Hydrothermal synthesis and catalysis of Nb2O5–WOx nanofiber crystal

Kazu Okumura *a, Takuya Tomiyama a, Shuhei Shirakawa a, Soichiro Ishida a, Takashi Sanada ab, Masazumi Arao b and Miki Niwa a
aDepartment of Chemistry and Biotechnology, Graduate School of Engineering, Tottori University, 4-101 Koyama-cho Minami, Tottori, 680-8552, Japan. E-mail: okmr@chem.tottori-u.ac.jp; Fax: +81-857-31-5684; Tel: +81-857-31-5257
bResearch Department, NISSAN ARC, LTD., Yokosuka, 237-0061, Japan. Fax: +81 046-866-5814; Tel: +81 046-866-7632

Received 31st August 2010 , Accepted 21st September 2010

First published on 18th October 2010


Abstract

Nb2O5–WOx synthesized by a hydrothermal method having a long nano-crystalline structure with ca. 10 nm diameter exhibited a high catalytic activity in the Friedel–Crafts alkylations and acylations when it was calcined in N2.


1. Introduction

Metal oxides that have an acidic character are promising catalysts in organic synthesis and gas-phase reactions. Many kinds of monolayer-type oxides such as WO3/ZrO2 and SO42/ZrO2 having an acidic character have been studied extensively.1 Another class of acidic oxides is mixed metal oxides including SiO2–Al2O3, SiO2–TiO2, and SiO2–ZrO2.2Nb2O5 mixed with WO3 or MoO3 may be categorized to the latter classes in which oxides with 6+ and 5+ valence states are mixed together. Hino et al. reported that Nb2O5–WO3 prepared by the coprecipitation method exhibited a high activity in alkane cracking and dehydration of ethanol.3 Recently, Tagusagawa and Domen et al. reported the mesopores Nb–W oxide catalyst active in the Friedel–Crafts alkylation and hydrolysis of sucrose.4 We have also found that Nb2O5–WO3 prepared by the coprecipitation method was active in the Friedel–Crafts alkylation.5 On the other hand, Nb2O5–WO3 attracted attention not only as an acidic catalyst but also as photocatalyst and functional materials such as electrochromic devices.6 By means of a conventional method, niobium tungstate with the tetragonal tungsten bronze structure has been synthesized by calcining the mixture of Nb2O5, NbO2, and WO3 at high temperatures.7 Similarly, Nb2O5–WO3 mixed oxide has been prepared by the coprecipitation method in which a mixed solution of niobium oxalate and ammonium tungstate was evaporated and thereafter calcined in air.8 In this study, we made an attempt to synthesize Nb2O5–WOx by a hydrothermal method with the aim of synthesizing a characteristic crystal that exhibits high catalytic activity. Hydrothermal synthesis has been applied to synthesize crystalline oxides with a high surface area and unique morphology.8 For instance, recently, Song et al. succeeded in synthesizing WO3 nanowires with this method.9 Unlike the usual synthesis method of Nb2O5–WO3 that requires a high temperature, hydrothermal synthesis is possible at a low temperature. Thus, hydrothermally synthesized Nb2O5–WOx (abbreviated as h-NbW) is expected to have a high surface area and enhanced intrinsic catalytic activity. Here, h-NbW was applied to the Friedel–Crafts alkylation of anisole with benzyl alcohol and to acylations performed using carboxylic acids as acylating agents. Its catalytic performance was compared with that of a catalyst prepared by the conventional coprecipitation method (abbreviated as c-NbW). In particular, acylations performed using carboxylic acids seem important to yield ketones from the viewpoint of green chemistry because water is the only side product, unlike conventional acylations carried out using carboxylic anhydrates and carboxylic chlorides that accompany the formation of carboxylic acid and HCl.10 In the present study, in addition to the synthesis of Nb2O5–WOx, the influence of the calcination method and Ar bubbling on the catalytic performance of Nb2O5–WOx was studied in detail.

