Promoting effect of ceria on the gas phase hydrogenation of crotonaldehyde over platinum catalysts

Article abstract of DOI:10.1006/jcat.1998.2199

The platinum-ceria interaction, induced after high temperature reduction of platinum based catalysts, has been used to improve the selectivity towards crotyl alcohol in the gas phase hydrogenation of crotonaldehyde. Three catalysts have been used in this

Full text of DOI:10.1006/jcat.1998.2199

JOURNAL OF CATALYSIS 178, 649–657 (1998)
ARTICLE NO. CA982199
Promoting Effect of Ceria on the Gas Phase Hydrogenation
of Crotonaldehyde over Platinum Catalysts
1
A. Sep u´ lveda-Escribano, F. Coloma, and F. Rodr ´ı guez-Reinoso
Departamento de Qu ´ı mica Inorg a´ nica, Universidad de Alicante, Apartado 99, E-03080 Alicante, Spain
Received February 20, 1998; revised May 14, 1998; accepted June 5, 1998
and the addition of a second metal such as tin (5–10) or iron
The platinum-ceria interaction, induced after high temperature (5,10,11) have been commonly used to increase the catalytic
reduction of platinum based catalysts, has been used to improve selectivity towards the unsaturated alcohol.
the selectivity towards crotyl alcohol in the gas phase hydro-
genation of crotonaldehyde. Three catalysts have been used in
this study, Pt/SiO2, Pt/CeO2, and Pt/CeO2/SiO2. They have been
characterized by X-ray diffraction and temperature-programmed
reduction, and their catalytic behavior in the gas phase benzene
hydrogenation of crotonaldehyde (2-butenal) over plat-
hydrogenation and crotonaldehyde hydrogenation reactions have
been determined after reduction treatments in flowing hydrogen
at 473 and 773 K. Results obtained in the gas phase hydrogenation
of crotonaldehyde show that the presence of ceria clearly promotes
It has been reported that the use of supports with
SMSI (strong metal-support interaction) properties pro-
duces an important enhancement of the selectivity towards
the unsaturated alcohols. Vannice et al. (2) compared the
inum dispersed on silica, alumina, and titania, after either
a high- or a low-temperature reduction. Whereas a 100%
selectivity for butanal was obtained with the Pt/SiO2 and
the hydrogenation of the carbonyl bond when the sample is re- Pt/Al2O3 catalysts, the high-temperature reduced Pt/TiO2
duced at high temperature, this being attributed to the presence of yielded a 37% of crotyl alcohol (2-butenol). The authors
oxygen vacancies or Ce3 cations in the neighbourhood of platinum attributed this effect to defect sites on titania created at the
+
particles, induced by the thermal treatment under hydrogen. The
promoting effect is more important for Pt/CeO2; crotonaldehyde
hydrogenation on this catalyst yields up to 80% crotyl alcohol
after reduction at 773 K, whereas it produces 100% butanal after
platinum–titania interface upon the reduction treatment at
high temperature. Similar results were obtained by Yoshi-
take et al. in a study on the deuteration of acrolein on plat-
inum supported on a series of partially reducible oxides:
reduction at only 473 K.
�c 1998 Academic Press
TiO2, Y2O3, ZrO2, and CeO2 (12).
Cerium oxide has become one of the most studied and
applied promoters in heterogeneous catalytic reactions dur-
ing the last years, mainly due to its important role in the
performance of three-way catalysts for the treatment of ex-
haust gas from automobiles(13). With regard to hydrogena-
tion reactions, cerium oxide has been used both as support
and additive in the hydrogenation of CO by metals such
as Pd and Rh (14–16), where it favours the formation of
oxygenates. Also, it has been reported that the addition of
CeO2 to Ru and Co (17) enhances the production of long
chain hydrocarbons and alkenes. In all cases, the effects of
ceria seem to be related to the activation of the C == O bond
through the interaction of the oxygen atom with oxygen
vacancies created at the ceria surface upon the reduction
treatment.
The main objective of this study has been to analyse the
promoting effect of ceria in the gas phase hydrogenation of
crotonaldehyde on platinum based catalysts, both when it is
used as support and when it is added as a promoter. To our
knowledge, this is the first study in which ceria dispersed
over an inert support has been used in the title or related
reactions.
1
. INTRODUCTION
Unsaturated alcohols constitute a very important group
of compounds for the chemical industry, as precursors to
prepare fine chemicals. They can be obtained by hydro-
genation of �,�-unsaturated aldehydes. This reaction can
produce the saturated aldehyde (through the hydrogena-
tion of the C == C bond), the unsaturated alcohol (through
the hydrogenation of the carbonyl C == O bond), the satu-
rated alcohol, and even the corresponding hydrocarbon. It
is not easy to achieve a profitable selectivity towards the de-
sired unsaturated alcohol with unpromoted metals, mainly
due to the fact that the C == C bond is preferentially reduced
by both thermodynamic and kinetic considerations. Thus,
monometallic platinum catalysts usually hydrogenate the
C == C bond with 100% selectivity, and it is necessary to pro-
mote the metal to obtain a certain selectivity towards the
hydrogenation of the carbonyl bond. Support effects (1–5)
1
Corresponding author. E-mail: REINOSO@UA.ES
649
0021-9517/98 $25.00
Copyright �c 1998 by Academic Press
All rights of reproduction in any form reserved.

