A simple discrimination of the promoter effect in alcohol oxidation and dehydrogenation over platinum and palladium

Article abstract of DOI:10.1016/j.jcat.2004.04.002

Several models have been suggested to interpret the positive effect of Bi, Pb, and other metal promoters in the oxidative dehydrogenation of alcohols over Pt-group metal catalysts, though the experimental proof to these models is mainly missing. Here we propose a new and simple approach for clarifying the promoter effect: the comparison of dehydrogenation in inert atmosphere and oxidative dehydrogenation in the presence of oxygen over promoted and unpromoted catalysts. As test reactions the transformations of 1-phenylethanol, 2-octanol, and cinnamyl alcohol to the corresponding carbonyl compounds in a slurry reactor are used. The observed promoter effects can be classified into three groups: (i) acceleration of the hydrogen abstraction (dehydrogenation) step, (ii) acceleration of a reaction involving the transfer of (chemisorbed) oxygen, e.g., oxidation of the coproduct hydrogen or a surface impurity, thus generating free active sites, and (iii) a combination of these two different processes. This discrimination revealed that the metal promoter may have a dual effect on the aerobic oxidation of alcohols, and this behavior may rationalize some former controversial interpretations. Besides, this study presents the first evidence for the promoting effect of Bi in alcohol dehydrogenation in the absence of oxygen.

Full text of DOI:10.1016/j.jcat.2004.04.002

Journal of Catalysis 225 (2004) 138–146
www.elsevier.com/locate/jcat
A simple discrimination of the promoter effect in alcohol oxidation and
dehydrogenation over platinum and palladium
Csilla Keresszegi, Tamas Mallat, Jan-Dierk Grunwaldt, and Alfons Baiker ∗
Department of Chemistry and Applied Biosciences, Swiss Federal Institute of Technology, ETH Hönggerberg, CH-8093 Zurich, Switzerland
Received 17 November 2003; revised 5 April 2004; accepted 5 April 2004
Available online 12 May 2004
Abstract
Several models have been suggested to interpret the positive effect of Bi, Pb, and other metal promoters in the oxidative dehydrogenation
of alcohols over Pt-group metal catalysts, though the experimental proof to these models is mainly missing. Here we propose a new and
simple approach for clarifying the promoter effect: the comparison of dehydrogenation in inert atmosphere and oxidative dehydrogenation
in the presence of oxygen over promoted and unpromoted catalysts. As test reactions the transformations of 1-phenylethanol, 2-octanol, and
cinnamyl alcohol to the corresponding carbonyl compounds in a slurry reactor are used. The observed promoter effects can be classified
into three groups: (i) acceleration of the hydrogen abstraction (dehydrogenation) step, (ii) acceleration of a reaction involving the transfer
of (chemisorbed) oxygen, e.g., oxidation of the coproduct hydrogen or a surface impurity, thus generating free active sites, and (iii) a
combination of these two different processes. This discrimination revealed that the metal promoter may have a dual effect on the aerobic
oxidation of alcohols, and this behavior may rationalize some former controversial interpretations. Besides, this study presents the first
evidence for the promoting effect of Bi in alcohol dehydrogenation in the absence of oxygen.
2004 Elsevier Inc. All rights reserved.
Keywords: Pt/Al O ; Pd/Al O ; Bismuth promoter; Alcohol dehydrogenation; 1-Phenylethanol; 2-Octanol; Cinnamyl alcohol; Aerobic oxidation
2
3
2 3
1. Introduction
tion with metals, oxides, phosphates, amines, or phos-
phines [17–21]. The promoter metals, such as Bi, Pb, Te,
Sn, and others, have generated the most interest [22–26].
The promoter metals alone are inactive in alcohol oxida-
tion under the mild conditions applied; still, they induce
sometimes spectacular rate enhancement [27–29] or a shift
in the product distribution [21,30,31]. Despite the intensive
effort by several research groups, the real nature of the pro-
moter effect is still debated. Several strategies have been
applied to understand the behavior of the bi- and trimetallic
catalysts, including the detailed kinetic analysis of the re-
action [4,32,33], in situ study of the oxidation state of the
metals by electrochemical methods [34–37] and X-ray ab-
sorption spectroscopy (XAS) [38–41], and the synthesis and
catalytic test of intermetallic compounds [42,43].
Transformation of alcohols to carbonyl compounds or
carboxylic acids with molecular oxygen over platinum-
group metal catalysts has long been the topic of academic
and industrial interest [1–5]. In some reactions, such as
the oxidation of aromatic alcohols, very high reaction rates
have been reported (e.g., [6,7]), whereas in other cases seri-
ous catalyst deactivation hinders the practical application of
the method [8–10]. According to the most accepted model,
the reaction obeys a dehydrogenation mechanism and the
coproduct hydrogen reacts with adsorbed oxygen to form
water [11]. Another important role of oxygen is the ox-
idative removal of surface impurities [12] formed in side
reactions such as alcohol degradation on the Pt-group metal
surface [13–16].
Here we report a new approach to clarify this mechanistic
point. We compare the performance of various promoted and
unpromoted catalysts in the oxidation of aromatic, aliphatic,
and allylic alcohols in the presence and absence of molecular
oxygen. The bimetallic catalysts were prepared by preferen-
tial deposition of Bi onto the surface of alumina-supported
There are numerous reports on the improvement of the
performance of supported Pt and Pd catalysts by promo-
*
Corresponding author. Fax: +41 1 632 1163.
E-mail address: baiker@chem.ethz.ch (A. Baiker).
0021-9517/$ – see front matter
doi:10.1016/j.jcat.2004.04.002
2004 Elsevier Inc. All rights reserved.

C. Keresszegi et al. / Journal of Catalysis 225 (2004) 138–146
139
Pt or Pd particles [12,44]. The advantage of this approach is
that during the promotion step the original structure of sup-
ported Pt and Pd is dominantly preserved and thus a direct
comparison of promoted and unpromoted catalysts is possi-
ble.
