Pt/Ga2O3 catalysts of selective hydrogenation of crotonaldehyde

Article abstract of DOI:10.1016/j.jcat.2007.06.021

Hydrogenation of crotonaldehyde in a gas phase at atmospheric pressure over Pt/Ga₂O₃ catalysts was studied. Two types of platinum precursors [Pt(acac)₂ and H₂PtCl₆] and two gallia supports (α-Ga₂

Full text of DOI:10.1016/j.jcat.2007.06.021

Journal of Catalysis 250 (2007) 195–208
www.elsevier.com/locate/jcat
Pt/Ga2O3 catalysts of selective hydrogenation of crotonaldehyde
E. Gebauer-Henke a, J. Grams a, E. Szubiakiewicz a, J. Farbotko a, R. Touroude b, J. Rynkowski a,∗
a
˙
Institute of General and Ecological Chemistry, Technical University of Łódz´, ul. Zeromskiego 116, 90-924 Łódz´, Poland
b
LMSPC, UMR 7515 du CNRS, ECPM-ULP, 25 rue Becquerel, 67087 Strasbourg Cedex 2, France
Received 27 February 2007; revised 11 June 2007; accepted 17 June 2007
Available online 27 July 2007
Abstract
Hydrogenation of crotonaldehyde in a gas phase at atmospheric pressure over Pt/Ga O catalysts was studied. Two types of platinum precursors
2
3
2
2
[Pt(acac) and H PtCl ] and two gallia supports (α-Ga O [60.7 m /g] and β-Ga O [2.2 m /g]) were used for catalyst preparation. The catalyst
2
2
6
2
3
2 3
5 wt% Pt/α-Ga O prepared from Pt(acac) precursor showed very high and stable selectivity to crotyl alcohol (91% and 63% at 10% and 70%
2
3
2
crotonaldehyde conversion, respectively). Because Pt-supported catalysts, showing a high selectivity, usually have relatively low activity, the most
interesting finding of this study is that the use of gallium oxide as a support for platinum makes it possible to obtain catalysts with significantly
increased C=O bond hydrogenation selectivity while maintaining high activity.
© 2007 Elsevier Inc. All rights reserved.
Keywords: Platinum; Selective hydrogenation; Crotonaldehyde; Gallium oxide; Unsaturated alcohol
1. Introduction
reaction to a preferred product, whereas the amount of unde-
sired products is decreased. The amount of energy required
for a given transformation is also reduced as the catalysts de-
crease the activation energy of the reaction [2]. Consequently,
hydrogenation processes based on heterogeneous catalysis have
been developed [3]. Hydrogenation of α,β-unsaturated alde-
hydes to unsaturated alcohols is difficult to realize, especially
under heterogeneous catalysis conditions, due to the much
greater susceptibility to hydrogenation of the C=C double
bond (578.8 kJ/mol) compared with the C=O double bond
(705.8 kJ/mol) for both thermodynamic and kinetic reasons [4].
The reaction of hydrogenation of α,β-unsaturated aldehy-
des may occur in different ways, as shown in Fig. 1. α,β-Un-
saturated aldehydes may be hydrogenated to the most desired
product, unsaturated alcohols (step 1), as well as to other sat-
urated products, butanal (steps 2 and 5) and butanol (steps 3
and 4).
The selectivity to unsaturated alcohols depends on various
factors, such as the structure of aldehydes, the nature of cat-
alysts, and the presence of additives, as well as the reaction
conditions [5]. Previous studies have shown that unpromoted
metals have specific selectivities to unsaturated alcohols. For
example, iridium and osmium are rather selective; palladium,
rhodium, and nickel are unselective or slightly selective; and
Selective hydrogenation of α,β-unsaturated aldehydes to
unsaturated alcohols has been a subject of many investigations
because of its great interest as an important step in the prepa-
ration of various fine chemicals and applications of unsaturated
alcohols in the flavor and fragrance industries and in pharma-
ceutical manufacturing [1]. The challenge lies in the selective
hydrogenation of the carbonyl bond while keeping the olefinic
bond unaffected. This can be achieved using stoichiometric
amounts of reducing agents, such as hydrides (e.g., AlLiH4,
NaBH4); however, such methods are useful only for a small-
scale production because they involve costly chemicals and are
hazardous to the environment.
The development of specific catalytic hydrogenation proc-
esses can replace the reduction using metal hydrides, which
is in agreement with one of the 12 principles of green chem-
istry: “Catalytic reagents (as selective as possible) are superior
to stoichiometric reagents.” Catalysis offers advantages over
stoichiometric reactions in terms of both increased selectivity
and energy minimization. The specificity of catalysts favors the
*
Corresponding author. Fax: +48 42 631 31 28.
E-mail address: jacryn@p.lodz.pl (J. Rynkowski).
0021-9517/$ – see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2007.06.021

