Journal of Catalysis 202, 163168 (2001)

doi:10.1006/jcat.2001.3274, available online at http://www.idealibrary.com on

Hydrogenation of Tetralin in the Presence of Sulfur over

-Al2 O3 -Supported Pt, Pd, and PdPt Model Catalysts
J. L. Rousset,,1 L. Stievano,,2 F. J. Cadete Santos Aires, C. Geantet, A. J. Renouprez, and M. Pellarin
Institut de Recherches sur la Catalyse, 2 Avenue A. Einstein, 69626 Villeurbanne Cedex, France; and Laboratoire de Spectrometrie Ionique et
Moleculaire, Universite Claude Bernard Lyon I, 43 Boulevard du 11 nov 1918, 69622 Villeurbanne Cedex, France

Received February 26, 2001; revised May 9, 2001; accepted May 9, 2001

Thus, great efforts have been devoted to develop sulfur

-Al2 O3 -supported Pd, Pt, and bimetallic PdPt catalysts have tolerant noble metal (generally Pt or Pd) catalysts with
been produced by laser vaporization of the respective bulk met- a high hydrogenation activity for the two-stage hydro-
als and alloys. Transmission electron microscopy and energy- genation process. The resistance to sulfur poisoning is
dispersive X-ray analysis experiments show that these bimetallic considered to come from the small bonding energy of the
clusters have a perfectly well defined stoichioemetry. Moreover, pure electron-acceptor sulfur atoms with the electron-deficient
or bimetallic clusters have similar size distributions. One has taken
metal clusters. Modifications of the electron-deficient
benefit of these features to study in detail the resistance to sulfur
nature of the metal clusters would be induced by changing
poisoning as a function of the composition and to verify to what
extent what is claimed in the literature, i.e., an improved resistance the geometry of the particles or the acidbase nature of
to sulfur poisoning of PdPt catalysts as compared to pure metals, the support or also by alloying with another metal.
is verified. The catalytic properties of Pd, Pt, Pd17 Pt83 , and Pd65 Pt35 In particular, studies have been focused on the sulfur tol-
in the hydrogenation of tetralin in the presence of varied amounts erance of PdPt alloy deposited on various supports such
of H2 S have been compared. Contrary to what is claimed, the co- as Al2 O3 (1, 2), SiO2 Al2 O3 (35), Al2 O3 B2 O3 (6), or HY
existence of Pd with Pt in the same particle has no influence on the zeolite (7) and industrial applications for hydrogenating
hydrogenation activity nor on the sulfur tolerance of the catalysts. aromatics and olefins in hydrocarbon feedstocks on PdPt
Hence, it has been shown that Pt has no electronic influence on the alloys were proposed and actually used in plants (810).
reactivity of Pd atoms. c 2001 Academic Press But the reasons for the high sulfur resistance of PdPt-
Key Words: bimetallic PdPt; laser vaporization; model catalysts;
supported catalysts invoked in the literature are still con-
tetralin hydrogenation; surface segregation.
troversial. On the one hand, Lin et al. (11) found that the
addition of Pd to the supported Pt catalysts inhibits the ag-
glomeration of Pt particles which could explain in part the
INTRODUCTION
better sulfur tolerance of the alloy even if this group has
The need for relatively low aromatic contents in liquid also invoked electronic interactions between Pd and Pt. On
fuels or solvents is continuously increasing. After a desul- the other hand, Yasuda et al. (4) claimed that the high sul-
furization and denitrogenation treatment, these products fur tolerance of the alloys is due to structural and electronic
usually still contain relatively high percentages of olefins effects rather than maintaining the metal dispersion. More-
and monocyclic or polycyclic aromatics, so that further hy- over, the situation is not really cleared up by the recent re-
drogenation is required to produce acceptable products. In- sults reported by Fujikawa et al. (12) in their study of the
deed, aromatics in diesel fuel not only generate undesired hydrogenation aromatics in hydrotreated light oil on Pd
emissions of particles in exhaust gases but also lower the Pt/SiO2 Al2 O3 catalysts. Indeed, they demonstrated that
cetane number. The presence of sulfur compounds in the the catalytic activity of bimetallic PdPt is significantly en-
products to be hydrogenated complicates the process since hanced with respect to the monometallic catalysts but also
sulfur constitutes a poison of the metal catalysts used for that all of them, pure or alloyed, are similarly affected by
hydrogenation. the concentration of the sulfur in the feed.
Finally these ill-matched results could be explained by
Cooper and Donnis (13), who revealed that the details of
1
To whom correspondence should be addressed. Fax: (00) 04 72 44 53 the preparation of a PdPt bimetallic catalyst and hence the
99. E-mail: rousset@catalyse.univ-lyon1.fr.
2
Current address: Laboratoire de reactivite de surface, Universite
degree to which Pd and Pt form alloyed clusters together
Pierre et Marie Curie-Paris VI, Place Jussieu, 75252 Paris Cedex 05, with the size of these clusters strongly influence the activity
France. as well as the sulfur tolerance of these catalysts.

