Palladium catalysts on alkaline-earth supports for racemization and dynamic kinetic resolution of benzylic amines

Article abstract of DOI:10.1002/chem.200600899

Palladium catalysts on alka-line-earth supports were studied as new heterogeneous catalysts for racemization of chiral benzylic amines such as 1-phenylethylamine. Particularly 5% Pd/BaSO₄ and 5% Pd/CaCO ₃ were able to selectively racemize amines, with minimal formation of secondary amines or hydrogenolysis to ethylbenzene. In contrast, these side reactions were pronounced on Pd/C. A reaction mechanism is proposed that is consistent with the reaction kinetics. The catalyst activity was found to depend on the number of available surface Pd atoms, determined by titration with CO. The selectivity crucially depends on the rate of condensation of the amine and the primary imine. which is highest on Pd/C. The racemization catalysts were combined in one pot with an immobilized lipase to perform dynamic kinetic resolution of chiral amines. High yields (up to 88%) of essentially enantiopure amides were obtained in a single step. The chemo-enzymatic catalyst system proved to be stable and could be reused without losing the initial activity.

Full text of DOI:10.1002/chem.200600899

DOI: 10.1002/chem.200600899
Palladium Catalysts on Alkaline-Earth Supports for Racemization and
Dynamic Kinetic Resolution of Benzylic Amines
[
a]
Andrei N. Parvulescu, Pierre A. Jacobs, and Dirk E. De Vos*
Abstract: Palladium catalysts on alka-
line-earth supports were studied as new
heterogeneous catalysts for racemiza-
tion of chiral benzylic amines such as
tent with the reaction kinetics. The cat-
alyst activity was found to depend on
the number of available surface Pd
atoms, determined by titration with
CO. The selectivity crucially depends
on the rate of condensation of the
amine and the primary imine, which is
highest on Pd/C. The racemization cat-
alysts were combined in one pot with
an immobilized lipase to perform dy-
namic kinetic resolution of chiral
amines. High yields (up to 88%) of es-
sentially enantiopure amides were ob-
tained in a single step. The chemo-en-
zymatic catalyst system proved to be
stable and could be reused without
losing the initial activity.
1-phenylethylamine. Particularly 5%
Pd/BaSO4 and 5% Pd/CaCO3 were
able to selectively racemize amines,
with minimal formation of secondary
amines or hydrogenolysis to ethylben-
zene. In contrast, these side reactions
were pronounced on Pd/C. A reaction
mechanism is proposed that is consis-
Keywords: amides
heterogeneous catalysis · kinetic
resolution · palladium
·
amines
·
AHCTREUNG
Introduction
such as homogeneous transition-metal catalysts, supported
[4]
metal catalysts, or solid acids. For amines reports on selec-
tive racemization or DKR are much more scarce. Homoge-
neous Ru-based catalysts were reported as being active in
Enantiomerically pure amines find many applications in the
pharmaceutical and agrochemical industries. They are used
as active intermediates, chemical building blocks, and chiral
[5]
racemization of chiral amines. Bäckvall et al. used the
Shvo complex catalyst, which showed good activity in race-
mization of a broad range of amines, with very few side re-
actions. In initial experiments, even with 5 mol% of the bi-
nuclear complex, a reaction temperature of 1108C was re-
[1]
auxiliaries. Usually, they are obtained by separation, for
example, diastereomeric crystallization or chiral chromatog-
raphy; by asymmetric hydrogenation of imines, enamines, or
[1,2]
oximes; or by kinetic resolution.
lution methods have been described that use enzymes as
A large number of reso-
[5]
quired, which impeded combination with an enzyme.
[3]
catalysts, but the resolution process is generally limited to
a maximum yield of 50%. Continuous racemization of the
remaining enantiomer combined with kinetic resolution,
called dynamic kinetic resolution (DKR), is an option to
overcome this problem. Like kinetic resolution, DKR starts
from the racemic mixture, but now the desired product en-
antiomer can be obtained with a theoretical yield of 100%.
While kinetic resolution is performed by an enzyme, the rac-
emization requires chemocatalysts. The DKR of secondary
alcohols is now well established and uses chemocatalysts
Changing the substituents on the cyclopentadienone ligand
allowed the reaction temperature to be decreased to 908C
and enabled one-pot combination with enzymatic resolu-
[6]
tion, but the long reaction time of three days is still disad-
vantageous.
In an alternative strategy for amine racemization, the re-
sidual amine enantiomer is treated at high temperature with
a mixture of hydrogen, ammonia, and the corresponding
[7]
ketone over solid metal oxide catalysts. As this process is a
gas-phase reaction, it is incompatible with enzymatic kinetic
resolution. Using a heterogeneous racemization catalyst in a
one-pot process is definitely more convenient, and such a re-
ported process uses Pd on charcoal as racemization catalytst
[
a] A. N. Parvulescu, Prof. Dr. P. A. Jacobs, Prof. Dr. D. E. De Vos
Centre for Surface Chemistry and Catalysis
Katholieke Universiteit Leuven
Kasteelpark Arenberg 23, 3001 Leuven (Belgium)
Fax : (+32)16-32-1998
in DKR of 1-phenethylamine in Et N. Unfortunately, the re-
3
action was quite slow (8 days) and side reactions decrease
[8]
the eventual yield of enantiopure amide. In a modified ap-
proach ketoximes are the starting materials, and the amines
E-mail: dirk.devos@biw.kuleuven.be
2034
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2034 – 2043

