High performances of Pt/ZnO catalysts in selective hydrogenation of crotonaldehyde

Article abstract of DOI:10.1006/jcat.1999.2635

Hydrogenation of crotonaldehyde in the gas phase, at atmospheric pressure and 353 K over Pt/ZnO catalysts, was studied. Two types of precursor, (Pt(NH₃)₄(NO₃)₂ and H₂PtCl₆, referred to as A

Full text of DOI:10.1006/jcat.1999.2635

Journal of Catalysis 188, 165–175 (1999)
Article ID jcat.1999.2635, available online at http://www.idealibrary.com on
High Performances of Pt/ZnO Catalysts in Selective
Hydrogenation of Crotonaldehyde
,1
M. Consonni, D. Jokic, D. Yu Murzin,† and R. Touroude
LERCSI, UMR 7515 du CNRS, ECPM, ULP, 25 rue Becquerel, 67087 Strasbourg Cedex 2, France; and †Laboratory of Industrial
Chemistry, Abo Akademi, FIN-20500, Turku, Finland
Received May 6, 1999; revised July 1, 1999; accepted July 5, 1999
pharmaceutical production, and fine chemicals (1). One
of the challenging tasks, from both scientific and indus-
trial points of view, is selective (chemio-, stereo-, regio-,
or enantio-) reduction of polyfunctional molecules taking
part in these processes.
Selective hydrogenation of , -unsaturated aldehydes is
an interesting model reaction because these molecules si-
Hydrogenation of crotonaldehyde in the gas phase, at atmo-
spheric pressure and 353 K over Pt/ZnO catalysts, was studied.
Two types of precursor, (Pt(NH3)4(NO3)2 and H2PtCl6, referred to
as A and B catalysts, respectively, were used for catalyst prepara-
tion. Before the catalytic experiments the catalysts were reduced at
different temperatures. The reducibility of the support and the cata-
lysts was followed by TPR. Catalysts were also analysed by XPS and
XRD. Rapid deactivation during time on stream was observed. The
A and B catalysts showed different dependence on the reduction
temperature. Thus, the A catalyst had the highest activity when
reduced at 473 K; a further increase in the reduction temperature
led to a decrease in the activity, but at 673 K both catalysts A and
B showed nearly the same activity. On the B catalyst, the crotyl
alcohol selectivity reached a value as high as 75–80%, whatever the
reduction temperature. The B catalyst was better dispersed than the
A catalyst and formed a PtZn alloy at low reduction temperature
(473 K). It contained about 5 wt% chloride, whatever the reduc-
tion temperature. In contrast, Pt metal particles were only formed
on the A catalyst, reduced at 473 K, and then showed low selec-
tivity in crotyl alcohol. However, when the reduction temperature
was increased, activity decreased and crotyl alcohol selectivity in-
creased parallel to Pt–Zn alloy formation. One can speculate that Pt
sites, when alloyed to Zn, formed Pt –Zn + entities, on which the
crotonaldehyde adsorbed by the carbonyl group rather than by the
==
==
==
multaneously contain C O and C C bonds which, more-
over, form a conjugated system. So selective reduction of
==
==
O to obtain the unsaturated alcohol while C C remains
C
unaffected is a difficult task to achieve.
In thisfield ofresearch, monometalliccatalystssupported
on Al2O3 or SiO2 lead mostly to the formation of saturated
aldehydes (2). To improve selectivity toward unsaturated
alcohols, the use of additives (promoters), bimetallic cata-
lysts, or easily reducible supports has been proposed. Dif-
ferent results can be obtained, depending on the nature of
the metal itself. The effects of promotion were recently dis-
cussed in detail by Ponec (3). Promoters are thought to be
needed for the activation of the carbonyl group due to the
formation of a chemical bond between the oxygen of this
group and the promoter in its cationic state. For crotonalde-
hyde hydrogenation in the gas phase, much attention has
been given to the performance of Pt–Sn catalysts for which
the selectivity in unsaturated alcohol was found to depend
on the Sn/Pt ratio, the preparation method, and more pre-
cisely the Pt–Sn or Pt–Sn oxide interactions (4). Another
interesting system is platinum deposited on titania (5–8).
It was speculated that after a high-temperature reduction
Pt–titania interfacial sites are created, which are responsi-
ble for this enhancement.
C
C double bond. On the B catalyst, the high selectivity observed,
whatever the reduction temperature, led us to assume that besides
the alloying effect, chlorine has an important promotor effect by
increasing the polarity of Zn + in the PtZn catalytic sites and facili-
tating the carbonyl adsorption. A reaction network and mechanism
were put forward. Kinetic models, developed from the proposed
elementary step mechanisms, were used to discuss the influence of
c
support and promoters on reaction selectivity.
1999 Academic Press
Key Words: platinum; hydrogenation; crotonaldehyde; Pt/ZnO;
Besides TiO2, other reducible oxides such as Nb2O5,
Y2O3, ZrO2, and CeO2 used as supports of platinum have
been found to contribute to changes in the catalytic ac-
tivity and selectivity in acrolein (9) and crotonaldehyde
(10) hydrogenation. As the selective hydrogenation of cro-
tonaldehyde is known to be a reaction sensitive to metal–
support interactions, it was interesting to test the properties
of Pt/ZnO catalysts for this reaction, knowing that ZnO has
shown promoting effects in several studies. In the water-gas
unsaturated alcohol.