163
0021-9517/01 $35.00
Copyright
c 2001 by Academic Press
All rights of reproduction in any form reserved.
164 ROUSSET ET AL.

Since the size, the structure, and the composition of sup- TABLE 1
ported bimetallic particles are difficult to control for ac- Metal Loading of the Catalysts and Mean Diameters and Disper-
tive phases obtained by chemical methods, the comparison sion of Supported Clusters As Deduced from TEM Experiments
of the catalytic properties of a bimetallic catalyst to those
of a monometallic one is far from straightforward. To ob- Mean Mean
tain valuable information on the bimetallic interactions, one Metal diameter,a diameter,
loading fresh sample used samples Dispersion
needs to prepare monometallic and bimetallic systems hav- Catalyst (wt%) (nm) (nm) (%)
ing approximately the same size, morphology, and overall
uniform and well-defined composition. We attempted to Pd100 0.038 4.0 6.1 14.6
prepare such model supported catalysts by using the soft Pd65 Pt35 0.032 3.6 5.5 15.2
Pd17 Pt83 0.08 3.4 5.2 16.0
landing deposition clusters over low-acidity -Al2 O3 and Pt100 0.048 3.4 5.1 16.2
to study their sulfur tolerance and catalytic behavior in
a
the vapor-phase hydrogenation of tetralin in the presence Mean diameter for the fresh samples, from Ref. (16).
of H2 S.

EXPERIMENTAL by using a gas-phase saturator system (corresponding flow

of tetralin, 5.48 108 mol/s). In order to investigate the
Catalysts Preparation by Using Laser Vaporization influence of H2 S on the activity of the catalysts, the H2 S
The laser vaporization cluster source has been exten- concentration was varied between 50 and 500 ppm. In or-
sively described elsewhere (14). In brief, the second har- der to avoid any H2 S consumption due to the reactivity
monic of a Nd:YAG pulsed laser is used to vaporize the of this gas with the walls of the reactor (15), the heated
metal from a rod to create a plasma. Cluster nucleation inner part was protected by a quartz tube whose diame-
and growth occur when a continuous flux of inert gas (a ter has been adjusted to minimize the resulting dead vol-
mixture of He and Ar) is introduced in the chamber. Dif- ume. The products of tetralin hydrogenation were analyzed
ferential pumping extraction of the clusters through a skim- every 30 min by gas chromatography. Cis and trans de-
mer yields a cluster beam which carries neutral and ionized calins were always obtained together with small amounts
species. After removal of the ionized clusters by electro- of naphthalene, while no detectable amounts of isomeriza-
static deflection, the neutral ones are deposited on the sub- tion or cracking products were observed. The amount of
strates. Typical deposition rates (equivalent thicknesses) catalyst (about 500 mg) was chosen in order to obtain a
of about 5 nm/min are obtained with neutral clusters. The conversion below 10%, ensuring the validity of a differ-
deposition rates were monitored using a quartz microbal- ential model for the determination of the specific rates. A
ance. In order to deposit the clusters on the support, in rapid deactivation of the catalyst is observed during the
the present case of a -alumina powder (Condea, Puralox first hour of the experiment. In order to check for further
SCFa-240 with a specific surface area of 240 m2 /g), a de- deactivation of the catalyst, the usual bracketing method
vice has been developed which stirs the powder in front was used: several experimental points were repeated af-
of the cluster beam. After deposition, the samples are ter the complete collection of data for a given set of reac-
air transferred, characterized, and used for catalytic reac- tion conditions. The deactivation of the catalyst appeared
tions. For these experiments Pd and Pt rods of 99.99% and always to be negligible, within the experimental errors.
99.95% purity respectively as well as Pd17 Pt83 and Pd65 Pt35 The activation energy of the catalysts was also determined
bimetallic rods obtained by melting Pd (99.99% purity) and by measuring the activity for 500 ppm of H2 S at several
Pt(99.95% purity) were used. About 3 g each of Pd, Pt, temperatures between 550 K and 610 K, in order to as-
Pd17 Pt83 , and Pd65 Pt35 -alumina-supported catalysts were certain that the reaction was carried out in the chemical
prepared. The metal contents of each catalyst, determined regime.
by atomic absorption spectroscopy, are reported in Table 1.
RESULTS

