Reac Kinet Mech Cat (2010) 101:141–151 DOI 10.1007/s11144-010-0207-x
Novel F–V2O5/SiO2 catalysts for oxidative dehydrogenation of propane Yuefeng Liu • Chengfa Jiang • Wei Chu Wenjing Sun • Zaiku Xie
•
Received: 2 February 2010 / Accepted: 15 May 2010 / Published online: 4 June 2010 Ó Akade´miai Kiado´, Budapest, Hungary 2010
Abstract The V2O5/SiO2 catalysts promoted by fluoride anion and prepared with different F adding sequences were investigated. The samples were characterized by X-ray photoelectron spectroscopy (XPS), N2 adsorption–desorption, scanning electron microscopy (SEM), X-ray powder diffraction (XRD), temperature-programmed reduction (TPR), oxygen temperature-programmed desorption (O2-TPD) techniques and oxidative dehydrogenation reaction measurements. The results indicated that the fluoride impregnation sequence had a significant effect on the structure and selectivity of propene. Co-impregnation prepared VOF–C catalyst showed a better performance in the ODP reaction. The C3H6 selectivity was 74.9% at 480 °C, compared to 60.9% of V–Si–O sample without promotion. The propane conversion and propene yield were also better. The fluoride addition enhanced the mobility of oxygen in the catalyst. Keywords Vanadia catalyst Fluoride promoter Impregnation sequence Isolation of active sites Oxidative dehydrogenation of propane
Introduction In recent years, oxidative dehydrogenation of propane (ODP) has been intensively investigated because of its significant advantages over the dehydrogenation process, which is presently utilized for producing alkene derivatives [1–10]. Vanadia-based catalysts were active for oxidative dehydrogenation of lower alkanes, particularly propane [2–5, 7]. When the active component was loaded on a suitable support with Y. Liu C. Jiang W. Chu (&) W. Sun Department of Chemical Engineering, Sichuan University, Chengdu 610065, China e-mail:
[email protected] Z. Xie Shanghai Research Institute of Petrochemical Technology, SINOPEC, Shanghai 201209, China
123
142
Y. Liu et al.
a proper amount, it led to promising yields of propene. Many different supports were investigated, including TiO2 [1, 2], SiO2 [3–5], c-Al2O3 [7, 8] and MgO [10]. Vanadia loading with monolayer dispersion differed from supports and depended on the binding ability of the active component and the support. SiO2 is always attractive because of its large surface area, high thermal stability, and mechanical strength. It has been reported that the isolated VOx species in the samples supported on SiO2 were built up to a surface density of ca. 2.6 V/nm2 [11]. With the other supports, it was difficult to achieve high vanadium dispersion without excessive vanadium loading decrease, with the catalytic activity consequently deteriorated [3]. The oxidative dehydrogenation reaction is always accompanied by the formation of considerable amounts of undesirable carbon oxides. The suppression of deep oxidation is one of the present difficulties that need to be solved. It is well know that the catalytic properties of oxide catalysts can be improved by the introduction of additives [1, 8, 12, 13]. There were also reports on anion doped metal oxide catalysts (such as F-), which showed better activity in the oxidative dehydrogenation of light alkanes [12, 13]. Luo [12] prepared catalysts using a ball mill method and tested for oxidative dehydrogenation from ethane to ethene. Much more active O-, anion vacancies and Ce4? active sites were observed after the addition of F, which resulted in higher ethene selectivity. Zhang [13] investigated the pure rare earth metal based catalysts with the intorduction of CeF3. The new type of modulator and release agent for active site had an excellent reactivity even with low element contents. Recently, Deng [14] and co-workers discovered the significant effects of Cu and Co introduction on the catalytic behaviors. The co-impregnated catalyst had a strong synergistic effect and appropriate surface content of active components. Shu [15] studied the relationship of the impregnation sequence of Pt and Ni and the catalytic activity. It was indicated that the Pt impregnated first had a significantly higher activity. To obtain a highly efficient oxidative dehydrogenation catalyst, we designed fluoride doped V2O5/SiO2 catalysts for the ODP reaction. The effects of different impregnation sequences of F–V2O5/SiO2 catalysts were investigated. The characterizations of X-ray photoelectron spectroscopy (XPS), N2 adsorption–desorption, Scanning electron microscopy (SEM), X-ray powder diffraction (XRD), temperature program reduction (TPR), and temperature program desorption (TPD) techniques and catalytic measurements have been operated to determine their properties and catalytic performances.
