Catalytic properties of Pt-Ge intermetallic compounds in the hydrogenation of 1,3-butadiene

Article abstract of DOI:10.1021/jp971117l

Catalytic properties of Pt-Ge intermetallic compounds, Pt₃Ge, Pt₂Ge, and PtGe, were studied for the H₂-D₂ equilibration and the hydrogenation of 1,3-butadiene. Powdered Pt, Ge, and the Pt-Ge intermetallic compounds were treated with H₂ at 873 K to reduce their surface oxidized in air. XPS spectra measured after the H₂ treatment showed the complete reduction into Pt⁰ and Ge⁰ and the stoichiometric surface composition for all the intermetallic compounds. The valence band XPS spectra suggested the electron transfer from Ge into Pt, which was also evidenced by the positive shift in the Ge3d XPS spectra. The activity of the intermetallic compounds for H₂-D₂ equilibration was much lower than that of pure Pt. Their low activity for the hydrogen dissociation resulted in the low activity for the hydrogenation of 1,3-butadiene. On the intermetallic compounds, no butane was formed at the initial stage of the reaction. After most of 1,3-butadiene was converted into butenes, the secondary hydrogenation of butenes into butane took place very slowly because of their intrinsically low activity for the hydrogenation of butenes. The hydrogenation with a mixture of H₂ and D₂ showed no inhibiting effect of 1,3-butadiene and 1-butene on the H₂-D₂ equilibration. Kinetic studies on the hydrogenation over Pt₃Ge revealed that the reaction rate was zero-order with respect to the partial pressure of hydrogen and first-order to that of 1,3-butadiene or 1-butene. The rate-determining step in the hydrogenation of 1,3-butadiene would be the adsorption of 1,3-butadiene. The strength of the adsorption was in the order of H₂ a?? 1,3-butadiene > 1-butene on Pt₃Ge. The weaker adsorption of butenes would result in the high selectivity to butenes in the hydrogenation of 1,3-butadiene.

Full text of DOI:10.1021/jp971117l

J. Phys. Chem. B 1997, 101, 5565-5572
5565
Catalytic Properties of Pt-Ge Intermetallic Compounds in the Hydrogenation of
1
,3-Butadiene
Takayuki Komatsu,* Shin-ichiro Hyodo, and Tatsuaki Yashima
Department of Chemistry, Tokyo Institute of Technology, 2-12-1 Ookayama, Meguro-ku, Tokyo 152, Japan
X
ReceiVed: March 28, 1997; In Final Form: May 6, 1997
Catalytic properties of Pt-Ge intermetallic compounds, Pt
equilibration and the hydrogenation of 1,3-butadiene. Powdered Pt, Ge, and the Pt-Ge intermetallic compounds
were treated with H at 873 K to reduce their surface oxidized in air. XPS spectra measured after the H
3
Ge, Pt
2
Ge, and PtGe, were studied for the H
2
-D
2
2
2
0
0
treatment showed the complete reduction into Pt and Ge and the stoichiometric surface composition for all
the intermetallic compounds. The valence band XPS spectra suggested the electron transfer from Ge into Pt,
which was also evidenced by the positive shift in the Ge3d XPS spectra. The activity of the intermetallic
2 2
compounds for H -D equilibration was much lower than that of pure Pt. Their low activity for the hydrogen
dissociation resulted in the low activity for the hydrogenation of 1,3-butadiene. On the intermetallic compounds,
no butane was formed at the initial stage of the reaction. After most of 1,3-butadiene was converted into
butenes, the secondary hydrogenation of butenes into butane took place very slowly because of their intrinsically
low activity for the hydrogenation of butenes. The hydrogenation with a mixture of H
inhibiting effect of 1,3-butadiene and 1-butene on the H -D equilibration. Kinetic studies on the
hydrogenation over Pt Ge revealed that the reaction rate was zero-order with respect to the partial pressure
2 2
and D showed no
2
2
3
of hydrogen and first-order to that of 1,3-butadiene or 1-butene. The rate-determining step in the hydrogenation
of 1,3-butadiene would be the adsorption of 1,3-butadiene. The strength of the adsorption was in the order
of H
2
3
. 1,3-butadiene > 1-butene on Pt Ge. The weaker adsorption of butenes would result in the high
selectivity to butenes in the hydrogenation of 1,3-butadiene.
Introduction
Sn supported on alumina, which is known to be a good catalyst
for the hydrocarbon reforming. In this catalyst system, Pt-Sn
IMC such as PtSn4, PtSn, and Pt2Sn have been proposed to be
the active species from the observation by transmission electron
Intermetallic compounds (IMC), sometimes called ordered
alloys, are known to be compounds between two or more metal
elements having simple stoichiometry. The most characteristic
feature of IMC is their specific crystal structure different from
that of their component metals, while usual alloys have the same
crystal structure as that of either component metals. The specific
structure sometimes gives them unique bulk properties, such
as shape memory effect, hydrogen storage ability, supercon-
ductivity, and so on. Therefore, IMC have been extensively
studied and utilized as new bulk materials. The surface
properties of IMC, however, have not attracted much attention
of scientists. In view of the catalysis, most of the investigations
9
microscopy. However, the presence of IMC has rarely been
reported though the formation of an alloy is often claimed under
10,11
reductive conditions.
This is because the bimetallic particles
are too small for observation by X-ray diffraction. Dautzenberg
1
2
et al. have detected the presence of PtSn on silica-supported
Pt-Sn catalysts after H2 reduction at 623 K for 3 h, while it
was detected on alumina supported one only after the reduction
1
3
at 923 K for 100 h. Bacaud et al. have investigated Pt-Sn/
Al2O3 by tin M o¨ ssbauer spectroscopy and proposed the presence
2
+
4+
of various Pt-Sn IMC in addition to the Sn and Sn ionic
1
on IMC have dealt with so-called hydrogen storage alloys
species.
because of their unique activity for hydrogen dissociation and
the possibility for the stored hydrogen to participate in the
reaction. Rare-earth-containing IMC are typical hydrogen
storage alloys studied for the catalytic reactions. Examples are
To clarify the catalytic properties of IMC would bring the
possibility of obtaining new catalyst systems as well as provide
information on the active species in bimetallic catalysts.
