Influence of support acid-base properties on the platinum-catalyzed enantioselective hydrogenation of activated ketones

Article abstract of DOI:10.1016/j.jcat.2010.02.012

The influence of support acidity and basicity was investigated in the enantioselective hydrogenation of methyl benzoylformate and ketopantolactone on cinchonidine-modified Pt/Al₂O₃-SiO₂ and Pt/Al₂O₃-Cs₂O catalysts. Two series of flame-derived 4.7 wt.% Pt/Al₂O₃ catalysts in which the acid-base properties of the support were systematically varied by introducing SiO₂ (5-80 wt.%) or Cs₂O (0.25-10 wt.%) were applied. Addition of SiO₂ improved the enantioselectivity with a maximum at 30 wt.%. Enantioselectivity correlated well with the acidity of the catalysts characterized by TPD of NH₃ and with the selectivity to hydrogenolysis of methyl cyclohexyl ketone to ethylcyclohexane. On the contrary, doping with Cs was detrimental to the formation of the (R)-alcohols and the drop in enantioselectivity could unambiguously be attributed to the basicity of the support characterized by TPD of CO₂. The critical impact of support ionicity on the electronic properties of Pt, and thus on the adsorption and interaction of the reaction components on the metal surface, was further proven by the good correlation between the enantioselectivities of all catalysts and the ratio of CO adsorbed in bridged to linear (B/L) geometry. The practical importance of our findings is demonstrated by the best ee (94 ± 0.5%) achieved so far in the industrially relevant hydrogenation of ketopantolactone to (R)-pantolactone.

Full text of DOI:10.1016/j.jcat.2010.02.012

Journal of Catalysis 271 (2010) 115–124
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Influence of support acid–base properties on the platinum-catalyzed
enantioselective hydrogenation of activated ketones
Fatos Hoxha a, Bjoern Schimmoeller b, Zdenek Cakl a, Atsushi Urakawa a,
Tamas Mallat a, Sotiris E. Pratsinis b, Alfons Baiker a,
*
a Institute for Chemical and Bioengineering, Department of Chemistry and Applied Biosciences, ETH Zurich, Hönggerberg, HCI, CH-8093 Zurich, Switzerland
b Particle Technology Laboratory, Department of Mechanical and Process Engineering, ETH Zurich, CH-8092 Zurich, Switzerland
a r t i c l e i n f o
a b s t r a c t
Article history:
The influence of support acidity and basicity was investigated in the enantioselective hydrogenation of
methyl benzoylformate and ketopantolactone on cinchonidine–modified Pt/Al2O3–SiO2 and Pt/Al2O3–
Cs2O catalysts. Two series of flame-derived 4.7 wt.% Pt/Al2O3 catalysts in which the acid–base properties
of the support were systematically varied by introducing SiO2 (5–80 wt.%) or Cs2O (0.25–10 wt.%) were
applied. Addition of SiO2 improved the enantioselectivity with a maximum at 30 wt.%. Enantioselectivity
correlated well with the acidity of the catalysts characterized by TPD of NH3 and with the selectivity to
hydrogenolysis of methyl cyclohexyl ketone to ethylcyclohexane. On the contrary, doping with Cs was
detrimental to the formation of the (R)-alcohols and the drop in enantioselectivity could unambiguously
be attributed to the basicity of the support characterized by TPD of CO2. The critical impact of support
ionicity on the electronic properties of Pt, and thus on the adsorption and interaction of the reaction com-
ponents on the metal surface, was further proven by the good correlation between the enantioselectiv-
ities of all catalysts and the ratio of CO adsorbed in bridged to linear (B/L) geometry. The practical
importance of our findings is demonstrated by the best ee (94 0.5%) achieved so far in the industrially
relevant hydrogenation of ketopantolactone to (R)-pantolactone.
Received 8 December 2009
Revised 6 February 2010
Accepted 9 February 2010
Available online 15 March 2010
Keywords:
Platinum
Ketopantolactone
Methyl benzoylformate
Asymmetric hydrogenation
Acid–base properties
Metal–support interaction
CO adsorption
Ó 2010 Elsevier Inc. All rights reserved.
1. Introduction
On the other hand, the support acid–base properties can control
the chemoselectivity in the hydrogenation of ketones on Pt and
The Pt–cinchona system is the most efficient heterogeneous
catalyst for the enantioselective hydrogenation of activated ke-
tones, and also the most studied among the chirally modified metal
catalysts (for recent reviews, see [1–8]). A major part of the re-
search aimed at finding the appropriate combination of catalyst,
modifier, and the reaction conditions to achieve practically useful
enantioselectivities and broaden the application range of catalyst
system. It has been attempted to improve the enantioselectivity
by varying the catalyst synthesis method and the nature of the
support. Besides alumina, various other supports have been tested,
such as carbon, SiO2, TiO2, CaCO3, zeolites, clays, and organic poly-
mers [9–18], and even unsupported Pt nanoparticles have been ex-
plored as an alternative [19,20]. The early activity decayed after the
broad recognition that none of the attempts led to a catalyst supe-
rior to a commercially available 5 wt.% Pt/Al2O3 catalyst (Engelhard
4759). Interestingly, there is no systematic study in the literature
regarding the support effect on the enantioselectivity of Pt, and
the acidity of the support – for example zeolites – was considered
to have only minor impact [15].
other noble metal catalysts [21–23]. The hydrogenolysis of the
CAO bond can be accelerated or suppressed by increasing the cat-
alyst acidity or basicity, respectively, with inorganic or organic
additives and thus, the selectivity toward the alcohol can be tuned
[24]. Note that cinchona–modified Pt provides always the alcohol
in the hydrogenation of activated ketones, due to the strongly basic
quinuclidine N of the alkaloid.
The influence of support ionicity, i.e. the acid–base properties,
on the geometric and electronic properties of Pt has been exten-
sively studied by Koningsberger’s group in recent years [25–32].
They changed systematically the ionicity of alumina by introducing
different dopants (Si, W, F, Cl, K, Rb, and Cs) via the sol–gel method.
A broad range of characterization techniques and DFT calculations
helped to elucidate the remarkable metal–support interaction.
They proved that the PtAH bond strength is higher on ionic (basic)
supports and lower on acidic and more covalent supports, which is
the origin of the changes in the reactivity of Pt toward hydrogena-
tion and hydrogenolysis.
Being inspired from the idea of controlling the reactivity of Pt by
fine tuning the ionic properties of the support, we aimed at extend-
ing this analysis to the enantioselection on Pt in the hydrogenation
of activated ketones. We prepared two series of Pt catalysts,
* Corresponding author. Fax: +41 44 632 11 63.
E-mail address: baiker@chem.ethz.ch (A. Baiker).