Al2O3 support. Reduction of the catalyst was performed in
the load lock, for 15 min in a hydrogen atmosphere at room
temperature, in a similar way as described previously [45].
The surface ratio of Bi3+ to Bi0 was determined by peak fit-
ting (after correction for background and energy shifting),
using the SPECSLAB program (Specs, Berlin). In order to
limit the number of fit parameters, the tabulated energy dif-
ferences between Bi 4f7/2 and Bi 4f5/2 (5.39 eV) were
kept constant. In addition, the full width at half-maximum
(FWHM) was constrained to one value.
2. Experimental
2.1. Materials
The Bi/Pt and Bi/Pd mass ratios in the bimetallic cata-
lysts were calculated from the edge jumps (absorption step,
i.e., difference in X-ray absorption coefficients times path-
length; ꢀµ · d) of X-ray absorption near-edge structure
(XANES) spectra using pellets of 13 mm diameter by com-
parison to reference pellets with known Bi, Pd, and Pt con-
centrations.
X-ray absorption spectroscopy (XAS) experiments at the
Pd K, Pt L3, and Bi L3 edges were recorded at beamline
X1 at HASYLAB (DESY in Hamburg) in the transmis-
sion mode using Si(311) and Si(111) double-crystal mono-
chromators. The raw data were energy-calibrated with the
respective metal foil (Pd K, Pt L3 edges), background-
corrected, and normalized using the WINXAS 2.1 program
package [46]. For exact determination of the X-ray absorp-
tion step (ꢀµ · d), all spectra were treated in the same way
using a linear background subtraction before the edge and a
parabolic correction above the edge.
For transmission electron microscopy (TEM), the sam-
ples were dispersed in ethanol and deposited onto a per-
forated carbon foil supported on a copper grid. The mea-
surements were performed on a CM30 microscope (Philips;
LaB6 cathode, operated at 300 kV). For determination of the
metal particle size (surface average diameter), different areas
were examined and about 700 particles were counted. Metal
dispersion was calculated from the metal particle size [47].
1-Phenylethanol (Aldrich, 98%), 2-octanol (Fluka,
ꢀ 99.5%), toluene (J.T. Baker, > 99.5%), cyclohexane
(Aldrich, > 99%), acetic acid (Fluka, ꢀ 99.8%), acetophe-
none (Fluka, ꢀ 99%), 2-propanol (J.T. Baker, > 99.5%),
NaHCO3 (Merck, ꢀ 99.7%), Bi(NO3)3 · 5H2O (Fluka,
99%), and high purity water (Merck) were used as received.
trans-Cinnamyl alcohol (Acros, 98%) was purified by re-
1
crystallization from petroleum ether (99.3% by H NMR
and GC). Ethylene glycol diacetate (Aldrich, > 99%) was
applied as internal standard and dodecylbenzenesulfonic
acid sodium salt (Fluka, techn.) as surfactant. Cyclohexene
(Merck, > 99%) and vinyl acetate (Aldrich, 99%) were used
as H acceptor for transfer dehydrogenation reactions. Gases
were of 99.999% purity (PANGAS).
The 5 wt% Pd/Al2O3 (Johnson Matthey 324) and 5 wt%
Pt/Al2O3 (Engelhard 4759) catalysts were used as received.
The bimetallic catalysts 0.75 wt% Bi–5 wt% Pd/Al2O3,
1.0 wt% Bi–5 wt% Pt/Al2O3, and 0.9 wt% Bi–5 wt%
Pt/Al2O3 were prepared according to a former recipe [44].
At first, 2.5 g 5 wt% Pd/Al2O3 or 5 wt% Pt/Al2O3 was pre-
reduced by hydrogen (40 ml min−1) in 200 ml distilled water
at room temperature under magnetic stirring. After 20 min
the pH was set to 3 with 3 ml acetic acid and the appropriate
amount of Bi(NO3)3 · 5H2O in 2% aqueous acetic acid so-
lution (∼ 10−3 M Bi3+) was dropped into the stirred slurry
within 30 min under hydrogen atmosphere. After 5 min, 3 ml
2-propanol was added in order to keep the metals in a re-
duced state during filtration. The system was flushed with
nitrogen; the catalyst was filtered off and washed to neutral
with a 1% aqueous 2-propanol solution without contacting
it with air, then suspended with a 0.05 M NaHCO3 solution,
and washed with water to neutral. The catalyst was dried un-
der vacuum at room temperature.
2.3. Alcohol dehydrogenation
Oxidation and dehydrogenation of 1-phenylethanol,
trans-cinnamyl alcohol and 2-octanol were carried out in a
flat-bottomed, magnetically stirred 100-ml glass reactor at 1
bar. Gases (argon, nitrogen, hydrogen, oxygen, and air) were
saturated with the corresponding solvent before entering the
reactor. The air (oxygen) flow rate (20–60 ml min−1) and
stirring speed (500–1000 min−1) were set to ensure working
in the mass-transport-limited regime and avoid overoxida-
tion of the catalyst [2,4,48].
2.2. Methods for catalyst characterization: XPS, XAS, and
TEM
The X-ray photoelectron spectroscopy (XPS) measure-
ments were performed on a Leybold Heraeus LHS11 MCD
instrument using MgKα (1253.6 eV) radiation [45]. The
powdered sample was pressed into a sample holder, evac-
uated in the load lock at room temperature to 10−6 mbar
and transferred to the analysis chamber at a typical pressure
of 10−9 mbar. The peaks were energy-shifted to the binding
energy of Al 2p (74.7 eV) to correct for the charging of the
Method I
One hundred milligrams catalyst (without any pretreat-
ment), 1.0 g 1-phenylethanol or 2-octanol, and 30 ml solvent
(cyclohexane or toluene) were given into the reactor. Air was
replaced by Ar, the reactor was immersed into a preheated
oil bath (reaction temperature: 80 ◦C), and stirring started
(750 min−1).

