Dispersion and catalytic activity of molybdenum supported on Y zeolites

Dispersion and catalytic activity of molybdenum supported on Y zeolites

Journal of Molecular Catalyszs, 40 (1987) 83 83 - 92 DISPERSION AND CATALYTIC ACTIVITY OF MOLYBDENUM SUPPORTED ON Y ZEOLITES TAKAYUKI Department K...

690KB Sizes 2 Downloads 35 Views

Recommend Documents

No documents
Journal of Molecular Catalyszs, 40 (1987)

83

83 - 92

DISPERSION AND CATALYTIC ACTIVITY OF MOLYBDENUM SUPPORTED ON Y ZEOLITES TAKAYUKI Department

KOMATSU and TATSUAKI YASHIMA* of Chemzstry, Tokyo Znstztute of Technology,

(Received May 5,1986,

Meguro-ku, Tokyo (Japan)

accepted October 24, 1986)

Summary The dispersion of molybdenum m Mo/HNa-Y, prepared from MOM and HNa-Y zeohtes, and its effect on the catalytic activity were investigated From the amount of oxygen chemisorbed at ‘77 K, the number of surface MO atoms was estunated When HNa-Y with a high degree of proton exchange was used as a support, MO species were dispersed to form a monolayer or less on zeohte, that is, the dispersion was -100% When the degree of proton exchange m Y zeohte was low, Moo species migrated to form aggregates, which resulted m a lower dispersion The activity of Mo/HNa-Y catalysts for the hydrogenation of ethene varied with the dispersion as well as the oxidation state of molybdenum The active site was thought to be MO’, and that its aggregation caused the mcrease m the activity In the case of the metathesis of propene, it was found that the active site was the shghtly aggregated Moo species

Introduction Reduced Mo03/A1203 catalysts are active for various reactions, such as the metathesis of propene [ 1,2], the hydrogenation of ethene [3] or propene [4,5], the polymerization of ethene [6], the hydrogenolysis of propane [ 71, the hydrodesulfurization of thiophenlc compounds [ 81, and so on The active molybdenum species for these reactions have an oxidation number below +6, smce the activity of the reduced Mo0s/A1203 is higher than that of Mo6+ species m calcined catalysts. However, clear and systematic conclusions as to the relation between the oxidation state of molybdenum and its activity for these reactions have never been obtained. In order to clarify this relation, it is necessary to determme the dispersion of molybdenum as well as to prepare MO species which have defmite oxidation states

*Author to whom correspondence should be addressed 0304-5102/87/$3,50

0 Elsevler Sequola/F’rmted m The Netherlands

84

The dispersion of molybdenum m reduced Mo0,/A120, has been studied quahtatlvely by Raman [9, lo] and UV [ 10,111 spectroscoples, which suggest the existence of Moo3 and Al,(MoO&, On the other hand, the chemlsorptlon of oxygen can provide quantitative mformatlon about the dispersion of molybdenum. Reduction of the supported MOO, leads to the formation of 0, chemlsorptlon sites. These chemlsorptlon sites are thought to be the coordmatlvely unsaturated &es (CUS) on Mo4+ [4,12] or parts of these CUS [ 131. Parekh and Weller [ 141 have determined the surface area of MO species supported on alumma from the amount of chemlsorbed oxygen, regardmg the MO species after reduction as crystalline MOO* Only about a quarter of the alumma surface was covered with the MO species, even if the content of MOO, exceeded monolayer coverage. They have suggested that the dispersion of molybdenum is low. Other mvestlgators [4,12,14,15] have reported that the amount of chemlsorbed oxygen for the reduced Mo0,/A1203 corresponds to an O/MO atomic ratio of 0 21 0.34 These results also suggest the low dispersion of molybdenum. Recently, a new class of supported molybdenum catalysts has been prepared from molybdenum complexes, such as MOM [16 - 181, Mo(~T-C+H&~ [ 19,201, Mo~(x-C~H~)~ [21], etc. In these catalysts, high and homogeneous dispersion of molybdenum can be achieved because these catalysts are prepared through chemical interaction between the complex molecule and the surface of the support In addition, the oxldatlon state of molybdenum can easily be controlled over a wide range In the case of the Mo03/Al,03 catalyst, however, MO species with an oxidation number of less than +4 cannot easily be obtamed by ordinary reduction, that is, hydrogen treatment at about 800 K. Thus the new type of catalysts has the advantage m clarlfymg the relation between the oxldatlon state of molybdenum and its catalytic activity compared with conventional Mo03/Al,03 catalysts We have already prepared the molybdenum supported on zeohte from MOM and HNa-Y zeohte and reported that the average oxldatlon number of molybdenum can be controlled from +0 3 to +3 8 by varying the proton concentration of HNa-Y and the decomposltlon temperature of MOM adsorbed on HNa-Y [22]. The MO species urlth the low oxldatlon number, probably MO+, 1s found to be the active site for the polymenzatlon of ethene [22], and the Moo species to be the active site for the metathesis of propene [23]. The dispersion of molybdenum 1s thought to be very high when HNa-Y with a high degree of proton exchange 1s used as a support [23] Therefore when Mo/HNa-Y catalysts are used it may be possible to elucidate the catalytic properties of molybdenum over the mde range of its oxldatlon state and dispersion. In this study, we have measured the amount of chemlsorbed oxygen and determined the dlsperslon of molybdenum m Mo/HNa-Y prepared from Mo(CO), and HNa-Y with various degrees of proton exchange In addition, we have tried to clanfy the effect of not only the oxldatlon state but also the dispersion of molybdenum on the catalytic actlvlty for the hydrogenation of ethene and the metathesis of propene

