Asymmetric hydrogenation on platinum: Nonlinear effect of coadsorbed cinchona alkaloids on enantiodifferentiation

Article abstract of DOI:10.1016/S0021-9517(02)00123-9

Prominent nonlinear effects in enantioselectivity were observed with a transient technique when ethyl pyruvate was hydrogenated over Pt/Al ₂O₃ in the presence of two cinchona alkaloids, which alone afford the opposite enantiomers of ethyl lactate in excess. The changes in reaction rate and ee, detected after injection of the second alkaloid, varied strongly with type and amount of the alkaloid, and with the order of their addition to the reaction mixture. For example, under ambient conditions in acetic acid cinchonidine (CD) afforded 90% ee to (R)-ethyl lactate and addition of equimolar amount of quinidine (QD) reduced the ee to (R)-ethyl lactate only to 88%, though QD alone provided 94% ee to (S)-lactate in a slightly faster reaction. The stronger adsorption of CD on Pt in the presence of hydrogen and acetic acid was proved by UV-vis spectroscopy. The different adsorption strengths result in an enrichment of CD on the Pt surface and also in a crucial difference in the dominant adsorption geometries. CD is assumed to adsorb preferentially via the quinoline rings laying approximately parallel to the Pt surface. In this position it can interact with ethyl pyruvate during hydrogen uptake and control the enantioselectivity. The weaker adsorbing QD adopts mainly a position with the quinoline plane being tilted relative to the Pt surface and these species are not involved in the enantioselective reaction. Competing hydrogenation of the alkaloid, and steric and electronic interactions among the adsorbed species, can also influence the alkaloid efficiency and the product distribution. Hydrogenation of the quinoline rings at low alkaloid concentration resulted in unprecedented swings in the enantiomeric excess.

Full text of DOI:10.1016/S0021-9517(02)00123-9

Journal of Catalysis 216 (2003) 276–287
www.elsevier.com/locate/jcat
Asymmetric hydrogenation on platinum: nonlinear effect of coadsorbed
cinchona alkaloids on enantiodifferentiation
W.-R. Huck, T. Bürgi, T. Mallat, and A. Baiker ∗
Laboratory of Technical Chemistry, Swiss Federal Institute of Technology, ETH–Hönggerberg,CH-8093 Zürich, Switzerland
Received 8 July 2002; revised 2 September 2002; accepted 11 September 2002
Abstract
Prominent nonlinear effects in enantioselectivity were observed with a transient technique when ethyl pyruvate was hydrogenated over
Pt/Al O in the presence of two cinchona alkaloids, which alone afford the opposite enantiomers of ethyl lactate in excess. The changes
2
3
in reaction rate and ee, detected after injection of the second alkaloid, varied strongly with type and amount of the alkaloid, and with the
order of their addition to the reaction mixture. For example, under ambient conditions in acetic acid cinchonidine (CD) afforded 90% ee to
(R)-ethyl lactate and addition of equimolar amount of quinidine (QD) reduced the ee to (R)-ethyl lactate only to 88%, though QD alone
provided 94% ee to (S)-lactate in a slightly faster reaction. The stronger adsorption of CD on Pt in the presence of hydrogen and acetic acid
was proved by UV–vis spectroscopy. The different adsorption strengths result in an enrichment of CD on the Pt surface and also in a crucial
difference in the dominant adsorption geometries. CD is assumed to adsorb preferentially via the quinoline rings laying approximately parallel
to the Pt surface. In this position it can interact with ethyl pyruvate during hydrogen uptake and control the enantioselectivity. The weaker
adsorbing QD adopts mainly a position with the quinoline plane being tilted relative to the Pt surface and these species are not involved in
the enantioselective reaction. Competing hydrogenation of the alkaloid, and steric and electronic interactions among the adsorbed species,
can also influence the alkaloid efficiency and the product distribution. Hydrogenation of the quinoline rings at low alkaloid concentration
resulted in unprecedented swings in the enantiomeric excess.
2003 Elsevier Science (USA). All rights reserved.
Keywords: Enantioselective; Asymmetric; Hydrogenation; Cinchona alkaloids; Platinum; Ethyl pyruvate; Adsorption; Transient behavior
1. Introduction
atives [32,33], oximes [34], and imines [35]. As a result of
intensive research by various groups in the past years, 90–
98% enantiomeric excess (ee) has been achieved in several
reactions.
Chirally modified metals are the most effective hetero-
geneous catalysts for enantioselective hydrogenation reac-
tions (for recent reviews see [1–8]). Originally, this research
field was triggered mainly by Japanese groups [9–13]. Inten-
sively investigated catalysts are the Ni–tartaric acid [14–18],
the Pt–cinchona [11], and the Pd–cinchona [19] and Pd–
vinca alkaloid [7,20] systems. In these catalyst systems the
source of chirality is a strongly adsorbing organic com-
pound that creates a chiral environment on the metal sur-
face for the hydrogenation reaction. Among these chiral
modifiers, cinchona alkaloids have the broadest application
range, including Pt-catalyzed hydrogenation of various acti-
vated ketones [21–28], and the Pd-catalyzed hydrogenation
of α, β-unsaturated carboxylic acids [29–31], pyrone deriv-
An intriguing question is the nature of enantiodifferen-
tiation on the metal surface. Some mechanistic models have
been proposed for interpretation of the metal–reactant–chiral
modifier interaction for the commonly used model reaction,
the hydrogenation of ethyl pyruvate to ethyl lactate (for a
review see [36]). According to present knowledge, good
enantioselectivity can be expected only when the adsorption
mode of reactant and chiral modifier allows their appropriate
interaction during hydrogen uptake in the enantiodiscrimi-
nating step. Obviously, in situ adsorption studies are indis-
pensable for understanding the mechanism of enantioselec-
tion.
Various surface science techniques and H/D exchange
have been applied to clarify the adsorption mode of cin-
chonidine (CD) [37–40] and its “anchoring” moiety quino-
line [39,41,42] on Pt metals. The NEXAFS, STM, XPS,
*
Corresponding author.
E-mail address: baiker@tech.chem.ethz.ch (A. Baiker).
0021-9517/03/$ – see front matter
doi:10.1016/S0021-9517(02)00123-9
2003 Elsevier Science (USA). All rights reserved.