However, the catalytic properties of IMC other than those
containing rare-earth elements have never been clarified satis-
2
the reports on LaNi5 for the synthesis of ammonia and for the
3
hydrogenation of ethylene, LaCu5 for the synthesis of metha-
1
4
4
5
6
factorily. Bahia and Winterbottom have reported the hydro-
genation of 1,3-butadiene on ZrPd3. They found that ZrPd3 had
high selectivity for the formation of butenes compared with Pd
sponge, which catalyzed the hydrogenation of butenes into
butane after 1,3-butadiene was completely converted into
nol, ThNi5 and CeAl2 for the methanation of CO, and LaCu2
7
for the decomposition of 4-methyl-2-pentanol. As the rare-
earth elements are easily oxidized, the surface of their IMC could
not be the genuine surface with the intermetallic structure. In
the case of LaNi5, the surface is covered with La2O3 or La(OH)3
1
5
8
butenes. Llorca et al. have prepared PtSn IMC on silica by
the interaction of cis-[PtCl2(PPh3)2] and SnCl2 with the hydroxyl
groups of silica and found that the supported PtSn has higher
activity than usual Pt-Sn/silica for the reverse water-gas shift
reaction and lactic acid formation from CO2, ethylene, and H2O.
We have already reported on the catalytic properties of Co-
containing IMC, CoGe, CoGe2, CoSn, and CoAl.16 These
compounds gave much lower activity for H2-D2 equilibration
than did pure Co. However, they catalyzed the hydrogenation
and Ni; that is, the surface is similar to Ni metallic particles
supported on La2O3.
On the other hand, supported bimetallic catalysts have been
studied to enhance the activity, selectivity, and durability of
their parent, monometallic catalysts. A typical example is Pt-
*
Corresponding author. FAX: +81-3-5734-2758. E-mail: komatsu@
chem.titech.ac.jp.
X
Abstract published in AdVance ACS Abstracts, June 15, 1997.
S1089-5647(97)01117-6 CCC: $14.00 © 1997 American Chemical Society

5566 J. Phys. Chem. B, Vol. 101, No. 28, 1997
Komatsu et al.
of acetylene with a high selectivity to ethylene, while pure Co
produced almost exclusively butane with a much higher rate.
The unique selectivity of the Co-Ge IMC for the partial
hydrogenation was found to result from their very low activity
for the hydrogenation of ethylene.
The purpose of this study is to clarify the catalytic properties
of Pt-Ge intermetallic compounds compared with those of
parent Pt and Ge. The results together with those on Co-Ge
IMC reported earlier16 would provide further and fundamental
insights into the catalytic properties of IMC consisting of
transition and typical elements. Platinum is one of the most
active metal catalysts, while germanium is almost inactive
compared with the transition metals. The combination of these
two elements, which are greatly different in the physicochemical
nature, may generate unique catalytic properties as found in the
case of Co-Ge IMC. The Pt-Ge binary system has six
intermetallic phases, Pt3Ge, Pt2Ge, Pt3Ge2, PtGe, Pt2Ge3, and
17
PtGe2. Among them, we used Pt3Ge, Pt2Ge, and PtGe because
Ge-rich IMC may have a very low catalytic activity. In fact,
CoGe2 is less active than CoGe for the hydrogenation of
Figure 1. Temperature-programmed reduction of powdered Pt (0.50
g), Ge (0.050 g), and Pt-Ge intermetallic compounds (0.50 g).
1
6
acetylene. The hydrogenation of 1,3-butadiene was carried
out to compare the hydrogenating activity for the conjugated
diene with that for the monoolefin because the geometrical
environment of Pt atoms, such as the distance between Pt atoms,
as well as their electronic nature, may reflect on the catalysis
in the hydrogenation of two distant unsaturated bonds. In
addition, Ge is also an effective additive like Sn for supported
Pt reforming catalysts, and the formation of Pt-Ge alloy on
alumina18 has been reported. Therefore, the detailed study on
the catalytic properties of Pt-Ge IMC might shed light also on
the active species for the Pt reforming catalysts.
and D2 (99 at. %, 13.3 kPa) circulated through 0.50 g (Pt3Ge,
Pt2Ge, PtGe, and Ge) or 0.25 g (Pt) of the pretreated catalysts.
The extent of the equilibration reaction was monitored by the
change in the fraction of HD in hydrogen, HD/(H2 + HD +
D2), measured with the mass spectrometer. For hydrogenation
reactions, a mixture of 1,3-butadiene (1,3-BD) and H2 or
1-butene and H2 was circulated through 0.50 g of the pretreated
catalyst. Initial pressures were 6.6 kPa (1,3-BD and 1-butene)
and 19.9 kPa (H2), respectively. The extent of the reaction was
monitored by the decrease in total pressure measured with a
silicon diaphragm sensor, and the change in gas composition
was measured with GC. H2 and D2 were purified by passing
through the Mn/SiO2 and silica gel columns. 1,3-BD and
1-butene were purified by repeating a freeze-pump-thaw cycle.
The hydrogenation of 1,3-BD (6.6 kPa) with the mixture of H2
(10.0 kPa) and D2 (10.0 kPa) was carried out in a similar
manner.
Experimental Section
Catalyst Preparation. Pt-Ge intermetallic compounds
IMC), Pt3Ge, Pt2Ge, and PtGe, were prepared by arc melting
(
the mixture of stoichiometric amounts of platinum (Tanaka
Kikinzoku Kogyo, 99.95%) and germanium (Soekawa Chemi-
cals, 99.99%) granular metals under 400 Torr of argon. Ingots
thus obtained were crushed in air and filtered into fine particles
with diameters of less than 90 µm. Powders of pure Pt
Results and Discussion
(
9
Soekawa Chemicals, 99.9%) and pure Ge (Soekawa Chemicals,
9.99%) were also filtered to have diameters of less than 90
Bulk crystal phases of the prepared Pt-Ge intermetallic
compounds (IMC) were analyzed by the powder X-ray diffrac-
tion. The XRD patterns showed almost exclusively the peaks
of each aimed IMC. No diffraction peaks of oxide phases were
observed.
µm and used as reference catalysts.