85

Expernnental Catalyst preparation Na-Y zeohtes (Toyo Soda Ind Co., Lot Y-30, &/Al = 2 88) were treated with 0.05 N or 0 5 N NH&l aqueous solution at 298 K or 336 K to form NH4Na-Y. After calcmation at 743 K m au, H(x)Na-Y was obtamed, where x is the percent degree of proton exchange In this study, we prepared the [email protected])Na-Y zeohtes with x = 14 - 82%. Na-Y was washed with pure water, followed by drying at 393 K m an The preparation of Mo/HNa-Y zeohtes was designed based on the procedure reported by Gallezot et al [17]. A specific amount of HNa-Y placed m a quartz tube was heated m vacua at various temperatures to yield dehydrated HNa-Y The desired amount of MOM which corresponded to the amount of the saturated adsorption, that is, two molecules per supercage of HNa-Y [24], was added to the dehydrated HNa-Y under argon atmosphere. After the argon gas was pumped out for 20 s, the tube was put m a thermostatted oven at 333 K and allowed to stand for 15 h to adsorb Mo(CO), on HNa-Y Mo/HNa-Y was obtamed by heating Mo(CO),/HNa-Y m vacua at 573 K. The content of molybdenum m the Mo/HNa-Y was determined by atomic absorption spectroscopy Oxygen chemlsorp taon The amount of chemisorbed oxygen was measured with a static system. A specific amount of HNa-Y (0.4 g) was placed m a quartz tube, and Mo/HNa-Y was prepared m the same manner as described above. At a certam temperature, a known amount of oxygen was mtroduced onto Mo/HNa-Y and the first adsorption isotherm was obtained The Mo/HNa-Y was usually evacuated for 30 mm at the same temperature When the adsorption was carried out at 77 K, the evacuation temperature was 195 K. The second isotherm was then obtamed at the same adsorption temperature. The amount of chemisorbed oxygen was calculated from the difference between the two isotherms and was expressed m terms of the number of adsorbed oxygen atoms per MO atom (O/MO) Catalytic reaction The hydrogenation of ethene and the metathesis of propene on MO/ HNa-Y were carried out at 274 K with a closed circulation system of 230 ml dead volume. In the case of the hydrogenation of ethene, the Mo/HNa-Y catalyst was prepared zn situ from 0.040 g of HNa-Y m the same manner as described above. Traces of oxygen were removed from hydrogen by passage through a trap of Mn/Si02. The mitral pressures of both ethene and hydrogen were 135 torr The reaction products were analyzed by gas chromatography The extent of the reaction was momtored by measurmg the pressure of the system

86

In the case of the metatheses of propene, the Mo/HNa-Y catalyst was prepared m sztu from 0.20 g of HNa-Y The mitral pressure of propene was 190 torr The reaction products were analyzed by gas chromatography [23].