INTRODUCTION
Hydrogenation is used on a large scale in the chemical
and petrochemical industry, the food-processing industry,
1 To whom correspondence should be addressed.
165
0021-9517/99 $30.00
c
Copyright
1999 by Academic Press
All rights of reproduction in any form reserved.

166
CONSONNI ET AL.
shift reaction (11) and the methanation of carbon mono- Catalytic Tests
xide (12), which were conducted over Pt/ZnO catalysts,
Catalytic tests were carried out in a glass reactor, oper-
the activity increased with the reduction temperature treat-
ment; that was explained by an atomic rearrangement at the
metal–semiconductor interface (11) and by a significant li-
gand effect of Zn in the formed PtZn alloy (12). A marked
decrease in the catalytic activity with a significant increase
in selectivity was observed in butadiene hydrogenation to
olefins when the Pd/ZnO catalyst was reduced above 423 K.
It was suggested that Pd crystallites were “decorated” by
metallic Zn, leading to less availability of Pd adsorption
sites (13). Exceptional performance of Pd/ZnO catalysts in
steam reforming reactions, dehydrogenation of methanol
into methyl formate (14, 15), and hydrogen production by
partial oxidation of methanol (16) has been reported.
This paper is devoted to the hydrogenation of croton-
aldehyde in the gas phase on Pt deposited on ZnO, using
different metal precursors and different reduction tempera-
tures.
ated at atmospheric pressure. The total gas flow, controlled
by a flow meter, was varied by changing the pumping rate
at the end of the flow line. Crotonaldehyde (CROTAL)
supplied by Fluka puriss and stored in argon was used
as received. A known quantity of CROTAL (50–200 L)
was drawn up from the bottle using a tight syringe and in-
troduced through a vaccine cap into a reservoir, installed
on-line, and maintained at 273 K; therefore, CROTAL at
constant partial pressure (8 Torr) was carried over the cata-
lyst by the hydrogen flow (30–40 cc/min). The H2 gas was
first purified by passing it through a trap, maintained at
room temperature, containing the Pt/Al2O3 catalyst mixed
with zeolite to remove oxygen and water. Further purifi-
cation was made through an MnO trap at 293 K, installed
just before the CROTAL reservoir. Beyond the CROTAL
reservoir, the gas line was thermostated at about 333 K to
avoid any condensation. The stability of the CROTAL pres-
sure and the duration of the experiment were controlled
by two catharometers inserted upstream and downstream
with respect to the reactor, enabling the CROTAL flow rate
(0.22–0.32 mol s 1) to be measured. The reaction prod-
ucts were drawn off the flow line at different times dur-
ing the catalytic run and analyzed by gas–liquid chromato-
graphy (GLC) with a 30-m long, 0.5461 10 3-m diameter
DB-Wax column (J&W Scientific), at 358 K using a flame
ionization detector.
Before each catalytic experiment, the catalyst (100–
200 mg) was reduced at the desired temperature for 4 h and
cooled under H2 flow to the reaction temperature (353 K).
The reaction activities were calculated using the for-
mula A = ( F/ω), being the CROTAL conversion, F the
CROTAL flow in mol/s, and ω the weight of platinum in
grams. The selectivity to the different products, crotyl alco-
hol (CROTOL), butanal, butanol, hydrocarbons, and some
side products, was calculated as the molar ratio of the se-
lected product to the total products formed. The sensitivity
factors are taken as 1 for CROTAL, CROTOL, butanal,
butanol, and side products and 1.4 for the hydrocarbons.
The identification of the various hydrocarbons and side
products was occasionally made using a DB 624 chromato-
graphic column connected to a mass spectrometer (Carlo
Erba).
2. EXPERIMENTAL
Catalyst Preparation
Two Pt/ZnO catalysts, A and B, were prepared by wet
1
impregnation of ZnO (Asturienne des Mines, 42.6 m2 g
precalcined at 673 K) using aqueous solutions of respec-
tively tetraammineplatinum(II) nitrate (Pt(NH3)4(NO3)2)
and hexachloroplatinic acid (H2PtCl6); both salt and acid
were supplied by Strem Chemicals. The water was slowly
eliminated by evaporation on a hot plate. The samples
were dried overnight at 383 K, calcined in air at 673 K
for 4 h, and stored until use. Atomic absorption analyses
(AAS) were performed on the calcined samples (at CNRS,
Vernaison, France) to evaluate the metal and chlorine load-
ings (Table 1).
TABLE 1
Catalyst Description
A catalyst
B catalyst
H2PtCl6
ZnO support
Precursor
Pt(NH3)4(NO3)2
Asturienne
des Mines
(BET: 42.6 m2/g)
Pt content
(wt% )
5.0
5.0
It was checked that ZnO alone had no activity for croton-
aldehyde hydrogenation after H2 pretreatment up to 673 K.
Cl content (wt% )
Calc. at 673 K
Red. at 473 K
Red. at 773 K
Red. at 873 K
0.016
5.35
0.016
0.013
(Cl/Pt) atomic = 5.8
X-Ray Photoelectron Spectroscopy Analysis (XPS)
6.49
(Cl/Pt) atomic = 7.1
X-ray photoelectron spectroscopy (XPS) analyses were
performed using a VG ESCA III spectrometer with
MgK radiation (1253.6 eV) as an incident beam with-
out a monochromator. Before the analysis was conducted,
the samples were treated at different temperatures at
5.89
(Cl/Pt) atomic = 6.4
<0.009
0.21–0.72
(inhomogeneous)

HYDROGENATION OF CROTONALDEHYDE ON Pt/ZnO CATALYST
TABLE 2
167
atmospheric pressure in air or hydrogen in a preparation
space connected to the analysis chamber.