´
´
6
50
SEPULVEDA-ESCRIBANO, COLOMA, AND RODRIGUEZ-REINOSO
. EXPERIMENTAL 3. RESULTS AND DISCUSSION
One of the support materials used in this study was 3.1. Characterization of the Fresh Catalysts
2
SiO2 (Silicagel 0.2–0.5 mm, pore diameter: � 6 nm, from
X-ray diffraction has been used to elucidate the degree
of platinum-support interaction in the catalysts after the
calcination treatment, and its effect on the state of the cata-
lysts after reduction in hydrogen. As indicated in the Ex-
perimental section, the fresh catalysts were stabilized by
calcination at 673 K. It has been evidenced by Raman spec-
troscopy studies that calcination of Pt/CeO2 catalysts re-
sults in the formation of Pt-O-CeO2 species that further
stabilizes the metal against sintering (18,19). These species
are destroyed upon reduction, but their formation as a first
step may influence the posterior behaviour of the reduced
catalysts and, at least, may hinder the platinum sintering,
yielding catalysts with better metal dispersion. Figure 1a
shows the DRX pattern of the Pt/SiO2 catalyst, in which the
2
� 1
Acros), with a BET surface area (N2 77 K) of 400 m g .
The Pt/SiO2 catalyst was prepared by the incipient wet-
ness method, impregnating the support with an ethano-
lic solution of H2PtCl6 of the appropriate concentration
to load about 1 wt% Pt. Then, the sample was dried at
3
83 K and calcined in air at 673 K for 4 h. The CeO2
support was prepared as follows: an aqueous solution of
NH4)2CO3 was slowly added to a magnetically stirred boil-
(
ing aqueous solution of Ce(NO3)3 � 6H2O (purity: 99.95% ,
from Ventron) at 423 K; the white precipitate of cerium
carbonate formed was filtered, washed with deionized wa-
ter, and dried overnight at 373 K in air; then, it was heated
�
1
in air at 5 K min up to 773 K and calcined at this tem-
2
� 1
perature for 4 h. It had a BET surface area of 70 m g .
The Pt/CeO2 catalyst was prepared as described above
for Pt/SiO2. Finally, a third support was prepared by im-
pregnating the silica support with the minimum amount of
aqueous solution of Ce(NO3)3 � 6H2O of the appropriate
concentration to obtain a CeO2 loading of 5 wt% . Then,
the sample was dried overnight in air at 383 K and cal-
cined at 673 K for 4 h. Platinum was introduced in the
same way as the other catalysts to prepare catalyst Pt/
CeO2/SiO2.
�
Pt(111) reflection at 2� = 39.65 and the Pt(200) reflection
�
at 2� = 47.36 appear as two intense peaks, indicating a rel-
atively large metal particle size. Figure 1b shows the XRD
pattern of Pt/CeO2/SiO2. Now, the diffraction peaks corre-
sponding to the ceria phase appear along with the peak for
�
the Pt(111) reflection at 39.65 , which is broader and worse-
defined than in Pt/SiO2. Finally, Fig. 1c shows the DRX pat-
tern of catalyst Pt/CeO2; in this case, the peak of platinum
�
at 39.65 is hardly discernible, indicating that, after calci-
nation of Pt/CeO2, platinum remains as a disperse phase.
From the comparison of these figures it can be concluded
that platinum is reduced to the metallic state in Pt/SiO2
after the calcination treatment in air at 673 K, whereas it
remains mainly in a disperse phase in the Pt/CeO2 sample.
X-ray diffraction patterns of samples, both just calcined
and reduced, were obtained with a JSO Debye-Flex 2002
system, from Seifert, fitted with a Cu cathode and a Ni filter,
�
� 1
and using a 2 min scanning rate.
Temperature-programmed reduction measurements on
calcined catalysts were carried out in a U-shaped quartz
3
� 1
cell, using a 5% H2/He gas flow of 50 cm min and about
.15 g of sample. Hydrogen consumption was followed by
0
on-line mass spectrometry and calibrated by carrying out
the reduction of CuO and assuming that it is completely
reduced to metallic copper.
Before the determination of their catalytic behaviour,
the catalysts were reduced in situ in flowing hydrogen
3
� 1
(
50 cm min ) at 473 K or 773 K for 10 h, with a heat-
�
1
ing rate of 5 K min . Benzene hydrogenation was stud-
ied in the temperature range from 323 to 398 K, with a
reactant mixture containing purified hydrogen and ben-
zene (Merck, reagent grade) in a H2/C6H6 ratio of 11.0.
Crotonaldehyde hydrogenation was performed at 313 K.
The reduced catalysts were contacted with a reaction mix-
ture (total flow: 50 cm min ; H2/CROALD ratio of 26)
formed by passing a hydrogen flow through a thermosta-
bilized saturator (293 K) containing the unsaturated alde-
hyde. A Carbowax 20 M 58/90 25–30-m capillary column
was used in both reactions to separate the reactants and the
reaction products, which were analyzed by gas chromato-
graphy.
3
� 1
FIG. 1. X-ray diffraction patterns of calcined catalysts: (a) Pt/SiO2;
(b) Pt/CeO2/SiO2; (c) Pt/CeO2.