Characterization. Temperature-programmed reduction (TPR)
of IMC and their component metals, Pt and Ge, crushed in air
was carried out under flowing H2(5%)/N2 gas mixture with a
TCD detector to measure the consumption of hydrogen. Trace
amounts of oxygen and water in the H2/N2 gas mixture were
removed by passing through the columns of Mn/SiO2 and silica
gel. The temperature was raised from 298 to 873 K at a heating
rate of 10 K min . The weight of the sample was 0.50 g,
except for the pure Ge (0.050 g). XPS spectra of IMC, Pt, and
Ge were obtained with an ESCALAB 220i spectrometer (Fisons
Instruments) with an X-ray source of Mg KR. The powdered
sample was pressed into a pellet and put into a pretreatment
chamber. After reduction with flowing hydrogen (101 kPa) at
While the IMC were crushed in air to get the fine powder,
their surface must be oxidized to some extent. In order to reduce
the oxidized surface to obtain the genuine surface of the
intermetallic phase, each IMC was treated with hydrogen at high
temperatures. Temperature-programmed reduction (TPR) of the
powdered IMC and their component metals was carried out to
know the optimum temperature for the complete reduction
without a significant sintering, which will decrease the surface
area. Figure 1 shows the TPR profiles of Pt, Ge, Pt Ge, Pt Ge,
-
1
3
2
and PtGe. Pure Pt powder gave two peaks at 390 and 720 K
corresponding to the hydrogen consumption. As Pt is known
to be easily reduced around room temperature, the 390 K peak
should be assigned to the reduction of the oxidized Pt surface,
though its small peak area indicates that the surface was not
oxidized extensively. The 720 K peak might result from the
hydrogen storage into the bulk. Pure Ge powder gave a large
amount of hydrogen consumption above 800 K though only an
one-tenth amount was used compared with the other samples.
This indicates that the surface was extensively oxidized in air
8
73 K for 1 h, the sample was transferred into the spectrometer
in vacuo to measure the spectra. Binding energies were obtained
using the C1s peak (284.5 eV) as a standard.
Catalytic Reaction. Catalytic reactions were carried out
using a glass circulation system connected to a quadrupole mass
spectrometer (Ulvac, MASSMATE 100). Before each reaction,
the sample was pretreated with 26.6 kPa of hydrogen at 873 K
for 1 h followed by an evacuation at 873 K for 10 min. H2-
D2 equilibration was performed with a mixture of H2 (13.3 kPa)
4+
and that the reduction of the oxidized Ge, probably Ge , needs

Catalytic Properties of Pt-Ge Compounds
J. Phys. Chem. B, Vol. 101, No. 28, 1997 5567
Figure 2. Pt4f XPS spectra measured before (- - -) and after (s) H
treatment at 873 K for 1 h.
2
Figure 3. Ge3d XPS spectra measured before (- - -) and after (s) H
treatment at 873 K for 1 h.
2
much higher temperature than that of the Pt reduction (390 K).
A small peak at 740 K might correspond to the reduction of
Ge .
The TPR of the powdered IMC gave different profiles from
those of the parent Pt and Ge. Pt3Ge and Pt2Ge gave a broad
peak at 390 K and a large peak at 670 K. It seems clear from
the profile of Pt that the broad peak results from the reduction
of Pt. The large peak could correspond to the reduction of Ge.
However, the peak temperature was much lower than that of
the pure Ge reduction. Obviously, the surface Ge atoms strongly
interact with Pt, resulting in the easy reduction. As we used
peak at 390 K (Figure 1). An increase in total intensity after
the treatment, as was also observed for the other samples, could
be ascribed to the slight sintering during the treatment at 873
K to make the surface flat to some extent.
2
+
Pure Ge powder showed two peaks at 29.6 and 33.2 eV,
0
which correspond to Ge (29.6 eV) and GeO2 (33.0 eV),
respectively. After the H2 treatment, the peak at 33.2 eV
completely disappeared, while the peak at 29.6 eV increased in
intensity. It is clear that the H2 treatment at 873 K totally
4
+
0
reduces Ge at the surface into Ge .
In the cases of Pt3Ge and Pt2Ge, the relative intensity of Pt4f7/2
peak was a little weaker before the treatment. This indicates
0
.5 g of IMC and 0.05 g of pure Ge for the TPR measurement,
the amount of the oxidized Ge is small on Pt3Ge and Pt2Ge
compared with the pure Ge. From these results, the TPR
profiles of Pt3Ge and Pt2Ge cannot be the sum of the profiles
of Pt and Ge. This could be an indication that no significant
phase separation into platinum metal and germanium oxide
occurs at the surface of these IMC. Pt metallic particles on
some oxide supports sometimes provide a significant amount
of dissociated hydrogen onto the oxide during the hydrogen
2
+
the presence of a small amount of Pt whose Pt4f7/2 peak
0
overlaps the Pt4f5/2 peak of Pt . The spectra after the treatment
were essentially the same as the spectrum of pure Pt. Ge3d
spectra gave a main peak at 32.6 eV with a smaller peak (Pt3Ge)
or a shoulder (Pt2Ge) at 30.5 eV. The H2 treatment eliminated
4
+
the Ge peak (32.6 eV) and gave a peak at 29.9 eV. Though
0
this binding energy is a little higher than that of Ge (29.5 eV),
2
+
1
9
it is far from Ge
(ca. 31 eV). We conclude that the H2
treatment, accelerating the reduction of the support oxide.
treatment at 873 K on Pt3Ge and Pt2Ge is sufficient to reduce
Therefore, there remains the possibility for the phase separation
if the same phenomenon occurs at the surface of IMC.
0
surface Ge species oxidized in air into Ge .
PtGe gave three peaks in its TPR profile. The reduction of
Pt is represented by the peak at 390 K as in the cases of Pt3Ge
and Pt2Ge, though its peak area was significantly large,
indicating that Pt in PtGe is oxidized easier than Pt in the other
Pt-rich IMC. The other two peaks at 720 and >870 K could
arise from Ge species. The two-step reduction suggests that
PtGe before the treatment exhibited the different spectra from
those of Pt3Ge and Pt2Ge. Pt4f peaks were very weak,
suggesting that the surface is covered with Ge species.
