16604362

(1)

Ammoxidation of propane over V-Sb and V-Sb-Nb mixed oxides

M.O. Guerrero-Pe´rez

1

, M.V. Martı´nez-Huerta

2

, J.L.G. Fierro, M.A. Ban˜ares

*

Instituto de Cata´lisis y Petroleoquimica, CSIC, Marie Curie 2, E-28049-Madrid, Spain

Received 15 July 2005; received in revised form 7 September 2005; accepted 15 September 2005 Available online 3 November 2005

Abstract

Pure VSbO4phase is not active for propane ammoxidation, unless some segregateda-Sb2O4is present on its surface. Surface vanadium may be

necessary to activate propane. For the ternary Sb-V-Nb oxide catalysts: Sb-V-O, Nb-V-O and Sb-Nb-O phases may form. SbNbO4is inert and forms

at the expense of efficient Sb-V-O and V-Nb-O phases. The extent of formation of these phases strongly depends on the preparation method. The Sb precursor determines if V-Sb or Nb-Sb is the preferred interaction. Tuning the interaction of Sb with the other elements affects the catalyst structure and performance.

Keywords:V-Sb-Nb-O; V-Sb-O; Oxidation; Ammoxidation; Propane; Acrylonitrile; Structure–activity relationship; In situ Raman; XRD

1. Introduction

Acrylonitrile has been industrially produced from propylene as an important intermediate for the preparation of fibers, important polymers as styrene–acrylonitrile, acrylonitrile– butadiene–styrene, and other valuables chemicals [1]. The direct conversion of propane into acrylonitrile by reaction with oxygen and ammonia is an alternative route to the conventional propylene ammoxidation since propylene is typically more expensive than propane. The economic implications of this new route are very important, but the activation of propane is the limiting step[2]. Since the adsorption rate of propane is near ten times smaller than that of propylene [3] the conversion of propane is at least ten times smaller than propylene [4]. In addition, the reaction conditions to activate the C–H bond in propane are more energy demanding, which has a negative effect on selectivity.

There are several studies about the catalysts used for propane ammoxidation, but most of the works concentrate on two types of catalysts [1,5]: the antimonates with rutile

structure [6,7] and the molybdates [1–18]. Both catalyst systems usually incorporate vanadium as an essential ingre-dient. Mo-V catalysts modified with Nb and Te may afford near 50% yield to acrylonitrile[2] and acrylonitrile yields of 60% have been reported for the antimonates system[9]. Despite the large relevance of these catalysts, the nature of the active site is still not fully understood. Sb-V-O catalysts with an excess of V are highly active and selective for propane oxidative dehydrogenation (ODH) while an excess of Sb affords Sb-V-O catalysts that are more efficient for propane ammoxidation [19].

Several preparation methods have been reported for the Sb-V-O-based catalysts and it is concluded that the preparation method has an important effect on the structure and on the performance [12,20,21]. The methods most usually studied correspond to patents filed in 1988 by Standard Oil Co.[9–11]and it consists on keeping in reflux during several hours an aqueous suspension of NH4VO3or V2O5and Sb2O3. Others studies use solutions of V4+ in oxalic acid to which antimonic acid is added [22]or even prepare catalysts with a solution of VCl3 and SbCl5on HCI [8,23]. Brazdil et al. have also developed a sol–gel method for this kind of catalysts [21]. The aim of all these preparation methods is to provide an intimate contact between Sb5+and V4+ in order to improve their redox reaction to form VSbO4phase [24], which is directly related to acrylonitrile formation [2,13,19–24], and provides the catalyst with a high specific area. More recently, mechanical activation has also proved to be an interesting method to prepare VSbO4phases[25].

www.elsevier.com/locate/apcata

* Corresponding author. Tel.: +34 91 585 4788; fax: +34 91 585 4760.

E-mail address:[email protected] (M.A. Ban˜ares).