Results and discussion The chemlsorptron of oxygen has been applied to certain Moos supported catalysts m order to gam mformatlon about the dispersion of molybdenum. In the case of MoOa/Al,Os catalysts, it has been reported that the manner of oxygen adsorption differs slgmficantly depending on the adsorptron temperature [ 141. We have already reported [23] that the manner of oxygen adsorption for Mo/HNa-Y differs when the adsorption is carried out at 77 K and 298 K In order to determine the effect of the adsorption temperature clearly, we measured the amount of chemlsorbed oxygen on Mo/HNa-Y at varrous temperatures The results are shown m Fig. 1. Irrespective of the temperature, the amount of chemlsorbed oxygen on Mo/H(74)Na-Y was nearly constant (O/MO = 1.2) Assuming that one oxygen atom is chemlsorbed on one MO atom, the dispersion of MO corresponds to -lOO%, though parts of the oxygen seemed to be chemlsorbed in the ratio of O/MO 1. It has been reported for the MoOs/A1203 catalysts that one oxygen atom 1s chemlsorbed on one CUS [4], that is, the stoichiometry of the chemlsorptlon IS O/MO = 1 In the Mo0s/A1203 catalysts, the oxldatlon numbers of MO are mainly +4 and +5, so that this stolchlometry of chemisorption cannot be applied directly to Mo/HNa-Y which contains MO species with oxldatlon numbers lower than +4 Yermakov and Kuznetsov [19] and Iwasawa and Ogasawara [25] have reported that on the catalysts prepared from x-ally1 complexes, oxygen is consumed on Mo2+ species m the ratio of O/MO = 1 at room temperature In the case of the MO metal, however, oxygen 1s adsorbed on the (100) face m the ratio O/MO = 1 5 at room temperature [ 261 We have already studied [23] the dispersion of MO with UV diffuse reflectance spectroscopy. The MO species m the oxidized catalyst, Mo6+/ H(65)Na-Y, consist almost exclusively of tetrahedrally coordinated MO We have concluded that the MO species m this catalyst are highly dispersed on zeohte, and correspond to monolayer coverage or less It seems that the redispersion of MO atoms m aggregates does not occur durmg the oxidation of Mo/H(65)Na-Y Therefore, the dlsperslon of MO before oxidation is nearly 100% The MO species supported on H(74)Na-Y which has a degree of proton exchange slmllar to H(65)Na-Y probably has a dispersion of -100% Therefore, we conclude that the O/MO stoichiometry is 1 for the chemuorption of oxygen at 77 K In the case of Mo/Na-Y, Frg 1 clearly shows that a lower amount of oxygen is chemisorbed compared with Mo/H(74)Na-Y This indicates the low dispersion of molybdenum m Mo/Na-Y The O/MO ratio increased with mcreasmg adsorption temperature As mentioned previously [23], when a

87 15

o,5_/ 10

10 e $

P 2 05

0

1

I

100

200 TemPerature/K

Fig 1 Effect Mo/H(74)Na-Y

I 300

0’

-

0

!_

20

I

I

I

40

60

80

Degree of proton exchange/X

of adsorptlon temperature on the amount of oxygen chemlsorbed on (0) and Mo/Na-Y (0) Dehydration temperature of zeohtes was 573 K

Fig 2 Effect of the degree of proton exchange m Y zeohtes on the amount of chemlsorbed oxygen Dehydration temperatures of zeohtes were 573 K (0) and 773 K (0) Oxygen was adsorbed at 77 K

