Crotonaldehyde hydrogenation by gold supported on TiO2: Structure sensitivity and mechanism

Article abstract of DOI:10.1016/j.jcat.2004.01.033

The catalytic properties of two series of Au/TiO₂ catalysts prepared by deposition-precipitation with NaOH (DP NaOH) (～3 wt%) and by deposition-precipitation with urea (DP Urea) (～8 wt%) were evaluated for the reaction of crotonaldehyde hydrogenation at atmospheric pressure. There is no difference in activity (molg_Au^-1s^-1) and selectivity between the DP Urea and DP NaOH samples for a given activation treatment. This is due to the fact that the DP Urea and DP NaOH samples exhibit a similar gold particle size distribution, although the gold loading in DP Urea catalysts is much higher than in DP NaOH. The DP Urea samples were reduced under H₂ at different temperatures (120-500°C) or treated in air at 300°C with various flow rates, to vary the average particle size within a large range, 1.7 to 8.7 nm. The selectivity to crotyl alcohol (selective hydrogenation of the carbonyl bond), in the 5-50% conversion range, is high, 60-70 %, and is independent of the reduction temperature, and almost constant as a function of the particle size. In contrast, the TOF depends on the gold particle size, drastically increasing when the gold particle size is ～2 nm. These characteristic features of Au/TiO₂ catalysts in this reaction are compared with those of Pt/TiO₂. The possible adsorption modes of crotonaldehyde are discussed. Hydrogen dissociation is proposed to be the rate-determining step, and to take place on the low-coordinated atoms of the gold particles.

Full text of DOI:10.1016/j.jcat.2004.01.033

Journal of Catalysis 223 (2004) 328–339
www.elsevier.com/locate/jcat
Crotonaldehyde hydrogenation by gold supported on TiO2:
structure sensitivity and mechanism
a
a
b
c,∗
Rodolfo Zanella, Catherine Louis, Suzanne Giorgio, and Raymonde Touroude
a
Laboratoire de Réactivité de Surface, UMR 7609 CNRS, Université Pierre et Marie Curie, 4 place Jussieu, 75252 Paris cedex 05, France
b
1
CRMC2 CNRS, Campus de Luminy, case 913, 13288 Marseille cedex, France
c
Laboratoire des Matériaux, Surfaces et Procédés pour la Catalyse, UMR 7515 CNRS, ECPM/ULP, 25 rue Becquerel, 67087 Strasbourg cedex 2, France
Received 8 September 2003; revised 12 January 2004; accepted 26 January 2004
Graham Hutchings
Abstract
The catalytic properties of two series of Au/TiO catalysts prepared by deposition-precipitation with NaOH (DP NaOH) (∼ 3 wt%) and
2
by deposition-precipitation with urea (DP Urea) (∼ 8 wt%) were evaluated for the reaction of crotonaldehyde hydrogenation at atmospheric
−
1
−1
s ) and selectivity between the DP Urea and DP NaOH samples for a given activation
Au
pressure. There is no difference in activity (mol g
treatment. This is due to the fact that the DP Urea and DP NaOH samples exhibit a similar gold particle size distribution, although the gold
loading in DP Urea catalysts is much higher than in DP NaOH. The DP Urea samples were reduced under H at different temperatures (120–
2
◦ ◦
00 C) or treated in air at 300 C with various flow rates, to vary the average particle size within a large range, 1.7 to 8.7 nm. The selectivity
5
to crotyl alcohol (selective hydrogenation of the carbonyl bond), in the 5–50% conversion range, is high, 60–70 %, and is independent of the
reduction temperature, and almost constant as a function of the particle size. In contrast, the TOF depends on the gold particle size, drastically
increasing when the gold particle size is ∼ 2 nm. These characteristic features of Au/TiO catalysts in this reaction are compared with those
2
of Pt/TiO . The possible adsorption modes of crotonaldehyde are discussed. Hydrogen dissociation is proposed to be the rate-determining
2
step, and to take place on the low-coordinated atoms of the gold particles.

2004 Elsevier Inc. All rights reserved.
Keywords: Gold; Catalysts; Metal nanoparticles; Titania; Catalyst preparation; Deposition-precipitation; Hydrogenation of crotonaldehyde; Transmission
electron microscopy; HRTEM
1
. Introduction
properties of gold, probably because of the very low activ-
ity of gold supported on oxides when the metal particles are
large. The position of gold in catalysis drastically changed at
the end of the 1980s, when Haruta and co-workers were able
to prepare small gold particles supported on oxides and dis-
covered that these catalysts were active in CO oxidation at
low temperatures [8,9]. Gold has been more recently tested
again in hydrogenation reactions, and as reported by Bond
and Thompson [10], it also seems a very promising catalyst
for this type of reaction.
Ni-, Pd-, or Pt-based catalysts are very active catalysts in
hydrogenation reactions because of the high ability of these
metals to dissociate hydrogen and because the hydrogena-
tion reaction rate (r) is positively dependent on the hydrogen
α
H
coverage (r = kΘ with α > 0). In comparison to these met-
als, gold has been classified as a poor catalyst to dissociate
H2 molecules.
Early studies showed that supported gold can catalyze
various hydrogenation reactions, such as those of linear
alkenes [1–3], alkynes [3,4], cyclohexene [5,6], and ace-
tone [7]. Then, little attention was paid to the hydrogenation
Indeed, supported gold catalysts were active and selec-
tive in partial and selective hydrogenation reactions. Jia et
al. [11] showed that Au/Al2O3 catalyzed the hydrogenation
of acetylene to ethylene with 100% selectivity in the tem-
◦
perature range 40–250 C and with a maximum activity for
*
Corresponding author.
an average gold particle size of 3.0 nm. Okumura et al. [12]
investigated the reaction of hydrogenation of 1,3-butadiene
E-mail address: touroude@chimie.u-strasbg.fr (R. Touroude).
Associated with the Universities of Aix-Marseille II and III.
1
0021-9517/$ – see front matter  2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2004.01.033