196
E. Gebauer-Henke et al. / Journal of Catalysis 250 (2007) 195–208
Fig. 1. Selective hydrogenation of crotonaldehyde: catalytic pathways.
Fig. 2. Reaction sites for hydrogenation of α,β-unsaturated aldehydes on metallic-promoted catalysts [24].
platinum, ruthenium and cobalt are moderately selective. These
trends were confirmed by Cordier for cinnamaldehyde hydro-
genation (Os > Ir > Pt > Ru > Rh > Pd) [6]. Different geo-
metric and electronic properties of metals can affect hydrogena-
tion activities and selectivities by influencing not only adsorp-
tion, but also surface reactions. The adsorption phenomenon
itself is limited by many factors, such as the mode of reactant
adsorption on a specific metal surface. Different crystal phases
of metals have different geometrical constraints, which explains
the need to study the performance of distinct crystallographic
faces. Platinum has been used most intensively as an active
metal in chemoselective hydrogenations. Among the nonno-
ble metals, special attention has been given to the supported
cobalt catalysts. This approach, which was initiated by Japanese
scientists [7–10], was developed in the works of Hutchings et
al. [11–13], Rodrigues and Bueno [14–16], and in our previ-
ous paper [17]. Recently, interesting properties of bimetallic
Pt-Co/SiO2 were reported by Borgna et al. [18]. Nonetheless,
systems based on platinum are still the most frequently explored
catalysts of crotonaldehyde hydrogenation. Pt is a very active
metal in the hydrogenation of the C=C double bond but is not
intrinsically selective to unsaturated alcohols. The selectivity to
the desired product greatly depends on the steric and electronic
structure of the active phase and reactant [19].
Ca, Cd) ionic compounds, (b) transition metal (Ti, V, Cr, Mn,
Fe) compounds, and (c) nontransition element (Ge, Ga, Sn)
compounds, with an increased promoting effect in the order
(a) < (b) < (c) [20].
Earlier work [21,22] has shown the role and effects of pro-
moters and metal–support interaction in the reaction of hydro-
genation of α,β-unsaturated aldehydes. As a general rule, ac-
tivity and selectivity to unsaturated alcohols are improved when
adding Sn, Ge, Ga, Fe, and others to active transitional metals
(e.g., Pt, Ru, Ni) and when a partially reducible support like
TiO2 is used [23]. The promoting effect is seen as a strong in-
teraction between the carbonyl group and a positively charged
center (e.g., Feδ+, Snδ+, TiOδx+) (Fig. 2) [24].
Gallium has been mentioned as an effective promoter in the
reduction of carbonyl group in Pt-based catalysts [5]. Englisch
et al. [22] showed that transitional metals like Ga can be used
as promoters in catalysis for selective hydrogenation of α,β-
unsaturated aldehydes. Transition metals used in catalysts have
partially filled d-orbitals. When the platinum d-orbital is filled,
the metal is relatively catalytically inactive, as in the cases of
silver, gold, and copper. The interstitial electron densities in
metals filling d-orbitals are low, and these metals are not ea-
ger to form bonds. However, the electronic structure of metal
can be changed by adding a second metal, leading to, for in-
stance, the formation of an alloy, by changing the metal particle
size and enhancing the interaction with the support. This in-
teraction decreases electron density at the carbonyl bond of
α,β-unsaturated aldehyde and increases its reactivity. The re-
Enhanced selectivity can be obtained by an increase in C=O
bond activity, with a simultaneous decrease in C=C bond ac-
tivity. This can be achieved by using various promoters, which
were divided into three groups: (a) alkali metal (Na, K, La,

E. Gebauer-Henke et al. / Journal of Catalysis 250 (2007) 195–208
197
duction to unsaturated alcohol occurs if the sorption of the C=O
group occurs at the Pt–gallium oxide interface, where the car-
bonyl group can be activated by Lewis acid sites and dissociated
hydrogen can be supplied by Pt.
inductively coupled plasma (ICP) analyses were performed on
calcined samples to evaluate metal loadings.
2.3. Catalytic tests
Based on the literature [25] and our preliminary studies, we
suggest that Ga2O3 should be considered a partly reducible ox-
ide of potential usefulness as a catalyst support. Consequently,
it seemed interesting to test the catalytic properties of Pt/Ga2O3
systems in the reaction of selective hydrogenation of croton-
aldehyde, never before studied in this context. The aim of this
work was to determine the physicochemical characterization
and catalytic performance of Pt/Ga2O3 systems in the reaction
of crotonaldehyde hydrogenation in a gas phase.
Catalytic tests were carried out in a glass reactor, operat-
ing at atmospheric pressure. The total gas flow controlled by a
flowmeter varied with the changes in the pump rate at the end of
the flow line. Pure crotonaldehyde (Fluka), stored in argon, was
used as received. A specific quantity (50–250 µL) was drawn
up from the bottle using a tight syringe and introduced through
a vaccine cap into a reservoir installed online and maintained
at 0 ◦C: therefore, aldehyde at constant partial pressure (8 Torr)
was carried over the catalyst by hydrogen flow (50–60 cc/min).
Beyond the aldehyde reservoir, the gas line was thermostatted
at about 60 ◦C to avoid any condensation. The reaction prod-
ucts were drawn off of the flow line at different times during
the catalytic run and analyzed by gas chromatography at 85 ◦C,
using a flame ionization detector.
2. Experimental
2.1. Support preparation
Two gallium oxide supports were used: α-Ga2O3 SBET
=
Catalyst samples (5–100 mg) were reduced in situ at 100,
200, or 300 ◦C for 1 h and cooled under H2 before the reaction
carried out at 50 and 80 ◦C.
60.7 m2/g, mesoporous and β-Ga2O3 (Aldrich 21506-6,
99.99%, SBET = 2.2 m2/g, nonporous). α-Ga2O3 was prepared
in our laboratory by dissolving metallic gallium in concentrated
nitric acid (V), followed by precipitation of Ga(OH)3 with NH3
aq., drying at 120 ◦C, and calcination at 500 ◦C for 3 h in air.
The polymorphic forms of supports were determined by XRD.
The reaction activities were calculated using the formula
A = αF/ω, where α is the crotonaldehyde conversion (in %),
F is the flow rate of crotonaldehyde (in mol/s), and ω is the
weight of platinum (in g). The selectivity to different products
(crotyl alcohol, butanal, butanol, and hydrocarbons) was cal-
culated as the molar ratio of the selected product to the total
amount of products formed. The sensitivity factors are taken as
1.4 for the hydrocarbons and 1 for the other products.
It was checked that both supports alone, α- and β-Ga2O3,
after H2 pretreatment up to 300 ◦C, had no activity in croton-
aldehyde hydrogenation.
2.2. Catalyst preparation
Pt/Ga2O3 catalysts were prepared by a wet-impregnation
method of α- and β-Ga2O3, using an aqueous solution of
hexachloroplatinic acid (H2PtCl6; P.O. Ch Gliwice, Poland)
and a methanolic solution of platinum acetyl acetate [Pt(acac)2;
Aldrich]. Water and methanol were evaporated slowly on a hot
plate. The samples were dried at 120 ◦C for 2 h and then cal-
cined in air at 200 ◦C for 2 h.
2.4. Nitrogen adsorption (BET) measurements
The composition of catalysts and other abbreviations are
presented in Table 1 (α and β, polymorphs of gallium oxide;
A, platinum acetylacetonate; H, hexachloroplatinic acid; 5, 2,
1, and 0.5, percentages of the active phase). Atomic absorp-
tion spectroscopy (AAS) (at CNRS, Vernaison, France) and
BET surface areas and total pore volume for all catalysts
and supports were obtained using a Sorptomatic 1900 (Carlo
Erba Instruments), after heating the samples at 200 ◦C under
vacuum.
2.5. TG–DTA–MS measurements
Table 1
Catalysts composition
A thermogravimetric (TG) analyzer, equipped with a dif-
ferential thermal analysis (DTA) device (Seteram SETSYS-
16/18) and combined inline with a quadrupole mass spectrom-
eter (MS) (Thermostar, Balzers) were used for the temperature-
programmed reduction of the catalyst active-phase precursor.
The TG–DTA–MS measurements were carried out with a ca.
20 mg sample, at a linear temperature increase of 10 ◦C per
minute to a temperature range of 25–1100 ◦C.
Catalysts
Symbols
Pt precursors
Pt conc.
(wt%)
Cl conc.
(wt%)
a
5 wt% Pt/α-Ga O
α5A
β5A
α5H
β5H
α2A
β2A
α1A
β1A
α0.5A
β0.5A
Pt(acac)
Pt(acac)
5.16
–
–
2
3
2
a
5 wt% Pt/β-Ga O
4.79
2
3
3
3
3
3
3
3
2
a
5 wt% Pt/α-Ga O
H PtCl
2
5.25
1.31
<0.2
2
6
a
5 wt% Pt/α-Ga O
H PtCl
2
5.38
2
6
b
2 wt% Pt/α-Ga O
Pt(acac)
Pt(acac)
Pt(acac)
Pt(acac)
Pt(acac)
Pt(acac)
1.97
–
–
–
–
–
–
2
2
2
2
2
2
2
b
2 wt% Pt/α-Ga O
1.87
2
b
1 wt% Pt/α-Ga O
0.97
2
b
2.6. XRD measurements
1 wt% Pt/α-Ga O
0.99
2
b
0.5 wt% Pt/α-Ga O
0.49
2
3
3
b
0.5 wt% Pt/α-Ga O
0.47
2
XRD analyses were carried out in Siemens D 5000 polycrys-
talline diffractometer using CuKα radiation and X’PERT PRO
MPD PAN polycrystalline diffractometer. Patterns of supports
a
AAS analysis.
ICP analysis.
b