certam amount of oxygen ISintroduced onto Mo/Na-Y at 298 K, the oxygen uptake increases gradually for more than 40 h and the O/MO ratro becomes more than twice that measured at 77 K These results suggest that the oxldatlon of MO m bulk phase occurs at higher temperatures Consequently, m order to determine the number of surface MO atoms, It is better to adsorb oxygen at 77 K. The low O/MO ratios for Mo/Na-Y imply that it 1s possible to control the dlsperslon of MO by varymg the proton concentration of HNa-Y. Figure 2 shows the amount of oxygen chemlsorbed at 77 K on Mo/HNa-Y with various degrees of proton exchange m Y zeohtes. When the degree of proton exchange was 65 or 74%, the O/MO ratio was about 1, which indicates the high dispersion of MO. However, the O/MO ratio decreased with decreasmg degree of proton exchange below 65%. Thus IS attnbuted to the aggregation of MO. In addition, the dehydration temperature of HNa-Y had an effect on the amount of chemlsorbed oxygen, that IS, the O/MO ratio was low when the dehydration temperature was high Dehydration at 773 K causes a decrease m the concentration of the OH groups of HNa-Y, so that It corresponds to lowermg the degree of proton exchange. Thus the lower proton concentration m zeohtes leads to a lower dlsperslon of MO, probably because the free Moo atoms which are formed by the decomposition of Mo(C0)6 and are not oxidized by the OH groups can migrate to form aggregates The existence of Moo species m Mo/Na-Y has been reported from the results of m sdu XPS spectra [24]. We have already measured [23] the UV spectra of MO species m the oxldlzed catalyst, Mo6+/Na-Y, and found that most of them have the octahedral coordmatlon Consequently, the MO species are aggregated to form bulk oxide, Moos. This result 1s consistent with the

88 TABLE 1 Amounts of oxygen chemlsorbed on various molybdenum supported catalysts Catalyst

Amount of chemlsorbed oxygen (O/MO)

Reference

Mo/[email protected])Na-Ya reduced Mo6+/H(74)Na-Yb reduced MoOXJAI~O~ reduced MoOs/S102 reduced MoOB/CeOz

0 78 0 38 021-034 0 26 0 29

this work this work [4,12,14,15]

aAON = +2 3 bMo/H(74)Na-Y

was oxldlzed at 673 K followed by reduction at 773 K,

1271 [271

AON = +2 6

results of the oxygen chemlsorptlon. Thus we can control the dispersion of MO as well as the oxldatlon state of MO by changmg the proton concentration of HNa-Y zeohtes. Table 1 shows the amount of chemlsorbed oxygen (O/MO) measured with two Mo/HNa-Y samples m which MO species with higher oxldatlon states are contamed The O/MO ratios measured with Moo3 supported catalysts [4,12,14,15,27] are also shown m Table 1. In the case of MO/ H(74)Na-Y oxidized to Mo6+ and subsequently reduced with Hz at 773 K, the O/MO ratlo was rather low compared urlth Mo/H(82)Na-Y, though two samples had nearly the same AON of MO Therefore, the aggregation might occur during the oxldatlon and/or the reduction at higher temperatures The supported MOO, catalysts exhibited lower O/MO ratios than the reduced Mo6+/H(74)Na-Y, which indicates a more extensive aggregation of MO species m the former catalysts. It was found that the Mo/HNa-Y system has an advantage m obtammg a high dlsperslon of MO compared with the supported Moo3 catalysts. The hydrogenation of ethene was carried out with various kinds of Mo/HNa-Y m order to clmfy the effect of the oxldatlon state and dispersion of MO on the catalytic actlvlty When a 1 1 mixture of ethene and hydrogen was introduced onto Mo/HNa-Y at 274 K, the pressure decreased lmmedlately to approach half of Its mrtlal value. After the decrease m the pressure became negligible, trace amounts of propane and methane were detected m the gas phase m addition to ethane However, the total amount of these byproducts corresponded to 0 4 mol% of the amount of produced ethane, so that the side reactlons can be ignored. Therefore, the rate of the pressure decrease represents the rate of the ethene hydrogenation We measured the hydrogenation rates with vmous mltlal pressures of one reactant, while keepmg the pressure of the other reactant constant The results indicated that the hydrogenation of ethene on Mo/HNa-Y was fast order with respect to the pressure of hydrogen and -zero order with respect to that of ethene Therefore, the actlvlty of Mo/HNa-Y 1s described m terms of the

