Highly active and stable n-pentane isomerization catalysts without noble
metal containing: Al- or Ga-promoted tungstated zirconia
Xiao-Rong Chen,a,bYu-Qiao Du,aChang-Lin Chen,a,* Nan-Ping Xu,aand Chung-Yuan Moub
aCollege of Chemistry and Chemical Engineering, Nanjing University of Technology, Nanjing, 210009, China bDepartment of Chemistry, National Taiwan University, 1 Roosevelt Road, Section 4, Taipei, Taiwan
Received 14 June 2006; accepted 29 June 2006
This paper reports on the isomerization of n-pentane over Al- or Ga-promoted tungstated zirconia (WZ) in the presence of hydrogen. The catalytic activity was significantly improved with the addition of Al or Ga to WZ (AWZ or GWZ). It was found that both AWZ and GWZ catalysts have higher activities for the isomerization of n-pentane: conversion is over 70% and the selectivity to iso-pentane reaches92% at 215 °C. These catalysts exhibit excellent stability and the deactivation is undetected for 1000 h operation. The promoted-WZ catalyst is a mixed oxide solid acid catalyst without noble metal. Furthermore, the alkanes isomerization catalyst is halogen-free, which is environmentally friendly. The promoted-WZ was characterized by Fourier-transformed infrared spectroscopy (FT-IR), NH3adsorption microcalorimetry and X-ray photoelectron spectroscopy (XPS). The
remarkable activity and selectivity for n-pentane isomerization are due to enhanced strong acid sites and redox properties in the promoted-WZ.
KEY WORDS: tungstated zirconia; aluminum; gallium; n-pentane isomerization.
1. Introduction
Isomerization of light n-alkanes (n-pentane and
n-hexane) is important for the production of clean and high-octane number fuels. The major commercial cata-lysts for light n-alkanes isomerization are Pt on chlori-nated alumina or Pt/H-mordenite [1,2]. Pt on chlorinated alumina has high catalytic activity at low temperature (115–150°C) where the production of branched isomers is favored in the equilibrium of pro-duction distribution [3]. However, this catalyst suffers from chlorine loss during the isomerization process and requires constant addition of chlorine-containing compounds. This chlorine-containing catalyst is also subjected to stringent environmental control. While Pt/H-mordenite does not have these disadvantages, but it requires higher reaction temperature (260 °C) which is thermodynamically unfavorable for the formation of branched isomers.
Extensive researches have been devoted to the search for an environmentally friendly catalyst that can operate at low temperature. It has been known that modified zirconia catalysts offer as a replacement for halogen-containing catalysts since they exhibit good potential for
n-alkanes isomerization. Sulfated zirconias (SZ) have been found to be a strong solid acid for C4/CS/C6
isomerization at low temperature [4–7]. SZ system, however, has the disadvantages of deactivation and sulfur loss during reaction and regeneration [8]. Tung-stated zirconia (WZ) catalysts, first reported by Hino and Arata [9] for n-butane isomerization, are sought as a nice alternative to SZ system and have been studied extensively [10–12]. WZ catalysts appear to be more suitable for practical application because of their supe-rior stability under reducing and oxidizing conditions and thus regenerability. Whereas, the non-promoted WZ is less active, its catalytic properties can be greatly improved by promotion with noble metals (Pt or Pd) and some metal oxides [13–15]. A12O3-doped WZ
(AWZ) catalyst has been claimed as an efficient catalyst for the skeletal isomerization of n-butane [16]. Previ-ously, we have reported Ga-promoted WZ (GWZ) greatly improved the catalytic properties for n-butane isomerization [17]. At present, there has been no report yet on the n-pentane isomerization over Al- or Ga-promoted WZ catalysts. Compared to the major commercial catalysts, Al- or Ga-promoted WZ catalysts are composed of mixed oxides without noble metal and are halogen-free catalysts. So the cost of Al- or Ga-promoted WZ catalysts is low and isomerization of n-pentane over this catalyst is a green process. In this paper, we report the n-pentane isomerization over Al- or Ga-promoted WZ catalysts. The stability of AWZ and GWZ catalysts is also investigated for extended time-on-stream, up to 1000 h test.