Hydrogenation of Crotonaldehyde on 5% Pt/ZnO (A: Metal
Precursor, Pt(NH3)4(NO3)2; B: Metal Precursor, H2PtCl6)
To obtain the Pt(4f) signals, the deconvolution proce-
dure was applied in the 64- to 99-eV BE range, taking into
account the Zn(3p) signals and their satellites (due to the
MgK 2, 3 X-ray excitation source). The Pt(4f 7/2) BE was
referred to Zn(3p3/2) BE at 88.3 eV obtained after the fit-
ting procedure. Theoretical curves were adjusted to fit the
peaks by a homemade computer program. This procedure
has been applied to the raw data points. A Lorentzian func-
tion was convoluted with an experimental Gaussian curve
(G = 0.8) and a Shirley background (17) was subtracted.
Then values of BE, (half-width at half maximum of the
Lorentzian curve), and peak area were deduced. The rela-
tive element intensities (Pt, Zn, Cl) were calculated from
peak surface ratio measurements, corrected by differences
in escape depths (a root square approximation was used)
and in cross-section (using Scofield’s data (18)).
Catalyst
A
B
T. red.
A ( mol/s/g of Pt)
473
7
573
1.8
673
1.5
473
4
573
3.3
673
1.6
Selectivity
HC
<1
63
15
21
1
<1
40
12
45
2
<1
47
4
49
1
1
15
1
78
5
1a
15
1
75
8b
<2
15
<1
81
2
Butanal
Butanol
CROTOL
Side products
a Butadiene, butenes, and butane.
b Chlorinated products (2–3% ), CROTOL isomers (3–4% ), and dienic
polymer (1–2% ).
on stream to reach a quasi steady state after about 2 h with
a relatively fast and profound deactivation. This was ob-
served with both A and B catalysts (Figs. 1A and 1B). In
Table 2 are reported the activities and selectivities obtained
in this quasi steady state after 2 h on stream. The A catalyst
had the highest activity when reduced at 473 K; a further
increase in the reduction temperature led to a decrease in
the activity, but at 673 K both catalysts, A and B, showed
nearly the same activity. The most remarkable result was
the difference in selectivities obtained with the two cata-
lysts. On the A catalyst, the crotyl alcohol selectivity was
low after 473 K reduction (21% ) and increased up to 45
and 49% after 573 and 673 K reduction. On the B cata-
lyst the crotyl alcohol selectivity reached a value as high as
75–80% , whatever the reduction temperature. It could be
noticed that some amounts of side products were formed,
mainly on the B catalyst reduced at 573 K. An attempt was
made to identify some of these using the GC mass coupling
analysis. It revealed the presence of chlorinated products,
the crotyl alcohol isomer, and dienic polymers. The produc-
tion of these side products was constant during all the time
on stream. The hydrocarbon analysis showed a mixture of
butane, butenes, and butadiene. In Fig. 2 we reported the
selectivities, recalculated by taking into account the direct
hydrogenated products, butanal, CROTOL, butanol, and
hydrocarbons, obtained on the B catalyst asa function ofthe
time on stream. It is remarkable that these selectivities were
independent of the reduction temperature. In the first min-
utes of the run, products from consecutive reaction, such as
butanol and hydrocarbons, were formed in slightly higher
quantities, giving then lower selectivity in CROTOL, but
the latter stabilized very quickly at 82 3% . This high value
was obtained in a conversion range going from 5 to 20% .
The XRD results are reported in Figs. 3A and 3B in the
X-Ray Diffraction Analysis (XRD)
XRD analyses were carried out in a Siemens D 5000
polycrystalline diffractometer using CuK radiation. Some
spectra were recorded in a rapid scanning mode (2.0 s/step,
step size 0.05 in 2 ) in a large 2 range (5 –100 ). Such
spectra showed all the lines of the zincite structure of ZnO
(JCPDS 36-1451) and other lines characteristic of the Pt
metal (JCPDS 4-0802) or the Pt–Zn alloy (JCPDS 6-0604);
the other PtxZny phases reported in the literature (JCPDS
23-0466, 12-0612, 6-0619, and 6-0584) were not found. For
better accuracy in the measurement of Pt and Pt–Zn lines,
the spectra were recorded in the 38.5–42.5 2 range; this
corresponds to the region of Pt and Pt–Zn (111) lines
(respectively, 39.752 and 40.790), with 0.01 in step size
and 5.0 s/step. The mean size of the particles (t) was re-
lated to the pure X-ray broadening using Scherrer’s for-
˚
mula (t = K /ε cos , where = 1.5406 A, ε = angular width
at half-height expressed in radians, K = Scherrer constant
(taken = 0.9 (19)), and = Bragg angle).
Temperature-Programmed Reduction (TPR)
The reducibility of the support and the catalysts was fol-
lowed by TPR in a gas flow system with 1.5% H2/Ar using
an X-Sorb-(S) instrument (GIRA society). First, the sample
(100 mg) was calcined in air at 673 K for 4 h, then cooled to
room temperature and flushed with Ar ( 20 min), followed
by a 1.5% H2/Ar mixture (20 min); after this the TPR was
performed, at 8 K/min until a temperature of 873 K was
reached, where the sample was kept for 1 h.
