High-throughput kinetic study of hydrogenation over palladium nanoparticles: Combination of reaction and analysis

Article abstract of DOI:10.1002/chem.200701780

The hydrogenation of 1-acetylcyclohexene, cyclohex-2-enone, nitrobenzene, and trans-methylpent-3-enoate catalyzed by highly active palladium nanoparticles was studied by high-throughput on-column reaction gas chromatography. In these experiments, catalysis and separation of educts and products is integrated by the use of a catalytically active gas chromatographic stationary phase, which allows reaction rate measurements to be efficiently performed by employing reactant libraries. Palladium nanoparticles embedded in a stabilizing polysiloxane matrix serve as catalyst and selective chromatographic stationary phase for these multiphase reactions (gas-liquid-solid) and are coated in fused-silica capillaries (inner diameter 250 μm) as a thin film of thickness 250 nm. The palladium nanoparticles were prepared by reduction of palladium acetate with hydridomethylsiloxane-dimethylsiloxane copolymer and self-catalyzed hydrosilylation with methylvinylsiloxane-dimethylsiloxane copolymer to obtain a stabilizing matrix. Diphenylsiloxane-dimethylsiloxane copolymer (GE SE 52) was added to improve film stability over a wide range of compositions. Herein, we show by systematic TEM investigations that the size and morphology (crystalline or amorphous) of the nanoparticles strongly depends on the ratio of the stabilizing polysiloxanes, the conditions to immobilize the stationary phase on the surface of the fused-silica capillary, and the loading of the palladium precursor. Furthermore, hydrogenations were performed with these catalytically active stationary phases between 60 and 100°C at various contact times to determine the temperature-dependent reaction rate constants and to obtain activation parameters and diffusion coefficients.

Full text of DOI:10.1002/chem.200701780

FULL PAPER
DOI: 10.1002/chem.200701780
High-Throughput Kinetic Study of Hydrogenation over Palladium
Nanoparticles: Combination of Reaction and Analysis
Oliver Trapp,* Sven K. Weber, Sabrina Bauch, Tobias Bäcker, Werner Hofstadt, and
[
a]
Bernd Spliethoff
Abstract: The hydrogenation of 1-ace-
tylcyclohexene, cyclohex-2-enone, ni-
trobenzene, and trans-methylpent-3-
enoate catalyzed by highly active palla-
dium nanoparticles was studied by
high-throughput on-column reaction
gas chromatography. In these experi-
ments, catalysis and separation of
educts and products is integrated by
the use of a catalytically active gas
chromatographic stationary phase,
which allows reaction rate measure-
ments to be efficiently performed by
employing reactant libraries. Palladium
nanoparticles embedded in a stabilizing
polysiloxane matrix serve as catalyst
and selective chromatographic station-
ary phase for these multiphase reac-
tions (gas–liquid–solid) and are coated
in fused-silica capillaries (inner diame-
ter 250 mm) as a thin film of thickness
250 nm. The palladium nanoparticles
were prepared by reduction of palladi-
um acetate with hydridomethylsilox-
ane–dimethylsiloxane copolymer and
wide range of compositions. Herein, we
show by systematic TEM investigations
that the size and morphology (crystal-
line or amorphous) of the nanoparti-
cles strongly depends on the ratio of
the stabilizing polysiloxanes, the condi-
tions to immobilize the stationary
phase on the surface of the fused-silica
capillary, and the loading of the palla-
dium precursor. Furthermore, hydroge-
nations were performed with these cat-
alytically active stationary phases be-
tween 60 and 1008C at various contact
times to determine the temperature-de-
pendent reaction rate constants and to
obtain activation parameters and diffu-
sion coefficients.
self-catalyzed
methylvinylsiloxane–dimethylsiloxane
copolymer to obtain stabilizing
hydrosilylation
with
a
matrix. Diphenylsiloxane–dimethylsi-
loxane copolymer (GE SE 52) was
added to improve film stability over a
Keywords: gas chromatography
hydrogenation kinetics
nanoparticles · palladium
·
·
·
Introduction
molecular level. Nowadays, various high-throughput screen-
ing techniques are available, for example, (time-resolved)
IR thermography, resonance-enhanced multiphoton ioniza-
[3]
The discovery of highly productive and selective catalysts
and the expansion of synthetic techniques and methods in
chemistry is of substantial interest for both science and in-
dustry. For the directed design of catalysts, comprehensive
kinetic and thermodynamic data of existing catalytic systems
are necessary to understand how the catalytic mechanism
[
4]
[5]
tion (REMPI), scanning mass spectrometry, visual detec-
[6]
[7]
tion with reactive dyes, UV/Vis spectroscopy, fluores-
cence-based assays, nondispersive IR (NDIR) analysis,
circular dichroism (CD) for screening of enantioselective
[8]
[9]
[
10]
catalysts, and acoustic-wave sensor systems,
which allow
[1,2]
might be controlled by structural parameters.
This re-
the investigation of comprehensive catalyst libraries. Chro-
[11]
quires the measurement of large sets of kinetic data for a
broad variety of substrates. From these data, models can be
developed and refined to identify rate-controlling elementa-
ry steps to gain insight into the reaction mechanism on a
matographic techniques
have the advantage that even
complex reaction mixtures can be separated and quantified
without any tagging with marker molecules to make ana-
lytes detectable by a specific detector. Consistent enhance-
ments in the field of parallelized kinetic measurements of
[12]
catalysts with minute material consumption are achieved
with microfluidic devices, which combine chemical synthesis
[
a] Dr. O. Trapp, S. K. Weber, S. Bauch, T. Bäcker, W. Hofstadt,
B. Spliethoff
Max-Planck-Institut für Kohlenforschung
Kaiser-Wilhelm-Platz 1, 45470 Mülheim an der Ruhr (Germany)
Fax : (+49)208-306-2995
[13]
and analysis on the same chip. However, a major draw-
back of commonly used batch reactors is the limitation of
studying only one reaction per run. The reaction, separation,
and quantification of the educts and products to determine
E-mail: trapp@mpi-muelheim.mpg.de
Chem. Eur. J. 2008, 14, 4657 – 4666
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4657

conversions have to be performed consecutively because
competing reactions lead to indefinable reaction kinetics.
Recently, we reported a strategy that allows the synchronous
combination of catalysis and separation in capillaries to per-
form high-throughput reaction rate measurements (on aver-
the Langmuir–Hinshelwood mechanism. From these experi-
ments, the diffusion coefficients of the substrates in the
polysiloxane matrix were also obtained.
ꢀ
1
age 147 reactionsh ) of reactant libraries for hydrogenation
Results and Discussion
[14]
over highly active palladium nanoparticles.
We focused on hydrogenation reactions as this chemical
transformation is one of the most important and widely
Preparation of Pd nanoparticles embedded in a polysiloxane
matrix: To control the growth of Pd nanoparticles and to
prevent them from agglomeration, surfactants are typically
used, for example, quaternary ammonium salts, which form
a monolayer on the surface of the nanoparticles and stabi-
[15]
used catalytic reactions in synthetic chemistry. Hydroge-
nations can be catalyzed by stabilized Pd and Pt nanoparti-
[16]
cles, which have become the focus of increasing interest
in catalysis in recent years due to their high activity under
[16d]
lize them by steric and electrostatic effects.