*To whom correspondence should be addressed. E-mail: [email protected]
DOI: 10.1007/s10562-006-0146-3
2. Experimental
2.1. Catalyst preparation
AWZ and GWZ catalysts synthesis procedures were
described in our previous report [18]. Zr(OH)4 was
prepared from zirconia nitrate solution by adding drop-wise ammonium hydroxide solution up to pH 9–10 and then refluxed for 24 h. The precipitated hydrogel was filtered and washed repeatedly until the filtered solution is neutral. The gel was dried and impregnated with aqueous ammonium tungstate (Acoros). In the synthesis of Al or Ga-promoted WZ catalyst, appropriate amount of Al(NO)3or Ga(NO)3 was added to the WZ slurry.
The resultant suspension was refluxed, dried and cal-cined at 800°C. The W loading was 15 wt%, unless otherwise noted. Al content for all AWZ catalysts was 0.5 wt% and Ga content in GWZ was 1.0 wt%.
2.2. Catalyst characterization
Pyridine-adsorbed Fourier-transformed infrared (FT-IR) spectra were conducted on a Nicolet 550
Spec-trometer instrument. Samples were treated at 400°C for
1 h under a vacuum of 103Pa and introduced pyridine
at room temperature. The system was then evacuated and the FT-IR spectra were recorded at 300°C. Microcalorimetric studies of the adsorption of NH3
were performed on a heat-flow microcalorimeter of the Tian–Valven type. A known amount of the probe molecular (1–10 lmol) was exposed stepwise to satu-rated adsorption at 150°C. X-ray photoelectron spec-troscopic analyses of samples was performed on VG Scientific SCALAB 250 fitted with a monochromatic AlKaradiation X-ray resource, under a residual pressure
of 10)9–10)10 Torr.
2.3. Catalytic test
The isomerization ofn-pentane was carried out in a
fixed-bed flow reactor. Two gram of 40-mesh catalyst was charged into the reactor and activated at 450°C under flowing dry air for 3 h. After catalyst pretreat-ment, the reactor was cooled to reaction temperature, and then pressurized with H2at 2.0 MPa or other
set-ting pressure. n-Pentane was fed into the reactor and hydrogen and n-pentane flows was adjusted to give a 3 H2/n-C5 molar ratio at n-pentane weight hour space
velocity (WHSV) of 1.0 h)1. An on-line gas chromato-graph equipped with FID was used to analyze the reaction products.
3. Results and discussion 3.1. Catalyst characterization 3.1.1. FT-IR spectra
The nature of acid sites in the catalyst was determined
by pyridine-adsorbed FT-IR. Figure1 compares the
FT-IR spectra of WZ, AWZ and GWZ after pyridine adsorption at 300 °C. Bro¨nsted and Lewis acid sites were found on all samples. For WZ and GWZ, there is no obvious difference in the intensity of band at 1540 cm)1 (Bro¨nsted acid sites) or 1450 cm)1 (Lewis acid sites). In the case of AWZ, the intensity of band of Lewis acid sites is enhanced. These spectra indicate that the addition of Al to WZ caused an increase of Lewis acid sites.
3.1.2. NH3adsorption microcalorimetry
NH3 adsorption microcalorimetry was used to
investigate the effect of Al or Ga on the surface acidities of WZ. The adsorption microcalorimetry results are summarized in table 1. The total number of acid sits for WZ and AWZ is about the same, but adding Al to WZ increases the initial heat of adsorption from 199 to 237 kJ mol)1. The numbers of weak acid sites (different heat <100 kJ mol)1) and moderate acid sites (different heat between 100 and 170 kJ mol)1) show no obvious difference between WZ and AWZ. Although there are few strong acid sites (different heat >170 kJ mol)1) for WZ and AWZ, the number of strong acid sites is increased from 9.1 to 14.9 lmol g)l after adding Al to WZ. For Ga-promoted WZ, both the total number of acid sits and the initial heat of adsorption are increased. Adding Ga to WZ also appears to generate more strong acid sites than those on AWZ and WZ. It was reported that the catalytic activity of zirconia-based catalysts for alkane isomerization is related to the strength of the acid sites. Hua and Sommer [19] suggested the strong acid sites are active sites for alkane isomerization. The calorimetric results suggest a correlation of the activity difference between promoted and non-prompted WZ to the number of strong acid sites. The addition of small amounts of Al or Ga to WZ increases the acid site Figure1. Pyridine adsorption FT-IR spectra of samples: (1) WZ, (2)
strength and thus produces a more active alkanes isomerization catalyst.