RESULTS
When the catalytic test was performed under crotonalde- 38.5–42.5 2 range. On one hand, the A catalyst, calcined at
hyde and H2 constant flow, at constant temperature (353 K), 673 K or reduced at 473 K, showed similar Pt lines, charac-
the overall activity was found to decrease during the time teristic of Pt metal; the particles can be estimated to 15 nm.

168
CONSONNI ET AL.
FIG. 1. (A) Hydrogenation of crotonaldehyde on 5% Pt/ZnO, A catalyst (metal precursor: Pt(NH3)4(NO3)2). (B) Hydrogenation of crotonaldehyde
on 5% Pt/ZnO, B catalyst (metal precursor: H2PtCl6).
When the reduction temperature was increased to 573 K ginning of ZnO reduction when the shape of the XAES
the Pt line intensity decreased while the Pt–Zn alloy line lines was analyzed as described in the literature (11). For
grew to become the most prominent line after 673 K reduc- the Pt(4f) lines, an example of the curve fitting is shown
tion temperature, revealing the presence of 18-nm PtZn in Fig. 4 and the quantitative results of all the samples are
particles. On the other hand, with the B catalyst, reduced reported in Table 3. When the catalysts were reduced, there
at 473 K, no line corresponding to Pt and PtZn was vis- was no noticeable change in the binding energy as the re-
ible; it was only after 673 K reduction temperature that a duction temperature increased; the (70.8 0.1)-eV value of
well-defined line appeared, indicating the presence of PtZn the binding energy corresponds to Pt in the reduced state.
particles, sized at 9 nm. After 473 K reduction temperature Differently, regarding the values, a narrowing of the line
the metal particles were too small to give a diffraction pat- width was clearly observed for the A catalyst: decreased
tern. Therefore, from XRD analysis, no conclusion can be from 0.45 to 0.39 and 0.23 when the reduction temperature
drawn as to the nature of these particles at this low reduc- was increased from 473 K to 573 and 673 K, respectively. On
tion temperature.
the B catalyst, with a reduction temperature as low as 473 K,
The XPS spectra showed no change in the binding ener- a low value (0.27) was obtained and decreased down to
gies and line shapes of all the XPS and XAES lines corre- 0.23 for 673 K reduction temperature. In the calcined state,
sponding to the Zn element on both catalysts and whatever the peaks were not unique: in the A catalyst, the Pt lines
the reduction treatment, on the condition that the reduc- corresponded mainly to the Pt metal (60% ) while in the B
tion temperature was less than 773 K. This result allowed catalyst 80% of Pt gave a component at 3.5 eV BE higher
us to take any Zn lines as an internal reference. It is only than the metal one. XPS studies of the oxygen–platinum
from the 773 K reduction treatment that we observed a be- (20) and chlorine–platinum (21) compounds reported in

HYDROGENATION OF CROTONALDEHYDE ON Pt/ZnO CATALYST
169
FIG. 2. Hydrogenation of crotonaldehyde on 5% Pt/ZnO, B catalyst (metal precursor: H2PtCl6) reduced at 473, 573, and 673 K.
the literature led us to assign the highest BE Pt component area of platinum on the B catalyst, meaning that smaller
to a chlorinated or oxychlorinated compound rather than particles have been formed on B than on A; this result was
PtO2 species since a large amount of chlorine remained on in accordance with XRD analyses.
the surface as discussed below. The other secondary com-
Chlorine was present in large amounts in the B cata-
ponents in both catalysts were located between 1 and 2 eV lyst, whatever the reduction temperature (between 5 and
higher than the metal one (72 and 72.4 eV) and should cor- 7 times the number of platinum atoms). Small amounts of
respond to some PtOx species, in which the Pt oxidation chlorine existed in the A catalyst after reduction at 473 K,
degree was hardly defined.
which disappeared at 673 K reduction temperature, but one
The (Pt/Zn) atomic ratios were about twice as high on the has to notice that the chlorine atoms were also present
B catalyst as on the A catalyst, revealing a larger surface in the calcined ZnO support and in the same proportion
TABLE 3
XPS Analyses of Pt(4f ) and Cl(2p) Lines
100 (Pt/Zn)
atomic ratio
BE (Pt(4f )7/2
)
(Pt(4f ))
100 (Cl/Zn)
atomic ratio
(eV)
(Lorentzian half-width)
theoretical = 2.2
A
B
A
B
A
B
A
B
Calc. at 673 K 4 h
70.7
72.4
72.0
74.2
0.50
0.50
0.37
0.32
1.2
0.6
0.9
3.2
7
22
Red. at 473 K 2 h
Red. at 573 K 2 h
Red. at 673 K 2 h
Red. at 673 K 4 h
70.7
70.7
70.9
70.9
70.8
70.7
70.7
70.7
0.45
0.39
0.34
0.23
0.27
0.25
0.23
0.23
1.8
1.8
1.7
1.4
3.7
3.7
2.9
2.9
4
2
ε
ε
20
19
14
19

170
CONSONNI ET AL.
FIG. 3. (A) XRD analysis on 5% Pt/ZnO, A catalyst (metal precursor: Pt(NH3)4(NO3)2). (B) XRD analysis on 5% Pt/ZnO, B catalyst (metal
precursor: H2PtCl6).
as that in the calcined A catalyst. The AAS analyses con- Pt/Al2O3 catalysts were concerned (23):
firmed the results from XPS analyses, as can be seen from
Table 1.