140
C. Keresszegi et al. / Journal of Catalysis 225 (2004) 138–146
Method II
the mixture was stirred for 15 min. After 16 h separation,
One hundred milligrams catalyst was prereduced at room
a sample from the organic layer was analyzed by GC.
temperature by flowing hydrogen for 10 min in 30 ml water
containing 10 mg surfactant (dodecylbenzenesulfonic acid
sodium salt). Then the reactor was purged with nitrogen
(5 min), and the catalyst was reoxidized with air (10 min).
Air was replaced by Ar and the reactor was immersed into a
preheated oil bath (reaction temperature: 55 ◦C). The 1.0 g
1-phenylethanol or 2-octanol was injected to the reaction
mixture through a septum and stirring started (750 min−1).
2.4. Catalytic hydrogenation
The hydrogenation reactions were carried out in a paral-
lel pressure reactor system Endeavor (Argonaut Technolo-
gies) with eight mechanically stirred 15-ml stainless-steel
reactors equipped with glass liners and mechanical mixing
(750 min−1). Variation of the mixing frequency indicated
no external mass-transport limitation in the slow reactions.
Intraparticle diffusion effects were unlikely but could not
be excluded. Generally, the slurry containing 10 mg cata-
lyst (without prereduction) and 0.5 g acetophenone in 5 ml
toluene was stirred at room temperature and 10 bars for 2 h.
For hydrogenationin acidic medium, 0.25 ml acetic acid was
added to the slurry as well and the reaction was stopped after
1 h. Conversion and product composition were determined
by GC analysis as described above.
Method III
The pretreatment procedure was started according to
method II but after reoxidation of the catalysts the reaction
was started in an air flow (20 ml min−1).
Method IV
For the dehydrogenation of 1.0 g cinnamyl alcohol only
50 mg catalyst (without prereduction) was applied in 20 ml
toluene, in the presence of ethylene glycol diacetate as in-
ternal standard. Air was replaced by Ar, the reactor was
immersed into a preheated oil bath (reaction temperature:
65 ◦C), and stirring started (1000 min−1).
3. Results
3.1. Catalyst characterization
Method V
Some important features of the mono- and bimetallic
catalysts and their abbreviations are collected in Table 1.
The unpromotedPd/Al2O3 and Pt/Al2O3 catalysts possessed
medium dispersion and monomodal particle size distribution
as determined by TEM. The Bi-promoted catalysts were pre-
pared by a method that has been shown to afford dominantly
bimetallic particles [12]. The basic idea is that under ambient
One hundred milligrams catalyst (without any pretreat-
ment), 1.0 g 1-phenylethanol, hydrogen acceptor (cyclohe-
xene or vinyl acetate; H acceptor/alcohol molar ratio 2), and
30 ml cyclohexane were given into the reactor. Air was re-
placed by Ar, the reactor was immersed into a preheated oil
bath (reaction temperature: 80 ◦C), and stirring was started
(500 min−1). The reactions were stopped after 1 h.
The reaction mixtures were analyzed by GC (Thermo
Quest Trace 2000, equipped with an HP-FFAP capillary
column and an FID detector). Products were identified by
GC-MS, and by GC analysis of authentic samples. From the
reaction mixtures containing nonaqueous solvents (methods
I, IV, and V) samples were periodically withdrawn during
dehydrogenation and analyzed. In aqueous phase reactions
(methods II and III) only the final products were analyzed.
To the aqueous slurry containing the catalyst, 3 ml saturated
sodium chloride solution and 20 ml toluene were added and
conditions Bi3+ cannot be reduced to Bi0 by hydrogen on the
surface of the support but it is readily reduced on the surface
of the Pt or Pd particles to form metal adatoms [26]. Thus,
a well-prepared and reduced catalyst contains only small
amounts of Bi3+ on the support, originating from the incom-
plete removal (washing) from the high surface area support
at the end of catalyst preparation.
Fig. 1 shows the X-ray photoelectron spectra of Bi in the
0.75 Bi–Pd/Al2O3 and 1.0 Bi–Pt/Al2O3 catalysts after re-
duction by hydrogen at room temperature. In both cases the
Table 1
Characteristics of the mono- and bimetallic catalysts
a
b
c
Catalyst
Abbreviation
Pd/Al O
Pt/Al O
2 3
Origin
Metal particle size (nm)
Metal dispersion
5 wt% Pd/Al O
5 wt% Pt/Al O
Johnson Matthey 324
Engelhard 4759
3.4
2.8
0.34
0.40
2
3
2 3
2
3
0
3+
d
Surface Bi /Bi ratio
e
0.75 wt% Bi–5 wt% Pd/Al O
0.75 Bi–Pd/Al O
2 3
Synthesized
73/27
63/37
35/65
2
3
e
1.0 wt% Bi–5 wt% Pt/Al O
1.0 Bi–Pt/Al O
Synthesized
2
3
3
2
3
3
e
0.9 wt% Bi–5 wt% Pt/Al O
2
0.9 Bi–Pt/Al O
Synthesized
2
a
Bi contents were calculated from the edge jumps of XANES spectra.
Determined by TEM.
Calculated from the metal particle size.
Determined by XPS analysis.
b
c
d
e
By Bi promotion (see Experimental section).

C. Keresszegi et al. / Journal of Catalysis 225 (2004) 138–146
141
Table 3
Hydrogenation of acetophenone (2, Scheme 1) to 1-phenylethanol (1); for
conditions see Experimental section
Catalyst
Conversion (%)
Toluene
Toluene + acetic acid
a
b
Pd/Al O
0.75 Bi–Pd/Al O
7.6
24
2
3
a
b
3.2
17
32
2
3
c
Pt/Al O
–
–
2
3
b
1.0 Bi–Pt/Al O
2
14
3
a
No by-product could be detected by GC.
By-product: α-methylbenzyl acetate.