R. Zanella et al. / Journal of Catalysis 223 (2004) 328–339
329
over Au/Al2O3, Au/SiO2, and Au/TiO2. All the Au catalysts
were 100% selective to butenes with 65–75% of selectivity
to 1-butene. They found that the reaction was almost insen-
sitive to the size of the Au particles (2.5 and 5 nm) and to the
nature of the oxide support.
One of the most studied reactions of hydrogenation with
gold catalysts is the selective hydrogenation of α, β-unsatu-
rated aldehydes [12–17]. Indeed, the selectivity in unsatu-
rated alcohols, the desired products, is difficult to achieve
quantum size effects, i.e., to alteration of the metallic prop-
erties of the smallest gold particles. In a more recent study,
Claus and colleagues [17] prepared three Au/ZrO2 catalysts
by the same method of deposition-precipitation, using vari-
ous pH’s, to obtain catalysts with various gold particle sizes
(4.0, 6.9, and 7.7 nm). They found that the TOF decreased
by nearly one order of magnitude, and the selectivity to allyl
alcohol increased (from 15 to 35%) when the particle sizes
increased. On the other hand, they observed by HRTEM,
multiply twinned particles whose number increased with in-
(
i) because thermodynamics favors the hydrogenation of the
−1
◦
C=C over the C=O bond by about 35 kJ mol and (ii) be-
cause for kinetic reasons, the reactivity of the C=C bond
is higher than that of the C=O bond. Hence, on conven-
tional supported hydrogenation catalysts (Pt, Ru, Ni, etc.),
the main products are the saturated aldehydes. On supported
gold catalysts, interesting selectivities to unsaturated alcohol
were found, even if the results obtained from one research
group to another are not always fully consistent. The selec-
tive hydrogenation of crotonaldehyde (but-2-enal) was first
studied by Bailie and Hutchings [14] at atmospheric pres-
creasing reduction temperature (300 to 500 C) or reduction
time (3 to 18 h). They pointed out the fact that when the
number of multiply twinned particles increased, both the se-
lectivity and the TOF decreased. They attributed the higher
activity of Au/TiO2 compared to that of Au/ZrO2 (TOF× 2)
to the fact that the gold particles were round shaped on TiO2
and facetted on ZrO2 (average particle size of ∼ 5 nm on
both supports). Round-shaped gold particles on TiO2 would
contain a higher proportion of low-coordinated surface sites,
which would explain the higher TOF.
◦
sure and 250 C, with Au/ZnO and Au/ZrO2 prepared by
As shown above, the influence of the gold particle size
and of the nature of the support on the activity and selectivity
for the hydrogenation of α, β-unsaturated aldehydes is not
clear, yet. To attempt to clarify the influence of the first pa-
rameter, we decided to investigate the dependence of the cat-
alytic properties with gold particle size in Au/TiO2 catalysts
prepared by a single method, the deposition-precipitation
with urea [18]. The main advantage of this method over the
method of deposition-precipitationwith NaOH developed by
Haruta and colleagues [19,20] is that all the gold of the so-
lution is deposited on TiO2, so higher gold loading can be
reached with the same gold particle size as with Haruta’s
method, and the gold loading of the catalysts can be easily
controlled. The results obtained are therefore compared with
those obtained on Au/TiO2 catalysts prepared by deposition-
precipitation with NaOH. The reaction studied is the hy-
drogenation of crotonaldehyde (CH3–CH=CH–CH=O) at
atmospheric pressure, under the same experimental condi-
tions as those used in former studies for Pt/TiO2 catalysts
[21,22]. Hence, comparisons of the catalytic behavior be-
tween platinum and gold catalysts could be made, especially
the influence of the particle size and that of the reduction
temperature, since these two parameters are the key parame-
ters which govern the activities and selectivities of Pt/TiO2
catalysts.
coprecipitation, and with Au/SiO2 prepared by impregna-
tion. They found that Au/ZnO was more selective for C=O
bond hydrogenation and that the addition of thiophene acted
as a promoter for selective hydrogenation. Haruta and col-
leagues [12] studied the same reaction but at high pressure
◦
(
1.0 MPa) and 227 C, with Au/Al2O3 and Au/TiO2 cata-
lysts prepared by deposition-precipitation with NaOH and
gold particles sizes of 4.6 and 3.4 nm, respectively, and
Au/SiO2 (5.3 nm) prepared by grafting of gold acetyl ace-
tonate. They also found that the selectivity depends on the
nature of the support. The highest selectivity to crotyl alco-
hol was obtained over Au/TiO2 (≈ 25% instead of < 10%
for the other samples). Bailie et al. [15] studied the influ-
ence of the particle size on the selectivity in the Au/ZnO
system. Gold particle sizes were increased by increasing
the gold loading in the samples (2 wt% (2–4 nm), 5 wt%
(
4–5 nm), and 10 wt% (bimodal distribution: 1–3 and 6–
1
5
3
0 nm)), or by increasing the reduction temperature of the
◦
◦
wt% Au/ZnO catalyst (250 C (4–6 nm), 300 C (3–8 nm),
◦
◦
50 C (10–15 nm), and 400 C (up to 20 nm)). They re-
ported that the selectivity to crotyl alcohol increased when
the gold loading or the reduction temperature increased, and
they proposed that the active sites for selective hydrogena-
tion of crotonaldehydeto crotyl alcohol were associated with
the presence of large Au particles. Claus et al. [16] also stud-
ied the effect of the particle size in a series of three Au/TiO2
catalysts (1.1, 2.0, and 5.3 nm) prepared by three different
preparation methods in selective hydrogenation of acrolein
2. Experimental
◦
under high pressure (2 MPa) and at 240 C. They observed
2.1. Catalyst preparation
a drastic decrease in the TOF by two to three orders of mag-
nitude and a decrease of selectivity when the size of gold
particles decreased from 2.0 to 1.1 nm, and constant val-
ues of TOF and selectivity for average particle sizes of 2.0
and 5.3 nm. They proposed that the structure sensitivity of
the reaction in the case of small gold particles was due to
2
−1
,
Titania Degussa P25 was used as support (45 m g
nonporous, 70% anatase and 30% rutile, purity > 99.5%)
and solid HAuCl4 · 3H2O (Acros) as gold precursor. Before
◦
preparation, TiO2 was previously dried in air at 100 C for
at least 24 h.

