Effect of Ag promoter on redox properties and catalytic performance of Ag-Mo-P-O catalysts for oxidative dehydrogenation of propane

Effect of Ag promoter on redox properties and catalytic performance of Ag-Mo-P-O catalysts for oxidative dehydrogenation of propane

Applied Surface Science 220 (2003) 117–124 Effect of Ag promoter on redox properties and catalytic performance of Ag-Mo-P-O catalysts for oxidative d...

157KB Sizes 1 Downloads 4 Views

Applied Surface Science 220 (2003) 117–124

Effect of Ag promoter on redox properties and catalytic performance of Ag-Mo-P-O catalysts for oxidative dehydrogenation of propane Xin Zhanga,*, Hui-lin Wanb, Wei-zheng Wengb, Xiao-dong Yib a

State Key Laboratory of C1 Chemistry and Technology, Department of Chemistry, Tsinghua University, Beijing 100084, PR China b State Key Laboratory of Physical Chemistry for Solid Surfaces, Department of Chemistry, Xiamen University, Xiamen 361005, PR China Received 18 March 2002; accepted 26 May 2003

Abstract The influence of addition of sliver ions on the redox properties and catalytic performance of molybdenum-phosphate catalyst in oxidative dehydrogenation (ODH) of propane has been investigated. MoO3 and AgMoO2PO4 have been detected in Ag-doped Mo-P-O catalysts. The redox properties have been characterized by H2-temperature-programmed reduction (TPR), electronic paramagnetic resonance (EPR) and X-ray photoelectron spectroscopy (XPS). The catalytic performance of MoO3, Ag2O, AgMoO2PO4 and Ag-doped Mo-P-O catalysts has also been examined. The incorporation of Ag improved the number of Mo5þ species and the reducibility of the catalysts. The Ag-doped catalysts favor activity and selectivity in ODH of propane higher than the undoped catalyst. The best catalyst was obtained with Ag/Mo ratio of 0.3 and MoO3/AgMoO2PO4 ratio of 1.1. The influence of promoter Agþ on redox properties and catalytic performance is due to forming redox couple ‘‘Ag0 þ Mo6þ , Agþ þ Mo5þ ’’ and the synergetic effect originating from ‘‘coherent interface’’ between MoO3 and AgMoO2PO4. # 2003 Elsevier B.V. All rights reserved. Keywords: Oxidative dehydrogenation of propane; Redox properties; Promoter Ag; Molybdenum-phosphate catalyst

1. Introduction The oxidative dehydrogenation (ODH) of alkane to corresponding alkene has recently been of great interest in catalysis chemistry [1–3]. In the case of ODH of propane, a variety of catalysts have been developed so far. The catalysts based on vanadium oxides and molybdenum oxides are considered to be effective in ODH of propane [4–14].

* Corresponding author. Fax: þ86-01-62792122. E-mail address: [email protected] (X. Zhang).

The ODH of propane usually takes place Mars van Krevelen mechanism involving in the oxidation state vary of cation and the transfer of oxygen species. The redox properties have obviously important effect on the activity and selectivity in ODH of propane [15,16]. The metal molybdate AMoO4 (A: Ni, Co, Mg, Mn, Zn) and Ni0.45Co0.45X0.06MoO4 (X: P, Bi, Fe, Cr, V, Ce) were investigated for ODH of propane. Addition of redox elements in the catalysts enhanced the propane conversion [17]. Valenzuela et al. [18] studied the effect of different promoters on the performance of VMgO in ODH of propane. They considered the easier the reducibility of V ions, the higher the selectivity to

0169-4332/$ – see front matter # 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0169-4332(03)00837-7

118

X. Zhang et al. / Applied Surface Science 220 (2003) 117–124

propene. Abello et al. [19] reported that the Mo6þ reducibility of MgO/A12O3-supported molybdenum catalyst paralleled to catalyst activity. Chen et al. [20] reported the effect of alkali (Cs, K, Li) oxides on the structure and on the catalytic behavior of MoOx/ZrO2 catalyst in ODH of propane. The presence of alkali inhibits the initial reduction of MoOx and increasing reduction activation energies. The turnover rate of ODH of propane decreased monotonically with increasing A/Mo atomic ratio and with increasing basicity of the alkali oxide. In present work, a detail study on the effect of promoter Ag on catalytic performance and redox properties of Mo-P-O catalyst has been performed. In this way, Ag-doped and undoped catalysts have been prepared and characterized by X-ray diffraction (XRD), H2-temperature-programmed reduction (TPR), electronic paramagnetic resonance (EPR), X-ray photoelectron spectroscopy (XPS), and their catalytic performance has been investigated. Difference in redox properties and in catalytic performance is discussed.