0021-9517/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2010.02.012

116
F. Hoxha et al. / Journal of Catalysis 271 (2010) 115–124
supported on Si- and Cs-doped Al2O3, by single-step flame spray
pyrolysis [33]. We choose this technique because of its simplicity
and exceptional reproducibility in the synthesis of multi-compo-
nent catalysts [34–36]. The 4.7 wt.% Pt/Al2O3–SiO2 and 4.7 wt.%
Pt/Al2O3–Cs2O catalysts used in the present study contain 0–
80 wt.% SiO2 and 0–10 wt.% Cs2O, related to the alumina support.
The preparation and characterization of the catalysts by N2 absorp-
tion, X-ray diffraction (XRD), CO chemisorption, electron micros-
copy, and NH3 and CO2 temperature-programmed desorption
(TPD) has been reported elsewhere [33]. A remarkably good corre-
lation was found between the surface acidity of the Pt/Al2O3–SiO2
catalysts and their activity toward hydrogenolysis in the hydroge-
nation of acetophenone. On the other hand, already 0.25 wt.% Cs2O
hindered the hydrogenolysis of 1-phenylethanol to ethylbenzene
[33].
Scheme 2. Hydrogenation and hydrogenolysis of cyclohexyl methyl ketone.
Here, we report the enantioselective hydrogenation of ketopan-
tolactone and methyl benzoylformate in the presence of cinchoni-
dine as chiral modifier of Pt (Scheme 1). The two model substrates
TOF values were calculated based on the number of surface Pt
atoms [33].
were selected based on the absence of an H atom in a-position to
the keto-carbonyl group to avoid the aldol condensation and thus
minimize the side reactions induced by Pt, alumina, and the basic
alkaloid [7,37–39]. In case of ketopantolactone, also decarbonyla-
tion by Pt was barely detectable by ATR-IR spectroscopy [40]. To
help understanding the origin of changes in enantioselectivity,
characterization of the catalysts was completed with the chemose-
lective hydrogenation of cyclohexyl methyl ketone (Scheme 2) and
a DRIFT spectroscopic study of CO adsorption, a proven technique
to follow variations in the electronic properties of Pt with the sup-
port ionicity [31,41–44].
2.2. Diffuse reflectance infrared Fourier transform spectroscopy – CO
adsorption
Diffuse reflectance infrared Fourier transform spectroscopy
(DRIFTS) was used to study CO adsorption on the catalysts. The
analysis was carried out at 297 K with an EQUINOX 55 spectrome-
ter (Bruker Optics) equipped with a liquid nitrogen-cooled HgCdTe
detector. Multiple samples (max. 5), separated by quartz wool,
were placed without dilution in a plug-flow DRIFTS cell [45], allow-
ing identical experimental conditions for all samples. The off-gas of
the cell was analyzed by a mass spectrometer. Spectra were col-
lected by averaging 200 scans at 4 cmÀ1 resolution. The standard
reductive treatment procedure mentioned in the next chapter
was used to reduce the catalysts in flowing 20 vol.% H2/He
(40 ml/min). CO adsorption was monitored over 60 min in flowing
10 vol.% CO/He (40 ml/min), followed by flowing He (40 ml/min)
for 30 min. The peak areas were determined after deconvolution
of the spectra [46]. The spectrum of the reference catalyst E4759
is reported elsewhere [47].
2. Experimental
2.1. Materials
The anhydrous solution of ketopantolactone 1a (Roche AG, 99%)
in toluene was prepared by azeotropic distillation of toluene at re-
duced pressure. Cyclohexyl methyl ketone (Alfa Aesar, 95%) was
carefully distilled in vacuum before use. Other chemicals were
used as received: methyl benzoylformate 2a (Acros, 99%), cincho-
nidine (CD) (Fluka, P98% alkaloid), and toluene (Acros, extra dry,
water < 30 ppm).
2.3. Catalytic hydrogenation
The catalysts were always reduced at elevated temperature
prior to use. According to the standard procedure, the catalyst
was heated under flowing nitrogen up to 400 °C in 30 min, fol-
lowed by a reduction in flowing hydrogen for 60 min at the same
temperature, and finally cooling down in hydrogen in 30 min.
Hydrogenation of methyl benzoylformate at atmospheric pres-
sure was carried out in a 100-ml glass reactor equipped with mag-
netic stirrer, septum for sample collection or substrate addition,
and an inlet and outlet line for gas flowing. The reactor allowed
avoiding the contact of the freshly reduced catalyst with air, since
the reductive heat treatment and the hydrogenation reaction could
be carried out in the same reactor. After introducing the solvent
and modifier into the reactor containing the freshly reduced cata-
lyst under hydrogen at room temperature, the mixture was stirred
for 1 min. The reaction was started by introducing the substrate
and immediately switching on the stirring. Under standard condi-
The preparation and structural characterization of the flame-
made Pt catalysts has been reported elsewhere [33]. Aluminum
sec. butoxide (Aldrich, >98%), hexamethyldisiloxane (Fluka,
>98%), and cesium acetylacetonate (Fluka) dissolved in 2-ethylhex-
anoic acid (Aldrich, >98%, 0.5 M) were used as aluminum, silicon,
and cesium precursors, respectively. The chemical and textural
properties of the applied catalysts are collected in Table 1. The
tions, 16.15 lmol Pt catalyst (63 mg), 20.4 lmol modifier,
4.14 mmol substrate, and 20 ml solvent were stirred magnetically
(1000 rpm) at 20 °C for 4 h. In all experiments, the Pt/modifier/sub-
strate molar ratios were kept constant at 1/1.26/256.
Hydrogenation of ketopantolactone was carried out at elevated
pressure in a 25-ml stainless steel Parr autoclave equipped with a
16-ml glass liner and a PTFE cover, and a magnetic stirrer. The
autoclave was equipped also with a valve for sample collection
Scheme 1. The major enantiomers in the hydrogenation of ketopantolactone and
methyl benzoylformate over 4.7 wt.% Pt supported on SiO2- or Cs2O-doped Al2O3
modified by cinchonidine (CD).

F. Hoxha et al. / Journal of Catalysis 271 (2010) 115–124
117
Table 1
Chemical and textural properties of Pt/Al2O3–SiO2 and Pt/Al2O3–Cs2O catalysts containing 4.7 wt.% Pt. Catalysts are designated as Pt/Al–xSi and Pt/Al–xCs, respectively, where x is
the content of the dopant in wt.%.