330
R. Zanella et al. / Journal of Catalysis 223 (2004) 328–339
ꢀ
ꢀ
3
i
2
i
Table 1
ds = nid / nid where ni is the number of particles
i
Au/TiO catalysts used in the present study
2
of diameter d .
Sample
Conditions of preparation
Au loading (wt%)
XANES (X-ray absorption near-edge structure) measure-
◦
ments of the Au/TiO samples were performed at the Au
DP Urea 1 h
DP Urea 2 h
DP Urea 16 h
DP Urea 16 h/90
DP NaOH 2 h
DP urea for 1 h at 80 C
6.9
7.5
7.5
7.3
2.8
2
◦
DP urea for 2 h at 80 C
LIII edge at the XAS 13 beam line of the DCI storage
◦
DP urea for 16 h at 80 C
DP urea for 16 h at 90 C
DP NaOH for 2 h at 80 C
ring (operating with positrons at 1.85 GeV and a mean
ring current of 300 mA) of the LURE synchrotron radi-
ation facility (Orsay, France). The samples were diluted
with cellulose and pressed into a pellet. The XANES spec-
tra were recorded at room temperature, in transmission
mode using two argon-filled ionization chambers and a
channel-cut Si(111) monochromator. For each sample, the
XANES spectra were scanned twice with 0.5 eV steps from
◦
◦
2
.1.1. Preparation by deposition-precipitation with NaOH
(
DP NaOH samples)
One hundred milliliters of an aqueous solution of HAuCl4
4.2 × 10−3 M) was heated to 80 ◦C. The pH was adjusted
(
1
1,870 to 12,020 eV. The energy was calibrated using a Au
to 7 by dropwise addition of NaOH 1 M, 1 g of TiO2 was
dispersed in the solution, and then the pH was readjusted
metal foil. The XANES spectra of references, gold foil, and
Au(OH)3 were also recorded. After background correction,
the XANES spectra were normalized at the end of the first
EXAFS oscillation.
◦
to 7 with NaOH. The suspension thermostated at 80 C was
vigorously stirred for 2 h.
Activation treatments and catalytic reactions were car-
ried out in a glass flow reactor, operating at atmospheric
pressure as described elsewhere [24]. Before catalytic ex-
periments, each catalyst (10–50 mg) was reduced at vari-
2
.1.2. Preparation by deposition-precipitation with urea
(
DP Urea samples)
One gram of TiO2 was added to 100 ml of an aqueous so-
−3
lution containing HAuCl4 (4.2×10 M) and urea (0.42 M).
◦
ous temperatures, between 120 and 500 C, either under H2
The initial pH was ∼ 2. The suspension thermostated at
−1
◦
−1
flow (50 ml min for 10 min at RT, heating (5 C min ),
◦
8
0 C was vigorously stirred for 1, 2, or 16 h. Urea decom-
1
h at the chosen temperature, and cooling down to the re-
position leads to a gradual rise in pH.
◦
action temperature, 120 C) or under air flow (50, 35, or
2
3
These sequences of preparation in solution were per-
formed in the dark, in blackened flasks, since light is known
to decompose the gold precursors and to reduce them. The
solids gathered after centrifugation were washed with wa-
ter, centrifuged four times, and then dried under vacuum at
1
ported in [18]. The samples were stored away from light
under vacuum in a desiccator or under Ar for transportation
to Strasbourg where the catalytic reaction was performed.
According to Zanella et al. [23] gold is still mainly in oxida-
tion state III after drying.
−1
◦
−1
ml min , 10 min at RT, heating (2 C min ), 4 h at
◦
00 C, and cooling down to RT). In the latter, catalysts were
afterward purged with Ar at RT and then submitted to a treat-
◦
ment at 120 C under H2, as described above.
Crotonaldehyde supplied by FLUKA puriss and stored in
argon was used as received. Experiments were performed
with a crotonaldehyde/H2 pressure ratio fixed to 8/752
◦
00 C for 2 h in the dark, according to the procedure re-
(
Torr).
The reaction products were analyzed on line by gas-liquid
FID chromatography (GLC, with a 30-m-long and 0.5461-
mm-diameter DB-Wax column (J&W Scientific) heated at
Table 1 summarizes data for the various Au/TiO2 cata-
lysts used in this study.
◦
8
5 C). The relative sensitivity factors using the FID detector
are taken from the Dietz Tables [25] as 1 for crotonaldehyde,
crotyl alcohol, butanal, and butanol and 1.4 for hydrocar-
bons.
2
.2. Techniques
The reaction activities (µmols 1
−
g
−1
) were calculated ac-
Au
Chemical analysis was performed by inductively cou-
cording to the following formula: A = αF/ω, where α is the
pled plasma atomic emission spectroscopy at the CNRS
Center of Chemical Analysis (Vernaison, France). The Au
weight loading of the samples was expressed in grams of
crotonaldehyde conversion, F, the crotonaldehyde flow rate
−1
in mol s , and ω, the weight of gold in grams. The turnover
frequency (TOF), number of molecules of crotonaldehyde
converted per surface atom of gold particles and per second,
was calculated with the assumption that gold particles were
cuboctahedral with a hexagonal face in contact with the tita-
nia surface. The selectivities to the different products, crotyl
alcohol, butanal, butanol, and hydrocarbons were expressed
as the molar ratio of the selected product to the all products
formed. The crotonaldehyde, used without further purifica-
tion, contained 0.1–0.3% butanal; therefore, the selectivity
values were always compared at conversions higher than 4%
to minimize the errors.
◦
Au per gram of sample calcined at 1000 C: wt% Au =
[
mAu/(mAu + mTiO )] × 100.
2
The average size of gold particles in the various sam-
ples, analyzed after the catalytic reaction, was determined
by transmission electron microscopy (TEM) with a Jeol
2
000FX electron microscope. The histograms of particle
sizes were established from the measurement of 300 to
000 particles. The size limit for the detection of gold
1
particles on TiO2 was about 1 nm. The average particle
diameter ds was calculated using the following formula:

R. Zanella et al. / Journal of Catalysis 223 (2004) 328–339
331
No activity was observed on the support alone tested un-
der the same experimental conditions as Au/TiO2 catalysts,
◦
more precisely in the 80–160 C temperature range. The
comparison of the catalytic behavior of all the Au/TiO2 cat-
◦
alysts was made for a reaction temperature of 120 C. The
weight of catalyst and the flow rate of the products were
consequently adjusted to convert 5–10% products in order
to compare the activities and selectivities in a similar range
of conversions.
3
. Results
3
.1. General features on activity and product distributions
All the catalysts (DP Urea and DP NaOH) show a de-
Fig. 1. Hydrogenation of crotonaldehyde (PH = 752 Torr, P
2
crotonaldehyde
◦
=
8 Torr, T
= 120 C). Activity versus time on stream of DP
reaction
activation period during the first 30–40 min of time on
stream as reported in Fig. 1 for DP Urea 16 h reduced under
H2 at different temperatures. The product distributions are
also the same for all the catalysts, with crotyl alcohol and
butanal as main products, and butanol and butenes as mi-
nor products. They remain quasi constant with the time on
stream, even during the deactivation period (Fig. 2). From
a kinetic point of view, the constant distribution of reaction
products indicates that the deactivation process is indepen-
dent of these products. Therefore, the activity evolutions
could be fitted by curves following the exponential equation,
At = a3 + a1 ∗ exp(−a2 ∗ t), where ai are only related to the
deactivation constants. This was detailed elsewhere for a Pt
catalyst, which exhibited the same features [26]. Examples
of curve fittings are shown on Fig. 1. The activities reported
in Tables 2 and 3, and further discussed, correspond to the
activity in the steady state, calculated from the equation with
t = ∞ (Activity = A∞ = a3).
Urea 16 h reduced under H at different temperatures (experimental
2
points + curve fittings (continuous lines)).
Fig. 2. Hydrogenation of crotonaldehyde (PH = 752 Torr, P
2
crotonaldehyde
◦
=
8 Torr, T
= 120 C). Selectivities versus time on stream of DP
reaction
◦
◦
Urea 16 h reduced under H at 120 C (black symbols); at 300 C (gray
symbols); at 500 C (white symbols).
2
◦
3
.2. Catalyst regeneration and reproducibility
3
.3. Product distributions and activation energy
We checked that the deactivation was reversible with cat-
◦
alyst DP Urea 16 h reduced at 300 C under H2. After about
The product distributions obtained with the Au/TiO2 cat-
4
5 min of reaction, i.e., deactivation, the catalyst treated
alysts are remarkable since they consist of 60–70% of crotyl
◦
again under H2 for 1 h at 300 C exhibits exactly the same
behavior on stream as the fresh catalyst (Fig. 3a). This in-
dicates that deactivation is due to adsorbed species and not
to metal sintering during the reaction. It can be deduced that
the gold particles that were measured after reaction are of
the same size as just after thermal pretreatment.
We also checked the reproducibility of the experiments
of reactivity. The curves of activity versus reaction time of
two DP Urea 2 h samples arising from the same batch and
alcohol, whatever the studied Au/TiO catalyst and the tem-
2
perature of pretreatment (Fig. 2). A selectivity-conversion
diagram was obtained for catalyst DP Urea 16 h in gradu-
ally decreasing flow rates during the time on stream in order
to increase the conversion (Fig. 4). This diagram shows that
the selectivity to butanal remains unchanged up to 50% con-
version while those in butenes and butanol slightly increase
at the expense of that in crotyl alcohol.
To check if butanal could be formed by isomerization of
crotyl alcohol as observed on Pt-based catalysts [27], the hy-
drogenation of crotyl alcohol was investigated for the DP
Urea 16 h catalyst. It was observed that crotyl alcohol re-
acted at a similar rate as crotonaldehyde, to give rise to
90% of butenes, 3.9% of butanal, and 6.1% of butanol. The
◦
reduced at 300 C under H2 are superimposable (Fig. 3b).
Noteworthy is the fact that the conditions of sample stor-
age described under Experimental are the determining ones
to obtain reproducible experiments, because of the trend of
gold particles toward sintering under ambient air.