Experimental Catalyst preparation The catalysts were prepared by different impregnation sequences of precursors. (1) Co-impregnation: SiO2 was impregnated by a solution containing NH4VO3 (adding C2H2O4 in 2:1 M ratio to NH4VO3) and NH4F in 1:1 M ratio of V:F. The precursor was dried at 110 °C for 12 h, and then calcined in air at 550 °C for 4 h. This
123
Novel F–V2O5/SiO2 catalysts
143
catalyst was designated VOF–C. (2) Impregnation of support with vanadia and fluoride: the support was impregnated first with NH4VO3 and C2H2O4 solutions, and the obtained solid was dried and calcined, and then impregnated with the solution of NH4F. The impregnation process and conditions were the same as those of coimpregnation method. The obtained sample was denoted V-1st. (3) Impregnation of support first with fluoride then vanadium: the prepared process and conditions were the same as those of V-1st sample, except the adding order of vanadium and fluoride precursors were the opposite. The prepared catalyst was denoted F-1st. (4) Impregnation of support only with vanadium: SiO2 was impregnated by a solution containing NH4VO3 (adding C2H2O4 in 2:1 M ratio to NH4VO3). The precursor was dried at 110 °C for 12 h, and then calcined in air at 550 °C for 4 h. The catalyst was designated as V–Si–O. The V2O5 content in the catalyst was expressed as the weight ratio of V2O5/(V2O5?SiO2). For these catalysts, the content of V2O5 was 14 wt%. Reaction measurement for propane oxidative dehydrogenation (ODP) Propane oxidative dehydrogenation was performed in a fixed-bed reactor with a continuous flow at atmospheric pressure and in the temperature range of 400– 600 °C. 200 mg catalyst was packed in a quartz tube which connected with a thermocouple in the middle of catalyst bed. The feed consisted of 10% C3H8, 5% O2, and 85% Ar, which were mixed in a composition to yield the stochiometric ratio of propane to oxygen as 2 for the ODP reaction. The total flow rate of gas reactant was 50 mL/min. The reactant mixture passed a static mixer and was pre-heated up to 400 °C prior to entering the reactor. The feed and product components out of the reactor were analyzed by an on-line gas chromatograph with a thermal conductivity detector (TCD) and a flame ionization detector (FID). A TDX01 column was used for the separation of O2, CO, CH4 and CO2, and a [DNBM?ODPN] column was used for the analysis of CH4, C2H4, C2H6, C3H6 and C3H8. Characterization methods X-ray photoelectron spectroscopy (XPS) experiments were performed on a XSAM 800 spectrometer with an Al–Ka (1486.6 eV) X-ray source. The binding energies (BE) were calibrated relative to the C 1s peak from carbon contamination of the samples at 284.6 eV. The specific surface area of carrier and catalysts were determined by N2 adsorption–desorption experiments at -196 °C on a Quantachrome Nova 1000e apparatus. Before measurements, the sample was degassed in vacuum at 300 °C for 3 h. The surface morphologies were observed by means of a scanning electron microscopy (PHILIP, XL30ESEM). X-ray powder diffraction (XRD) patterns were ´˚ recorded on the DX-2000 diffractometer using Cu–Ka (k = 1.54056 A ) radiation between 10° and 70°. In the experiments of temperature-programmed reduction (TPR), the steps were similar to those in our previous works [16, 17]. The catalyst (50 mg) was placed into a quartz reactor, and the sample was saturated in a gas stream of 5.0% (v) H2 in nitrogen at a total flow rate of 30 mL/min. After stabilization, the temperature was increased from 100 to 800 °C at a rate of 10 °C/min. For O2-TPD studies, 200 mg of the sample was placed in the middle of a quartz micro-reactor with
123
144
Y. Liu et al.
6 mm inner diameter. The outlet was analyzed on-line by Multi-absorption instrument (TP5080). The heating rate was 10 °C/min and the temperature range was from 50 to 750 °C.