198
E. Gebauer-Henke et al. / Journal of Catalysis 250 (2007) 195–208
were recorded in the scanning mode (2.0 s/step, step size: 0.05◦
2θ, 2θ range 20◦–100◦). Spectra of the catalysts (after calcina-
tion at 200 ◦C) were recorded in the same scanning mode but
in a 2θ range of 38◦–48.5◦ (the region characteristic of for Pt,
PtOx, and some lines characteristic of α- and β-Ga2O3).
The mean crystallite size of Ga2O3 samples was related to a
pure X-ray broadening by Scherer’s formula, D = kλ/β cosθ.
The measured line broadening b was corrected for the instru-
√
mental broadening B, using the equation β = B2 − b2. The
particle shape factor k was taken as 0.9. Phase identification
was made based on JCPDS files.
2.7. TPR measurements
All TPR measurements were performed on an Altamira ap-
paratus using an Ar–H2 mixture (5 vol% H2) and a flow rate
of 50 cc/min, at a temperature ramp rate of 10 ◦/min. Before
TPR, the samples (100 mg) were calcined in air at 200 ◦C for
2 h.
Fig. 3. Activity versus time on stream for 5 wt% Pt/Ga O (temp. red. =
2
3
◦
◦
300 C, temp. reaction = 80 C, conv. < 20%).
2.8. ToF-SIMS measurements
The ToF-SIMS investigations (α5A, β5A and α2A, β2A)
were performed in the static mode using an ION–TOF instru-
ment (TOF SIMS IV) equipped with a 25-kV pulsed 69Ga+
primary ion gun. To obtain the plain surface of catalysts, pow-
der samples were tableted before the measurements. For both
catalyst samples, five spectra were collected (from different ar-
eas) to determine the distribution of Pt on the support. The
195Pt−/71GaO− and 35Cl−/OH− intensity ratios were calcu-
lated.
2.9. TEM
Fig. 4. Selectivity versus time on stream for 5 wt% Pt/Ga O (temp. red. =
2
3
◦
◦
300 C, temp. reaction = 80 C, conv. < 20%).
High-resolution TEM images were recorded on a TOPCOM
002B electron microscope, operating at 200 kV, with a point-to-
point resolution of 0.18 nm. Some catalyst grains obtained after
different catalytic pretreatments performed in the catalytic ap-
paratus were ground and diluted in an ethanolic solution. One
drop of this solution previously dispersed in an ultrasonic tank,
was deposited onto a Cu grid coated by a holed carbon film
and dried in air. Various regions of the grid were observed and
particle sizes were measured from the observation of 250–350
particles. The following formula was used to calculate the mean
3. Results and discussion
3.1. Catalytic tests
Detailed studies of the reaction of crotonaldehyde hydro-
genation in a gas phase were carried out to determine the in-
fluence of polymorphic form of the support (gallium oxide
[α, β]), platinum precursor [Pt(acac)2, H2PtCl6], catalyst re-
duction temperature, reaction temperature, and platinum load-
ing on the catalytic performance of the Ga2O3-supported Pt
catalysts under study. For most of the catalysts, we checked the
catalytic performance at 180 min time on stream. We found that
after 40–60 min, the activity did not change and remained sta-
ble. That is why for all of the catalytic data presented in this
paper, we limited the time on stream axis to 90 min.
Figs. 3 and 4 show the activity and selectivity of the 5 wt%
Pt/Ga2O3 catalysts, illustrating the influence of different gallia
polymorphs and catalyst precursors. During the first 45 min of
time on stream, decreased activity can be seen, but subsequently
the reaction reached a quasi-steady state. This effect is espe-
ꢀ
ꢀ
surface diameter: ds = nidi3/ nidi2, where ni is the num-
ber of particles of diameter di.
2.10. FTIR measurements
FTIR measurements were obtained with a Shimadzu 8501
spectrometer using a quartz cell fitted with NaCl windows and
an external furnace. Catalyst samples (80 mg) were prepared
in the form of 25-mm-diameter discs. After the reduction in
hydrogen at 300 ◦C, the cell was evacuated for 30 min at room
temperature. Pulses of CO or crotonaldehyde were introduced
at 25 ◦C, and spectra were recorded.