2. Experimental 2.1. Catalyst preparation Ag-Mo-P-O catalysts with different Ag/Mo/P mole ratio were prepared by the method of grinding (NH4)H2PO4, MoO3 and Ag2O powder with certain amount of deionized water. The sample was dried overnight in air at 383 K. After dry, the sample was calcinated at 623 K for 5 h and then at 823 K for 12 h. AgMoO2PO4 was prepared as described in [21]. 2.2. Characterization Powder X-ray diffraction pattern was obtained by using Rogaku Rotflex D/Max-C diffractometer with Cu Ka (l ¼ 0:15406 nm) radiation. Temperature-programmed reduction experiments were carried out using TPR-mass spectrometer (MS) apparatus. The exit gas was analyzed on-line by quadrupole mass spectrometer (Omin-star GSD3000). Each sample of ca. 100 mg was pretreated in a 20 ml/mm flow of O2 at 823 K for 30 min and then cooled at room temperature (r.t.). After that, He was

switched to purge oxygen in apparatus at r.t. Subsequently, the sample was exposed to a 20 ml/min 5%H2/N2 flow, and warmed at rate of 10 K/mm. X-ray photoelectron spectroscopy measurements were carried out using VG ESCLAB MK-II spectrometer Al Ka (1486.6 eV, 10.1 kV). The spectra were referenced with respect to the C 1s line at 284.7 eV. The electronic paramagnetic resonance spectra were obtained by Brucker 200D-SCR spectrometer at r.t. The klystron frequency was 9.6 GHz and magnetic field modulation was 100 KHz. Each sample was ca. 200 mg. 2.3. Catalytic testing Catalytic test was carried out at atmospheric pressure in a continuous flow system with a fixed bed quartz tubular reactor (i:d: ¼ 6 mm). To minimize possible homogeneous reactions, the reactor volume was reduced by filling the reactor up and down the catalyst bed with quartz wool. The feedstock and products were analyzed with two on-line gas chromatograph operating three columns, carbon molecular sieve column and Al2O3 column impregnating squalane (102-GC, TCD) for the separation of C3H8, C3H6, C2H4, C2H6, CO, CO2, GDX-103 (103-GC FID) for the separation of acrolein, acetone and propanal, etc. The exit gases were heated to 393 K to prevent condensation. The catalysts were allowed to equilibrate under the reaction conditions for at least 30 min.

3. Results 3.1. Characterization 3.1.1. XRD characterization The XRD patterns of Ag-Mo-P-O catalysts and AgMoO2PO4 are shown in Fig. 1. The MoO3 and AgMoO2PO4 are main phases in Ag-Mo-P-O catalysts. The ratio of IMoO3 ð0 2 1Þ =IAgMoO2 PO4 ð2 3 0Þ is used to indicate the relative amount of the two phases. Note that the amount of AgMoO2PO4 increases with the Ag content increasing (Table 1). 3.1.2. TPR characterization TPR profiles of the catalysts are present in Fig. 2. Mo0.5P0.3Ox had two reduction peaks at ca. 865 and

X. Zhang et al. / Applied Surface Science 220 (2003) 117–124

119

Fig. 1. (A) XRD pattern of Ag-Mo-P-O catalysts; (B) XRD pattern of AgMoO2PO4.

Table 1 Properties of Ag-Mo-P-O catalysts Catalyst

IMoO3 ð0 2 1Þ = IAgMoO2 PO4 ð2 (XRD)

Mo0.5P0.3Ox AgMoO2PO4 Ag0.1Mo0.5P0.3Oy 0.8 Ag0.3Mo0.5P0.3Oy 1.1 Ag0.5Mo0.5P0.3Oy 1.3

3 0Þ

Relative H2 consumption (H2-TPR) (a.u.)

Relative concentration of Mo5þ (EPR) (a.u.)

40.2 228.6

27.2 31.5

291.3 392.0

47.3 55.5

Fig. 2. TPR profile of Mo0.5P0.3Ox, AgMoO2PO4 and Ag-Mo-P-O catalyst.