2. Experimental

Nb2O5–WOx was synthesized by a hydrothermal method; typically, a solution of ammonium paratungstate, (NH4)10W12O41·5H2O (supplied by Wako Chemicals, 2.03 g), dissolved in 60 ml of water was mixed in a flask with ammonium niobium oxalate, NH4[NbO(C2O4)2(H2O)]·xH2O (supplied by Companhia Brasileira de Metalurgia e Mineração (CBMM) co., 0.263 g) dissolved in 10 ml of water and then bubbled with N2. The mixed solution was placed in a Teflon-sealed autoclave. The hydrothermal synthesis was carried out at 443 K while the bottle was continuously rotated at a speed of 15 rpm. The typical synthesis time was 48 h. After the synthesis, the formation of a white solid precipitate was observed at the bottom of the autoclave. The precipitate was calcined in an N2 flow (50 ml min−1) or in air by using a furnace at 773 K for 2 h prior to using it for catalytic reactions. Dark blue and light yellow solids were obtained after the calcination in N2 and air, respectively. For comparison, Nb2O5–WOx was prepared by the coprecipitation method. Hydrated niobium oxide, Nb2O5·nH2O (n = 3.8, measured by thermogravimetric analysis) kindly supplied by CBMM Co. was dissolved into an oxalic acid solution heated at 353 K. Similarly, (NH4)10W12O41 was thoroughly dissolved in deionized water at 353 K. Both the solutions were mixed and stirred vigorously at 353 K and then evaporated to dryness. Thus-obtained materials were calcined at 773 K for 3 h in an N2 flow. The molar ratio between Nb and W of c-NbW was 1[thin space (1/6-em)]:[thin space (1/6-em)]2.3.

Field emission scanning electron microscope (FE-SEM) images were taken by means of a JEOL JSM-6701F microscope with an acceleration voltage of 5 kV. Transmission electron microscope (TEM) images were taken by means of a HITACHI H-9000UHR microscope with an acceleration voltage of 300 kV. In order to confirm the special distribution of Nb and W, we also carried out the EDX analyses by means of a HITACHI HF-2000 microscope (acceleration voltage of 200 kV) equipped with a Kevex SIGMA spectrometer and a FEI Tecnai G2 F20 microscope (acceleration voltage of 200 kV) equipped with an EDAX r-TEM spectrometer. The specimen for TEM observation was prepared by the crushing method, and the TEM observations were carried out at room temperature. The crystalline structure was analyzed by XRD in ambient conditions using a Rigaku Ultima IV X-ray diffractometer with Cu Kα radiation. Data of N2 adsorption isotherms were collected with BELSORP-max. Samples were dehydrated at 573 K under vacuum prior to the measurements.

Synchrotron radiation experiments were performed at the BL01B1 station with the approval of the Japan Synchrotron Radiation Research Institute (JASRI) (Proposal No. 2009A1055). A Si(111) single crystal was used to obtain a monochromatic X-ray beam. The measurement was carried out in a quick mode. For the collection of W L3-edge data, the ion chambers filled with N2 and N2(50%)/Ar(50%) were used for I0 and I, respectively. For the collection of Nb-K edge data, the ion chambers filled with Ar and Ar(75%)/Kr(25%) were used for I0 and I, respectively. The energy was calibrated using W and Nb foils for the collection of W-L3 and Nb-K edge XAFS data, respectively. The data were analyzed using the REX2000 ver. 2.5.9 program (Rigaku Co.). Fourier transformation of k3χ(k) data were performed in a k range of 20–175 nm−1 for the analysis of the W L3-edge EXAFS spectra.

The acid property of synthesized Nb2O5–WOx was measured by means of TPD of ammonia with an equipment of Japan Bel TPD-1-AT(NH3). The sample was evacuated at 673 K prior to the measurement. 13.3 kPa of ammonia were equilibrated with the pretreated sample at 373 K. The TPD data were collected with the temperature ramping rate of 10 K min−1. A mass spectrometer was used to measure the desorbed NH3. In the measurement, m/z = 16 was monitored to analyze the desorbed NH3. This mass number was used instead of m/z = 17 in order to prevent the interference caused by water.