The protect-
[17]
mild reaction conditions.
However, multiphase catalytic
ing shell also controls the interaction of the substrate with
the active site on the nanoparticles and consequently the
catalytic activity. Taking this into consideration and making
the protecting shell compatible with a chromatographic sta-
tionary phase, so that the catalytic system can be investigat-
ed by on-column reaction gas chromatography, the following
strategy was developed. Pd nanoparticles were embedded in
a polysiloxane matrix by using vinylpolysiloxane and hydri-
systems (gas–liquid–solid), which are typically used in indus-
trial applications, are difficult to investigate as the interac-
tion of the reactant with the catalyst is controlled by the
mass transfer between the different phases, and therefore
the apparent reaction rate is composed of the inherent reac-
tion rate and diffusion rates. To reduce this effect, the inter-
facial area has to be increased. Microstructured reaction sys-
tems intrinsically have a high specific interfacial area per
volume (ainter =2/r), which is only dependent on the radius
of circular reaction channels, that is, for microcapillaries
with inner diameters (i.d.) between 250 and 100 mm the spe-
cific interfacial area per volume ranges from 16000 to
[20]
dopolysiloxane, which acts as reducing agent
and cross-
linked by hydrosilylation at the same time. Therefore, we
prepared a methylvinylsiloxane–dimethylsiloxane copolymer
(MVPS, 4.5% Si(O)
A
H
R
U
G
A
H
R
U
Pd(OAc) to the vinyl groups in diethyl ether. To reduce
AHCTREUNG
2
2
ꢀ3
2+
0
4
0000 m m .
Pd to Pd and to obtain an inert protecting matrix for the
Pd nanoparticles, a hydridomethylsiloxane–dimethylsiloxane
Furthermore, nanoparticles show extraordinary catalytic
[21]
activity because of high surface-to-volume ratios. The parti-
cle size and the morphology of Pd nanoparticles, as well as
the nature of the stabilizing matrix to prevent agglomera-
tion, correlate with their catalytic activity.
copolymer
(HMPS, 25.7% Si(O)(CH )H groups) was
A
H
R
U
G
3
added to this mixture. Cross-linking with the MVPS was
achieved by hydrosilylation, catalyzed in the presence of Pd-
1
A
H
R
U
G
2
There are numerous exam-
ples describing hydrogena-
[17]
tions and C-C coupling reac-
[18]
tions
catalyzed by Pd nano-
particles or formation of Pd
nanoparticles from Pd com-
plexes in the course of the reac-
tion. Despite the large number
of publications, further insights Scheme 1. Preparation of highly active Pd nanoparticles embedded in a polysiloxane matrix.
into the dependency of these
quasi-homogeneous reactions
on the size and morphology of the nanoparticles are neces-
sary to understand how their structure controls the reaction
mechanism.
spectroscopic measurements, we could confirm that the per-
centage of Si(O)(CH )H groups starting from 16.1%, exem-
plified for a mixture of Pd
25.7% Si(O)(CH )H groups), and MVPS (6 mg, 4.5%
Si(O)(CH )(CH=CH ) groups), decreased to 2.6%.
AHCTREUNG
3
[19]
A
H
E
N
Herein, we show by systematic TEM investigations that
the size and morphology (crystalline or amorphous) of the
nanoparticles strongly depends on the ratio of the stabilizing
polysiloxanes, the activation temperature to immobilize the
stationary phase on the surface of the capillary, and the
loading of the Pd precursor. Hydrogenations were per-
formed with these catalytically active stationary phases be-
tween 60 and 1008C at various contact times by on-column
reaction gas chromatography to determine temperature-de-
pendent reaction rate constants and to obtain activation pa-
ACHTREUNG
3
A
H
R
U
G
ACHTREUNG
3
2
We observed by TEM measurements that the size and
morphology of the Pd nanoparticles depends on the ratio of
the stabilizing polymers, the activation conditions (under a
hydrogen atmosphere, temperature), and the concentration
of the Pd precursor. The obtained Pd nanoparticles stabi-
lized by a 3:1 ratio of the two polysiloxanes (HMPS/MVPS)
are spherical and crystalline with a narrow size distribution
[
14]
of (3.2ꢁ0.7) nm
determined by TEM (Figure 1a). After
°
°
°
rameters (DG , DH , DS ) with a kinetic model based on
treatment with hydrogen at an elevated temperature
4658
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 4657 – 4666

Hydrogenation over Palladium Nanoparticles
FULL PAPER
Figure 1. a) Overview TEM image of Pd nanoparticles showing the homogeneous size distribution; b) TEM image of amorphous nanoparticles; c) size
distribution of Pd nanoparticles with a ratio of HMPS/MVPS 1:2and mean diameter (3.0 ꢁ1.0) nm; and d) size distribution of Pd nanoparticles depend-
ing on the ratio of the stabilizing polysiloxanes before heat treatment under a hydrogen atmosphere.
(
2008C), we also found amor- Table 1. Particle sizes and distributions and their dependency on the polysiloxane ratio before and after heat
[a]
[22]
treatment under a hydrogen atmosphere. The conversions C, rate constants k, and Gibbs activation energies
phous nanoparticles
(Fig-
°
DG are given for the hydrogenation of cyclohex-2-enone.
ure 1b) with an increased size
of (3.6ꢁ1.6) nm, which can be
attributed to the formation of
Pd hydride and also agglomera-
tion processes at these elevated
temperatures. Surprisingly, the
size is also controlled by the 11:1
ratio of the two polysiloxanes.
To study this effect systemati- (0.3 mg). The Pd concentration was 1.1”10 molcm fused-silica capillary for the catalytic testing at 608C
°
Ratio
HMPS/MPVS
Particle size
after preparation [nm]
Particle size
2008C/H [nm]
C
[%]
k
DG
ꢀ
1
ꢀ1
2
A
H
R
U
G
]
A
C
H
T
R
E
U
N
G
[kJmol ]
1
1
1
2
:11
:23.0
:1
5.5ꢁ2.8
ꢁ1.0
11.0ꢁ6.2
3.6ꢁ1.6
3.4ꢁ1.5
5.5ꢁ1.8
5.9ꢁ2.5
66.4
88.24.85
49.4
37.1
82.5
2.27
79.6
77.5
3.4ꢁ1.1
5.4 ꢁ2.0
5.4ꢁ2.0
1.56
1.07
3.87
80.7
81.7
78.2
:1
[
a] Samples were prepared from the mixture of HMPS and MPVS (24 mg), GE SE 52 (24 mg), and Pd
A
C
H
T
R
E
U
N
G
(OAc)
2
ꢀ
10
ꢀ1
2
and 100 kPa H , by using a 3 cm capillary.
cally, we prepared 40 Pd nano-
particle samples by variation of
both the mixture of the two
polysiloxanes and the Pd loadings (Table 1). It is important
to note that for these experiments we added the slightly
polar stationary phase GE SE 52to the polysiloxane mixture
to improve the stability of coatings on the inner surface of
fused-silica capillaries. The addition of the GE SE 52has the
effect of making the particle sizes larger because of the dilu-
tion effect of the two cross-linking polysiloxanes.