PROMOTING EFFECT OF CERIA OVER PLATINUM
651
tion profiles of CeO2 (dashed line) and Pt/CeO2 (full line).
The TPR profile of CeO2 shows two peaks, centered at 830
� � 1
1
(
0.44 mmol H2 g ) and 1200 K (1.4 mmol H2 g ), which
are assigned to the reduction of CeO2 to Ce2O3. The first
peak, at a lower temperature, has been related to the reduc-
tion of the surface oxygen layer of CeO2, whereas the peak
at a higher temperature is assigned to the reduction of bulk
oxygen (20). The total hydrogen consumption (1.9 mmol
�
1
H2 g ) is lower than that corresponding to the complete
�
1
reduction of CeO2 to Ce2O3 (2.9 mmol H2 g ), indicat-
ing that the reduction is not complete in the temperature
range used (up to 1300 K). The TPR profile obtained with
Pt/CeO2 is also similar to those found in the literature (21-
2
2); The first intense peak centered at 570 K (0.65 mmol
�
1
H2 g ) is assigned to the surface reduction of ceria in close
contact with the metal, as well as the platinum reduction
(
breakdown of Pt-O-CeO2 species created upon calcina-
�
1
tion). The second peak, at about 665 K (0.14 mmol H2 g )
FIG. 2. X-ray diffraction patterns of catalyst Pt/CeO2/SiO2 reduced is related to the surface reduction of CeO2 not in close con-
at 473 (a) and 773 K (b).
tact with platinum. It has to be noted that this peak appears
at a lower temperature than for pure CeO2, this being re-
lated to the easier hydrogen availability by spillover from
An intermediate situation is obtained in Pt/CeO2/SiO2, in
the metal surface. Finally, the peak at 1170 K (1.84 mmol
which platinum placed close to the CeO2 particles could
�
1
H2 g ) corresponds to the bulk reduction of CeO2. In this
case, the total amount of hydrogen consummed (2.7 mmol
be interacting with it and forming Pt-O-CeO2 species, the
remaining platinum being in the metallic state as in the un-
promoted Pt/SiO2 catalyst.
�
�
1
g ) is closer to the one theoretically expected (2.9 mmol
1
g ), this indicating the easier reduction of ceria in the pres-
ence of platinum.
X-ray diffraction patterns of samples reduced at 473 K
under hydrogen are similar to those obtained for the cal-
cined samples, indicating that the low temperature reduc-
tion treatment does not affect the phase composition of the
catalysts, at least at a level discernible by this technique.
However, when the catalysts are reduced at 773 K, the
XRD pattern of sample Pt/CeO2/SiO2 does change appre-
ciably, those of Pt/CeO2 and Pt/SiO2 remaining unmodified.
Figure 2 compares the XRD patterns of catalyst Pt/CeO2/
SiO2 after reduction at 473 (a) and 773 K (b). It can be
seen how the diffraction peaks corresponding to the CeO2
phase are worse-defined after the high temperature reduc-
tion treatment. These changes could be ascribed to the par-
tial reduction of ceria taking place in the presence of hy-
drogen and the metal phase at this temperature, with the
subsequent formation of oxygen vacancies. The relatively
high ceria dispersion in this catalyst facilitates the observa-
tion of the changes by XRD. The same process is likely to
occur in the Pt/CeO2 catalysts but, in this case, the amount
of ceria affected is small, compared to the bulk support, and
the changes are not reflected in the XRD pattern.
Figure 4 shows the TPR profile of catalyst Pt/CeO2/SiO2.
The same features as for Pt/CeO2 can be observed, but with
a downward temperature shift for peaks centered at lower
(
500 K) and higher temperatures, the later being split into
3
.2. Temperature-Programmed Reduction
The degree ofreduction of ceria in the promoted catalysts
after the different reduction treatments has been estimated
by analyzing the TPR curves obtained with a 5% H2/He
flow. Figure 3 shows the temperature-programmed reduc-
FIG. 3. TPR curves (H2 consumption) for CeO2 (dotted line) and
Pt/CeO2 (full line).

´ ´
SEPULVEDA-ESCRIBANO, COLOMA, AND RODRIGUEZ-REINOSO
6
52
supported rhodium catalysts (26), because of the structure
insensitive character of this reaction (27). Thus, the cata-
lytic activity in this reaction will only depend on the num-
ber of metal atoms available at the surface of the catalysts
and would not be affected by the metal particle size or the
crystallographic plane. Under this assumption, the turnover
rate (number of benzene molecules transformed per sec-
ond and per surface metallic atom) should be the same,
at a given reaction temperature, for the different catalysts.
The amount of surface metal atoms could be then estimated
from the catalytic activity (reaction rate) in benzene hydro-
genation at a given temperature if the turnover number at
the mentioned temperature is known. With this approach,
Fajardie et al. (26) have obtained a very good correlation
between the amount of exposed metal atoms determined
by benzene hydrogenation and that calculated from hydro-
gen chemisorption in chlorine containing samples, where
hydrogen spillover from the metal atoms to the ceria sup-
port is inhibited (28). In chlorine-free samples, hydrogen
chemisorption was affected by hydrogen spillover to the
support and, thus, benzene hydrogenation gave more accu-
rate percentages of metal exposed. Also cyclohexane de-
hydrogenation, another structure insensitive reaction over
platinum, has recently been used to assess the accessible
metallic area in model three-way catalysts (29), although in
this case, the reduction treatment of the catalysts was car-
ried out at just 573 K. This temperature may be too low as to
produce the SMSI effect between ceria and the noble met-
als. Indeed, the determination of the exposed metal area by
using these reactions could be affected by this metal–ceria
interaction.