E. Gebauer-Henke et al. / Journal of Catalysis 250 (2007) 195–208
199
◦
Fig. 5. Selectivity versus time on stream for α5A catalyst (temp. red. = 300 C,
◦
temp. reaction = 80 C, conv. < 20%).
Fig. 6. Activity and selectivity versus time on stream for α5A catalyst, reduced
◦
at different temperature (temp. reaction = 80 C, conv. < 20%).
cially clearly observed for the most active catalysts obtained
from the Pt(acac)2 precursor. The activity of the catalysts pre-
pared from the chlorine-containing catalyst H2PtCl6 was poor.
According to previous work [26–31], the presence of chlorine
in the precursor can have a significant and diverse affect on
the properties of the catalysts. For example, Pt/ZnO catalysts
obtained using H2PtCl6 solution show better selectivity to de-
sired crotyl alcohol than those prepared from Pt(NO3)2(NH3)4
[26,27], whereas the opposite is true for Pt/SnO2 catalysts [28].
A more detailed discussion of the effect of residual chloride in
the chemoselective hydrogenation of unsaturated aldehydes has
been provided by Mäki-Arvela et al. [19].
200, or 300 ◦C. As an example, the effect of the reduction tem-
perature on the catalytic activity and selectivity of the most
active catalyst α5A is presented in Fig. 6.
For all of the catalysts investigated in this work, the best
catalytic performance was obtained for the samples reduced at
300 ◦C. The comparison of activity and selectivity in the reac-
tion of crotonaldehyde hydrogenation for all of the investigated
catalysts reduced at 100 and 300 ◦C is presented in Table 2. All
values in this table are related to the steady state of the reaction
for conversion <20%.
The abatement of the reaction temperature below 80 ◦C re-
sults in decreased activity (as can be expected) as well as in
selectivity to crotyl alcohol to a certain extent. An example of
the efficiency of catalysts α5A and β5A for the reaction carried
out at 50 and 80 ◦C is presented in Fig. 7.
Up to now, all catalytic results have been reported for conver-
sions <20% (usually ca. 10%) in all cases. Proper experimental
parameters ensuring low conversions were chosen to limit the
effects leading to the formation of a completely hydrogenated
product, butanol. However, it is noteworthy that most of the cat-
alysts supported on gallia studied in this work showed very high
selectivity to crotyl alcohol up to 70% conversion. Fig. 8 illus-
trates the dependence of selectivity as a function of conversion
for the most active catalyst, α5A. The data were taken at the
steady-state regime. Considering the extremely high catalytic
activity of this catalyst, such results are remarkable.
Our catalysts prepared from H2PtCl6 were also relatively se-
lective to crotyl alcohol (63% at the activity 2.4 µmol s
−1),
−1
g
Pt
but their much worse catalytic performance compared with that
obtained from Pt(acac)2, made us focus our attention on fur-
ther studies concerning Pt(acac)2 as a precursor. Fig. 5 shows
changes of selectivity to the main products. After the first
50 min, the reaction reached a quasi-stable state. The signifi-
cantly increased selectivity to crotyl alcohol accompanied by
the decreased selectivity to butanol and butanal can be clearly
observed during the first 10 min of the reaction.
In parallel, the activity dropped significantly, although this
process is more extended (Fig. 3). This may suggest that at
the beginning of the reaction, the most active centers on which
the C=C double-bond hydrogenation predominates were sup-
pressed, resulting in decreased activity but increased selectivity
to crotyl alcohol.
To sum up the catalytical studies, we can conclude that
gallia-supported platinum catalysts, prepared from platinum
acetylacetonate as a precursor, showed good performance in
the reaction of selective hydrogenation of crotonaldehyde. The
optimum parameters are the temperature of catalyst reduction
(300 ◦C) and the temperature of reaction (80 ◦C). The 5 wt%
Pt/α-Ga2O3 catalyst (α5A) showed the best catalytic proper-
ties, but lower-loaded catalysts containing 1 and 2 wt% of
platinum showed promising catalytic performance as well (Ta-
ble 2).
Catalysts supported on α-Ga2O3 show significantly higher
activity and selectivity compared with those supported on
β-Ga2O3. This may connected with a much greater specific
surface area of α-Ga2O3 compared with β-Ga2O3, leading to
a difference in platinum particles size (3.5 nm on α-Ga2O3 vs
10.9 nm on β-Ga2O3).
The catalyst reduction temperature, especially in the case of
catalysts supported on reducible supports exhibiting the SMSI
effect, can influence the catalytic properties to a great extent.
Before the reaction, the catalysts were reduced in situ at 100,

200
E. Gebauer-Henke et al. / Journal of Catalysis 250 (2007) 195–208
Table 2
◦
Hydrogenation of crotonaldehyde on Pt/Ga O catalysts at 80 C
2
3
Catalysts
Reduction temperature
◦
◦
100 C
300 C
Act.
Selectivity (%)
Act.
Selectivity (%)
HC
Butanal
Butanol
CROTOL
HC
Butanal
Butanol
CROTOL
α5A
β5A
α5H
β5H
α2A
β2A
α1A
β1A
α0.5A
β0.5A
109
50
3
1
2
3
2
2
3
1
1
–
–
8
11
7
10
15
12
26
31
27
36
8
8
83
79
61
58
63
57
50
45
43
39
159
75
8
2
2
3
4
2
5
1
1
–
–
4
8
14
13
4
5
10
20
24
25
24
28
24
33
27
89
80
63
59
69
62
51
50
48
42
29
30
20
28
23
23
30
25
2
2
78
72
67
65
58
55
100
80
70
70
60
61
9
20
25
19
31
−1
−1
Note. Act.—activity (µmol s
g
). HC—hydrocarbons. CROTOL—crotyl alcohol.
Pt
studies published by others [22,32–37]. Bearing in mind that
one must be very careful when comparing results presented
of different studies due to unavoidable differences in cata-
lyst preparation, experimental conditions, and other parameters,
we believe that our results—especially those obtained on α5A
catalyst—are exceptionally good and promising.
To throw more light on the special catalytic performance of
some catalysts reported in this paper, we carried out physico-
chemical characterization using XRD, TPR, XPS, HRTEM and
ToF-SIMS.
3.2. XRD
For our Ga2O3 sample, XRD patterns show characteristic
lines at 2θ = 24.5◦, 33.8◦, 36.0◦, 40.3◦, 41.4◦, 50.3◦, 55.1◦,
63.4◦ and 64.8◦, corresponding to the diffractions (012), (104),
(110), (006), (113), (024), (116), (214), and (300). These are
in fact identical to those given in ICDS file (00-006-0503) and
the literature [38], demonstrating that we are dealing with pure
α-Ga2O3. Likewise, the commercial β-Ga2O3 shows charac-
teristic lines according to ICDS file (01-076-0573). The mean
crystallite sizes of α-Ga2O3 and β-Ga2O3 were estimated based
on the broadening of XRD lines using Scherer’s formula; these
were ca. 16 nm and 22 nm, respectively. XRD patterns of the
catalysts calcined at 200 ◦C and reduced at 100 and 300 ◦C re-
lated to the pure support are shown in Fig. 9. For catalyst β5A,
both calcined and reduced, a broad line at 2θ about 39.8◦ is ob-
served, which can be ascribed to the Pt(111) diffraction (ICDS
00-004-0802). In contrast to the β5A catalyst, XRD patterns of
the β5H catalyst show a Pt(111) diffraction line for only the re-
duced samples. Indeed, in the calcined sample, platinum occurs
in the form of oxychlorine species. The different intensities of
samples reduced at 100 and 300 ◦C may suggest that the re-
duction of these catalysts at temperatures as low as 100 ◦C is
not complete, as confirmed by TPR results. In addition, an in-
crease in the reduction temperature may lead to an increase in
platinum crystallites.
Fig. 7. Activity (full symbols) and selectivity to crotyl alcohol (empty symbols)
for α5A and β5A catalysts after reduction at 300 C (conv. < 20%).
◦
◦
Fig. 8. Selectivity versus conversion for α5A (temp. red. = 300 C, temp.
◦
reaction = 80 C).
Table 3 compares our results with those reported in our
previous studies of different catalytic systems [26,28–30] and
Interpretation of XRD patterns of the catalysts supported on
α-Ga2O3 is more complex, because the reflex of α-Ga2O3 (006)