1_{Actual address: Departamento de Ingenierı´a Quı´mica, Universidad de}

Ma´laga, Campus de Teatinos, E-29071-Ma´laga, Spain. 2

Actual address: Departamento de Quı´mica Fı´sica, Universidad de La Laguna, c/Astrofı´sico Francisco Sa´nchez s/n, E-38071 La Laguna, Tenerife, Spain.

(2)

Grasselli [1,5] reports on ternary mixed metal oxide catalysts with the formula VSbxMyOz, where M stands for

W[13,26], Te[9,13], Ni[13,27], Mg[13], Nb[13], Sn[28,29],

Bi [30], Al [26,32,33], Fe [34], and Ti [35], among others.

Usually the metal M forms a solid solution with the rutile structure of VSbO4forming mixed phases with Sb and V, like Al1 xSbVxOx[32]or Sb0.9V0.9 xWxO4[26]or V-Sb-Ti-O[58]. Niobium compounds and materials such as niobium pentoxide, niobic acid, niobium phosphate, and mixed oxides containing niobia are shown to be useful on catalysis as supports and as promoters for different reactions[37–41]. Specially, catalysts containing niobium have been found active in dehydrogenation reactions [42–44] and the dehydrogenation of propane is commonly proposed as the first step in propane ammoxidation [23,45–47], so Nb seems to be a key element on propane ammoxidation to acrylonitrile, specially mixed oxides contain-ing Nb (Mo-V-Nb-Te catalysts) have been found active and selective propane ammoxidation[13,48–50].

A third element may also be present as the cation of an oxide support. The use of an oxide to support/promote VSbO4 is particularly interesting, for instance, alumina-supported VSbO4

[31]; it performs like bulk VSbO4 with a more economic

composition [51]. Other supports, like niobia, significantly promote the performance of VSbO4[52]. These catalysts are prepared by impregnation of the oxide support. It would be desirable to assess direct synthesis of efficient V-Sb-Nb oxide catalysts for propane ammoxidation. However, the use of niobia, vanadia and antimony supports may lead to efficient Nb-promoted Sb-V-O phases or to inert SbNbO4phases. This manuscript evaluates the relevance of different precursors on the preparation of V-Sb-Nb oxide catalysts.

2. Experimental

Bulk Sb-V-O catalysts (SV series) were prepared drying in a rotatory evaporator at 808C NH4VO3(Sigma) and antimony that was kept in solution according to a method developed in our laboratory[49]. The resulting solid was dried at 1158C for 24 h and then calcined in air at 6608C for 12 h and at 7508C for 24 h. Four catalysts were prepared SV1, SV3, SV5, and SV7 with Sb/V molar ratios of 1, 3, 5, and 7, respectively.

Sb-V-Nb-O catalysts (SVN series) were prepared with three different methods. The Method 1 uses an aqueous solution with SbCb, NH4VO3, and Nb2(C2O4)5 and two catalysts were prepared SVN1A and SVN1B (Table 1), with different molar ratios. In this method a solution of Nb2(C2O4)5[50]was added to an aqueous solution of an appropriate amount of NH4VO3 and SbCl3at 608C. The solution was kept under stirring for 2 h at this temperature, and then was separated by centrifugation. The samples were dried at 1208C for 12 h and calcined in air at 4008C for 5 h. The Method 2 uses an organic solution and was used for the preparation of SVN2 catalyst, a suspension of Nb2(C2O4)5in CH2Cl2was added to a solution of SbCl5and (C5O2H8)2VO in CH2Cl2. The solution was stirred at 608C for 2 h and the solvent was evaporated. The catalyst powder was dried at 1208C for 16 h and calcined in air at 5008C for 5 h. Method 3 uses the same molar ratio and methodology than 2, but Nb2O5is used as the precursor of Nb, catalyst SVN3 was prepared according this method.

Raman spectra were run with a single-monochromator Renishaw System 1000 equipped with a thermoelectrically cooled CCD detector ( 738C) and holographic super-Notch filter. The holographic Notch filter rejects the elastic scattering. The samples were excited with the 514 nm Ar line; spectral resolution was ca. 3 cm 1and the spectrum acquisition time was 300 s. The spectra were obtained under dehydrated conditions (ca. 390 K) in a hot stage (Linkam TS-1500). Raman spectro-scopy provides information about the bulk of the material.

