The reaction of methanol oxidation on MoOJTi02 catalysts shows different behaviour depending on the surface coverage with molybdena and on the specific surface area of the Ti02 support. The nature of the Ti02 support (anatase or mtile) seems to play only a minor role. Two different zones can be distinguished which correspond to a molybdena content higher and lower than a theoretical monolayer (mnl). The catalytic behaviour of samples with a low molybdena content ( < 1 mnl) depends largely on the specific surface area and on the surface coverage of the Ti02 support. Catalysts with a high molybdenum content ( > 1 toni) exhibit similar catalytic properties. The formation of methyl formate is observed only on high specific surface area catalysts and it involves methoxy groups formed by the reaction of methanol with the OH groups on the surface of the TiO2 support.
Trang 1Applied Catalysis A: General, 96 ( 1993 ) 279-288 279 Elsevier Science P u b l i s h e r s B.V., A m s t e r d a m
A P C A T A2455
Methanol oxidation on MoO3/TiO2 catalysts
K Briickman and B Grzybowska
Institute of Catalysis and Surface Chemistry, Polish Academy of Sciences, 30-239 Krakdw (Poland)
and
M Che and J.M Tatibou~!t
Laboratoire de R~activitd de Surface et Structure, URA CNRS 1106, Universitd P et M Curie,
4 Place Jussieu, 75252 Paris Cedex 05 (France)
(Received 8 September 1992, revised manuscript received 15 December 1992 )
A b s t r a c t
The reaction of methanol oxidation on M o O J T i 0 2 catalysts shows different behaviour depending on the surface coverage with molybdena and on the specific surface area of the Ti02 support The nature
of the Ti02 support (anatase or mtile) seems to play only a minor role Two different zones can be distinguished which correspond to a molybdena content higher and lower than a theoretical monolayer (mnl) The catalytic behaviour of samples with a low molybdena content ( < 1 mnl) depends largely on the specific surface area and on the surface coverage of the Ti02 support Catalysts with a high molyb- denum content ( > 1 toni) exhibit similar catalytic properties The formation of methyl formate is ob- served only on high specific surface area catalysts and it involves methoxy groups formed by the reaction
of methanol with the OH groups on the surface of the TiO2 support
Keywords: methanol oxidation; MoOa/TiO2
I N T R O D U C T I O N
Over the last decade transition metal oxides, in particular molybdenum and vanadium oxides, dispersed on a carrier have attracted much attention due to their wide range of applications in a variety of reactions Special interest has been paid to low-loading systems with dispersed active oxide At these loadings
a bidimensional, monolayer-type species can be formed, whose properties are considerably modified with respect to unsupported oxides [ 1 ] Of these sys- tems, oxides supported on titanium dioxide have been widely studied, following the successful application of V2OJTi02 catalysts in o-xylene oxidation to
Correspondence to: Prof B Grzybowska, I n s t i t u t e of Catalysis and Surface Chemistry, Polish Academy of Sciences, ul Niezapominajek 3, 30-239, Krak6w, Poland Tel ( + 48-12 )365737, fax ( + 48-12 )365281
0926-860X/93/$06.00 © 1993 Elsevier Science Publishers B.V All right.~ reserved
Trang 2280 K Briickman et al./Appl Catal A 96 (1993) 279-288
phthalic anhydride [2,3] One of the important problems in these studies is the dependence of the structure and properties of the dispersed phase on the type of titania: (anatase or rutile), and on the specific surface area of this support Less attention has been paid to MoO,~/TiO2 in this respect, with most
of the papers on this system being concerned with structural studies on anatase titania [ 4-11 ], and only a few devoted to catalytic oxidation [ 11-18 ]
In a previous paper [ 11 ] catalysts containing MoO.~ supported on low spe- cific surface area titania (anatase and rutile) have been characterized by dif- ferent techniques and tested in butene-1 and o-xylene oxidation Marked dif- ferences were found, in the structure and properties of molybdenum species, between catalysts with a low (submonolayer) and those with a high content of MOO3 On the other hand for the same molybdenum content relatively small differences in the type of molybdenum species on anatase and rutile titania were observed