3.1.3. X-ray photoelectron spectroscopy
The W 4fXPS spectra of WZ, AWZ and GWZ are
shown in figure 2 and the fitting results of the XPS are compiled in table 2. The W 4f XPS spectra of WZ can be deconvoluted into three doublet at 35.6, 34.5 and 33.5 eV, assigned to W6+, W5+and W4+ [20]. The
relative concentrations of W6+, W5+ and W4+in WZ are 28, 42 and 30%, respectively. Compared to WZ, the W 4f spectra of AWZ and GWZ could not be fitted to reveal the presence of W4+, which resulted in the rela-tive concentration of W6+ and W5+increased. The W 4f XPS measurement showed that the tungsten species are more reducible in AWZ and GWZ than in WZ because of the enrichment of w6+. We have also found that Al or Ga-doped WZ is more easily reduced than WZ as using H2-TPR technique [18]. The enrichment of
W6+density can also influence the electronic properties of WOx. It involves the dissociation of H2 and the
migration of H atom to WOx, domain, which stabilize
protons (Hd+) to form Bro¨nsted acid sites by delocal-izing the compensating electron density among the W6+ Lewis acid centers [21].
The surface W/Zr ratios of samples are also listed in table 2. The surface W/Zr is calculated from the ratio of the integrated area of W 4f XPS peaks to the integrated area of Zr 3d XPS peaks (not shown) and consideration of the atomic sensitivity factor of W and Zr. The surface W/Zr ratio in WZ is 0.23. Conversely, the surface W/Zr ratios in AWZ and GWZ decrease to 0.086 and 0.084. The results showed that Al and Ga facilitate the dis-persion of WOxon the surface of zirconia.
3.2. Catalytic reaction performance
3.2.1. Catalytic activity and product distribution
Isomerization reaction of n-pentane was carried out
at 215 °C in the presence of H2. Figure 3 shows the
conversion with time on stream over the promoted and non-promoted catalysts. Catalyst deactivation was not observed during the test period of 6 h. The promoted catalysts exhibit higher n-pentane conversion than WZ. Under the identical reaction condition, the catalytic Table 1
NH3adsorption microcalorimetry results of distribution of acid site strength
Catalyst Initial heat/kJ mol)1 Acid sites/lmol g)1
Total <100 kJ mol)1 100–170 kJ mol)1 >170 kJ mol)1
WZ 199 252.7 205.1 38.5 9.1
AWZ 237 259.3 205.2 39.2 14.9
GWZ 214 275.6 210.7 47.3 17.6
Figure2. W 4f XPS spectra of (1) WZ, (2) AWZ and (3) GWZ.
Table 2 W 4f XPS fitting results
Sample BE (eV) Assignment W (%) W/Zr
WZ 35.8 W6+ 28 0.23 34.8 W5+ 42 33.4 W4+ 30 AWZ 35.7 W6+ 53 0.086 34.7 W5+ 47 GWZ 35.5 W6+ 61 0.084 34.5 W5+ 39
activity of WZ is low; n-pentane conversion is only 41%. The catalytic conversions of pentane for AWZ and GWZ are greatly improved to 71 and 72%, respectively. The product distributions of n-pentane isomerization over WZ, AWZ and GWZ catalysts are compared in table 3. In contrast to WZ, AWZ and GWZ catalysts not only display much higher catalytic activity, but also favor the formation of pentane. The yield of iso-pentane over AWZ and GWZ reaches 65 and 66%, much more than that over WZ. The higher catalytic activity of AWZ and GWZ is also reflected in the for-mation of a little higher amount of cracking products, C4). The cracking products are negligible over WZ
because of its lower catalytic activity.
For comparison with the major commercial catalysts forn-pentane isomerization, the results of Pt–C1/A12O3
and Pt/HM are also listed in table 3. Pt/HM shows the highest reaction temperature of 250°C for n-pentane isomerization where thermodynamic constrains give the lowest yield of iso-pentane. AWZ and GWZ show the lower reaction temperature of 215°C and improved n-pentane conversion and the yield of iso-pentane. Although the reaction temperature over AWZ and GWZ is higher than Pt–Cl/A12O3, AWZ and GWZ
have n-pentane isomerization activities approaching that of Pt–Cl/A12O3. However, chloride is corrosive to the
facility and harmful to the environment. AWZ and GWZ catalysts are novel halogen-free and Pt-free solid acid catalysts that give a green process for n-pentane isomerization. Both Pt–Cl/A12O3 and Pt/HM require
noble metal Pt to maintain the catalytic activities and stabilities. AWZ and GWZ, composed of mixed metal oxides without using noble metal, may be cheaper and sulfur-tolerant.