Catalyst
SiO2 in the support
(wt.%)
Cs2O in the support,
(wt.%)
Al2O3 in the support
(wt.%)
BET surface area
Average Pt particle size
(nm)
Pt dispersion
(%)
(m2 gÀ1
)
Pt/Al
Pt/Al–5Si
–
5
–
–
–
–
–
–
–
–
0.25
0.5
3
5
10
100
95
90
85
77.5
70
50
20
99.75
99.5
97
138
128
123
126
136
139
157
211
208
234
257
243
227
1.9
3.2
4.3
4.0
5.5
4.3
7.8
13.5
2.0
2.2
2.1
1.6
1.8
60.3
35.8
26.1
28.5
20.6
26.6
14.5
8.4
56.2
51.9
55.1
69.0
61.5
Pt/Al–10Si
Pt/Al–15Si
Pt/Al–22.5Si
Pt/Al–30Si
Pt/Al–50Si
Pt/Al–80Si
Pt/Al–0.25Cs
Pt/Al–0.5Cs
Pt/Al–3Cs
Pt/Al–5Cs
Pt/Al–10Cs
10
15
22.5
30
50
80
–
–
–
–
–
95
90
or substrate addition. The treatment of the catalyst was carried out
in a fixed-bed reactor following the standard procedure. At the end,
the freshly reduced catalyst was purged with nitrogen for 10 min
at room temperature to remove the excess of hydrogen and then
transferred immediately to the autoclave. During the hydrogena-
tions of ketopantolactone at 100 bar, the proper amount of catalyst
tion of 5–80 wt.% SiO2 into the bulk of the non-porous alumina
diminished the crystallinity, and the SiO2-rich samples were amor-
containing 1.41 lmol Pt (5.5 mg), 1.7 lmol modifier, 1.61 mmol
substrate, and 5 ml solvent were stirred magnetically (1000 rpm)
at 18 °C for 30 min. The Pt/modifier/substrate molar ratios were
kept constant at 1/1.21/1142. The pressure was held at a constant
value with a pressure regulator valve.
The conversion and enantioselectivity (ee) were determined by
GC analysis, using an HP 6890 gas chromatograph and a Chirasil–
DEX CB (Chrompack 7502 25 m  0.25 mm  0.25
lm) capillary
column. All experiments were carried out at least twice, and the
average values are shown in the figures. Always the (R)-enantio-
mer was formed in excess without any detectable side product.
The estimated standard deviation of ee was about 0.5%.
Hydrogenation of ketopantolactone under optimized conditions
was carried out with the CD–modified Pt/Al-22.5Si catalyst (for
designation of catalysts see Table 1) at À10 °C and 100 bars in
5 ml toluene. The Pt/modifier/substrate molar ratios were 1/0.03/
718.
Hydrogenation of cyclohexyl methyl ketone was investigated at
3 bar in the same 25-ml Parr autoclave as described previously. The
Pt/substrate molar ratio was kept constant at 1/101. The proper
amount of catalyst containing 4.9 lmol Pt, 0.5 mmol substrate,
and 6 ml solvent were stirred magnetically (1000 rpm) at 20 °C
for 120 min. The pressure was held at a constant value with a con-
stant pressure regulator valve. The conversion and product compo-
sition were determined by GC analysis, using a Termo Quest Trace
2000 gas chromatograph and
a HP-FFAP (Agilent HP-FFAP
30 m  0.32 mm  0.25
l
m) capillary column. Bicyclohexyl was
used as internal standard (0.1 mmol).
3. Results and discussion
3.1. Catalyst properties – chemoselective hydrogenation of cyclohexyl
methyl ketone
The abbreviation and composition of the catalysts, as well as
their BET surface area and mean Pt particle size, are collected in Ta-
ble 1. A detailed characterization using nitrogen adsorption, XRD,
CO chemisorption for Pt particle size determination, tempera-
ture-programmed desorption (TPD) of NH3 and CO2, and electron
microscopy (TEM and STEM) combined with EDX analysis have
been reported elsewhere [33]. Summarizing it shortly, incorpora-
Fig. 1. Influence of the SiO2 content of the Pt/Al–Si catalysts on the selectivity to
CAO bond hydrogenolysis (to ethylcyclohexane, EC, Scheme 2) during the hydro-
genation of cyclohexyl methyl ketone (CMK) over the Pt/Al2O3–SiO2 catalysts. Data
refer to 50% conversion of CMK, and empty squares represent the reference catalyst
E4759.

118
F. Hoxha et al. / Journal of Catalysis 271 (2010) 115–124
phous. Temperature-programmed desorption of ammonia revealed
a maximum acidity at 22.5–30 wt.% SiO2 in the support. The sup-
port acidity had a strong influence on the properties of the PtAH
system, as evidenced by the hydrogenation of acetophenone. The
catalysts were highly active, and at 50% conversion, the selectivity
to the hydrogenolysis byproducts ethylbenzene and ethylcyclo-
hexane increased with increasing SiO2 content by more than 20-
fold.
The mean Pt particle size remained in the range 1.6–2.2 nm in
the catalysts doped with 0.25–10 wt.% Cs2O. The basicity increased
rapidly with Cs2O-doping, as confirmed by the temperature-pro-
grammed desorption of carbon dioxide. The dopant was probably
located in the surface region of the catalyst particles and strongly
influenced the chemoselectivity in acetophenone hydrogenation:
only 0.25 wt.% Cs2O ceased the hydrogenolysis of 1-phenylethanol
to ethylbenzene, and 5 wt.% Cs2O blocked the formation of ethylcy-
clohexane from cyclohexyl methyl ketone.
A limitation of acetophenone hydrogenation as a test reaction is
that the competing saturation of the phenyl ring may distort the
determination of the relative rates of hydrogenolysis and hydroge-
nation of the CAO bond. To avoid this effect, we completed the
characterization of the Pt/Al2O3–SiO2 catalysts with the hydroge-
nation of cyclohexyl methyl ketone (CMK, Scheme 2), which reac-
tion is in fact a part of the reaction route in acetophenone
hydrogenation. CMK can be hydrogenated to the saturated alcohol
1-cyclohexylethanol (CE) or hydrogenolyzed to ethylcyclohexane
(EC). A control experiment showed that hydrogenolysis of the ali-
phatic alcohol under the reaction conditions was not detectable
even after 20 h, in agreement with former reports [48,49]. As illus-
trated in Fig. 1a, all SiO2-doped catalysts were more active than the
reference Pt/Al2O3, which is the mostly used catalyst in the enan-
tioselective hydrogenation of ketones. There is no clear correlation
between the TOFs related to the number of surface Pt sites and the
SiO2 content of the mixed oxide support at low SiO2 content, where
the mean Pt particle size varied only in the relatively narrow range
2–5.5 nm (Table 1). There is a maximum in CAO bond hydrogenol-
ysis selectivity at around 20–30 wt.% SiO2 in the support, and all
SiO2-doped samples were more selective than the undoped Pt/Al
and the reference catalyst Pt/Al2O3.
3.2. Enantioselective hydrogenation over Pt/Al2O3–SiO2 catalysts
Variation of enantioselectivity with the support composition in
the hydrogenation of ketopantolactone over the CD–modified Pt/
Al–Si catalyst series is plotted in Fig. 2a. With increasing SiO2 con-
tent in the support up to 30 wt.%, the enantioselectivity to (R)-pan-
tolactone increased by 16%. Higher silica content was detrimental
but even those catalysts were more enantioselective than the ref-
erence Pt/Al2O3. A possible explanation for this behavior is that
too small Pt particles cannot accommodate the bulky CD–substrate
interacting complex and give lower ees [37,50], and doping with Si
increases the average Pt size (Table 1). However, literature shows
contradictory data on the particle size effect [12,51], and Fig. 2b
shows that this explanation might be valid only at low SiO2 con-
tents. Particularly at around the optimum enantioselectivity, the
Pt particle size cannot be the critical parameter.