332
R. Zanella et al. / Journal of Catalysis 223 (2004) 328–339
Table 2
Gold particle size and catalytic activity in hydrogenation of crotonaldehyde of DP Urea16 h and DP NaOH 2 h as a function of the reduction temperature
under H
2
Catalysts
Reduction
Average
Standard
Activity
TOF
S
crotyl alcohol
(%)
◦
−6
−1 −1
−3 −1
temperature ( C)
particle size ds (nm)
deviation on ds (nm)
(10 mol g
s
)
(10
s
)
Au
DP Urea 16 h
120
1.7
1.7
1.9
2.1
2.6
0.40
0.49
0.46
0.59
0.58
19.9
20.3
22.4
16.7
8.3
6.4
6.5
7.8
6.3
3.7
65
66
67
68
64
200
300
400
500
DP NaOH 2 h
120
1.3
1.7
1.8
2.1
2.5
0.31
0.40
0.41
0.54
0.60
22.8
19.8
20.7
13.2
6.3
6.0
6.3
6.9
5.0
2.7
66
66
69
70
62
200
300
400
500
Table 3
◦
Gold particle size and catalytic activity in hydrogenation of crotonaldehyde as a function of the DP Urea samples, pretreated at 300 C in H or air
2
Catalysts
Pretreatment
Average
Standard
Activity
TOF
S
crotyl alcohol
(%)
−
6
−1 −1
−3 −1
particle size ds (nm)
deviation on ds (nm)
(10 mol g
s
)
(10
s
)
Au
−1
DP Urea 1 h
DP Urea 2 h
DP Urea 16 h
DP Urea 16 h/90
H (50 ml min
2
)
2.4 and 6.7
5.1
12.1
22.4
18.8
59
65
67
68
2.1
1.9
1.6
0.50
0.46
0.37
4.6
7.8
5.6
−
1
DP Urea 2 h
Air (50 ml min
Air (12 ml min
)
)
2.8
3.5
4.0
8.7
0.58
0.74
0.86
2.76
5.2
2.9
2.3
0.7
2.5
1.7
1.5
0.9
60
60
53
−1
−
1
Air (5 ml min
Air (50 ml min
)
1
−
a
DP Urea 1 h
a
)
nd
The value was not reliable due to the too low conversion (∼ 1%) (see experimental section).
Fig. 4. Hydrogenation of crotonaldehyde (PH = 752 Torr, P
2
crotonaldehyde
◦
=
8 Torr, T
= 120 C). Selectivities versus conversion on DP Urea
reaction
◦
1
6 h/90 reduced under H at 120 C.
2
◦
Fig. 3. Experiments of (a) regeneration of DP Urea 16 h reduced at 300 C
under H and of (b) reproducibility of the catalytic reaction with DP Urea
2
◦
butenes. This could explain the increased of selectivity to
butenes at the expense of that in crotyl alcohol when the con-
version of crotonaldehyde increased. This also indicates that
the butanal observed in the crotonaldehyde hydrogenation
products is directly formed from the reactant since the iso-
merization of crotyl alcohol into butanal is a minor reaction.
Scheme 1 can be proposed.
2
h reduced at 300 C under H ; white symbols for the first experiment and
2
black symbols for the second one.
butenes formed from crotyl alcohol consisted of 73% of but-
1
-ene, 7% of trans-but-2-ene, and 20% of cis-but-2-ene. This
butene distribution is similar to that obtained from butadiene
hydrogenation on the same catalyst (unpublished results).
These results suggest that crotyl alcohol transforms by de-
hydration into butadiene, which further hydrogenates into
The activation energy of the crotonaldehyde hydrogena-
tion was deduced from the experiments performed in the
◦
−1
80–160 C temperature range. A value of 36.3 kJ mol was

R. Zanella et al. / Journal of Catalysis 223 (2004) 328–339
333
(A)
(B)
◦
◦
Fig. 5. TEM images and size histograms of gold particles of DP Urea 16 h reduced under H at 120 C (A) and at 500 C (B).
2
3
.4. Influence of the reduction temperature under H2
The influence of the reduction temperature under H2 (120
◦
to 500 C) on the catalytic properties was studied with cat-
alysts DP Urea 16 h and DP NaOH 2 h. Both Au/TiO2
catalysts show the same evolution of the catalytic proper-
ties versus the reduction temperature with roughly the same
activities and TOFs (Table 2). They exhibit almost the same
activities and TOF’s whether they have been reduced at 120,
Scheme 1. Reaction pathway of the hydrogenation of crotonaldehyde with
Au/TiO catalysts; plain and dotted arrows represent the primary and sec-
2
ondary reactions, respectively, and the lengths of the arrows are roughly
proportional to the reaction rates.
◦
−
1
200, and 300 C, and lower activity and TOF after reduction
found, which is slightly higher than the 21–34 kJ mol re-
◦
−1
at 400 and 500 C.
ported by Vannice and Sen [28] or 28 kJ mol reported by
Raab and Lercher [29], obtained on Pt/TiO2 catalysts for the
same reaction.
It is important to note that the average sizes of gold parti-
cles in both catalysts are very close to each other. In addition,
the average sizes of both types of catalysts are almost the
On Au/TiO2 catalysts, the product distributions were also
independent of the reaction temperature in the studied tem-
◦
same for reduction temperatures between 120 and 400 C
◦
◦
◦
perature range (80–160 C). However, at 160 C, the consec-
utive reaction of crotyl alcohol became detectable, since the
butene selectivity slightly increased from 2 to 6 % at the ex-
pense of the crotyl alcohol selectivity.
(∼ 1.8 nm), and slightly larger after reduction at 500 C
(∼ 2.5 nm). Figs. 5A and 5B show the TEM images and the
size histograms of gold particles of DP Urea 16 h reduced
◦
under H2 at 120 and 500 C, respectively.