It is therefore interesting to study methanol oxidation on MoOffTiO,~ cat- alysts This reaction, an important route to formaldehyde synthesis on ferric molybdate catalysts may also yield, as has been shown recently, methyl for- mate on catalysts containing molybdena or vanadia dispersed on carriers [17- 22] In addition to its practical importance, the oxidation of methanol is a convenient model reaction for probing catalyst centres of various character; the formation of various side products accompanying oxidation to formalde- hyde such as methylal (dimethoxymethane) - - (CH:30)2CH2, dimethyl ether
- - (CH3)20, or methyl formate - - HCOOCH3 can provide information both
on the nature of centres and on their mutual arrangement [23] It has been shown that this reaction is structure sensitive on unsupported molybdenum oxide [23 ] and is dispersion-dependent on MoOffSiO,~ catalysts [ 19 ] which selectively form methyl formate on highly dispersed samples Methanol oxi- dation on MoOffTiO~ has been reported recently by Soung Kim et al [17] and Matsuoka et al [18], their studies being, however, limited to TiO.~ with a relatively high specific surface area (50 me/g), containing both anatase and rutile modification of titania, and to relatively low reaction temperatures
In the present work methanol oxidation was studied for MoO~ deposited on low specific surface area anatase and rutile titania and for catalysts containing MoO3 on anatase TiO2 of high specific surface area T h e decomposition of isopropanol was measured on the same catalysts, this reaction being taken as
a probe of redox and acidic centers [24,25] and as an indicator of the type of molybdenum dispersion [ 26 ]
EXPERIMENTAL
Three different TiO2 supports denoted AN-7, RT-5 and AN-48 were used for the catalysts' preparation Anatase AN-7 (7 m'~/g) and rutile (RT-5; 5 m2/ g) were supplied by Chemical Works "Police" (Poland) TiO2 anatase AN-48
Trang 3K Briickrnan et al./Appl Catal A 96 (1993) 279-288 281
with high surface area (48 m2/g) was prepared by hydrolysis of titanium iso- propoxide AN-7 and RT-5 were found to contain potassium and phosphorus impurities with a surface concentration of 5-8 at.-%, estimated from X-ray photoelectron spectroscopy (XPS) measurements No impurities were de- tected on AN-48 TiO2 support
Two oxygen peaks with a binding energy (BE) of 530 _+ 0.2 and 531.5 _+ 0.2
eV were observed by the XPS technique on the AN-48 titania, whereas only one single peak at 530 _+ 0.2 eV was found on the AN-7 and RT-5 samples The value of the BE at 530 eV is characteristic of the lattice oxygen in most tran- sition metal oxides, the higher value may be due to oxygen in the hydroxyl groups on the oxide surface [27,28]
The MoOJTiO2 catalysts were prepared by incipient wetness impregnation
at pH = 6 of the TiO2 support with an appropriate amount of an aqueous so- lution of ammonium heptamolybdate, followed by evaporation of water, and drying at 120 ° C Samples were calcined for 8 h in a stream of air at 400°C for Mo/AN-48 samples and at 500 °C for Mo/AN-7 and Mo/RT-5 catalysts The Mo/AN-7 and Mo/RT-5 samples were the same as those used in [11] One sample containing 1 mnl of MOO:, was also prepared on AN-7 titania washed for 8 h in a stream of water at 50°C: the XPS analysis of the washed support did not reveal potassium and phosphorus impurities on the support surface The molybdenum content in the samples varied between a fraction of a mono- layer (mnl) and about 3 mnl One monolayer calculated with the assumption that (ibMoO.~=15.4 ~k 2 [29] corresponds to 6.5-10 '8 molybdenum atoms per 1
m 2 It should be remembered that the maximal coverage of a titania anatase surface with dispersed MoOx species obtained experimentally with the grafting technique does not exceed about 0.6 theoretical monolayer [5] The X-ray diffraction (XRD) analysis revealed the presence of crystalline MoO3 only in the sample Mo/AN-48 with the highest molybdena content (2.4 mnl; 15 wt.-
% MoO,~) The BET surface areas of the Mo/TiO2 samples were equal within 0.5 m2/g to those of the pure supports