X-ray diffraction patterns were recorded on a Siemens Krystalloflex D-500 diffractometer using Cu Ka radiation (l= 0.15418 nm) and a graphite monochromator. Working conditions were 40 kV, 30 mA, and scanning rate of 28min 1 for Bragg’s angles (2u) from 5 to 708. Nitrogen adsorption isotherms ( 1968C) were recorded on an automatic Micro-meritics ASAP-2000 apparatus. Prior to the adsorption experiments, samples were outgassed at 413 K for 2 h. BET areas were computed from the adsorption isotherms (0.05<P/ Po<0.27), taking a value of 0.164 nm2for the cross-section of the adsorbed N2 molecule at 1968C. ICP analyses were performed in a 3300 Perkin-Elmer apparatus. Around 100 mg of sample were dissolved with nitric, chloride, and fluoride acids in a microware oven.

Table 1

Precursors used during synthesis, composition and BET area of catalysts

Catalysts Compositiona _{BET surface area (m}2_g 1₎ _Method _Precursors

Sb-V-O catalysts

SV1 VSb 0.6 – NH4VO3, Sb soluble complex

SV3 VSb3 0.5 – NH4VO3, Sb soluble complex

Sb-V-Nb-O catalysts

SVN1A V0.1Sb0.3Nb0.3 72 1 SbCl3, NH4VO3, Nb2(C2O4)5

SVN1B V0.1Sb0.3Nb0.1 19 1 SbCl3, NH4VO3, Nb2(C2O4)5

SVN2 V0.2Sb0.3Nb0.5 8 2 SbCl5, (C5O2H8)2VO, Nb2(C2O4)5

SVN3 V0.2Sb0.1Nb0.4 12 3 SbCl5, (C5O2H8)2VO, Nb2O5

(3)

Activity measurements were performed using a conven-tional microreactor with on-line gas chromatograph equipped with a flame ionization and thermal conductivity detector. The correctness of the analytical determinations was checked for each test by verification that the carbon balance (based on the propane converted) was within the cumulative mean error of the determinations (10%). Tests were made using 0.2 g of sample with particle dimensions in the 0.25–0.125 mm range. The axial temperature profile was monitored by a thermocouple sliding inside a tube inserted into the catalytic bed. Tests were made using the following feedstock: 25% O2, 9.8% propane, 8.6% ammonia, and helium balance. The total flow rate was 20 ml min 1 corresponding to a gas-space velocity (GHSV) near 3000 h 1. The quantity of catalyst and total flow were determined in order to avoid internal and external diffusion contributions. Yield and selectivity values were determined on the basis of the moles of propane feed and products, considering the number of carbon atoms in each molecule.

3. Results

3.1. Sb-V-O system

Table 1represents the composition and BET area values for

the Sb-V-O catalysts. The BET area values are quite low for these catalysts, probably due to the high calcination tempera-ture. The XRD patterns of the fresh catalyst are displayed in

Fig. 1. SV1 pattern (Sb/V = 1) corresponds to the VSbO4phase

(JCPDS file 16-0600), the patterns of SV3, SV5, and SV7 catalysts (Sb/V = 3, 5, and 7, respectively) correspond to VSbO4anda-Sb2O4phases (JCPDS file 11-0694 and 32-0042). The Raman spectra of the fresh and used catalyst are rep-resented inFig. 2. The spectrum of the fresh SV1 catalyst pre-sents a band centered at 880 cm 1, which has been assigned to the Sb0.92V0.92O4phase[19]. The weak Raman band near 1020 cm 1 is not sensitive to humidity, so this band must correspond to

the VSbO4phase and not to surface vanadium oxide species. However, this band is quite broad, and it is not possible to provide a conclusive absence of surface vanadium oxide species.