The TPR profiles of the A and B catalysts were totally
Pt(OH)4Cl2 + 3H2 → Pt + 4H2O + 2HCl.
different (Fig. 5). On the B catalyst a large peak of H2 con- We could estimate that similar species were formed on this
sumption appeared between 473 and 573 K with an H2/Pt calcined H2PtCl6/ZnO catalyst, knowing that Cl atoms were
ratio equal to 3.3. In the Pt/Al2O3 systems, such a high- present on the surface (from AAS and XPS analyses) and
temperature reduction peak was usually assigned to the re- high BE was found for the Pt(4f) line in XPS. Different
duction of oxychlorinated platinum species (22). Moreover, from the B catalyst, on the A catalyst a small consumption
a stoichiometry, H2/Pt = 3, has already been found and at- of H2 occurred at 393 K, corresponding to H2/Pt equal to
tributed to the presence of Pt(OH)4Cl2 when chlorinated 0.13. This suggested that the particles in this A catalyst were

HYDROGENATION OF CROTONALDEHYDE ON Pt/ZnO CATALYST
171
FIG. 4. XPS analysis, example of fit in Pt(4f )–Zn(3p) BE range (5% Pt/ZnO, B catalyst (metal precursor: H2PtCl6) reduced 4 h at 673 K).
FIG. 5. TPR on 5% Pt/ZnO (after calcination at 673 K).

172
CONSONNI ET AL.
TABLE 4
surrounded by platinum oxide while the bulk was formed
of platinum metal, as already reported in the literature (24).
Moreover, this result corroborated the preceding analyses
by XRD and XPS: the XRD spectrum, representative of the
particle bulk, revealed the presence of Pt metal for the cal-
cined A catalyst while XPS results in which the contribution
of the surface was important showed both oxide and metal
platinum on the A catalyst and only chlorinated species on
Kinetic Parameters in the Deactivation Process
(r = a3 + a1 exp( a2t))
Catalyst
a1
a2
a3
ssa
A, 473 K
A, 573 K
A, 673 K
31.31 1.57
18.53 0.65
1.74 0.14
16.80 1.20
10.10 2.50
2.54 0.27
0.041 0.004
0.047 0.003
0.024 0.004
0.029 0.006
0.011 0.007
0.003 0.001
6.99 0.63
1.73 0.19
1.38 0.06
3.99 1.00
0.70 3.16
0
0.99
0.13
0.003
0.49
0.42
0.033
the B catalyst. The reduction of ZnO on the A catalyst was B, 473 K
apparent between 523 and 873 K with two maxima at 658
and 813 K. A part of the reduced Zn was engaged in the
B, 573 K
B, 673 K
a Sum of squares of the relative deviations.
Pt–Zn alloy formation. This phenomenon also occurred on
the B catalyst since the baseline after the platinum species
reduction peak had not recovered the zero level, indicating
that a continuous reduction process occurred; moreover,
in the 473–573 K temperature range the excess hydrogen
which is not involved in the reduction of Pt(OH)4Cl2
(H2/Pt > 3) is used for the reduction of ZnO to form the
PtZn alloy.
Table 4; a1 + a3 represents the initial rate and a3 the rate
at infinite time. It is worth noting that the B catalyst, af-
ter reduction at 473 K, retains significant activity at infinite
time, knowing that it shows an exceptional crotyl alcohol
selectivity.
Kinetics
DISCUSSION
For further discussion let us briefly analyze the reaction
network and reaction kinetics. The kinetic network was
originally proposed by Simonik and Beranek (26) and fur-
ther developed in (6). In the original treatment of Simonik
and Beranek, isomerization from crotyl alcohol into bu-
tyraldehyde was assumed. In (25), we even supposed that
the opposite isomerization reaction from butyraldehyde
into crotyl alcohol takes place, due to specific features of the
support which was applied and hence to the reactivity of the
Deactivation
Experimental data on selectivity dependence on time on
stream (Fig. 3) showed that, for the B catalyst, the selectiv-
ities, measured in the 5–20% conversion range, were inde-
pendent of the reduction temperature pretreatments and
then of the activities.
As we pointed out (25), from a kinetic point of view, such
a behavior implies that deactivation phenomena and com-
plex reaction kinetics can be treated separately. Similarly
to our previous consideration (25) one can speculate that
catalyst decay, at least for the B catalyst, is due to the for-
mation of coke deposits from the starting products parallel
to the main hydrogenation reaction. We assume that coke
(or carbon deposit) is produced from adsorbed crotonalde-
hyde. Deactivation and self-regeneration proceed simulta-
neously; moreover, these steps are essentially slower than
the reaction steps. This consideration allows one to apply
a steady-state hypothesis only to hydrogenation but not to
the deactivation steps. A detailed kinetic analysis was pre-
sented in (25). Here, we present only the final equation for
the reaction rate,
==
C
O bond. However, calculations in (25) showed that this
isomerization step can be neglected. In the present paper
for the more general treatment we assume that the isomer-
ization step is reversible. The kinetic considerations in (6)
were somewhat simplified, as it was assumed that there ex-
ists only one equilibrium constant for crotonaldehyde ad-
sorption. However, this molecule can be adsorbed by an
==
olefinic bond as well as by a C O bond. Moreover, both
unsaturated bonds can be bounded to the metal surface, en-
hancing crotonaldehyde adsorption by the resonance effect
(delocalization of electron density).
The reaction network is presented in the following
scheme:
r = a3 + a1 exp( a2t),
[1]
where
a1 = ksr0/(ks + k s); a2 = ks + k s; a3 = k r0/(ks + k s).
s
[2]
Here, ks and k denote the deactivation and self-rege-
s
neration rate constants and r0 the rate in deactivation-free
conditions.
