Euaggelia Skliria,
Ioannis N. Lykakisb and
Gerasimos S. Armatas*a
aDepartment of Materials Science and Technology, University of Crete, Voutes, 71003, Heraklion, Greece. E-mail: garmatas@materials.uoc.gr
bDepartment of Chemistry, Aristotle University of Thessaloniki, University Campus, 54124, Thessaloniki, Greece
First published on 15th January 2014
Ordered mesoporous composite catalysts consisting of nanocrystalline tetragonal ZrO2 and heteropolytungstic clusters, i.e. 12-phosphotungstic (PTA) and 12-silicotungstic (STA) acids, were prepared via a surfactant-assisted co-polymerization route. According to the X-ray diffraction, transmission electron microscopy and N2 physisorption measurements, the resultant materials possess a well-defined mesoscopic order ranging from wormhole to hexagonal pore structure and exhibit large internal surface area (126–229 m2 g−1) and quite narrow pore size distribution (ca. 2.2–2.6 nm in diameter). Energy dispersive X-ray microanalysis and infrared spectroscopy confirms that the heteropoly clusters are well dispersed within the zirconia matrix, while preserving intact their Keggin structure. The inclusion of PTA and STA clusters in the mesoporous framework has a beneficial effect on the catalytic activity of these materials. Although zirconium oxide and heteropoly acids alone show little catalytic activity, the ZrO2–PTA and ZrO2–STA heterostructures exhibit surprisingly high activity in hydrogen peroxide mediated oxidation of 1,1-diphenyl-2-methylpropene under mild conditions. Indeed, the mesoporous ZrO2–STA composite sample loaded with 5 wt% STA shows a conversion rate that is 17 times higher than the mesoporous ZrO2. The catalytic activity of these materials is related to the spatial distribution of heteropoly acids in zirconia matrix and possible synergistic interactions between the incorporated Keggin units and Zr(IV) oxohydroxide species.
Over the past few years, extensive research efforts have been made to immobilize POMs on a high-surface-area solid support as an efficient strategy for achieving high catalytic activity and reusability.5 Compared to bulk heteropoly acids, supported POM clusters have various advantages, which are their large surface area, good accessibility to the reactants, and easier separation from the reaction mixture.6 In addition, the metal oxide surfaces could have a beneficial effect in catalytic processes through complementary or synergistic interactions with the POM compounds. For example, recently we indicated that incorporation of [PW12O40]3− Keggin clusters into the cobalt oxide matrix can dramatically improve the catalytic activity of mesoporous Co3O4 towards N2O decomposition.7 Therefore, the synthesis of high-surface-area mesoporous materials that combine the catalytic activity of POMs with the redox activity and durability of metal oxides is a major direction in catalysis research.
Impregnation is one of the most attractive synthetic routes towards mesoporous POM-based materials. This approach involves infiltration of soluble POM compounds within the pore channels of a previously synthesized host material, such as silica,8 alumina,9 titania and zirconia.10 Although this is a simple yet efficient method, the impregnation process is often plagued by poor dispersion of POM units and pore blocking effects of the inorganic support. In addition, the impregnated POM clusters show moderately low chemical stability when anchored on surfaces with strong basic character (e.g. Al2O3, MgO),11 while they are prone to leaching in polar solvents when anchored on weak acid or neutral supports (e.g. SiO2).12 Nevertheless, both the unprecedented molecular transport through nanopores and high hydrolytic stability of mesophase are valuable prerequisites for heterogeneous catalysis. A promising synthetic strategy to this end is the sol–gel inclusion of POMs in the pore walls of a mesoporous metal oxide material. This approach entails cross-linking polymerization of metal oxide and POM compounds in the presence of surfactants and allows controlled synthesis of mesoporous heterostructures with modulated POM/metal-oxide composition. Unlike to the impregnated analogues, the POM compounds in these materials constitute an integral part of the inorganic framework, forming rather a solid-solution structure.13 Because of the excellent catalytic properties of POM clusters and the large pore surface of the host structure, mesoporous solids of this type are anticipated to have interesting perspectives in the applications as catalysts, exhibiting high activity and reusability.14 While silica-based POM composites are easy to realize,15 the synthesis of mesoporous heterostructures from transition-metal oxide and POM compounds still remains a challenge. The assembly of such complex materials requires an adjustment of the interaction between heteropoly acids and organic templates. Otherwise disordered mesophases of organic–inorganic aggregates due to the strong binding of heteropoly anions to the surfactant micelles might be formed.16 Therefore, specific reaction conditions according to the choice of metal sources, template molecules, and POM anions are essential in order to achieve ordered POM-based mesoporous frameworks.