The Raman spectra of catalysts with antimony excess (SV3, SV5, and SV7) present modes at 190, 255, 372, 451, and 716 cm 1 that correspond to the a-Sb2O4 phase [51], these bands are very strong while VSbO4 possesses a low Raman section; therefore, Fig. 2B blows-up the 530–1030 cm 1 window. The weak bands at 610, 650, 710, 749, and 820 cm 1that correspond to thea-Sb2O4phase are evident. Raman spectra of all catalysts with antimony excess show the bands that correspond to the a-Sb2O4phase. The SV1, SV3, and SV5 catalysts exhibit the bands of VSbO4 phase; these bands are particularly weak for catalyst SV7. The Raman spectra show no appreciable changes for the used catalysts.

These catalysts exhibit a steady-state operation for at least 20 h. Fig. 3 represents the yields to different products for

Fig. 1. XRD patterns of fresh Sb-V-O catalysts.

(4)

catalysts SV1, SV3, SV5, and SV7 at 5008C. The catalyst with the highest vanadium content (SV1, with Sb/V = 1) is the most active and presents the highest yield to acrylonitrile; however, selectivity to acrylonitrile is rather low. Sb excess leads to less active, but more acrylonitrile selective, catalysts. The catalysts SV5 is most selective towards acrylonitrile.

3.2. Sb-V-Nb-O system

XRD patterns of fresh and used Sb-V-Nb-O catalysts are represented inFig. 4. The catalyst SVN2 presents a diffraction pattern that corresponds to the phases Nb2O5(JCPDS 30-0873), a-Sb2O4(JCPDS 11-0694), and VSbO4(JCPDS file 16-0600). After reaction, the diffraction peaks of VSbO4become more intense. Catalyst SVN3 exhibits a pattern quite similar to that of SVN2, it presents the pattern of Nb2O5(JCPDS file 7-0061), while after reaction it presents those of Nb2O5anda-Sb2O4. SVN1A is amorphous but it exhibits the pattern of SbNbO4 phase after reaction (Fig. 4) (JCPDS 16-0907). Fresh and used SVN1B catalyst present the pattern of SbNbO4anda-Sb2O4 phases, and the diffraction peaks of SbNbO4 become more intense for the used sample.

Fig. 5shows the Raman spectra of fresh and used dehydrated

Sb-V-Nb-O catalysts. The Raman spectra of fresh and used SVN2 and SVN3 are very similar with broad bands near 900, 700, and 200 cm 1, typical of Nb2O5 TT crystalline phase [52,53]. The Raman spectrum of the fresh SVN1A catalyst (Fig. 5) shows broad Raman bands that correspond to Nb2O5 phase but after reaction it exhibits Raman bands at 190, 255, 372, and 451 cm 1, typical ofa-Sb2O4phase[51]and features near 845, 686, and 621 cm 1that correspond to SbNbO4phase [52,53]. Spectra of fresh and used SVN1B catalyst show Raman bands that correspond toa-Sb2O4and SbNbO4phases. Mixed V-Nb-O phases, if present, could not be observed. These possess weak Raman bands[54].

Fig. 3. Yields (%) to different products at 5008C for catalysts SV1, SV3, SV5, and SV7. Reaction conditions: 20 h time-on-stream, total flow 20 ml min 1, feed composition (vol.%): C3H8/O2/NH3(9.8:25:8.6), 200 mg of catalysts.

Fig. 4. XRD patterns of fresh and used Sb-V-Nb-O catalysts.

Fig. 5. Raman spectra of fresh and used Sb-V-Nb-O catalysts.

Fig. 6. Yields (%) to different products at 5008C for catalysts SVN2, SVN1A, SVN1B, and SVN3. Reaction conditions: total flow 20 ml min 1_{, feed} com-position (vol.%): C3H8/O2/NH3(9.8:25:8.6), 200 mg of catalysts.

(5)

Fig. 6 represents the yields to the different products for catalysts SVN2, SVN1A, SVN1B, and SVN3, at 5008C. Catalyst SVN3 is much more active than SVN2, SVN1A, and SVN1B, but at 5008C it is very selective towards propylene while catalysts SVN2 is the most selective towards acrylonitrile.

