Tungsten and Tungsten Carbide Supported on Activated Carbon

Filter By. Menu Switch Switch View Sections. Previous Results. Next Results. ACS2GO © 2019. ← → → ←. Hide Menu Back ...
0 downloads 0 Views 73KB Size
1752

Langmuir 2001, 17, 1752-1756

Tungsten and Tungsten Carbide Supported on Activated Carbon: Surface Structures and Performance for Ethylene Hydrogenation C. Moreno-Castilla,*,† M. A. Alvarez-Merino,†,‡ F. Carrasco-Marı´n,† and J. L. G. Fierro§ Departamento de Quı´mica Inorga´ nica, Facultad de Ciencias, Universidad de Granada, 18071 Granada, Spain, and Instituto de Cata´ lisis y Petroleoquı´mica, CSIC, Cantoblanco, 28049 Madrid, Spain Received September 26, 2000. In Final Form: November 27, 2000 Tungsten catalysts supported on an activated carbon were prepared, with different metal loading, by using either ammonium tungstate or tungsten hexacarbonyl as metal precursor compounds. The catalysts so prepared were heat treated in H2 flow at 748 and 973 K or in He flow at 1223 K. The heat-treated catalysts were characterized by N2 adsorption at 77 K and mercury porosimetry, to determine their surface area and porosity, and by XPS and XRD, to evaluate the chemical state and dispersion of the active phase. These catalysts were tested in the hydrogenation reaction of ethene. The treatments in H2 yielded an active phase composed of metallic tungsten and tungsten trioxide, whereas the treatment in He generated a mixture of metallic tungsten, tungsten carbide, and tungsten trioxide. The catalysts containing tungsten carbide were the most active in the hydrogenation of ethene, even at a temperature as low as 273 K, although their rate of deactivation was dependent on reaction temperature; the lower the reaction temperature, the higher the rate of deactivation.

Introduction 1

In a previous work, we studied the characteristics of different tungsten catalysts supported on activated carbon after their heat treatment, between 623 and 1223 K, in He flow. One of the results obtained was that the O/W atomic ratio dropped below 3 and that different nonstoichiometric oxides appeared as the temperature was raised. At 1223 K, a mixture of tungsten, tungsten carbide, and tungsten trioxide was detected. This result is interesting and shows the important role that carbon materials can play as supports for metals that can yield carbides, by moderate heat treatments, since this could be a method to obtain metal carbides dispersed on a high surface area support. The catalytic properties of tungsten carbide have attracted interest since Levy and Boudart2 drew some parallels, from a qualitative point of view, between WC and Pt. Thus, the use of WC as well as other transition metal carbides as heterogeneous catalysts has grown.2-17 * Corresponding author. Fax: 34-958-248526. E-mail: cmoreno@ ugr.es. † Universidad de Granada. ‡ Present address: Departamento de Quı´mica Inorga ´ nica y Orga´nica, Facultad de Ciencias, Universidad de Jae´n, Jae´n, Spain. § CSIC. (1) Alvarez-Merino, M. A.; Carrasco-Marı´n, F.; Fierro, J. L. G.; Moreno-Castilla, C. J. Catal. 2000, 192, 363. (2) Levy, R. B.; Boudart, M. Science 1973, 181, 547. (3) Kojima, I.; Miyazaki, E.; Inoue, Y.; Yasumoru, I. J. Catal. 1979, 59, 472. (4) Kojima, I.; Miyazaki, E.; Inoue, Y.; Yasumoru, I. J. Catal. 1982, 73, 128. (5) Vidick, B.; Lemaitre, J.; Leclercq, L. J. Catal. 1986, 99, 439. (6) Oades, R. D.; Morris, S. R.; Moyes, R. B. Catal. Today 1991, 10, 379. (7) Park, K. Y.; Seo, W. K.; Lee, J. S. Catal. Lett. 1991, 11, 349. (8) Woo, H. C.; Park, K. Y.; Kim, Y. G.; Nam, I. S.; Chung, J. S.; Lee, J. S. Appl. Catal. 1991, 75, 267. (9) Qin, D. Y.; Gao, Z. J. Nat. Gas Chem. 1992, 1, 30. (10) Oyama, S. T. Catal. Today 1992, 15, 79. (11) Iglesia, E.; Ribeiro, F. H.; Boudart, M.; Baumgantner, J. E. Catal. Today 1992, 15, 307.