Catalytic Activity Measurements: Tetralin Hydrogenation Transmission Electron Microscopy Experiments

in the Presence of H2 S
TEM experiments have been performed on each catalyst
The hydrogenation of tetralin (1,2,3,4-tetrahydronap- at three different stages: (1) fresh samples, (2) pretreated
hthalene) was performed in a fixed-bed gas-flow microreac- samples heated for 2 h at 573 K under hydrogen flow, and
tor at 573 K, at a constant H2 pressure of 4.6 MPa (45 atm). (3) after tetralin hydrogenation. The results for the fresh
The total flow rate was 56 ml/min. The partial pressure of sample have been published in a recent paper reporting
tetralin (Fluka, purity >99%) was kept constant at 6.1 kPa on the hydrogenation of toluene (16) on identical catalysts.
 TETRALIN HYDROGENATION ON Al2 O3 -SUPPORTED PdPt MODEL CATALYSTS 165

Hydrogenation of Tetralin in the Presence of H2 S

The thio-resistance of the catalysts can be evaluated by
measuring the reaction rate at different sulfur concentra-
tions. The hydrogenation of tetralin over the four catalysts
in the presence of sulfur was carried out in the 50 to 500 ppm
H2 S range. Figure 2 shows the effects of the H2 S partial
pressure on decalin formation. As expected, all reaction
rates decrease with an increase in the partial pressure of
H2 S. Assuming a pseudo order of reaction with respect to
the pressure of H2 S, one can fit the reaction rates by the
following equation,

TOF = k[P(H2 S)]n , [1]