Methanol oxidation was carried out at atmospheric pressure in a flow mi- croreactor at 260°C, using a mixture of methanol, oxygen and helium in the ratio 7/16/77 (mol-%), with a total flow-rate of 34 ml/min A detailed descrip- tion of the apparatus and of the products analysis was published elsewhere [19] The methanol conversion was kept below 20% by adjusting the weight
of the catalyst sample
Isopropanol decomposition was studied by the pulse method [26] at 170 ° C Pulses of 2/~l pure isopropanol were injected in the carrier gas (dried helium;
30 m l / m i n ) , which passed through the catalyst (0.1 g) The reaction products were analyzed on-line by gas chromatography
RESULTS AND DISCUSSION
The catalytic activities and selectivities in methanol oxidation at 260 ° C for the three series of preparations are presented in Table 1 Formaldehyde is the
Trang 4282 K Briickman et al,/Appl Catal A 96 (1993) 279-288
TABLE 1
Oxidation of m e t h a n o l on MoO,~/Ti02 catalysts at 260 °C
~MF, methyl formate
bML, methylal
cSample prepared on washed AN-7
main reaction product for all the catalysts studied with a selectivity in the range 60-96% Much higher selectivities to this product as compared with the data of Soung Kim et al [ 17] may be due to the higher reaction temperature used in the present studies On the other hand Matsuoka et al [ 18] found formaldehyde to be a main reaction product for similar types of catalysts at 225°C The common feature of the studied catalysts is the low selectivity to carbon oxides (CO + CO2 < 4% ), the only exception being the samples pre- pared on low surface area rutile titania with a MoO~ content below the mono- layer The selectivities to other products of methanol transformation (meth- ylal, methyl formate and dimethyl ether) depend considerably on the type of TiO~ support and on the molybdenum loading Methyl formate and dimethyl ether are formed with a significant selectivity only on AN-48 catalysts at sub- monolayer coverage with molybdena, their quantities decreasing with the mo- lybdenum loading Methylal formation follows a different trend for the high and low specific surface area samples: it decreases with molybdenum content for Mo/AN-48 samples and increases with this parameter for Mo/AN-7 and Mo/RT-5 preparations The sample prepared on washed AN-7 shows a similar distribution of products to the samples containing low quantities of molybde- num dispersed on AN-7 with potassium and phosphorus impurities: higher selectivities to CO2 are, however, observed on potassium-free preparation Fig
1 gives the rates of formation of various products as a function of the number
Trang 5K Briickrnan et al./Appl Catal A 96 (1993) 279-288
o
b
071
05
03
01
equivalent MoO 3 monoIayers
283
O04
~" 0 0"3 • A N - 7 RT 5
002
eouivotent MoO 3 mono[oyers
Fig l Activity of MoOJTi02 catalysts in methanol oxidation as a function of molybdena content
Reaction temperature 260°C (a) Rate of formaldehyde formation: ( [ ] ) Mo/AN-48; ( O ) Mo/
AN-7; ( A ) Mo/RT-5 (b) Rate of side products formation: ( I , O, • ) methylal; ([J, O, ~ )
methyl formate Denotation of catalysts as in Fig la
of equivalent molybdena monolayers The rates are expressed as the number
of moles of methanol reacted to a given product, per molybdenum atom As can be seen the most significant differences are observed between the catalysts
of low molybdena content prepared on high specific surface area titania AN-
48 and those obtained from low surface area AN-7 and RT-5 At a molybdena content below one monolayer, the total activity (see Table 1 ) and the rate of formaldehyde formation are much higher for M o / A N - 4 8 samples, when com- pared with both M o / A N - 7 and M o / R T - 5 preparations At higher coverage ( > 1 mnl) the activities of all the catalysts approach similar values The type
of polymorphic modification of titania of low specific surface area also affects, though to a smaller extent, the activity of the dispersed molybdena phase The total rate of methanol consumption and the rate of formaldehyde formation in the submonolayer region are higher for rutile supported catalysts as compared with anatase AN-7 samples, and their values are similar for both types of sam-