W.-R. Huck et al. / Journal of Catalysis 216 (2003) 276–287
277
LEED, and IR studies were carried out far from the con-
ditions of enantioselective hydrogenation; hence, the obser-
vations cannot easily be correlated with the catalytic results.
In contrast, ATR-IR spectroscopy allows the choice of con-
ditions very close to those of catalytic hydrogenation. A re-
cent study on Pt/Al2O3 in the presence of solvent and hydro-
gen revealed three different adsorption modes of CD [43,44].
These species differ from each other not only by their ad-
sorption strength but also by the adsorption geometries: the
quinoline rings of the alkaloid may adopt a tilted or an ap-
proximately parallel orientation relative to the Pt surface. It
was proposed that the latter π-bonded species is involved
in enantiodifferentiation, in agreement with two mechanistic
models developed for α-ketoester hydrogenation [36,45,46].
Another approach to investigating the role of modifier ad-
sorption is to apply mixtures of modifiers [47]. When the
chiral modifiers alone afford the opposite enantiomers in ex-
cess, the product distribution in the presence of both mod-
ifiers can provide useful hints concerning their adsorption
on the metal surface under reaction conditions, that is, dur-
ing interaction with the reactant and hydrogen. The small
deviations from the expected linear correlation in the hydro-
genation of ethyl pyruvate in the presence of CD–CN and
QN–QD mixtures (Scheme 1) were attributed to different
adsorption strength of the alkaloids [47]. The considerable
nonlinearity detected when CN and NPE (1-{1-naphthyl}-2-
{1-pyrrolidinyl}ethanol)were applied in equimolar amounts
was explained by the low stability against saturation of the
naphthyl group, the anchoring moiety of NPE [48].
A different strategy, based on transient experiments, is
applied in the present study. Ethyl pyruvate hydrogenation
was used as a test reaction. It was started in the presence
of a cinchona alkaloid or its simple derivative (MeOCD),
and after 30 min an equimolar amount of a second alkaloid
was injected that alone affords the opposite enantiomer
of ethyl lactate in excess. The transient behavior in rate
and ee after the disturbance was analyzed to study the
adsorption and interactions of cinchona alkaloids on Pt.
The remarkable nonlinearity in enantioselectivity is traced
mainly to differences in strength and geometry of adsorption
of the two alkaloids.
2. Experimental
Acetic acid (AcOH, Scharlau Chemie), toluene (J.T.
Baker), quinoline (Fluka), 1,2,3,4-tetrahydroquinoline (Flu-
ka), and 5,6,7,8-tetrahydroquinoline (TCI, Japan) were used
as received. Tetrahydrofurane (THF, J.T. Baker) was re-
fluxed over LiAlH4 and distilled before use. Ethyl pyruvate
(Fluka) was carefully distilled in vacuum before use; the
ethyl lactate content after distillation was only 0.1% by GC
analysis.
The commercially available cinchona alkaloids contained
a high proportion of impurities (for the abbreviations, see
Scheme 1): CD (92%; impurities: 1% QN, 7% QD, deter-
mined by HPLC, Fluka), CN (85%; impurity: 15% 10,11-
dihydro-CN, determined by HPLC, Fluka), QD (> 90%;
impurity: < 10% 10,11-dihydro-QD, determined by HPLC,
Fluka), and QN (> 95%; impurity: < 5% 10,11-dihydro-
QN, determined by HPLC, Fluka). Recrystallization of CD
from toluene or 2-propanol was attempted but—according
to NMR analysis—no significant enrichment was achieved.
MeOCD was prepared according to a known method [30].
Elementary analysis and NMR spectroscopic data were in
1
good agreement with the structure of MeOCD. H and 13
C
NMR spectra were measured using a DPX 300 spectrometer.
The 5 wt% Pt/Al2O3 catalyst (Engelhardt 4759; BET
surface area, 100 m2 g−1; metal dispersion after reductive
heat treatment, 0.27, determined by TEM) and Al2O3 (sup-
port of the catalyst Engelhardt 4759) were treated with
flowing H2 for 60 min at 400 ◦C and cooled to room
temperature in hydrogen for 30 min. After flushing with Ar,
the catalyst was transferred to the reactor within 10 min.
Hydrogenation of ethyl pyruvate (1) was carried out in
a magnetically stirred (1000 min−1) 100-ml glass reactor.
In a standard procedure, 20 mg of catalyst was prereduced
in 10 ml of solvent in flowing H2 for 5 min, at 1 bar and
room temperature. Then 1.7 µmol of modifier or modifier
mixture was added in 0.5 ml of solvent. After a 5-min
preadsorption period the reaction was started by introducing
2 ml (16.5 mmol) of ethyl pyruvate. In some experiments
(see Figs. 1, 2, 8, and 9) an additional modifier (1.7 µmol in
0.5 ml of solvent) was added to the reaction mixture after 30
min of reaction time.
Scheme 1. Hydrogenation of ethyl pyruvate over 5 wt% Pt/Al O in the
presence of cinchona alkaloids. CD, MeOCD, and QN afford (R)-lactate in
excess, while CN and QD provide the (S)-enantiomer as the major product.
2
3

278
W.-R. Huck et al. / Journal of Catalysis 216 (2003) 276–287
The conversion and enantioselectivity {ee = |R(%) −
S(%)|} were determined by an HP 6890 gas chromatograph
using a Chirasil-DEX CB column (Chrompack). The es-
timated error in the determination of the ee is less than
0.5%. The actual or incremental ee was calculated as
ꢀee = (ee1y1 − ee2y2)/(y1 − y2), where y represents the
yield to ethyl lactate, and index 2 refers to a sample subse-
quent to sample 1.
UV–vis measurements were used to determine the amount
of CD and QD remaining in solution after the preadsorp-
tion period. Measurements were performed in transmission
mode on a CARY 400 spectrophotometer using a 1-cm-path-
length quartz cuvette. Spectra are given in absorbance units
with neat acetic acid serving as the reference. Pretreatment
of the catalyst and support was performed as described for
catalytic experiments (see above).
Structure optimization of the modifiers was performed at
the AM1 semiempirical level using GAUSSIAN98 [49].
3. Results
3.1. Enantioselective hydrogenation of ethyl pyruvate:
transient experiments with CD and QD
Under standard conditions in THF, hydrogenation of
ethyl pyruvate over 5wt% Pt/Al2O3 modified by CD af-
forded 82% ee to (R)-ethyl lactate. When QD was ap-
plied as chiral modifier, the (S)-enantiomer formed in 52%
excess. To study the competition between these modifiers
in the enantiodifferentiation, the reaction was started with
Pt/Al2O3 modified by CD. After a 30-min reaction time one
molar equivalent QD (related to CD) was injected into the
reaction mixture (Fig. 1). Clearly, the reaction rate was not
affected by QD and the enantioselectivity decreased only by
about 2% after addition of the second modifier. In the control
experiment the hydrogenation was started with Pt/Al2O3
modified by QD and a solution containing one equivalent
CD was injected after 30 min. Addition of CD resulted in
an immediate decrease of ee and enhancement of reaction
rate. The time-dependent changes of the calculated actual or
incremental ee (ꢀee) revealed that CD controlled the enan-
tiodiscriminating step on the Pt surface within 2–3 min after
its addition. Apparently, equivalent amount of CD rapidly
replaced QD on the Pt surface in the enantioselective hydro-
genation reaction, indicating stronger adsorption of CD.
Fig. 1. Transient behavior of ethyl pyruvate hydrogenation over 5 wt%
Pt/Al O in tetrahydrofurane. Influence of the addition of one molar
equivalent QD or CD after a 30-min reaction time to the reaction mixture
containing CD or QD, respectively. (a) Enantiomeric excess (ee) and
incremental ee (ꢀee) vs time; (b) conversion vs time. Standard conditions.
2
3
3.2. Influence of solvents and acids on the competition
between the alkaloids
The above two experiments were repeated in toluene and
acetic acid. In both solvents, addition of QD to the reaction
mixture barely influenced the performance of Pt/Al2O3
modified by CD, similarly to the reaction carried out in
THF (Fig. 1). More interesting is the reverse case, the
replacement of QD by CD (Fig. 2). The time-dependent
Fig. 2. Solvent effect on the exchange of QD by CD during hydrogenation
of ethyl pyruvate over 5 wt% Pt/Al O . Standard conditions, addition of
2
3
one equivalent CD after a 30-min reaction time.