Several methods to prepare WC are proposed in the literature,18-25 and the most used involve reaction of the solid precursor, either as metal or oxide, with a gas phase containing the reducing and /or the carburizing agents. In general, these methods yield WC samples with a small surface area and with polymeric carbon deposited on their surface. The main objective of this work was to determine the chemical structures of the tungsten species supported on activated carbon and then to investigate the performance of these structures in the hydrogenation reaction of ethylene in which they were found to be active.3,4 These catalysts were obtained by heat treatment of the fresh supported catalysts at 1223 K in He flow. Their behavior in the above reaction was also compared to that of metallic tungsten supported on activated carbons, which were (12) Keller, V.; Wehrer, P.; Garin, F.; Ducros, R.; Maire, G. J. Catal. 1995, 153, 9. (13) Ramanathan, S.; Oyama, S. T. J. Phys. Chem. 1995, 99, 16365. (14) Leclercq, L.; Almazouari, A.; Dufour, M.; Leclercq, G. In The Chemistry of Transition Metal Carbides and Nitrides; Oyama, S. T., Ed.; Blackie: Gasgow, 1996; p 345. (15) Keller, V.; Wehrer, P.; Garin, F.; Ducros, R.; Maire, G. J. Catal. 1997, 166, 125. (16) Garin, F.; Keller, V.; Ducros, R.; Muller, A.; Maire, G. J. Catal. 1997, 166, 136. (17) Claridge, J. B.; York, A. P. E.; Brungs, A. J.; Ma´rquez-Alvarez, C.; Sloan, J.; Chi Tsang, S.; Green, M. L. H. J. Catal. 1998, 180, 85. (18) Boudart, M.; Oyama, S. T.; Leclercq, L. Stud. Surf. Sci. Catal. 1981, 7A, 578. (19) Oyama, S. T.; Haller, G. Catalysis. Specialist Periodicals Report; Bond, G. L., Wedd, G., Eds.; The Chemical Society: London, 1981. (20) Leclercq, L.; Provost, M.; Pastor, H.; Grimblot, J.; Hardy, A. M.; Gengembre, L.; Leclercq, G. J. Catal. 1985, 117, 384. (21) Ledoux, M. J.; Phan-Huu, C.; Guille, J.; Dunlop, M.; Hantzer, S.; Martin, S.; Weibel, M. Catal. Today 1992, 15, 263. (22) Katrid, A.; Hemming, F.; Hilaire, L.; Wehrer, P.; Maire, G. J. Electron Spectrosc. 1994, 68, 589. (23) Leclercq, G.; Kamal, M.; Lamonier, J. F.; Feigenbaum, L.; Malfoy, P.; Leclercq, L. Appl. Catal., A 1995, 121, 169. (24) Leclercq, G.; Kamal, M.; Girandon, J. M.; Devassine, P.; Feigenbaum, L.; Leclercq, L.; Frennet, A.; Bastin, J. M.; Lo¨fberg, A.; Decker, S.; Dufour, M. J. Catal. 1996, 158, 142. (25) Decker, S.; Lo¨fberg, A.; Bastin, J. M.; Frennet, A.Catal. Lett. 1997, 44, 229.

10.1021/la001367k CCC: $20.00 © 2001 American Chemical Society Published on Web 02/10/2001