Zirconia (ZrO2) is an important n-type semiconductor material with great potential applications in redox and acid catalysis.17 Recently, we demonstrated that well-ordered mesoporous frameworks consisting of nanocrystalline ZrO2 and Keggin-type H3PMo12O40 acids can be prepared via a soft-templating chemistry.18 This work used non-ionic surfactant Pluronic P123 to direct the formation of hexagonal mesoporous ZrO2-based structures. These materials showed remarkable catalytic activity in H2O2-mediated oxidation of selected alkenes, which is comparable to that of homogeneous 12-phosphomolybdic acids. Herein, we report on the synthesis of a new type of mesoporous composite materials consisting of zirconium oxide and 12-phosphotungstic (H3PW12O40, PTA) or 12-silicotungstic (H4SiW12O40, STA) acid components. To produce these ZrO2–POM composites with different loading of POM clusters (i.e. ranging from 2 to 20 wt%), we utilized evaporation-induced cooperative self-assembly of Zr(IV) oxide dichloride and heteropolytungstic acid compounds in the presence of the non-ionic surfactant polyoxoethylene cetyl ether (Brij-58) as structure-directing agent. The products possess a ZrO2–PTA and ZrO2–STA solid-solution structure with large and accessible surface area and narrow pore size distribution. More important, we also show that, although mesoporous zirconia and POM compounds alone have little catalytic activity, the ZrO2–POM composite frameworks exhibit surprisingly high activity in oxidation of 1,1-diphenyl-2-methylpropene under mild conditions. Indeed, the ZrO2–STA sample loaded with 5 wt% STA is about 17 times more active than pure mesoporous ZrO2, while 12-silicotungstic acid along shows negligible activity under similar conditions.
Sample | BET Surface (m2 g−1) | Pore volumea (cm3 g−1) | Pore size (nm) | Unit cell (nm) | WTb (nm) |
---|---|---|---|---|---|
a Cumulative pore volume at relative pressure p/p0 = 0.95. In parenthesis: the micropore volume obtained by the DR plots.b The framework wall thickness given by WT = a0 − D, where D is the pores size. | |||||
meso-ZrO2 | 118 | 0.06 (0.04) | 2.5 | 4.3 | 1.8 |
ZrPTA(2) | 126 | 0.07 (0.05) | 2.4 | 4.5 | 2.1 |
ZrPTA(5) | 160 | 0.07 (0.04) | 2.3 | 4.6 | 2.3 |
ZrPTA(10) | 207 | 0.09 (0.05) | 2.3 | 5.1 | 2.8 |
ZrPTA(20) | 210 | 0.09 (0.05) | 2.3 | 5.6 | 3.3 |
ZrSTA(2) | 140 | 0.08 (0.05) | 2.5 | 4.7 | 2.2 |
ZrSTA(5) | 155 | 0.08 (0.05) | 2.6 | 5.1 | 2.5 |
ZrSTA(10) | 229 | 0.11 (0.06) | 2.6 | 5.8 | 3.2 |
ZrSTA(20) | 218 | 0.12 (0.07) | 2.6 | 6.2 | 3.6 |
The wide-angle XRD patterns of the mesoporous ZrPTA(w) and ZrSTA(w) suggested the presence of nanocrystalline ZrO2, as indicated from the broad reflections in the 2θ = 30–60° range (Fig. 2). The XRD peaks can be assigned to the (011), (110), (012), (112) and (121) Bragg reflections of the tetragonal phase of ZrO2 (JCPDS card no. 50-1089). As for the domain size of ZrO2 crystallites, it is difficult to estimate from the peak-width analysis of X-ray diffractions because of the low intensity of the peaks. However, their broadness obviously indicates crystallite size in the nanoscale range; typically below 30 nm.22 We did not observed any peak in the XRD patterns that could be assigned to the crystalline phase of H3PW12O40 and H4SiW12O40 compounds. This observation supports the notion that both PTA and STA heteropoly acids are homogeneously dispersed in mesoporous structures. However, given that the lower limit of detection of powder XRD is approximately 3–5 wt%, the presence of tiny amounts of nanometer-sized POM aggregates within the samples cannot be preclude.