W.-R. Huck et al. / Journal of Catalysis 216 (2003) 276–287
279
variations in the actual or incremental ee (ꢀee) indicate that
in toluene and THF the changes are faster than in acetic acid
by a factor of about 10. In acetic acid the ethyl pyruvate
were similar in the presence of QD or CD alone (0.39
and 0.36 molh−1 (gcat)−1, respectively) and no significant
change in rate was observed by injection of the second
modifier.
The differences in the exchange rate might be correlated
with the solvent polarity. According to the mostly used
N
empirical solvent polarity scale [50,51], toluene (ET
=
N
0.1) and THF (ET = 0.21) are weakly polar solvents and
N
acetic acid (ET = 0.65) is more polar than the commonly
N
used polar solvents dimethylformamide (ET = 0.4) or
N
acetonitrile (ET = 0.46). Furthermore, polar compounds—
and particularly carboxylic acids—adsorb strongly on Pt and
can block the active sites from the catalytic reaction [52,53].
However, further experiments are needed to confirm these
possibilities.
Another feasible explanation is that formation of an
alkaloid–acid ion pair leads to slow exchange in acetic acid.
We have shown recently that in the presence of a carboxylic
acid solvent or additive CD forms ion pairs [54,55] and
these ion pairs may be the real chiral modifiers of Pt or
Pd [56,57]. The quinuclidine N and quinoline N atoms of
CD are medium and weakly basic (pKa = 10.0 and 5.8,
respectively). Excess acetic acid (pKa = 4.7) protonates
the quinuclidine N while trifluoroacetic acid (pKa = 0.3)
protonates both N atoms [58,59].
To clarify the role of acidity, the experiment in THF
(Fig. 2) was repeated and either 100 eq. AcOH or 25 eq.
trifluoroacetic acid (related to both modifiers) was added
to the reaction mixture. None of the acid additives induced
a significant change in the exchange rate of the modifiers.
Hence, protonation and building of a cyclic ion pair cannot
be the reason for the slow exchange of modifiers on the Pt
surface when using acetic acid as a solvent.
Fig. 3. UV–vis spectra of CD and QD (a), and quinoline, 1,2,3,4-
tetrahydroquinoline, and 5,6,7,8-tetrahydroquinoline. (b) Concentrations:
0.17 mmolL in acetic acid.
−1
3.3. UV–vis study of cinchona alkaloid adsorption
A possible explanation for the striking changes in ee
presented in Figs. 1 and 2 is the different adsorption strength
of the alkaloid modifiers on Pt. As no data are available in
the literature, the relative adsorption strength of CD and QD
in the presence of hydrogen was analyzed by an indirect
UV–vis spectroscopic method. The spectra of CD and QD
in acetic acid are shown in Fig. 3a. The presence of the
methoxy group has a strong effect on the electronic structure
of the quinoline moiety, that is, the chromophore moiety of
the molecule. The resulting shift in the UV spectrum allows
the quantitative analysis of CD and QD mixtures. The UV
spectrum of quinoline (Fig. 3b) resembles very much that of
CD, confirming that the UV spectrum of CD is determined
by the quinoline chromophore.
Fig. 4. Illustration to determination of the alkaloid concentration by UV–vis
−1
spectroscopy. Spectrum a represents the solution of 0.034 mmolL CD
−1
and 0.034 mmolL QD in acetic acid before treatment. Spectrum b was
measured after a 5-min hydrogenation of the solution over Pt/Al O (stan-
2
3
−1
dard conditions). Spectrum c is the sum of the spectra of 0.017 mmolL
−1
Fig. 4 illustrates determination of the individual concen-
tration of the modifiers remaining in solution after filter-
ing off the catalyst or support. Curve a represents the spec-
CD and 0.024 mmolL
QD. Spectrum d is the difference between the
measured (b) and synthesized (c) spectra and indicates the presence of hy-
drogenated alkaloids.

280
W.-R. Huck et al. / Journal of Catalysis 216 (2003) 276–287
Scheme 2. Hydrogenation of the vinyl group and the quinoline rings of cinchonidine (CD) over 5 wt% Pt/Al O in acetic acid.
2
3
trum of an equimolar mixture of CD and QD containing
0.034 mmolL−1 from each alkaloid. After stirring the al-
kaloid solution for 5 min in hydrogen over Pt/Al2O3, the
alkaloid concentration in solution decreased (curve b). Next,
a synthetic spectrum, c, was generated by linear combina-
tion of the spectra of individual alkaloids being present in
the appropriate concentration, in this case by summing up
the spectra of 0.017 mmolL−1 CD and 0.024 mmolL−1 QD.
This process was accepted when the difference between the
measured and synthesized spectra did not exhibit the typical
absorption bands of CD and QD at 315 and 338 nm, respec-
tively. Curve d is the difference between spectra b and c,
and indeed this spectrum does not show the typical features
of CD and QD. The difference spectrum, however, shows a
band at around 275 nm, which belongs to modifiers possess-
ing a partially hydrogenated quinoline ring (Scheme 2). The
latter band was found in all experiments where the catalyst
and modifier were pretreated with hydrogen. Since this band
does not overlap with the characteristic features of CD and
QD at 300–350 nm, its presence does not hinder quantifica-
tion of CD and QD in solution.
The drop in alkaloid concentration detected by UV–vis
may be attributed to adsorption on the Al2O3 support and
on Pt, and to partial or complete saturation of the quinoline
rings. It is commonly assumed that saturation of the quino-
A test with CD–QD mixtures of known concentrations
revealed that the reliability of the method is high: at
a concentration of 0.034 mmolL−1 the relative error is
estimated to be less than 2%. (A 2% change in the modifier
concentration, used to synthesize the spectra, resulted in
the appearance of clear features of the modifiers in the
difference spectra.)
3.4. Competitive adsorption of CD and QD on Pt/Al2O3
Fig. 5 illustrates the competitive adsorption of CD and
QD during the catalyst premodification procedure. Equal
amounts of the two alkaloids were dissolved in acetic acid
and the slurry containing Pt/Al2O3 was stirred in hydrogen
for 5 min (preadsorption step in the standard procedure).
In Fig. 5a the amount of alkaloids remaining in solution is
plotted as a function of the initial total concentration of the
alkaloids. The dashed line indicates the initial concentration
for each modifier. Clearly, less CD than QD can be detected
in solution after contact with the catalyst.
Fig. 5. Competitive adsorption of alkaloids on Pt in the presence of
hydrogen. (a) Concentration of CD and QD remaining in acetic acid after
preadsorption of equimolar mixtures of the alkaloids (5 min in hydrogen
over Pt/Al O , standard conditions). Dashed line: initial concentration (no
2
3
loss). The amount of adsorbed and/or hydrogenated alkaloids (b) and their
molar ratio (c).