Results and discussion Catalyst activity in propane oxidative dehydrogenation The propane conversions at the temperature of 480 and 540 °C of the catalysts with different fluoride added sequences are displayed in Table 1. It can be seen that the increase of the conversion with the temperature was usually accompanied by the decrease of the propene selectivity for the samples. The propene selectivity of 64.16% and propane conversion of 14.76% were obtained over VOC–F at 540 °C. Compared with V–Si–O, a 9.89% enhancement of propene selectivity was observed, a slight increase in C3H8 conversion was detected too. The yield of propane was 5.24% over the V-1st and 6.04% over the F-1st at 540 °C, lower than that of VOF– C. Meanwhile, the propene selectivity of VOF–C was equal to the V-1st and higher than the F-1st sample. The catalyst prepared by the co-impregnation method (VOF– C sample) displayed better catalytic performance in the ODP to propene. The relationship between the propane conversion and propene selectivity of typical catalysts are shown in Fig. 1. The trend of selectivity change with the conversion in our work was the same as in the other which reported vanadia-based catalysts [3, 5]. The decrease of the C3H6 selectivity was related to the increase of the conversion, which indicated that propene was formed and then further oxidized to COx [3]. The activity of the catalysts with the fluoride doping was higher than that of the undoped sample. The studies on the oxidative dehydrogenation reaction also showed that fluoride addition on the catalyst surface was beneficial for the isolation of surface active centers and decreased the production oxidation [12, 13]. The mechanism of oxidative dehydrogenation of propane to propene was studied in the previous reports [4, 8]. The commonly accepted mechanism was that propane transformed into propene by losing two hydrogen atoms on the active oxygen which were provided by the metal oxide catalysts. However, there was a competition
Table 1 Oxidative dehydrogenation of propane over catalysts of different fluoride adding sequence Sample
480 °C
540 °C
Conv. (C3H8)/%
Sel. (C3H6)/%
Yield (C3H6)/%
Conv. (C3H8)/%
Sel. (C3H6)/%
Yield (C3H6)/%
V–Si–O
7.93
60.90
4.83
11.52
54.27
6.25
V-1st
4.35
73.03
3.18
8.31
63.10
5.24
F-1st
4.48
60.13
2.69
11.31
53.37
6.04
VOF–C
8.74
74.93
6.55
14.76
64.16
9.47
Reaction conditions: V (C3H8):V (O2):V (Ar) = 10:5:85; GHSV = 9000 h
123
-1
Novel F–V2O5/SiO2 catalysts
90
Propene selectivity (%)
Fig. 1 C3H6 selectivity as a function of C3H8 conversion of typical catalysts. Reaction conditions: V (C3H8): V (O2): V (Ar) = 10:5:85; GHSV = 9000 h-1; 400–600 °C
145
75
60
VOF-C V-Si-O
45
30 0
4
8
12
16
20
Propane conversion (%)
pathway, in which C3H8 reacts with adsorbed oxygen or lattice oxygen, provided by the neighboring electrophilic surface to produce COx as major products, leading to a decrease in the selectivity of propene. The addition of fluoride could effectively insulate the neighboring electrophilic activity oxygen, resulting in the termination of the deep oxidation of propene. A more practical mechanism described propane oxidative dehydrogenation over V2O5/SiO2 by the DFT method [6, 18, 19]. Fig. 2 shows the reaction mechanism of ODP over the F–V2O5/SiO2 catalyst. Since Fdistributed and replaced O2- at random and the location of inert F- in the configuration was unimportant [12], we structured the configuration and assumed that only one of the O2- was replaced by F- as it is displayed in Fig. 2. First, the propane molecule is adsorbed at the active oxygen sites on the V2O5 surface. The C–H bond breaks as homolytic cleavage usually takes place at the active oxygen sites [18] (path 1 in Fig. 2). Next, another hydrogen atom abstraction from a CH3 group of C3H7 via path 2 directly leads to propene, meanwhile, the water produced in path 3 is attached to a V4? site. Further oxidation of propene was prevented significantly when fluoride was added into the metal oxide catalysts. Therefore, in our experiment, after fluoride was added into the catalyst, the selectivity of the product was obviously improved (from 60.90 to 74.93% at 480 °C). In their report, path 3 had the lowest energy among the reaction pathways [19], which was the most feasible pathway in our reaction. In this way, a H2O molecule is formed and dissociated from catalyst active site, makes the catalyst lose a large number of lattice oxygens and form the oxygen vacancies. In an O2 atmosphere (Path 5 in Fig. 2), O2 will be adsorbed to the oxygen vacancy to form active oxygen, which achieves the regeneration of the catalyst. But generally, the catalyst cannot recover the original structure, which is one of the main reasons for the decrease of catalytic activity. Surface element analysis of catalysts The XPS results of the VOF–C catalyst before and after reaction at 540 °C are shown in Table 2. The BE of V 2p3/2 was ca. 517.6 eV over both samples,