4. Discussion

Sb-V-based oxide structures are directly related to propane ammoxidation reaction. In particular, VSbO4appears closely related to efficient catalysts[55]. In excess of Sb, the VSbO4/a -Sb2O4system is more selective to acrylonitrile formation. SV1 is active (Fig. 3), to the production of carbon oxides, propylene and acrylonitrile; the higher activity at lower Sb/V atomic ratios is probably related to the higher vanadium content [55–57]. Bulk Sb-V-O samples with Sb/V>1 are more selective to the formation of acrylonitrile[1,2,19,55]. Our results are consistent with those works: samples possessing VSbO4 and a-Sb2O4 phases, like SV3, SV5, and SV7, are more selective to acrylonitrile formation than the pure VSbO4phase, as can be observed inFig. 3. The excess of Sb renders the Sb-V-O system more selective; however, total conversion is much lower. Surface vanadium oxide species, on the other hand, appear critical to activate propane [49,58]. Surface vanadium oxide species also activate conversion of propane towards oxidative dehydrogenation reaction, even in the presence of Sb[58,59]. The effect of Sb/V atomic ratio is not clearly understood yet. Nilsson et al.[19]have found that the best rate for formation of acrylonitrile is achieved with catalysts with Sb/V = 2. Centi and Mazzoli[8]reported catalysts with Sb/V>1 as the most active on acrylonitrile formation and Chiang and Lee[60]concluded that the best activity results are achieved with catalysts with Sb/ V = 1. In general, mixeda-Sb2O4/VSbO4phases are selective to acrylonitrile formation; the results presented are in accordance with this point. However, the origin of these differences on the Sb/V values may be due to the fact that surface composition does not usually correspond to the bulk composition. Obviously, catalytic performance is directly affected by the surface composition of the catalysts, where the molecules react. There appears to be a delicate interaction between segregated Sb oxide, surface vanadium oxide species, and VSbO4 that renders the system efficient for propane ammoxidation [49]. It may be due to a cooperative effect between these phases, as it has been pointed out previously [19,26]. Operando Raman-GC studies during propane ammox-idation on supported Sb-V systems suggest that such cooperation may be due to a migration cycle of Sb5+species that facilitate the redox cycle of vanadium species [49,58], which involves surface V5+ species and lattice V3+ sites in

VSbO4[49,55].

The interplay between Sb and V for propane ammoxidation is strongly affected by the presence of a third element [1,5,9,13,26–36]. Niobia is particularly interesting since many possible interactions may take place, leading to V-Nb-O, V-Sb-O, and Sb-Nb-O phases[52,53], some of which are present in catalysts developed by Mitsubishi[17,18]and by bp[21]. The V-Nb-O phases are efficient for propane ammoxidation in

niobia-supported vanadia catalysts[52]and SbNbO4does not appear active for propane ammoxidation [53].

The bulk V-Sb-Nb-O catalysts have been prepared follow-ing different preparation methods to differently affect their structures.

SVN2 and SVN3 are prepared by interaction between vanadium acetyl acetonate and SbCl5with different niobium precursors. They possess similar structures; according to combined XRD and Raman spectroscopy: VSbO4and Nb2O5 phases are present, but not SbNbO4. The absence of the VSbO4 diffraction pattern for SVN3, visible to Raman spectroscopy, must be due to the small size of its domains. Some antimony remains segregated asa-Sb2O4; which becomes more evident in used samples. Thus, the trend of antimony to migrate to the catalysts surface observed with supported Sb-V-O catalysts [45,51], is also observed in the present work for bulk Sb-V-Nb-O catalysts. These catalysts perform propane ammoxidation better that the SV series and the other SVN catalysts. This is consistent with the beneficial effect of niobia on the performance reported for VSbO4[52]. Thus, the presence of both VSbO4 and Nb2O5 account for the better performance, compared to SV series.