Moreover, the Pt4f5/2 peak was higher than the Pt4f7/2 peak,
0
and no Ge peak was observed, demonstrating the deep oxidation
of both Pt and Ge compared with Pt3Ge and Pt2Ge. This deep
oxidation coincides with the larger TPR peaks for PtGe than
those for Pt3Ge and Pt2Ge (Figure 1). However, the H2
treatment changed the spectra into almost similar ones obtained
for Pt3Ge and Pt2Ge, with an exception that the Pt4f spectrum
contained a Ge Auger peak around 77 eV. Therefore, PtGe
was also found to be reduced completely by the H2 treatment
2
+
Ge species exist in a comparable amount at the surface to
4+
give the 720 K peak or that Ge species having little interaction
with Pt coexists at the surface to give the peak at higher
temperatures.
From the above results, we chose the pretreatment temperature
of 873 K to reduce the oxidized surface, though the reduction
of Ge and PtGe may not be complete at the temperature.
Therefore, the oxidation state of the surface Pt and Ge atoms
was next examined by X-ray photoelectron spectroscopy.
Figures 2 and 3 show XPS spectra of Pt4f and Ge3d regions,
respectively. The spectra were measured before and after the
in situ H2 pretreatment at 873 K for 1 h. In the case of pure Pt
powder, two peaks at 71.3 and 74.6 eV were observed before
the treatment. These peaks were essentially the same as Pt4f7/2
20
at 873 K. In the case of Pt-Ge supported on alumina, H2
4
+
0
treatment at 923 K reduced a part of Ge into Ge , which was
proposed to form Pt-Ge alloy. Though the reduction of Ge in
Pt-Ge/Al2O3 was much easier than that of Ge in Ge/Al2O3, a
strong interaction between Ge and alumina might hinder the
reduction of Ge because the unsupported Ge was completely
0
reduced into Ge at 873 K in this study.
In the XPS spectra after the H2 treatment, the binding energies
of Pt4f7/2 peaks were almost the same among pure Pt (71.3 eV)
and IMC (71.3-71.4 eV), while those of Ge3d peaks were
slightly shifted from pure Ge (29.6 eV) to IMC (29.9 eV). The
higher Ge3d binding energy observed on IMC suggests that a
(71.3 eV) and Pt4f5/2 (74.6 eV) peaks of Pt foil. The binding
energies and peak shape did not change after the H2 treatment.
This indicates that no appreciable oxidation occurs at the surface
of Pt powdered in air, which was expected from the small TPR

5568 J. Phys. Chem. B, Vol. 101, No. 28, 1997
Komatsu et al.
Figure 4. Valence band XPS spectra measured after H
2
treatment at
Figure 5. H
2
-D
2 3
equilibration on Pt (O, 298 K), Ge (0, 573 K), Pt Ge
8
73 K for 1 h.
(9, 523 K), Pt
2
Ge (2, 523 K), and PtGe (b, 573 K).
TABLE 1: Surface Composition of Pt-Ge Intermetallic
hydrogen during the H2 pretreatment might result in the lower
equilibrium value of the HD fraction. Pt3Ge did not show any
activity for HD formation at 298 K. It required 523 K to
catalyze the reaction with a similar rate of HD formation to
that obtained on Pt at 298 K. Pt2Ge exhibited a little lower
activity at the same temperature, 523 K, while higher temper-
ature (573 K) was needed to get a comparable reaction rate on
PtGe. The fraction of HD on these IMC approached its
equilibrium value, 48.7% at 523 K or 48.9% at 573 K. Pure
Ge powder was the least active and gave a very low rate of HD
formation even at 573 K. The active site in the Pt-Ge IMC
Compounds
Pt composition/at. %
sample
stoichiometry
before reduction
after reduction
Pt
Pt
PtGe
3
Ge
Ge
75a
67
50
74
62
5
79
67
51
2
a
6-78 at. % in phase diagram.17
7
significant electron transfer from Ge to Pt occurs on IMC to
lower the electron density of Ge. We measured the valence
band XPS spectra to know further the electronic state in IMC
16
should be Pt atoms. We have already reported the much lower
activity of Co-Ge intermetallic compounds for the H2-D2
equilibration than that of pure Co.
(Figure 4). Pure Pt and IMC gave a peak near the Fermi level,
while pure Ge did not show any significant peak. The peak
near the Fermi level can be assigned to Pt5d orbitals. The peak
on IMC was obviously shifted from pure Pt with increasing Pt
content. This could correspond to the electron transfer from
Ge into Pt, which is in agreement with the peak shift in the
Ge3d spectra (Figure 3). A similar peak shift has been reviewed
for the valence band spectra of many IMC, such as Ni-Al,
Ni-Mg, Pd-Al, and Pd-La systems.21 In addition to the peak
shift, it is clear that the peak became narrower with Ge content.
The distance between Pt atoms should be longer in IMC than
that in pure Pt with fcc structure because of the dilution by Ge
atoms as well as the change in the crystal structure. This results
in the less overlap of Pt5d orbitals and the narrower band width.
Table 1 shows the surface composition of IMC calculated
from the core level XPS spectra. In the cases of Pt3Ge and
Pt2Ge, the surface Pt compositions were almost the same as
their stoichiometric values and did not change significantly by
the H2 treatment. Therefore, we conclude that Pt3Ge and Pt2Ge
have the genuine intermetallic phase even at their surface after
the H2 treatment at 873 K. On the other hand, PtGe exhibited
very low surface Pt composition before the H2 treatment as
expected from its weak Pt spectrum (Figure 2). This surface
enrichment in Ge before the reduction implies the presence of
the surface Ge species having little interaction with Pt. Such a
Ge species was already assumed to explain the two-step TPR
profile (Figure 1). The H2 treatment increased the Pt composi-
tion to reach the stoichiometric value. Therefore, we conclude
that PtGe also has the genuine surface intermetallic phase after
the H2 treatment.
Though we did not measure the specific surface areas of the
powdered catalysts, the activity for the H2-D2 equilibration was
found to decrease to a great extent with decreasing Pt content
except for the difference between Pt3Ge and Pt2Ge. We can
conclude that the activity of Pt-Ge IMC and their component
metals for the dissociation of hydrogen is in the order of Pt .