A better correlation is found when the ee is plotted against the
acidity of the catalysts, using the hydrogenolysis selectivity to eth-
ylcyclohexane at 50% conversion of cyclohexyl methyl ketone
(Fig. 1b). Already a visual analysis reveals a striking similarity be-
tween enantioselectivity and chemoselectivity shown in Figs. 2a
and 1b, respectively. To quantify this correlation, we defined the
highest hydrogenolysis selectivity of the Pt/Al–22.5Si as unity,
and thus arrived at a scale of ‘‘normalized acidity” of this catalyst
series (Fig. 1b). The enantioselectivity to (R)-pantolactone shows
a good linear correlation (R2 = 0.93) with the normalized acidity
Fig. 2. Hydrogenation of ketopantolactone over CD–modified Pt/Al–Si catalysts.
Enantioselectivities determined at full conversion are plotted as a function of SiO2
content in the support (a), Pt particle size (b), and ‘‘normalized acidity” of the
catalysts (c). Standard conditions, toluene, 100 bar, 18 °C; open symbols refer to the
reference catalyst E4759. The ‘‘normalized acidity” is derived from the hydrogen-
olysis selectivity of the catalysts in the hydrogenation of methyl cyclohexyl ketone
(Fig. 1b).

F. Hoxha et al. / Journal of Catalysis 271 (2010) 115–124
119
as illustrated in Fig. 2c. This correlation suggests that there is a
common origin of the enantioselectivity to (R)-pantolactone and
the hydrogenolysis selectivity to ethylcyclohexane in the hydroge-
nation of the two structurally very different ketones: the variation
of the support acidity that controls the properties of the Pt–H sys-
tem via the metal–support interaction [21–23,28,30].
in the range 0.15–0.38 sÀ1, and all catalysts were more active than
the reference catalyst E4759 (TOF: 0.13 sÀ1).
The potential of catalysts synthesized by flame spray pyrolysis
is well illustrated by the high activities of the Si-doped catalysts
and also by the better enantioselectivity achieved in the hydroge-
nation of ketopantolactone (above 10 wt.% SiO2) and methyl ben-
zoylformate (whole series), compared to the enantioselectivity of
the commonly used Pt/Al2O3 E4759 (Figs. 2a and 3a, open
symbols).
A similar trend was observed in the hydrogenation of methyl
benzoylformate over the same catalyst series (Fig. 3a). The
improvement in enantioselectivity was smaller in this reaction,
probably because the conditions selected are closer to the opti-
mum (92% in the best case of Pt/Al–30Si vs 84% for the Pt/Al cata-
lyst). We arrived at the same interpretation of the data concerning
the origin of enhanced enantioselectivity by SiO2-doping: the in-
crease in particle size can explain only the changes below 4–
5 nm (Fig. 3b), while there is a reasonably good correlation be-
tween the ee and the normalized acidity of the catalysts (Fig. 3c).
This reaction was sufficiently slow to allow determination of the
initial reaction rates by GC analysis. The rates of alcohol formation
were constant up to high conversion for each catalyst. By increas-
ing the Si content in the support, the hydrogenation rate decreased
(Fig. 3d), parallel to the increase of the Pt particle size (Table 1).
The TOF values, related to the fraction of surface Pt atoms, varied
3.3. Enantioselective hydrogenation over Pt/Al2O3–Cs2O catalysts
In the hydrogenation of ketopantolactone over Cs-doped Pt/
Al2O3 catalysts, the enantioselectivity dropped sharply from 67%
to 26% ee by increasing the Cs content in the support (Fig. 4). Al-
ready very small amounts of Cs in the support (0.25–0.5 wt.%
Cs2O) had a detrimental effect. Since the average Pt particle size
in this catalyst series varied in a narrow range 1.6–2.2 nm (Table
1), the potential role of this parameter is negligible. In contrast,
the basicity of the catalysts characterized by the maximum tem-
perature of the CO2 desorption peak (TPD analysis [33] revealed
an inverse correlation: the position of the peak maximum in-
Fig. 3. Hydrogenation of methyl benzoylformate over CD–modified Pt/Al–Si catalysts. Enantioselectivity is plotted as a function of the SiO2 content in the support (a), the
mean Pt particle size (b), and the catalyst acidity (c). The initial reaction rates (d) and the enantioselectivities at full conversion were determined under standard conditions in
toluene, at 1 bar and 20 °C. Open symbols refer to the reference catalyst E4759.

120
F. Hoxha et al. / Journal of Catalysis 271 (2010) 115–124
creased monotonously in the whole concentration range investi-
gated, particularly below 1 wt.%. In other words, the higher the ba-
sicity of the support, the lower is the enantioselectivity in carbonyl
hydrogenation, independent of the Pt particle size. A comparison of
Figs. 2a and 4 reveals a remarkably higher efficiency of Cs as dop-
ant, which is attributed to its probable location on the support
[33]: while Si may enter into the alumina lattice and be well dis-
tributed there, Cs2O is probably enriched in the outer layer of alu-
mina and thus exerts a stronger effect on the Pt particles.
The negative influence of Cs2O on the enantioselectivity was
significant also in the hydrogenation of methyl benzoylformate,
as shown in Fig. 5. Only 10 wt.% Cs2O in the support reduced the
enantioselectivity from 84% to 31%. On the other hand, no unam-
biguous correlation was found between the hydrogenation rate
and the Cs2O content in the support. The rate of alcohol formation
remained constant up to high conversion for each catalyst and the
TOF, related to the fraction of surface Pt atoms, varied in the range
0.18–0.35 sÀ1. The higher activity of both Si- and Cs-doped catalyst
series, compared to the reference E4759 Pt/Al2O3, is attributed to
the non-porous structure of flame-made catalysts and the absence
of intra-particle mass transfer limitation [33].
Fig. 5. The influence of Cs2O-doping on the enantioselectivities determined at full
conversion of methyl benzoylformate; standard conditions, toluene, 1 bar, 20 °C.
The high energy signal is split into two narrow intensive signals,
3.4. IR study of CO adsorption
at ca. 2085 cmÀ1 and 2065 cmÀ1
, and a broad band at ca.
2030 cmÀ1. The first two signals are assigned to CO bound to highly
coordinated Pt atoms such as {1 1 1} and {1 0 0} terraces, respec-
tively, whereas the latter signal is attributed to CO adsorbed on
low coordinated Pt sites at step-edges, corners, and defects
[46,47,52–56].