Trang 6284 K Briickman et al./Appl Catal A 96 (1993) 279-288
pies at concentrations of MOO.3 exceeding a monolayer coverage A similar ef- fect of the titania modification was observed by Tatibou~t [30 ] in the oxida- tion of methanol on vanadia-titania samples A small amount of methyl formate was observed on Mo/RT-5 samples, whereas on Mo/AN-7 this product was absent Similar behavior was reported by Van Hengstum et al [ 20 ] for vanadia supported on low surface area anatase and rutile titania, though Tatibou~t did not observe any marked change in the formation of this product [30]
The above results suggest that the structure and properties of molybdena dispersed on a TiO2 surface were different in the submonolayer region, de- pending on the specific surface area of the support and to a smaller extent on the modification of titania On the other hand the catalytic behavior of all the catalysts was similar for a coverage of MoO3 higher than a theoretical mono- layer, indicating that the dispersed molybdena species were probably identical Some information about the mode of dispersion can be provided by mea- surements of isopropanol decomposition, which occurs to acetone on dehydro- genating, basic centers, and to propene on dehydrating, acidic sites [ 24 ] It has been shown previously [26], that this reaction can be applied as a probe of the mode of dispersion of vanadia and molybdena on low surface area titania The negligible yield of propene observed in the submonolayer region and the high acetone yield increasing with an active oxide content up to maximum at a monolayer content, were considered as an indicator of the formation of a bi- dimensional, monolayer structure of dispersed oxide on titania The increase
in dehydration to propene at a higher content of dispersed oxide was assigned
to the appearance of bulk, tridimensional vanadia or molybdena, since for un- supported V205 and MoO:~, dehydration is the predominant reaction In Fig 2 the results of isopropanol decomposition measurements are presented for the three series of MoOffTiO2 catalysts The activity of pure supports in this re-
025
o 0.20
015
E
~ OlO
o
005
A N -.!~
"-I
05 1 115 2 25
equivatent MoO 3 monotoyers
12
~ o ~ o8 ~
05 E
04 )
02
Fig 2 Isopropanol decomposition on MoO3/TiO~ catalysts at 170°C as a function of molybdena content ([-], O, A ) acetone; ( 1 , 0 , • ) propene Denotation of catalysts as in Fig 1
Trang 7K Briickman et al./Appl Catal A 96 (1993) 279-288 285
action was found to be negligible under the conditions studied As can be seen
in the submonolayer range, the rate of acetone formation is higher for the high surface area sample as compared with both the low specific surface Mo/AN-7 and Mo/RT-5 samples and its maximum, corresponding to the full coverage with the monolayer structure is obtained at a lower molybdenum content than for the other two series This indicates a better dispersion of molybdena in the submonolayer region on the high surface area titania A comparison of Figs
la and 2 shows a parallel course for the formation of acetone and formaldehyde
on MoO3/TiO2, confirming that methanol oxidation to formaldehyde occurs
on dehydrogenating centers The electron spin resonance (ESR) data for the Mo/An-7 and Mo/RT-5 [11] and the Mo/AN-48 [31] samples in the sub- monolayer range suggest that these centers are molybdenyl species M=O, in strongly distorted, coordinatively unsaturated octahedra The increase of the rate of formaldehyde formation per atom with increasing coverage of titania with these species is consistent with the mechanism which requires the pres- ence of two M=O bonds in close vicinity for oxidation to formaldehyde [ 19,32 ] The decrease in the rate of formaldehyde formation coincides with the increase
in propene formation in isopropanol decomposition, indicating the appearance
of bulk molybdena and hence the smaller dispersion of this oxide No distinct correlation is, however, observed between the rate of propene formation (a measure of the concentration of acidic centers) on the different samples and the overall rate of methanol oxidation
According to Louis et al [ 19 ] the formation of methyl formate on Mo/SiO~ catalysts requires the presence of both isolated oxo-molybdenum sites and methoxy groups on the support The adsorbed formaldehyde species formed