Pt3Ge g Pt2Ge . PtGe . Ge. The huge difference in the
activity between Pt and Pt-Ge IMC demonstrates that no phase
separation into Pt and Ge particles takes place at the surface of
the IMC. This again supports the regeneration of the genuine
intermetallic surface by the H2 treatment. The difference in
activity might be caused by the change in the electronic state
of Pt as can be seen in the valence band XPS spectra (Figure
4), because the shift of the Pt5d band to the higher binding
energy makes the spectrum relatively closer to that of Au, which
is much less active as a catalyst. However, the difference in
the geometric circumstances around the active Pt sites may not
be excluded to be the cause of the huge activity change.
Compared with Pt, the low activity of Pt-Ge IMC for the
H2-D2 equilibration implies their different activity and selectiv-
ity for the reactions containing hydrogen, for example, hydro-
genation. We have found the high selectivity for the partial
hydrogenation of acetylene into ethylene on Co-Ge IMC. In
this study, Pt-Ge IMC were used as a catalyst for the
hydrogenation of 1,3-butadiene (1,3-BD). Figure 6 shows the
results obtained on pure Pt; gas phase composition against the
reaction time (a) and the selectivity to each product against the
conversion of 1,3-BD (b). An excess amount of H2 was
supplied for the total conversion of 1,3-BD into butane. At
any reaction time, the decrease in total pressure agreed well
with the consumption of H2 calculated from the hydrogenated
products. Therefore, the formation and accumulation of coke
or oligomer on the surface can be neglected for the discussion
on the selectivity. 1-Butene, trans-2-butene, cis-2-butene, and
butane were the reaction products. The formation of butane
was obvious even at the initial stage of the reaction. This is a
1
6
The catalytic properties of Pt-Ge IMC were first studied for
H2-D2 equilibration. Figure 5 shows the change in the fraction
of HD in gas phase with reaction time. Catalysts were pretreated
with H2 at 873 K for 1 h to obtain the specific metallic or
intermetallic surface as mentioned above. On pure Pt, HD was
produced readily at 298 K. However, the HD fraction was
leveled off at about 40% and did not reach its equilibrium value,
4
7.4%, at this temperature. A significant amount of stored

Catalytic Properties of Pt-Ge Compounds
J. Phys. Chem. B, Vol. 101, No. 28, 1997 5569
Figure 8. Gas phase composition (a) and selectivity (b) in the
Figure 6. Gas phase composition (a) and selectivity (b) in the
hydrogenation of 1,3-butadiene on Pt at 298 K. 1,3-BD (9), 1-butene
hydrogenation of 1,3-butadiene on Pt
same as those in Figure 6.
2
Ge at 523 K. Symbols are the
(O), trans-2-butene (4), cis-2-butene (0), and butane (b).
was formed with a much slower rate than that of the initial
formation of butenes. It is clear that Pt3Ge has the high
selectivity for the butene formation in the hydrogenation of 1,3-
BD. The 1-butene/2-butene ratio at 50% conversion was 2.0.
After 1,3-BD was totally consumed, the isomerization of
1
-butene into 2-butenes gradually occurred.
As shown in Figure 8, Pt2Ge exhibited a similar tendency to
Pt3Ge at 523 K, that is, butane was scarcely formed until most
of 1,3-BD was converted into butenes, though an induction
period was observed for the butene formation. The rate of
1
-butene isomerization after 1,3-BD conversion exceeded 95%
was slightly higher, and the 1-butene/2-butene ratio was 2.9.
Because the activity and selectivity of Pt2Ge after the induction
period were close to those of Pt3Ge, a certain transformation of
surface structure might occur from the genuine Pt2Ge surface
structure into a Pt3Ge like structure during the induction period.
PtGe was less active than Pt3Ge and Pt2Ge. Therefore, the
reaction was carried out at 573 K as shown in Figure 9. Butane
was not formed significantly, and the 1-butene/2-butene ratio
at 50% conversion was estimated as 3.9. In the case of Pt alloy,
2
4
Bortolini and Massardier have reported that when Pt was
alloyed with Fe (Pt80Fe20), the activity for the hydrogenation
of 1,3-BD increased and that the selectivity to butenes also
increased from 58 to 89 mol % at 50% conversion. This
increase in selectivity was not so evident compared with Pt-
Ge IMC, where the selectivity was almost 100 mol %, though
the activity decreased much by forming IMC. This difference
in the effect of second element may result from the change in
the Pt crystal structure in the IMC as well as the difference in
the electronic nature of the second element.
From the above results, the characteristic catalysis by Pt-
Ge IMC in the hydrogenation of 1,3-BD is summarized as
follows compared with the pure Pt and Ge:
(1) The activity decreased in the order of Pt . Pt3Ge g Pt2Ge
> PtGe . Ge.
Figure 7. Gas phase composition (a) and selectivity (b) in the
hydrogenation of 1,3-butadiene on Pt
same as those in Figure 6.
3
Ge at 523 K. Symbols are the
typical result on Pt-supported catalysts.22,23 The rate of butane
formation increased with increasing reaction time especially after
most of 1,3-BD was reacted. Finally, gas phase contained
exclusively butane at 120 min of the reaction time. 1-Butene
was the main product in butene isomers. The isomerization of
1-butene to 2-butenes occurred very slowly; a 1-butene/2-butene
ratio at 50% conversion was about 3.0, which was slightly higher
2
2
than the value (1.9) obtained on Pt/Al2O3 at 288 K. The
activity of pure Ge was very low. Only 1.6% conversion was
obtained at 573 K for 2 h of the reaction.
The Pt-Ge IMC gave a comparable conversion to that on Pt
at 298 K when the reaction temperature was raised to 523 K
and higher. Figure 7 shows the result on Pt3Ge obtained at
(2) Very low activity for butane formation in the presence or
even in the absence of 1,3-BD resulted in the high selectivity
to butenes. The activity order was in good agreement with that
in H2-D2 equilibration (Figure 5). Therefore, the low activity
of Pt-Ge IMC for the 1,3-BD hydrogenation can be explained
by their low activity for the dissociation of hydrogen. In order
5
23 K. Butane was not formed at the initial stage of the
reaction, which is totally different from the result on Pt (Figure
). After most of 1,3-BD was converted into butenes, butane
6

5570 J. Phys. Chem. B, Vol. 101, No. 28, 1997
Komatsu et al.