SVN1A and SVN1B are prepared by interaction between NH4VO3and SbCl3and niobium oxalate. They possess similar structures, both possess SbNbO4 but not VSbO4 or Nb2O5, according to Raman spectroscopy (Fig. 5) and XRD patterns (Fig. 4). The formation of VSbO4is not observed for SVN1A or SVN1B (from SbCl3); in these catalysts; antimony reacts preferentially with niobium. SbNbO4phase is inert for propane ammoxidation[52,53]. The presence of SbNbO4and absence of VSbO4accounts for their poor ammoxidation performance. Apparently, niobia precursor is not critical for Sb-V-Nb interactions since niobia oxalate may result in either Nb2O5or SbNbO4. However, the oxidation state of the antimony chloride follows a trend: SbCl5 interacts preferentially with vanadia; SbCl3, with niobia. Using SbCl5, Sb, and V interact forming mixed phases like VSbO4 and Nb2O5 crystalline phase, regardless of niobium precursor (SVN2, SVN3). Such interac-tion is consistent with the redox reacinterac-tion between Sb5+and V4+to form VSbO4 [21,24]. In these catalysts, the VSbO4 phase becomes more evident after propane ammoxidation, so this phase is further formed during reaction. The in situ formation of this phase during propane ammoxidation has been reported for alumina-supported Sb-V-O catalysts using operando Raman-GC methodology[24].

The interaction between Sb, V, and Nb can be tuned by the oxidation state of antimony chloride. SbCl3promotes the Sb-Nb interaction, which leads to SbSb-NbO4. This phase is not active for propane ammoxidation; but, in addition, the interaction between Sb and Nb prevents Sb-Vor Nb-V interactions. VSbO4 and V-Nb-O phases are efficient for propane ammoxidation. Therefore, this system would not be efficient. On the other hand, SbCl5promotes the Sb-V interaction, forming VSbO4and segregating niobia as Nb2O5on calcination. Niobia-promoted VSbO4performs better than VSbO4[52]. The exact nature of these catalysts if not fully understood, the promotion could be due to the acidic sites of niobia and to the combination of V and

(6)

Nb into V-Nb-O phases, which are efficient for propane ammoxidation[52], along with VSbO4. Therefore, it is possible to prepare Nb-promoted Sb-V-O catalysts in a single step by appropriate control of the precursors.

SVN1 A, SVN1 B, and SVN2 use niobium oxalates, SVN3 uses Nb2O5suspension. All the catalysts prepared with SbCl3 possess SbNbO4. For a given niobium precursor (oxalate), SVN1A and SVN1B possess SbNbO4, but SVN2 does not. The interaction between Sb and V is stronger with SbCl5, while the interaction between Nb and Sb appears stronger with SbCl3.

5. Conclusions

Pure VSbO4 phase does not appear to be efficient for propane ammoxidation to acrylonitrile. It appears that some segregated Sb oxide is necessary for the catalytic cycle. The combination of VSbO4witha-Sb2O4enhances the selectivity to acrylonitrile, especially when Sb/V is near 5. The presence of surface vanadium species cannot be determined in detail, but it maybe important, since it would be at the surface, which is directly exposed to the reactants.

For V-Sb-Nb oxide catalysts, the niobium precursor does not appear to have a critical effect; however, the antimony precursor shows a trend:

(a) SbCl3precursor results in preferential antimony–niobium interaction; the formation of the SbNbO4phase is observed and the catalyst is neither active nor selective to acrylonitrile formation. Antimony interacts with niobium at the expense of the VSbO4anda-Sb2O4active phases. (b) SbCl5precursor favors the redox reaction with vanadium to

form the VSbO4 active phase. Nb improves the catalytic behavior with respect to the Sb-V-O mixed oxide catalysts. The presence of Nb-V-O phases could not be detected but its presence may not be ruled out either.

Acknowledgments

This research was partially funded by the Ministry of Education and Science, Spain (Project MAT-2002-04000-C02-01). M.O.G.-P. thanks the Ministry of Science and Technology of Spain, for a doctorate studies fellowship. M.V.M.-H. acknowledges theJuan de la Ciervaprogram of the Ministry of Science and Technology of Spain for financial support. The authors thank Dr. M.A. Vicente at the University of Salamanca for his help with the XRD analysis and his helpful discussions.