Tungsten Catalysts Supported on Activated Carbon

Langmuir, Vol. 17, No. 5, 2001 1753

obtained from the same fresh catalysts but by heat treatment in H2 flow. Experimental Section The activated carbon used was obtained from almond shells by N2 carbonization and steam activation as reported elsewhere.1 This sample will be referred to in the text as S. The catalysts were prepared either from ammonium tungstate or tungsten hexacarbonyl precursors as explained in detail elsewhere.1 The catalysts prepared from ammonium tungstate will be referred to in the text as W, and those prepared from the hexacarbonyl as HW. In both cases the tungsten content of the catalysts is indicated by the number following the letters W or HW. Before characterization, the catalysts were heated either in a H2 flow, at 2 K min-1, up to 748 and 973 K or in a He flow up to 1223 K. In all cases, the maximum temperature was held for 4 h. Both the support and some selected catalysts were characterized, to determine their surface area and pore texture, by N2 adsorption at 77 K and mercury porosimetry up to 4190 kg cm-2. In this case a mercury porosimeter, Quantachrome model Autoscan 60, was used, which allowed the volume of pores wider than 3.6 nm to be measured. X-ray diffraction (XRD) patterns were recorded with a Phillips PW1710 diffractometer using Cu KR radiation. The JCPDS files were used to assign the different diffraction peaks observed. Mean crystallite size was obtained, in some cases, by applying Sherrer’s equation to the half-height of the main diffraction peak. The line broadning of this peak was corrected with polycrystalline graphite. X-ray photoelectron spectroscopy measurements (XPS) were carried out with a VG Escalab 200 R spectrometer with Mg KR source (hν ) 1253.6 eV) and hemispherical electron analyzer. Prior to analysis, the samples were heated in situ at temperatures equal to or below 748 K in He flow. When the temperature of the heat treatment was higher than 748 K, the treatment was carried out in a quartz reactor under He flow. Once the treatment had finished, the sample was cooled to room temperature under He flow, impregnated with n-octane, to avoid its contact with air, and transferred to the pretreatment chamber of the XPS instrument. After the in situ heat treatment, all the samples were evacuated at high vacuum and room temperature and then introduced in the analysis chamber. A base pressure of 10-9 Torr was maintained during data acquisition. Survey and multiregion spectra were recorded at C1s, O1s, and W4f photoelectron peaks. Each spectral region of the photoelectron of interest was scanned several times to obtain good signal-to-noise ratios. The spectra obtained, once the background signal was corrected, were fitted to Lorentzian and Gaussian curves to obtain the number of components, the position of the peaks, and their areas. The carbon 1s electron binding energy corresponding to graphitic carbon was referenced at 284.9 eV for calibration.26 The hydrogenation reaction of ethylene was performed in a quartz microreactor at atmospheric pressure with 0.3 g of catalyst, which was previously heat treated in situ, either in H2 or He, for 4 h. Products were separated by gas chromatography using a Chromosorb 102 column and analyzed by flame ionization detector. A flow (55 mL min-1) of the reactant mixture, which consisted of C2H4-H2-He with proportions of 0.01:0.09:0.90 by volume, was used. The reaction temperature ranged between 273 and 623 K.

Results and Discussion Surface Area and Porosity of the Supported Catalysts. Surface area and porosity of the support and supported catalysts were studied by N2 adsorption at 77 K and by mercury intrusion porosimetry. The BET equation was applied to N2 adsorption isotherms thus allowing the N2 surface area (SN2) to be calculated. The volume of pores with a diameter between 3.6 and 50 nm (V2) and wider than 50 nm (V3) were obtained by mercury porosimetry. The V2 volume corresponds to the mesopore volume, although, in fact, the latter includes pores with a diameter between 2 and 50 nm.27 V3 corresponds to the macropore volume. The changes in surface area and

Figure 1. Variation of SN2 with tungsten content of catalysts treated at 748 K in H2.

Figure 2. Variation of meso- and macropore volumes (V2 and V3, respectively) with tungsten content of catalysts treated at 748 K in H2. Table 1. Surface Area and Porosity of Supported Catalysts after Their Heat Treatment in He Flow at 1223 K for 4 h catal

SN2, m2 g-1

V2, cm3 g-1

V3, cm3 g-1

W4.8 W23.1 HW6.1

792 ( 7 555 ( 5 898 ( 11

0.13 0.07 0.15

0.25 0.21 0.29

porosity of the catalysts, when heat treated at 748 K in H2, are depicted in Figures1 and 2, respectively. Figure 1 shows a linear decrease of SN2 with increasing tungsten content. The decrease in surface area calculated from the slope of the straight line was 0.20 m2 µmol-1 of W. Pore volumes V2 and V3 also decreased, to a lesser extent, with increasing W content, as shown in Figure 2. Similar results were obtained when the catalysts were heat treated in H2 at 978 K; in this case the decrease in surface area with tungsten content was 0.21 m2 µmol-1 of W. Results obtained with some of the catalysts heat treated at 1223 K in He flow are compiled in Table 1. For catalysts from series W the surface area and porosity decreased with the increase in tungsten content. The decrease in these parameters with regards to the support was lower for sample HW6.1 than for sample W4.8. Chemical State and Dispersion of the Metallic Phase. The chemical state and dispersion of the metallic phase were studied by XRD and XPS after the supported catalysts were heat treated in H2 or He flow. XRD (26) Yue, Z. R.; Jiang, W.; Wang, L.; Gardner, S. D.; Pittman, C. U., Jr. Carbon 1999, 37, 1785. (27) Gregg, S. J.; Sing, K. S. W. In Adsorption Surface Area and Porosity, 2nd ed.; Academic Press: London, 1982.