3.2.2. Effect of reaction pressure on the catalytic activity The reaction pressure has a great effect on the
cata-lytic activities of AWZ and GWZ for n-pentane
isom-erization. Figure 4 shows n-pentane conversion greatly depends on the reaction pressure. For both AWZ and GWZ, n-pentane conversion and the cracking products, C4), show the similar trend, they pass through a
maxi-mum at a reaction pressure of approximate 2.0 MPa and slightly decreased as the reaction pressure was fur-ther increased to 2.5 MPa. Because of the highest cracking products at the reaction pressure of 2.0 MPa, the selectivity to iso-pentane is 92%, slightly lower than those at other setting pressure. Pentane isomeri-zation is an equimolecular reaction. The change of reaction pressure has no effect on pentane isomerization theoretically. Actually, the side reaction, such as the cracking of pentane or its oligomer, can increase the number of molecules. It is suggested that increasing reaction pressure can suppress the side hydrogenolysis reaction. However, Kuba et al. [22] suggested the reac-tion pathway in n-pentane isomerizareac-tion over WZ catalysts by two modes: monomolecular and bimolecu-lar. For both modes, iso-pentane is formed from isom-erization intermediate and isomisom-erization intermediate is originally derived from the dehydrogenation of n-pen-tane. Too high the reaction pressure is a disadvantage in the dehydrogenation of n-pentane. It seems 2.0 MPa is the optimum reaction pressure.
3.2.3. Effect of tungsten content on the catalytic activity
Figure5 shows the effect of W content on the
cata-lytic activity for n-pentane isomerization over AWZ. At the same calcination temperature of 850 °C, the activity of AWZ catalyst depends strongly on W content. AWZ Figure3. Catalytic activity of WZ, AWZ and GWZ catalysts for
n-pentane isomerization as a function of time. P = 2 MPa, T = 215°C, H2/n-C5= 3 and WHSV = 1 h)1.
Table 3
Comparison of the catalytic performance of WZ, AWZ, GWZ and some commercial catalysts
Catalyst WZ AWZ GWZ Pt–Cl/Al2O3 Pt/HM
T/°C 215 215 215 <160 250 P/MPa 2 2 2 3 3 Space velocity 1 g g)1h)1 1 g g)1h)1 1 g g)1h)1 2 v v)1h)1 1 g g)1h)1 H2/n-C5H12 3 3 3 – 2.5 n-C5H12conversion (wt%) 41 71 72 77 a 63 Total cracking C4)(wt%) 0.5 2.6 3.0 – 0.5 i-C5H12(wt%) 40 65 66 76a 61 Selectivity to i-C5H12(wt%) 98 92 92 90* 97
Reference This study This study This study 1 2
catalyst is almost inactive below 5 wt% W content; n-pentane conversion is only10%. Increasing W content leads to a large improvement in catalytic activity. The maximum conversion is found at 10–15 wt% W loading. Further increasing W content, the catalytic activity of AWZ shows a small decline. It was reported that the maximum catalytic activity of WZ for o-xylene isomer-ization occurred at the WOxsurface density higher than
the theoretical monolayer capacity of WZ catalysts [23]. Baertsch [24] suggested the acidity and number of Bro¨nsted acid sites presented on WZ increased with the loading of WOxto monolayer coverage.
3.2.4. Effect of calcination temperature on the catalytic activity
The effect of calcination temperature of AWZ on the catalytic activity for n-pentane conversion is shown in
figure6. AWZ catalysts with 10 and 15 wt% W
con-tent show the higher catalytic acidity from figure 5. So,
two samples of AWZ with 10 and 15 wt% W content respectively were used to investigate the effect of cal-cination temperature. From figure 6, the optimum calcination temperature is 850°C for AWZ of 10 wt% W content; and the optimum calcination temperature is 800 °C for AWZ of 15 wt% W content. At the calci-nation temperature of 850°C, the catalytic activity of AWZ of 10 wt% W content is slightly higher than that of AWZ of 15 wt% W content. As the calcination temperature decreased to 800°C, the catalytic activity of AWZ of 10 wt% W content sharply decreases to a much lower level than that of AWZ of 15 wt% W content. The results are qualitatively in accord with the conclusions that increasing WOx surface concentration
resulted in the lowering of optimum calcination temperature of WZ catalysts for o-xylene isomerization [24].