By-products: α-methylbenzyl acetate, 4 and 5.
b
c
good agreement with the proposed structure of the catalysts:
a considerable fraction of surface Pd and Pt atoms is cov-
ered by Bi, and upon exposure to air these Bi atoms prevent
the oxidation of Pd and Pt atoms below them. The white line
at the Pt L3 edge decreased significantly from Pt/Al2O3 to-
ward 0.9 Bi–Pt/Al2O3 and 1.0 Bi–Pt/Al2O3, the latter thus
containing mainly metallic platinum. Also in the case of 0.75
Bi–Pd/Al2O3, Pd was dominantly in the metallic state as de-
scribed previously [40,41].
Additionally, the presence of Bi on the surface of Pd and
Pt particles was evidenced by the catalytic hydrogenation
of acetophenone (Table 3). Though the product distribution
varied with the catalyst and solvent compositions, the con-
version was always lower on the Bi-promoted catalysts (by
19–58%) relative to the monometallic reference catalysts. As
Bi is inactive in the reaction, the drop in conversion confirms
the partial coverage of surface Pt and Pd sites by Bi.
To sum up, characterization of the bimetallic catalysts
supports the expectation that Bi has been deposited domi-
nantly on the surface of alumina-supported Pt and Pd parti-
cles. This fraction of Bi is expected to influence the rate and
selectivity of alcohol dehydrogenation.
Fig. 1. X-ray photoelectron spectra of Bi in 0.75 Bi–Pd/Al O (upper
2
3
spectrum) and 1.0 Bi–Pt/Al O (lower spectrum with larger shoulder at
2
3
165.9 eV) catalysts after reduction by hydrogen under ambient conditions.
Table 2
XPS analysis of Bi 4f -peaks in the bimetallic catalysts after reduction by
hydrogen under ambient conditions
0
3+
Bi
Bi
Bi 4f
Bi 4f
Bi 4f
Bi 4f
5/2
7/2
5/2
7/2
0.75 Bi–Pd/ Binding energy (eV) 157.2
162.6
35.2
2.95
160.4
15.2
2.95
165.8
11.6
2.95
Al O
Extent (%)
38.0
2.95
2
3
FWHM (eV)
1.0 Bi–Pt/ Binding energy (eV) 157.4
162.7
30.0
2.7
160.6
19.2
2.7
165.9
18.2
2.7
Al O
Extent (%)
32.6
2.7
2
3
FWHM (eV)
0.9 Bi–Pt/ Binding energy (eV) 156.4
161.8
15.8
1.96
158.8
37.8
1.96
164.4
27.0
1.96
Al O
Extent (%)
19.4
1.96
2
3
FWHM (eV)
majority of Bi promoter was in a metallic state at room tem-
perature (Table 1): 63% Bi0 in 1.0 Bi–Pt/Al2O3 and 73%
Bi0 in 0.75 Bi–Pd/Al2O3. As discussed above, the reducible
fraction of bismuth is located on the Pt or Pd particles. The
fraction of unreducible bismuth corresponds to Bi3+ species
present on alumina; this fraction is assumed to be ineffective
in the catalytic dehydrogenation of alcohols. More details of
the analysis of Bi 4f signals in the bimetallic catalysts can
be found in Table 2.
The surface composition of 0.75 Bi–Pd/Al2O3 was de-
termined by XPS as well. Compared to the unpromoted
Pd/Al2O3 the Pd concentration on the surface is lowered
from 0.9 to 0.6%. In the case of 0.9 Bi–Pt/Al2O3 catalyst
the extent of reducible Bi is smaller (35%, Tables 1 and 2);
a considerable part of Bi is deposited on the alumina support.
The total Bi content of the bimetallic catalysts was es-
timated from the edge jumps in the XANES spectra (Ta-
ble 1). Furthermore, XAS measurements revealed that in
the samples exposed to air the Bi constituent was always
in an oxidized state (Bi3+), while the noble metal compo-
nents were mainly in a reduced state. These results are in
3.2. Alcohol dehydrogenation: preliminary screening
The dehydrogenation of 1-phenylethanol, 2-octanol, and
cinnamyl alcohol has been selected as test reactions. These
reactants were commonly used in former studies of the aero-
bic oxidation [1–3] and transfer dehydrogenation [49–51] of
alcohols over platinum-group metal catalysts to represent the
transformation of aromatic, aliphatic, and allylic alcohols to
the corresponding carbonyl compounds.
The product distribution observed in the dehydrogenation
of 1-phenylethanol (1) to acetophenone (2) in Ar is shown
in Scheme 1. Three by-products could be detected on Pt-
based catalysts, which formed by hydrogenolysis of the C–O
bond (3) or saturation of the aromatic ring (4, 5). The latter
reactions were absent on Pd-based catalysts that were also
more active than the Pt-based catalysts.
The hydrogenation and hydrogenolysis-type side reac-
tions were even more extensive in the dehydrogenation of
cinnamyl alcohol (6, Scheme 2). Under identical conditions
the reactivity of cinnamyl alcohol was lower than that of 1-

142
C. Keresszegi et al. / Journal of Catalysis 225 (2004) 138–146
Fig. 2. Dehydrogenation of 1-phenylethanol (1) in Ar or in air over
0.75 Bi–Pd/Al O , 1.0 Bi–Pt/Al O , Pd/Al O , and Pt/Al O (aqueous
2
3
2
3
2
3
2 3
medium, method II in Ar and method III in air; 3 h).
achieved under the best conditions, in agreement with for-
mer reports [8,44,53]. Note that the by-products formed by
hydrogenation and hydrogenolysis even in the presence of
oxygen indicate H coverage of the active sites and thus
confirm the dehydrogenation mechanism of alcohol oxida-
tion [52].
Scheme 1. Dehydrogenation of 1-phenylethanol (1) to acetophenone (2) and
the by-products detected in Ar. Selectivities (in %) achieved by Pd/Al O
2
3
(first value) and Pt/Al O (second value) are indicated below the formulas
2
3
(method I, cyclohexane, 3 h).