References

[1] R.K. Grasselli, in: G. Ertl, et al. (Eds.), Handbook in Catalysis, vol. V, Wiley-VCH, 1997, p. 2302.

[2] R.K. Grasselli, Top. Catal. 23 (2003) 5.

[3] M. Bowker, C.R. Bicknell, P. Kerwin, Appl. Catal. A 136 (2) (1996) 205. [4] V.D. Sokolovskii, A.A. Davydov, O.Y. Ovsitser, Catal. Rev. 37(3) (1995) 425. [5] R.K. Grasselli, Catal. Today 49 (1999) 141.

[6] R. Catani, G. Centi, F. Trifiro`, R.K. Grasselli, Ind. Eng. Chem. Res. 31 (1992) 107.

[7] G. Centi, F. Marchi, S. Perathoner, Appl. Catal. A 149 (1) (1997) 225. [8] G. Centi, P. Mazzoli, Catal. Today 28 (4) (1996) 351.

[9] A.T. Guttmann, R.K. Grasselli, J. Brazdil, F. James, 047949 [US4788317] (1988).

[10] L.C. Glaeser, J. Brazdil, F. James, 887478 [US4788173] (1988). [11] A.T. Guttmann, R.K. Grasselli, J. Brazdil, F. James, 724226 [US4746641]

(1988).

[12] H.W. Zanthoff, W. Gru¨nert, S. Buchholzs, M. Heber, L. Stievano, F.E. Wagner, G. Wolf, J. Mol. Catal. 162 (2000) 443.

[13] H.W. Zanthoff, S. Shafer, G.U. Wolf, Appl. Catal. A 164 (1997) 105. [14] Y.C. Kim, W. Ueda, Y. Morooka, Catal. Today 13 (4) (1992) 673. [15] S.I. Woo, J.S. Kim, Stud. Surf. Sci. Catal. 92 (1995) 191. [16] J.S. Kim, S.I. Woo, Appl. Catal. A 110 (2) (1994) 207.

[17] T. Ushikubo, K. Oshima, A. Kayou, M. Vaarkamp, M. Hatano, J. Catal. 169 (1) (1997) 394.

[18] T. Ushikubo, K. Oshima, K. Atsushi, M. Hatano, Stud. Surf. Sci. Catal. 112 (1997) 473.

[19] R. Nilsson, T. Lindblad, A. Andersson, J. Catal. 148 (2) (1994) 501. [20] G. Centi, P. Mazzoli, S. Perathoner, Appl. Catal. A 165 (1–2) (1997) 273. [21] J.F. Brazdil, M.A. Toft, J.P. Bartek, R.G. Teller, R.M. Cyngier, Chem.

Mater. 10 (1998) 4100.

[22] S.B. Derouane-Abd Hamid, G. Centi, P. Pal, E. Derouane, Top. Catal. 15 (2–4) (2001) 161.

[23] G. Centi, F. Marchi, Stud. Surf. Sci. Catal. 10 (1996) 277.

[24] M.O. Guerrero-Pe´rez, M.A. Ban˜ares, Chem. Commun. 12 (2002) 1292. [25] U.A. Schubert, F. Anderle, J. Spengler, J. Zuhlke, H.J. Eberle, R.K.

Grasselli, H. Kno¨zinger, Top. Catal. 15 (2001) 195.

[26] J. Nilsson, A.R. Landa-Ca´novas, S. Hansen, A. Anderson, J. Catal. 186 (1999) 442.

[27] T.J. Cassidy, M. Pollastri, F. Trifiro, J. Catal. 172 (1) (1997) 55. [28] S. Albonetti, G. Blanchard, P. Burattin, F. Cavani, F. Trifiro, [US5686381]

(1997).

[29] S. Albonetti, G. Blanchard, P. Burattin, F. Cavani, S. Masetti, F. Trifiro, Catal. Today 42 (3) (1998) 283.

[30] G. Blanchard, G. Ferre, [US5336804] (1994).