Equation [1] was used for fitting experimental data. The
sum of squares of the relative deviations served as an ob-
ject function. Results of the simulations are presented in

HYDROGENATION OF CROTONALDEHYDE ON Pt/ZnO CATALYST
173
The proposed reaction mechanism is
(and neglecting isomerization step 13), one arrives at the
following equations which relate the selectivities toward B
and C,
SB = d PB/d PA = LB MB PB/PA ,
SC = d PC/d PA = LC MC PB/PA ,
[3]
[4]
where
and
1
LC
=
,
[5]
[6]
1 + k10 (KAB + KABC)/k30 (KAC + KABC
)
LB = 1 LC,
k40
MC = LC
,
.
[7]
[8]
k0 (1 + KABC/KAC
)
)
3
k20
MB = LB
k0 (1 + KABC/KAB
1
Constants k10 are effective, as they contain hydrogen pres-
sure dependence. Constantsofequilibrium stepsare related
to adsorption/desorption coefficients
kAB
kAC
kAB + kAC
+ k
KAB
=
;
KAC
=
;
KABC
=
k
AB
.
k
k
AB
AC
AC
[9]
An analysis of Eqs. [3] and [4] reveals that the initial
selectivity (at low conversions) depends on the LC value.
The selectivity profile as a function of conversion depends
more on the MC value. For parallel-consecutive reactions
the lower the value of MC, the less pronounced is the C
selectivity dependence on the conversion.
It is quite clear from Eq. [5] that the LC value depends on
both kinetic (k1/k3) and adsorption factors (KAB + KABC)/
(KAC + KABC).
Here, A, B, C, and P denote crotonaldehyde, butyraldehyde,
crotyl alcohol (desired product), and butanol, respectively.
AB represents adsorption of crotonaldehyde via an olefinic
bond, AC via the carbonyl bond, and ABC via both bonds,
respectively. Steps 2, 4, etc. are not necessarily elementary
reactions; they can consist of several elementary reactions
with only one of them being rate-determining. In the above
mechanism, Z is a surface site. On the right-hand side of
the equations of steps, stoichiometric numbers (27, 28) for
the different routes (N(1), etc.) are given.
In addition to these reaction steps, there exist also deac-
tivation steps, however these are not depicted in this mech-
anism. Here, we can only speculate about the types of ad-
sorbed species which lead to side products. Hydrocarbons
and dienic polymers can arise from adsorbed crotonalde-
hyde, while chlorinated products can be formed from any
adsorbed species.
It follows from the literature and from our own exper-
imental data that the selectivity in crotyl alcohol strongly
depends on the type of support as well as on the reduction
temperature for the case of reducible supports. It is also
evident that in some cases an increase in selectivity is asso-
ciated with a decrease in activity (A catalyst, for example).
On one hand, however, as with the B catalyst, although
the selectivity is almost independent of the reduction tem-
perature, the overall activity in hydrogenation varied. On
the other hand, at exactly the same overall activity for the
A and B catalysts, the selectivities differed dramatically.
Although the kinetic factor does not represent the overall
catalytic activity but rather the ratio of hydrogenation con-
stants of olefinic and carbonyl bonds, it seems feasible to
speculate that the kinetic factor is less dependent than the
adsorption factor on the type of support.
Experimental data on selectivity show clearly that the
selectivity patterns for crotyl alcohol and butyraldehyde
follow similar trends; therefore, the isomerization reaction
between these molecules can be neglected in kinetic model-
Promotion Mechanism
Experimental data on crotonaldehyde hydrogenation
ing. Deriving kinetic equations from the mechanism above show a very strong influence of the catalyst support on the

174
CONSONNI ET AL.
reaction selectivity. Potential promotion effects in the hy-
The B catalyst was much better dispersed than the A
drogenation of , -unsaturated aldehydes were analyzed catalyst, certainly due to the formation of oxychlorinated
by Ponec (3). Promoters were defined as nonmetallic ad- species during the calcination step.
ditives which, being inactive by themselves, influence the
activity and the selectivity:
The noble metal in the catalyst formed an alloy with Zn.
A reduction temperature as high as 573 K was necessary
for the A catalyst to form an alloy. That was clearly shown
from XRD spectra and could also be deduced from XPS
analyses in which the narrowing of the Pt 4f lines occurred
parallel to the increase in the reduction temperature and
then to the alloy formation. In fact, the line width in the
XPS line is related to the electronic structure as it was al-
ready observed in PdPb (30) or PdSb (31) alloys. Since we
observed a narrow Pt(4f) line for the B catalyst after 473 K
reduction temperature, we can conclude that the alloy has
already been formed at this low reduction temperature.
On the A catalyst, after 473 K reduction, the catalytic
behavior of Pt metal particles dispersed on ZnO showed
(1) If the nonmetallic additive is reduced, it can form an
alloy with the main metal and then beneficial effects can be
due to ensemble size effects, ensemble composition effects,
and ligand or electronic influences.
(2) The ionic compounds could influence the surface
composition of the support.
(3) There could be changes in the morphology and the
size of the particles on a support.
(4) Promotion can also be of a chemical origin.
(5) Promoters can influence the accessibility of the spe-
cial sites.