FIG. 4. TPR curve (H2 consumption) for Pt/CeO2/SiO2.
two peaks at 950 and 1075 K). The peak centered at 700 K is
again assigned to surface reduction of ceria not in close con-
tact with platinum. The fact that this reduction takes place
at a temperature somewhat higher than for Pt/CeO2 could
be due to the lack of continuity of the ceria surface (the sys-
tem is composed of ceria particles deposited on the silica
support), thus inhibiting the mobility of hydrogen spilt over
from the metallic phase. In a similar way, the splitting of the
high temperature peak would be due to the bulk reduction
of ceria placed in the proximity of platinum (950 K) and at a
larger distance (1075 K). In conclusion, it seems clear from
these figures that the LTR treatment (473 K) of Pt/CeO2
and Pt/CeO2/SiO2 does not produce ceria reduction; how-
ever, the HTR treatment (773 K) does produce the surface
reduction of ceria, with the creation of oxygen vacancies
Despite a large number of investigations, there is not a
complete agreement about the origin of the metal–ceria in-
teraction produced after high temperature reduction. Dec-
oration of metal particles by cerium suboxides patches
(
classic SMSI effect), formation of metal–cerium alloy and
and the presence of Ce3 ions. Also, the surface reduction
after the HTR treatment seems to be complete for Pt/CeO2
but not for Pt/CeO2/SiO2, where the spillover of hydrogen
from the metallic phase is hindered by the discontinuity of
the ceria surface.
+
electronic transfer from the metallic phase to the reduced
ceria have been invoked. The first two possibilities would
not affect the structure insensitive character of benzene
hydrogenation or cyclohexane dehydrogenation, but any
electronic effect of reduced ceria on the metal atoms would
modify their intrinsic activity; in this case, the use of the
mentioned reactions to determine the amount of metal ex-
3
.3. Benzene Hydrogenation
It is not easy to assess the metallic dispersion in supported posed would be questionable. To avoid this possibility, i.e.,
metal catalysts containing ceria (23). First, the poor con- the existence of electronic effects as the origin of the SMSI
trast between cerium and metals such as platinum makes it effect, it has been preferred in this study to use the catalytic
difficult to detect highly dispersed metal particles in trans- activity for benzene hydrogenation as a reference to ana-
mission electron microscopy images. On the other hand, lyze the effect of the reduction treatments on the catalytic
hydrogen or carbon monoxide chemisorption cannot be behavior towards a very different reaction, crotonaldehyde
used because these gases can also be adsorbed on the ceria hydrogenation.
phase, especially in the presence of metals, giving rise to
Table 1 reports the catalytic activity for benzene hydro-
�
1
� 1
partial reduction of ceria (24,25). The benzene hydrogena- genation (�mol s gPt ) and the apparent activation en-
tion reaction has been recently used as a tool for estimat- ergies obtained, for each catalyst, after reduction at 473
ing the percentage of metal exposed on low loaded ceria and 773 K. The increase in the reduction temperature for

PROMOTING EFFECT OF CERIA OVER PLATINUM
653
TABLE 1
account for these results; thus, the electron deficiency in
platinum would strength the platinum-benzene bond, hin-
dering the benzene hydrogenation. In conclusion, the pres-
ence of ceria hinders the platinum sintering upon the cal-
cination treatment of the catalysts in such a way that after
the LRT treatment (no SMSI effect) the ceria-containing
catalysts are more active for benzene hydrogenation than
the unpromoted Pt/SiO2 catalyst. After reduction at 773 K,
the SMSI effect induced in the ceria-containing catalysts
causes an important decrease in activity. This decrease is
even higher for Pt/SiO2 but, in this case, it has been at-
tributed to platinum encapsulation by the silica support.
This different deactivation mechanism will be better ap-
preciated after the analysis of the catalytic behavior in the
Benzene Hydrogenation at 373 K; Catalytic Activities and Appa-
rent Activation Energies after Reduction at 473 and 773 K
Reduction
temperature
(K)
Activity
Ea
(kJ mol
�
1
� 1
� 1
Catalyst
Pt/SiO2
(�mol s gPt
)
)
473
159.3
8.4
415.5
54.7
203.3
62.3
50
60
42
52
37
40
7
73
473
73
473
73
Pt/CeO2/SiO2
Pt/CeO2
7
7
Pt/SiO2 causes an important decrease in its catalytic activ- gas phase hydrogenation of crotonaldehyde, where, in ad-
ity; as no SMSI effect can be expected in this case, this effect dition to the catalytic activity, the product selectivity will
can be assigned to a loss of accessible platinum atoms. How- also be affected by the presence of ceria.
ever, as mentioned above, no differences could be appre-
ciated in the XRD patterns of the catalysts after reduction 3.4. Crotonaldehyde Hydrogenation
at both temperatures; i.e., the decrease in the number of
surface platinum atoms is not due the presence of larger
platinum particles after reduction at 773 K. Then, these re-
sults can only be explained by some degree of encapsulation
of the platinum particles by the silica support after the HTR
treatment.
