Applied Catalysis, 44 (1988) l-9 Elsevier Science Publishers B.V., Amsterdam -
Printed in The Netherlands
Ruthenium-Palladium Catalysts: the Effect of Palladium on the Catalytic Behaviour of Ruthenium MARGARITA
VINIEGRA*, VICTOR ARROYO and RICARDO Gi)MEZ
Depto. de Quimica, Universidad Autdnoma Metropolitana-Iztapalapa, A. P. 55- 534, M&co 09340 D.F. (Mexico) (Received 2 October 1987, revised manuscript received 16 May 1988)
ABSTRACT Ru-Pd/SiO, catalysts were characterized by hydrogen chemisorption and their catalytic activity was obtained for the hydrogenation of benzene at atmospheric pressure. A synergistic effect appeared in the bimetallic catalysts with respect to monometalic ruthenium. This enhancement in the catalytic activity is explained in terms of the dilution of ruthenium ensembles by palladium. This increased the resistance to self poisoning of the active surface. It is also suggested that palladium serves as a source of hydrogen atoms for the reaction. Catalytic activity experiments performed in the presence of thiophene showed that this molecule inhibits the formation of undesired carbonaceous residues. INTRODUCTION
Bimetallic catalysts have been the subject of a great deal of interest in the area of scientific studies and potential industrial applications. These systems have greatly contributed to the identification of the factors which determine the activity of metals. It seems that the relative importance of these factors depends on the type of reaction selected to characterize the catalytic activity of bimetallic catalysts, The ensemble theory is often used to explain the decrease in the ability of alloys to rupture C-C bonds, as this reaction requires sites which consist of several metal atoms [ 11, For reactions which involve the formation of C-H bonds, the modification of the reaction rate as a function of alloy composition has sometimes been explained in terms of either an electronic or a ligand effect [ 2,3], The hydrogenation of benzene has been shown to be a facile reaction over several metals [4,5]. The synergistic effect found in bimetallic catalysts towards this reaction has been taken as evidence of electronic interactions between the metals in the alloy [ 6,7]. Nevertheless, it has been pointed out that the synergism found in alloy catalysts for hydrogenation reactions may have some explanation other than the ligand effect [ 81.
0 1988 Elsevier Science Publishers B.V.
With the aim of clarifying this particular point, we studied the hydrogenation of benzene in order to characterize the catalytic activity of rutheniumpalladium bimetallic catalysts. This system was chosen because both, ruthenium and palladium in combination with other metals, have been shown to modify the catalytic activity of hydrogenation reactions [ 2,3,7]. EXPERIMENTAL
Catalysts were prepared by the co-impregnation technique from aqueous solutions of RuC1,.3H,O and PdC12.2H20 (ICN Pharmaceuticals, Plainview, NY, U.S.A.) in appropriate amounts to give a total loading of 4 wt.-% metal. The atomic fraction of palladium was varied between 0.0 and 0.85. The support used was silica (Ketjen F-2,180 m’/g) obtained from Akzo Chemie, which had previously been calcined in air at 723 K for 12 h. Following the evaporation of the solvent, the samples were dried in air at 383 K for 12 h. The catalysts were reduced for 3 h in flowing hydrogen at 573 K for the palladium catalysts and at 673 K for the ruthenium and the bimetallic catalysts. The catalysts were cooled to room temperature in flowing hydrogen and stored in bottles for subsequent use. Metal dispersions were obtained by hydrogen chemisorption at 343 K in a conventional glass volumetric apparatus. Prior to the chemisorption experiments the catalysts were reactivated in flowing hydrogen at 673 K for 2 h, outgassed at the same temperature for the same period of time, and then cooled under vacuum to 343 K for the adsorption studies. The pressure in the system was found to be 1.10e5 Torr (1 Torr= 132 Pa) or better, before hydrogen was admitted to the cell. The total hydrogen uptake was obtained by extrapolating the linear portion of the isotherm to zero pressure. The hydrogenation of benzene was performed in a conventional differential flow reactor coupled to a gas chromatograph in order to analyze the products of the reaction. The reaction was carried out at atmospheric pressure and 323 K with a hydrogen/benzene partial pressure ratio of 9.5 and a flow-rate of 3.6 l/h. The conversion of benzene (%C) was kept below 15% by varying the mass of the samples. The only product detected under these conditions was cyclohexane. Experiments with thiophene were also performed, as described above, by adding 20 ppm of thiophene to the reactant mixture. RESULTS
Table 1 lists the series of ruthenium-palladium catalysts prepared in addition to the metal dispersions which are defined as H/ (Ru + Pd), i.e. the ratio of the total hydrogen atoms chemisorbed with respect to the total ruthenium plus palladium atoms. Dispersions were small for all of the catalysts (mean crystallite size in the range 5.0-8.0 nm) therefore, dispersion effects were not likely to appear in the bimetallic systems.