on oxo-molybdenum groups spill-over onto the silica surface where they react with methoxy groups to form a hemiacetal intermediate, which is further transformed into methyl formate The decrease in the rate of methyl formate formation with an increase in molybdenum coverage, which is observed in the present work for high surface area Mo/AN-48 samples, could be explained in
a similar way Methoxy groups are formed on the titania surface by the reaction
of surface hydroxyl groups with methanol:
- T i O H (s) + CH:,OH -~-Ti0CH~ (5) + H20
The concentration of OH surface groups, located on the molybdenum-free TiO2 surface, decreases when molybdenum coverage increases, inducing a de- crease in the methoxy group concentration and hence in the rate of methyl formate formation The absence of methyl formate on the Mo/AN-7 catalysts
at submonolayer coverage suggests that the uncovered titania is not able to form methoxy groups by the reaction given above, due to either the absence of
OH groups on molybdenum-free, low surface area anatase, or to the blocking
of these groups by a potassium impurity Such an explanation is consistent with the observation by XPS of two forms of oxygen on the high surface titania
Trang 8286 K Bri2ckman et aL/Appl Catal A 96 (1993) 279-288
AN-48 which can be ascribed to OH groups and to lattice oxygen, whereas only one single oxygen peak corresponding to lattice oxygen is observed on low sur- face area samples T h e fact t h a t the sample prepared on the washed AN-7 titania shows a similar distribution of products to the preparations on the sup- port containing potassium and phosphorus, suggests that these impurities are not responsible for the absence of methyl formate on the low surface area sam- ples Moreover as can be seen from Table 1 pure titania AN-48 is capable of forming dimethyl ether, a product requiring adsorption of methanol in the form of methoxy groups T h e selectivity to dimethyl ether decreases with in- creasing content of deposited molybdena No dimethyl ether formation is found
on low specific surface area samples of either pure supports or of mixed MoO:~- Ti02 catalysts, supporting the hypothesis that the OH groups capable of meth- anol adsorption in the form of methoxy groups on the support surface are ab- sent in this case It should be mentioned t h a t a similar effect of a decreasing amount of methyl formate with an increasing coverage of titania with depos- ited phase and an increase of this product on high specific surface area samples was observed by Forzati et al for vanadia-titania catalysts [21 ]
The decrease of the rate of methylal formation with molybdenum content
in the submonolayer region on Mo/AN-48 catalysts, similar to t h a t of methyl formate formation suggests, that the methoxy groups on the carrier are also involved in the reaction leading to methylal on these catalysts Methylal is only formed on low surface area catalysts M o / A N - 7 and M o / R T - 5 for concentra- tions above the monolayer coverage i.e in the region where the presence of bulk molybdena is expected T h e mechanism of methylal formation in this case could be analogous to t h a t proposed for pure MoO3 [23,33], involving two types of molybdenum centers in vicinity: the dehydrogenating M=O site and the exposed Mo 6+ ion of Lewis acid properties
CONCLUSIONS
The activity and selectivity in methanol oxidation on molybdena deposited
by the impregnation technique on different titania supports are dependent on the specific surface area of the support, on the molybdena loading and to a smaller extent on the kind of polymorphic modification of TiO2 Two different ranges can be distinguished corresponding to a molybdena content lower and higher than a theoretical monolayer of MOO3 In a submonolayer region the total rate of methanol consumption is considerably higher for high specific area catalysts indicating better dispersion of MoOz T h e formation of methyl formate and methylal is observed in addition to formaldehyde - - the main reaction product T h e decrease of methyl formate and methylal production with coverage of titania with molybdena species is explained by the mechanism that involves the participation of methoxy groups formed by the reaction of methanol with OH groups on the support T h e methoxy groups on the titania