Benzylation of anisole (Friedel–Crafts alkylation) was carried out over the Nb2O5–WOx catalysts (catalyst weight: 0.02 g). The reaction was performed using 10 g of anisole and 0.675 g (6.25 mmol) of benzyl alcohol in an oil bath at 353 K in an atmosphere of N2 or with Ar bubbling (Ar: 99.9999%, 30 ml min−1). Acylation of anisole with various straight-chain carboxylic acids (Friedel–Crafts acylations) was carried out in a manner analogous to the benzylation of anisole using 10 g of anisole and 2 mmol of carboxylic acids in an oil bath at 413 K (catalyst weight: 0.1 g). For recycling, the catalyst was separated by filtration followed by washing with anisole. It was then repeatedly used for further reaction without pretreatment. The products were analyzed with a gas chromatograph (GC-2010, Shimadzu) equipped with a capillary column (InertCap 5) and an FID detector. In the analysis, tridecane was used as an internal standard.

3. Results and discussion

3.1. Hydrothermal synthesis of Nb2O5–WOx

Nb2O5–WOx was hydrothermally synthesized using a mixed solution of (NH4)10W12O41 and NH4[NbO(C2O4)2(H2O)]. In Fig. 1a, the amount of the obtained solid precipitate is plotted as a function of the synthesis time. A steep increase in the amount of the precipitate was observed in less than 12 h. The yield of the obtained h-NbW was calculated from the loaded amount of (NH4)10W12O41·5H2O and NH4[NbO(C2O4)2(H2O)]·xH2O was found to be 72% and 92% after 48 h and 336 h from the beginning of the synthesis, respectively. In contrast, no precipitate was found to form over (NH4)10W12O41·5H2O without the addition of Nb source or NH4[NbO(C2O4)2(H2O)]·xH2O without the addition of W source, implying that the combination of W and Nb was necessary to form a solid precipitate. In Fig. 1b, the molar ratio between W and Nb of the synthesized oxide is plotted as a function of the synthesis time. The composition of Nb2O5–WOx was measured by the inductively coupled plasma (ICP) method after melting Nb2O5–WOx with KOH at 773 K and then dissolving the obtained residue in water. Fig. 1b indicates that the W/Nb ratio tended to decrease with an increase in the synthesis time, meaning an increase in the amount of Nb incorporated into Nb2O5–WOx.
Dependence of (a) the amount of solid and (b) W/Nb ratio plotted as a function of time for hydrothermal synthesis.
Fig. 1 Dependence of (a) the amount of solid and (b) W/Nb ratio plotted as a function of time for hydrothermal synthesis.

3.2. Structural and acidic characterizations

Fig. 2 shows the FE-SEM images of h- and c-NbW with the magnitude of ×100[thin space (1/6-em)]000. The image of c-NbW calcined in N2 was almost shapeless (Fig. 2a). In contrast to this, the formation of fiber-like shape, which aggregate and make bundles, was clearly seen in the image of h-NbW calcined in N2 (Fig. 2b). The maximum length of the fibers was at least several micrometres. The diameter of the fibers was ca. 10 nm. The fiber-like shape was almost collapsed after grinding in a mortar or calcination in air at 773 K to give aggregates of round-shaped particles (Fig. 2c and d). These FE-SEM images indicated that the formation of nanofiber was realized only when h-NbW was calcined in N2. TEM images of h-NbW calcined in an N2 flow are shown in Fig. 3. In agreement with the FE-SEM images (Fig. 2b), crystallites with fiber-like shape are clearly visible in the images. The electron diffraction patterns taken from the nanofibers indicated that the fiber was single-crystalline which has tetragonal structure similar to tetragonal tungsten bronze (TTB), Na1.2Nb1.2W0.8O6, with a preferential growth in the [001] direction.11 However, the broadness of the diffraction spots as shown in the inset of Fig. 3 suggested the low crystallinity of the synthesized fiber. The appearance of streaks also suggested that single fiber contains several planar defects. Fig. 4 shows FFT images of the TEM lattice images of the region a–c in Fig. 3. No significant difference in the patterns was found in these images, indicating the homogeneous structure in a Nb–W nanofiber. In order to measure the distribution of Nb and W atoms, EDX semi-quantitative analysis in different regions of Fig. 5 with area from 5 to 500 nm φ was made using Nb-Kα and W-Lα lines. The measured composition of the sample was listed in Table 1. No substantial deviation from the ratio Nb[thin space (1/6-em)]:[thin space (1/6-em)]W = 1[thin space (1/6-em)]:[thin space (1/6-em)]9.1 was found before and after the calcination in N2. That is to say, the average molar ratio of as prepared and the calcined samples were calculated to be 8.9 and 9.3, respectively, which almost agreed with the bulk composition (W/Nb = 9.1) of the Nb–W measured by ICP (Fig. 1b, 48 h). Further, EDX spectra images (mapping) of Nb-Kα and W-Lα of h-NbW calcined in N2 were displayed in Fig. 6. We confirmed that Nb and W were homogeneously distributed in the nanofiber, suggesting the formation of solid solutions.