334
R. Zanella et al. / Journal of Catalysis 223 (2004) 328–339
We checked by XANES that gold is already fully reduced
◦
at a pretreatment temperature as low as 120 C for both types
of catalysts. Indeed, the XANES spectra of DP Urea 16 h
◦
and DP NaOH 2 h reduced under H2 at 120 or 200 C are
0
the same, and they are similar to that of the Au reference of
gold foil (Fig. 6).
3
.5. Influence of the gold particle size
The influence of the gold particle size on the catalytic
properties was systematically investigated with a series
of DP Urea catalysts prepared during various deposition-
precipitation times (1, 2, and 16 h) and temperatures (80 or
◦ ◦
Fig. 6. XANES spectra of DP Urea 16 h and DP NaOH 2 h reduced under
9
0 C) and pretreated at 300 C. As shown in Table 3, for
◦
H at 120 and 200 C and XANES spectra of references, Au(OH) , and Au
2
3
the series of catalysts reduced under H2, the average particle
size decreases when the deposition-precipitation time or the
preparation temperature increases, in the same way as when
the samples were treated under air [18]. DP Urea 1 h, which
contains the largest gold particles, shows the lowest activity
in the hydrogenation of crotonaldehyde. The bimodal size
distribution did not make possible the calculation of the av-
metal foil.
erage particle size, and therefore that of the TOF. However,
DP Urea 1 h gives a monomodal size distribution (8.7 nm)
when it is pretreated with air at 300 C. Intermediate particle
◦
sizes (2 to 4 nm) were obtained with DP Urea 2 h pretreated
(A)
(B)
◦
Fig. 7. TEM image and size histogram of gold particles in DP Urea 2 h pretreated at 300 C under H (A) and under air (B).
2

R. Zanella et al. / Journal of Catalysis 223 (2004) 328–339
335
Fig. 8. Activity (A) and TOF (B) versus the average particle size of DP Urea 16 h pretreated under H ("), DP Urea 2 h pretreated under H (2), DP Urea 2 h
2
2
and DP Urea 1 h pretreated under air (F), and DP NaOH 2 h pretreated under H (1).
2
◦
the same initial preparation after pretreatment at 500 C un-
with air at various flow rates (Table 3). Figs. 7A and 7B show
the TEM images and the size histograms of gold particles of
der H . The particle sizes are now perfectly facetted with
2
a DP Urea 2 h sample pretreated under H2 and under air at
facets corresponding to the (111) and (100) planes. They are
also stable under electron beam irradiation. The shape is the
same on all the observed particles, and the interface is flat.
◦
3
00 C, respectively. It shows that the H2 pretreatment leads
to smaller gold particles than the air pretreatment. Table 3
shows that the catalytic activity increases when the particle
size decreases, which is consistent with the increasing metal
surface area.
On Fig. 9B, the gold particle is in epitaxy on the TiO sub-
2
strate. The (111) plane of gold is parallel to the (110) plane
of the rutile phase of the support.
Figs. 8A and 8B report the activity and TOF, respectively,
versus the average particle size of all the DP Urea and DP
NaOH catalysts described above (from Tables 2 and 3). This
figure clearly shows that both activities and TOFs decrease
when the average gold particle size increases, and that both
of them are much higher when the Au/TiO2 catalysts contain
gold particles with average sizes of ∼ 2 nm. It also indi-
cates that there is no effect of the preparation method, or
of the conditions of reduction on the catalytic properties, in
agreement with a previous study on the reaction of CO oxi-
dation [23].
4
. Discussion
The main features observed during this study on croton-
aldehyde hydrogenation over Au/TiO2 catalysts are the fol-
lowing.
(
i) Regarding the reaction.
–
–
The Au/TiO2 catalysts are highly selective to crotyl
alcohol, which arises from the selective hydrogena-
tion of the C=O bond: 60–70% for particle sizes
lower than 4 nm and a slightly lower selectivity when
the particle size increases (Tables 2 and 3).
The selectivities do not depend on the conversion, at
least up to 50% (Fig. 4), or on the time on stream, i.e.,
on the state of deactivation of the catalysts (Fig. 2), or
on the reduction temperature (Tables 2).
Concerning the product distributions, we can note a slight
decrease in the selectivity to crotyl alcohol from 60–70 to
5
0–60% when the average particle size increases (Tables 2
and 3).
3
.6. HRTEM observations
The morphology, structure, and orientation of the gold
particles on the TiO2 support after different reduction tem-
peratures have been determined by HRTEM. Fig. 9A shows
a 2-nm gold particle from a DP Urea sample pretreated
– Both the activity and the TOF increase when the gold
particle size decreases (Fig. 8).
(ii) Regarding the catalysts.
◦
at 200 C under H2. The interface is not completely flat.
– There is no influence of the preparation method,
deposition-precipitation with NaOH or urea, on the
catalytic properties since similar trends are obtained
in deactivation, and the DP Urea and DP NaOH cat-
alysts show the same TOF, activity (expressed per
Among all the observed particles, no preferential orienta-
tion could be observed. The particles are single crystals and
present some tiny facets and round parts. They are stable un-
der the electron beam. Fig. 9B corresponds to particles from