3 TABLE 1 Bulk composition and dispersion of Ru-Pd/SiO,
Atomic fraction of Pd
Ru1.00 RuO.90 RuO.75 RuO.50 RuO.25 RuO.15
4.0 3.6 3.0 2.0 1.0 0.6
0.0 0.4 1.0 2.0 3.0 3.4
0.00 0.10 0.25 0.50 0.75 0.85
0.14 0.14 0.13 0.18 0.22 0.22
Fig. 1 shows the catalytic activity for the hydrogenation of benzene. The data points correspond to reaction times of 20 min. The rates are plotted as a function of catalyst composition and are expressed as turnover frequencies (molecules/site/s). A synergistic effect appeared, since a maximum in the activity was observed for the catalyst which had a composition of 0.75 ruthenium. Palladium was found to be practically inactive under the experimental conditions used. It is worthwhile noting here that the values of turnover frequencies (TOF) were calculated in terms of the total number of surface atoms, i.e. ruthenium plus palladium atoms, since hydrogen chemisorption cannot distinguish between them. Monometallic ruthenium, as well as bimetallic catalysts, deactivated steadily at 323 K, and the catalytic decay followed the hyperbolic relationship pro-
Fig. 1. Turnover frequencies of Ru-Pd/SiO, catalysts as a function of bulk composition. Reaction temperature = 353 K, flow-rate = 3.6 l/h. Data points correspond to reaction times of 20 min.
75 t (mid
Fig. 2. Self deactivation
cording to the hyperbolic
at 353 K in the hydrogenation
law. Key: (o )Rul.OO; ( l )RuO.SO; (0 )RuO.75).
posed by Germain and Maurel [91 as shown in Fig. 2. In order to make a more quantitative analysis of the deactivation process, the model of Levenspiel [101 was applied to obtain the rate constant of deactivation, as was also done for Pd/Al,O, [111and Rh/A1203 [ 121 catalysts. Taking into account the experimentally determined reaction rate order of zero with respect to benzene, the hyperbolic decay corresponded to a deactivation order of two. The application of Levenspiel’s model led to the following equation [ 111
where k, is the experimentally determined slope as obtained from the hyperbolic law, kd is the deactivation rate constant, /zh is the rate constant for the hydrogenation reaction, Cb is the benzene concentration (mol per I), Fb is the feed rate of benzene (mol per s ) , w is the catalyst weight (g) andp is the order relative to benzene for deactivation. Considering that the feed rate as well as the concentration of benzene were kept constant in all of the experiments, one can write: &=k,(WP(KJ
and therefore: &i=k(kh)(W)
5 TABLE 2 Catalytic activity and deactivation constants for Ru-Pd/SiOz benzene at 353 K Catalyst Ru1.00 Ru0.90 RuO.75 Ru0.50 Ru0.25 Ru0.15
catalysts in the hydrogenation of
TOF x lo’* (molecules/site/s)
99.0 8.6 7.6 9.3 14.0 24.0
4.1 4.1 3.2 1.6 1.8 0.7
40.6 3.5 2.5 1.5 2.5 17.3
76.0 74.0 63.8 23.0 21.2 8.5
*Initial activity (time zero of reaction).