As described above, the MVPS coordinates the Pd
A
C
H
T
R
E
U
N
G
(OAc)
2
through vinyl groups while the HMPS is needed for the re-
2
+
0
duction of Pd to Pd and the cross-linking of the polymer
matrix. Taking into account that the content of Si(O)(CH )-
(CH=CH ) groups in the MVPS is 4.5% and the content of
ACHTREUNG
3
AHCTREUNG
2
Si(O)(CH )H groups in the HMPS is 25.7%, a 1:6 ratio of
A
H
R
U
G
3
the two polymers (HMPS/MVPS) should lead to complete
cross-linking. However, HMPS is also needed to reduce
Chem. Eur. J. 2008, 14, 4657 – 4666
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4659

O. Trapp et al.
2
+
0
Pd to Pd , and therefore to achieve completion of the re-
action an HMPS/MVPS ratio ranging from at least 1:1 to
3
let formation, which indicates destruction of the polymer
film. However, to compare the catalytic properties of the Pd
nanoparticles with the varying polysiloxane compositions,
the slightly polar stationary phase GE SE 52(polysiloxane
containing 5% phenyl groups) was added to the mixture.
On-column catalysis experiments were performed by cou-
pling the Pd nanoparticle microcapillaries (length=2.02,
2.90, 3.84, 5.10, 6.15 cm) between a preseparation capillary
(1 m) and a separation column (25 m), which were installed
in a gas chromatograph. The purpose of the preseparation
column was to thermally equilibrate the reactants and to
spatially separate the educts of the injected compound li-
brary, which enabled high-throughput kinetic investigations
due to the absence of competing reactions. Hydrogen was
used as a reactive carrier gas. Reaction educts and products
were detected by flame ionization detection (FID) for quan-
tification and identified by quadrupole ion-trap mass spec-
trometry. The educt library consisting of four unsaturated
and functionalized compounds (1-acetylcyclohexene, cyclo-
hex-2-enone, nitrobenzene, and trans-methylpent-3-enoate)
was simultaneously injected onto this column configuration
at different temperatures (60–1008C) and inlet pressures
(60–100 kPa) to vary the reaction time and to obtain tem-
perature-dependent kinetic data (Figure 2).
:1 is required. The size distribution of the nanoparticles
measured by TEM for five selected ratios of the two stabi-
lizing polysiloxanes (HMPS/MVPS 1:11, 1:2, 1:1, 2:1, 11:1)
indicates that a 1:2mixture leads to smaller Pd nanoparti-
cles with a narrow size distribution (Figure 1c and Table 1).
Pd nanoparticles prepared with a polysiloxane ratio of
HMPS/MVPS 1:11 show an increased particle size and a
broader size distribution. To study the effect of hydrogen on
the size and morphology of the Pd nanoparticles, Cu grids
for TEM analysis were coated with the viscous brownish-
gray polysiloxanes and then treated with hydrogen at 2008C.
It was found that high HMPS and low MVPS concentrations
and vice versa, as well as high Pd precursor concentrations,
lead to pronounced agglomeration of the particles with a
broad size distribution ((11.0ꢁ6.2) nm; see Table 1).
The dependency of nanoparticle formation on the particle
size and distribution of the two polysiloxanes can be ex-
2
+
0
plained by kinetic effects. The reduction of Pd to Pd is
compared to the hydrosilylation leading to cross-linking and
stabilization of the nanoparticles in a fast process. It can be
assumed that in a first step, seeds of small Pd nanoparticles
are formed which agglomerate to form the resulting nano-
particles. This process is con-
trolled by diffusion. The MVPS
used here is less viscous than
HMPS, which should allow
easier diffusion of the nanopar-
ticles resulting in the observed
agglomeration to larger parti-
cles. Samples prepared with a
high HMPS concentration form
Pd nanoparticles quickly, which
can be observed by a darkening
of the solution. The agglomera-
tion itself is controlled by diffu-
sion and by the rate of the
cross-linking of the two polysi-
loxanes, which dynamically in-
creases the viscosity and de-
creases diffusion of the parti-
cles.
Figure 2. Expanded image (2–9 min) of the gas chromatographic separation and mass traces of the on-column
hydrogenation of the four substrates at 708C and 100 kPa.
Catalytic studies by on-column
reaction chromatography: To perform kinetic investigations
of hydrogenation over the prepared Pd nanoparticles, solu-
tions of the polysiloxanes with the embedded Pd nanoparti-
cles were used to coat fused-silica capillaries (i.d. 250 mm,
film thickness 250 nm) by the static method described by
By the use of educt libraries and measurement of reaction
rate constants of the spatially separated compounds, an ex-
traordinarily high throughput can be realized. Recently, we
demonstrated that for a library consisting of 22 compounds
a throughput of 5880 reactions in 40 h could be achieved.
Here, we screened systematically varied sets of fused-silica
capillaries coated with Pd nanoparticles embedded in a
polysiloxane matrix, by changing the ratio of stabilizing pol-
ysiloxanes (see Table 1) and the Pd loading, under different
reaction conditions (temperature, contact time) and with
variation of the column length.
[23]
Grob. For the activation of the Pd nanoparticles and to
permanently bond the polysiloxanes to the fused-silica capil-
lary by reaction of surface silanol groups with the remaining
hydridosiloxane groups, the fused-silica capillaries were
ꢀ1
heated to 2008C at a rate of 0.5 Kmin under slow hydro-
gen flow. It was found that polymers prepared with an
excess of HMPS gave more stable films not leading to drop-
4660
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 4657 – 4666

Hydrogenation over Palladium Nanoparticles
FULL PAPER
The conversions and rate constants for the hydrogenation
of cyclohex-2-enone varied in a broad range for the individ-
ual polysiloxane ratios (see Table 1). These data (in total
370 experiments were performed) can be interpreted by
taking the particle sizes and the degree of cross-linking of
the polysiloxanes into consideration. Higher conversions are
observed for small particles than for larger particles. Howev-
er, the reaction rate constant is also influenced by the
degree of cross-linking, which controls the accessibility of
the nanoparticles by the substrates. For nanoparticles only
stabilized by the viscosity of the linear polysiloxane chains,
higher conversions, even for larger nanoparticles, are ob-
served (1:11 and 11:1 for HMPS/MVPS). To make a com-
promise between particle size, stabilization, and activity, a
polysiloxane ratio of 3:1 for HMPS/MVPS was chosen for
further experiments, because under these conditions the Pd
nanoparticles do not agglomerate when heated.
Figure 3. Educt conversion at 808C and 80 kPa versus the column length
for trans-methylpent-3-enoate (^), cyclohex-2-enone (
~), and 1-acetylcyclohexene (”).