In the dehydrogenation of 2-octanol to 2-octanone no by-
products could be detected by GC analysis, independent of
the catalyst composition but the reaction was very slow on
all catalysts tested. This observation is in agreement with
the generally poor activity of Pt-group metal catalysts in the
dehydrogenation and oxidative dehydrogenation of aliphatic
alcohols [54–57]. A specific difficulty of the dehydrogena-
tion of 2-octanol was that the results were irreproducible
with the Pt/Al2O3 catalyst in an apolar organic medium. Pre-
sumably, the coproduct water remained on the hydrophilic
catalyst surface and after a short initial period the catalyst
powder adhered to the reactor wall and the reaction stopped.
This difficulty could be eliminated by working in a water–
surfactant system.
The reaction conditions (solvent, temperature, catalyst
pretreatment) of the three test reactions were varied in a
broad range to obtain a general overview on the influence of
Bi promotion of Pt and Pd and the role of oxygen in the pro-
moter effect. In the following sections some representative
examples will be shown that help understanding the influ-
ence of oxygen and bismuth promotion.
3.3. Dehydrogenation of 1-phenylethanol
Scheme 2. Reaction network in the dehydrogenation of cinnamyl alcohol
(6) over Pt- and Pd-based catalysts in Ar.
Dehydrogenation of 1-phenylethanol to acetophenone is
a good example on the promoting effect of Bi. In this re-
action Bi promotion increased the rate and selectivity of
both Pt/Al2O3 and Pd/Al2O3 catalysts, independent of the
presence or absence of oxygen, and the reaction conditions.
A comparative example is shown in Fig. 2, where the re-
actions were carried out under mild conditions, in a water–
surfactant system. Interestingly, the 1.0 Bi–Pt/Al2O3 cata-
phenylethanol; the difference is (partly) attributed to decar-
bonylation of cinnamaldehyde (7) and dihydrocinnamalde-
hyde (11) [52]. The extent of these side reactions increased
at higher temperature and in apolar solvents (e.g., cyclo-
hexane). In contrast, the presence of oxygen suppressed
by-product formation in the dehydrogenation of both alco-
hols 1 and 6, and over 99% selectivity to 2 or 7 could be

C. Keresszegi et al. / Journal of Catalysis 225 (2004) 138–146
143
Fig. 4. Dehydrogenation of 2-octanol in Ar or in air over 0.75 Bi–Pd/Al O ,
2
3
1.0 Bi–Pt/Al O and unpromoted Pd/Al O and Pt/Al O (aqueous
medium, method II in Ar and method III in air; 3 h). The selectivity to
2-octanone was always 100%.
2
3
2
3
2 3
Fig. 3. Dehydrogenation of 1-phenylethanol (1) over Pt/Al O and 0.9
2
3
Bi–Pt/Al O . The argon atmosphere was replaced by air (flow rate:
2
3
−1
50 ml min ) after 65 min (method I, toluene).
lyst had good selectivity even in Ar where the hydrogen co-
product was available in high concentrations for the hydro-
genation and hydrogenolysis-typeside reactions (Scheme 1).
An important point is that the rate enhancement achieved
in air by Bi promotion should be considered only as a qual-
itative measure of the reaction rate. In the aerobic oxidation
of alcohols over Pt-group metal catalysts the reactions are
commonly carried out under mass-transport-limited condi-
tions to avoid the rapid “overoxidation” and deactivation of
the catalyst [2,4,48,58]. A further difficulty of a compara-
tive study is that applying the same conditions, including
the same rate of oxygen supply for all reactions, this rate
may be too low for an active catalyst and it performs well
below its intrinsic activity while it may be too high for a
poorly active catalyst leading to its overoxidation and deacti-
vation. This effect is illustrated by the transient experiments
in Fig. 3. Dehydrogenation on Pt/Al2O3 in Ar afforded only
6.4% conversion in 65 min. After replacing Ar by air the cat-
alyst completely lost its activity and the final conversion was
still below 7%. In contrast, Bi promotion more than doubled
the rate of 1-phenylethanol conversion in Ar (13.5% conver-
sion in 65 min), and after switching to air the reaction was
complete within less than 50 min.
Fig. 5. Oxidation of cinnamyl alcohol (6, Scheme 2) over Pd/Al O and
2
3
0.75 Bi–Pd/Al O (toluene, method IV). The Ar atmosphere was replaced
2
3
−1
by air (flow rate: 60 ml min ) after 55 min. Details of the product distrib-
ution are shown in Table 4.
3.5. Dehydrogenation of cinnamyl alcohol
Dehydrogenation of cinnamyl alcohol (6, Scheme 2) un-
der Ar represents a case where no rate acceleration could
be achieved by Bi promotion. Dehydrogenation of allylic
alcohols on Pt and Pd is relatively fast in the presence
of a hydrogen acceptor, such as oxygen [52,59] or an
olefin [51,54,60]. Here, under Ar (i.e., in the absence of
any hydrogen acceptor), the reaction on Pd/Al2O3 slowed
down after about 10 min, indicating some catalyst deactiva-
tion (Fig. 5). Replacing Ar by air accelerated the dehydro-
genation enormously and over 99% conversion was reached
within 18 min. The likely reasons for the catalyst deactiva-
tion in Ar and reactivation in air are decarbonylation-type
side reactions (Scheme 2 and Table 4) and the rapid removal
of CO from the Pd surface, respectively [52]. Partial cover-
age of the Pd surface by Bi decreased the rate in Ar only to a
minor extent but in air the negative effect was considerable.
3.4. Dehydrogenation of 2-octanol
Bi promotion enhanced the reactivity of both Pd- and
Pt-based catalysts in the presence and absence of oxygen,
as illustrated in Fig. 4 by reactions carried out in a water–
surfactant system. Note that relatively mild conditions had
to be chosen to avoid full conversion with the best catalyst,
1.0 Bi–Pt/Al2O3, but under these conditions the unpromoted
catalysts afforded only 1% conversion in Ar. The rate ac-
celeration achieved by Bi promotion was more pronounced
with Pt/Al2O3 than with Pd/Al2O3 both in Ar and in air.

144
C. Keresszegi et al. / Journal of Catalysis 225 (2004) 138–146
Table 4
◦
Transformation of cinnamyl alcohol in toluene at 65 C (method IV). For identification of the products 7–13 see Scheme 2
Entry
Catalyst
Atmosphere
Time
(min)
Conv.