where k and n are respectively the rate constant and the

pseudo order. The turnover frequencies (TOFs) of pure Pt
and pure Pd samples as well as the corresponding fitting
curves are displayed in Fig. 3. The TOFs are based on sur-
face metallic atoms deduced from TEM experiments and
corresponding size histograms. The different k and n values
corresponding to the four catalysts are reported in Table 2.
FIG. 1. Size distribution histograms of PdPt clusters after reaction. The effects of temperature on the activity were also inves-
Pure Pd (black), Pd65 Pt35 (dark gray), Pd17 Pt83 (light gray), and pure Pt tigated and the corresponding Arrhenius plots and param-
(white). eters are given in Fig. 4 and Table 2, respectively.
The activation energies obtained from the Arrhenius
plots were found to be approximately 70 kJ/mol for all the
The size histograms of the different catalysts after reaction systems, i.e., high enough to exclude the presence of a dif-
are presented in Fig. 1. The mean sizes of the four fresh fusion regime.
samples (16) are reported for comparison in Table 1. Af-
ter reaction, analysis of the size histograms (Fig. 1) shows
that the mean diameter of the particles is larger than that
of the fresh catalysts (Table 1), remaining however simi-
lar to those of the preheated samples before reaction. The
catalytic properties of the catalysts do not vary appreciably
during the catalytic test, indicating that the observed in-
crease of the mean particle size has already occurred during
the annealing treatment and not during the reaction. This
sintering, occurring at rather low temperature, is at first
sight surprising for these low-loaded samples, compared
to catalysts prepared by conventional methods. Actually,
the clusters deposited by the so-called soft landing mode
have a weak chemical interaction with the support and are
thus able to diffuse easily at the surface. The composition of
individual particles for both Pd17 Pt83 and Pd65 Pt35 bimetal-
lic samples was determined by EDX-S analysis with 1-nm2
electron probes which were defocused to the size of each
analyzed cluster. The homogeneity of composition has been
checked for both systems and found equal to that of the va-
porized rod. Identical results obtained on the same kind of
catalysts have been published recently (14, 16) and, conse-
quently, the corresponding figures will not be presented in FIG. 2. Effects of H2 S partial pressure on reaction rate. Pure Pd (n),
the present paper. Pd65 Pt35 (,), Pd17 Pt83 (s), and pure Pt(h).
166 ROUSSET ET AL.

In a recent work, low-energy ion scattering (LEIS) was em-

ployed to study PdPt bimetallic clusters deposited on flat
supports and a strong enrichment in Pd at the surface of the
bimetallic particles was evidenced. In parallel with LEIS
experiments, the modelling of the segregation process was
performed using Monte-Carlo simulations (14) and the de-
duced surface concentrations were in good agreement with
the experiments. In the present case, the metal loading is too
low to measure the surface composition with a reasonable
accuracy by LEIS. Thus, the surface compositions have been
predicted on the basis of our theoretical model, thoroughly
described elsewhere (17), since it has been shown to give
reliable results. The present results have been obtained with
the procedure described in a recent work (16). Following
this procedure, one calculates the surface Pd concentration
and, consequently, the number of Pd surface atoms for each
cluster size. Taking into account the size histogram of the
Pd17 Pt83 -supported catalyst (Fig. 1), a Pd surface concentra-
tion of 59% has been calculated. Concerning the Pd65 Pt35
alloy, the calculations have shown that the surface is al-
most entirely covered by Pd atoms whatever the size of the
clusters. Finally, one has to note that the large Pd surface
segregation is associated with a Pd depletion in the second
atomic layer (17).

FIG. 3. Reaction rates of pure Pd (s) and pure Pt (h) catalysts in the
presence of H2 S. The solid lines represent the best-fit data curves using
Eq. [1] (see text).

DISCUSSION

Surface Composition of the Clusters

The surface composition of alloys and especially the com-
position of the topmost surface layer are generally differ-
ent from that of the bulk one, due to segregation processes
(17, 18). A first step in the understanding of the reactivity
of alloys is the determination of the surface composition.

TABLE 2
Activation Energies and Pseudo Order Rate Constants for
Tetralin Hydrogenation in the Presence of H2 S Derived Respec-
tively from Arrhenius Plot (Fig. 4) and Best-Fit TOF Curves Using
Eq. [1]

Activation
energy
Catalyst (kJ/mol) Rate constant k (s1 ) pseudo order n

Pd100 70.1 25.0 0.38

Pd65 Pt35 71.2 18.6 0.34
Pd17 Pt83 73.1 28.4 0.67 FIG. 4. Arrhenius plot for hydrogenation of tetralin on pure Pd (s),
Pt100 68.9 65.6 0.88 Pd65 Pt35 (4), Pd17 Pt83 (h), and pure Pt (,). (Partial pressure of H2 S =
500 ppm.)
 TETRALIN HYDROGENATION ON Al2 O3 -SUPPORTED PdPt MODEL CATALYSTS 167