The IR spectroscopic study of CO adsorption is the mostly used
method to gain insight into the electronic properties and morphol-
ogy of platinum. In order to help comparing our results to litera-
ture data, we extended the characterization of the Pt/Al–Si and
Pt/Al–Cs catalyst series with this method. To allow a direct com-
parison of these data with catalytic results, the samples were pre-
reduced in a flowing H2/He mixture at 400 °C. Mass spectrometric
analysis of the exhaust gas indicated that the stabilized reduced
state of the catalysts was reached within 30 min (the concentration
of the exhaust gas mixture was constant showing low water con-
tent). Due to the importance of SiO2-doped catalysts in enantiose-
lective hydrogenations, the following discussion will focus on the
characteristics of this series.
The DRIFT spectra of CO adsorbed on the reduced Pt/Al, Pt/Al-
30Si, and on Pt/Al-10Cs samples are shown in Fig. 6 as representa-
tive examples. Two modes of CO adsorption were identified from
the spectra: the bands around 2080 cmÀ1 are commonly attributed
to a linear, single bonded (on-top) species, and the broad and less
intensive signal at ca. 1830 cmÀ1 is attributed to bridged CO [33].
SiO2-doping modifies not only the signal intensities but also the
band shapes. With increasing SiO2 content, the bands correspond-
ing to CO linearly bound to highly coordinated Pt{1 1 1} and
Pt{1 0 0} terraces are enhanced in relative intensity and shifted
to higher wavenumbers in the range 2090–2095 cmÀ1 and 2067–
2080 cmÀ1, respectively (Fig. 7a and b). The ratio of CO linearly ad-
sorbed on Pt{1 1 1} and Pt{1 0 0} terraces varies in a narrow range
for all SiO2-doped catalysts, almost independent of the SiO2 con-
tent (Fig. 7d). Interesting is the fact that the band intensity corre-
sponding the CO bound to low coordinated Pt atoms reaches a
maximum at around 15–30 wt.% SiO2 content (not shown). Varia-
tion of the band position and the area of bridged CO is also shown
in Fig. 7a and c. The wavenumber shifts from 1802 to 1874 cmÀ1
with increasing the SiO2 content up to 50 wt.% SiO2 but the catalyst
with the highest SiO2 content (Pt/Al–80Si) does not follow this
trend. In contrast to the changes in linearly bound CO, the band
area of bridged CO decreases by increasing the SiO2 content in
the support.
Similar to several other characteristics of the two catalyst series,
the CO adsorption spectra on the Cs2O-doped catalysts displayed
the opposite trend compared to those of the SiO2-doped samples.
The intensity of the signals corresponding to linear adsorption de-
creased, while that of the bridged CO band increased with increas-
ing the Cs content in the support. Parallel to these changes, the
signals assigned to CO bound to {1 1 1} and {1 0 0} terraces and
to low coordinated Pt atoms were shifted to lower wavenumbers,
reaching a minimum at 5 wt.% Cs2O. To illustrate the opposite ef-
fects of doping alumina with Si or Cs, we plotted in Fig. 8 the inten-
sity ratio of bridged (B) to linearly (L) adsorbed CO (B/L) as a
function of the dopant content in the support. Despite of some sig-
nificant fluctuation in the B/L ratio, the inverse trend induced by
the basic and acidic additives is clearly discernible. The increasing
B/L intensity ratio with increasing Cs2O content in the support is in
agreement with former CO adsorption studies on Pt supported on
K-modified Al2O3, where K favored the bridged adsorption mode
of CO on Pt on the expense of the linear (on-top) CO species
[42,46].
Fig. 4. Enantioselectivities determined at full conversion of ketopantolactone on
Cs2O-doped Pt/Al2O3 catalysts; standard conditions, toluene, 100 bar, 18 °C.

F. Hoxha et al. / Journal of Catalysis 271 (2010) 115–124
121
tone with the variations in the particle size and the acid–base
properties of the catalysts. By introducing up to 80 wt.% SiO2 into
the support, the mean Pt particle size increased considerably (Ta-
ble 1). We have discussed previously that the flame spray pyroly-
sis conditions chosen allowed reasonable Pt particle size only for
the synthesis of the catalyst series covering the alumina-rich re-
gion [33]. Nevertheless, when considering the range 5–30 wt.%
SiO2, which is the most interesting concerning the catalytic re-
sults (Figs. 2 and 3), the mean Pt particle size varied in a rela-
tively narrow range of 4.35 1.05 nm. In both test reactions, the
highest ee was obtained at 30 wt.% SiO2, and the mean Pt size
of this catalyst was 4.3 nm. Figs. 2b and 3b illustrate that – con-
sidering the whole concentration range – the correlation between
enantioselectivity and mean Pt size is poor. It has been shown
that the enantioselectivity in the hydrogenation of activated ke-
tones depends on the Pt particle size and the optimum is at or
above 3–4 nm [11,61]. Note, however, that the suggested correla-
tion is far not satisfying. For example, tiny Pt nanoparticles of less
than 2 nm average diameter [19] provided higher enantioselectiv-
ities than those catalysts described in the previously mentioned
reports.
The other important feature of the SiO2-doped catalysts is the
support acidity. TPD of ammonia indicated the highest acidity at
22.5 and 30 wt.% SiO2 contents, and the data correlate well with
the hydrogenolysis activity of the Si-doped catalysts during
methyl cyclohexyl ketone hydrogenation (Fig. 1b). Figs. 2c and
3c display the good correlation between the enantioselectivity
and the normalized acidity, derived from the selectivity to the
hydrogenolysis product in methyl cyclohexyl ketone hydrogena-
tion (Fig. 1b).
Fig. 6. DRIFT spectra of CO adsorption on (A) Pt/Al and the deconvoluted
components of the CO bands, (B) Pt/Al–30Si, and (C) Pt/Al–10Cs samples after
prereduction at 400 °C in flowing H2. The fitted spectrum and the deconvoluted
components are represented with dashed and gray lines, respectively.
For the Cs2O-doped catalysts, the origin of the support effect is
clear. In this series of catalysts (0–10 wt.% Cs2O), the synthesis con-
ditions were ideal, and the average Pt particle size varied in the
narrow range 1.9 0.3 nm (Table 1). Hence, the dramatic drops
in the enantioselectivity to (R)-alcohols in the two test reactions
(Figs. 3 and 4) can be unambiguously related to the increased ba-
sicity of the doped catalysts, as determined by TPD of CO2 [33].
To get further support to our interpretation of the outstanding
enantioselectivity of the SiO2-doped catalysts, we plotted in
Fig. 9 the enantioselectivities of both catalyst series (SiO2- and
Cs2O-doped) as a function of the B/L ratio of CO adsorption. (In or-
der to better illustrate the effect of silica-promotion, the variation
of ee with the SiO2 content is plotted also separately in Fig. 9b.) The
unique advantage of the B/L ratio is that it gives a common scale to
characterize both the acidic and basic properties of the catalysts.