E. Gebauer-Henke et al. / Journal of Catalysis 250 (2007) 195–208
201
Table 3
Comparison of catalytic activities and selectivities for different Pt/support catalysts in selective hydrogenation of crotonaldehyde in a gaseous phase
Catalysts
Pretreatment
( C)
Reaction
Steady-state
activity
Selectivity to
crotyl alcohol (%)
(Cv < 20%)
Crotyl alcohol
selectivity (%)
(Cv > 50%)
Reference
This paper
◦
temperature
◦
( C)
a
Pt/Ga O
Red. 300
80
180
85–90
70
2
3
b
0.1
a
Pt/SnO
Red. 443 K
Red. 400
80
80
6
70–75
85–90
50
80
[28]
[26]
2
a
Pt/ZnO
10
0.006
b
a
Pt/CeO
Red. 700
80
80
10, 4
87
80
[29]
[32]
2
b
+calc. 400 K
0.006
+rered. 500
a
Pt/CeO -SiO
Red. 500
Red. 300
21
2
2
a,c
68
a,d
Pt-Sn/SiO
80
80
2.2
70
64
60
[33]
[34]
2
a
Pt/TiO
Red. 400
100
2
+calc. 300
+rered. 200
a
Pt/AC
80
80
80
7
35
82
[35]
[30]
[22]
OX
a
Pt/ZnCl /SiO
Red. 400
Red. 400
6.5
2
2
b
Pt/SiO
Pt-Sn/SiO
Pt-Ga/SiO
Pt-Bi/SiO
Pt-La/SiO
Pt-Pb/SiO
0.017
5
2
b
0.029
30, 5
49
6
7
3
2
b
0.013
2
b
0.022
2
b
0.013
2
b
0.0053
2
a
Pt/ZnO /TiO
Red. 500
Red. 500
80
80
42
50
35
[36]
[37]
2
2
a
Pt/TiO
34
2
a
Pt/SiO
95
11
59
7
50
2
a
Pt-Sn/SiO
800
2
−1
a
−1
Activity calculated as (µmol s
g
).
Pt
b
c
d
−1 −1
Activity calculated as TOF (molec site
s ).
Calculated considering E = 21 kJ [33].
a
).
−1
−1
Initial rate (µmol s
g
Pt
at 2θ = 40.3◦ is almost in line with that coming from platinum.
The only differences observed can be related to the broaden-
ing and intensities of lines, the trend of which is similar to that
discussed for catalysts supported on β-Ga2O3.
occurring mainly on the catalyst surface. A two-stage charac-
ter of this process can be seen. The diverse susceptibility to
the reduction of the surface support species can be explained
assuming that species being in the intimate contact with plat-
inum are reduced more easily at lower temperatures, whereas
those suffering due to a lack of metal–support contact need
spilled-over hydrogen for reduction. It can be clearly seen that
reduction of the support occurs in a significantly lower tem-
perature range for the catalysts supported on α-Ga2O3, which
have a much higher specific surface area compared with those
supported on β-Ga2O3. In such a case, the intimate contact of
platinum and Ga2O3 is preferential, due to better dispersion of
metal on the surface.
TPR profiles of the catalysts prepared from the chlorine-free
precursor Pt(acac)2 do not contain the distinct low-temperature
peaks. This means that either Pt(acac)2 undergoes at least par-
tial decomposition to metallic platinum during catalyst precal-
cination at 200 ◦C or/and PtOx is reduced at room temperature
before the start of TPR. PtOx, especially in the form of large
3.3. TPR and TG–DTA–MS
TPR profiles of the catalysts are presented in Fig. 10. For
the catalysts obtained from hexachloroplatinic acid (α5H and
β5H), a large peak with a maximum at about 160 ◦C (α5H)
and 200 ◦C (β5H) is observed. This peak can be assigned to the
reduction of oxychlorine platinum species formed during pre-
liminary drying and/or calcination of the catalysts precursor at
200 ◦C. A shift of the maximum toward lower temperatures in
the case of the catalyst α5H may be due to better dispersion of
oxychlorine species connected with the higher specific surface
of α-Ga2O3. After termination of this peak, the hydrogen con-
sumption is continued to about 800 ◦C. This can be attributed to
the partial reduction of Ga2O3 promoted by metallic platinum

202
E. Gebauer-Henke et al. / Journal of Catalysis 250 (2007) 195–208
(a)
(b)
(c)
(d)
Fig. 9. XRD diffractograms of (a) α5A, (b) α5H, (c) β5A, (d) β5H.
crystallites, can be reduced at room temperature [39–41]. The
ascending character of the TPR curves of catalysts α5A and
β5A observed at very low temperatures (at the beginning of
the TPR process at room temperature) may be connected with
the further reduction of small platinum oxide crystallites, but
also may indicate surface Pt-promoted support reduction. Until
recently, it was believed that gallium oxide cannot be reduced
by hydrogen below 600 ◦C [42]. However, Zheng et al. [38]
showed that some part of Ga2O3 can be reduced at a low tem-
perature (200–250 ◦C). The H2:Ga2O3 molar ratio calculated
by these authors is 0.1 for α-Ga2O3 and 0.045 for β-Ga2O3.
Promoted by platinum, the surface reduction of gallia may oc-
cur at much lower temperatures.
Fig. 11 presents the thermal analysis of Pt(acac)2 carried
out in air atmosphere. The TG curve shows that the decom-
position occurs directly to the metallic platinum, begins below
200 ◦C, and reaches a maximum at about 255 ◦C. Although the
Fig. 10. TPR for 5 wt% catalysts.