Tetralin Hydrogenation
To examine the sulfur tolerance of the noble metal cata-
lysts, variable amounts of H2 S were added during the hy-
drogenation reaction. Comparing the TOF measured on
the different samples, a first important result comes out:
platinum is the most active catalyst at low sulfur concen-
tration as shown in Fig. 2. In contrast, pure Pd becomes
slightly more active than Pt for high sulfur contents. As a
consequence of the observed pseudo order of reaction with
respect to H2 S reported in Table 2, one notes that inhibition
effects decrease in the following order: Pt100 > Pd17 Pt83 >
Pd65 Pt35 Pd100 . Additionally, one notes that the activity
curve of the Pd65 Pt35 alloy is almost superimposed with that
of the pure Pd sample in the whole range of sulfur concen-
tration. However, even if the surfaces of both systems are
entirely covered by Pd atoms, the Pd surface atoms of the
Pd65 Pt35 catalyst are surrounded by Pt atoms lying in the
layer underneath. These observations lead to the conclusion
that neighboring Pt atoms have no influence on the reactiv-
ity of Pd atoms nor on their sulfur tolerance: the bimetallic
catalyst behaves as a monometallic supported Pd catalyst.
This is consistent with the results obtained with the same
catalysts in the hydrogenation of toluene where no synergy
between Pd and Pt was observed (16). For the Pd17 Pt83 al-
loy, which is the only catalyst with a mixed PdPt surface,
the value of the pseudo order (Table 2) lies between those
FIG. 5. Comparison of the reaction rate of Pd17 Pt83 catalyst (h) with
of pure Pt and pure Pd. To illustrate quantitatively its be- those obtained by adding the TOF of the pure Pd and Pt catalysts weighted
havior, we have represented in Fig. 5 the TOF dependence by the respective Pd (59%) and Pt (41%) surface concentrations (s).
on the H2 S concentration for this alloy and that obtained by
adding the TOF of the pure metals weighted by the respec- Au alloy and three atoms over PdCu alloy. We have also
tive Pd (59%) and Pt (41%) surface concentrations. Fitting found an active site composed of a pair of atoms in the case
this predicted curve with Eq. [1] leads to a pseudo order of toluene hydrogenation over PdPt catalyst (16). There-
value of 0.61, very close to 0.67, the value measured fore, the results obtained in the present work can be sum-
on the Pd17 Pt83 catalyst. This indicates that the reactivity marized as follows: (i) The catalytic behavior of the PdPt
of the alloy in the presence of sulfur can be perfectly ac- alloys in the presence of sulfur can be perfectly explained
counted for by an additivity of the catalytic properties of by an additivity of the properties of the two pure metals. (ii)
the pure metals. However, the activity of the alloys is not a The observed activities of the two alloys are smaller respec-
simple weighted average of those of the two components. tively than that of Pt for low H2 S concentration and than
Indeed, one has to take into account an ensemble size effect that of pure Pd for high sulfur content. This established lack
of the active sites. More precisely, the nearly equal values of synergy between Pt and Pd seems in contradiction with
of the pseudo orders derived above imply that the ratio of many recent works devoted to PdPt catalysts supported on
the measured TOF of the alloy to the simulated one is al- acidic support prepared by chemical methods (3, 57, 21)
most constant whatever the sulfur content. Actually, this excepted those of Qian et al., who studied tetralin hydro-
ratio was found equal to 0.64 0.05. A value smaller than genation over Pd, Pt, and PdPt supported on alumina (2).
1 is not surprising and can be explained by a dilution ef- This could indicate that these new catalysts have a weaker
fect of the Pt ensembles, constituting the active sites, by Pd interaction with the support, explaining their different cata-
atoms or vice versa, leading to a reactivity of mixed PdPt lytic properties.
sites lower than that resulting from a simple additivity of
pure Pd and Pt activities. Indeed, it has been shown in sev- CONCLUSION
eral publications that the active site in the hydrogenation of
aromatics is often composed of several atoms. For example, If the results reported in the literature often do not agree
Cadenhead and Masse (19) and Leon y Leon and Vannice among themselves, the reason, as outlined by Stanislaus
(20), who studied benzene hydrogenation, determined re- and Cooper (22), is probably the simultaneous influence of
spectively an active site composed of two atoms over Pd many ill-defined parameters. According to these authors,
168 ROUSSET ET AL.