            FE-SEM images of (a) c-NbW calcined in N2, (b) h-NbW calcined in N2, (c) h-NbW calcined in N2 followed by grinding with a mortar, and (d) h-NbW calcined in air at 773 K.
Fig. 2 FE-SEM images of (a) c-NbW calcined in N2, (b) h-NbW calcined in N2, (c) h-NbW calcined in N2 followed by grinding with a mortar, and (d) h-NbW calcined in air at 773 K.

The TEM image and the electron diffraction pattern (inset) of h-NbW calcined in an N2 flow.
Fig. 3 The TEM image and the electron diffraction pattern (inset) of h-NbW calcined in an N2 flow.


            FFT images of the region a–c in the TEM image of Fig. 3.
Fig. 4 FFT images of the region a–c in the TEM image of Fig. 3.


            TEM images of h-NbW: (a and b) as prepared and (c and d) calcined in an N2 flow at 773 K. EDX semi-quantitative analyses were made in regions indicated by red squares.
Fig. 5 TEM images of h-NbW: (a and b) as prepared and (c and d) calcined in an N2 flow at 773 K. EDX semi-quantitative analyses were made in regions indicated by red squares.
Table 1 Atomic (molar) ratio of Nb[thin space (1/6-em)]:[thin space (1/6-em)]W in different regions of Nb2O5–WOx fibers in Fig. 5
Image no. Region W/Nb atomic ratio
a 1 8.75
a 2 10.95
b 1 8.96
b 2 8.44
b 3 7.79
b 4 8.57
c 1 9.56
c 2 9.50
d 1 9.78
d 2 8.31



Bright field and EDX spectrum images of h-NbW calcined in N2. (a) Bright field, EDX mapping of (b) Nb-Kα and (c) W-Lα.
Fig. 6 Bright field and EDX spectrum images of h-NbW calcined in N2. (a) Bright field, EDX mapping of (b) Nb-Kα and (c) W-Lα.

X-Ray diffraction (XRD) patterns of h-, c-NbW and reference samples are presented in Fig. 7. The pattern of h-NbW calcined in air at 773 K (d) was close to that of monoclinic crystalline WO3 (e), indicating crystalline WO3 and amorphous Nb2O5 were segregated. In contrast to this, the pattern of h-NbW calcined in N2 was almost featureless, but intense diffraction peaks appeared at 2θ = 23.3 and 47.7°, which were tentatively assignable to the (001) and (002) planes from the comparison with the XRD pattern of TTB structure, respectively.11 Intensity of these diffraction lines became smaller after grinding in a mortar (b), probably due to the disruption of the nanofiber as observed in FE-SEM images (Fig. 2c). Although the intensity of the peaks was broad and smaller, a similar pattern was observed in the XRD of c-NbW (c).12



            Powder XRD patterns of Nb2O5–WOx and reference samples. (a) h-NbW calcined in N2, (b) h-NbW calcined in N2, followed by grinding with a mortar, (c) c-NbW calcined in N2, (d) h-NbW calcined in air, (e) WO3, and (f) Nb2O5.
Fig. 7 Powder XRD patterns of Nb2O5–WOx and reference samples. (a) h-NbW calcined in N2, (b) h-NbW calcined in N2, followed by grinding with a mortar, (c) c-NbW calcined in N2, (d) h-NbW calcined in air, (e) WO3, and (f) Nb2O5.