123
146
Y. Liu et al.
Fig. 2 Reaction mechanism of ODP over fluoride-containing metal oxide catalysts
Table 2 Surface element molar ratio of VOF–C catalyst of the fresh sample (a) and the sample after reaction at 540 °C (b) from XPS measurement Sample Binding energy (eV) V 2p
Surface element molar ratio (%)
O 1s OI
OII
F 1s
V
O
F
Si
(a)
517.6
531.0 (4.1%)
533.3
687.7
0.77
69.13
2.42
27.68
(b)
517.6
531.0 (3.9%)
533.1
687.7
0.92
66.17
0.63
32.28
corresponding to the V5? species. The BE of the main O 1s peak (OII) at 532– 533 eV corresponds to the oxygen in SiO2; a peak of small intensity at 530–531 eV (OI) can be attributed to the oxygen in VOx phases [20, 21]. The results of semiquantitative calculations indicated that the F atom concentration on the surface of the VOF–C fresh sample, compared with the VOF–C sample after reaction at 540 °C, decreased from 2.42 to 0.63%. Parts of the fluorine may be volatilized as vanadium fluoride-oxide at higher temperature. However, such volatilization does not impact the catalytic performance. Surprisingly, the vanadium concentration increased from 0.77 to 0.92%. This was attributed to the volatilization of a small part of the vanadium, while more interior vanadium species can be transferred to the
123
Novel F–V2O5/SiO2 catalysts
(a)
600
225
Volume (mL/g)
SiO 2
Volume (mL/g)
Fig. 3 N2 adsorption– desorption isotherms of the catalysts and carrier. (a) V–Si– O; (b) VOF–C; (c) V-1st; (d) F-1st
147
(b)
450
300
(c) 150
150
(d)
0
0.0
0.2
0.4
0.6
0.8
1.0
P/P0
75
0.0
0.2
0.4
0.6
0.8
1.0
P/P 0
surface, which compensates the lost vanadium and increased the vanadium concentration. Structural property and surface morphology analysis The N2 adsorption–desorption isotherms of various catalysts and carrier are illustrated in Fig. 3. The isotherms of the four samples could be classified as type IV according to the IUPAC classification [22, 23]. When P/P0 \ 0.7, the ascending trend of adsorption line was slow without observable inflexion or flat. When P/P0 [ 0.7, the adsorption isotherm line grew faster, an obvious abruption of adsorption and desorption lines appeared. The low or high of relative pressure related to the separation point in adsorption and desorption isotherms illustrates the wide or narrow of mesopore distribution: a higher relative pressure of separation point disclosed a widespread mesopore distribution, while lower relative pressure suggested a narrow mesopore distribution [22]. As shown in Fig. 3 and Table 3, supporting V2O5 on SiO2 has a condensation of the relative pressure compared with SiO2, which resulted in the lower of the specific surface area of the V–Si–O sample. Similar results were obtained on other fluoride doped catalysts. The relative pressure of sample isotherms shifted from P/P0 = 0.6–0.7 to P/P0 = 0.7–0.8, which indicated that the surface area and pore diameter became lower. Besides, the specific surface area and pore diameter of the co-impregnation catalyst (VOF–C) were larger than the stepwise impregnated catalysts V-1st and F-1st (Table 3). Fig. 4 displays the morphology of VOF–C and V-1st. As shown in Fig. 4, the particles on the surface of VOF–C sample were uniform, while the particles on the surface of V-1st were aggregated. Considering the data of N2 adsorption– desorption, it was found that the active component has got into the SiO2 hole after the stepwise impregnation. This resulted in a further block of hole in V-1st, causing the decrease of the catalyst specific surface area, pore volume and pore diameter (Table 3). X-ray diffraction of three samples and the carrier are presented in Fig. 5. No diffraction peaks of fluoride were observed in the figure. Only smaller V2O5 crystallites (JCPDS #72-0598) were detected by X-ray diffraction on SiO2 support.