W.-R. Huck et al. / Journal of Catalysis 216 (2003) 276–287
281
line rings of the alkaloid leads to weaker adsorption on Pt
[8,60,61] and also on Pd [32]. Thus, the hydrogenated (de-
structed) alkaloid is expected to be replaced by intact mole-
cules from solution. Adsorption on the catalyst support was
investigated by repeating the experiments with the support
alone (after the same pretreatment procedure as that used for
the catalyst). At the highest concentration (0.17 mmolL−1
)
the amount of alkaloid adsorbed on the support was less
than 0.0017 mmolL−1, corresponding to about 4% of the
amount (0.041 mmolL−1) removed from solution by pre-
reduced Pt/Al2O3. At lower initial concentration the frac-
tion adsorbed on the support increased. At a concentration
of 0.0068 mmolL−1 the fraction adsorbed on the support
was about 30% of the total amount removed from solution.
It is important that no clear preference of adsorption of CD
or QD on the support could be observed. A possible explana-
tion is that on Pt, the alkaloids adsorb via the quinoline ring
system, while on alumina, the basic quinuclidine N atom is
assumed to be the favored “anchoring” moiety.
The adsorption on the support was taken into account,
and Fig. 5b shows the amount of each modifier that either
adsorbed on Pt or had its quinoline rings saturated. In Fig. 5c
the same amount of alkaloids are presented as the molar ratio
of CD to QD. At low concentrations this ratio is close to
one and increases with the initial concentration up to 2.5.
Hence, the preferential disappearance of CD (by stronger
adsorption and/or faster hydrogenation) on Pt is important
only at sufficiently high concentrations.
3.5. Hydrogenation of the alkaloids
Fig. 6. Hydrogenation of CD (a) and QD (b) over Pt/Al O in acetic
2
3
acid, followed by UV–vis spectroscopy. Spectra of the initial solutions
(0.068 mmolL ) are shown as reference.
Hydrogenation of the aromatic rings of CD (Scheme 2)
and QD disturbs the quantitative determination of alkaloid
adsorption by UV–vis spectroscopy. The importance of
hydrogenation of the aromatic rings of modifiers during
the adsorption study is addressed next. It is known that
hydrogenation of the vinyl group is much faster than
that of the aromatic rings, but this transformation barely
influences the enantioselectivity [61–63]. Hydrogenation
of the aromatic and heteroaromatic rings of the quinoline
moiety of CD and QD has been followed by UV–vis
spectroscopy (Fig. 6). After 45 min no intact alkaloid
molecule could be detected in acetic acid. In both cases
a band at around 275 nm appeared, which is assigned to
species with the quinoline moiety partially hydrogenated
(compare to the spectra of the reference compounds in
Fig. 3b). The intensity of this peak is higher for QD than
for CD.
−1
An important observation is that the rates of disappear-
ance (partial hydrogenation) of the original alkaloid mole-
cules are about the same, independent of the presence (QD)
or absence (CD) of the methoxy group on the quinoline ring.
This is illustrated in Fig. 7, where the absorbance at 315 nm
for CD and at 338 nm for QD, as well as the absorbance of
the signals at 275 nm, are plotted as a function of reaction
time. The markers at zero time refer to the reference solu-
Fig. 7. Kinetic analysis of the hydrogenation of CD and QD in acetic acid,
derived from the data in Fig. 6. The actual concentrations are determined
from the absorbances at 315 nm (CD), 338 nm (QD), and 275 nm (partial
hydrogenation products). The two values at t = 0 refer to the reference
solution and the sample, which was only manually shaken for a few
seconds.

282
W.-R. Huck et al. / Journal of Catalysis 216 (2003) 276–287
Table 1
Influence of prehydrogenation of CD (in the absence of reactant) on the rate
a
and enantioselectivity of ethyl pyruvate hydrogenation
b
Prehydrogenation
period (min)
Rate
(molh
ee
−1 −1
g
)
(%)
0
30
60
0.32
0.30
0.27
88 (R)
87 (R)
79 (R)
a
−1
Standard conditions, 0.068 mmolL CD in acetic acid.
b
Determined after a 60-min reaction (at ca. 30–35% conversion).
tions (no treatment), and to samples that were not stirred
but only shaken manually for a few seconds. Curve fitting
and extrapolation to zero time indicated practically the same
initial rates (3.6 and 3.7 µmolh−1 for CD and QD, respec-
tively, under standard conditions, at 0.068 mmolL−1 initial
modifier concentration). In contrast, Fig. 5 shows that from
equimolar mixtures of the two modifiers CD disappears pref-
erentially.
To clarify the impact of saturation of the quinoline rings
of the alkaloids, we carried out the hydrogenation of ethyl
pyruvate after different preadsorption (prehydrogenation)
periods (Table 1). In the standard procedure the preadsorp-
tion step, in which the reduced catalyst is stirred with the
alkaloid modifier in hydrogen before addition of ethyl pyru-
vate, is 5 min. Here this period was varied between 0 and 60
min. The initial concentration of CD was the same as that
used in the UV–vis study (Figs. 6 and 7). After a 30-min
prehydrogenation of CD in the absence of ethyl pyruvate,
the ee was 87%, barely smaller than without any prehydro-
genation (Table 1). According to the UV–vis study in Fig. 6a,
after a 30-min prehydrogenation only 4% of initial amount
of CD was present in solution. Prehydrogenation for 60 min
resulted in complete conversion of the quinoline rings of CD
(Fig. 6a) but in the subsequent reaction the ee to (R)-lactate
dropped only by 9%. Similarly, the changes in initial rate of
pyruvate hydrogenation were also small. Note that the par-
tial loss of ee due to competitive hydrogenationof CD during
enantioselective hydrogenations over Pt [8,60,61,64] and Pd
[32,65] is a well-known phenomenon.
According to NMR analysis, at 36% saturation of the
quinoline rings of CD in acetic acid the ratio of homoaro-
matic and heteroaromatic hydrogenation products (CD∗ to
CD∗∗ in Scheme 2) was 2.5 to 1. (Note that in nonacidic
solvents only CD∗ formed.) The presence of heteroaromatic
hydrogenation products of CD and QD was confirmed by
UV–vis spectroscopy in Fig. 6 but the method is not sensi-
tive to the homoaromatic hydrogenation products (Fig. 3b).
It is known [66] that hydrogenation of the heteroaromatic
ring of quinoline and its derivatives is remarkably faster on
most catalysts and in most solvents, but in strongly acidic
medium on Pt the homoaromatic ring is reduced almost ex-
clusively [67,68].
Fig. 8. Influence of modifier concentration on the transient behavior of
ethyl pyruvate hydrogenation over Pt/Al O . (a) Addition of one molar
equivalent CD after a 30-min reaction carried out in the presence of QD;
(b) addition of one molar equivalent QD after a 30-min reaction carried
out in the presence of CD. Standard conditions; acetic acid; amounts of the
modifiers: 1.7, 0.17, and 0.017 µmol.
2
3
3.6. Effect of modifier hydrogenation on the transient
behavior of pyruvate hydrogenation
Saturation of the quinoline rings of the modifiers has a
strong influence on the transient behavior of ethyl pyruvate
hydrogenation. This effect is best illustrated by repeating
the transient experiment shown in Fig. 2 with different
amounts of modifiers in acetic acid. Considering at first the
highest amount of modifier (1.7 µmol; Fig. 8b), the initial
incremental ee of 90% to (R)-lactate in the CD-modified
reaction decreased by only 2% after addition of equivalent
amount of QD. On the other hand QD, which provided 94%
ee to (S)-lactate before addition of CD, was replaced by CD
very slowly (Fig. 8a). According to UV–vis analysis, 76%
of the modifier was intact after the 5-min preadsorption time
and we assume that further hydrogenation of the quinoline
rings is considerably slower after addition of large excess of
ethyl pyruvate.