![]() | ||
Fig. 2 Wide-angle XRD patterns of mesoporous (a) ZrPTA(w) and (b) ZrSTA(w) composite materials. The XRD pattern of mesoporous zirconia (meso-ZrO2) is also given in panel (a). |
Direct examination of mesoporous ZrSTA(5) and ZrSTA(10) samples with TEM reveals extended mesoscopic pore ordering with uniform pores. As seen in Fig. 3, the ZrSTA(5) shows a poorly ordered, wormhole-like, pore structure with regular pores, while the ZrSTA(10) shows discernible domains of hexagonally arranged cylindrical pore channels, suggesting p6mm pore structure; although small deformations of the hexagonal honeycomb structure can be observed. Detailed analysis of TEM images indicated that both samples have an average pore-to-pore distance of ∼5 nm and pore diameter of ∼2 nm. This gives estimation for the framework wall thickness of about 3 nm, which is in good agreement with the results of the XRD and N2 porosimetry (see below). Furthermore, apart from XRD, the structure of crystalline phase was investigated by selected-area electron diffraction (SAED). SAED characterization suggests the poorly crystalline nature of zirconia, showing a broad and diffusive Debye–Scherrer diffraction ring (inset of Fig. 3a and c). According to the wide-angle XRD data, the electron diffraction ring could be tentatively assigned to the overlapped pair of (101) and (002) reflections of ZrO2 with tetragonal structure. Energy dispersive X-ray spectroscopy (EDS) acquired on TEM confirms the chemical composition of the as-prepared materials. As shown in Fig 3b and d (insets), the EDS spectra taken across a thin area of mesoporous structures indicate an average atomic ratio of Zr:
W close to 97.3
:
2.7 and 94.8
:
5.2 for ZrSTA(5) and ZrSTA(10), respectively. In excellent agreement with the stoichiometry of reactions, these atomic ratios correspond to a weight percentage of STA content at ∼5.3 wt% for ZrSTA(5) and ∼10.1 wt% for ZrSTA(10). These results give evidence of the molecular dispersion and complete incorporation of heteropoly clusters into the mesoporous ZrO2 structure.