1754

Langmuir, Vol. 17, No. 5, 2001

Moreno-Castilla et al.

Figure 5. Variation of the mean particle size with tungsten content for catalysts treated at 973 K in H2: Series W, 0; series HW6.1, 9. Figure 3. XRD patterns of support and catalyst W14.9 treated at 748 K in H2.

Figure 4. XRD patterns of support and catalyst W14.9 treated at 973 K in H2.

diffraction patterns of some catalysts after their heat treatments in H2 at 748 and 973 K are shown, as an example, in Figures3 and 4, respectively, together with that corresponding to the support S. The main 2θ diffraction peaks of tungsten are at 40.26, 58.27, and 73.19° with relative intensities of 100, 15, and 24, respectively. All catalysts from series W after their heat treatments in H2 at 748 K showed the main diffraction peaks of tungsten. This result indicates that the interaction of the tungsten oxide particles with the surface of the activated carbon is not as strong as with Al2O3, since on this oxide the tungsten oxide particles were not reduced at 773 K.28,29 From Sherrer’s equation applied to the main diffraction peak of tungsten a mean particle size of 5.2 nm was obtained for all catalysts from series W, independent of the tungsten content. In the case of catalyst HW6.1, its XRD pattern did not show any diffraction peak corresponding to the metallic phase indicating that it is well dispersed on the support and with a particle size smaller than 4 nm. When the catalysts of series W were heat treated in H2 at 973 K, the diffraction peaks corresponding to metallic tungsten were also observed. The mean tungsten particle size obtained from Sherrer’s equation linearly increased with the tungsten content between 11.1 and 17.5 nm, as shown in Figure 5. The diffraction peaks obtained after heat treatment at 973 K are more intense and narrower (28) Biloen, P.; Pott, G. T. J. Catal. 1973, 30, 169. (29) Ng, K. T.; Hercules, D. M. J. Phys. Chem. 1976, 80, 2094.

Figure 6. Curve-fitted W4f core level spectra for catalysts treated at 973 K in H2.

than in the case of the heat treatment at 748 K (Figures 4 and 3, respectively), which means that there would be a greater proportion of tungsten particles and these would have a size distribution closer to the mean particle size. All these results indicate that tungsten particles were sintered when the treatment temperature increased. XRD of catalyst HW6.1 heat treated in H2 at 973 K also showed the diffraction peaks of tungsten, with a mean particle size of 11.3 nm, which also fit the straight line found between the mean particle size and tungsten content of catalysts from series W, depicted in Figure 5. In the case of this catalyst, sintering was stronger than for catalysts from series W, probably due to the greater mobility of the tungsten carbonyl than the ammonium tungstate. Catalysts reduced at 973 K in H2 were further analyzed by XPS. The spectra of the W 4f level are shown in Figure 6, which also includes the curve-fitted spectra. The BEs of the W 4f7/2, the percentages of the different metallic species, and the surface atomic ratios (O/W)s are compiled

Tungsten Catalysts Supported on Activated Carbon

Langmuir, Vol. 17, No. 5, 2001 1755

Table 2. BE Values of the W 4f7/2 Level and Surface Atomic Ratio, (O/W)s, for Catalysts Heat Treated in H2 Flow at 973 K for 4 h catal

W 4f7/2

(O/W)s

catal

W 4f7/2

(O/W)s

W4.8

31.5 (40)a 35.6 (60) 31.5 (45) 35.7 (55)

1.68

W23.1

1.17

1.36

HW6.1

31.5 (71) 35.7 (29) 31.6 (40) 35.7 (60)

W14.9

1.52

a

The values in parentheses correspond to the percentage of each peak.