[31] A. Andersson, S.L.T. Andersson, G. Centi, R.K. Grasselli, M. Sanati, F. Trifio`, Stud. Surf. Sci. 75 (1993) 691.

[32] J. Nilsson, A.R. Landa-Canovas, S.A. Hansen, A. Andersson, J. Catal. 160 (2) (1996) 244.

[33] R. Catani, G. Centi, F. Trifiro, R.K. Grasselli, Ind. Eng. Chem. Res. 31 (1) (1992) 107.

[34] H. Roussel, B. Mehlomakulu, F. Belhadj, E. van Steen, J.M.M. Millet, J. Catal. 205 (2002) 97.

[35] A. Wickman, L.R. Wallenberg, A. Anderson, J. Catal. 194 (2000) 153. [36] G. Xiong, V.S. Sullivan, P.C. Stari, G.W. Zajac, S.S. Trail, J.A. Kaduk, J.T.

Golab, J.F. Brazdil, J. Catal. 230 (2005) 317.

[37] K. Tanabe, S. Okazaki, Appl. Catal. A: Gen. 133 (1995) 191. [38] I. Nowak, M. Ziolek, Chem. Rev. 99 (1999) 3603.

[39] M. Ziolek, Catal. Today 78 (2003) 47. [40] K. Tanabe, Catal. Today 78 (2003) 65. [41] T. Ushikubo, Catal. Today 78 (2003) 79.

[42] D.A.G. Aranda, A.L.D. Ramos, F.B. Passos, M. Schmal, Catal. Today 28 (1–2) (1996) 119.

[43] R.H.H. Smite, K. Seshan, H. Leemreize, J.R.H. Ross, Catal. Today 16 (1993) 513.

[44] R.H.H. Smits, K. Seshan, J.R.H. Ross, L.C.A. van der Oetelaar, H.J.H.M. Hlewegen, M.R. Anantharaman, H.H. Brongersma, J. Catal. 157 (1995) 584. [45] H.W. Zanthoff, S.A. Buchholz, Catal. Lett. 49 (3–4) (1997) 213. [46] S.A. Buchhotz, H.W. Zanthoff, ACS Symp. Ser. 638 (1996) 259. [47] S. Albonetti, G. Blanchard, P. Burattin, T.J. Cassidy, S. Masetti, F. Trifiro,

Catal. Lett. 45 (1–2) (1997) 119.

[48] M. Vaarkamp, T. Ushikubo, Appl. Catal. A: Gen. 174 (1–2) (1998) 99. [49] M.O. Guerrero-Pe´rez, M.A. Ban˜ares, Catal. Today 96 (2004) 265. [50] R. Burch, R. Swarnakar, Appl. Catal. A: Gen. 70 (1991) 129. [51] M.O. Guerrero-Pe´rez, J.L.G. Fierro, M.A. Vicente, M.A. Ban˜ares, J. Catal.

206 (2002) 339.

[52] M.O. Guerrero-Pe´rez, J.L.G. Fierro, M.A. Ban˜ares, Catal. Today 78 (2003) 387.

(7)

[53] M.O. Guerrero-Pe´rez, J.L.G. Fierro, M.A. Ban˜ares, Phys. Chem. Chem. Phys. 5 (2003) 4032.

[54] I.E. Wachs, Y. Chen, J.-M. Jehng, L.E. Briand, T. Tanaka, Catal. Today 78 (2003) 13.

[55] A. Anderson, S. Hansen, A. Wickman, Top. Catal. 15 (2–4) (2001) 103. [56] M.A. Ban˜ares, Catal. Today 51 (1999) 319.

[57] V.V. Guliants, Catal. Today 51 (1999) 255.

[58] M.A. Ban˜ares, M.O. Guerrero-Pe´rez, J.L.G. Fierro, G.G. Corte´z, J. Mater. Chem. 12 (11) (2002) 3337.

[59] V.Yu. Bychkov, M.Yu. Sinev, V.P. Vislovskii, Kinet. Catal. 42 (2001) 574.