120

X. Zhang et al. / Applied Surface Science 220 (2003) 117–124

988 K; AgMoO2PO4 also exhibited two peaks at 819 and 975 K which are correspond to the step-wise reduction process of Mo6þ ) Mo4þ and Mo4þ ) Mo0 [22,23]. The TPR profiles of Ag-Mo-P-O catalysts are different. In these cases, three reduction peaks were observed. The reduction peaks near 771–763 K may be due to Mo6þ ) Mo5þ reduction, and the reduction peaks near 897–845 K can attribute to Mo5þ ) Mo4þ reduction, and the reduction peaks near 975–915 K are Mo4þ ) Mo0 reduction. The H2 con-

sumption is in the order Mo0:5 P0:3 Ox < AgMoO2 PO4 < Ag0:3 Mo0:5 P0:3 Oy < Ag0:5 Mo0:5 P0:3 Oy catalyst (Table 1).

Fig. 3. (A) EPR spectra of the catalysts; (B) EPR spectra of Ag0.3Mo0.5P0.3Oy catalysts treated under different conditions: (a) the used catalyst; (b) the fresh catalyst; (c) the catalyst treated in O2 flow at 773 K for 15 min and then sealed in glass tube.

Fig. 4. (A) XPS spectra of Mo3d on MoO3 and the catalysts (a): Mo0.5P0.3Ox, (b): AgMoO2PO4, (c): Ag0.1Mo0.5P0.3Oy, (d): Ag0.3Mo0.5P0.3Oy, (e): Ag0.5Mo0.5P0.3Oy, (f): MoO3. (B) XPS spectra of Ag 3d on the Ag0.3Mo0.5P0.3Oy and Ag0.3Mo0.5P0.3Oy catalysts.

3.1.3. EPR characterization The presence of Mo5þ and change in the amount of the catalysts were investigated by EPR. The spectra of the catalysts exhibited signal with g1 ¼ 1:94 and g2 ¼ 1:89, which were attributed to Mo5þ [24] (Fig. 3). Any hyperfine structure due to Mo5þ

X. Zhang et al. / Applied Surface Science 220 (2003) 117–124

(S ¼ 1=2, I ¼ 5=2) are not observed, indicating that Mo5þ in the catalyst are not enough apart from one another and the strong interface among Mo5þ has eliminated the hyperfine structure, that is, the concentration of Mo5þ are quite high. The formulate ½Ii ðDHpp Þ2i =½IðDHpp Þ2 standard sample is used to calculate relative concentration of Mo5þ in the catalysts, in which Ii is intensity of peaks and DHpp is line width at peak to peak maximum [25]. The results of relative concentration of Mo5þ in the catalysts are summarized in Table 1. The relative concentration of Mo5þ increases in the order of Mo0:5 P0:3 Ox < AgMoO2 PO4 < Ag0:3 Mo0:5 P0:3 Oy < Ag0:5 Mo0:5 P0:3 Oy catalyst. Fig 3(B) gives the EPR spectra of Ag0.3Mo0.5P0.3Oy catalyst treated under different conditions. Comparing with the fresh catalyst, the EPR spectrum of used catalyst show a broaden peak, indicating interference of Mo5þ become stronger and the concentration of Mo5þ become higher after reaction. After the catalyst was treated in O2 flow at 773 K for 15 mm and then sealed in a glass tube, the intensities of the catalyst were remarkably less than that of the fresh one, in other words, the concentration of Mo5þ was quite low. The above results suggest Mo5þ is involved in activation of molecular oxygen and oxidation of propane. 3.1.4. XPS characterization The XPS spectra of Mo 3d on the catalysts are shown in Fig. 4(A). Mo0.5P0.3Ox and AgMoO2PO4, respectively show the binding energy (BE) of Mo 3d5/2

121

at 233.8 and 233.6 eV, which are higher than Ag-MoP-O catalysts. It can be seen that the BE of Mo 3d5/2 on Ag-Mo-P-O catalysts decreases from 233.4 to 233.0 eV with Ag content increasing. Fig 4(B) gives the XPS spectra of Ag 3d of Ag0.3Mo0.5P0.3Oy and Ag0.5Mo0.5P0.3Oy catalysts. The BE of Ag 3d on Ag0.3M0.5Po0.3Oy is slightly higher than that on Ag0.5Mo0.5P0.3Oy. The shoulder peaks at 368.2 eV is observed, which may be due to Ag0 3d. The positive BE shift with increase oxidation state are determined by the electronegative difference between the metal atom and cation. The cation has lower electron density in the valence region than the metal atom, which therefore reduces the screening of core-level electrons from their nucleus. This results in higher BEs for the core electrons of the cations compared to the metal. In the case of BE shift of Ag, the negative BE shift was observed. The BE shift of Ag 3d is governed by the factors than electronegativity difference such as lattice potential, work functions changes and extra-atomic relaxation energies, which result in the negative BE shift with oxidation state increase [26,27]. 3.2. Catalytic performance The results of catalytic performance of ODH of propane are shown in Table 2. The reactor did not contribute to ODH of propane, so the homogeneous ODH of propane could be avoid under our experiment conditions. Both MoO3 and Ag2O had no significant