(%)
Yield (%)
Selectivity (%)
7
8
9
10
11
12
13
7
8
1
2
3
4
Pd/Al O
Argon
Air
Argon
Air
Argon
Air
Argon
Air
55
+18
55
8.3
99.1
6.4
50.3
0.6
1.5
0.5
37.7
4.5
64.3
3.7
35.8
0.6
1.4
3.4
32.6
1.3
10.6
< 0.1
0.1
0
0.2
0.3
1.4
2.6
0
0
0
0
0
0.7
< 0.1
0.3
0
0
0
0
< 0.1
1.1
0
0.4
0
0
< 0.1
< 0.1
< 0.1
0.4
0
0
55
65
58
71
99
96
87
99
40
33
20
21
1
4
13
1
2
3
< 0.1
< 0.1
0.1
0
0
0.75 Bi–Pd/Al O
2
3
+60
a
a
a
a
Pt/Al O
55
2
3
+130
55
+60
1.0 Bi–Pt/Al O
0.4
37.3
0.1
0.4
0
0
0
0
0
2
3
< 0.1
a
Due to the very low conversion (< 1%) determination of selectivity is uncertain.
Table 5
Effect of olefins as hydrogen acceptors on the dehydrogenation of 1-phe-
nylethanol (1) over Pd/Al O and Pt/Al O (method V)
2
3
2
3
H acceptor
Yield (selectivity) to 2 (%)
Pd/Al O
2
Pt/Al O
2
3
3
No
33 (94)
72 (100)
41 (100)
15 (81)
3 (100)
0 (–)
Cyclohexene
Vinyl acetate
a
a
No yield even after 5 h.
the solution. Introduction of air (after 55 min, Figs. 5 and 6)
can oxidize CO to CO2 that desorbs from the surface but the
oxidative removal of hydrocarbon fragments or dimerized
and oligomerized species is not expected. This could be the
reason for the poor performance of unpromoted Pt/Al2O3
(1.5% final conversion, Table 4, entry 3; this point is not
shown in Fig. 6). Bi promotion had a dramatic effect on
the activity but only in the presence of molecular oxygen.
Probably, Bi contributes to the oxidative removal of surface
impurities or even diminishes their formation due to an en-
semble effect (i.e., larger active site ensembles are required
for the side reactions).
The assumption that Pt is more sensitive to poisoning
by hydrocarbon residues than Pd is supported by additional
experiments. For these experiments the dehydrogenation of
1-phenylethanol in cyclohexane under Ar was chosen as
this reaction was apparently not affected by catalyst poi-
soning (Fig. 2). The dehydrogenation reaction in Ar was
repeated in the presence of cyclohexene and vinyl acetate
as organic hydrogen acceptors (Table 5). Contrary to the ex-
pected higher dehydrogenation rate [50,51,54], these olefins
poisoned Pt/Al2O3. The reference reactions with Pd/Al2O3
showed the expected behavior: higher rate and selectivity
in transfer dehydrogenation with hydrogen acceptors, com-
pared to simple dehydrogenation in Ar.
Fig. 6. Dehydrogenation of cinnamyl alcohol (6) over Pt/Al O and 1.0
2
3
Bi–Pt/Al O (toluene, method IV). The Ar atmosphere was replaced by air
2
3
−1
(flow rate: 60 ml min ) after 55 min. Details of the product distribution
are shown in Table 4.
The diminished reactivity is attributed to the lower number
of surface Pd sites available for hydrogen adsorption and
alcohol dehydrogenation in the Bi-promoted catalyst, as in-
dicated by a drop of the hydrogenation activity (Table 3) and
by XPS analysis (Fig. 1, Table 2). Promotion of Pd/Al2O3 by
Bi changed the product distribution and resulted in a small
enhancement of selectivity to the key product cinnamalde-
hyde (7).
Table 4 and Fig. 6 show that in comparison to Pd/Al2O3,
Pt/Al2O3 is much more sensitive to the destructive adsorp-
tion of reactant and products: less than 1% conversion was
achieved in 55 min in Ar and the reaction was barely accel-
erated by introducing air. Bi promotion of Pt/Al2O3 had a
positive effect only after replacement of Ar by air, leading to
higher rate and excellent selectivity. The good performance
of Bi-promoted Pt in the aerobic oxidation of cinnamyl al-
cohol has already been reported [44,59].
We attribute the poor performance of Pt/Al2O3 to poison-
ing by CxHy-type fragments formed in the decarbonylation
type side reactions (Scheme 2) [16]. It is very likely that
only a fraction of these species desorbs from the surface as
olefins. For example, propenylbenzene (10) and styrene (13)
could be detected in significant amounts by GC analysis of
4. Discussion
Numerous mechanistic models have been suggested for
the interpretation of the promoting effect of metal additives
(Bi, Pb, Sn, Te, etc.) on the oxidation of alcohols and poly-

C. Keresszegi et al. / Journal of Catalysis 225 (2004) 138–146
145
ols to the correspondingcarbonyl compounds and carboxylic
5. Conclusions
acids. Adopting the dehydrogenation mechanism of alcohol
oxidation [2,3,11], two important groups of these models
emerge.
(i) The promoter may directly influence the rate and
(regio)selectivity of the alcohol dehydrogenation reaction.
A plausible example is the oxidative dehydrogenation of
polyfunctional alcohols in 2-position, where the high regios-
electivity has been attributed to complex formation between
a neighboring Pt-group metal–promoter metal bimetallic site
and the reactant [1,17,61].
(ii) The second group comprises those models which as-
sume that the promoter does not influence directly the de-
hydrogenation step but affects the adsorption and transfer of
oxygen, including the oxidation of the coproduct hydrogen,
the oxidative removal of surface impurities [21,40,62,63],
and the improved resistance of the Pt-group metal against
“overoxidation” [64] (i.e., the coverage of surface sites by
oxygen leading to a slow down of alcohol oxidation [48,65]).