Figure 7. Variation of surface atomic ratio, (W/C)s, against total atomic ratio, (W/C)t, for catalysts from series W treated at 973 K in H2.

in Table 2. All catalysts present binding energy values of around 31.5 and 35.6 eV, which correspond to metallic tungsten and W(VI), respectively. The proportion of metallic tungsten increased with the tungsten content, as shown in Table 2, in both the percentage of each BE peak and the (O/W)s ratio. WO3 was not detected by XRD. This would either be due to its particle size being smaller than 4 nm or it being dispersed in a thin layer covering the surface of the tungsten particles. Information about supported metallic particles can be obtained from the metal/support surface atomic ratio.30-32 Thus, Figure 7 shows the relationship between the surface atomic ratio, (W/C)s, and the total atomic ratio, (W/C)t, for the catalysts from series W. The linear relationship found between the two magnitudes seems to indicate that the metallic phase is uniformly dispersed on the support over the concentration range studied. XRD and XPS results of catalysts heat treated at 1223 K in He were given and widely discussed elsewhere.1 XRD of catalysts from series W showed the diffraction peaks of tungsten, tungsten carbide, and a nonstoichiometric carbide with composition W6C2.54. This was because the carbon support reduced the tungsten oxide particles to tungsten at 1223 K and this reacted with the support yielding different carbides. The XRD pattern of catalyst HW6.1 showed no diffraction peaks corresponding to the metallic phase, due to its higher dispersion. The XPS spectra of the W 4f level showed a mixture of W, WC, and WO3 at BEs values of 31.6-31.8, 32.2-32.4, and 35.5-35.6 eV, respectively, which agree with those reported in the literature. These values are compiled in Table 3 together with the percentage of each component. The amount of WO3 in the catalysts decreased as the tungsten content increased, and so the proportion of W (30) Kerkhof, F. P. J. M.; Moulijn, J. A. J. Phys. Chem. 1979, 83, 1612. (31) Fung, S. C. J. Catal. 1979, 58, 454. (32) Hilbrig, F.; Go¨bel, H. E.; Kno¨zinger, H.; Schmelz, H.; Lengeler, B. J. Phys. Chem. 1991, 95, 6973.

Table 3. BE Values of the W 4f7/2 Level for Catalysts Heat Treated in He Flow at 1223 K for 4 h catal W4.8 W14.9 W23.1 HW6.1

W 4f7/2 31.7 (7)a 31.6 (7) 31.7 (25) 31.8 (13)

32.4 (16) 32.2 (33) 32.2 (40) 32.4 (25)

35.5 (77) 35.5 (60) 35.5 (35) 35.6 (62)

a The values in parentheses correspond to the percentage of each peak.

Figure 8. Variation of the activity of catalyst W4.8, treated at 748 K in H2, in the hydrogenation of ethylene as a function of reaction temperature and reaction time at 598 K.

Figure 9. Variation of the activity of catalyst W4.8, treated at 973 K in H2, in the hydrogenation of ethylene as a function of reaction temperature and reaction time at 623 K.

and WC underwent similar increases. It is noteworthy that catalyst HW6.1 has a greater proportion of WC than either catalysts W4.8 or W14.9, which could be due to its greater dispersion favoring the transformation of tungsten trioxide into tungsten and tungsten carbide. WO3 was not detected by XRD, as in the above case, probably because it was well dispersed on the support or in a thin layer covering the surface of the metallic particles. Hydrogenation of Ethylene. Some of the catalysts studied above were tested in the hydrogenation reaction of ethene in which ethane was the sole reaction product. The behavior of catalyst W4.8 in this reaction after its heat treatment in H2 at 748 K is shown in Figure 8. This catalyst became active at a reaction temperature of around 500 K and reached its maximum activity at 598 K. Once at this temperature, there was a decrease (around 33%) in activity with reaction time, reaching the steady state after 1 h. Results obtained after this catalyst was heat treated at 973 K in H2 are shown in Figure 9. In this case, the catalyst became active at around 523 K and reached its maximum activity at 623 K. Both temperatures were

1756

Langmuir, Vol. 17, No. 5, 2001

Figure 10. Variation of the activity of catalyst W4.8, treated at 1223 K in He, in the hydrogenation of ethylene as a function of reaction time. Reaction temperatures: 273 K, O; 423 K, ∆; 623 K, 0; 623 K, after regeneration at 1223 K for 1 h, 9.