After the reduction treatment at 473 K, catalyst Pt/
CeO2/SiO2 shows the highest catalytic activity, more than
twice that of Pt/SiO2. This result agrees, at least from a
qualitative point of view, with data obtained from XRD
measurements, and it is an evidence of the stabilizing ef-
fect of ceria against platinum sintering in an oxidizing en-
vironment (22). The catalytic activity is lower in Pt/CeO2,
probably because an important fraction of platinum is inter-
acting with the support and it is not active for the benzene
hydrogenation reaction.
Figure 5 shows the evolution of the overall activity
�moles of crotonaldehyde transformed per gram of plat-
(
inum and second) of catalysts Pt/SiO2, Pt/CeO2/SiO2, and
Pt/SiO2 in the gas phase hydrogenation of crotonaldehyde,
as a function of time on stream at 313 K, after being re-
duced at 473 K (LRT). Two different processes take place
simultaneously during the first stages of the reaction: the re-
action itself and the adsorption of the reactant (and, likely,
also the products) on the support. The adsorption equilib-
rium is reached after about 5 min on stream, depending on
the porous texture of the support and, consequently, ini-
tial reaction rates cannot be measured. During this stage, a
When catalysts are reduced at high temperature, 773 K, a
drastic decrease in the activity for benzene hydrogenation is
observed for the two ceria-containing catalysts. This effect
is related to the strong metal–support interaction induced
upon the thermal treatment. As commented above, several
explanations have been invoked to explain these results.
For example, it has been argued that the decrease in both
hydrogen chemisorption and benzene hydrogenation ac-
tivity observed in Pt/CeO2 catalysts after high temperature
reduction can be due to the formation of an intermetallic
compound such as Pt5Ce. This would cause the dilution of
platinum and the consequent decrease of the amount of
accessible platinum atoms (30). The coverage of platinum
crystallites by ceria patches, after reduction, has also been
invoked as the origin of the SMSI in the Pt-CeO2 system
(
31), and it could also explain the above-mentioned results.
Also an electronic transfer from the platinum particles to
FIG. 5. Evolution of catalytic activity for crotonaldehyde hydrogena-
the partially reduced ceria, which would stabilize the an- tion for samples Pt/SiO , Pt/CeO /SiO , and Pt/CeO as a function of time
2
2
2
2
ionic vacancies created after the reduction treatment, could on stream at 313 K, after reduction at 473 K.

´ ´
SEPULVEDA-ESCRIBANO, COLOMA, AND RODRIGUEZ-REINOSO
6
54
very high crotonaldehyde consumption is observed, which Pt/SiO2 and Pt/CeO2 show a very close activity (1.0 and
�
1
� 1
is not matched by the evolution of products. Once the car- 0.8 �mol s gPt , respectively), whereas Pt/CeO2/SiO2 re-
�
1
� 1
bon balance is obtained, i.e., once the amount of croton- mains more active (7.5 �mol s gPt ), despite the stronger
aldehyde leaving the reactor is equal to that fed minus that deactivation.
transformed into the products, reaction rates deeply de-
The hydrogenation ofa molecule such ascrotonaldehyde,
crease during, at least, 20 min on stream. This deactiva- containing two conjugated double bonds (C == C and C == O)
tion of the catalysts has been related to the decarbonyla- can yield a series of products, mainly depending on the pref-
tion of the reactant molecule yielding carbon monoxide, erential hydrogenation of one of them. Hence, the hydro-
which is irreversibly adsorbed on platinum at the reaction genation of the olefinic carbon-carbon double bond yields
conditions (32). Unfortunately, limitations in the column butanal, whereas the hydrogenation of the carbonyl C == O
used for the GC analysis made it impossible to determine bond yields crotyl alcohol:
the presence of C3 hydrocarbons in our system, given that
CH3-CH == CH-CHO + H2 → CH3-CH2-CH2-CHO
CH3-CH == CH-CHO + H2 → CH3-CH == CH-CH2OH.
they cannot be separated from butane. It can be seen in
Fig. 5 that catalytic activities after 6 min on stream fol-
low the order Pt/CeO2/SiO2 > Pt/SiO2 > Pt/CeO2. Catalyst
Pt/CeO2/SiO2 was also the most active for benzene hydro- However, these compounds can be further hydrogenated
genation, and this can be ascribed to its higher platinum to produce butanol and butane:
dispersion, assuming the absence of SMSI effect after re-
CH3-CH2-CH2-CHO + H2 → CH3-CH2-CH2-CH2OH
CH3-CH == CH-CH2OH + H2 → CH3-CH2-CH2-CH2OH
duction at 473 K. However, Pt/SiO2 and Pt/CeO2 showed a
similar activity for benzene hydrogenation after reduction
at 473 K (somewhat higher for Pt/CeO2), but the ceria-
supported catalyst is much less active for crotonaldehyde
hydrogenation. This can be seen better in Fig. 6, where the
ratios between the corresponding activities (after 6 min on
CH3-CH2-CH2-CH2OH + H2
CH3-CH2-CH2-CH3 + H2O.
