Co and Ni reacted with MA to form CoAl2O4 and NiAl2O4 spinels, and some BaCO3 formed on MA for Rh/Ba-MA. All the obtained catalysts had high specific surface area, large pore volume, and nanopores. Rh2O3 particles of around 1 nm in size were well dispersed on the supports. All of these factors caused the mesoporous catalysts to exhibit excellent catalytic activity. For CO oxidation, the activity of the catalysts followed the sequence Rh/Mn-MA > Rh/Fe-MA ≈ Rh/Cu-MA > Rh/Co-MA > Rh/Ni-MA > Rh/MA > Rh/Ba-MA. For N2O decomposition, the activity followed the sequence Rh/Co-MA > Rh/Ni-MA ≈ Rh/Fe-MA ≈ Rh/Ba-MA > Rh/MA > Rh/Cu-MA > Rh/Mn-MA. Although the reasons for the differences in activity between the different catalysts are not yet explicitly clear, we have successfully demonstrated the application of mesoporous MOx-Al2O3 as a support for Rh2O3 catalysts. Although here we chose CO oxidation and N2O decomposition as probe reactions to compare the catalytic activity of the samples, we believe that these catalysts may be useful for other reactions, considering the broad usefulness of Rh2O3 and Al2O3 in catalysis.
2016, Vol. 37 Issue (1): 73-82
PDF (1519 KB)
2016, Vol. 37 Issue (1): 73-82
PDF (1519 KB)