336
R. Zanella et al. / Journal of Catalysis 223 (2004) 328–339
◦
temperature than bulk gold, ∼ 330 C for particles of
◦
∼
2 nm instead of 1060 C for bulk [30], and since
they are very close to each other, especially in the
DP Urea samples because of the high Au loading
(
average distance between the particles of ∼ 5 nm;
Figs. 5 and 7). It may be noted that in the case of
Au/ZnO prepared by coprecipitation, Hutchings and
colleagues [15] reported evidence for particle sinter-
ing, from 4 to 20 nm, when the reduction temperature
◦
increased from 250 to 400 C.
–
The size of the gold particles is not affected by the
reaction, and catalyst deactivation is not due to metal
sintering. This has also been reported by Hutchings
and colleagues [15] in the case of Au/ZnO for the
same reaction.
In order to interpret the catalytic results, it is desirable
to compare these results with the catalytic behavior of the
Pt/TiO2 catalysts in the selective hydrogenation of croton-
aldehyde.
4
.1. Comparison with Pt/TiO2 in crotonaldehyde
hydrogenation
Compared to Pt/TiO2, which was studied under the same
reaction conditions (R. Touroude, unpublished results and
Ref. [21]), it must be pointed out that the Au/TiO2 catalysts
are much less active, by about two orders of magnitude. In
contrast, the selectivities to unsaturated alcohol are higher
with gold catalysts, and the evolution with the reduction tem-
perature and the particle size is completely different. Indeed,
Vannice and Sen [28], and then other groups [21,29,31],
showed that both the selectivity and TOF increased when
◦
the reduction temperature of Pt/TiO2 increased up to 500 C,
which is not the case with Au/TiO2. However, on Pt/TiO2,
the selectivity to crotyl alcohol did not exceed 40–50% af-
ter high-temperature reduction. The interpretation, generally
admitted, proposed first by Vannice and Sen [28] is that the
carbonyl bond of the unsaturated aldehydes is activated on
Fig. 9. HRTEM image of a gold particle seen in profile view of DP Urea
◦
◦
after calcination under H at (A) 200 C, (B) 500 C (the view axis, per-
2
pendicular to the image, is in the [110] direction).
0
the sites located at the Pt –TiOx interface, involving oxy-
gram of Au), and product selectivities. This was also
observed in a previous study on the reaction of CO
oxidation [23]. Noteworthy is that if the activities are
expressed per gram of catalyst, the DP Urea samples
are more active since they contain a higher Au loading
gen vacancies. These sites result from the so-called strong
metal support interactions (SMSI) which are described as
0
the migration of TiO entities onto the Pt particles [32].
x
These sites exist on Pt/TiO after reduction at a low tempera-
2
◦
ture (200 C), and their number increases when the reduction
temperature increases up to 500 C. An additional interpre-
◦
(
∼ 8 wt%) than the DP NaOH samples (∼ 3 wt%).
–
–
For a given temperature, the metal gold particles are
smaller when the catalysts are pretreated under hy-
drogen than under air (Table 3). In addition, gold is al-
ready fully reduced at 120 C under hydrogen (Fig. 6)
whereas under air, a temperature of 200 C must be
tation was proposed by Lercher and colleagues [21] who at-
tributed the higher selectivity and activity found in catalysts
containing large Pt particles, i.e., after high-temperature re-
duction, to the fact that large particles contain a high frac-
tion of Pt atoms on face (111). This face would favor the
adsorption of crotonaldehyde via the carbonyl bond to the
detriment of the adsorption via the olefinic bond. In a for-
mer theoretical work [33], Delbecq and Sautet had already
proposed that face (111) better adsorbs the C=O bond than
the C=C one.
◦
◦
reached to obtain fully reduced gold particles [23].
The average particle sizes in the catalysts do not dras-
tically increase when the reduction temperature in-
creases (Tables 2). This is remarkable since nanopar-
ticles of gold are supposed to melt at a much lower

R. Zanella et al. / Journal of Catalysis 223 (2004) 328–339
337
On Au/TiO2 catalysts, the evolution of the activity and
selectivity versus the reduction temperature and the parti-
cle size is completely different. The same high selectivity
Pt/ZnO exhibits a larger crotyl alcohol yield than Pt/TiO2
because of the easy formation of a PtZn alloy and of the
presence of remaining chlorine atoms on the support.
Hence, the results obtained by Hutchings and collea-
gues [15], Claus et al. [16], and us are not consistent from
the point of view of the influence of the gold particle size,
but the nature of the support in [15] and the experimental re-
action conditions [15,16] (temperature and/or pressure) are
not similar either. Now, let us go further with a few more
comments regarding our results.
(
60–70%) was obtained for all the reduction temperatures
◦
between 120 and 500 C (Table 2), and whether the thermal
pretreatments were performed under hydrogen or air (Ta-
ble 3). This last point deserves a comment. Lin and Vannice
[
34] established by O2 chemisorption that oxygen vacancies
were present on titania after Au/TiO2 reduction under H2,
and that their concentration markedly increased with the re-
duction temperature. The fact that our catalytic results do
not show any dependence with the gas used for reduction
4
.3. Comments on the reaction mechanism of
(
H2 or air) indicates that the oxygen vacancies play no role
crotonaldehyde hydrogenation on Au/TiO2 catalysts
in the crotonaldehyde hydrogenation on Au/TiO2 catalysts.
It may also be noted that the HRTEM images (Fig. 9) do not
show any decoration of the gold particles by TiOx entities.
Moreover, in contrast to Pt/TiO2 catalysts, the activity and
TOF of Au/TiO2 catalysts significantly decrease when the
particle sizes increase (Fig. 8), and the selectivities are only
slightly smaller for the largest particles (50–60%) (Table 3).
Therefore, it turns out to be difficult to apply the concepts
proposed for Pt/TiO2 catalysts, to explain the high selectiv-
ity to crotyl alcohol for Au/TiO2 catalysts.
The selectivity to C=O versus C=C hydrogenation is
mainly dictated by the extent of activation of the C=O
group [15]. Therefore, the adsorption mode of crotonalde-
hyde is the determining factor, which orientates the selectiv-
ity toward the formation of the monohydrogenatedproducts.
The possible adsorption modes of α, β-unsaturated aldehy-
des on monocrystalline metal surfaces were described by
Delbecq and Sautet [33,36]. If the adsorption is restricted
2
to that of the C=C moiety of the molecule (η of di-σCC-
CC
type) (Scheme 2), hydrogenation would only lead to butanal.
4
.2. Comparison with published data on the effect of gold
Conversely, if the adsorption is restricted to that of the C=O
particle size on selective hydrogenation of α, β-unsaturated
aldehydes
bond (η2 of di-σCO-type), or to that of the oxygen lone
CO
pair (atop), only crotyl alcohol could form. In the case of
gold catalysts, both products are formed, and we must then
assume either the coexistence of two independent C=C and
C=O (or O atop) adsorption modes or a unique adsorption
As reported in the Introduction, several teams have at-
tempted to correlate the hydrogenation properties with the
gold particle sizes [15,16]. The results presented in this pa-
per, which cover a large range of particle sizes, from 1.7 to
3
mode involving both C=C and C=O (or O), such as the η
4
or η modes (Scheme 2).
8
.7 nm, with monomodal size distributions, are somewhat
In contrast to platinum-based catalysts, the presence of
different from the published results.
First, we observed a significant increase of activity and
TOF when the gold particle size decreases, especially in the
2
sites, which would hydrogenate the C=C bond only (η CC
structure), is unlikely on gold catalysts. Indeed, gold cata-
lysts are active for the partial hydrogenation of butadiene
1
–2 nm range (Fig. 8). The only study covering this range of
small sizes is that of Claus et al. [16] on Au/TiO2 catalysts
1.1, 2.0, and 5.3 nm), but, as noted in the Introduction, they
[
12,37–41] or acetylene [11], but very poorly active for the
hydrogenation of butene or ethylene. We found for the DP
Urea 16 h catalyst that butadiene reacted twice as fast as
crotonaldehyde under the same conditions of pressure and
temperature, but this catalyst was totally inactive for butene
hydrogenation (unpublished results). This last result makes
(
obtained the opposite results, i.e., a large decrease of activity
for the smallest particle sizes. However, it may be noted that
the authors proposed an alternative explanation for the low
activity of the samples containing 1-nm particles, the bury-
ing of the metal particles into the support, due to the fact that
the sample had been prepared by the sol–gel method.
2
unlikely the η CC adsorption mode of olefin on gold parti-
cles, and therefore, that of crotonaldehyde.
Second, a constant selectivity to unsaturated alcohol (60–
0%) was obtained within a large size range, and whatever
On the other hand, the fact that the selectivities to crotyl
alcohol and butanal are almost constant whatever the gold
particle size and the reduction temperature is not in favor
of the coexistence of two types of adsorption modes involv-
ing two types of catalytic sites whose proportions depend on
the structure/size of the gold particles. Therefore, we rather
propose that a single mode of adsorption of crotonaldehyde
governs the selectivity, either η3 or η4 adsorption mode
(Scheme 2), and that the addition of two hydrogen atoms
gives rise to both butanal and crotyl alcohol. The selectiv-
ities to each product would be determined by the relative
7
the gold loading and the conditions of reduction. Our results
drastically contrast with the conclusion of Hutchings and
colleagues [15], that the active sites for the selective hydro-
genation of crotonaldehyde to crotyl alcohol are associated
with the presence of large Au particles. It may be noted that
their study did not cover the same particle range (4 to 20 nm).
In addition, they used a different support, ZnO, which may
also have an influence on the reactivity of gold. Indeed, in
the case of supported Pt catalysts, one of us [35] showed that