→
stream) for crotonaldehyde hydrogenation and for benzene And crotyl alcohol may also isomerize to yield butanal:
hydrogenation are plotted. After reduction at 473 K cata-
CH -CH == CH-CH OH → CH -CH -CH -CHO.
lysts Pt/SiO2 and Pt/CeO2/SiO2 present a very similar ratio,
but that of Pt/CeO2 is much lower. Vannice and Sen found
similar results when comparing silica- and titania-sup-
ported platinum catalysts, the latter being reduced at low
temperature (473 K) (2). After 90 min on stream, when
a nearly steady-state activity has been achieved, catalysts
3
2
3
2
2
The extent of these secondary reactions is determined by
the overallconversion, in such a waythat selectivitytowards
crotyl alcohol decreases as the conversion degree increases
(
2,4,7,8).
The hydrogenation of crotonaldehyde over the three
catalysts under study after reduction at 473 K, in the exper-
imental conditions used, yields only butanal; i.e., only the
olefinic C == C bond is hydrogenated. This is the expected be-
haviour for monometallic platinum catalysts in the absence
of any support effect (2,4), although some authors have re-
ported that an important selectivity to crotyl alcohol can
be obtained with Pt/SiO2 catalysts if platinum particles are
large enough (32). This latter behaviour is not observed
here even although the platinum dispersion is very low,
especially for Pt/SiO2. Thus, it can be concluded that the
presence of ceria does not affect the selectivity of platinum
after low temperature reduction (473 K).
The catalytic behaviour of ceria containing catalysts dras-
tically changes after high temperature reduction (773 K).
Figure 7 shows the evolution of the overall activity with
time on stream at 313 K after reduction at 773 K. Now,
the activities after 6 min on stream follow the order Pt/
CeO2 > Pt/CeO2/SiO2 > Pt/SiO2. That is, after reduction at
high temperature, the presence of ceria promotes the cata-
lytic activity in a very important way. The activity ratios
between crotonaldehyde hydrogenation and benzene hy-
drogenation after this reduction treatment are plotted in
FIG. 6. Activity ratios for crotonaldehyde hydrogenation and ben-
zene hydrogenation for catalysts reduced at 473 and 773 K.

PROMOTING EFFECT OF CERIA OVER PLATINUM
655
Figure 8 plots the evolution of the selectivity towards
crotyl alcohol with time on stream, at 313 K, after reduc-
tion at 773 K; in all cases, the overall conversion degree is
below 8% (some values are indicated in the figure). The dif-
ferent behaviour of each catalyst is clearly evidenced. As
mentioned above, unlike Pt/SiO2 both Pt/CeO2/SiO2 and
Pt/CeO2 yield crotyl alcohol, the latter to a greater extent.
Production of crotyl alcohol with Pt/CeO2/SiO2 is initially
verylow, but it increaseswith time on stream to yield around
1
0% after 20 min. On the other hand, the carbonyl bond
is nearly selectively hydrogenated on Pt/CeO2 in the first
stages of the reaction, yielding around 80% crotyl alcohol
and, then, selectivity decreases. This loss of selectivity, slow
during the first 50 min and then faster, may be due to a dif-
ferent deactivation mechanism. It has to be noted that the
selectivity values presented are not affected by the conver-
sion degree at which they have been obtained, given that
the expected effect is a decrease of the selectivity towards
crotyl alcohol as the conversion degree increases, due to the
occurrence of secondary reactions.
Some recent papers have dealt with the beneficial effect
of using titania as support for platinum to prepare catalysts
which are very selective towards the formation of crotyl
alcohol from crotonaldehyde (2, 32). This effect, which is at-
tained mainly after reduction at high temperature (773 K),
is attributed to the removing of oxygen atoms from TiO2 to
form titania suboxides, TiOx, which can decorate the plat-
inum particles. Thus, crotonaldehyde would adsorb onto
platinum through the C == C double bond, but the presence
FIG. 7. Evolution of catalytic activity for crotonaldehyde hydrogena-
tion for samples Pt/SiO2, Pt/CeO2/SiO2, and Pt/CeO2 as a function of time
on stream at 313 K and after reduction at 773 K.
Fig. 6; all of them are enhanced with regard to those ob-
tained after reduction at 473 K, but the increase is higher
for Pt/SiO2. This could be due to a particle size or morphol-
ogy effect by which larger particles would be more active
(
32). Deactivation with time on stream strongly affects all
the catalysts, but it is faster for Pt/SiO2, in such a way that
it loses 95% of its initial activity within the first 20 min on
stream. After 90–100 min, when a nearly constant activity
has been achieved, Pt/CeO2/PtSiO2 remains as the most ac-
tive catalyst due to its slower deactivation as compared to
Pt/CeO2.