Fig. 8 shows N2 adsorption isotherms of h- and c-NbW. The N2 adsorption isotherms categorized to the type II were obtained in these samples. The Brunauer–Emmett–Teller (BET) surface area of h-NbW calcined at 773 K in an N2 flow was 63 m2 g−1, which was much higher than that calcined in air (16 m2 g−1), h-NbW ground with a mortar (39 m2 g−1) and c-NbW (15 m2 g−1). The low specific surface area of h-NbW calcined in air may be due to the formation of crystalline WO3 as confirmed by the XRD pattern shown in Fig. 7d. Thus, it is noted that the hydrothermal synthesis coupled with calcination in an N2 flow was effective in increasing the surface area of Nb2O5–WOx due to the persistence of fiber-like structure.


N2 adsorption isotherms of (a) c-NbW calcined in N2, (b) h-NbW calcined in N2, (c) h-NbW calcined in N2, followed by grinding with a mortar, and (d) h-NbW calcined in air. Closed symbols: adsorption branches; open symbols: desorption branches.
Fig. 8 N2 adsorption isotherms of (a) c-NbW calcined in N2, (b) h-NbW calcined in N2, (c) h-NbW calcined in N2, followed by grinding with a mortar, and (d) h-NbW calcined in air. Closed symbols: adsorption branches; open symbols: desorption branches.

Fig. 9 shows the W L3-edge EXAFS spectra of h-NbW k3χ(k) and their Fourier transforms. The EXAFS spectrum of h-NbW calcined in air was similar to that of crystalline-WO3, in agreement with the XRD data. In the W L3-edge EXAFS Fourier transform of h-NbW calcined in N2, the peak assignable to the W–O–W bond could be seen at 0.37 nm (Fig. 9b). This feature was partially similar to that of WO2, suggesting the formation of W4+ species. This peak corresponded to the characteristic rapid oscillations that can be observed at 80–170 nm−1 in the k3χ(k) spectra (Fig. 9a). Fig. 10a gives the W L3-edge XANES of h-NbW calcined at 773 K in N2 and in air, together with those of the reference samples. The steep white line appeared at 10[thin space (1/6-em)]209 eV. This is attributed to the electronic transition from 2p3/2 to the vacant 5d orbital. Therefore, the white line reflects the electronic state of the vacant d orbitals of the W atom.13 The intensity of the white line of h-NbW calcined in air was comparable to that of WO3, meaning that the valence of the uncalcined sample was 6+. Unlike this, the intensity of h-NbW calcined in N2 was comparable to that of WO2, suggesting that W was reduced to tungsten oxide with a lower valence state (4+) as a result of calcination in N2 at 773 K. In agreement with the XANES data, the color of h-NbW changed from white (uncalcined sample) to dark blue as a result of calcination in N2, while the color of h-NbW calcined in air was light yellow. Fig. 10b presents the Nb K-edge XANES of h-NbW calcined in N2 and air, together with that of bulk Nb2O5. The spectra of h-NbW were similar to that of Nb2O5, suggesting that the valence of niobium oxide was 5+. Fig. 11 shows the temperature-programmed desorption (TPD) of ammonia from the prepared samples. A broad desorption peak of ammonia is observed in these spectra. The acid amount was calculated from the desorbed ammonia. The acid amounts of h-NbW samples that were calcined in N2 and air were determined to be 0.25 and 0.09 mol kg−1, respectively, indicating that the calcination of the sample in N2 was effective in enhancing the acid amount of h-NbW. The amount decreased to 0.19 mol kg−1 after grinding with a mortar. The acid amount of c-NbW calcined in N2 was 0.21 mol kg−1. The amount of c-acid NbW was much lower than that of h-NbW.


W L3-edge EXAFS of h-NbW calcined in N2, air and reference samples: (a) k3χ(k) and (b) Fourier transforms.
Fig. 9 W L3-edge EXAFS of h-NbW calcined in N2, air and reference samples: (a) k3χ(k) and (b) Fourier transforms.

(a) W L3- and (b) Nb K-edge XANES of h-NbW calcined in N2 and in air and those of reference samples.
Fig. 10 (a) W L3- and (b) Nb K-edge XANES of h-NbW calcined in N2 and in air and those of reference samples.


            Temperature-programmed desorption of ammonia from (a) h-NbW calcined in air, (b) h-NbW calcined in N2, (c) h-NbW calcined in N2, followed by grinding with a mortar and (d) c-NbW calcined in N2.
Fig. 11 Temperature-programmed desorption of ammonia from (a) h-NbW calcined in air, (b) h-NbW calcined in N2, (c) h-NbW calcined in N2, followed by grinding with a mortar and (d) c-NbW calcined in N2.