FULL PAPER
that are formed are resolved in situ by using an enzyme in
combination with a Pd/C racemization catalyst. The reaction
is still slow (5 days) and additives are required to prevent
Table 1. Screening results with various heterogeneous hydrogenation cat-
alysts in racemization of (S)-1-phenylethylamine.
[
a]
Entry Catalyst
Conv. YieldR-amine Sel.R-amine eeS-amine
[9]
[%]
[%]
[%]
[%]
the formation of large amounts of secondary products.
1
2
3
4
5
6
7
8
9
1
ideal catalyst
5% Pd/BaSO
5% Pd/C
5% Ru/C
5% Rh/C
5% Ir/CaCO
5% Pt/C
50
41
100
64
100
92
10
100
3
50
30
0
6
0
2
0
0
1
100
73
0
9
0
2
0
0
33
–
0
34
–
70
–
55
100
–
98
100
37
100
The activity of Pd in racemization of amines was first sug-
gested by Murahashi et al. as a side reaction when Pd black
was used for catalytic alkyl-group exchange between pri-
4
[10]
mary and secondary amines.
After screening different
types of heterogeneous hydrogenation catalysts in the race-
mization of (S)-1-phenylethylamine, we were drawn to the
possible use of Pd on alkaline-earth supports as racemiza-
tion catalyst. These catalysts are known to be selective for
the hydrogenation of alkynes to alkenes, of dienes to mono-
3
8% Pd/C
5% Ir/C
A
H
R
U
G
0
1% Pd/ZSM-5
30% Ni+1.26% Pd/SiO
0
87
29
–
6
0
11
2
7
0
[11]
1
2
5% Ru/BaSO
[a] Reaction conditions: 0.2 MPa H
S)-1-phenylethylamine, 4 mL toluene, three days, 708C.
4
olefins, for the synthesis of hydroxamic acids, and for dia-
[12]
stereoselective hydrogenation of imines.
However, their
2
pressure, 40 mg catalyst, 0.33 mmol
(
activity was not yet investigated in the racemization of opti-
cally active amines.
Herein we explore the potential of Pd on alkaline-earth
supports as catalysts for racemization of optically active
amines. Based on a detailed analysis of the reaction kinetics,
a sound mechanism was elucidated, and the parameters that
steer the racemization activity and selectivity were revealed.
By using an immobilized lipase, racemization and kinetic
resolution were combined in a one-pot DKR process. Reuse
of the chemocatalyst/enzyme system was also investigated.
A preliminary report on some aspects of this work appeared
Table 2. Pd on alkaline-earth supports versus Pd/C in racemization of
[
a]
(
S)-1-phenylethylamine.
Entry
Catalyst
Conv.
YieldR-amine
Sel.R-amine
eeS-amine
[
%]
[%]
[%]
[%]
1
2
3
4
5
5% Pd/BaSO
5% Pd/CaCO
4
41
59
58
65
30
21
35
35
–
73
36
60
53
–
34
33
10
0
3
5% Pd/SrCO
5% Pd/BaCO
5% Pd/C
3
3
100
–
[13]
recently.
[
a] Reaction conditions: 0.2 MPa H
2
pressure, 40 mg catalyst, 0.33 mmol
(
S)-1-phenylethylamine, 4 mL toluene, three days, 708C.
Results and Discussion
tion of Pd on different alkaline-earth supports, and includes
5% Pd/C as a reference. These data confirm that the best
combination comprises Pd as active noble metal, and a salt
of an alkaline-earth metal as basic support.
Selection of a racemization catalyst and optimization of con-
ditions: Racemization of (S)-1-phenylethylamine was chosen
as test reaction (Scheme 1). The optimum catalyst should
Based on this knowledge, we attempted to improve the
catalytic performance by tuning the reaction conditions,
such as hydrogen pressure. The racemization likely occurs
[5,8,9,10]
by dehydrogenation/hydrogenation (Scheme 2),
and
Scheme 1. Racemization of (S)-1-phenylethylamine
effect a 50% conversion of the starting enantiomer with
100% selectivity to the opposite stereoisomer, and thus
result in an ee of 0%. From the library of heterogeneous
catalysts screened in Table 1, Pd appears indeed to be active
in the racemization of chiral benzylic amines. The perfor-
mance obtained with 5% Pd/BaSO most closely approaches
4
the ideal behavior (Table 1, entry 2). Clearly, Pd is more
suitable for the racemization of amines than other noble
metals such as Ir, Pt, Rh, and Ru (Table 1, entries 4–7). The
basicity of the support also seems important. Without any
optimization of the reaction conditions, 5% Pd/BaSO gave
4
an acceptable selectivity of 73%, while 5% Pd/C exhibited
no selectivity at all. Beside the R enantiomer, ethylbenzene
and some amine condensation products were identified as
side products. Table 2 compiles results obtained by deposi-
Scheme 2. Reaction mechanism and kinetic constants for racemization of
(S)-1-phenylethylamine.
Chem. Eur. J. 2007, 13, 2034 – 2043
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2035