&
), nitrobenzene
(
The dependency of the conversion on the column length
is the prerequisite for on-column reaction chromatography
[24]
to determine kinetic data. The contact time should corre-
late with the length of the capillary and the catalytic activity
must not change. It is improtant to note that this correlation
is only given for diluted samples, because overloading of the
[25]
stationary phase leads commonly to nonlinear effects.
To prove that the contact time correlates with the length
of the capillary, the hydrogenation reactions of the sub-
strates were performed at 808C by using different capillary
lengths (2.02, 2.90, 3.84, 5.10, 6.15 cm) at a constant Pd load-
ꢀ
11
ꢀ1
ing of 7.28”10 molcm (see Table 2; medium concentra-
Figure 4. Educt conversion versus the Pd loading for trans-methylpent-3-
enoate (^), cyclohex-2-enone (
), nitrobenzene (~), and 1-acetylcyclo-
hexene (”).
&
Table 2. Variation of the Pd concentration (50% GE SE 52, HMPS/
MVPS 3:1).
Capillary length
cm]
Pd
[mg]
Pd loading
[10 molcm ]
ꢀ
11
ꢀ1
[
which lead to larger particles because of agglomeration, the
relative surface area is decreased compared to small nano-
particles and therefore a nonlinear decrease of the catalytic
activity is expected.
5
2
2
.10
.90
.02
0.05
0.10
0.20
3.64
7.28
14.56
To investigate this dependency in greater detail, we per-
formed temperature- and flow-dependent conversion meas-
urements to obtain three-dimensional datasets (temperature,
reaction time, conversion) for each compound. All measure-
ments were repeated three times at 60, 70, 80, 90, and
1008C and hydrogen inlet pressures of 60, 70, 80, 90, and
100 kPa. Note that the resulting conversion surface functions
are not flat planes, which can be attributed to the fact that
the residence time strongly depends on the separation effi-
ciency of the reactor column. The separation efficiency is a
function of the temperature and the flow rate and is de-
scribed by the van Deemter equation (see below). The eval-
uation of these datasets gives the activation parameters,
which can be correlated with the Pd loading of the capilla-
ries and also the composition of the polysiloxanes
(Figure 5).
tion of the three Pd concentrations investigated, HMPS/
MVPS 3:1), which resulted in different reaction times in the
range of 76 to 1690 ms (Figure 3). For longer capillaries
(
>6 cm) or higher Pd concentrations, we observed complete
conversion for some of the substrates, for example, 99.9%
conversion for trans-methylpent-3-enoate on a 6.15 cm capil-
lary.
Additionally, we systematically varied the Pd nanoparticle
loading in the fused-silica capillaries (see Table 2) and per-
formed catalytic tests with the four hydrogenation educts.
We found that in this concentration range, the conversion of
the educts and hence the catalytic activity of the Pd nano-
particles linearly correlates with the Pd concentration of the
Pd precursor. We can conclude that the hydrogenation pro-
ceeds according to a first-order reaction with regard to Pd.
Figure 4 shows the conversion as a function of the Pd load-
ing on the capillary for the different substrates. However, it
has to be pointed out that for increasing Pd concentrations,
Determination of reaction rate constants and activation pa-
rameters: The hydrogenation of unsaturated compounds
Chem. Eur. J. 2008, 14, 4657 – 4666
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4661

O. Trapp et al.
Considering our experimental
conditions in on-column reac-
tion chromatography, the fol-
lowing assumptions can be
made. 1) H is used as a reac-
2
tive carrier gas at a high flow
ꢀ
1
rate in the range of mLmin ,
and thus there is an excess of
H compared to the substrates
2
(
typically
H /substrate
2
9
9.99999:0.00001). 2) The ad-
sorption of H2 on the Pd sur-
face is fast because the activa-
tion energy for this process is
very low and H2 easily passes
the polysiloxane matrix. This
means that there is a steady
state of hydrogen adsorbed on
the Pd nanoparticles, which
does not influence the intrinsic
hydrogenation rate of the sub-
strate. 3) Compared to the sub-
strate there is an excess of Pd
nanoparticles, so that there are
enough reactive sites available
Figure 5. 3D plots of the conversion of the hydrogenations of nitrobenzene (a), cyclohex-2-enone (b), 1-acetyl-
cyclohexene (c), and trans-methylpent-3-enoate (d) at different temperatures (60–1008C) and inlet pressures for the adsorption of the reac-
(
60–100 kPa).
tion educt. 4) The hydrogena-
tion of the substrates can be
considered as an irreversible
over Pd surfaces is typically described by a Langmuir–Hin-
process to the hydrogenated product. Dehydrogenation in
the presence of Pd is only expected at higher temperatures,
typically higher than at least 1508C, because here the limit-
[26]
shelwood mechanism. To use this model for the evalua-
tion of the experimental data, the following concept has
been elaborated to adapt it to the on-column reaction chro-
matographic experiment. The Langmuir–Hinshelwood
mechanism is depicted in Scheme 2a, in which the substrates
are represented as a double bond and the catalytically active
sites by an asterisk in the following mechanistic description,
considering, in particular, the diffusion-limited steps (ad-
sorption and desorption of educts and products).
[27]
ing step is the desorption of hydrogen from the metal.
Consideration of these points simplifies the model in the on-
column reaction chromatographic setup to the key steps de-
picted in Scheme 2b. The advantage of this chromatographic
experiment is that the transport by diffusion of the substrate
to the catalytically active site can be considered orthogonal
to the hydrogenation process itself. The distribution equili-
bria of the educts and products are obtained by the reten-
tion parameters, and diffusion coefficients can be deter-
mined by temperature- and flow-dependent measurements
of the separation efficiency by using the retention times and
peak width of the individual substrates. Therefore, the
model depicted in Scheme 2b can be applied to every chro-
matographic theoretical plate, which can be considered as a
catalytic chromatographic reactor (Scheme 2c). This means
that with the known retention parameters given the resi-
dence time of the substrate, the reaction time is defined and
reaction rate constants can be calculated from the conver-
sion data by using a first-order reaction rate law according
to Equation (1):
ꢀ
dðsubstrateÞ
ð1Þ
¼
k2½substrateŠ
dt
Scheme 2. Hydrogenation of the substrates described by a Langmuir–
Hinshelwood mechanism. Substrates are represented as a double bond
(=) and the catalytically active sites by an asterisk (*).
This expression is independent of the absolute concentra-
4662
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 4657 – 4666

Hydrogenation over Palladium Nanoparticles
FULL PAPER
tions of the substrates. Notably, the conversions were deter-
mined from the data for gas chromatography with FID,
which gives signal intensities that can be correlated to the
ionizable carbon atoms in a specific molecule. This means
that the signal intensities of the hydrogenated product and
educt are the same and therefore no correction factor has to
be applied.
was achieved according to the Langmuir–Hinshelwood
mechanism.