3.3. Friedel–Crafts alkylations

Table 2 lists the data of the Friedel–Crafts alkylation of anisole with benzyl alcohol carried out over the h- and c-NbW catalysts. First, the influence of the gas atmosphere on the calcination of h-NbW was investigated. h-NbW calcined in air was completely inactive in the reaction (entry 1), probably due to the formation of crystalline WO3, as evidenced by XRD and EXAFS. However, the conversion of benzyl alcohol increased from 0% to 34% through the calcination of h-NbW in N2 (entry 3), indicating that the calcination in N2 in place of air was effective in enhancing the catalysis of h-NbW. Second, the influence of Ar bubbling (30 ml min−1) over the h-NbW catalyst calcined in N2 was examined. From the comparison of the data of entries 3 and 5, a remarkable enhancement in the catalytic activity was confirmed. That is, the conversion of benzyl alcohol increased from 34% to 99% through the application of Ar bubbling into the solution. The ratio of the p and o-alkylated products was 58[thin space (1/6-em)]:[thin space (1/6-em)]42, as analyzed by the GC. Formation of the multiply alkylated products was not observed. The amount of dibenzyl ether (byproduct) formed was less than 4% in every reaction. The reaction was performed after the addition of a dried molecular sieve in order to check the possibility accelerating catalysis through the removal of H2O accompanied by Ar bubbling as a result of a shift in the reaction equilibrium (entry 4). Although the conversion of benzyl alcohol increased from 34% to 51% by the addition of the molecular sieve, the extent of improvement was much lower than that obtained by Ar bubbling (entry 5), indicating that the effect of H2O removal by Ar bubbling was limited. Although the remarkable effect of Ar bubbling on the catalytic activity is not fully understood at this stage, one hypothesis is that Ar was effective in keeping the air-free conditions; thus, the low valence state of W oxide and the fiber-like shape were maintained during reactions as evidenced by TEM and FE-SEM. In agreement with this, the addition of 0.3% O2 in Ar resulted in severe deactivation of the catalyst (conv. = 18%, entry 7). Third, the effect of preparation method was examined. As compared with N2-calcined samples, h-NbW exhibited much higher conversion and TON (entry 5, conv. = 99%, TON= 1600) in comparison with c-NbW (entry 8, conv. = 1%, TON = 20) and H-β zeolite (entry 10, conv. = 13%, TON = 36). The h-NbW ground with a mortar was almost inactive in the reaction (entry 9, conv. = 2%, TON =25) despite the presence of substantial acid amount as measured by NH3-TPD (0.19 mol kg−1). The fact suggested that the nanofiber structure was important in the evolution of catalytic activity. In order to examine the possibility for the dissolution of W in anisole, the concentration of W in anisole was measured by the ICP method after the reaction performed using h-NbW. The dissolved amount of W was less than 0.2% of the total W present in the catalyst. Furthermore, no reaction proceeded after the addition of substrates into the filtrate. These facts suggested that the reaction proceeded on the solid h-NbW catalyst. As a whole, the following three prerequisites are necessary to evolve high activity in a Nb2O5–WOxcatalyst: (1) hydrothermal synthesis, (2) calcination in N2 and (3) Ar bubbling during reactions.
Table 2 Friedel–Crafts alkylation over Nb2O5–WOx catalystsa
ugraphic, filename = c0jm02882g-u1.gif
Entry Catalyst Calcination condition Bubblingf Benzyl alcohol conv. (%) Benzyl anisole yield (%) Dibenzyl ether yield (%) Material Balance (%) TONg
a Reaction conditions: anisole (92.5 mmol), benzyl alcohol (6.2 mmol), catalyst (0.02 g), 353 K, 3 h. b Dried molecular sieve (1.0 g) was added. c Catalyst weight, 0.005 g. Ar, 80 ml min−1. d h-NbW was ground in a mortar for 30 min. e H-β-20 (Si/Al2 = 25), supplied by PQ company. f Total flow rate was 30 ml min−1. g Turnover numbers calculated based on the conversion of benzyl alcohol and acid amounts determined by NH3-TPD.
1 h-NbW Air 0 0 0 100 0
2 h-NbW Air Ar 0 0 0 100 0
3 h-NbW N2 34 26 3 98 550
4b h-NbW N2 51 41 3 97 830
5 h-NbW N2 Ar 99 89 4 98 1600
6c h-NbW N2 Ar 68 60 4 100 4400
7 h-NbW N2 Ar + 0.3% O2 18 15 2 100 290
8 c-NbW N2 Ar 1 1 0 100 20
9d h-NbW N2 Ar 2 2 0 100 25
10 H-β zeolite e N2 Ar 13 5 2 96 36