The Fourier transform-infrared (FT-IR) spectra of mesoporous ZrPTA(w) and ZrSTA(w) materials show a series of absorption peaks in the 700–1100 cm−1 region that are similar to the spectra of pure heteropoly acids, i.e. H3PW12O40 and H4SiW12O40. The IR spectra of high PTA-loaded ZrPTA(w) (w = 5, 10 and 20 wt%) samples show the characteristic WOd and W–Ob–W (i.e., between the corner-sharing WO6 octahedral) absorption bands of 12-phosphotungstic clusters at ∼949–952 and ∼865–872 cm−1, respectively, consistent with the presence of these clusters in composite materials (Fig. 4a).23 These vibration bands are rather obscure on mesoporous ZrPTA(2) due to the low content (2 wt%) of PTA clusters. In comparison with the bulk PTA, the absorption peaks of W
Od and W–Ob–W vibrations are remarkable red-shifted (by over 30 cm−1), probably due to the strong conjugated interactions between the terminal (W
Od) and bridged (W–Ob–W) oxygen atoms of (PW12O40)3− anions and zirconia matrix. The absorptions at ∼1083 and ∼1026 cm−1 can be assigned as P–Oa stretching vibrational bands of the distorted PO4 central units in [PW12O40]3− and lacunary [PW11O39]8− Keggin clusters, respectively.24 Given the relative strong basic character of zirconia surface, the formation of lacunary Keggin structures in these samples is a distinct possibility. Correspondently, the IR analysis of ZrSTA(w) showed absorption peaks at ∼1024, ∼952 and ∼870 cm−1, which are reasonably assigned to the Si–Oa, W
Od and W–Ob–W stretching vibrations of the [SiW12O40]4− Keggin cluster, respectively (Fig. 4b).25 The remarkable red shift of the W
Od and W–Ob–W absorption bands (ca. 15–30 cm−1) compared with those for pure STA, suggests that these clusters may also engages in strong interactions with the Zr–OH species through the W
Od terminal and W–Ob–W bridging bonds. The strong absorptions at ∼756 cm−1 and ∼623 cm−1 are attributed to the stretching vibrations of Zr–O and Zr–O–Zr of tetragonal zirconia, respectively.26
The nanoporosity of the ZrPTA(w) and ZrSTA(w) heterostructures was determined by nitrogen physisorption measurements. As shown in Fig. 5, all samples display a combination of type I and type IV adsorption isotherms associated with a distinct condensation step at relative pressure (p/po) ∼0.15–0.2, characteristic of mesoporous solids with uniform pores.27 Analysis of the adsorption data with the Brunauer–Emmett–Teller (BET) method indicated large surface areas in the range of 126–229 m2 g−1 and total pore volumes in range of 0.07–0.12 cm3 g−1. It can be observed that the surface area of these materials is strongly related to the POM loading and it increases as the POM fraction in the pore walls increases, see Table 1. This is quite peculiar considering the heavy elements of heteropoly acids (W) that composed the frameworks. It is worth noting that mesoporous ZrO2 exhibits a slightly lower BET surface area (118 m2 g−1) and pore volume (0.06 cm3 g−1), see ESI, Fig. S1.† This signifies the role of heteropoly acids in self-assembly processes, in which the [PW12O40]3− and [SiW12O40]4− anions are possibly engaged in specific van der Waals and/or hydrogen bonding interactions with the surfactant micelles. Therefore, a ZrO2–POM/surfactant composite phase can be stabilized, leading in self-assembly of these hybrid complexes into three-dimensional ZrO2–POM mesostructures with long-range order. This assumption is also consistent with the structural deformation in low POM-loaded ZrPTA(w) and ZrPTA(w) (w = 2 and 5 wt%) samples recommended by XRD and TEM results. To acquire more information of the porous structure of composite materials, the Dubinin–Radushkevich (DR) micropore analysis was applied in the low relative pressure region (p/po < 0.15). By means of this method, the micropore volume is estimated to be in range of 0.05–0.07 cm3 g−1 that is coincided with a structural microporosity of about 50 to 60%. These data, along with the type-I shape of N2 adsorptions, readily confirm to the presence of interconnected pore network constructed from both mesopore and micropore channels. Table 1 summarizes all the textural parameters of mesoporous zirconia (meso-ZrO2) and ZrPTA(w) and ZrSTA(w) composite materials.