E. Gebauer-Henke et al. / Journal of Catalysis 250 (2007) 195–208
203
Fig. 11. TG–DTA–MS profile of Pt(acac) decomposition.
2
temperature conditions of decomposition of pure and support
deposited catalyst precursors are not always the same, the eas-
iness of Pt(acac)2 decomposition explains the occurrence of
metallic platinum in the calcined catalysts prepared from this
precursor.
a characteristic structure seems to “force” a regular distribu-
tion of platinum. As we demonstrated earlier, the best catalytic
performance, taking into consideration both activity and selec-
tivity to crotyl alcohol, is shown by catalyst α5A. As shown in
Fig. 12, independent of the pretreatment of this catalyst (i.e.,
calcination at 200 ◦C and reduction at 100 and 300 ◦C), in all of
these cases the particle sizes of platinum crystallites are similar
within quite a narrow range (2–4.5 nm). The catalyst reduced
at 300 ◦C shows significantly higher activity and selectivity to
crotyl alcohol compared with that reduced at 100 ◦C, whereas
the activity of the catalyst just calcined is minimal. Thus, it is
obvious that the presence of metallic platinum, as well as crys-
tallite size, influence the catalytic behavior of the catalyst under
discussion. A careful analysis of the HRTEM pictures of the re-
duced samples, especially those reduced at 300 ◦C, seems to
indicate the migration of partially reduced GaOx on the top of
platinum species, according to a well-known SMSI mechanism.
To rationalize such a suggestion, we carried out SIMS investi-
gations.
ToF-SIMS measurements were carried out for the catalysts
reduced at 300 ◦C. Table 5 presents intensity ratios Pt−/O− on
the surface of samples as prepared and after the removal of a
few top monolayers. In all cases, the value of Pt−/O− was lower
for subsurface regions. For catalysts α2A and α5A, this value
dropped just 5.9- and 2.3-fold, respectively, whereas for cata-
lysts β2A and β5A, 8.0- and 4.2-fold decreases were observed.
The difference between the amount of Pt present on the surface
and in the deeper layers is considerably smaller for Pt/α-Ga2O3.
This probably means that Pt supported on β-Ga2O3 is situated
mainly on top of the support surface, whereas in the second
case, Pt crystallites are modified by a thin layer of partially
reduced α-Ga2O3 and are situated slightly deeper in the sub-
surface region.
3.4. HRTEM and ToF-SIMS measurements
Figs. 12–14 show examples of representative HRTEM im-
ages of some catalysts under study, both calcined at 200 ◦C and
reduced at 100 and 300 ◦C. Based on these images, we esti-
mated the distribution of platinum particles and the main size
of platinum crystallites; the results are presented in Table 4.
Catalysts prepared from chlorine-free precursor α2A and
α5A show high, homogeneous platinum dispersion, similar in
both catalysts. An application of H2PtCl6 as an active-phase
precursor leads to a significant increase in platinum crystal-
lite size. The effect of chlorine on the nature of the catalysts
and their activity and selectivity in the reaction of selective
hydrogenation of α,β-unsaturated aldehydes is diversified and
complex [19,26–29,31,43–45]. Residual chlorine can diminish
the metal dispersion and enhance [26,27,46] or inhibit [31] for-
mation of an alloy between platinum and metal originating from
the reducible support. In the part of this paper describing cat-
alytic properties of the studied catalysts, we showed that our
Pt/Ga2O3 systems obtained from H2PtCl6 precursor are much
less active, particularly due to the much larger platinum parti-
cles compared with those prepared from the chlorine-free pre-
cursor, although their selectivity was still high. That is why we
focused our attention on such catalysts, especially α5A.
HRTEM images of the catalysts supported on α-Ga2O3 re-
veal a very interesting structure of this polymorphic modifi-
cation. We can see a characteristic, hexagonal structure [47]
of the cells. Relatively uniform platinum particles are regu-
larly distributed in the center of the support cells. Thus, such
The prevailing opinion is that in crotonaldehyde hydrogena-
tion, selectivity to the unsaturated alcohol increases with the

204
E. Gebauer-Henke et al. / Journal of Catalysis 250 (2007) 195–208
(a)
(b)
(c)
◦
◦
Fig. 12. HRTEM image and particle size histograms of α5A: (a) after calcination, (b) after reduction at 100 C, (c) after reduction at 300 C.
increase in metal particles [19]. Englisch et al. [48] concluded
that in the case of large particles, the prevalent dense Pt(111)
surface planes of Pt constrains sorption of the C=C double
bond, which enhances the selectivity to hydrogenation of the
carbonyl group. On the other hand, the abundance of highly
exposed metal atoms in small particles favors an unrestricted
sorption of both double bonds, facilitating hydrogenation of the
C=C bond. However, only combined effects of metal particles
nature and support and/or promoter interaction may improve
the selectivity to a great extent. In the case of partly reducible
supports, such as TiO2, due to the SMSI effect, the presence
of coordinatively unsaturated Ti atoms strengthens the inter-
action of catalyst with the C=O bond and thus enhances its
hydrogenation. Englisch et al. stated that both the particle size
and decoration of Pt particles by titania suboxides are addi-
tive effects, leading to increased selectivity. Other work has
investigated a promoting effect of gallium in silica-supported
platinum catalysts, improving the selectivity toward crotyl al-
cohol [22]. They suggest that gallium is present in the catalyst
in the form of oxide electron pair acceptor sites, which activate