89

first-order rate constant for the pressure of hydrogen determined m the m&ml stage of the reactlon (1.5 mm of reactIon) Figure 3 shows the relation between the activity and the average oxldatlon number (AON) of MO. Various Mo/HNa-Y (AON = +0.4 - +2.8) were prepared by changmg both the degree of proton exchange m Y zeohtes and their dehydration temperatures, as reported previously [23] The activity mcreased with decreasmg AON of MO when H(36)Na-Y, H(65)Na-Y or H(74)Na-Y was used as a support After oxygen ~eatment at 77 K, which Increased the AON to +3.6, the a&n&y of Mo/H(74)Na”Y decreased to nearly zero These results Imply that the MO species m low oxtdatlon state are the active sites for the ethene hydrogenation Several studies m relation to the active site for ethene hydrogenation have been carried out with supported MO catalysts In the case of Moos/ A1203, Lombard0 et crl, 131 have found that the reduction of MO causes higher activity. They considered the CUS on Mo4+ to be the actrve site. In the case of MoO~/~-~O~, however, when MO species were reduced to an AON of less than +4, the actn&y increased several fold higher than that of the Mo4+ species 1283. In the case of catalysts prepared from the ~-ally1 complex of MO, Mo2’ was the most active among the MO species with oxldatlon numbers of +2 - +6 [29]. Brenner [30] has reported for catalysts delnved from Mo(CO), and alumma that the Moo subcarbonyl species is active, but that further decomposltlon accompamed by oxldatlon of MO lowers the actlvlty. In the above studies, the lower oxldatlon state of MO always provides the higher a&v&y for the ethene hydrogenation This tendency ISconsistent with our results w&b the Mo/HNa-Y catalysts When Na-Y or H(14)Na-Y was used as a support, the actn&y was lower than that of Mo/H(36)Na-Y and Mo/H(74)Na-Y, which had AONs of MO m a sumlar range (+0.3 - +0 6) The MO species m Mo/Na-Y and MO/ H(14)Na-Y exhibited a low dispersion of MO, as shown m Fig. 2. To clarify the effect of the dispersion of MO, the actlvky per surface MO atom was

0

1

2

3

Average oxicJotionnmber

Fig 3 Change m actlvlty for ethene hydrogenation with average oxldatlon number of MO Percent degrees of proton exchange m Y zeohtes were 0 (m), 14 (a), 36 (*), 65 (0) and 74 (A) Mo~H(74)Na-Y treated wkh oxygen at 77 K (x) was also Included

90

calculated from the amount of chemlsorbed oxygen (O/MO) at 77 K The results are shown m Fig. 4 The plots for the activity were obviously dlvlded mto two groups. The lower activity group contamed Mo/H(65)Na-Y and Mo/H(74)Na-Y, where the dispersion of MO was about 100%. The higher activity group contamed Mo/Na-Y, Mo/H(14)Na-Y and Mo/H(36)Na-Y. In these catalysts, the dispersion of MO was low and the formation of Moo aggregates was suggested, as mentioned above. It is concluded that the active sites for ethene hydrogenation are the Moo species, and that when the Moo species form metallic aggregates, the actlvlty of the surface Moo atoms becomes higher. We have studied previously the catalytic actnnty of Mo/HNa-Y for the metathesis of propene [23]. The Mo/HNa-Y catalysts were prepared by the same procedure as that used m the hydrogenation of ethene. In the case of Mo/H(36)Na-Y, Mo/H(65)Na-Y and Mo/H(74)Na-Y, the activity for the metathesis increased wth decreasmg AON of MO. It was suggested that the oxldatlon number of the most active MO species for the propene metathesis was less than +4, probably 0 We have also observed that the activity is not m proportion to the amount of chemlsorbed oxygen. In order to determine the effect of the dispersion of MO on the activity for the metathesis of propene, we calculated the activity per surface MO atom m the same manner as for the hydrogenation of ethene The results are shown m Fig 5. In the [email protected] of low AON, Mo/H(36)Na-Y exhibited the highest actlvlty. The MO species m Mo/H(36)Na-Y have a medium dispersion (O/MO = 0.6 - 0 7) compared with those m Mo/H(74)Na-Y (O/MO = 1) and Mo/Na-Y (O/MO = 0 2 - 0 3). Consequently, we suggest that the activity for the propene metathesis mcreases with the formation of small Moo aggregates, but that the activity decreases with the formation of large aggregates by the further aggregation of Moo The propene metathesis is believed to proceed through

1

I

I

1

2

3

Average oxldotion number

1

2

3

Average oxidation number

Fig 4 Actlvlty per surface MO atom for ethene hydrogenation (symbols as m Fig 3) The amount of chemlsorbed oxygen (O/MO) was measured at 77 K Fig 5 Actlvlty per surface MO atom for propene metathesis (symbols as m Fig 3) The amount of chemlsorbed oxygen (O/MO) was measured at 77 K