![]() | ||
Fig. 5 N2 adsorption–desorption isotherms at 77 K for mesoporous ZrPTA(w) and ZrSTA(w) materials. Insets: the corresponding NLDFT pore size distributions. |
The pore size distribution calculated from the adsorption data based on the non-local density function theory (NLDFT) is quite narrow, indicating a uniform pore structure with an average pore diameter of ∼2.2–2.6 nm. Since the mesoporous materials with different POM content show similar pore size, the continuous increase in unit cell size (ao) of ZrPTA(w) and ZrSTA(w) samples implies an relative increase of framework wall thickness. From a combination of NLDFT pore sizes and XRD lattice constants, we obtained a pore wall thickness from 2.1 to 3.3 nm for ZrPTA(w) and from 2.2 to 3.6 nm for ZrSTA(w). Based on similar calculations, the framework wall thickness of mesoporous zirconia (meso-ZrO2) was estimated to be ca. 1.8 nm. The increase of the wall thickness with subsequent increase of the POM loading is consistent with the integration of these heteropoly acids into the framework of composite materials. This is in agreement with the XRD and TEM-EDS data.
Next, efforts were directed to investigate the catalytic performance of composite catalysts featuring different POM loading. The oxidations of 1 using hydrogen peroxide as oxidant was monitored by CG-MS analysis and the results are shown in Fig. 6 and S3 of the ESI.† It is evident that the catalytic activity for both ZrPTA(w) and ZrSTA(w) composites is strongly related to the amount of the incorporated POM clusters. Namely, the 2–5 wt% POM-loaded samples showed a moderate-to-high conversion of 1, while composite materials containing large amount of heteropoly acids (10–20 wt%) exhibited a dramatic drop in conversion yield. It is noteworthy that employing pure mesoporous zirconia and heteropoly acids as catalysts gave poor activity under similar reaction conditions (∼25% and less than 2% in 30 min, respectively). Based on the 1 consumption, the ZrSTA(5) catalyst achieves the highest activity, giving >99% conversion of oxidative products in only 30 min. Meanwhile, control experiment showed that in the absence of catalyst no reaction takes places. The present data suggest that, apart to the surface area, the elemental PTA/ZrO2 and STA/ZrO2 composition has a profound influence on the catalytic oxidation processes. It seems that ZrSTA(w) and ZrPTA(w) are really synergistic catalysts, wherein Keggin anions and Zr(IV) oxohydroxide (ZrOx(OH)y) species potentially act as cooperative catalytic sites for the activation of CC bonds of 1 substrate. This is a different from the case of pure ZrO2 and heteropolytungstic acids.
Table 2 summarizes the results of the catalytic oxidation of 1 on heterogeneous ZrPTA(w) and ZrSTA(w) as well as homogeneous PTA and STA catalysts. In all cases the corresponding epoxide 1b and ketone 1c were formed as major oxidative products, accompanying with small amount of the ketone 1a and diol 1d. Benzophenone 1a is a cleavage product of the corresponding 1,2-dioxitane alkene, which is a non stable intermediate.19 While, a notable amount of ketone 1c should be formed through 1,2-phenyl shift29 of the corresponding diol 1d, as verified by GC-MS and H1 NMR measurements,30 as well as by control catalytic experiments.31 As seen in Table 2, all the ZrPTA(w) and ZrSTA(w) catalysts promote the oxidation of 1 alkene with moderate-to-good chemoselectivity (36–58%) for 1c.