E. Gebauer-Henke et al. / Journal of Catalysis 250 (2007) 195–208
205
(a)
(b)
Fig. 13. HRTEM image and particle size histograms of: (a) α5H, (b) β5H.
the C=O bond by coordination to the oxygen of the carbonyl
group.
bridged adsorption on two Pt atoms. The intensity and strength
of the bridging band CO is very weak and hard to see [51]. This
band disappears after a 10-min evacuation at room temperature,
whereas the linear band is removed after outgassing at 100 ◦C.
For polyfunctional molecules, such as crotonaldehyde, sev-
eral possible adsorption structures on the platinum catalysts
surface can exist [50,52]:
In our studies, we demonstrated that gallium oxide is a par-
tially reducible oxide. As a support, it behaved like TiO2 dec-
orating platinum particles with gallium suboxides. This effect
is strengthened by a well-established promoting affect of gal-
lium on the selectivity of carbonyl hydrogenation [5]. Another
important factor contributing to the very high selectivity of the
catalysts under study (especially catalyst α5A) is stabilization
of the Pt dispersion by placing Pt particles in regular hexagonal
cells of the α-Ga2O3 support.
• For adsorption through the C=C double bond, two struc-
tures are possible. These are designated πCCη2 and di-
σCCη2, corresponding to the adsorption band at 1641 and
1543, 1452 cm−1, respectively;
3.5. Infrared spectroscopy
• For adsorption through the C=O double bond, three struc-
tures are possible, ascribed as πCOη2, di-σCOη2, and the η1
“on-top” adsorption by an oxygen atom, corresponding to
The FTIR analyses were carried out for α5A and β5A cat-
alysts. The characteristic absorption bands of crotonaldehyde
and CO were the same for both catalysts. Figs. 15 and 16
present FTIR spectra for α5A catalyst. IR spectroscopy of ad-
sorbed carbon monoxide may be used to provide information
on the size and morphology of metallic particles and the extent
of interaction with the support [49,50]. Fig. 15 shows spectra
obtained for the sample reduced at 300 ◦C and exposed to CO
at 25 ◦C, which contain two broad absorption bands at 2065
and 1822 cm−1. The first is assigned to linearly adsorbed CO
on a single Pt0 atom, whereas the second characterizes the CO
respective absorption bands at 1693, 1660, and 1680 cm−1
;
• The structure where the adsorption occurs involves both
double bonds, C=C and C=O, ascribed as η4, with bands
at 1693 and 1641 cm−1
.
The experimentally obtained FTIR spectra of adsorbed cro-
tonaldehyde are presented in Fig. 16. Adsorption at room tem-
perature resulted in the formation of absorption bands at 1718,
1683, 1614, 1440, 1371, and 1162 cm−1. The intensity of the
absorption bands at 1718, 1614, and 1440 cm−1 decreased after

206
E. Gebauer-Henke et al. / Journal of Catalysis 250 (2007) 195–208
(a)
(b)
◦
◦
Fig. 14. HRTEM image and particle size histograms of α2A: (a) after reduction at 100 C, (b) after reduction at 300 C.
Table 4
Main size for platinum crystallites
Catalyst
α5A
α2A
α5H
β5A
β2A
β5H
◦
Calc. 200 C
3.4 nm
2.8 nm
3.2 nm
–
–
–
–
–
–
–
–
–
–
◦
Red. 100 C
3.3 nm
3.7 nm
◦
Red. 300 C
23.5 nm
10.9 nm
25.9 nm
Table 5
−
−
3
The values of Pt /O intensity ratios (×10 ) calculated for α2A, β2A, α5A,
◦
β5A catalysts after reduction at 300 C
α5A
α2A
β5A
β2A
(a) As prepared
(b) After removal
a few top monolayers
a/b ratios
2.5
1.1
3.6
0.61
14.0
3.3
12
1.5
2.3
5.9
4.2
8
Fig. 15. FTIR spectra of CO on α5A: (1) “fresh” sample, (2) adsorption of CO
at RT, (3) evacuation at RT, 10 min, (4) evacuation at RT, 30 min.
evacuation at room temperature. Because of the very close po-
sition of these bands to analogous ones in the crotonaldehyde
gas phase, they can be assigned to νC=O, νC=C, and δas(CH3)
in the crotonaldehyde molecule. The absorption bands at 1683,
1371, and 1162 cm−1 remained in the spectrum after evacua-
aldehyde are adsorbed on the catalyst surface by one platinum
atom. The FTIR spectra show no absorption bands responsible
for crotonaldehyde adsorption on the two Pt atoms (structures
η4, di-σCCη2, and di-σCOη2). This may suggest that platinum
atoms have a close interaction with the support. Thus, FTIR
analysis seems to confirm platinum decoration with a thin layer
tion. These bands can be attributed to νC=O, δC–H, and γCH
3
vibrations in the coordinatively bonded crotonaldehyde, respec-
tively. These results demonstrate that the molecules of croton-