3.4. Friedel–Crafts acylations

h-NbW calcined in N2 was applied to the Friedel–Crafts acylation of anisole by using various carboxylic acids. The catalyst was calcined at 773 K in an N2 flow prior to the reactions. Table 3 lists the data of the acylations performed with Ar bubbling. h-NbW exhibited high activity in the reaction performed using various carboxylic acids, with an exception of benzoic acid (entry 9). p-Acylated ketones were obtained as primary products, where the molar ratio of p- and o-acylated products was 98[thin space (1/6-em)]:[thin space (1/6-em)]2. The yield of esters was less than 3% in every reaction. In marked contrast to h-NbW, in agreement with the data of alkylations given in Table 2, c-NbW was almost inactive in the reaction (entry 4). In addition, the yield of ketones was almost three-times higher than that obtained with H-β zeolite (supplied by PQ Co., Si/Al2 = 25, entry 5). This fact indicated the superior nature of h-NbW calcined in N2. The recycle use of the h-NbW catalyst was possible, as can be seen in the reaction data using butanoic acid (entry 2).
Table 3 Friedel–Crafts acylations of anisole with various straight-chain carboxylic acidsa
ugraphic, filename = c0jm02882g-u2.gif
Entry R Catalyst Carboxylic acid conv. (%) Ketone yield (%) Methyl ester yield (%) Phenyl ester yield (%) Material balance (%) TONe
a Reaction conditions: anisole (92.5 mmol), carboxylic acid (2.0 mmol), catalyst (0.1 g), 413 K, 3 h, Ar bubbling (30 ml min−1). b 2nd run. c Benzoic acid. d H-β-20 (Si/Al2 = 25), supplied by PQ company. e Turnover numbers calculated based on the conversion of carboxylic acids and acid amounts determined by NH3-TPD.
1 C3H7 h-NbW 93 82 1 0 91 98
2b C3H7 h-NbW 90 76 1 0 87 95
3 C5H11 h-NbW 97 93 0 0 103 102
4 C5H11 c-NbW 0 0 0 0 100 0
5 C5H11 H-β zeolite d 46 35 0 0 90 7
6 C7H15 h-NbW 91 82 1 1 93 96
7 C9H11 h-NbW 98 89 0 3 94 103
8 C11H23 h-NbW 77 62 1 2 85 81
9 C6H5c h-NbW 33 22 0 0 90 35


Taking account of the W L3-edge XANES data, TEM and SEM images, it is postulated that the formation of partially reduced tungsten atoms (probably W4+) in the crystalline Nb2O5–WOx during the course of calcination in an N2 flow was responsible for the evolution of catalysis in the Friedel–Crafts reactions. This assumption is in agreement with the result reported by several groups in that partially reduced tungsten oxides exhibit the Brønsted acid character.14 Probably, the oxalate anion or ammonium cation acted as a reductant for tungsten oxide during calcination of h-NbW in N2. Consistent with the assumption, it has been reported that the formation of tungsten bronze through the decomposition of ammonium paratungstate changed depending on the gas employed. Namely, the calcination of ammonium paratungstate in He and O2 resulted in the formation of partially reduced tungsten bronze and crystalline triclinic WO3, respectively.15 Therefore, it is inferred that the role of N2 is to keep the inert atmosphere while reducing nanofibers with oxalate anions or ammonium cations.