Figure 9. Gas phase composition (a) and selectivity (b) in the
hydrogenation of 1,3-butadiene on PtGe at 573 K. Symbols are the
same as those in Figure 6.
Figure 10. Gas phase composition in the hydrogenation of 1-butene
on Pt
3 2
Ge at 523 K (a), Pt Ge at 523 K (b), and PtGe at 573 K (c).
to clarify the reason for the high selectivity of Pt-Ge IMC for
the butene formation, the hydrogenation of 1-butene was
performed under the equivalent conditions used for the 1,3-BD
hydrogenation.
1
-butene (O), trans-2-butene (4), cis-2-butene (0), and butane (b).
represented by the distance between the adjacent Pt atoms differs
among the Pt-Ge IMC. The larger difference in their activities
for the 1,3-BD hydrogenation might be an evidence for the
geometrical effect because 1,3-BD usually needs two Pt sites,
while 1-butene can be adsorbed on a Pt site.
For Co-Ge IMC, CoGe, and CoGe2, we have reported16 the
unique catalysis for the selective formation of ethylene in the
hydrogenation of acetylene comparing with the complete
hydrogenation to form ethane on pure Co catalyst. A significant
retardation of hydrogen dissociation by ethylene found on CoGe
has been proposed to be the main reason for the selective
formation of ethylene. The unique selectivity of the Pt-Ge
IMC in the 1,3-BD hydrogenation may also be caused by the
change in the hydrogen dissociation activity in the presence of
hydrocarbons.
Figure 11 shows the hydrogenation of 1,3-BD on Pt with a
mixture of H2 and D2 at 298 K. The fraction of HD in hydrogen
(Figure 11a) increased with reaction time, but its formation rate
was much lower than that without 1,3-BD (broken line) as
shown in Figure 5. The strong adsorption of 1,3-BD may retard
the dissociative adsorption of hydrogen. In Figure 11b, the
change in 1,3-BD fraction in the hydrogenation with H2 was
also shown by the broken line using the data in Figure 6. It
should be noted that the rate of the 1,3-BD conversion was lower
with H2/D2 than that with H2, showing the normal isotope effect
on the reaction rates.
On pure Pt, 1-butene was hydrogenated quickly into butane
within 0.5 min at 298 K with an evolution of the heat of reaction.
It is clear that Pt catalyzes the hydrogenation of 1-butene much
easier than that of 1,3-BD. The high activity for the 1-butene
hydrogenation must result in the simultaneous formation of
butane during the 1,3-BD hydrogenation. The rate of butane
formation shown in Figure 6, however, was much slower than
that measured separately without 1,3-BD. The inhibiting effect
of adsorbed 1,3-BD on the hydrogenation of 1-butene is obvious.
On the other hand, the rate of the 1-butene hydrogenation on
Pt-Ge IMC was lower than that of the 1,3-BD hydrogenation.
Figure 10 shows the results of the 1-butene hydrogenation on
Pt3Ge, Pt2Ge, and PtGe. The comparison between the conver-
sion rate of 1,3-BD in Figures 7, 8, or 9 and the formation rate
of butane in parts a, b or c of Figure 10 clearly demonstrates
the low activity of IMC for the 1-butene hydrogenation. PtGe
catalyzed mainly the isomerization into 2-butenes. In the cases
of Pt3Ge and Pt2Ge, the rate of butane formation in the 1,3-BD
hydrogenation after most of 1,3-BD was converted was almost
the same as that in the 1-butene hydrogenation on each catalyst.
It is concluded, therefore, that the main reason for the high
selectivity of IMC for the butene formation in the 1,3-BD
hydrogenation is their intrinsically low activity for the butene
hydrogenation, though the slight poisoning effect of coexisting
1
,3-BD to retard the butane formation was apparent on Pt3Ge
and Pt2Ge. The poisoning species could be the adsorbed 1,3-
BD, because the poisoning effect was not obvious after the
complete conversion of 1,3-BD. This development of activity
was seen in the rate of isomerization of 1-butene into 2-butenes
as well as in the rate of butane formation. Pt2Ge did not give
the induction period in the 1-butene hydrogenation (Figure 10b).
The induction period observed in the 1,3-BD hydrogenation
On the other hand, the HD formation was not affected by
1,3-BD on Pt3Ge as shown in Figure 12a compared with that
without 1,3-BD (broken line). The adsorption of 1,3-BD could
be weaker than that of hydrogen. A slight decrease in HD
fraction after 15 min of the reaction could correspond to the
change in the equilibrium caused by the exchange between gas
phase D2 or HD and H in the hydrocarbons. The effect of
1-butene was also examined in a similar manner, and it was
found that 1-butene did not affect the hydrogen dissociation,
either. Similar effects of coexisting 1,3-BD and 1-butene were
also observed on Pt2Ge. Therefore, there must be a sufficient
amount of hydrogen dissociatively adsorbed on Pt3Ge and Pt2Ge
for the hydrogenation. The adsorption of 1,3-BD and 1-butene
(Figure 8), therefore, could not be caused by butenes formed
in the reaction but by 1,3-BD.
The difference in the activities of Pt-Ge IMC for the
hydrogenation of 1,3-BD as can be seen in Figures 7a, 8a, and
9a was relatively large compared with that for the hydrogenation
of 1-butene shown in Figure 10. The geometry of Pt atoms

Catalytic Properties of Pt-Ge Compounds
J. Phys. Chem. B, Vol. 101, No. 28, 1997 5571
Figure 13. Dependence of the rate of 1,3-butadiene hydrogenation
on the partial pressure of hydrogen for Pt at 373 K (O), Pt
K (9), and Pt Ge at 483 K (2). P(C ) ) 2.7 kPa.
3
Ge at 483
2
4 6
H
2 2
Figure 11. Hydrogenation of 1,3-butadiene with H /D mixture on Pt
at 298 K. Change in the fraction of HD in hydrogen (a) and gas phase
composition (b) with reaction time. Broken lines indicate the HD
formation without 1,3-BD (a) and the 1,3-BD fraction in the hydro-
2
genation with H (b). Symbols are the same as those in Figure 6.