W.-R. Huck et al. / Journal of Catalysis 216 (2003) 276–287
283
Starting the experiment with 10-fold lower modifier
concentration (0.17 µmol CD or QD), only about 33% of
the modifiers were still intact after a 5-min preadsorption, as
indicated by the UV–vis analysis. This amount was enough
to maintain the same initial ee (90% with CD and 95% with
QD). However, the exchange of QD by CD was faster and
the final ee was lower (Fig. 8a). A more interesting behavior
was observed when the reaction was started with 0.17 µmol
CD, and after 30 min 0.17 µmol QD was added (Fig. 8b).
The incremental ee dropped rapidly and in the extreme
point 18% actual ee to the (S)-enantiomer was produced,
and then it changed back to the (R)-enantiomer in excess
(58%). A possible (simplified) explanation for this unusual
transient behavior is that the quinoline rings, the “anchoring”
part of CD, were mostly converted to partially hydrogenated
rings (Scheme 2) by the time QD was added to the reaction
mixture. CD∗ and CD∗∗ are assumed to adsorb much more
weakly than the original alkaloid and, apparently, more
weakly than QD. Hence, the added QD rapidly replaced the
hydrogenationproducts of CD on the Pt surface. Explanation
for the next part, the shift back to (R)-lactate as the main
enantiomer, might be that QD was slowly hydrogenated and
the partially saturated products were replaced by CD∗ and
CD∗∗. It is also possible that intact CD, adsorbed on the
alumina support, migrated to the Pt particles and replaced
QD and its partially hydrogenated derivatives.
A further 10-fold decrease in the initial amount of
modifier (0.017 µmol) led to significant changes in the initial
ee. UV–vis measurements revealed that at this concentration
no original (intact) modifier was present in solution after
the 5-min preadsorption period. The exchange of this small
amount of hydrogenated QD by CD (Fig. 8a) was very
fast and around 80% ee to (R)-lactate was obtained after
2–3 min. The subsequent slow decrease in incremental ee
by about 3–4% is attributed to the slow hydrogenation
of CD. The reverse case, addition of 0.017 µmol QD
to a hydrogenation reaction started with 0.017 µmol CD
(Fig. 8b), showed a behavior similar to that described above
for 0.17-µmol modifier. Here, due to the lower modifier
concentrations, the changes in incremental ee were faster
and an extreme value of 23% ee to (S)-lactate was reached
15 min after addition of QD.
Fig. 9. Transient behavior of ethyl pyruvate hydrogenation over Pt/Al O
2
3
after addition of a second modifier, which provides the opposite enantiomer
in excess. (a) Addition of one molar equivalent QD to reactions that were
started with CD, MeOCD, or QN; (b) addition of one molar equivalent
CD, MeOCD, or QN to reactions that were started with QD. Standard
conditions; acetic acid; 1.7 µmol from each modifier.
initially used modifier was dominant in controlling the enan-
tioselection and (R)-lactate formed in excess.
Fig. 9b presents the reverse case, the transient behavior
of those reactions, which were started with QD as modifier,
and after 30 min one equivalent of CD, MeOCD, or QN
was added. The initial ee was 94% to (S)-lactate and after
addition of the second modifier, which alone provides the
opposite enantiomer, the incremental ee successively shifted
to (R)-lactate as the major product. Another interesting
point is the outcome of the reaction in the presence of
equimolar amounts of QD and QN. Independent of the order
of addition, i.e., whether the reaction was started with QD
or QN, the final ee approaches around 50% ee to (R)-
lactate. These two alkaloids possess identical “anchoring”
moieties (6-methoxy-quinoline fragment, Scheme 1); thus
the difference in their efficiency as chiral modifiers of Pt
should be attributed to their different configuration.
3.7. Comparison of different cinchona alkaloids
The transient experiments in Fig. 1 were repeated with
different cinchona alkaloids in acetic acid using high modi-
fier concentration (0.17 mmolL−1) in order to minimize the
effect of modifier hydrogenation. In Fig. 9a the hydrogena-
tion of ethyl pyruvate was started with Pt/Al2O3 modified
by CD, MeOCD, or QN. MeOCD and QN provided 95%
initial ee to (R)-lactate, higher than that achieved with CD
(90% ee). After a 30-min reaction time one equivalent QD
was added to the reaction mixture. The incremental ee to
(R)-lactate in the reactions modified by CD, MeOCD, or QN
decreased by 2, 27, and 40%, respectively, but in all cases the
A small deviation from linearity has already been found
for the QD–QN and CD–CN couples when using mixtures