123
148
Y. Liu et al.
Table 3 Composition and texture property of the prepared catalysts Sample
V2O5 content (wt%)
Surface area (m2/g)
Pore volume (mL/g)
Pore size (nm)
SiO2
–
301.7
0.945
12.53
V–Si–O
14
225.8
0.413
7.31
VOF–C
14
207.2
0.364
7.03
V-1st
14
189.8
0.285
6.00
F-1st
14
171.0
0.213
4.99
Fig. 4 Surface morphology of the prepared samples. (a) VOF–C; (b) V-1st
Thus, it was realized that most of the V2O5 species were highly dispersed on the SiO2 in this catalyst system. The V2O5 diffraction peaks of the V-1st and F-1st samples were the strongest, which were ascribed to the stepwise impregnation and resulted in the agglomeration of active component. For VOF–C catalysts, the diffraction peaks of V2O5 crystallites became weaker, indicating that fluoride doping in the catalyst using co-impregnation was synergized with the active component, which avoided V2O5 further cluster. The result was also confirmed by the SEM analyses (Fig. 4). Temperature-programmed measurements of the samples TPR tests were used to achieve a better understanding of the catalyst reducibility and possible interaction of active components [24]. The reduction behaviors of the catalysts were significantly influenced by the fluoride adding sequence as shown in Fig. 6. From Fig. 6, H2-TPR profiles of V–Si–O sample show a single reduction peak observed at 618 °C, which could be attributed to the reduction of V5? to V3? [25]. Compared with the reduction behavior of mere V2O5 supported on SiO2, the reduction temperatures of the fluoride doped over V2O5/SiO2 catalysts shifted to lower region. This indicated that the addition of fluorine resulted in easy reduction of the active V2O5 component. The H2-TPR profile of the VOF–C catalyst exhibited
123
Novel F–V2O5/SiO2 catalysts
149
Fig. 5 XRD patterns of the prepared catalyst
V2O5
Intensity (a.u.)
F-1st V-1st VOF-C SiO2
10
20
30
40
50
60
70
2 (° )
Fig. 6 H2-TPR profiles of the samples with different fluoride adding sequence
603
644
Intensity (a.u.)
614
F-1st
598 618
V-1st VOF-C V-Si-O 400
500
600
700
800
Temperature (°C)
an intensive reduction peak with a maximum temperature around 598 °C, which was less than the temperature of reduction of the main species in V–Si–O catalyst for 20 °C. For the V-1st catalyst, the H2-TPR profile similar to the V–Si–O catalyst showed an intensive reduction peak with a maximum at the temperature about 614 °C. Compared to this, two different vanadium reduction peaks were observed in the case of F-1st. The reduction peak at 603 °C was attributed to the incomplete reduction of V5? to V3? (partial reduction to V4?), whereas the peak at 644 °C could be assigned with the reduction of V4? to V3? and/or a presence of nonreducible V5? species [26]. Arena et al. [27] investigated the reduction capacity of VOx/SiO2 catalysts. They reported that the lower temperature peak was ascribed to the reduction of monolayer or low-oligomeric VOx species on the surface, while the higher temperature peak attributed to a portion of bigger-sized V2O5 phases. O2-TPD profiles of catalysts with different fluoride adding sequences are shown in Fig. 7. There were two O2 desorption peaks at ca. 395 and 656 °C in the V–Si–O
123
150
Y. Liu et al.
Fig. 7 O2-TPD profiles of different catalyst samples. (a) V–Si–O; (b) VOF–C; (c) V-1st; (d) F-1st
Intensity (a.u.)
(d) (c)
(b) (a) 200
300
400
500
600
700
800
Temperature ( °C)
catalyst, while only one main O2 desorption peak for fluoride added catalyst. In the present work, the desorption peaks of surface adsorbed oxygen were not very obvious. In the sample V–Si–O, observed desorption peaks (a peak) ascribed to surface oxygen in the low temperature region (\450 °C) and a high-temperature peak ([600 °C, b peak) attributed to desorption of lattice oxygen were observed. The results were similar to the previous reports [26]. Two overlapped desorption peaks were observed in the sample VOF–C. It was noteworthy that the b peak transferred to the low-temperature region, which formed a large overlapped peak with the a peak. Two observable peaks in both V-1st and F-1st samples were also overlapped but broader. The temperature corresponding to the desorption of O2 (ad) and O2-(ad) was low while that of lattice oxygen was harder and at higher temperature [28]. It was reported that [9] the desorption temperature ranging from 500 to 800 °C corresponded to mobile oxygen, such as O2-, O- and O22-, and not the conventionally considered lattice oxygen. The addition of fluorine increased the transition between surface oxygen and lattice oxygen, enhanced the mobility of oxygen species in catalysts.