E. Gebauer-Henke et al. / Journal of Catalysis 250 (2007) 195–208
207
[2] P.T. Anastas, L.B. Bartlett, M.M. Kirchhoff, T.C. Williamson, Catal. To-
day 55 (2000) 11.
[3] J. March, March’s Advanced Organic Chemistry, McGraw–Hill Ko-
gakusha, Tokyo, 1997, p. 829.
[4] V. Ponec, Appl. Catal. A 149 (1997) 27.
[5] S. Nishimura, Handbook of Heterogeneous Catalytic Hydrogenation for
Organic Synthesis, Wiley, New York, 2001.
[6] P. Gallezot, D. Richard, Catal. Rev. Sci. Eng. 40 (1&2) (1998) 81.
[7] Y. Nitta, Y. Hiramatsu, T. Imanaka, J. Catal. 126 (1990) 235.
[8] Y. Nitta, K. Vero, T. Imanaka, Appl. Catal. A 56 (1989) 9.
[9] Y. Nitta, T. Kato, T. Imanaka, Stud. Surf. Sci. Catal. 78 (1993) 83.
[10] C. Ando, H. Kurokawa, H. Miura, Appl. Catal. 185 (1999) 181.
[11] M.B. Padley, C.H. Rochester, G.J. Hutchings, J. Catal. 148 (1994) 438.
[12] J.E. Bailie, C.H. Rochester, G.J. Hutchings, J. Chem. Soc. Faraday Trans.
93 (1997) 2331.
[13] J.E. Bailie, G.J. Hutchings, H.A. Abdullah, J.A. Anderson, C.H. Ro-
chester, Phys. Chem. Chem. Phys. 2 (2002) 283.
[14] E.L. Rodrigues, J.M.C. Bueno, Appl. Catal. 232 (2002) 147.
[15] E.L. Rodrigues, A.J. Marchi, C.R. Apesteguia, J.M.C. Bueno, Stud. Surf.
Sci. Catal. 130 (2000) 2087.
[16] E.L. Rodrigues, J.M.C. Bueno, Appl. Catal. 257 (2004) 201.
[17] F. Djerboua, D. Benachour, R. Touroude, Appl. Catal. 282 (2005) 123.
[18] A. Borgna, B.G. Anderson, A.M. Saib, H. Bluhm, M. Hävecker, A. Knop-
Gericke, A.E.T. Kuiper, Y. Tamminaga, J.W. Niemantsverdriet, J. Phys.
Chem. B 108 (2004) 17905.
[19] P. Mäki-Arvela, J. Hájek, T. Salmi, D.Yu. Murzin, Appl. Catal. A
292 (1&2) (2005) 1.
[20] T.B.L.W. Marinelli, S. Nabuurs, V. Ponec, J. Catal. 151 (1995) 431.
[21] A.B. da Silva, E. Jordão, M.J. Mendes, P. Fouilloux, Appl. Catal. A 148
(1997) 253.
Fig. 16. FTIR spectra of C H O on α5A: (1) “fresh” sample, (2) adsorption of
4
6
C H O at RT, (3) evacuation at RT, 10 min, (4) evacuation at RT, 30 min, (5)
4
6
◦
evacuated at 150 C.
of a partially reduced support. In our spectra, only a band at
1683 cm−1 is observed, which can be attributed to the “on-top”
η1 position, connected with high catalyst selectivity in the hy-
drogenation of α,β-unsaturated aldehydes [50,52].
4. Conclusion
[22] M. Englisch, V.S. Ranade, J.A. Lercher, J. Mol. Catal. A 121 (1997) 69.
[23] G. Neri, L. Mercadante, C. Milone, R. Pietropaolo, S. Galvano, J. Mol.
Catal. A 108 (1996) 69.
[24] B. Coq, F. Figueras, Coord. Chem. Rev. 178–180 (1998) 1753.
[25] S.E. Collins, M.A. Baltanás, J.L. Garcia Fierro, A.L. Bonivardi, J. Catal.
221 (2002) 252.
[26] F. Ammari, J. Lamotte, R. Touroude, J. Catal. 221 (2004) 32.
[27] M. Consonni, D. Jokic, D.Yu. Murzin, R. Touroude, J. Catal. 188 (1999)
165.
[28] K. Liberková, R. Touroude, J. Mol. Catal. A 180 (2002) 221.
[29] M. Abid, G. Ehret, R. Touroude, Appl. Catal. 217 (2001) 219.
[30] F. Ammari, C. Milone, R. Touroude, J. Catal. 235 (2005) 1.
[31] M. Abid, R. Touroude, Catal. Lett. 69 (2000) 139.
[32] P. Concepcion´, A. Corma, J. Silvestre-Albero, V. Franco, J.Y. Chane-
Ching, J. Am. Chem. Soc. 126 (2004) 5523.
Platinum catalysts supported on gallium oxide appear to
be very promising in the reaction of selective hydrogenation
of crotonaldehyde to crotyl alcohol. The use of gallium oxide
as a platinum support increases the C=O bond hydrogenation
significantly while maintaining high activity, whereas Pt-based
catalysts are usually reported with high selectivity but have poor
activity. The best catalytic performance (high activity and selec-
tivity) is given by the 5 wt% Pt/Ga2O3 catalyst prepared from
the chlorine-free precursor: platinum acetylacetonate supported
on α-Ga2O3. We have shown that the use of gallium oxide as
a support for platinum increases the C=O bond hydrogena-
tion selectivity considerably while maintaining high activity.
These particular properties of Pt/α-Ga2O3 catalysts are due to
the synergetic effect of the following advantages for selective
hydrogenation of carbonyl group elements:
[33] J.L. Margitfalvi, A. Tompos, I. Kolosova, J. Valyon, J. Catal. 174 (1998)
246.
[34] C. Raab, J.A. Lercher, Catal. Lett. 18 (1993) 99.
[35] F. Coloma, A. Sepulveda-Escribano, F. Rodríguez-Reinoso, Appl. Catal.
A 123 (1995) L1.
[36] A.M. Ruppert, T. Paryjczak, Appl. Catal. A 320 (2007) 80.
[37] A.B. Merlo, G.F. Santori, J. Sambeth, G.J. Siri, M.L. Caselle, O.A. Fer-
retti, Catal. Commun. 7 (2006) 204.
[38] B. Zheng, W. Hua, Y. Yue, Z. Gao, J. Catal. 232 (2005) 143.
[39] D.E. Resasco, G.L. Haller, J. Catal. 82 (1982) 279.
[40] H. Lieske, G. Lietz, H. Spindler, J. Volter, J. Catal. 81 (1993) 8.
[41] F. Le Normand, A. Borgna, T.F. Garetto, R. Apesteguia, B. Moraweck,
J. Phys. Chem. 100 (1996) 9068.
(a) surface reducibility of gallium oxide, leading to covering
(decoration) of platinum by gallium suboxides;
(b) promoting effect of gallium;
(c) regular distribution of platinum particles on α-Ga2O3 by
specific hexagonal structure of the support.
Acknowledgments
[42] M. Saito, S. Watanabe, I. Takahara, M. Inaba, K. Murata, Catal. Lett.
89 (3–4) (2003) 213.
[43] L.-P. Tiainen, P. Mäki-Arvela, A.K. Neyestanaki, T. Salmi, D.Yu. Murzin,
React. Kinet. Lett. 78 (2003) 251.
[44] J. Hájek, N. Kumar, P. Mäki-Arvela, T. Salmi, D.Yu. Murzin, I. Paseka,
T. Heikkilä, E. Laine, P. Laukkanen, J. Väyrynen, Appl. Catal. A 251
(2003) 385.
The authors thank Dr. Waldemar Maniukiewicz and Dr.
Joanna Bojarska for XRD analysis. This work was supported
by Grant 3 T09B 11326 (0112/T09/2004/26).
References
[45] B. Bachiller-Baeza, A. Guerrero-Ruiz, I. Rodríguez-Ramos, Appl. Catal.
A 192 (2000) 289.
[1] R.L. Augustine, Heterogeneous Catalysis for the Synthetic Chemist,
Dekker, New York, 1996.
[46] J. Silvestre-Albero, F. Coloma, A. Sepúlveda-Escribano, F. Rodriguez-
Reinoso, Appl. Catal. A 304 (2006) 159.

208
E. Gebauer-Henke et al. / Journal of Catalysis 250 (2007) 195–208
[47] R. Roy, V.G. Hill, E.F. Osborn, J. Am. Chem. Soc. 74 (1952) 710.
[48] M. Englisch, A. Jentys, J.A. Lercher, J. Catal. 166 (1997) 25.
[49] F. Coloma, J.M. Coronado, C.H. Rochester, J.A. Anderson, Catal. Lett. 51
(1998) 155.
[50] F. Delbecq, P. Sautet, J. Catal. 152 (1995) 217.
[51] P. Bazin, O. Saur, J.C. Lavalley, M. Daturi, G. Blanchard, Phys. Chem.
Chem. Phys. 7 (2005) 187.
[52] A. Dandekar, M.A. Vannice, J. Catal. 183 (1999) 344.