284
W.-R. Huck et al. / Journal of Catalysis 216 (2003) 276–287
of these modifiers [47]. The small shift in ee related to
the calculated value was attributed to stronger adsorption
of the alkaloid affording (R)-lactate in excess. The present
transient experiments confirmed this behavior also for the
CD–CN couple: the incremental ee approached very slowly
about 20% to (R)-lactate, independent of the modifier which
was added to the slurry after 30 min (not shown). It is
interesting that the CN–MeOCD combination was the only
example in this study where equimolar mixtures of modifiers
afforded racemic ethyl lactate (about 90 min after injection),
though the modifiers alone provide the opposite enantiomers
in excess.
Fig. 10. Simplified illustration of the enantioselective hydrogenation of
ethyl pyruvate over a hypothetical flat Pt surface modified by an equimolar
mixture of CD and QD. Ethyl pyruvate interacts with the strongly adsorbed
CD; the carbonyl groups of the reactant and the quinoline rings of the
alkaloid lay almost parallel to the surface. Hydrogen uptake in this position
leads to the formation of (R)-lactate. QD is adsorbed more weakly via its
quinoline N-atom, adopting a tilted position of the aromatic rings relative
to the surface (“spectator” species for enantioselective hydrogenation).
4. Discussion
Hydrogenation of ethyl pyruvate over Pt/Al2O3 has
been studied in the presence of two cinchona alkaloids,
which alone provide the opposite enantiomers of ethyl
lactate in excess (Scheme 1). A transient method has been
applied: the reaction was started with one alkaloid and after
establishment of the rate and enantioselectivity a second
alkaloid in an equimolar amount was injected. The changes
in rate, ee, and incremental ee, observed after addition of the
second modifier, depended dominantly on the alkaloid pair
and the order of their addition but the alkaloid concentration
and the solvent also played some role (Figs. 1, 2, 8, and 9).
though it cannot fully account for the observed large nonlin-
earity in enantioselectivity.
To explain this nonlinearity we have to take into account
also the adsorption mode of alkaloid on the Pt surface. In
situ ATR-IR spectroscopy in the presence of solvent and hy-
drogen revealed three different adsorption modes [43,44].
A strongly adsorbed π-bonded species that adsorbed via its
quinoline rings approximately parallel to the Pt surface was
assumed to interact with ethyl pyruvate in the enantiodis-
criminating step. This position of the reactant is necessary
for the uptake of two hydrogen atoms from the Pt surface.
The other two alkaloid species adsorbed weaker and adopted
a position in which the quinoline rings of the alkaloid were
tilted relative to the Pt surface. In this adsorption mode the
alkaloid may be considered a spectator species, which serves
as a reservoir but is not involved in the crucial reactant–
modifier interaction.
We propose that when CD–QD modifier mixtures are
applied in the hydrogenation of ethyl pyruvate, most of the
alkaloid that adsorbs stronger (CD) will occupy a position in
which the quinoline rings are bound via π-bonding to the Pt
surface (Fig. 10). The weaker adsorbing component (QD)
will mainly adopt a “tilted” position. By this, the overall
adsorption free energy is minimized.
4.1. The role of strength and geometry of alkaloid
adsorption
For the interpretation of the results we can start from the
remarkable nonlinearity uncovered in the case of the CD–
QD alkaloid pair. In acetic acid QD is slightly more effective
than CD: both the reaction rate (0.39 vs 0.36 molh−1
(gcat)−1) and the enantioselectivity (94% to (S)-lactate vs
90% to (R)-lactate) were higher with QD when using the
modifiers alone. Still, after addition of QD to the reaction
mixture already containing CD, the ee to (R)-lactate dropped
only by 2% and the effect on the reaction rate was negligible
(Fig. 1). Apparently, QD has barely any influence on the
enantiodifferentiating step in the presence of equimolar
amounts of CD. This conclusion has been confirmed by
repeating the experiment with the reverse order of alkaloid
addition.
A feasible explanation for this nonlinearity is the differ-
ent adsorption strength of the alkaloids on Pt. The adsorption
of CD–QD mixtures on alumina-supported Pt from acetic
acid solutions in the presence of hydrogen was measured
by UV–vis spectroscopy. The values were corrected with the
amounts adsorbed on the support but partial hydrogenation
of the aromatic rings disturbed the analysis. At the high-
est total modifier concentration (0.17 mmolL−1) the con-
tribution of modifier hydrogenation was the smallest and the
molar ratio of CD to QD on the Pt surface was about 2.5.
This result supports the preferential adsorption of CD on Pt,
4.2. Steric and electronic interactions
A refined consideration has to take into account also
the various steric effects and electronic interactions among
the surface species present on Pt. During reaction the
surface is covered by the reactant, modifier(s), hydrogen,and
solvent, but also by degradation products originated from
solvent [43] and reactant [69,70]. Interaction of the modifier
or a mixture of two modifiers with the other adsorbed species
may increase or decrease the nonlinearity, but this effect
cannot be reliably estimated yet. Let us consider only one
point, the possible modifier–modifier interactions on the
Pt surface. The importance of these interactions (“dimer
formation” [71]) has been proposed and it can be expected
also from the well-known nonlinear effects in homogeneous

W.-R. Huck et al. / Journal of Catalysis 216 (2003) 276–287
285
of ethyl pyruvate on Pt/Al2O3 modified by mixtures of CN
and NPE [48]. Rapid hydrogenation of the naphthyl rings
of NPE resulted in a remarkable nonlinearity, as the hydro-
genated modifier could not compete with CN for the active
sites on Pt.
4.4. Comparison to nonlinear effects in homogeneous
asymmetric catalysis
Finally, it should be stressed that the observed nonlinear
behavior cannot be related to the classical nonlinear effect
(NLE), as originally defined by Kagan in his pioneer work
in asymmetric homogeneous catalysis (see [72,73] and refer-
ences therein). Their basic observation is that in many enan-
tioselective reactions the ee in the product is not proportional
to the ee in the chiral auxiliary or ligand. This asymmetric
amplification or depletion is attributed to molecular inter-
actions (associations) involving two enantiomers of the (in-
pure) auxiliary.
In contrast, our observations are related to hydrogenation
reactions in which diastereomer pairs (e.g., QD and QN) or
even chemically different chiral modifiers (e.g., CD and QD)
are applied. Besides, the nonlinear behavior is mainly traced
to differences in the strength and geometries of modifier
adsorption on Pt and only partly to modifier–modifier
interactions. It remains a challenge for future research
to investigate the importance of NLE in heterogeneous
catalysis, that is, carrying out a reaction in the presence of
two enantiomers of a chiral modifier.
Fig. 11. Energetically optimized structures of QN and QD.
asymmetric catalysis (see Kagan’s pioneer work in this field
[72,73]). Figure 11 illustrates the energetically optimized
structure of the diastereomeric pair QN and QD located
over a hypothetical flat metal surface. In this case the so-
called anchoring moieties of the alkaloids are the same; thus,
their different adsorption strength on Pt cannot simply be
attributed to the methoxy functional group on the quinoline
rings, as was the case for the CD–QD pair. Still, it is easily
understandable that the different structure of the alkaloids
may have a strong influence on their mutual interaction, on
the interactions with other surface species, and thus also on
the adsorption strength. This is the explanation for the ca.
50% ee to (R)-lactate obtained with equimolar mixtures of
these modifiers (Fig. 9). Note that the smaller nonlinearity
reported earlier by Wells’ group [47] is presumably due
to the different reaction conditions applied (high pressure,
other solvent).
4.3. Influence of hydrogenation of cinchona alkaloids
5. Conclusions and outlook
We have to consider also the influence of competing
hydrogenation of the modifier (Scheme 2) on the adsorp-
tion strength. This phenomenon and its unfavorable conse-
quences on enantioselectivity when using either cinchona-
modified Pt [8,60,61] or Pd [32,65] have been thoroughly
discussed in the recent literature. It is commonly assumed
(though not yet proved) that partial hydrogenationof the aro-
matic ring system of the modifier leads to weaker adsorption
on Pt and Pd, and to a successive loss of ee. Here we proved
(Figs. 6 and 7) that the rate of saturation of one of the quino-
line rings is the same for CD and QD, at least under the con-
ditions applied. That is, partial hydrogenation of the alka-
loid as an explanation can be excluded for the strong non-
linear effect depicted in Fig. 1. Still, partial transformation
of the alkaloids can have a dramatic effect on the transient
behavior of ethyl pyruvate hydrogenation when the modi-
fiers are applied in low concentration. The unusual shift of ee
from (R)- to (S)-lactate and than back to (R)-lactate (Fig. 8)
is attributed to the combined effects of different adsorption
strengths of the modifiers and their hydrogenated products.
A distinctly different case applies when the two mod-
ifiers possess anchoring moieties with substantially differ-
ent resistance against hydrogenation. This is the explanation
for a former observation: the high-pressure hydrogenation
Pairs of cinchona alkaloids, which afford opposite enan-
tiomeric products when used in the asymmetric hydrogena-
tion of ethyl pyruvate, have been applied in transient experi-
ments. Hydrogenation in the presence of equimolar amounts
of cinchonidine (CD) and quinidine (QD) provided 88% ee
to (R)-lactate, as compared to 90% ee to (R)-lactate with CD
alone. The strong nonlinear effect is traced to differences in
the adsorption strength, and partly to steric and electronic
effects.
The observed nonlinearity is of great practical importance
in heterogeneous asymmetric synthesis, as cheap but low pu-
rity chiral compounds can be used as effective chiral mod-
ifiers. In the present study, the commercial CD contained a
considerable amount of QD (in our sample, 7%) and smaller
amount of quinine (QN) (1%). CD and QN afford the op-
posite enantiomer in excess compared to that induced by
QD in all known hydrogenation reactions catalyzed by Pt
or Pd. Our observations can rationalize how the commercial
CD–QD–QN mixture can afford 94–98% ee in the hydro-
genation of 2-pyrones [33], α-ketoacetals [26,27], α, α, α-
trifluoromethyl ketones [74], and α-ketoesters [75]. Still, in
the design of chirally modified metal catalysts for new reac-
tions special care has to be given to the application of pure