of coordinatively unsaturated Ti cations at the platinum–
titania interface would activate the C == O bond through the
interaction with the electron pair donor of the oxygen atom
From these results, no relevant effect of ceria on the cata-
lytic behaviour of platinum for the title reaction seems
to be easily envisaged. After reduction at low tempera-
ture, catalytic activities at 313 K obtained with Pt/SiO2 and
Pt/CeO2/SiO2 are consistent with results found in the lit-
erature for similar Pt/SiO2 catalysts (2) but the Pt/CeO2
catalyst is much less active than expected. On the other
hand, reduction at 773 K increases the catalytic activity
for crotonaldehyde hydrogenation as compared to benzene
hydrogenation, with Pt/CeO2 and Pt/CeO2/SiO2 reaching
nearly the same values. However, ceria does play its role
in modifying the intrinsic selectivity of platinum in the title
reaction. It has been mentioned above that the three cata-
lysts produced only butanal after reduction at 473 K. When
the reduction treatment is carried out at 773 K, Pt/SiO2
also yields only butanal, but the ceria-containing catalysts
produce both butanal and crotyl alcohol; i.e., these cata-
lysts are able to hydrogenate the carbonyl C == O bond. No
more hydrogenated products (neither butanol nor butane)
are obtained, and thus, a higher selectivity to crotyl alcohol
means a lower yielding of butanal.
FIG. 8. Evolution of selectivity towards crotyl alcohol for catalysts
Pt/SiO2, Pt/CeO2/SiO2, and Pt/CeO2 at 313 K and after reduction at 773 K.

´ ´
SEPULVEDA-ESCRIBANO, COLOMA, AND RODRIGUEZ-REINOSO
6
56
in the carbonyl group. In this way, the carbonyl group is
hydrogenated, by hydrogen adsorbed on platinum or spilt
over from it to the Pt-TiOx interface, at a higher rate than
if low reduction temperatures are used, and enhanced se-
lectivities towards crotyl alcohol are obtained.
4. CONCLUSIONS
Cerium dioxide is able to promote the synthesis of crotyl
alcohol from crotonaldehyde over platinum catalysts, by
favouring the hydrogenation of the carbonyl bond. This
effect, which is obtained only after high temperature re-
duction (773 K) of the catalysts, can be explained by the
existence of oxygen vacancies, or coordinatively unsatu-
This model can also be used to explain the observed
behaviour of the ceria-containing catalysts in this study:
creation of oxygen vacancies on ceria upon the high tem-
perature reduction treatment, which are able to activate the
carbonyl bond. Poisoning of these sites by strong adsorption
of reactant and/or products would account for the loss of
selectivity with time on stream, which is also accompanied
by the decrease in the overall activity. This classical strong
metal–support interaction (SMSI) effect, which is well es-
tablished for titania supported catalysts, is not as straight-
forward in the Pt/CeO2 system and, as mentioned above,
the formation of a Pt-Ce alloy has also been invoked to ex-
plain the different behaviour of high temperature reduced
Pt/CeO2 catalysts (30). Results presented in this paper can
add some light to this controversy.
rated Ce3 cations, near the platinum particles after the
reduction treatment. The interaction of the oxygen atom of
the carbonyl group in crotonaldehyde with these vacancies
can activate it, thus favouring its hydrogenation by hydro-
gen atoms adsorbed on the platinum particles or spilt over
from them to the Pt-CeOx interface. The main result is an
enhanced selectivity towards crotyl alcohol, which is as high
as 80% for Pt/CeO2 during the first stages of the reaction.
The selectivity is lower for the ceria-promoted platinum
catalyst, Pt/CeO2/SiO2, probably due to the lower amount
of ceria interacting with the platinum particles, compared
with the ceria-supported platinum catalyst.
+
Thus, the decrease in catalytic activity for benzene hydro-
genation can be explained by: (i) the formation of Pt-Ce
alloy, with the dilution of the active metal (Pt) by an in-
active one (Ce); (ii) the decoration of platinum particles
by patches of ceria suboxides; and (iii) electronic effects
on platinum due to the partial reduction of ceria. The dilu-
tion of platinum with cerium (alloy formation) would affect
the catalytic activity for crotonaldehyde hydrogenation as
well. But it is difficult to understand how it can influence
so strongly the selectivity towards the hydrogenation of the
carbonyl bond, unless the preferred adsorption mode of
crotonaldehyde onto the alloy surface is completely differ-
ent from that on platinum. On the other hand, no evidences
of platinum decoration by ceria have been obtained by
HRTEM in the Pt/CeO2 system (33), but the platinum parti-
cles are indeed surrounded by ceria (mostly in the Pt/CeO2
catalyst), whose surface is partially reduced. In any case, it
REFERENCES
1
. Richard, D., Gallezot, P., Neibecker, D., and Tkatchenko, Catal. Today
6, 171 (1989).
2
3
. Vannice, M. A., and Sen, B., J. Catal. 115, 65 (1989).
. Blackmond, D. G., Oukaci, R., Blanc, B., and Gallezot, P., J. Catal. 131,
401 (1991).
4
. Coloma, F., Sep u´ lveda-Escribano, A., Fierro, J. L. G., and Rodr ı´ guez-
Reinoso, F., Appl. Catal. A: Gen. 150, 165 (1997).
5. Da Silva, A. B., Jordao, E., Mendes, M. J., and Fouilloux, P., Appl.
Catal. A: Gen. 148, 253 (1997).
. Poltarzewski, Z., Galvagno, S., Pietropaolo, R., and Staiti, P., J. Catal.
02, 190 (1986).