Figure 14. Dependence of the rate of 1,3-butadiene hydrogenation
on the partial pressure of 1,3-butadiene for Pt at 373 K (O), Pt Ge at
83 K (9), and Pt Ge at 483 K (2). P(H ) ) 13.3 kPa.
3
4
2
2
TABLE 2: Reaction Orders with Respect to the Partial
Pressure of Each Reactant
1
,3-BD hydrogenationa
1-butene hydrogenationb
catalyst
Pt
P(H
2
)
P(C
4
6
H )
P(H
2
)
4 8
P(C H )
1.0
0
0.9
0
1.0
0.4
Pt
Pt
3
Ge
Ge
0
0.6
1.0
0.7
2
a
Reaction temperatures were 373 K for Pt and 483 K for Pt
) ) 13.3 kPa and P(C
kPa. Reaction temperature was 523 K, and standard pressures were
P(H ) ) 13.3 kPa and P(C ) ) 2.7 kPa.
3
Ge and
H ) ) 2.7
Pt
2
Ge. Standard pressures were P(H
2
4 6
b
2
4 8
H
reaction rate thus measured was reproduced within 5% error
under the same reaction conditions. Figures 13 and 14 show
the dependence of the initial reaction rate on the partial pressures
of H2 and 1,3-BD, respectively, for Pt, Pt3Ge, and Pt2Ge. From
the slope of each line, the reaction orders for each catalyst were
calculated and are listed in Table 2 with the results obtained
for the hydrogenation of 1-butene in a similar manner.
Figure 12. Hydrogenation of 1,3-butadiene with H
2 2
/D mixture on
Pt Ge at 523 K. Change in the fraction of HD in hydrogen (a) and gas
3
phase composition (b) with reaction time. Broken lines indicate the
HD formation without 1,3-BD (a) and the 1,3-BD fraction in the
hydrogenation with H (b). Symbols are the same as those in Figure
2
6
.
might be the rate-determining step in their hydrogenation. From
Figure 12b, the inverse isotope effect was clearly observed on
Pt3Ge in the hydrogenation of 1,3-BD compared with the
hydrogenation with H2 (broken line). This will be discussed
later.
On Pt metal, the reaction rate was first-order to P(H2) and
zero-order to P(C4H6). This indicates that 1,3-BD is adsorbed
on Pt much more strongly than H2 and that the rate-determining
step is the adsorption of H2. This is a typical result for Pt-
22
The above results imply that the strength of the adsorption
is the important factor governing the selectivity in the hydro-
genation. Therefore, the kinetic study was carried out to know
the reaction order with respect to the reactants. To obtain better
reproducibility of the activity, the catalyst was used repeatedly
with an evacuation at the reaction temperature for 5 min between
the reactions. The weight of the catalyst was controlled to give
less than 5% conversion of 1,3-BD, and the reaction temperature
was selected to obtain the best reproducibility of the activity.
Pt2Ge did not show the induction period after the first run. The
supported catalysts. On Pt3Ge, however, the reaction rate was
zero-order to P(H2) and first-order to P(C4H6). It is revealed
that the adsorption of 1,3-BD is much weaker than that of H2,
the former step being the rate-determining step. This is the
opposite result to that on Pt. In the case of the 1-butene
hydrogenation, the reaction orders were zero for P(H2) and one
for P(C4H8), which indicates the weak adsorption of 1-butene
as in the case of 1,3-BD. The weak adsorption of 1,3-BD and
1-butene agrees well with the negligible effect of 1,3-BD and
1-butene on the H2-D2 equilibration as mentioned above.

5572 J. Phys. Chem. B, Vol. 101, No. 28, 1997
Komatsu et al.
On Pt2Ge, the reaction orders in the 1,3-BD hydrogenation
Conclusions
were 0.9 for P(H2) and 0.4 for P(C4H6) and those in the 1-butene
hydrogenation were 0.6 for P(H2) and 0.7 for P(C4H8). These
results suggest that Pt2Ge does not adsorb the particular molecule
strongly among 1,3-BD, 1-butene, and H2 to cover the surface
active sites. Because the activity and selectivity of Pt2Ge for
the reactions tested in this study were not so much different
from those of Pt3Ge, we expected the strong adsorption of H2
on Pt2Ge as was observed on Pt3Ge. The comparable strength
of the adsorption of 1,3-BD and H2 and the negligible effect of
The catalytic properties of Pt-Ge intermetallic compounds
for the hydrogenation of 1,3-butadiene are summarized as
follows compared with those of pure Pt.
(1) The activity for the hydrogenation is much lower than
that of pure Pt, and it decreases with decreasing the Pt content,
that is, in the order of Pt3Ge > Pt2Ge > PtGe. Their low
hydrogenation activity results from their low activity for the
dissociation of hydrogen.
(
2) The Pt-Ge intermetallic compounds have the high
1
,3-BD on the H2-D2 equilibration might be explained simul-
selectivity to butenes because of their lower activity for the
butene hydrogenation than that for the 1,3-butadiene hydrogena-
tion, while pure Pt has low selectivity to butenes due to its
extremely high activity for the butene hydrogenation.
taneously by the idea that the adsorption site of H2 is different
from that of 1,3-BD or that HD is formed through the adsorbed
reaction intermediate in the hydrogenation of 1,3-BD in addition
to the simple exchange between the dissociated H and D atoms.
The different reaction orders obtained between Pt3Ge and Pt2Ge
suggest that the surface structure and/or the electronic states of
Pt are different on these IMC. Therefore, the surface transfor-
mation into the Pt3Ge like structure proposed to explain the
induction period in the 1,3-BD hydrogenation on Pt2Ge should
be excluded. The induction period might correspond to the
formation of surface active species from the adsorbed 1,3-BD.
From the above results, the reaction mechanism is discussed
to explain the high catalytic selectivity of Pt3Ge for the
formation of butenes in the hydrogenation of 1,3-BD. On Pt,
(3) In the case of Pt3Ge, the lower activity for the butene
hydrogenation is caused by the much stronger adsorption of H2
and weaker adsorption of butene than that of 1,3-butadiene.