286
W.-R. Huck et al. / Journal of Catalysis 216 (2003) 276–287
modifiers, because other catalyst systems may be more sen-
sitive to the optical purity of the modifier.
[31] I. Kun, B. Török, K. Felföldi, M. Bartók, Appl. Catal. A 203 (2000)
71.
[32] W.-R. Huck, A. Mallat, A. Baiker, J. Catal. 193 (2000) 1.
[33] W.R. Huck, T. Mallat, A. Baiker, New J. Chem. 26 (2002) 6.
[34] K. Borszeky, T. Mallat, R. Aeschimann, W.B. Schweizer, A. Baiker,
J. Catal. 161 (1996) 451.
Albeit a single chiral modifier is generally used to transfer
the chiral information, it could be challenging to explore
whether a mixture of two different modifiers can afford
better enantioselectivities in some reactions. This possibility
remains the target of future studies.
[35] G. Szöllösi, I. Kun, M. Bartók, Chirality 13 (2001) 619.
[36] A. Baiker, J. Mol. Catal. A 163 (2000) 205.
[37] K.E. Simons, A. Ibbotson, P. Johnston, H. Plum, P.B. Wells, J. Ca-
tal. 150 (1994) 321.
[38] G. Bond, P.B. Wells, J. Catal. 150 (1994) 329.
[39] A.F. Carley, M.K. Rajumon, M.W. Roberts, P.B. Wells, J. Chem. Soc.
Faraday Trans. 91 (1995) 2167.
Acknowledgments
Financial support by the Swiss National Science Founda-
tion is greatly acknowledged. Thanks are also due to Engel-
hard (Italy) for providing pure alumina support material.
[40] T. Evans, A.P. Woodhead, A. Gutierrez-Sosa, G. Thornton, T.J. Hall,
A.A. Davis, N.A. Young, P.B. Wells, R.J. Oldman, O. Plashkevych, O.
Vahtras, H. Agren, V. Carravetta, Surf. Sci. 436 (1999) L691.
[41] J.M. Bonello, R.M. Lambert, Surf. Sci. 498 (2002) 212.
[42] J.M. Bonello, R. Lindsay, A.K. Santra, R.M. Lambert, J. Phys. Chem.
B 106 (2002) 2672.
[43] D. Ferri, T. Bürgi, J. Am. Chem. Soc. 123 (2001) 12074.
[44] D. Ferri, T. Bürgi, A. Baiker, Chem. Commun. (2001) 1172.
[45] K.E. Simons, P.A. Meheux, S.P. Griffiths, I.M. Sutherland, P. Johnston,
P.B. Wells, A.F. Carley, M.K. Rajumon, M.W. Roberts, A. Ibbotson,
Recl. Trav. Chim. Pays-Bas 113 (1994) 465.
[46] O. Schwalm, B. Minder, J. Weber, A. Baiker, Catal. Lett. 23 (1994)
271.
[47] K.E. Simons, P.A. Meheux, A. Ibbotson, P.B. Wells, Stud. Surf. Sci.
Catal. 75 (1993) 2317.
[48] M. Schürch, T. Heinz, R. Aeschimann, T. Mallat, A. Pfaltz, A. Baiker,
J. Catal. 173 (1998) 187.
References
[1] A. Baiker, J. Mol. Catal. A 115 (1997) 473.
[2] H.U. Blaser, H.P. Jalett, M. Müller, M. Studer, Catal. Today 37 (1997)
441.
[3] P.B. Wells, A.G. Wilkinson, Top. Catal. 5 (1998) 39.
[4] G.J. Hutchings, Chem. Commun. (1999) 301.
[5] T. Sugimura, Catal. Surv. Jpn. 3 (1999) 37.
[6] T. Osawa, T. Harada, O. Takayasu, Top. Catal. 13 (2000) 155.
[7] A. Tungler, G. Fogassy, J. Mol. Catal. A 173 (2001) 231.
[8] M. von Arx, T. Mallat, A. Baiker, Top. Catal. 19 (2002) 75.
[9] Y. Izumi, Adv. Catal. 32 (1983) 215.
[49] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb,
J.R. Cheeseman, V.G. Zakrzewski, J.A. Montgomery, R.E. Stratmann,
J.C. Burant, S. Dapprich, L.M. Millam, A.D. Daniels, K.N. Kudin,
M.C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi,
B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, G.A.
Petersson, P.Y. Ayala, Q. Cui, K. Morokuma, D.K. Malick, A.D.
Rabuck, K. Raghavachari, J.B. Foresman, J. Cioslowski, J.V. Ortiz,
A.G. Baboul, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I.
Komaromi, R. Gomperts, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-
Laham, C.Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe,
P.M.W. Gill, B. Johnson, W. Chen, M.W. Wong, J.L. Andres, C.
Gonzalez, M. Head-Gordon, E.S. Replogle, J.A. Pople, A.7 ed.,
Gaussian Inc., Pittsburgh, PA, 1998.
[10] A. Tai, T. Harada, in: Y. Iwasawa (Ed.), Tailored Metal Catalysts,
Reidel, Dordrecht, 1986, p. 265.
[11] Y. Orito, S. Imai, S. Niwa, J. Chem. Soc. Jpn. (1979) 1118.
[12] W.M.H. Sachtler, in: R.L. Augustine (Ed.), Catalysis of Organic
Reactions, Dekker, New York, 1985, p. 189.
[13] M. Bartók, J. Czombos, K. Felföldi, L. Gera, G. Göndös, À. Molnàr,
F. Notheisz, I. Pàlinkó, G. Wittman, À.G. Zsigmond, Stereochemistry
of Heterogeneous Metal Catalysts, Wiley, Chichester, 1985.
[14] M.A. Keane, Can. J. Chem. 72 (1994) 372.
[15] P. Kukula, L. Cerveny, J. Mol. Catal. A 185 (2002) 195.
[16] T. Sugimura, S. Nakagawa, A. Tai, Bull. Chem. Soc. Jpn. 75 (2002)
355.
[17] T. Osawa, S. Sakai, T. Harada, O. Takayashu, Chem. Lett. (2001) 392.
[18] A. Wolfson, S. Geresh, M.V. Landau, M. Herskowitz, Appl. Catal.
A 208 (2001) 91.
[50] C. Reichardt, Solvents and Solvent Effects in Organic Chemistry,
VCH, Weinheim, 1988.
[19] Y. Nitta, Y. Ueda, T. Imanaka, Chem. Lett. (1994) 1095.
[20] A. Tungler, T. Tarnai, T. Màthé, G. Vidra, J. Petró, R.A. Sheldon,
in: M.G. Scaros, M.L. Prunier (Eds.), Catalysis of Organic Reactions,
Dekker, New York, 1995, p. 201.
[21] B. Török, K. Balàzsik, G. Szöllösi, K. Felföldi, M. Bartók, Chirality 11
(1999) 470.
[22] H.U. Blaser, H.P. Jalett, Stud. Surf. Sci. Catal. 78 (1993) 139.
[23] G.-Z. Wang, T. Mallat, A. Baiker, Tetrahedron Asymmetry 8 (1997)
2133.
[24] M. Schürch, N. Künzle, T. Mallat, A. Baiker, J. Catal. 176 (1998) 569.
[25] E. Toukoniitty, P. Maki-Arvela, M. Kuzma, A. Villela, A.K.
Neyestanaki, T. Salmi, R. Sjoholm, R. Leino, E. Laine, D.Y. Murzin,
J. Catal. 204 (2001) 281.
[51] J.-L.M. Abboud, R. Notario, Pure Appl. Chem. 71 (1999) 645.
[52] J.W. Nicoletti, G.M. Whitesides, J. Phys. Chem. 93 (1989) 759.
[53] P.N. Rylander, in: W.H. Jones (Ed.), Catalysis in Organic Syntheses,
Academic Press, New York, 1980, p. 155.
[54] D. Ferri, T. Bürgi, A. Baiker, J. Chem. Soc. Perkin Trans. 2 (1999)
1305.
[55] D. Ferri, T. Bürgi, A. Baiker, J. Chem. Soc. Perkin Trans. 2 (2002)
437.
[56] M. von Arx, T. Bürgi, T. Mallat, A. Baiker, Chem. Eur. J. 8 (2002)
1430.
[57] W.-R. Huck, T. Bürgi, T. Mallat, A. Baiker, J. Catal. 205 (2002) 213.
[58] S. Budàvari, The Merck Index, Whitehouse Station, New York, 1996.
[59] W.-R. Huck, T. Bürgi, T. Mallat, A. Baiker, J. Catal. 200 (2001) 171.
[60] C. LeBlond, J. Wang, A.T. Andrews, Y.-K. Sun, Top. Catal. 13 (2000)
169.
[26] M. Studer, S. Burkhardt, H.U. Blaser, Chem. Commun. (1999) 1727.
[27] B. Török, K. Felföldi, K. Balàzsik, M. Bartók, Chem. Commun. (1999)
1725.
[61] M. Bartók, G. Szöllösi, K. Balàzsik, T. Bartók, J. Mol. Catal. A 177
(2002) 299.
[28] M. von Arx, T. Mallat, A. Baiker, Tetrahedron Asymmetry 12 (2001)
3089.
[62] P.A. Meheux, A. Ibbotson, P.B. Wells, J. Catal. 128 (1991) 387.
[63] H.U. Blaser, H.P. Jalett, D.M. Monti, A. Baiker, J.T. Wehrli, Stud.
Surf. Sci. Catal. 67 (1991) 147.
[29] Y. Nitta, A. Shibata, Chem. Lett. (1998) 161.
[30] K. Borszeky, T. Bürgi, Z. Zhaohui, T. Mallat, A. Baiker, J. Catal. 187
(1999) 160.