the resistance to poisoning of these catalysts should de- 4. Yasuda, H., Matsubayashi, N., Sato, T., and Yoshimura, Y., Catal. Lett.
pend to a large extent on the nature of the support, and the 54, 23 (1998).
size, the structure, and the location of the metal particles, 5. Navarro, R. M., Pawelec, B., Trejo, J. M., Mariscal, R., and Fierro,
J. L. G., J. Catal. 189, 184 (2000).
which in turn are controlled by the preparation conditions. 6. Yasuda, H., Kameoka, T., Sato, T., Kijima, N., and Yoshimura, Y.,
Therefore, the catalytic performances of these systems can- Appl. Catal. 185, 199 (1999).
not be easily compared, especially if their structural char- 7. Yasuda, H., and Yoshimura, Y., Catal. Lett. 46, 43 (1997).
acterizations have not been carefully performed. Thanks to 8. Kovach, S. M., and Wilson, G. D., U.S. Patent 3943053, Ashland Oil
Inc. (1974).
the laser vaporization technique, the as-prepared pure Pd,
9. Minderhoud, J. K., and Lucien, J. P., U.S. Patent 4960505, Shell Oil Co.
pure Pt, and PdPt catalysts have the same structure and the (1988).
bimetallic particles have a well-defined composition. This 10. Winquist, B. H. C., Milam, S. N., Murray, B. D., and Ryan, R. C., EP
was indeed the key point which allowed us to easily com- Appl. 0519573, Shell IRM (1992).
pare their catalytic properties and to affirm that alloying 11. Lin, T., Jan, C., and Chang, J., Ind. Eng. Chem. Res. 34, 4284 (1995).
12. Fujikawa, T., Tsuji, K., Mizuguchi, H., Godo, H., Idei, K., and Usui, K.,
Pd with Pt has no intrinsic influence on the activity and re-
Catal. Lett. 63, 27 (1999).
sistance to sulfur poisoning, compared to the pure metals, 13. Cooper, B. H., and Donnis, B. B. L., Appl. Catal. A 137, 203 (1996).
at least under the reaction conditions used in the present 14. Rousset, J. L., Renouprez, A., and Cadrot, A. M., Phys. Rev. B 58,
work. 2150 (1998).
To investigate to what extent an effect of the support is at 15. Labruyere, F., Lacroix, M., Schweich, D., and Breysse, M., J. Catal.
167, 464 (1997).
the origin of the thio-resistance observed by many authors,
16. Rousset, J. L., Stievano, L., Cadete Santos Aires, F. J., Geantet, C.,
these clusters obtained by using the laser vaporization tech- Renouprez, A. J., and Pellarin, M., J. Catal. 197, 335 (2001).
nique will be, in the near future, deposited on supports of 17. Rousset, J. L., Bertolini, J. C., and Miegge, P., Phys. Rev. B 53, 4947
different acidic strengths. (1996).
18. Surface segregation phenomena (P. A. Dowben and A. Miller, Eds.).
CRC Press, Boca Raton, FL, 1990.
REFERENCES 19. Cadenhead, D. A., and Masse, N. G., J. Phys. Chem. 70, 3558
(1966).
1. Jan, C., Lin, T., and Chang, J., Ind. Eng. Chem. Res. 35, 3893 (1996). 20. Leon y Leon, C. A., and Vannice, M. A., Appl. Catal. 69, 305
2. Qian, W., Shirai, H., Ifuku, M., Ishiara, A., and Kabe, T., Energy Fuels (1991).
14, 1205 (2000). 21. Lee, J., and Rhee, H., J. Catal. 177, 208 (1998).
3. Fujikawa, T., Idei, K., Ebihara, T., Mizuguchi, H., and Usui, K., Appl. 22. Stanislaus, A., and Cooper, B. H., Catal. Rev. Sci. Eng. 36, 75
Catal. A 192, 253 (2000). (1994).