Table 2 Catalytic performance of oxidative dehydrogenation of propane on the catalysts Catalyst

Temperature (K)

Conversion (%) C3H8

Selectivity (%)

Yield (%)

C3H6

C3–O

C2

COx

C3H6

Blank run Ag2O MoO3 Mo0.5P0.3Ox AgMoO2PO4 Ag0.1Mo0.5P0.3Oy

773 773 773 773 773 773

– – – 2.1 4.3 6.2

5.7 18.0 75.2

7.7 12.0 8.4

15.9 22.6 6.3

70.7 45.7 10.1

0.1 0.8 4.6

Ag0.3Mo0.5P0.3Oy

753 773 793

8.5 10.9 15.8

74.6 67.9 36.5

5.9 10.4 3.7

1.5 3.2 15.6

18.0 18.5 44.2

6.3 7.4 5.8

Ag0.5Mo0.5P0.3Oy

773

14.4

22.5



25.2

52.3

3.2

Reaction conditions: C3H8/O2/N2 ¼ 3/1/4, 2400 ml/(gcat h), atmospheric pressure. C3–O: acrolein, acetone, propanal; C2: ethane and ethane; acetaldehyde; COx: CO and CO2.

122

X. Zhang et al. / Applied Surface Science 220 (2003) 117–124

activity of ODH of propane. Mo0.5P0.3Ox exhibited 5.7% selectivity in propene with 2.1% propane conversion. Apparently, the addition of Ag into Mo0.5P0.3Ox improved the catalytic performance of ODH of propane. Propene, C3–O, acetaldehyde, C2 alkane and COx were products in ODH of propane on Ag-Mo-P-O catalysts. Significant different on activity and selectivity in propene could be observed depending on Ag content in the catalysts. In fact, propene conversion increased and selectivity in propene decreased with Ag content increasing, in addition, the higher selectivity in COx was observed on AgMo-P-O catalyst with the higher Ag content. The reaction temperature had great effect on activity and products distribution of ODH of propane. Selectivity in propene decreased and propane conversion increased with temperature increasing, and selectivity in COx was favorable at higher temperature. The maximal 7.4% yield of propene with 67.9% selectivity was obtained on Ag0.3Mo0.5P0.3Oy catalyst at 773 K. AgMoO2PO4 showed 18.0% selectivity in propane and 4.3% propane conversion.

4. Discussion 4.1. Redox properties The addition Ag dose modify the redox properties but also the number of Mo5þ species in these catalysts. The difference in redox properties of these catalysts was found by TPR, EPR and XPS. TPR results indicated that the addition Ag improved the reducibility of Ag-Mo-PO catalysts. XPS and EPR results showed the concentration of Mo5þ increased in Ag-Mo-P-O catalysts with Ag content increasing, indicating that Ag enhance the reduction of Mo6þ to unsaturated coordination Mo. The variation of the reducibility and the number of Mo5þ species of these catalysts could be due to the interaction between Mo and Ag in Ag-Mo-P-O catalysts. Mo6þ/Mo5þ and Agþ/Ag0 exist in Ag-Mo-P-O catalysts. The redox potential of Mo6þ/Mo5þ is about 0.3 eV. In relation to Agþ/Ag0, the redox potential is about 0.8 eV [28]. Both Ag and Mo are able to activate/store oxygen and transform/release oxygen species. Hence, the redox couple Ag0 þ Mo6þ , Agþ þ Mo5þ can form, improving the transfer of electron and oxygen species.