The general trend apparent in both test reactions is the improving
enantioselectivity with decreasing B/L ratio. In other words, the
acidic support favors and the basic support diminishes the enantio-
selection on the Pt surface, as compared to the performance of Pt/
Al2O3 (Pt/Al). The best ees were reached in a relatively narrow
range of the B/L ratio at around 0.25 (Fig. 9). Although some scat-
tering is obvious in Fig. 9, these deviations do not detract from the
basic correlation. For example, the Pt/Al-80Si catalyst is an outlier
with a B/L of ꢀ0.1, but as we discussed previously, the conditions
chosen for catalyst synthesis are not optimal for samples of high
silica content and these catalysts behave differently, compared to
the rest of the series.
The shift in the CO band position is often interpreted as evi-
dence of changes in the electronic properties of supported metals
[43,57]. The exact position of the absorption bands depends on
the particle size [58], surface coverage [59], and electronic changes
in the metal structure [57]. Since the particle size of the Pt/Al–Si
catalyst series varies with the SiO2 content (Table 1) and the CO
adsorption bands are broad, the shifts in the band position proba-
bly reflect both the changes in the electronic properties and in the
size of the Pt particles.
In a theoretical analysis of the electronic parameters that deter-
mine the coordination of CO to a transition-metal surface, Van San-
ten [60] demonstrated that the intensity ratio of bridged to linear
(B/L) bonded CO is a critical parameter. A growing B/L ratio indi-
cates an increase of the electron density on the supported Pt parti-
cles, which phenomenon is typical for Pt on a support doped with a
base [43,44]. Opposite is the situation for Pt particles located on an
acidic support, where the electron density on Pt is diminished with
increasing the support acidity, and this change is reflected by a
smaller B/L ratio. The B/L ratio was used in several studies by Kon-
ingsberger and coworkers as a key measure for the electronic state
of supported Pt [31,41,42]. It was found for all supports that the IR
bands of adsorbed CO shifted to higher wavelengths, and the B/L
ratio decreased with increasing acidity of the support. Both effects
were attributed to a decrease in the electronic charge of the sup-
port oxygen atoms and hence to a decrease of the electron density
on the supported Pt particles [43,44].
In the interpretation of the metal–support interaction, it is com-
monly assumed that the shift in the electronic properties of Pt with
the acid–base properties of the support controls the characteristics
of the PtAH system [31]. Considering the role of the electronic
properties of Pt in the enantioselection, we may go a step further
and speculate that probably also the adsorption mode and strength
of the modifier and the substrate are influenced by the acid–base
properties of the support, although this assumption needs
confirmation.
Next we will discuss the application of the B/L ratio as a unique
parameter in the interpretation of the enantioselectivity of both
SiO2- and Cs2O-doped Pt/Al2O3 catalysts.
3.5. Origin of the support effect on the enantioselectivity
In Section 3.2, we tried to correlate the enantioselectivities in
the hydrogenation of methyl benzoylformate and ketopantolac-

122
F. Hoxha et al. / Journal of Catalysis 271 (2010) 115–124
Fig. 7. Influence of SiO2-doping on the characteristics of CO adsorption on the reduced catalysts, as derived from the deconvoluted IR spectra. (a) Shift of the wavenumber of
CO linearly absorbed on Pt{1 1 1} (j) and {1 0 0} (h) terraces, and bridged adsorption of CO (N). (b) Variation of the band area of CO linearly absorbed on Pt{1 1 1} (j) and
{1 0 0} (h) terraces. (c) Variation of the band area of bridged CO adsorption. (d) The ratio of CO linearly absorbed on Pt{1 1 1} and Pt{1 0 0} terraces.
3.6. Ketopantolactone hydrogenation under optimized conditions
An intriguing question is whether the above described correla-
tion between the support acidity/basicity and the enantioselectiv-
ity can lead us to practically useful applications. In general, it is not
very difficult to find a set of reaction conditions (sometimes far
from the optimum), which allows a higher performance of the pro-
posed new catalyst compared to the reference catalyst. The only
real test of superiority is the comparison of the performance of
the two catalysts under the conditions optimized separately for
each of them.
The enantioselective hydrogenation of benzoylformates has
been thoroughly investigated in the past decade [62–64]. Years-
long optimization of the reaction conditions, catalyst pretreatment,
and the modifier structure by Bartók and coworkers allowed 98%
ee at 25 bar and 0 °C in the presence of acetic acid, using dihydroc-
inchonidine as the chiral modifier of Pt/Al2O3 and the ethyl ester of
the substrate [65–67]. To our knowledge, this is the highest value
achieved with a chirally modified metal catalyst, and there are only
two more examples to 98% ee using cinchona–modified Pt [19,68].
Hydrogenation of ketopantolactone has also been a popular test
reaction for chirally modified Pt [69–72], and the highest ee of
Fig. 8. Intensity ratio of bridged and linearly adsorbed CO (B/L) as a function of
dopant content (Cs2O or SiO2) in the support;
D – Pt/Al–Si series, h – Pt/Al–Cs
series, d – undoped Pt/Al2O3. Data are extracted from the deconvoluted IR spectra.

F. Hoxha et al. / Journal of Catalysis 271 (2010) 115–124
123
Fig. 10. Influence of the molar ratio of CD to the total amount of Pt on the
enantioselectivity in the hydrogenation of ketopantolactone over the Pt/Al–22.5Si
catalyst. Standard conditions, toluene, 100 bar, 18 °C.
persion of Pt in the two catalysts is also different, but this effect is
relatively small. Repeating the best reaction in Fig. 10 at À10 °C
gave 94 0.5% ee, which value was well reproducible. It is ex-
pected that a careful re-optimization including a broader range
of
reaction
parameters
can
provide
even
better
enantioselectivities.
4. Conclusions
The effect of the support on catalyst performance, or more spe-
cifically on the Pt electronic properties in heterogeneous catalysis,
is an extensively studied – and debated – field. It is still a challeng-
ing task to localize the origin of the phenomenon and separate it
from the disturbance caused by, for example, the different impuri-
ties in the catalyst systems compared.
In this study, we applied two series of flame-made Pt/Al2O3 cat-
alysts, where the acidic or basic properties of the support were var-
ied by the introduction of SiO2 and Cs2O, respectively. Changing the
acid–base properties of the support allowed to vary systematically
the electronic properties of Pt, characterized by the bridged to lin-
ear (B/L) ratio of chemisorbed CO. A striking correlation was
uncovered between the enantioselectivities achieved in the hydro-
genation of ketopantolactone and methyl benzoylformate, and the
acid–base properties of the catalysts reflected by the ratio B/L of
chemisorbed CO. This correlation indicates that the acid–base
properties of the support are a key factor for the optimization of
the enantioselectivity in the hydrogenation of activated ketones.