The high activity of the Pd nanoparticles is corroborated
by the activation parameters, which show low activation en-
°
°
thalpies DH and negative activation entropies DS that cor-
respond to a restrained transition state. Comparison of the
activation parameters between 1-acetylcyclohexene, cyclo-
hex-2-enone, and trans-methylpent-3-enoate shows that the
less-hindered alkene (trans-methylpent-3-enoate) has the
Residence times of the hydrogenation educts were deter-
mined with a reference column by using the same composi-
tion of the stationary phase as for the catalytically active ca-
pillary columns. Reaction rate constants, k, of hydrogenation
were determined by application of Equation (1) to the con-
°
lowest activation enthalpies DH and activation entropies
°
DS compared to the more hindered alkenes, for example,
1-acetylcyclohexene. For cyclohex-2-enone, the reaction
°
°
version data. The Gibbs free activation energies DG (T)
rates are mainly dominated by the activation entropy DS ,
were calculated according to the Eyring equation [Eq. (2)]
which results in high productivities.
with kB as the Boltzmann constant (k =1.380662”
As already pointed out, the reaction rate constants k in-
crease with higher Pd loadings, which can be attributed to a
linear dependency on the rate law of the active sites of the
catalysts. This can be correlated to the Pd concentration
under the described diluted conditions. We also found that
the reaction rate constants were lower for the stabilizing
matrix with GE SE 52as the additive compared to the Pd
nanoparticles, which are only stabilized by HMPS and
B
ꢀ
23
ꢀ1
1
0
JK ), T as the reaction temperature (K), h as Planckꢁs
ꢀ
34
constant (h=6.62617”10 Js,) and R as the gas constant
ꢀ
1
ꢀ1
(
1
R=8.31441 JK mol ). The statistical factor k was set to
.0.
ꢀ
ꢁ
k1h
°
DG ðTÞ ¼ ꢀRT ln
ð2Þ
kk T
B
[14]
MVPS. This finding suggests that the phenyl groups of the
stationary phase GE SE 52have an inhibiting effect on the
hydrogenation reaction, probably caused by stabilization of
the nanoparticles. Such stabilization effects for Pd nanopar-
ticles have been also observed and used by Kobayashi
Activation parameters were calculated from temperature-
dependent measurements. The activation enthalpy DH of
°
the reaction was obtained from the slope and the activation
°
[17a]
entropy DS from the intercept of the Eyring plot (ln
A
H
R
U
(k /T)
et al.
1
ꢀ
1
as a function of T ). Deviations of the activation parame-
°
°
ters DH and DS were calculated by error-band analysis of
the linear regression with a level of confidence of 95%.
Table 3 summarizes the conversions, reaction rate con-
stants k, and activation parameters (DG , DH , DS ) of the
hydrogenation reactions of the four substrates for different
Pd loadings. By applying pseudo-first-order reaction kinet-
ics, a very good agreement with respect to the substrates
Determination of diffusion constants: In microfluidic sys-
tems, the control of mixing is often challenging, because dif-
fusion rates contribute to the apparent reaction rates. In the
present setup, diffusion processes can be quantified and ex-
perimentally controlled by the polysiloxane used. The van
°
°
°
[28]
Deemter equation
diffusion coefficients:
[Eq. (3)] was employed to determine
Table 3. Selected results of the on-column hydrogenations over highly active Pd nanoparticles for three different palladium loadings.
[
a]
[b]
°[c]
°
°
[d]
Substrate
Product
Pd loading
10 molcm
C
k
[s
DG
[kJmol
DH
DS
r
ꢀ
11
ꢀ1
ꢀ1
ꢀ1
ꢀ1
ꢀ1
ꢀ1
[
]
[%]
A
H
R
U
G
]
A
H
R
U
G
]
A
H
R
N
[kJmol
]
[JK mol
]
A
H
R
U
G
3.64
7.29
14.6
1.20.03
17.2
65.0
91.3
68.3
ꢁ0.6
69.4ꢁ0.3
76.5ꢁ2.8
ꢀ77ꢁ1
ꢀ49ꢁ1
ꢀ19ꢁ1
0.999 (0.028)
0.999 (0.034)
0.993 (0.102)
0.29
1.05
84.1
82.1
3.64
7.29
14.6
10.8
49.7
100.0
0.94
3.83
–
79.9
75.8
–
39.3ꢁ0.9
53.5ꢁ0.7
ꢀ136ꢁ5
ꢀ75ꢁ1
0.997 (0.045)
0.995 (0.073)
–
[
e]
[e]
[e]
[e]
[e]
–
–
3.64
7.29
14.6
3.6
11.7
58.0
0.07
0.14
0.65
86.1
83.6
80.4
37.4ꢁ0.5
47.4ꢁ0.2
57.2ꢁ1.7
ꢀ163ꢁ6
ꢀ121ꢁ1
ꢀ78ꢁ3
0.999 (0.023)
0.991 (0.098)
0.996 (0.065)
3.64
7.29
14.6
16.4
32.4
88.0
3.35
4.00
13.9
76.7
77.3
75.239.8
28.5ꢁ0.4
64.5ꢁ0.8
ꢁ2.5
ꢀ162ꢁ6
ꢀ43ꢁ1
ꢀ44ꢁ1
0.999 (0.010)
0.996 (0.077)
0.999 (0.015)
[
a] Conversion C at 608C on a 2cm fused-silica column. [b] Reaction rate constant at 60 8C. [c]Gibbs activation energy at 258C. [d]Correlation factor
and residual standard deviation (s.d.) of the linear regression of the Eyring plot. [e]Conversion is 100% for most data points under these conditions.
Chem. Eur. J. 2008, 14, 4657 – 4666
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4663

O. Trapp et al.
B
(C ). D and D are coefficients of molecular diffusion in
S
M
S
H ¼ A þ u þ Cu
ð3Þ
the mobile and stationary phases, respectively, d , the inner
c
diameter of the capillary, and d , the film thickness of the
f
in which H is the height equivalent to a theoretical plate, u,
the mean velocity of the mobile phase, A, the eddy diffu-
sion, B, the coefficient of longitudinal diffusion, and C, the
coefficient for the mass transfer. Interestingly, van Deemter
originally studied the heat and mass transport in fixed cata-
lyst beds and later transferred these derivations to chroma-
tography, which are well-known to optimize and character-
ize the efficiency of a chromatographic system. Therefore,
the effective plate height H=l/N, obtained from the capilla-
ry length l (24 m) divided by the effective plate number N,
was plotted as a function of the velocity u of the carrier gas
at 608C for the four substrates (see Figure 6). The effective
stationary phase. The retention factor k’=(t ꢀt )/t is cal-
R
M
M
culated from the retention time tR of the substrate and the
hold-up time t , determined by co-injection of methane.
M
From the coefficients of the van Deemter plots, the diffusion
coefficients of the substrates in the mobile phase (D ) and
M
in the stationary phase (D ) were calculated. These diffusion
S
coefficients are summarized in Table 4.
Table 4. Diffusion coefficients of the substrates in the mobile phase (D
and in the stationary phase (D ) at 608C.