The promoter effect in the dehydrogenation of 1-phenyl-
ethanol on 0.75 Bi–Pd/Al2O3 (Fig. 2) can be interpreted
by a model belonging to the first group. Addition of Bi to
Pd/Al2O3 enhanced the rate of reaction by a factor of 1.3,
independent of the presence or absence of oxygen. Clearly,
Bi accelerated the dehydrogenation step and its influence
cannot be connected to the presence of oxygen. Dehydro-
genation of 2-octanol (Fig. 4) provides a further example to
this case. Interestingly, in this reaction the rate acceleration
by Bi promotion was even higher in Ar (a factor of 3 and
6) than in air (1.3 and 2.4) for Pd/Al2O3 and Pt/Al2O3, re-
spectively. The likely reason is the very low activity of the
catalysts in Ar that increases the relative error of the deter-
mination of conversions.
Interpretation of the frequently observed promoter effect
in the aerobic oxidation of alcohols is a demanding task. Un-
ambiguous experimental evidence to support a model is rare.
A detailed kinetic analysis is troublesome and not always
conclusive. In situ techniques are not widely accessible. We
propose a simple test to clarify the role of promoter: the
comparison of alcohol dehydrogenation in the presence and
absence of molecular oxygen. In case there is no promoter
effect in inert atmosphere, the role of promoter cannot be
attributed to improvement in the dehydrogenation step. On
the other hand, a strong rate enhancement or a shift in the
product distribution in the absence of oxygen is an evidence
of a direct role of promoter in alcohol dehydrogenation and
an indication that the role of promoter is not limited to, for
example, a better oxygen transfer or improved oxygen toler-
ance of the noble metal component.
Acknowledgments
We gratefully acknowledge HASYLAB (DESY, Ham-
burg) for offering beamtime, Dr. F. Krumeich for TEM mea-
surements, and ETH Zurich for the financial support.
References
[1] H. van Bekkum, in: F.W. Lichtenthaler (Ed.), Carbohydrates as Or-
ganic Raw Materials, VCH, Weinheim, 1991, p. 289.
[2] T. Mallat, A. Baiker, Catal. Today 19 (1994) 247.
[3] M. Besson, P. Gallezot, Catal. Today 57 (2000) 127.
[4] J.H.J. Kluytmans, A.P. Markusse, B.F.M. Kuster, G.B. Marin, J.C.
Schouten, Catal. Today 57 (2000) 143.
Dehydrogenationof cinnamyl alcohol on 1.0 Bi–Pt/Al2O3
is a good example of the second group of models: the pro-
moter effect is clearly connected to the presence of oxygen
and Bi has no influence on the alcohol dehydrogenation in
Ar (Fig. 6). On the basis of former studies [52] it is proba-
ble that Pt/Al2O3 and Bi–Pt/Al2O3 are poisoned by alcohol
degradation products in Ar, and Bi deposited on the surface
of Pt particles accelerates the oxidative removal of surface
impurities when switching from Ar to air.
Dehydrogenation of 1-phenylethanol on Pt-based cata-
lysts reveals a complex effect of Bi. Promotion of Pt/Al2O3
enhanced the reaction rate in Ar by a factor of 1.4 and in air
by a factor of 2.8 (Fig. 2). When using another Bi-promoted
catalyst under different conditions (Fig. 3), the relative rate
acceleration in air was even bigger, compared to the rate ac-
celeration achieved by Bi promotion in Ar. We can conclude
that when using these catalysts the role of Bi is not limited to
the acceleration of the dehydrogenation of 1-phenylethanol
but influences also an additional process that is connected
to the transfer of oxygen. This effect may be the enhanced
rate of oxidation of the coproduct hydrogen or the oxidative
removal of a strongly adsorbed surface impurity.
[5] J. Muzart, Tetrahedron 59 (2003) 5789.
[6] H. Lefranc, US patent 5,689,009 (1997), to Rhodia Chimie.
[7] K. Bauer, R. Moelleken, H. Fiege, K. Wedemeyer, German patent
2,620,254 A1 (1977), to Bayer AG.
[8] T. Mallat, Z. Bodnar, A. Baiker, in: S.T. Oyama, J.W. Hightower
(Eds.), Catalytic Selective Oxidation, Am. Chem. Society, Washing-
ton, DC, 1993, p. 308.
[9] J.C. Beziat, M. Besson, P. Gallezot, Appl. Catal. A 135 (1996) L7.
[10] A.P. Markusse, B.F.M. Kuster, J.C. Schouten, J. Mol. Catal. A 158
(2000) 215.
[11] K. Heyns, H. Paulsen, Adv. Carbohydr. Chem. 17 (1962) 169.
[12] T. Mallat, Z. Bodnar, A. Baiker, O. Greis, H. Strübig, A. Reller, J. Ca-
tal. 142 (1993) 237.
[13] J.L. Davis, M.A. Barteau, Surf. Sci. 197 (1988) 123.
[14] J.L. Davis, M.A. Barteau, J. Am. Chem. Soc. 111 (1989) 1782.
[15] R. Shekhar, M.A. Barteau, R.V. Plank, J.M. Vohs, J. Phys. Chem.
B 101 (1997) 7939.
[16] B.J. Wood, H. Niki, H. Wise, J. Catal. 26 (1972) 465.
[17] P.C.C. Smits, B.F.M. Kuster, K. van der Wiele, H.S. van der Baan,
Appl. Catal. 33 (1987) 83.
[18] T. Mallat, A. Baiker, Appl. Catal. A 79 (1991) 41.
[19] H. Kimura, A. Kimura, I. Kokubo, T. Wakisaka, Y. Mitsuda, Appl.
Catal. A 95 (1993) 143.
[20] M. Hronec, Z. Cvengrosova, J. Kizlink, J. Mol. Catal. 83 (1993) 75.
[21] H.H.C.M. Pinxt, B.F.M. Kuster, G.B. Marin, Appl. Catal. A 191
(2000) 45.