Catalyst | Conversionb (%) | Relative yieldc (1a![]() ![]() ![]() ![]() ![]() ![]() |
Kinetic constant (k) (h−1) |
---|---|---|---|
a Reaction conditions: 0.1 mmol of 1, 0.5 mmol of catalyst, 0.3 mmol of H2O2 (30 wt% in water), 3 mL ethanol 50 °C, 30 min.b Based on the consumption of alkene 1 measured by GC.c Products relative yield determined by GC-MS using nonane as internal standard. | |||
meso-ZrO2 | 25 | 10![]() ![]() ![]() ![]() ![]() ![]() |
0.010 |
ZrPTA(2) | 78 | 6![]() ![]() ![]() ![]() ![]() ![]() |
0.048 |
ZrPTA(5) | 87 | 6![]() ![]() ![]() ![]() ![]() ![]() |
0.064 |
ZrPTA(10) | 38 | 7![]() ![]() ![]() ![]() ![]() ![]() |
0.012 |
ZrPTA(20) | 5 | 12![]() ![]() ![]() ![]() ![]() ![]() |
0.001 |
PTA | <1 | ||
ZrSTA(2) | 65 | 6![]() ![]() ![]() ![]() ![]() ![]() |
0.041 |
ZrSTA(5) | >99 | 8![]() ![]() ![]() ![]() ![]() ![]() |
0.168 |
ZrSTA(10) | 40 | 6![]() ![]() ![]() ![]() ![]() ![]() |
0.017 |
ZrSTA(20) | 8 | 12![]() ![]() ![]() ![]() ![]() ![]() |
0.003 |
STA | <2 |
The time evolution of 1 conversion can be reasonably accounted by a pseudo-first order reaction model. This is reasonable if we account the excess of H2O2, so that its concentration could be considered constant during the reaction. The Fig. 7 shows the plots of ln(Ct/Co) (where, Co is the concentrations of 1 at the initial state of reaction and Ct is the concentration at the time t) as a function of time, by which the reaction rate constants (k) as a slope of the linear fits were obtained. The corresponding k values, shown in Table 2, indicate higher reaction rate on ZrSTA(5) compared to the other catalysts, despite the less ordered mesostructure and relative low surface area. Notably, the catalytic oxidation rate over ZrSTA(5) is about 17 times greater than mesoporous ZrO2 sample (0.168 h−1 vs. 0.010 h−1) under similar conditions. The enhancement of catalytic activity over ZrSTA(5) can be regarded as a result of synergistic interactions between the [SiW12O40]4− anions and ZrO2 matrix, as previously discussed. Nevertheless, the well dispersion of POM units and the open-framework structure are important virtues, which may also account for the high reactivity of these materials.
The ZrPTA(5) catalyst is stable and it can be reused for the next catalytic run without significant loss in catalytic activity. After catalysis, the solid ZrPTA(5) was separated from the reaction mixture by filtration, washed with fresh ethanol several times and dried at ambient conditions, and used for the next catalytic cycle. The results in Fig. S4 of the ESI,† show that the consumption of 1 remains as high as ∼96% and ∼99% after 30 min and 2 h reaction time, respectively, i.e. very close to the conversion levels achieved by the fresh catalyst. Nitrogen physisorption measurements on reused sample gave no sings of decomposition of mesostructure, indicating surface area of 149 m2 g−1 and total pore volume of 0.08 cm3 g−1 (Fig. S5†). Moreover, elemental EDS microanalysis and IR spectroscopy showed no change in the chemical composition and Keggin structure of the incorporated [SiW12O40]4− clusters of reused sample, thereby, confirming high durability of recyclability. In particular, the EDS spectra showed an average Zr/W atomic ratio of ∼97.3:
2.7, which corresponds to a STA content of ∼5.1 wt%, in accordance with the nominal composition of the as-prepared catalyst, see ESI, Fig. S6.† The IR spectrum displayed the characteristic Si–Oa, W
Od and W–Ob–W vibration bands at ∼1024, ∼941 and ∼882 cm−1 respectively, reflecting the Keggin structure of the 12-silicotungstic acids (Fig. S7 of the ESI†). Nevertheless, minimal decomposition or leaching of incorporated STA clusters (on the basis of EDS and IR analysis limitations) by the excess of H2O2 cannot be excluded.32
Footnote |
† Electronic supplementary information (ESI) available: N2 physisorption data for meso-ZrO2, catalytic data for ZrSTA(5) in different media, time evolution profiles of 1 oxidation over meso-ZrO2, ZrPTA(w) and ZrSTA(w) materials and PTA and STA compounds, and catalytic data, N2 adsorption–desorption isotherms and EDS and IR spectra of reused ZrSTA(5) catalyst. See DOI: 10.1039/c3ra46421k |
This journal is © The Royal Society of Chemistry 2014 |