The values of the experimental slope (k, ) , the initial rate of reaction ( kh ) , and the deactivation constant (kb) are reported in Table 2. The TOF values listed in this table, were computed from the initial rates of the reaction (&). It was observed that the monometallic ruthenium catalyst had the largest value of the deactivation rate constant. A sharp decrease was observed when 0.10 atomic fraction of palladium was added. The initial TOF values decreased smoothly on going from pure ruthenium to pure palladium. The results obtained for the hydrogenation reaction in the presence of 20 ppm of thiophene are shown in Table 3. The conversion followed an exponential decay characterized by a rate constant kd such that:
where kd represents the rate constant for the deactivation in the presence of TABLE 3 Deactivation constants and catalytic activity of Ru-Pd/SiOl benzene in the presence of 20 ppm of thiophene Catalyst
Ru1.00 Ru0.90 Ru0.75 RuO.50 Ru0.25 RuO.15
r0X 10” (moI/g cat/s)
6.4 3.3 6.5 3.7 2.0 0.7
2.6 2.0 1.7 2.0 2.0 1.8
at 353 K for the hydrogenation of
11.9 6.0 13.0 5.3 2.4 0.8
Benzene + thiophene
0.2 19.2 11.4 28.1 16.9 11.7
21.2 43.3 56.7 48.1 44.7 53.8
the hydrogenation of benzene by Kubicka [ 141. Therefore, the first observation that should be made is that, although the ruthenium-palladium phase diagrams do not predict the alloy formation, these metals interact in some way, possibly by forming bimetallic clusters. The question that requires an answer is the nature of the interaction responsible for changes in activity of ruthenium. Because the hydrogenation of benzene is a “facile” reaction, the geometric effect (dilution of ruthenium atoms) should not be an important point in this explanation. However, if we take into account the data reported in Table 2 attention must be focussed on the dilution of the ruthenium ensembles by palladium. First, there was a sharp decrease in the deactivation constant with the addition of palladium, and secondly the initial turnover frequencies decreased monotonically with the Ru atomic composition. Since palladium was found to be inactive the kb and TOF values are representative of the ruthenium atoms only. It is surprising that the synergism shown in Fig. 1 disappeared after the initial rates of reaction were calculated (TOF values of Table 2 ), and the activity decreased smoothly from pure ruthenium to pure palladium. We must remember that these initial rates were obtained from the deactivation curves (Fig. 2). Therefore, the different resistance to self poisoning of the catalysts in the series, appears to be the explanation for the “volcano curve” behaviour shown in Fig. 1. Palladium seems to increase the stability of the ruthenium ensembles in such way as to impede the total blockage of ruthenium atoms by carbonaceous residues (highly dehydrogenated species). Very similar observations were made and conclusions drawn by Sachtler and Somorjai [ 151 for the dehydrogenation of cyclohexane over platinum-gold catalysts. These authors showed that the amount of surface carbon present after the reaction decreased as the Au content of the catalyst increased. Therefore, and this seems to be general behaviour, alloying (or clustering) inhibits the side reactions that cause poisoning of the catalysts surface during hydrocarbon conversions, Although it seems quite clear that the presence of palladium increases the resistance to self poisoning, it is tempting to speculate a little and recall that palladium also adsorbs hydrogen, perhaps more easily than ruthenium itself and, therefore, it may supply hydrogen atoms to the ruthenium ensembles where the hydrogenation takes place. However more research is needed to clarify whether the reduction in size of the ruthenium ensembles by dilution with palladium or whether the formation of ruthenium-palladium dual sites is responsible for the increased stability of ruthenium-palladium bimetallic catalysts. The hydrogenation of benzene was also performed in the presence of thiophene, since it was reported previously that the rate constant of deactivation is highly sensitive to electronic modifications of the respective metal [ 161. In this sense, it was observed that the deactivation constant was invariable with the catalyst composition (Table 3), and since benzene was mainly hydrogen-