. Coloma, F., Sep u´ lveda-Escribano, A., Fierro, J. L. G., and Rodr ı´ guez-
Reinoso, F., Appl. Catal. A: Gen. 136, 231 (1996).
8. Coloma, F., Sep u´ lveda-Escribano, A., Fierro, J. L. G., and Rodr ı´ guez-
Reinoso, F., Appl. Catal. A: Gen. 148, 63 (1996).
. Marinelli, T. B. L. W., and Ponec, V., J. Catal. 156, 511 (1995).
0. Marinelli, T. B. L. W., Nabuurs, S., and Ponec, V., J. Catal. 151, 431
1995).
1. Beccat, P., Bertolini, J. C., Gauthier, Y., Massardier, J., and Ruiz, P.,
J. Catal. 126, 451 (1990).
6
1
7
9
1
1
1
3+
seems clear that the existence of oxygen vacancies or Ce
(
cations in the neighbourhood of platinum particles can ex-
plain the results obtained for the selectivity towards crotyl
alcohol. The formation of Pt-Ce alloys and/or electronic ef-
fectsmayhelp to decrease the activityfor the hydrogenation
of the C == C bond, as has been shown in the platinum–tin
2. Yoshitake, H., and Iwasawa, Y., J. Chem. Soc. Faraday Trans. 88, 503
(
1992).
1
1
3. Trovarelli, A., Catal. Rev.-Sci. Eng. 38, 439 (1996).
4. Diagne, Idriss, H., Hindermann, J. P., and Kiennemman, A., Appl.
Catal. 51, 165 (1989).
system (7,8), which in turn favours the production of crotyl
alcohol. The small induction period observed in catalyst 15. Lavalley, J. C., Saussey, J., Lamotte, J., Breault, R., Hindermann, J. P.,
and Kiennemann, A., J. Phys. Chem. 94, 5941 (1990).
Pt/CeO2/SiO2, during which the selectivity to crotyl alco-
1
1
6. Turner, M. L., Byers, P. K., and Maitlis, P. M., Catal. Lett. 26, 55
1994).
hol increases from nearly 0 to about 10% (Fig. 8), could be
explained by the partial oxidation of cerium atoms in the
alloy under reaction conditions to yield active sites for ac-
(
7. Guerrero-Ruiz, A., Sep u´ lveda-Escribano, A., and Rodr ı´ guez-Ramos,
I., Appl. Catal. A: General 120, 71 (1994).
tivation of the carbonyl bond. This effect is not observed in 18. Brogan, M. S., Dines, T. J., and Cairns, J. A., J. Chem. Soc., Faraday
Trans. 90, 1461 (1994).
9. Murrel, L. L., Tauster, S. J., and Anderson, D. R., Stud. Surf. Sci. Catal.
the Pt/CeO2 catalysts, because the amount of such sites af-
1
ter reduction at high temperature is high enough as to yield
a very important selectivity to crotyl alcohol. Anyway, this
is just a tentative explanation, given that the presence of
such an alloy is not proved at this stage.
71, 275 (1991).
2
2
0. Yao, H. C., and Yao, Y. F. Y., J. Catal. 86, 254 (1984).
1. Trovarelli, A., Dolcetti, G., De Leitenburg, C., and Kaspar, J., Stud.
Surf. Sci. Catal. 75, 2781 (1993).

PROMOTING EFFECT OF CERIA OVER PLATINUM
657
2
2. Diwell, A. F., Rajaram, R. R., Shaw, H. A., and Truex, T. J., Stud. on
Surf. Sci. Catal. 71, 139 (1991).
on Catalysis Automotive Pollution Control, Capoc3, Brussels, April,
1994,” p. 275.
2
2
3. Duplan, J. L., and Praliaud, H., Appl. Catal. 67, 325 (1991).
4. Jim, T., Zhou, Y., Mains, G. J., and White, J. M., J. Phys. Chem. 91, 5931
29. Rogemond, E., Essayem, N., Frety, R., Perrichon, V., Primet, M., and
Mathis, F., J. Catal. 166, 229 (1997).
(
1987).
30. Meriaudeau, P., Dutel, J. F., Dufaux, M., and Naccache, C., Stud. Surf.
Sci. Catal. 11, 95 (1982).
31. Golunski, S. E., Hatcher, H. A., Rajaram, R. R., and Truex, T. J., Appl.
Catal. B: Env. 5, 367 (1995).
32. Englisch, M., Jentys, A., and Lercher, J. A., J. Catal. 166, 25
(1997).
2
2
5. Le Normand, F., Hilaire, L., Kili, K., Krill, G., and Maire, G., J. Phys.
Chem. 92, 2561 (1988).
6. Fajardie, F., Temp e´ re, J-F., Dj e` ga-Mariadassou, G., and Blanchard, G.,
J. Catal. 163, 77 (1996).
2
7. Bond, G. C., Garin, F., and Maire, G., Appl. Catal. 41, 313 (1988).
2
8. Bernal, S., Calvino, J. J., Cifredo, G. A., Gatica, J. M., Perez Omil, 33. Datye, A. K., Kalakkad, D. S., Yao, M. H., and Smith, D. J., J. Catal.
J. A., Laachir, A., and Perrichon, V., in “Third International Congress 155, 148 (1995).