W.-R. Huck et al. / Journal of Catalysis 216 (2003) 276–287
287
[64] V. Morawsky, U. Prüsse, L. Witte, K.-D. Vorlop, Catal. Commun. 1
(2000) 15.
[70] D. Ferri, T. Bürgi, K. Borszeky, T. Mallat, A. Baiker, J. Catal. 193
(2000) 139.
[65] W.-R. Huck, T. Mallat, A. Baiker, Catal. Lett. 69 (2000) 129.
[66] S. Nishimura, Handbook of Heterogeneous Catalytic Hydrogenation
for Organic Synthesis, Wiley, New York, 2001.
[67] F.W. Vierhapper, E.L. Eliel, J. Org. Chem. 40 (1975) 2729.
[68] M. Hönel, F.W. Vierhapper, J. Chem. Soc. Perkin Trans. 1 (1980)
1933.
[71] J.L. Margitfalvi, E. Talas, M. Hegedüs, Chem. Commun. (1999) 645.
[72] D. Guillaneux, S.H. Zhao, O. Samuel, D. Rainford, H.B. Kagan,
J. Am. Chem. Soc. 116 (1994) 9430.
[73] C. Girard, H.B. Kagan, Angew. Chem. Int. Ed. Engl. 37 (1998)
2923.
[74] M. von Arx, T. Mallat, A. Baiker, Catal. Lett. 78 (2002) 267.
[75] M. Sutyinszki, K. Szöri, K. Felföldi, M. Bartók, Catal. Commun. 3
(2002) 125.
[69] T. Mallat, Z. Bodnar, B. Minder, K. Borszeky, A. Baiker, J. Catal. 168
(1997) 183.