91

a metallocyclobutane mtermedlate [31]. Therefore, the lower metathesis actlvltles of Mo/H(14)Na-Y and Mo/Na-Y probably result from sterlc hmdrance caused by large Moo aggregates located m the supercage of zeohtes. We conclude that not only the oxldatlon state but also the dispersion of molybdenum m Mo/HNa-Y prepared from Mo(CO), and HNa-Y zeohtes can be controlled by varying the proton concentration of HNa-Y The active site for the hydrogenation of ethene 1s MO’, and the aggregation of Moo leads to an mcrease m the actlvlty In the case of the metathesis of propene, the active site 1sthe slightly aggregated Moo species Acknowledgements Financial support for this work by the Asahl Glass Foundation for Industrial Technology 1sgratefully acknowledged. References 1 N Glordano, M Padovan, A Vaghl, J C J Bart and A Castellan, J Catal, 38 (1975) 1 2 J R Hardee and J W Hlghtower, J Catal, 83 (1983) 182 3 E A Lombardo, M Houalla and W K Hall, J Catal, 51 (1978) 256 4 W S Mlllman and W K Hall, J Catal, 59 (1979) 311 5 E A Lombardo, M L Jacono and W K Hall, J Catal, 64 (1980) 150 6 R V Morris, D R Waywell and J W. Shepard, J Less-Common Metals, 36 (1974) 395 7 R Nakamura, D Pooch, R G Bowman and R. L Burwell, Jr, J Catal, 93 (1985) 388 8 F E Massoth and G MurahDhar, Proc 4th Int Conf Chem Uses of Molybdenum, Golden CO, 1982, p 343 9 H. Jezlorowskl and H Knozmger, J Phys Chem , 83 (1979) 1166 10 N. Glordano, J C J. Bart, A Vaghl, A Castellan and G Martmottl, J Catai, 36 (1975) 81 11 L Wang and W K. Hall, J Catal, 77 (1982) 232 12 J Valyon and W K Hall, J CataZ, 84 (1983) 216 13 N K Nag, J Catal, 92 (1985) 432 14 B S Parekh and S W Weller, J Catal, 47 (1977) 100. 15 H -C Llu and S W Weller, J Catal, 66 (1980) 65 16 R G Bowman and R L Burwell, Jr , J Catal, 63 (1980) 463 17 P Gallezot, G Coudurler, M Prlmet and B Imehk, ACS Symp Ser, No 40 (1977) 144 18 S Abdo and R F Howe, J Phys Chem , 87 (1983) 1713, zbtd , 87 (1983) 1722. 19 Y I Yermakov and B N Kuznetsov, Prepr 2nd Japan-Souzet Catalyszs Semznar, Tokyo, 1973, p. 65. 20 Y Iwasawa and S Ogasawara, J Chem Sot , Faraday Trans 1, 74 (1978) 2968 21 Y Iwasawa, Y Sato and H Kuroda, J Catal, 82 (1983) 289, Y Iwasawa and M Yamaglshl, zbzd 82 (1983) 373. 22 T Yashnna, T Komatsu and S Namba, Proc 4th Znt Conf Chem Uses of Molybdenum, Golden CO, 1982, p 274 23 T Komatsu, S Namba and T Yashlma, Prepr Intern Symp Zeolzte Catal, Slofok, 1985, p. 251

92

24 T Komatsu, S Namba, T Yashnna, K Domen and T Omshl, J Mel Catal, 33 (1985) 345. 25 Y. Iwasawa and S. Ogasawara, J Chem Sot., Faraday Trans 1, 75 (1979) 1465. 26 R. Rlwan, C Gulllot and J Palgne, Surf Scz , 47 (1975) 183. 27 G. MuraliDhar, B E Concha, G L Bartholomew and C H Bartholomew, J Catal, 89 (1984) 274. 28 K Tanaka, K Mlyahara and K Tanaka, J Mel Catal, 15 (1982) 133,Proc 7th Znt Congr Catal, Tokyo, 1980, p. 1318 29 Y Iwasawa, M Yamaglshl and S Ogasawara, J Chem Sot, Chem Commun, (1980) 871 30 A Brenner, J Mol Catal, 5 (1979) 157. 31 J L Herlsson and Y Chauvm, Makromol Chem , 141 (1970) 161