M
)
S
2
ꢀ1
ꢀ10
2 ꢀ1
Substrate Product
D
M
[cm s
]
S
D [10 cm s ]
ꢀ
ꢁ
tRꢀt0
plate numbers, N=5.545
,
wh
0
1
.14
.91
1.82
2.79
0
0
.20
.81
0.55
31.1
Considering these diffusion coefficients and the small
paths for the substrates to migrate into the catalytically
active stationary phase (film thickness 250 nm), it can be
concluded that the reactions are not limited by diffusion.
This means that the reaction rate constants determined in
this on-column reaction chromatographic setup are directly
accessible and depend only on the probability of the sub-
strates to adsorb on Pd nanoparticles, which correlates with
the concentration of the active sites in the reaction volume.
Figure 6. Determination of diffusion coefficients from van Deemter plots
of the four substrates at 608C (precisely 2400 cm fused-silica column (in-
jector outlet to detector inlet, i.d. 250 mm)).
were calculated from the retention times t , peak width at
R
half height, w , of the individual substrates, and the hold-up
h
time, t , determined from methane injections.
0
These curves were fitted with the van Deemter equation
to obtain the constants B and C for the individual substrates.
For coated capillaries, the eddy diffusion (A term) can be
neglected because no additional migration paths caused by
different particle sizes, shapes, and porosity such as in in
packed columns have to be considered [Eq. (3)]. The
B term, representing the longitudinal diffusion, directly
yields the diffusion coefficients for the substrates in the
mobile phase at the given temperature [Eq. (4)].
Conclusion
We have investigated hydrogenation reactions over highly
active Pd nanoparticles by a synchronous combination of
catalysis and separation in microcapillaries for different
ratios of the stabilizing polysiloxanes, different activation
temperatures, and different Pd loadings. We obtained com-
prehensive kinetic data from on-column chromatographic
experiments by high-throughput reaction rate measurements
of reactant libraries. With these data, we gained further in-
sights into the influence of reaction parameters, for example,
the particle size and stabilizing matrix. This helped us to un-
derstand the parameters controlling the mechanism and the
kinetics of the investigated hydrogenation reactions over Pd
nanoparticles, for example, the steric hindrance of 1-acetyl-
cyclohexene compared to cyclohex-2-enone and trans-meth-
B
H ¼ u þðCM þ C Þu with
S
B ¼ 2DM,
ð4Þ
0
02
2
0
2
1
þ 6k þ 11k
dc
D
2k
df
D
CM
¼
ꢂ
, C ¼
ꢂ
0
2
S
0
2
9
6ð1 þ k Þ
M
3ð1 þ k Þ
S
The coefficient C for the mass transfer consists of the mass
transfer coefficient in the mobile (C ) and stationary phases
M
4664
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 4657 – 4666

Hydrogenation over Palladium Nanoparticles
FULL PAPER
ylpent-3-enoate results in a more negative activation entro-
py DS , which can be explained by a more restrained transi-
Preparation of Pd nanoparticles in a polysiloxane matrix and coating of
microcapillaries: Palladium acetate, HMPS, MVPS, and GE SE 52(Ma-
cherey-Nagel, containing 5% phenyl groups) were mixed in absolute di-
ethyl ether in the ratios given in Tables 1 and 2. Fused-silica capillaries
°
tion state. Furthermore, the influence of the nature of the
stabilizing matrix of the Pd nanoparticles on the catalytic ac-
tivity was investigated. The employed concept of on-column
reaction chromatography integrating catalyzed reactions and
separation efficiency can be generally applied to other cata-
lytic processes to characterize catalysts and materials in
comprehensive kinetic studies.
(
length 5 m, thermally deactivated at 2208C for 24 h) were coated with
1
these solutions. The reaction course was followed by H NMR spectros-
copy for the SiH group. H NMR (300 MHz, CDCl
0
nals of the H NMR spectrum, the total content of Si(O)
was determined.
1
3
, 2 58 C): d=ꢀ0.20–
.19 (m; Si(CH (n=1–3)), 4.61 ppm (s, SiH); by integration of the sig-
A
H
R
U
G
3 n
)
1
A
H
R
U
G
3
)H groups
[23]
Capillaries were coated by the static method described by Grob
to
obtain a uniform film with a thickness of 250 nm. Therefore, fused-silica
capillaries (i.d. 250 mm) were filled with the respective polymer solution
in diethyl ether and the solvent was removed by high vacuum after clos-
ing one end of the capillary. Then the capillaries were flushed with argon
and the polymer film was immobilized in a slow hydrogen stream while
Experimental Section
ꢀ
1
heating the capillary from 25 to 2008C at a rate of 0.5 Kmin . The tem-
perature was maintained for 8 h.
General methods and materials: Unless otherwise indicated, all reactions
were performed under an argon atmosphere by using standard Schlenk
techniques. All chemicals were obtained from Fluka (Buchs, Switzerland)
or Sigma–Aldrich (Steinheim, Germany) and used as received. GE SE 52
was purchased from Macherey–Nagel (Düren, Germany). All solvents
were dried by using standard techniques. Fused-silica capillaries (i.d.
On-column hydrogenation: On-column hydrogenation experiments were
performed on a Thermo Trace PolarisQ GC–MS apparatus equipped
with a split injector (2508C), a flame ionization detector (2508C), and a
quadrupole ion-trap mass spectrometer. For the hydrogenation of differ-
ent substrates, fused-silica capillaries coated with Pd nanoparticles em-
bedded in a polysiloxane matrix (2.02–6.15 cm”250 mm i.d., 0.25 mm film
thickness) were employed. These capillaries were coupled between a pre-
separation column (HP-5, 1 m”250 mm i.d.) and a separation column
2
50 mm, o.d. 365 mm) were purchased from Microquartz (Munich, Germa-
ny).
Analytical techniques: Nuclear magnetic resonance ( H NMR and
1
(
HP-5, 25 m”250 mm i.d.) to quantify the reaction mixture. Hydrogen
1
3
3
C NMR) spectra were recorded in CDCl on a Bruker DPX 300 spec-
was used as a carrier gas. All measurements were repeated three times at
each temperature (60, 70, 80, 90, and 1008C) and pressure (60, 70, 80, 90,
and 100 kPa). A total of 375 measurements were considered for the stat-
istical analysis of the reaction rate constants to obtain activation parame-
ters.
trometer (Rheinstetten, Germany) either at 300 and 75 MHz, respective-
ly, or on a Bruker AV 400 spectrometer (Rheinstetten, Germany) at 400
and 100 MHz, respectively. On-column reaction gas chromatographic
measurements were performed on a Thermo Trace PolarisQ GC–MS ap-
paratus equipped with a split injector (2508C) and a flame ionization de-
tector (2508C). Hydrogen (5.0) was used as reactive carrier gas. The
measurements were repeated three times at each temperature in steps of
1
0 K. Transmission electron micrographs were obtained on a Hitachi 7500
microscope operating with an acceleration voltage of 100 kV. High-reso-
lution transmission electron microscopy (HRTEM) images were obtained
on a Hitachi HF 2000 microscope with a cold field-emission source oper-
ating at 200 keV.