338
R. Zanella et al. / Journal of Catalysis 223 (2004) 328–339
Scheme 2. Adsorption modes of the molecule of crotonaldehyde on metal, from Ref. [36].
probabilities of addition of the hydrogen atoms to the differ-
ent positions in the molecule (O, Cα, Cβ, Cγ ).
as that reported by Buchanan and Webb [37] for the se-
lective hydrogenation of butadiene on Au/Al O catalysts,
2
3
−1
The dramatic increase of the TOF when the particle
size decreases (Fig. 8) suggests that the gold atoms in-
volved in the hydrogenation process are those located on
edges and corners. For instance, the relative number of
low-coordinated sites in cubooctahedral-shape particles [42]
varies in the same way as the TOF versus the particle size
36.5 kJ mol . This could support the hypothesis that butadi-
ene and crotonaldehyde hydrogenationkinetics are governed
by a common rate-determining step, the hydrogen disso-
ciation. In [37], the Au/Al2O3 catalysts were prepared by
impregnation (1 wt% of Au) and exhibited a broad distribu-
tion of particle sizes up to 80 nm with the largest population
of size in the 0–10 nm range. Because of the complete ab-
sence of activity of the support in butadiene hydrogenation,
Buchanan and Webb proposed a mechanism where hydro-
gen dissociated on alumina and migrated to gold particles to
hydrogenate the adsorbed butadiene. However, they did not
exclude the fact that the activation of H2 could take place
on tiny gold particles, which could not have been detected
by electron microscopy (the smallest particle size they could
detect with their microscope was not reported in their pa-
per).
(
Fig. 8). Besides this pure size effect, the evolution of the
particle morphology observed on the HRTEM micrographs
with the reduction temperature leads to the same conclu-
◦
sion: the gold particles obtained after reduction at 200 C
(
Fig. 9A) do not exhibit well-developed facets while those
◦
obtained after reduction at 500 C (Fig. 9B) are perfectly
facetted with low index planes. Therefore, the proportion
of low-coordinated sites is much lower after pretreatment at
◦
◦
5
00 C than at 200 C, and the decrease in catalytic activ-
ity and TOF when the reduction temperature increases from
◦
00 to 500 C is attributed to the smoothing of the outer
2
surface of the gold particles, i.e., to the decrease in the num-
ber of low-coordinated sites. This is in agreement with the
interpretation of Claus and colleagues [17] where they at-
tributed the higher activity of Au/TiO2 compared to that of
Au/ZrO2 to the fact that the gold particles were round shaped
on TiO2 and facetted on ZrO2. In another work [43] on selec-
tive decoration of the faces of gold particles by indium, they
proposed that the active sites favoring the adsorption of the
C=O group of acrolein are the edges of the gold nanoparti-
cles.
The poor catalytic activity in hydrogenation with gold
catalysts compared to Pd or Pt catalysts has been generally
attributed to the fact that gold hardly dissociates molecular
H2, as it has been established experimentally and theoreti-
cally [44–46]. However, several authors have unexpectedly
observed the dissociative chemisorption of H2 (D2) on gold
films. For instance, Okada et al. [47], who studied the ad-
sorption and reaction of hydrogen (deuterium) on thin gold
film grown on Ir (111), noted that surprisingly H2 (D2) can
dissociatively adsorb on these Au films. They attributed this
property to locally relaxed surface features like defects. On
thin gold film deposited on Pyrex glass, Stobinski et al. [48]
also found that H2 dissociative chemisorption occurs, pro-
viding that the thin Au films were deposited at a low tem-
perature (78 K) and were not sintered afterward. They con-
cluded that H2 chemisorption occurs only on surface atoms
of low-coordination number. These conclusions support our
proposals for the hydrogenation mechanism.
However, these low-coordinated sites are not prone to
4
favor the η -adsorption mode for which the reactant struc-
3
ture is rather planar or the η -adsorption mode for which
multiple metallic sites are involved (Scheme 2), but they
might be more adapted to dissociate the hydrogen mole-
cule. Therefore, we propose that the activity of gold cat-
alysts in hydrogenation is determined by the dissociation
step of molecular hydrogen. The apparent activation energy
−1
calculated in the present work, 36.3 kJ mol , is the same