According to Ponec (3), a promoter should activate the low selectivity in crotyl alcohol; however, when the reduc-
carbonyl group, but not bind the olefinic group. The most tion temperature is increased, one observes, parallel to the
likely explanation is that a positively charged cationic site formation of the PtZn alloy, a decrease in activity and an
is formed, leading to the formation of the following species: increase in crotyl alcohol selectivity. It is tempting to con+-
clude that the Pt sites, when alloying to Zn, form Pt –Zn
entities. These entities would not adsorb the crotonalde-
hyde molecule by binding to the olefinic bond but rather
by binding to the carbonyl bond. In fact, if the Pt atom is
==
electron-rich, it repels the electrons of the C C bond while
==
it is attracted by the electropositive carbon of the C
O
Recently, the influence of metal cations on the selective
hydrogenation of , -aldehydes (cinnamaldehyde and cro-
tonaldehyde) over polymer-stabilized platinum colloid was
studied. Yu et al. (29) added metal cations after the forma-
tion of a colloidal platinum catalyst, diminishing the pos-
sibility of influencing the size, shape, or structure of metal
colloidal particles. The addition of Fe3+ and Co2+ increased
both the activity and the selectivity in crotonaldehyde hy-
drogenation. A slight increase in selectivity was observed
when Zn2+ was added to the Pt–Fe3+ system; moreover, the
activity decrease and the selectivity increase could be cor-
related with the amount of added Zn2+. The mechanism of
the modification of metal cations to the Pt colloid (M = Fe,
Co, Ni, etc.) given by Yu et al. is presented below:
bond and the electronegative oxygen atom is bonded to the
electropositive Zn atom. Such a charge transfer could be
too small to be revealed by a Pt(4f) binding energy shift;
however, it has already been suggested by Li et al. (11) and
by Boccussi et al. (32).
Besides this alloying effect, it is clear that chlorine has a
promotor effect in this reaction since when both catalysts
contain quasi exclusively Pt–Zn particles, after 673 K re-
duction, the B catalyst gives a much higher crotyl alcohol
selectivity (80% ) than the A catalyst (50% ) with similar
activity. The difference in particle sizes (9 nm for B and
18 nm for A) can hardly be invoked to explain the differ-
ent catalytic behaviors. Nitta et al. (33) have also found a
promotor effect of chlorine in the unsaturated alcohol se-
lectivity in the hydrogenation of , -unsaturated aldehy-
des on cobalt catalysts. They concluded that the residual
chlorine in these catalysts affected both, the H2-reduction
step leading to a “favorable” crystallite size distribution
==
and the reaction stage depressing more strongly the C
C
==
bond than the C O bond hydrogenation reaction. Yu et al.
(34) supposed that the Cl anion acted only as spectator
ions in crotonaldehyde hydrogenation. The possible influ-
ence of Cl ions was not mentioned in the review of Ponec
Let us discuss our own experimental data, taking as a (3), probably due to the common belief that chloride acts
basis the considerations addressed above. merely as a poison. However, some examples exist show-
The Pt/ZnO catalysts prepared from Pt(NH3)4(NO3)2 ing the promotion effect of Cl in heterogeneous catalytic
(A) and H2PtCl6 (B) differentiate each other by the physi- reactions. For instance, the ethylene oxidation process was
cal characteristics and by the catalytic behavior as well.
strongly influenced by the chloride ions (35, 36); a decrease

HYDROGENATION OF CROTONALDEHYDE ON Pt/ZnO CATALYST
175
of oxygen adsorption on the silver catalysts was observed tion (80% crotyl alcohol selectivity in 5–20% conversion
with the increase of the chloride-covered fraction of the range) with a significant and stable activity along the time
surface. It was believed that Cl ions diminished the rate of stream after a short period of initial deactivation.
of undesired reaction, e.g., total oxidation, thus promoting
the main reaction—ethylene oxide formation. It was ob-
served that, at lower chloride concentrations, catalytic ac-
tivity even increased, while poisoning effects were observed
only at higher Cl concentrations. Electronic and geomet-
ric effects were thought to be responsible. Recently, it was
shown that modification of nickel catalysts by impregnation
with chlorine-containing compounds led not only to a de-
crease in the overall catalytic activity in thymol (3-methyl-
6-isopropylphenol) hydrogenation but also to changes in
selectivity. The latter was attributed to alteration of keto–
enol transformations of intermediate menthones and cor-
responding enols (37).
REFERENCES
1. Augustine, R. L., “Heterogeneous Catalysis in Organic Synthesis.”
Dekker, New York, 1995.
2. Touroude, R., J. Catal. 65, 110 (1980).
3. Ponec, V., Appl. Catal. A 149, 27 (1997).
4. Margitfalvi, J. L., Tompos, A., Kolosova, I., and Valyon, J., J. Catal.
174, 246 (1998).
5. Sen, B., and Vannice, M. A., J. Catal. 113, 52 (1988).
6. Makouangou, R. M., Murzin, D. Yu, Dauscher, A. E., and Touroude,
R. A., Ind. Eng. Chem. Res. 33, 1881 (1994).
7. Englisch, M., Jentys, A., and Lercher, J. A., J. Catal. 166, 25 (1997).
8. Claus, P., Schimpf, S., Scho¨del, R., Kraak, P., Mo¨rke, W., and Ho¨nicke,
D., Appl. Catal. A 165, 429 (1997).
9. Yoshitake, H., and Iwasawa, Y., J. Chem. Soc., Faraday Trans. 88(3),
503 (1992).