The practical importance of this study was demonstrated by the
highest enantioselectivity achieved in the hydrogenation of keto-
pantolactone. On the other hand, our study indicates that trace
amounts of basic impurities originating from the catalyst synthesis
may explain the sometimes huge differences in the performance of
supported Pt catalysts.
As a final note, there is a fundamental difference between the
effect of acidic supports and an acid additive or solvent. The sup-
port acidity affects the electronic state of Pt and thus influences
the enantioselectivity and reaction rate. In contrast, acid additives
protonate the basic N atom of the alkaloid modifier and they can
even form various linear or cyclic H-bonded complexes with the
alkaloid and the ketone [74], and these chemical changes influence
the reaction rate and enantioselectivity. To illustrate the differ-
ences, in the hydrogenation of ketopantolactone, acetic acid dimin-
ishes the ee [75] while the acidic support improves it.
Fig. 9. Variation of the ee with the electronic properties of Pt reflected by the B/L
ratio of CO adsorption on the a) Pt/Al, Pt/Al–Si, and PtAl–Cs catalyst series and b) Pt/
Al and Pt/Al–Si catalyst series. Filled symbols: ketopantolactone hydrogenation,
open symbols: methyl benzoylformate hydrogenation. ds: Pt/Al catalyst; jh: Pt/
Ai–Si catalysts, and ND: Pt/Al–Cs catalysts.
91.6% was achieved on the Pt/Al2O3 – cinchonidine system at
70 bar and À9 °C in toluene [73]. In that study, and also in the
hydrogenation of ethyl benzoylformate affording 98% ee, the com-
mercial Pt/Al2O3 (E4759) catalyst was applied, which is used as a
reference in the present study.
In the following, the hydrogenation of ketopantolactone to (R)-
pantolactone – an important intermediate in the industrial synthe-
sis of pantothenic acid – was chosen as the test reaction. To prove
the superior performance of Si-doped Pt/Al2O3 catalysts, the Pt/Al–
22.5Si sample was selected for a limited parameter study. This cat-
alyst possessed the highest acidity as indicated by TPD of ammonia
and hydrogenolysis of methyl cyclohexyl ketone [33]. A key
parameter in the optimization of ketopantolactone hydrogenation
was the CD concentration [73]. The optimum in the CD/Pt molar ra-
tio was 0.01 with the E4759 catalyst but using our Si-doped cata-
lyst, the ee could be improved by increasing this ratio by a factor of
3–10, under otherwise identical conditions (Fig. 10). The difference
probably arises from the higher amount of CD adsorbed on the
more acidic support, leading to a lower actual CD concentration
in solution, and thus on the Pt surface. The change can be compen-
sated by increasing the overall CD/Pt molar ratio. Note that the dis-

Hoxha et al. / Journal of Catalysis 271 (2010) 115–124
F.
124
[34] R. Jossen, M.C. Heine, S.E. Pratsinis, S.M. Augustine, M.K. Akhtar, Appl. Catal., B
Acknowledgments
69 (2007) 181.
[35] V. Tiwari, J. Jiang, V. Sethi, P. Biswas, Appl. Catal., A 345 (2008) 241.
[36] R. Strobel, A. Baiker, S.E. Pratsinis, Adv. Powder Technol. 17 (2006) 457.
[37] F. Hoxha, N. van Vegten, A. Urakawa, F. Krumeich, T. Mallat, A. Baiker, J. Catal.
261 (2009) 224.
[38] E. Tálas, L. Botz, J. Margitfalvi, O. Sticher, A. Baiker, J. Planar Chromatogr. 5
(1992) 28.
Financial support of this work by the Swiss National Science
Foundation and ETH Research Grants TH-09 06-2 and TH-41 06-1
is kindly acknowledged. We thank Johnson Matthey PLC (J.M.) for
supporting this project by providing the Pt precursor.
[39] D. Ferri, S. Diezi, M. Maciejewski, A. Baiker, Appl. Catal., A 297 (2006) 165.
[40] N. Bonalumi, T. Bürgi, A. Baiker, J. Am. Chem. Soc. 125 (2003) 13342.
[41] B.L. Mojet, J.T. Miller, D.E. Ramaker, D.C. Koningsberger, J. Catal. 186 (1999)
373.
[42] D.C. Koningsberger, D.E. Ramaker, J.T. Miller, J. de Graaf, B.L. Mojet, Top. Catal.
15 (2001) 35.
[43] T. Visser, T.A. Nijhuis, A.M.J. van der Eerden, K. Jenken, Y.Y. Ji, W. Bras, S.
Nikitenko, Y. Ikeda, M. Lepage, B.M. Weckhuysen, J. Phys. Chem. B 109 (2005)
3822.
[44] M. Lepage, T. Visser, A.M.J. van der Eerden, F. Soulimani, B.M. Weckhuysen,
Vib. Spectrosc. 48 (2008) 92.
[45] A. Urakawa, A. Baiker, Top. Catal. 52 (2009) 1312.
[46] M.J. Kappers, J.H. Vandermaas, Catal. Lett. 10 (1991) 365.
[47] E. Schmidt, W. Kleist, F. Krumeich, T. Mallat, A. Baiker, Chem.–Eur. J. 16 (2010)
2181.
[48] P. Geneste, Y. Lozano, C.R. Acad. Paris 280 (1975) 1137.
[49] P.N. Rylander, S. Starrick, Engelhard Ind. Tech. Bull. 7 (1966) 106.
[50] T. Marzialetti, M. Oportus, D. Ruiz, J.L.G. Fierro, P. Reyes, Catal. Today 133
(2008) 711.
[51] J.L. Margitfalvi, E. Tàlas, L. Yakhyaeva, E. Tfirst, I. Bertóti, L. Tóth, in: D.G.
Morrell (Ed.), Catalysis of Organic Reactions, M. Dekker, New York, 2003, p.
393.
[52] R.P. Eischens, W.A. Pliskin, S.A. Francis, J. Chem. Phys. 22 (1954) 1786.
[53] R.G. Greenler, K.D. Burch, K. Kretzschmar, R. Klauser, A.M. Bradshaw, B.E.
Hayden, Surf. Sci. 152/153 (1985) 338.
References
[1] T. Mallat, E. Orglmeister, A. Baiker, Chem. Rev. 107 (2007) 4863.
[2] M. Bartók, Curr. Org. Chem. 10 (2006) 1533.
[3] D.Y. Murzin, P. Maki-Arvela, E. Toukoniitty, T. Salmi, Catal. Rev. – Sci. Eng. 47
(2005) 175.
[4] F. Zaera, J. Phys. Chem. C 112 (2008) 16196.
[5] T. Bürgi, A. Baiker, Acc. Chem. Res. 37 (2004) 909.
[6] M. Studer, H.U. Blaser, C. Exner, Adv. Synth. Catal. 345 (2003) 45.
[7] M. von Arx, T. Mallat, A. Baiker, Top. Catal. 19 (2002) 75.