On the other hand, the addition of Ag gives rise to AgMoO2PO4 phase. Ag-Mo-P-O catalysts were mainly containing MoO3 and AgMoO2PO4 phases. The ratio of MoO3 to AgMoO2PO4 decreased as Ag content increasing. As regards the results of catalytic test, Ag-Mo-P-O catalysts show the higher activity and selectivity in propene than MoO3 and AgMoO2PO4. From the results of characterization, the reducibility and the concentration of Mo5þ of Ag-Mo-P-O catalysts are higher than those of MoO3 or AgMoO2PO4; in addition, the redox properties and the concentration of Mo5þ are relate to the ratio of MoO3 to AgMoO2PO4 in Ag-Mo-P-O catalysts. The above results indicated that the synergetic effect occurred between MoO3 and AgMoO2PO4, modifying the redox properties of Ag-Mo-P-O catalysts. The way explaining for the synergetic effect is to consider the ‘‘coherent interface’’ existing between AgMoO2PO4 and MoO3 because these two phases have close related structure. MoO3 is a complicate layer and zigzag raw structure, which is built by rather distorted MoO6 octahedron [29]. In the framework of AgMoO2PO4, Mo atoms locate in isolated octahedron, whose geometry corresponds to classical isolated MoO6 octahedron [21]. When the two phases are in contact, the crystallographic misfit can be low, and the interface is said to be ‘‘coherent interface’’. The atoms close to the interface are strained, so that their potential reactivity are greater than when they are inside the purse phase, and the diffusion of various species (atom, ions, electrons) across the interface is favored because the interface energy barrier is strongly lowered. Hence, the transfer of electron and the transfer of oxygen species should be improved by the synergetic effect [30–32]. At short range order in ‘‘coherent interface’’ microdomains of MoO3 phase inside (and/or at the surface of) AgMoO2PO4 can be formed during the transient state and be kinetically, but not thermodynamically, stable during the steady state. These microdomains are a way to account for the generation of new active sites and isolate sites for avoiding total oxidation [30–32]. The synergetic effect with respect to ‘‘coherent interface’’ has been reported in Ni-Mo-O catalysts for ODH of propane [30,32]. 4.2. Activity and selectivity in ODH of propane The addition of Ag into Mo0.5P0.3Ox catalysts greatly improved the catalytic performance of ODH

X. Zhang et al. / Applied Surface Science 220 (2003) 117–124

of propane. The Ag-doped catalysts favor the activity and selectivity in ODH of propane higher than the undoped one is due to the formation of redox couple ‘‘Ag0 þ Mo6þ , Agþ þ Mo5þ ’’ and synergetic effect originating from ‘‘coherent interface’’ between MoO3 and AgMoO2PO4. At increasing the Ag content in Ag-Mo-P-O catalyst, selectivity in propene decreased and propane conversion increased. The maximal 7.4% yield of propene with 69.7% selectivity in propene was obtained on Ag0.3Mo0.5P0.3Oy catalyst. Ag0.3Mo0.5P0.3Oy catalyst is a potential catalyst for ODH of propane. The yield of propene is maximal on Ag0.3Mo0.5P0.3Oy catalyst could be explained by an appropriate interaction between Ag and Mo and mutual covering the two phases. A clear correlation was observed between the catalytic activity/selectivity in ODH of propane and the redox and the number of Mo5þ species of Ag-Mo-P-O catalysts. From our results, the redox properties and the number of Mo5þ species increased as Ag content increasing. Propane conversion follows the same trend, while opposite is observed for the selectivity in propene. It seems reasonable to think that the activity and selectivity in ODH of propane is dependent on the variety of redox properties and the number of Mo5þ species of the catalysts. In a view of ODH of propane on Ag-Mo-P-O catalysts, the oxygen-deficient around Mo5þ reacted with molecular oxygen to form lattice oxygen species for propane activation and the Mo5þ was oxidized to Mo6þ. Propane absorbed on the lattice oxygen species and dehydrogenated to propene. The propane activation was improved by the increase of the concentration of Mo5þ and the enhancement of redox properties. On the other hand, the oxygen-deficient site also supplies adsorption site for product propene. Desorption of product propene will be mostly determined by the interaction of propene with the catalyst surface. The much more the number of Mo5þ and the much stronger redox properties will result in the consecutive oxidation of propene to COx. So selectivity in propene decreases with Ag content increasing.

5. Conclusions In conclusions, the addition Ag into the catalysts leads to forming redox couple ‘‘Ag0 þ Mo6þ ,

123

Agþ þ Mo5þ ’’ and the synergetic effect originating from ‘‘coherent interface’’ between MoO3 and AgMoO2PO4, modifying the redox and the concentration of Mo5þ in Ag-Mo-P-O catalysts catalytic performance. In the case of Ag-Mo-P-O catalysts, the catalytic performance depends on the redox and the concentration of Mo5þ in the catalysts. The enhancement of the redox and of the concentration of Mo5þ results in increasing propane conversion, and favoring the consecutive oxidation of propene to COx. Ag0.3Mo0.5P0.3Oy is a potential catalyst for the ODH of propane because of high yield in propene.