Figure4. Effect of reaction pressure on catalytic performance of (a) AWZ and (b) GWZ. Data obtained after 6 h on feed. T = 215°C, H2
/n-C5= 3 and WHSV = 1 h)1.
Figure5. Effect of W content on catalytic performance of AWZ. Data obtained after 6 h on feed. P = 2 MPa, T = 215°C, H2/n-C5= 3
and WHSV = 1 h)1.
Figure6. Effect of calcination temperture on catalytic performance of AWZ. Data obtained after 6 h on feed. P = 2 MPa, T = 215°C, H2/
n-C5= 3 and WHSV = 1 h -l
3.2.5. Stability of the catalysis
In order to evaluate the application potential of AWZ and GWZ catalysts, it is necessary to investigate their stability. The result of stability test of GWZ for 1000 h
is shown in figure7. No catalyst deactivation was
observed during the entire test period of 1000 h. n-Pentane conversion is stable at71%, the selectivity to iso-pentane reaches 92% and the cracking products, C4), is less than 3%. Figure 8 gives the stability of AWZ
for a 200 h test. Deactivation was not observed during the test period for AWZ catalyst. n-Pentane conversion is stable at72%, the selectivity to iso-pentane reaches 93% and the cracking products, C4), is less than 3%.
3.2.6. The nature of promotional effect
The promotion of catalytic activities of Al and Ga are through two different effects: (a) the formation of stronger acid sites (b) the increased redox activity of W. Both are important for the catalytic isomerization of alkanes. The pentane molecules are activated by
dehy-drogenation on well-dispersed polyoxotungstate via a
redox process. Then the resulting alkenes are protonated on the acidic sites. A stable and active catalyst for alkane isomerization requires the balanced action of the above two functions. It seems Al and Ga play
promo-tional roles in both the redox and acidic properties. The strengths of the acid sites in AWZ and GWZ are found to be stronger than those originally present on WZ. Al and Ga favor the dispersion of WOx on the surface of
zirconia. In our previous work [18], the crystal WO3was
observed on WZ and disappeared on AWZ and GWZ by UV–visible diffuse reflectance spectra. It is known that the crystal WO3is an inactive phase. The
well-dis-persed WOx can affect the acid sites distribution and
cause more active sites in catalysis.
Addition of Al or Ga to WZ also improves the redox
properties of W6+. The redox properties of W6+
in WOx/ZrO2play a crucial role as a redox initiator in
activation of alkanes. Kuba et al. [25] suggested that in tungstated zirconia alkane activation proceeds via a hemolytic C–H bond breaking by electron transfer to W6+ forming W5+. Occhiuzzi et al. [26] further found that in a reducing environment W5+ forms on small WOxclusters or polyoxotungstates. In. W03,the inactive
W4+ would form. Our investigation by XPS technique found that the Al or Ga addition do help the dispersion of tungstated on zirconia and lead to the more active W5+species. The improved redox pair of W6+/W5+in promoted WZ is accompanied by the easier activation of alkanes. With both improved initiation of alkane dehydrogenation and acidity in forming carbenium ion, the isomerization activities in promoted WZ are thus increased and more importantly well-balanced.
4. Conclusion
In this paper, we have shown that Al- and
Ga-pro-moted WZ catalyze isomerization of n-pentane with
high catalytic activity and selectivity. Their high cata-lytic activities remain very stable over a long period. For both AWZ and GWZ, the catalytic activity is greatly dependent on W content, calcination temperature and reaction pressure. Compared to the commercial cata-lysts, Al- or Ga-promoted WZ catalysts show n-pentane isomerization activities approaching that of Pt on chlorinated aluminum catalyst, and much higher than that of Pt/H-mordenite. AWZ and GWZ catalysts are a new kind of halogen-free solid acid catalysts that are friendly to the environment. Another notable charac-teristic of AWZ and GWZ catalysts is that they are composed of mixed oxides without noble metal.