4. Conclusions

Hydrothermal synthesis of Nb2O5–WOx was carried out over a mixed solution of Nb and W in an autoclave at 443 K. Synthesized Nb2O5–WOx had a long nano-crystalline structure with ca. 10 nm diameter as observed by FE-SEM and TEM. It exhibited a high catalytic activity in the Friedel–Crafts alkylations and acylations when it was calcined in N2. Ar bubbling was effective in enhancing the catalytic activity of N2-calcined Nb2O5–WOx. Comparison of the catalytic performance and structural characterizations revealed that the nanofiber structure with partially reduced tungsten oxide species was indispensable to afford high catalytic activity in the Friedel–Crafts reactions.

References

  1. M. Hino and K. Arata, Chem. Commun., 1988, 1259 RSC; X. M. Song and A. Sayari, Catal. Rev. Sci. Eng., 1996, 38, 329 CrossRef CAS.
  2. K. Tanabe, in Catalysis-Science and Technology, ed. J. R. Anderson and M. Boudart, Springer-Verlag, Berlin, 1981 Search PubMed; T. Shishido, T. Kitano, K. Teramura and T. Tanaka, Catal. Lett., 2009, 129, 383 Search PubMed.
  3. M. Hino, M. Kurashige and K. Arata, Catal. Commun., 2004, 5, 107 CrossRef CAS; M. Hino, M. Kurashige, H. Matsuhashi and K. Arata, Appl. Catal., A, 2006, 310, 190 CrossRef CAS.
  4. C. Tagusagawa, A. Takagaki, A. Iguchi, K. Takanabe, J. N. Kondo, K. Ebitani, S. Hayashi, T. Tatsumi and K. Domen, Angew. Chem., Int. Ed., 2010, 49, 1128 CrossRef CAS.
  5. K. Yamashita, M. Hirano, K. Okumura and M. Niwa, Catal. Today, 2006, 118, 385 CrossRef CAS.
  6. P. R. Bueno, C. Gabrielli and H. Perrot, Electrochim. Acta, 2008, 53, 5533 CrossRef CAS; C. Martin, I. Martin, V. Rives, G. Solana, V. Loddo, L. Palmisano and A. Sclafani, J. Mater. Sci., 1997, 32, 6039 CrossRef CAS; M. L. Sanjuan, A. Kuhn, M. T. Azcondo and F. Garcia-Alvarado, Eur. J. Inorg. Chem., 2008, 49 CrossRef CAS.
  7. F. Krumeich, M. Worle and A. Hussain, J. Solid State Chem., 2000, 149, 428 CrossRef CAS.
  8. W. Ueda, K. Oshihara, D. Vitry, T. Hisano and Y. Kayashima, Catal. Surv. Jpn., 2002, 6, 33 CrossRef CAS.
  9. X. C. Song, Y. F. Zheng, E. Yang and Y. Wang, Mater. Lett., 2007, 61, 3904 CrossRef CAS.
  10. K. Okumura, J. Jpn. Pet. Inst., 2009, 52, 219 Search PubMed.
  11. M. L. Sanjuán, A. Kuhn, M. T. Azcondo and F. García-Alvarado, Eur. J. Inorg. Chem., 2008, 49 CrossRef CAS.
  12. P. Botella, E. García-González, B. Solsona, E. Rodríguez-Castellón, J. M. González-Calbet and J. M. López Nieto, J. Catal., 2009, 265, 43 CrossRef CAS.
  13. S. Yamazoe, Y. Hitomi, T. Shishido and T. Tanaka, J. Phys. Chem. C, 2008, 112, 6869 CrossRef CAS.
  14. H. Belatel, H. Al-Kandari, F. Al-Kharafi, F. Garin and A. Katrib, Appl. Catal., A, 2007, 318, 227 CrossRef CAS; A. Benadda, A. Katrib, J. W. Sobczak and A. Barama, Appl. Catal., A, 2004, 260, 175 CrossRef CAS.
  15. O. Kirilenko, F. Girgsdies, R. E. Jentoft and T. Ressler, Eur. J. Inorg. Chem., 2005, 2124 CrossRef CAS; I.M. Szilágyi, F. Hange, J. Madarasz and G. Polol, Eur. J. Inorg. Chem., 2006, 3413 CrossRef CAS.

This journal is © The Royal Society of Chemistry 2011
Click here to see how this site uses Cookies. View our privacy policy here.