Conclusions V2O5/SiO2 catalysts promoted by fluoride anion and prepared with different F adding sequences were investigated in this work. The addition of fluorine increased the transition between surface oxygen and lattice oxygen, which enhanced the mobility of oxygen species in catalysts. Meanwhile, the isolation of surface active centers became available, resulting in an obvious improvement of C3H6 selectivity. The performance and structure of catalysts were greatly affected by the fluoride impregnation sequence and the preparation technique. For VOF–C catalyst, the active component was easily reduced and highly dispersed on the carrier than the catalysts prepared by using other fluoride adding sequence. The VOF–C catalyst prepared by co-impregnation yielded better catalytic performance for the ODP to propene.
123
Novel F–V2O5/SiO2 catalysts
151
Acknowledgements This work was supported by National Natural Science Foundation of China (20776089) and New Century Excellent Talent Project of China (NCET-05-0783). The authors would like to thank Dr. Dongge Tong, Dr. Hui Zhang, Huiyuan Xu and Jinjie Luo for their useful discussions and helps.
References 1. Garcia T, Solsona B, Murphy DM, Antcliff KL, Taylor SH (2005) J Catal 229:1 2. Zhaorigetu B, Li WZ, Kieffer R, Xu HY (2002) React Kinet Catal Lett 75:275 3. Rossetti I, Fabbrini L, Ballarini N, Olivaa C, Cavani F, Cericolab A, Bonelli B, Piumetti M, Garrone E, Dyrbeckd H, Blekkand EA, Forni L (2008) J Catal 256:45 4. Chen K, Khodakov AY, Yang J, Bell AT, Iglesia E (1999) J Catal 186:325 5. Klisin´ska A, Samson K, Gressel I, Grzybowska B (2006) Appl Catal A 309:10 6. Rozanska X, Kondratenko EV, Sauer J (2008) J Catal 256:84 7. Schwarz O, Frank B, Hess C, Schoma¨cker R (2008) Catal Commun 9:229 8. Jibrila BY, Al-Zahrania SM, Abasaeeda AE, Hughesb R (2004) Chem Eng J 103:59 9. Zhaorigetu B, Li WZ, Xu HY, Kieffer R (2004) Catal Lett 94:125 10. Mishakov IV, Vedyagin AA, Bedilo AF, Zaikovskii VI, Klabunde KJ (2009) Catal Today 144:284 11. Tian HN, Elizabeth IR, Wachs IE (2006) J Phys Chem B 110:9593 12. Luo JZ, Wan HL (1997) Appl Catal A 158:137 13. Zhang WD, Tang DL, Zhou XP, Wan HL, Tasi KR (1994) J Chem Soc Chem Commun 6:771 14. Deng SY, Chu W, Xu HY, Shi LM, Huang LH (2008) J Nat Gas Chem 17:369 15. Shu YY, Murillo LE, Bosco JP, Huang W, Frenkel AI, Chen JG (2008) Appl Catal A 339:169 16. Chu W, Chernavskii PA, Gengembre L, Pankina GA, Fongarland P, Khodakov AY (2007) J Catal 252:215 17. Zhang H, Chu W, Zhou J, Hong JP, Jing F (2008) React Kinet Catal Lett 94:139 18. Rozanska X, Fortrie R, Sauer J (2007) J Phys Chem C 111:6041 19. Fu H, Liu ZP, Li ZH, Wang WN, Fan KN (2006) J Am Chem Soc 128:11114 20. Rybarczyk P, Berndt H, Radnik J, Pohl MM, Buyevskaya O, Baerns M, Bru¨ckner A (2001) J Catal 202:45 21. Ruszel M, Grzybowska B, Ga˛sior M, Samson K, Gressel I, Stoch J (2005) Catal Today 99:151 22. Rouquerolt J, Aunir D, Fairbridge CW, Everett DH, Haynes JH, Pernicone N, Ramsay JDF, Sing KSW, Unger KK (1994) Pure Appl Chem 66:1739 23. Xu JQ, Chu W, Luo SZ (2006) J Mol Catal A 256:48 24. Khodakov AY, Chu W, Fongarland P (2007) Chem Rev 107:1692 25. Steinfeldt N, Mu¨ller D, Berndt H (2004) Appl Catal A 272:201 26. Berndt H, Martin A, Bru¨ckner A, Schreier E, Mu¨ller D, Kosslick H, Wolf GU, Lu¨cke B (2000) J Catal 191:384 27. Arena F, Giordano N, Parmaliana A (1997) J Catal 167:66 28. Jiang XY, Zhou RX, Pan P, Zhu B, Yuan XX, Zheng XM (1997) Appl Catal A 150:131
123