Acknowledgments
We are grateful to Professor Jianyi Shen from Nanjing University for help in the microcalorimetric study, Dr. J.Hh Wang of NTU for helps in XPS mea-surements. We also acknowledge the support of the National Science Foundation of China (NSFC) (Project 20476047), the SINOPEC (Project 104010), the Science and Technology Department of Jiangsu Province Figure7. Thousand hours stability test of n-pentane isomerization
over GWZ. P = 2 MPa, T = 220°C, H2/n-C5= 4 and
WHSV = 1 h)1.
Figure8. Two hundred hours stability test of n-pentane isomerization over AWZ. P = 2 MPa, T = 215°C, H2/n-C5= 3 and
(Project BG2002017) and the Key Laboratory of Material-Oriented Chemical Engineering of Jiangsu Province.
Reference
[1] B.W. Burbidge and J.K.R. Rolfe, Hydrocarbon Process, Petrol. Refiner 45 (1966) 168.
[2] H.W. Kouwenhoven and W.C. van Zijll Langhout, Chem. Eng. Progr. 67 (1971) 65.
[3] A. Chica and A. Corma, J. Catal. 187 (1999) 167.
[4] M. Hino, S. Kobayashi and K. Arata, J. Am. Chem. Soc. 101 (1979) 6439.
[5] W. Wang, C.L. Chen, N .P. Xu and C.Y. Mou, Green Chem. 4 (2002) 257.
[6] W. Wang, J.H. Wang, C.L. Chen, N.P. Xu and C.Y. Mou, Catal. Today 97 (2004) 307.
[7] C.J. Cao, S. Han, C.L. Chen, N.P. Xu and C.Y. Mou, Catal. Commun. 4 (2003) 511.
[8] G. Larsen, E. Lotero, R.D. Parra, L.M. Petkovic, H.S. Silva and S. Radhavan, Appl. Catal. A 130 (1995) 213.
[9] K. Arata and M. Hino, J. Chem. Soc. Chem. Commun. 12 (1988) 1259.
[10] J.C. Yori, C. Rvera and J.M. Pareta, Appl. Catal. A 163 (1997) 165.
[11] M. Scheithauer, T.K. Cheung, R.E. Jentoft, R.K. Grasselli, B.C. Gates and H. Kno¨zinger, J. Catal. 180 (1998) 1.
[12] T.N. Vu, J. van Gestel, J.P. Gilson, C. Cpllet, J.P. Dath and J.C. Duchet, J. Catal. 231 (2005) 468.
[13] P. Lukinskasa, S. Kuba, B. Spliethoff, B. Spliethoff, Grasselli, R.K. , B. Tesche and H. Kno¨zinger, Topics Catal. 23 (2003) 163. [14] S. Kuba, B.C. Gates, R.K. Grasselli and H. Kno¨zinger, Chem.
Commun. 4 (2001) 321.
[15] W.M. Hua and J. Sommer, Appl. Catal. A 232 (2002) 129. [16] S.T. Wong, T. Li, S. Cheng, J.F. Lee and C.Y. Mou, J. Catal. 215
(2003) 45.
[17] X.R. Chen, C.L. Chen, N.P. Xu and C.Y. Mou, Catal. Lett. 85 (2003) 177.
[18] X.R. Chen, C.L. Chen, N.P. Xu and C.Y. Mou, Catal. Today 93– 95 (2004) 129.
[19] W.M. Hua and J. Sommer, Appl. Catal. A 232 (2002) 129. [20] S.T. Wong, C.C. Hwang and C.Y. Mou, Appl. Catal. B 63 (2006)
1.
[21] D.G. Barton, M. Shtein, R.D. Wilson, S.L. Soled and E. Iglesia, J. Phys. Chem. B 103 (1999) 630.
[22] S. Kuba, P. Lukinskas, R. Ahmad, F.C. Jentoft, R.K. Grasselli, B.C. Gates and H. Kno¨zinger, J. Catal. 219 (2003) 376. [23] D.G. Barton, S.L. Soled, G.D. Meitzner, G.A. Fuentes and E.
Iglesia, J. Catal. 181 (1999) 57.
[24] C.D. Baertsch, S.L. Soled and E. Iglesia, J. Phys. Chem. B 105 (2001) 1320.
[25] S. Kuba, P. Lukinskas, R.K. Grasselli, B.C. Gates and H. Kno¨zinger, J. Catal. 216 (2003) 353.
[26] M. Occhiuzzi, D. Cordischi, D. Gazzoli, M. Valigi and P.C. Heydorn, Appl. Catal. A 269 (2004) 169.