R. Zanella et al. / Journal of Catalysis 223 (2004) 328–339
339
5
. Conclusion
[7] C.T.H. Stoddart, C. Kemball, J. Colloid Sci. 11 (1956) 532.
[8] M. Haruta, H. Kageyama, N. Kamijo, T. Kobayashi, F. Delannay, Stud.
Surf. Sci. Catal. 44 (1988) 33.
Thanks to the method of deposition-precipitation with
[9] M. Haruta, T. Kobayashi, H. Sano, N. Yamada, Chem. Lett. 2 (1987)
urea, it has been possible to prepare a series of Au/TiO2
samples with the same gold loading (∼ 8 wt%), and various
average particle sizes (1.7 to 8.7 nm), either by varying the
duration of the preparation, or by changing the conditions
of reduction (nature of the gas, temperature, flow rate). With
this series of Au/TiO2 catalysts, it has been possible to study
systematically the influence of the gold particle size on their
catalytic properties in the hydrogenation of crotonaldehyde,
4
05.
[
10] G.C. Bond, D.T. Thompson, Catal. Rev.-Sci. Eng. 41 (1999) 319.
[11] J. Jia, K. Haraki, J.N. Kondo, K. Domen, K. Tamaru, J. Phys Chem.
B 104 (2000) 11153.
12] M. Okumura, T. Akita, M. Haruta, Catal. Today 74 (2002) 265.
13] C. Milone, M.L. Tropeano, G. Gulino, G. Neri, R. Ingoglia, S. Gal-
vagno, Chem. Commun. (2002) 868.
[
[
[
14] J.E. Bailie, G.J. Hutchings, Chem. Commun. (1999) 2151.
[15] J.E. Bailie, H.A. Abdullah, J.A. Anderson, C.H. Rochester, N.V.
Richardson, N. Hodge, J.G. Zhang, A. Burrows, C.J. Kiely, G.J.
Hutchings, Phys. Chem. Chem. Phys. 3 (2001) 4113.
◦
under mild experimental conditions (∼ 100 C, 0.1 MPa),
already used in former studies on Pt-based catalysts. The
selectivity to carbonyl bond hydrogenation (crotyl alcohol
formation) was twice as high as for olefinic bond hydrogena-
tion (butanal formation). The selectivity was not sensitive
to the reduction temperature, as was the case for platinum
catalysts whose selectivity to crotyl alcohol was generally
much lower (< 50%) than that for gold. The gold catalysts
led to very reproducible results. It must be pointed out that
the Au/TiO2 catalysts were much less active by about two
orders of magnitude than Pt/TiO2 catalysts. The structure
sensitivity of the Au/TiO2 catalysts for the hydrogenation
of crotonaldehyde was clearly shown by the significant in-
crease in activity and TOF for gold particles of ∼ 2 nm. The
possible role of the oxygen vacancies of TiO2 at the metal-
support interface, whose number is known to increase with
the reduction temperature, was not evident as in the case of
Pt/TiO2. It is proposed that the dissociation of hydrogen on
the low-coordinated sites of gold particles (edges and cor-
ners) is the rate-determining step of this reaction, and that
[
16] P. Claus, A. Brückner, C. Mohr, H. Hofmeister, J. Am. Chem. Soc. 122
2000) 11430.
17] C. Mohr, H. Hofmeister, P. Claus, J. Catal. 213 (2003) 86.
(
[
[18] R. Zanella, S. Giorgio, C.R. Henry, C. Louis, J. Phys. Chem. B 103
2002) 7634.
(
[
19] S. Tsubota, D.A.H. Cunningham, Y. Bando, M. Haruta, Stud. Surf.
Sci. Catal. 91 (1995) 227.
20] S. Tsubota, M. Haruta, T. Kobayashi, A. Ueda, Y. Nakahara, Stud.
Surf. Sci. Catal. 72 (1991) 695.
[21] M. Englisch, A. Jentys, J.A. Lercher, J. Catal. 166 (1997) 25.
22] R. Makouangou, D.Yu. Murzin, A. Dauscher, R. Touroude, Ind. Eng.
[
[
[
[
Chem. Res. 33 (1994) 1881.
23] R. Zanella, S. Giorgio, C. Shin, C.R. Henry, C. Louis, J. Catal. 222
(2004) 357.
24] M. Consonni, D. Jokic, D.Yu. Murzin, R. Touroude, J. Catal. 188
(1999) 165.
25] W.A. Dietz, J. Gas Chromatogr. 5 (1967) 68.
26] M. Consonni, R. Touroude, D.Yu. Murzin, Chem. Eng. Technol. 21
[
[
(
1998) 7.
27] J. Simonik, L. Beranek, Collect. Czech., Chem., Commun. 37 (1972)
53.
[28] M.A. Vannice, B. Sen, J. Catal. 115 (1989) 65.
29] C.G. Raab, J.A. Lercher, Catal. Lett. 18 (1993) 99.
[30] P. Buffat, J.P. Borel, Phys. Rev. A 13 (1976) 2287.
[
3
3
[
crotonaldehyde is adsorbed on gold particles through a η or
4
η mode of adsorption.
[
31] P. Claus, S. Schimpf, R. Schödel, P. Kraak, W. Mörke, D. Hönicke,
Appl. Catal. A 165 (1997) 429.
[32] S.J. Tauster, Acc. Chem. Res. 20 (1987) 389.
Acknowledgment
[33] F. Delbecq, P. Sautet, J. Catal. 152 (1995) 217.
[
[
[
[
34] S. Lin, M.A. Vannice, Catal. Lett. 10 (1991) 47.
35] F. Ammari, J. Lamotte, R. Touroude, J. Catal. 221 (2004) 32.
36] F. Delbecq, P. Sautet, J. Catal. 211 (2002) 389.
Rodolfo Zanella is indebted to CONACYT (Mexico) and
SFERE (France) for his PhD grant, and to FESC, UNAM.
37] D.A. Buchanan, G. Webb, J. Chem. Soc., Faraday Trans. I 71 (1975)
134.
[38] C. Mohr, P. Claus, Sci. Progr. 84 (2001) 311.
References
[39] M. Shibata, H. Kimura, T. Masumoto, J. Chem. Soc. Jpn. 6 (1990) 628.
[
[
[
[
40] H.G. Ahn, H. Niiyama, J. Korean Ind. Eng. Chem. 10 (1999) 1003.
41] H.G. Ahn, H. Niiyama, J. Korean Ind. Eng. Chem. 12 (2001) 920.
42] R. Hardeveld, F. Hartong, Surf. Sci. 15 (1969) 189.
[1] S. Naito, M. Tanimoto, J. Chem. Soc., Faraday Trans. I 84 (1998)
4115.
43] C. Mohr, H. Hofmeister, J. Radnik, P. Claus, J. Am. Chem. Soc. 125
[
[
2] S. Naito, M. Tanimoto, Chem. Commun. (1988) 832.
3] P.A. Sermon, G.C. Bond, P.B. Wells, J. Chem. Soc., Faraday Trans.
I 75 (1979) 385.
(2003) 1905.
[
[
[
[
44] B. Hammer, J.K. Norskov, Nature 376 (1995) 238.
45] J. Harris, Surf. Sci. 221 (1989) 335.
46] A.G. Sault, R.J. Madix, C.T. Campbell, Surf. Sci. 169 (1986) 347.
47] M. Okada, M. Nakamura, K. Moritani, T. Kasai, Surf. Sci. 523 (2003)
[4] D.V. Sokol’skii, A. Ualikhanova, Russ. J. Phys. Chem. 56 (1982)
633.
[5] R.P. Chambers, M. Boudart, J. Catal. 5 (1966) 517.
218.
[6] J. Erkelens, C. Kemball, A.K. Galwey, Trans. Faraday Soc. 59 (1963)
[48] L. Stobinski, L. Zommer, R. Dus, Appl. Surf. Sci. 141 (1999) 319.
1181.