Trang 9K Briickman et al./Appl Catal A 96 (1993) 279-288 287
surface react with formaldehyde which spills over from the molybdena species
to give methyl formate On low specific surface area samples no formation of side products is observed in the submonolayer region, the activity in formal- dehyde formation being slightly higher for molybdena on futile samples
At a molybdena coverage higher than a monolayer, the catalytic behavior in methanol oxidation is quite similar for the studied catalysts, independent of the specific area of the support and TiO2 modification
The rate of formaldehyde formation can be correlated with the rate of iso- propanol decomposition to acetone, indicating that the same dehydrogenating centers are involved in these reactions In contrast, no distinct correlation is observed between the rate of the formation of methanol oxidation products and the rate of isopropanol dehydration to propene, which is known to occur
on acidic centers
ACKNOWLEDGEMENTS
The participation of Mrs E Filipek-Labuda, M.Sc., in the preparation of the catalysts and with the isopropanol measurements is gratefully acknowl- edged The authors also wish to express their thanks to Dr J Stoch for pro- viding the X P S measurements of the supports
REFERENCES
1 G.C Bond, S Flamerz and R Shrukri, Discuss Faraday Soc., 87 (1989) 65
2 B Grzybowska, Catal Today, 1 (1987) 157
3 G.C Bond and S Flamerz, Appl Catal., 71 ( 1991 ) 1; and references therein
4 K.Y.S Ng and E Gulari, ,1 Catal., 92 (1985) 340
5 G.C Bond, S Flamerz and L Van Wijk, Catal Today, 1 ( 1987 ) 229
6 M.I Zaki, B Vielhaber and H Kniizinger, J Phys Chem., 90 (1986) 3176
7 K Segawa, Du Soung Kim, Y Kurusu and I.E Wachs, in M.J Phillips and M Ternan
(Editors), Proc 9th Int Congr on Catalysis, Calgary, Canada, Vol 4, The Chemical Institute
of Canada, Ottawa, Canada, 1988, p 1960
8 G Ramis, G Busca and V Lorenzelli, Z Phys Chem N.F., 153 (1987) 1890
9 T Machej, B Doumain, B Yasse and B Delmon, J Chem Soc., Faraday Trans l, 84 ( 1988 )
3905
10 T Machej, J Haber, A.M Turek and J.E Wachs, Appl Catal., 70 ( 1991 ) 115
11 E Filipek, B Grzybowska, E Serwicka, Y Barbaux, J.P Bonnelle and J Grimblot, in G
Centi and F Trifirb (Editors), New Developments in Selective Oxidation, (Studies in Sur-
face Science and Catalysis, Vol 55), Elsevier, Amsterdam, 1990, p 767
12 T Ono, Y Nakagawa, H Miyata and Y Kubokawa, Bull Chem Soc Jpn., 57 (1984) 1205
13 A.J van Hengstum, Thesis, Twente University of Technology, Enschede, 1984, p 35
14 Y.C.I,iu, G.L Griffin, S.S Chan and I.E Wachs, J Catal.,94 (1985) 108
15 M Akimoto and E Echigoya, J Catal., 31 (1973) 278
16 M Ai, Bull Soc Chem Jpn., 49 (1976) 278
17 D Soung Kim, Y Kurusu, I.E Wachs, F.D Hardcastle and K Segawa, J Catal., 120 (1989)
325
Trang 10288 K Briickman et al./Appl Catal A 96 (1993) 279-288
18 Y Matsuoka, M Niwa and Y Murakami, J Phys Chem., 94 (1990) 1477
19 C Louis, J.M Tatibouiit and M Che, J Catal., 109 (1988) 354
20 A.J van Hengstum, J.G van Ommen, H Bosch and P.J Gellings, Proc 8th Int Congre~
on Catalysis, Berlin, Germany, Vol 4, 1984, p 297
21 F ForT~ti, E Tronconi, G Busca and P Tittarelli, Catal Today, 1 ( 1987 ) 209
22 M Niwa, H Yamada and Y Murakami, J Catal., 134 ( 1992 ) 331
23 J.M Tatibou~it and J.E Germain, J Catal., 72 {1981 ) 375
24 M Ai, Bull Chem Soc Jpn., 49 (1976) 328,
25 M Ai, Bull Chem Soc Jpn.,53 (1980) 3391
26 B Grzybowska-Swierkosz, Mater Chem Phys., 17 (1987) 121
27 C.D Wagner, W.M Riggs, L.E Davies, J.F Moulder and G.E Mullenberg (Editors), Hand- book of X-Ray Photoelectron Spectroscopy, Perkin-Elmer, Eden Prairie, MN, 1979
28 J Stoch, Poverkhnost (in Russian), 10 (1987) 68
29 T Fransen, P.C van Berge and P Mars, in B Delmon, P.A Jacobs and G Poncelet, (Edi- tors), Preparation of Catalyst.s I, (Studies in Surface Science and Catalysis, Vol 1 ), Elsevier, Amsterdam, 1976, p 405
30 J.M Tatibou~t, Bull, Soc Chim Fr., 1 (1986) 18
31 E Serwicka, unpublished results
32 J.N Allison and W.A Goddard |II, J Catal., 92 (1985) 127
33 J.M Tatibou~t, C.R Acad Sci Paris, Serie II, 297 (1983) 703