146
C. Keresszegi et al. / Journal of Catalysis 225 (2004) 138–146
[22] S. Szabo, I. Bakos, Ach-Models Chem. 133 (1996) 83.
[23] U.W. Hamm, D. Kramer, R.S. Zhai, D.M. Kolb, Electrochim. Acta 43
(1998) 2969.
[24] J. Clavilier, J.M. Feliu, A. Aldaz, J. Electroanal. Chem. 243 (1988)
419.
[25] E. Herrero, L.J. Buller, H.D. Abruna, Chem. Rev. 101 (2001) 1897.
[26] S. Szabo, Int. Rev. Phys. Chem. 10 (1991) 207.
[27] H. Lefranc, European patent 0,667,331 B1 (1999), to Rhodia Chimie.
[28] T. Mallat, Z. Bodnar, A. Baiker, Stud. Surf. Sci. Catal. 78 (1993)
377.
[29] F. Alardin, B. Delmon, P. Ruiz, M. Devillers, Catal. Today 61 (2000)
255.
[30] P.C.C. Smits, B.F.M. Kuster, K. van der Wiele, H.S. van der Baan,
Carbohydr. Res. 153 (1986) 227.
[31] M. Akada, S. Nakano, T. Sugiyama, K. Ichitoh, H. Nakao, M. Akita,
Y. Moro-oka, Bull. Chem. Soc. Jpn. 66 (1993) 1511.
[32] Y. Schuurman, B.F.M. Kuster, K. van der Wiele, G.B. Marin, Appl.
Catal. A 89 (1992) 31.
[33] L. Jelemensky, B.F.M. Kuster, G.B. Marin, Chem. Eng. Sci. 51 (1996)
1767.
[43] S. Karski, I. Witonska, J. Mol. Catal. A 191 (2003) 87.
[44] T. Mallat, Z. Bodnar, P. Hug, A. Baiker, J. Catal. 153 (1995) 131.
[45] J.D. Grunwaldt, M.D. Wildberger, T. Mallat, A. Baiker, J. Catal. 177
(1998) 53.
[46] T. Ressler, J. Synchrotron Rad. 5 (1998) 118.
[47] A. Borodzinski, M. Bonarowska, Langmuir 13 (1997) 5613.
[48] P.J.M. Dijkgraaf, M.J.M. Rijk, J. Meuldijk, K. van der Wiele, J. Ca-
tal. 112 (1988) 329.
[49] D. Kramer, in: E. Müller (Ed.), Methoden der Organischen Chemie
(Houben–Weil), Thieme, Stuttgart, 1973, p. 699.
[50] M. Hayashi, K. Yamada, S. Nakayama, Synthesis 11 (1999) 1869.
[51] M. Hayashi, K. Yamada, S. Nakayama, J. Chem. Soc., Perkin Trans. 1
(2000) 1501.
[52] C. Keresszegi, T. Bürgi, T. Mallat, A. Baiker, J. Catal. 211 (2002) 244.
[53] T. Mallat, Z. Bodnar, C. Brönnimann, A. Baiker, Stud. Surf. Sci. Ca-
tal. 88 (1994) 385.
[54] C. Keresszegi, T. Mallat, A. Baiker, New J. Chem. 25 (2001) 1163.
[55] R. Anderson, K. Griffin, P. Johnston, P.L. Alsters, Adv. Synth. Ca-
tal. 345 (2003) 517.
[56] K. Mori, K. Yamaguchi, T. Hara, T. Mizugaki, K. Ebitani, K. Kaneda,
J. Am. Chem. Soc. 124 (2002) 11572.
[57] K. Yamaguchi, N. Mizuno, Angew. Chem. Int. Ed. 41 (2002) 4538.
[58] H.E. van Dam, A.P.G. Kieboom, H. van Bekkum, Appl. Catal. 33
(1987) 361.
[34] R. DiCosimo, G.M. Whitesides, J. Phys. Chem. 93 (1989) 768.
[35] T. Mallat, A. Baiker, Catal. Today 24 (1995) 143.
[36] A.P. Markusse, B.F.M. Kuster, J.C. Schouten, Catal. Today 66 (2001)
191.
[37] V.R. Gangwal, B.G.M. van Wachem, B.F.M. Kuster, J.C. Schouten,
Chem. Eng. Sci. 57 (2002) 5051.
[59] K. Ebitani, Y. Fujie, K. Kaneda, Langmuir 15 (1999) 3557.
[60] M. Hayashi, K. Yamada, S. Nakayama, H. Hayashi, S. Yamazaki,
Green Chem. 2 (2000) 257.
[38] A.P. Markusse, B.F.M. Kuster, D.C. Koningsberger, G.B. Marin, Catal.
Lett. 55 (1998) 141.
[61] P. Fordham, M. Besson, P. Gallezot, Catal. Lett. 46 (1997) 195.
[62] T. Mallat, A. Baiker, J. Patscheider, Appl. Catal. A 79 (1991) 59.
[63] H. Kimura, K. Tsuto, T. Wakisaka, Y. Kazumi, Y. Inaya, Appl. Catal.
A 96 (1993) 217.
[39] H.H.C.M. Pinxt, B.F.M. Kuster, D.C. Koningsberger, G.B. Marin,
Catal. Today 39 (1998) 351.
[40] C. Keresszegi, J.-D. Grunwaldt, T. Mallat, A. Baiker, Chem. Commun.
(2003) 2304.
[64] M. Besson, F. Lahmer, P. Gallezot, P. Fuertes, G. Fleche, J. Catal. 152
(1995) 116.
[41] C. Keresszegi, J.-D. Grunwaldt, T. Mallat, A. Baiker, J. Catal. 222
(2004) 268.
[65] P.J.M. Dijkgraaf, H.A.M. Duisters, B.F.M. Kuster, K. van der Wiele,
J. Catal. 112 (1988) 337.
[42] M. Wenkin, C. Renard, P. Ruiz, B. Delmon, M. Devillers, Stud. Surf.
Sci. Catal. 108 (1997) 391.