Acknowledgements
Generous financial support by an Emmy Noether grant of the Deutsche
Forschungsgemeinschaft (DFG TR 542/3), the Max-Planck-Institut für
Kohlenforschung, the Fonds der Chemischen Industrie (FCI), and the
Merck Research Laboratories (Rahway, New Jersey, USA) is gratefully
acknowledged. S.K.W. thanks the Cusanuswerk for a doctorate scholar-
ship. We thank Dr. B. Tesche and A. Dreier for TEM, SEM, and EDX
measurements.
Hydridomethylsiloxane–dimethylsiloxane copolymer: A mixture of hexa-
methyldisiloxane (8.1 g, 50.0 mmol), polyhydridomethylsiloxane (6.0 g,
1
00.0 mmol), octamethylcyclotetrasiloxane (66.7 g, 225.0 mmol), china
clay (2g), and sulfuric acid (1 mL) was stirred at 100 8C under gentle
reflux for 5 d. During this time, the reaction mixture became increasingly
more viscous. After cooling the mixture to room temperature, the cata-
lyst (china clay+sulfuric acid) was removed by extraction with water and
filtration. Water and other volatile constituents of the filtrate were re-
moved by evaporation at 1208C under an oil-pump vacuum to give
[
1] P. L. Hansen, J. B. Wagner, S. Helveg, J. R. Rostrup-Nielsen, B. S.
Clausen, H. Topsoe, Science 2002, 295, 2053.
1
HMPS (62.6 g, 78.3%). H NMR (300 MHz, CDCl
3
, 2 58 C): d=ꢀ0.02–
13
0
.25 (m; Si
A
C
H
T
R
E
U
N
G
3 n
(CH ) (n=1–3)], 4.61 ppm (s; SiH); C NMR (100 MHz,
[
[
2] C. T. Campbell, Nature 2004, 432, 282.
CDCl
3
, 2 58C): d=ꢀ2.3, 1.3, 1.6 ppm; by integration of the signals of the
3] a) L. Razon, R. Schmitz, Chem. Eng. Sci. 1986, 41, 1561; b) F. C.
Moates, M. Somani, J. Annamalai, J. T. Richardson, D. Luss, R. C.
Willson, Ind. Eng. Chem. Res. 1996, 35, 4801; c) A. Holzwarth, H.-
W. Schmidt, W. F. Maier, Angew. Chem. 1998, 110, 2788; Angew.
Chem. Int. Ed. 1998, 37, 2644; d) M. T. Reetz, M. H. Becker, K. M.
Kühling, A. Holzwarth, Angew. Chem. 1998, 110, 2792; Angew.
Chem. Int. Ed. 1998, 37, 2647; e) S. J. Taylor, P. J. Morken, Science
1998, 280, 267; f) S. S. Lasko, R. J. Hendershot, J. Reed, Y. Fu, M.-F.
Fellmann, G. Oskarsdottir, C. M. Snively, J. Lauterbach in High-
Throughput Analysis: Spectroscopic Imaging in the Mid-Infrared Ap-
plied to High-Throughput Studies of Supported Catalyst Libraries
(Eds.: R. A. Potyrailo, E. Amis), Kluwer Academic, New York,
2003, p. 77; g) J. Klein, W. Stichert, W. Strehlau, A. Brenner, D.
Demuth, S. A. Schunk, H. Hibst, S. Storck, Catal. Today 2003, 81,
329; h) S. Cypes, A. Hagemeyer, Z. Hogan, A. Lesik, Comb. Chem.
High Throughput Screening 2007, 10, 2 5.
1
H NMR spectrum, the total content of Si(O)
mined to be 25.7%.
ACHTREUNG
3
(CH )H groups was deter-
Methylvinylsiloxane–dimethylsiloxane copolymer: Diethoxymethylvinyl-
silane (4.3 g, 26.9 mmol) and diethoxydimethylsilane (35.9 g, 242.0 mmol)
were mixed and sodium hydroxide (0.5 g) dissolved in a mixture of water
(
50 mL) and ethanol (50 mL) was added dropwise at 08C. The mixture
was stirred for 1 h, then heated to 608C for 5 h. Chlorotrimethylsilane
0.9 g, 8.1 mmol) was added slowly for end-capping and the mixture was
(
stirred for 24 h. The product was extracted with n-pentane and washed
with water. The organic phase was separated and the solvent was re-
moved by rotary evaporation. The product was dried under high vacuum
1
at 608C to give MVPS (14.8 g, 69.2%). H NMR (300 MHz, CDCl
3
,
2
58C): d=ꢀ0.03–0.09 (m; Si
A
H
R
U
G
3
)
n
(n=1–3)), 5.69–5.99 ppm (m; Siꢀ
1
3
CH=CH
5
2
3
, 2 58 C): d=136.1, 132.0, 131.7,
1
6.7, 17.3, 0.0 ppm; by integration of the signals of the H NMR spec-
trum, the total content of Si(O)
to be 4.5%.
A
H
R
U
G
3
)
A
H
R
U
G
2
) groups was determined
[4] S. M. Senkan, S. Ozturk, Angew. Chem. 1999, 111, 867; Angew.
Chem. Int. Ed. 1999, 38, 791.
Chem. Eur. J. 2008, 14, 4657 – 4666
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4665

O. Trapp et al.
[
5] a) P. Cong, R. D. Doolen, Q. Fan, D. M. Giaquinta, S. Guan, E. W.
Mcfarland, D. M. Poojary, K. Self, H. W. Turner, W. H. Weinberg,
Angew. Chem. 1999, 111, 507; Angew. Chem. Int. Ed. 1999, 38, 483;
b) M. Orschel, J. Klein, H.-W. Schmidt, W. F. Maier, Angew. Chem.
[16] a) H. Bçnnemann, W. Brijoux, R. Brinkmann, E. Dinjus, T. Joujen,
B. Korall, Angew. Chem. 1991, 103, 1344; Angew. Chem. Int. Ed.
Engl. 1991, 30, 1312; b) M. T. Reetz, W. Helbig, S. A. Quaiser, U.
Stimming, N. Breuer, R. Vogel, Science 1995, 267, 367; c) H. Bçnne-
mann, G. A. Braun, Angew. Chem. 1996, 108, 2 12 0 A; ngew. Chem.
Int. Ed. Engl. 1996, 35, 1992; d) H. Bçnnemann, R. M. Richards,
Eur. J. Inorg. Chem. 2001, 2455.
1
999, 111, 2961; Angew. Chem. Int. Ed. 1999, 38, 2791; c) S. Senkan,
K. Krantz, S. Ozturk, V. Zengin, I. Onal, Angew. Chem. 1999, 111,
965; Angew. Chem. Int. Ed. 1999, 38, 2794; d) M.-J. Kang, A.
2
Tholey, E. Heinzle, Rapid Commun. Mass Spectrom. 2001, 15, 1327;
e) H. Wang, Z. Liu, J. Shen, J. Comb. Chem. 2003, 5, 802.
6] a) A. C. Cooper, L. H. McAlexander, D.-H. Lee, M. T. Torres, R. H.