[8] P.B. Wells, A.G. Wilkinson, Top. Catal. 5 (1998) 39.
[9] D. Ruizand, P. Reyes, J. Chil. Chem. Soc. 53 (2008) 1740.
[10] H.U. Blaser, H.P. Jalett, D.M. Monti, A. Baiker, J.T. Wehrli, Stud. Surf. Sci. Catal.
67 (1991) 147.
[11] J.T. Wehrli, A. Baiker, M. Monti, H.U. Blaser, J. Mol. Catal. 61 (1990) 207.
[12] G.A. Attard, K.G. Griffin, D.J. Jenkins, P. Johnston, P.B. Wells, Catal. Today 114
(2006) 346.
[13] H.U. Blaser, D. Imhof, M. Studer, Stud. Surf. Sci. Catal. 108 (1997) 175.
[14] U. Bohmer, K. Morgenschweis, W. Reschetilowski, Catal. Today 24 (1995) 195.
[15] H.U. Blaser, H.P. Jalett, M. Müller, M. Studer, Catal. Today 37 (1997) 441.
[16] A. Bykov, V. Matveeva, M. Sulman, P. Valetskiy, O. Tkachenko, L. Kustov, L.
Bronstein, E. Sulman, Catal. Today 140 (2009) 64.
[17] Y. Orito, S. Imai, S. Niwa, N.G. Hung, J. Synth. Org. Chem. 37 (1979) 173.
[18] B. Török, K. Balàzsik, I. Kun, G. Szöllössi, G. Szakonyi, M. Bartók, Stud. Surf. Sci.
Catal. 125 (1999) 515.
[19] X. Zuo, H. Liu, M. Liu, Tetrahedron Lett. 39 (1998) 1941.
[20] H. Bönnemann, G.A. Braun, Angew. Chem., Int. Ed. 35 (1996) 1992.
[21] A.Y. Stakheev, L.M. Kustov, Appl. Catal., A 188 (1999) 3.
[22] K. Hayek, R. Kramer, Z. Paal, Appl. Catal., A 162 (1997) 1.
[23] G. Haller, D.E. Resasco, Adv. Catal. 36 (1989) 173.
[24] S. Nishimura, Handbook of Heterogeneous Catalytic Hydrogenation for
Organic Synthesis, Wiley, New York, 2001.
[25] D.C. Koningsberger, M.K. Oudenhuijzen, J. de Graaf, J.A. van Bokhoven, D.E.
Ramaker, J. Catal. 216 (2003) 178.
[26] M.K. Oudenhuijzen, J.A. van Bokhoven, D.E. Ramaker, D.C. Koningsberger, J.
Phys. Chem. B 108 (2004) 20247.
[27] Y.Y. Ji, A.M.J. van der Eerden, V. Koot, P.J. Kooyman, J.D. Meeldijk, B.M.
Weckhuysen, D.C. Koningsberger, J. Catal. 234 (2005) 376.
[28] M.K. Oudenhuijzen, J.A. van Bokhoven, J.T. Miller, D.E. Ramaker, D.C.
Koningsberger, J. Am. Chem. Soc. 127 (2005) 1530.
[29] D.E. Ramaker, M.K. Oudenhuijzen, D.C. Koningsberger, J. Phys. Chem. B 109
(2005) 5608.
[30] Y.Y. Ji, V. Koot, A.M.J. van der Eeirden, B.M. Weckhuysen, D.C. Koningsberger,
D.E. Ramaker, J. Catal. 245 (2007) 415.
[54] R.K. Brandt, M.R. Hughes, L.P. Bourget, K. Truszkowska, R.G. Greenler, Surf. Sci.
286 (1993) 15.
[55] K. McCrea, J.S. Parker, P.L. Chen, G. Somorjai, Surf. Sci. 494 (2001) 238.
[56] H. Harle, U. Metka, H.R. Volpp, J. Wolfrum, Phys. Chem. Chem. Phys. 1 (1999)
5059.
[57] G. Blyholder, J. Phys. Chem. 68 (1964) 2772.
[58] R. van Hardeveld, F. Hartog, Adv. Catal. 22 (1972) 75.
[59] F. Stoop, F.J.C.M. Toolenaar, V. Ponec, J. Catal. 73 (1982) 50.
[60] R.A. van Santen, J. Chem. Soc. Faraday Trans. I 83 (1987) 1915.
[61] D.Y. Murzin, E. Toukoniitty, React. Kinet. Catal. Lett. 90 (2007) 19.
[62] K. Szöri, M. Sutyinszki, K. Felföldi, M. Bartók, Appl. Catal., A 237 (2002) 275.
[63] K. Felföldi, K. Balàzsik, M. Bartók, J. Mol. Catal. A – Chem. 202 (2003) 163.
[64] S. Diezi, S. Reimann, N. Bonalumi, T. Mallat, A. Baiker, J. Catal. 239 (2006) 255.
[65] B. Török, K. Balàzsik, M. Török, K. Felföldi, M. Bartók, Catal. Lett. 81 (2002) 55.
[66] M. Sutyinszki, K. Szöri, K. Felföldi, M. Bartók, Catal. Lett. 81 (2002) 281.
[67] M. Sutyinszki, K. Szöri, K. Felföldi, M. Bartók, Catal. Commun. 3 (2002) 125.
[68] W.M. Huck, T. Mallat, A. Baiker, Adv. Synth. Catal. 345 (2003) 255.
[69] N. Künzle, R. Hess, T. Mallat, A. Baiker, J. Catal. 186 (1999) 239.
[70] S. Diezi, T. Mallat, A. Szabo, A. Baiker, J. Catal. 228 (2004) 162.
}
[71] M. Bartók, K. Balázsik, I. Bucsi, G. Szöllosi, J. Catal. 239 (2006) 74.
[72] T.A. Martinek, T. Varga, K. Balazsik, G. Szollosi, F. Fulop, M. Bartok, J. Catal. 255
(2008) 296.
[73] M. Schürch, N. Künzle, T. Mallat, A. Baiker, J. Catal. 176 (1998) 569.
[74] M. Schürch, O. Schwalm, T. Mallat, J. Weber, A. Baiker, J. Catal. 169 (1997) 275.
[75] M. von Arx, T. Bürgi, T. Mallat, A. Baiker, Chem. – Eur. J. 8 (2002) 1430.
[31] A.Y. Stakheev, Y. Zhang, A.V. Ivanov, G.N. Baeva, D.E. Ramaker, D.C.
Koningsberger, J. Phys. Chem. C 111 (2007) 3938.
[32] D.E. Ramaker, M. Teliska, Y. Zhang, A.Y. Stakheev, D.C. Koningsberger, Phys.
Chem. Chem. Phys. 5 (2003) 4492.
[33] B. Schimmoeller, F. Hoxha, T. Mallat, F. Krumeich, S.E. Pratsinis, A. Baiker, Appl.
Catal. A, Gen. 374 (2010) 48.