(4) The electron transfer from Ge to Pt5d orbitals as well as
the difference in the geometrical environment of Pt active sites
may result in the drastic change in the activity and selectivity
from those of pure Pt.
The low activity for hydrogenation and hydrogen dissociation
(1) and the high selectivity for partial hydrogenation (2) are
similar catalytic behaviors to those of Co-Ge intermetallic
compounds. This would provide general insights into the unique
catalytic properties of intermetallic compounds prepared by the
combination of transition and typical elements.
1
,3-BD and butenes are adsorbed more strongly than H2. The
normal isotope effect observed on the rate of the 1,3-BD
hydrogenation (Figure 11b) reasonably results from the differ-
ence in the zero-point energy of H2 and D2 because the
dissociation of hydrogen is the rate-determining step. The
poisoning effect of 1,3-BD on the hydrogenation of butenes is
evident from the fact that butane was formed slowly in the
presence of 1,3-BD (Figure 6) and that the hydrogenation of
References and Notes
(1) Wallace, W. E. CHEMTECH 1982, 12, 752.
(2) Takeshita, T.; Wallace, W. E.; Craig, R. S. J. Catal. 1976, 44, 236.
(3) Soga, K.; Imamura, H.; Ikeda, S. J. Phys. Chem. 1977, 81, 1762.
(4) Coon, V. T.; Takeshita, T.; Wallace, W. E.; Craig, R. S. J. Phys.
1
-butene was completed in a minute in the separate run.
Chem. 1976, 80, 1878.
Therefore, the strength of the adsorption is in the order of 1,3-
BD > butenes . H2 on Pt. The strongly adsorbed 1,3-BD
prevents hydrogen molecules from being adsorbed for the
hydrogenation. Butenes adsorbed relatively weakly do not
prevent the dissociative adsorption of hydrogen as strongly as
(5) Elattar, A.; Takeshita, T.; Wallace, W. E.; Craig, R. S. Science
977, 80, 1093.
1
(6) Luengo, C. A.; Cabrera, A. L.; MacKay, H. B.; Mayple, M. B. J.
Catal. 1977, 47, 1.
(7) Tkatchenko, D. B.; Branco, J.; de Matos, A. P. J. Phys. Chem.
995, 99, 5481.
1
(
(
8) Imamura, H.; Wallace, W. E. J. Phys. Chem. 1980, 84, 3145.
9) Chojnacki, T. P.; Schmidt, L. D. J. Catal. 1991, 129, 473.
1
,3-BD, resulting in their higher hydrogenation rate and the
higher selectivity to butane in the hydrogenation of 1,3-BD.
On the surface of Pt3Ge, 1,3-BD and 1-butene are adsorbed
more weakly than H2. The inverse isotope effect observed on
the hydrogenation of 1,3-BD (Figure 12b) and 1-butene can be
explained by their weak adsorption; that is, they can be adsorbed
on Pt3Ge slightly easier when the surface is covered by D instead
of H because of the weaker interaction of D with the surface.
From Figure 7a, the inhibiting effect of 1,3-BD was observed
on the formation of butane, that is, the hydrogenation of butenes.
This demonstrates the stronger adsorption of 1,3-BD than
butenes on Pt3Ge. As a result, the strength of the adsorption is
in the order of H2 . 1,3-BD > butenes. As the rate-determining
step is the adsorption of hydrocarbons, 1,3-BD has the advantage
of being easily hydrogenated compared with 1-butene. During
the hydrogenation of 1,3-BD, butenes formed from 1,3-BD are
easily desorbed into the gas phase and can hardly be adsorbed
again in the presence of hydrogen and 1,3-BD, leading to the
high selectivity to butenes. After 1,3-BD is completely
converted into butenes, the hydrogenation of butenes begins to
occur, though the rate of this hydrogenation is very low because
of the intrinsically low activity of Pt3Ge for the butene
hydrogenation.
(10) Merlen, E.; Beccat, P.; Bertolini, J. C.; Delivhere, P.; Zanier, N.;
Didillon, B. J. Catal. 1996, 159, 178.
11) Coloma, F.; Escribano, A. S.; Fierro, J.-L. G.; Reinoso, F. R. Appl.
Catal. A 1996, 136, 231.
12) Dautzenberg, F. M.; Helle, J. N.; Biloen, P.; Sachtler, M. H. J.
Catal. 1980, 63, 119.
(
(
(
(
13) Bacaud, R.; Bussiere, P.; Figueras, F. J. Catal. 1981, 69, 399.
14) Bahia, A.; Winterbottom, J. M. J. Chem. Tech. Biotechnol. 1994,
6
0, 305.
(15) Llorca, J.; de la Piscina, P. R.; Fierro, J.-L. G.; Sales, J.; Homs, N.
J. Catal. 1995, 156, 139.
16) Komatsu, T.; Fukui, M.; Yashima, T. 11th International Congress
on Catalysiss40th Anniversary. Stud. Surf. Sci. Catal. 1996, 101, 1095.
17) Binary Alloy Phase Diagrams, second edition; ASM International
and National Institute of Standards and Technology: Materials Park, OH,
990.
18) Garetto, T. F.; Borgna, A.; Apesteguia, C. R. 11th International
(
(
1
(
Congress on Catalysiss40th Anniversary. Stud. Surf. Sci. Catal. 1996, 101,
1155.
(
19) Khoobier, S. J. Phys. Chem. 1964, 68, 411. Benson, J. E.; Kohn,
H. W.; Boudart, M. J. Catal. 1966, 5, 307.
20) Bouwman, R.; Biloen, P. J. Catal. 1977, 48, 209.
(
(21) Fuggle, J. C.; Hillebrecht, F. U.; Zeller, R.; Zolnierek, Z.; Bennett,
P. A.; Freiburg, C. Phys. ReV. B 1983, 27, 2145.
(22) Bond, G. C.; Webb, G.; Wells, P. B.; Winterbottom, J. M. J. Chem.
Soc. 1965, 3218.
23) Boitiaux, J. P.; Cosyns, J.; Robert, E. Appl. Catal. 1987, 32, 145.
(24) Bertolini, J. C.; Massardier, J. Catal. Lett. 1991, 9, 183.
(