higher than in the above example, and the maximum activity obtained was around eight times lower than after the pretreatment at 748 K. Furthermore, activity decreased with reaction time, and after 5 h it was around 50% lower than in the above pretreatment. The higher activity of the catalyst after its pretreatment at 748 K than at 973 K would be due to the tungsten particles being smaller in the former, 5.2 nm, than in the latter, 11.1 nm. The reduction in activity with reaction time would be related with carbon deposition on the metal particles. The greater decrease in activity after heat treatment at 973 K would be due to the larger particle size of this catalyst favoring carbon deposition. Results obtained with catalyst W4.8 after its heat treatment in He flow at 1223 K are shown in Figure 10. In this case, the catalyst had a very high activity, much higher than after the above heat treatments, even at a reaction temperature as low as 273 K. But, at this reaction temperature, the catalyst was completely deactivated after 30 min. When the reaction temperature was increased to 423 K, the catalyst also had an activity similar to that for the above reaction temperature but was deactivated more slowly. Thus, after 100 min the activity was zero. At a reaction temperature of 623 K, the initial activity was maintained and deactivation was much slower. These results are quite interesting and indicate that after the treatment at 1223 K in He flow a very active catalyst was obtained. Although the initial activity was similar at reaction temperatures between 273 and 623 K, lower reaction temperatures corresponded to faster deactivation of the catalyst. Heat treatment of the catalysts at 1223 K in He flow yielded a mixture of tungsten, tungsten carbide, and tungsten trioxide, according to XRD and XPS data.1 Tungsten trioxide was inactive in the hydrogenation of ethylene, as confirmed experimentally. Therefore, the high initial activity of this catalyst would be due to the presence of tungsten carbide since, as commented above, tungsten carbide was found to be active, as catalyst, in the hydrogenation of ethylene. After 200 min on-stream at 623 K, catalyst W4.8 was regenerated by heating at 1223 K in He flow for 1 h, and as shown in Figure 10, the initial activity of this regenerated catalyst was slightly higher than that of the original one. This slight increase in activity can be due to an increase in WC content during regeneration of the catalyst. Figure 11 shows that the activity of catalyst HW6.1 after its pretreatment in He flow at 1223 K was much higher than that of catalyst W4.8, under the same conditions, which was due to both the high dispersion of

Moreno-Castilla et al.

Figure 11. Variation of the activity of catalyst HW6.1, treated at 1223 K in He, in the hydrogenation of ethylene as a function of reaction time. Reaction temperatures: 423 K, ∆; 423 K, after regeneration at 1223 K for 1 h, 2; 623 K, 0; 623 K, after regeneration at 1223 K for 1 h, 9.

Figure 12. Adsorption kinetics of ethylene on catalyst W4.8 treated at 1223 K in He. Adsorption temperature: 273 K, 0; 423 K, ]; 623 K, O.

the metallic phase and the higher WC content of the former catalyst, as shown by XRD and XPS, respectively. In this catalyst the trend toward deactivation with reaction time is similar to that found with catalyst W4.8. The increased rate of catalyst deactivation with decreasing reaction temperature can be due to a strong adsorption of the hydrocarbon either on the active sites of the catalyst33 or on the support and catalyst. Thus, the adsorption kinetics of ethylene at different temperatures on catalyst W4.8 pretreated in He flow at 1223 K were determined and the results obtained are depicted in Figure 12. The amount of C2H4 adsorbed at 273 K was, after 30 min, about six times greater than at both 423 and 623 K. Therefore, the decrease in reaction temperature would make the rate of hydrogenation of ethylene slower than the rate of its adsorption, with activity decay increasing with decreasing reaction temperature. Conclusions In the hydrogenation of ethylene, the catalysts containing WC were much more active than those containing only metallic tungsten as active phase, even at reaction temperatures as low as 273 K, although the deactivation of the catalysts was more rapid. The higher the reaction temperature the slower the deactivation rate. This deactivation with reaction time was due to the strong adsorption of the hydrocarbon on the support and catalyst, which is favored when the reaction temperature decreased. Acknowledgment. This work was supported by the DGESIC Project No. PB97-0831 LA001367K (33) Gonza´lez, P. N.; Villa-Garcı´a, M. A.; Brenner, A. J. Catal. 1989, 118, 360.