Crabtree, J. Am. Chem. Soc. 1998, 120, 9971; b) E. Reddington, A.
Sapienza, B. Gurau, R. Viswanathan, S. Sarangapani, E. S. Smotkin,
T. E. Mallouk, Science 1998, 280, 1735.
[17] a) J. Kobayashi, Y. Mori, K. Okamoto, R. Akiyama, M. Ueno, T. Ki-
tamori, S. Kobayashi, Science 2004, 304, 1305; b) D. Astruc, F. Lu,
J. R. Aranzaes, Angew. Chem. 2005, 117, 8062; Angew. Chem. Int.
Ed. 2005, 44, 7852.
[18] a) M. T. Reetz, E. Westermann, Angew. Chem. 2000, 112, 170;
Angew. Chem. Int. Ed. 2000, 39, 165; b) M. T. Reetz, J. G. D. Vries,
Chem. Commun. 2004, 1559; c) D. Astruc, Inorg. Chem. 2007, 46,
1884; d) K. Kçhler, W. Kleist, S. S. Prçckl, Inorg. Chem. 2007, 46,
1876.
[19] D. Teschner, E. Vass, M. Hävecker, S. Zafeiratos, P. Schnçrch, H.
Sauer, A. Knop-Gericke, R. Schlçgl, M. Chamam, A. Wootsch, A. S.
Canning, J. J. Gamman, S. D. Jackson, J. Mcgregor, L. F. Gladden, J.
Catal. 2006, 242, 2 6.
[20] B. P. S. Chauhan, J. S. Rathore, T. Bandoo, J. Am. Chem. Soc. 2004,
126, 849.
[
[
7] a) M. T. Reetz, A. Zonta, K. Schimossek, K. Liebeton, K.-E. Jaeger,
Angew. Chem. 1997, 109, 2961; Angew. Chem. Int. Ed. Engl. 1997,
3
3
6, 2830; b) M. T. Reetz, K.-E. Jaeger, Top. Curr. Chem. 1999, 200,
1; c) M. T. Reetz, K.-E. Jaeger, Chem. Eur. J. 2000, 6, 407; d) P. G.
Schultz, R. A. Lerner, Science 1995, 269, 1835; e) F. Zocher, M. M.
Enzelberger, U. T. Bornscheuer, B. Hauer, R. D. Schmid, Anal.
Chim. Acta 1999, 391, 345; f) E. Henke, U. T. Bornscheuer, Biol.
Chem. 1999, 380, 1029.
8] a) A. P. de Silva, H. Q. N. Gunaratne, T. Gunnlaugsson, A. J. M.
Huxley, C. P. McCoy, J. T. Rademacher, T. E. Rice, Chem. Rev. 1997,
[
[21] H. Figge, A. Deege, J. Kçhler, G. Schomburg, J. Chromatogr. 1986,
351, 393.
9
7, 1515; b) G. Zandonella, L. Haalck, F. Spener, K. Faber, F. Pal-
tauf, A. Hermetter, Chirality 1996, 8, 481; c) G. Gauglitz, Curr.
Opin. Chem. Biol. 2000, 4, 351; d) M. Olsen, B. Iverson, G. Geor-
giou, Curr. Opin. Biotechnol. 2000, 11, 331; e) R. P. Hertzberg, A. J.
Pope, Curr. Opin. Chem. Biol. 2000, 4, 445.
9] C. Hoffmann, A. Wolf, F. Schüth, Angew. Chem. 1999, 111, 2971;
Angew. Chem. Int. Ed. 1999, 38, 2800.
[22] W. Lu, B. Wang, K. Wang, X. Wang, J. G. Hou, Langmuir 2003, 19,
5887.
[23] K. Grob, Making and Manipulating Capillary Columns for Gas
Chromatography, Huethig, Heidelberg (Germany), 1986.
[24] a) T. Falk, A. S. -Morgenstern, Chem. Eng. Sci. 1999, 54, 1479;
b) T. D. Vu, A. Seidel-Morgenstern, S. Grüner, A. Kienle, Ind. Eng.
Chem. Res. 2005, 44, 9565.
[
[
[
[
[
10] R. A. Potyrailo, R. J. May, Rev. Sci. Instrum. 2002, 73, 1277.
11] O. Trapp, J. Chromatogr. A 2008, 1184, 160–190.
12] S. J. Haswell, P. Watts, Green Chem. 2003, 5, 240.
13] a) C. de Bellefon, N. Tanchoux, S. Caravieilhes, P. Grenouillet, V.
Hessel, Angew. Chem. 2000, 112, 3584; Angew. Chem. Int. Ed. 2000,
[25] A. Seidel-Morgenstern, J. Chromatogr. A 2004, 1037, 255.
[26] a) T. Haas, J. Gaube, Chem. Eng. Technol. 1989, 12, 45; b) M. Di S-
erio, V. Balato, A. Dimiccoli, L. Maffucci, P. Iengo, E. Santacesaria,
Catal. Today 2001, 66, 403; c) A. Schmidt, R. Schomäcker, Ind. Eng.
Chem. Res. 2007, 46, 1677.
3
9, 3442; b) J. Knight, Nature 2002, 418, 474; c) A. Müller, K. Drese,
H. Gnaser, M. Hampe, V. Hessel, H. Lçwe, S. Schmitt, R. Zapf,
Catal. Today 2003, 81, 377; d) K. Jähnisch, V. Hessel, H. Lçwe, M.
Baerns, Angew. Chem. 2004, 116, 410; Angew. Chem. Int. Ed. 2004,
[27] a) H. Wieland, Chem. Ber. 1912, 45, 484; b) N. Zelinsky, N. Pavlov,
Chem. Ber. 1923, 56B, 1249; c) R. P. Linstead, S. L. S. Thomas, J.
Chem. Soc. 1940, 1127.
4
3, 406; e) D. Belder, M. Ludwig, L.-W. Wang, M. T. Reetz, Angew.
[28] a) J. J. van Deemter, Ind. Eng. Chem. 1953, 45, 1227; b) J. J. van
Deemter, Ind. Eng. Chem. 1954, 46, 2300; c) J. J. van Deemter, F. J.
Zuiderweg, A. Klinkenberg, Chem. Eng. Sci. 1956, 5, 271; d) J. J. van
Deemter, Chem. Eng. Sci. 1961, 13, 143; e) S. J. Hawkes, J. Chem.
Educ. 1983, 60, 393.
Chem. 2006, 118, 2 52 3 A; ngew. Chem. Int. Ed. 2006, 45, 2463; f) D.
Janasek, J. Franzke, A. Manz, Nature 2006, 442, 374; g) S. J. Haswell,
Nature 2006, 441, 705.
[
[
14] O. Trapp, S. K. Weber, S. Bauch, W. Hofstadt, Angew. Chem. 2007,
1
19, 7447; Angew. Chem. Int. Ed. 2007, 46, 7307.
Received: November 12, 2007
Published online: April 2, 2008
15] M. R. Arnold, Chem. Eng. Rev. 1956, 48, 1629.
4666
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 4657 – 4666