Acknowledgements Thanks for the finical support by the Ministry of Science and Technology (G1999022408).

References [1] T. Blako, J.M. Lopez Nieto, Appl. Catal. A 157 (1997) 117. [2] M.M. Bettahar, G. Costentin, L. Savary, Appl. Catal. A. 145 (1996) 1. [3] H.H. Kung, Adv. Catal. 40 (1994) 1. [4] M.A. Chaar, D. Patel, M.C. Kung, J. Catal. 105 (1987) 483. [5] M.A. Chaar, D. Patel, M.C. Kung, J. Catal. 109 (1988) 463. [6] M.C. Kung, H.H. Kung, J. Catal. 134 (1992) 668. [7] P. Concepcion, J.M. Lopez Nieto, J. Perez-pariente, Catal. Lett. 19 (1993) 333. [8] K.T. Nguyen, H.H. Kung, Ind. Eng. Chem. Res. 30 (1991) 352. [9] D. Patal, P.J. Anderson, H.H. Kung, J. Catal. 125 (1990) 132. [10] C.T. Au, W.D. Zhang, H.L. Wan, Catal. Lett. 37 (1996) 241; W.D. Zhang, C.T. Au, H.L. Wan, Appl. Catal. A. 181 (1999) 63. [11] A. Corma, J.M. Lopez Nieto, N. Paredes, Stud. Surf. Sci. Catal. 72 (1992) 213; J. Castiglioni, P. Poix, R. Kieffer, Stud. Surf. Sci. Catal. 75 (1993) 2309. [12] Y.S. Yoon, W. Ueda, Y. Moro-oka, Catal. Lett. 35 (1995) 57. [13] K.H. Lee, Y.S. Soon, W. Ueda, Catal. Lett. 46 (1997) 267. [14] Y.S. Yoon, K. Suzuk, T. Hayakawa, Catal. Lett. 59 (1999) 165; M.C. Abello, M.F. Gomez, L.E. Cadus, Catal. Lett. 43 (1999) 229. [15] J.C. Vedrine, J.M.M. Millet, J.C. Volta, Catal. Today 32 (1996) 115. [16] P. Concepcion, A. Galli, J.M. Lopez Nieto, Top. Catal. 3 (1996) 451. [17] D.L. Stern, R.K. Grasselli, J. Catal. 167 (1997) 550. [18] R.X. Valenzuela, E.A. Mamedov, V. Cortes Corberan, Raect. Kinet. Catal. Lett. 55 (1995) 213.

124

X. Zhang et al. / Applied Surface Science 220 (2003) 117–124

[19] M.C. Abello, M.F. Gomez, L.E. Cadus, Catal. Lett. 53 (1998) 185. [20] K.D. Chen, S. Xie, A.T. Bell, J. Catal. 195 (2000) 244. [21] P. Kierkegaard, S. Holmen, Ark. Kemi. 23 (1963) 213. [22] A. Parmaliana, F. Arena, F. Frusteri, Stud. Surf. Sci. Catal. 110 (1997) 347. [23] F. Arena, A. Parmaliana, J. Phys. Chem. 100 (1996) 1995. [24] K.M. Sancier, T. Dozono, R. Wise, J. Catal. 123 (1971) 270. [25] Y.G. Yin, Duoxiang Cuihuaji de Yanjiu Fangfa, Huaxue Gongye Chubanshe, Beijing, 1988. [26] J.T. Wolan, G.B. Hoflund, Appl. Surf. Sci. 125 (1998) 251.

[27] S.W. Gaarenstroom, N. Winograd, J. Chem. Phys. 67 (1977) 15. [28] R.C. Weast, M.J. Astle, W.H. Beyer, Redox potentional, in: Handbook of Chemistry and Physics (1983–1984) [M], 64th ed., CRC Press, Boca Raton, FL, 1984, pp. D156–D163. [29] G. Mestl, T.K.K. Srinivasan, Catal. Rev. Sci. Eng. 40 (4) (1998) 451. [30] P. Courtine, E. Bordes, Stud. Surf. Sci. Catal. 110 (1997) 177. [31] R.G. Teller, M.A. Antonio, J.F. Brazdil, J. Solid State Chem. 64 (1986) 249. [32] J. Barrault, C. Batiot, L. Magaud, Stud. Surf. Sci. Catal. 110 (1997) 375.