10. Sepulveda-Escribano, A., Coloma, F., and Rodriguez-Reinoso, F.,
J. Catal. 178, 649 (1998).
In light of our previous considerations, we can specu-
late that Cl remaining in the catalyst after reduction can
polarize even stronger Zn +, increasing therefore the in-
+
==
teractions between Zn and O
C, thus increasing the 11. Li, W., Chen, Y., Yu, C., Wang, X., Hong, Z., and Wei, Z., in “Proceed-
ings, 8th International Congress on Catalysis, Berlin, 1984,” Vol. 5,
equilibrium constant of crotonaldehyde adsorption via the
p. 205. Dechema, Frankfurt-am-Main, 1984.
12. Boccuzzi, F., Chiorino, A., Ghiotti, G., Pinna, F., Strukul, G., and
Tessari, R., J. Catal. 126, 381 (1990).
13. Sarkany, A., Zsoldos, Z., Furlong, B., Hightower, J. W., and Guczi, L.,
J. Catal. 141, 566 (1993).
14. Takezawa, N., and Iwasa, N., Catal. Today 36, 45 (1997).
15. Iwasa, N., Masuda, S., Ogawa, N., and Takesawa, N., Appl. Catal. A
125, 145 (1995).
16. Cubeiro, M. L., and Fierro, J. L. G., J. Catal. 179, 150 (1998).
17. Shirley, D. A., Phys. Rev. B 5, 4907 (1972).
==
C
O bond. In terms of Eq. [5] this means that there is an in-
crease of the KAC value, leading under certain conditions to
the overall decrease of the second term in the denominator.
Let us make simple calculations neglecting the diad-
sorbed mode of crotonaldehyde and assuming that the ki-
netic factor is constant, whatever the experiment presented
here (KABC = 0, (km0 /kn0 ) = (km0 /kn0 ) where and repre-
sents two sets of experiments). One arrives at
18. Scofield, J. H., J. Electron Spectrosc. Relat. Phenom. 8, 129 (1976).
19. Klug, H. P., and Alexander, L. E., “X-ray Diffraction Procedure for
Polycrystalline and Amorphous Materials,” 2nd ed. Wiley, New York,
1974.
20. Hecq, M., Hecq, A., Delrue, J. P., and Robert, T., J. Less Common
Meta. 64, P25 (1979).
21. Escard, J., Pontvianne, B., Chenebaux, M. T., and Cosyns, J., Bull. Soc.
Chim. Fr. 11–12, 2399 (1976).
22. Lieske, H., Lietz, G., Spindler, H., and Volter, J., J. Catal. 81, 8 (1983).
23. Le Normand, F., Borgna, A., Garetto, T. F., Apesteguia, R., and
Moraweck, B., J. Phys. Chem. 100, 9068 (1996).
(KAC/KAB
(KAC/KAB
)
(1 LC ) LC
(1 LC) LC
=
.
)
This means that taking the different selectivities 50 and
80% , observed respectively for A and B catalysts, reduced
at 673 K, the ratio of adsorption constants is four times
higher on B catalysts containing chlorine than on A cata-
lysts.
CONCLUSION
24. McCabe, R. W., Wong, C., and Woo, H. S., J. Catal. 114, 354 (1988).
25. Consonni, M., Touroude, R., and Murzin, D. Yu, Chem. Eng. Technol.
21, 7 (1998).
26. Simonik, J., and Beranek, L., Collect. Czech. Chem. Commun. 37, 353
(1972).
Catalyst A, 5% Pt/ZnO prepared from Pt(NH3)4(NO3)2
which does not contain Cl , displays more or less conven-
tional mechanisms, typical of platinum deposited on re-
ducible supports in which one can assume alloy formation
27. Horiuti, J., Res. Inst. Catal. Hokkaido Univ. 5, 1 (1957).
28. Temkin, M. I., Adv. Catal. 28, 173 (1979).
and/or migration of reducible oxide onto the metal particles 29. Yu, W., Liu, H., Liu, M., and Tao, Q., J. Mol. Catal. 138, 273 (1999).
30. Goetz, J., Volpe, M. A., Sica, A. M., Gigola, C. E., and Touroude, R.,
J. Catal. 167, 314 (1997).
31. Van Attekum, P. M. Th., and Trooster, J. M., J. Phys. 9, 2287 (1979).
like in Pt/TiO2 catalysts. For such catalysts, which display
modest crotyl alcohol selectivity in crotonaldehyde hydro-
genation, there is a correlation between reduction temper-
32. Boccuzzi, F., Chiorino, A., and Guglielminotti, E., Surf. Sci. 368, 264
==
ature, decrease in activity, and increase in C O bond hy-
(1996).
drogenation selectivity. On the contrary, the B catalyst, 5%
Pt/ZnO prepared from H2PtCl6, contains Cl ions at any
reduction temperatures, from 473 to 673 K. It forms a Pt–
Zn alloy, even at low reduction temperature (473 K), and
then shows one of the best selectivities to crotyl alcohol
achieved so far in gas-phase crotonaldehyde hydrogena-
33. Nitta, Y., Hiramatsu, Y., and Imanaka, T., J. Catal. 126, 235 (1990).
34. Yu, W., Liu, H., and Tao, Q., Chem. Commun. 1773 (1996).
35. Kul’kova, N. V., Kamenski, D., and Bonchev, D., React. Kinet. Catal.
Lett. 7, 335 (1977).
36. Campbell, C. T., and Koel, B. E., J. Catal. 92, 272 (1985).
37. Allakhverdiev, A. I., Kul’kova, N. V., and Murzin, D. Yu., Catal. Lett.
29, 57 (1994).