D. E. De Vos et al.
the hydrogen pressure will strongly affect the equilibrium
between amine and imine. Experiments at hydrogen pres-
sures between 0 and 0.2MPa proved indeed that an opti-
mum hydrogen pressure exists that affords superior racemi-
zation selectivity and yield. The absence of hydrogen led to
a high ee because the imine cannot be hydrogenated back to
the amine. On the contrary, a high hydrogen pressure sup-
presses imine formation (Figure 1). A hydrogen pressure of
sis catalyst for secondary amines at low hydrogen pres-
[12,14]
sure.
However, pretreatment of Pd/C by washing with a
base or use of a basic solvent such as triethylamine did not
improve the selectivity to the (R)-amine (Table 4, entries 2
and 3). In contrast, pretreatment of Pd/CaCO with NaOH
3
increased the selectivity of racemization, which is indeed an
indication that a basic support is more suitable for racemiza-
tion of chiral amines (Table 4, entries 4 and 5). The data for
Table 4. Influence of pretreatment with a base of Pd catalysts in racemi-
[
a]
zation of (S)-1-phenylethylamine.
Entry
Catalyst
Conv.
Sel.R-amine
Sel.ETB
[%]
eeS-amine
[
%]
[%]
[%]
[
[
[
b]
c]
d]
1
2
3
4
5
5% Pd/C
5% Pd/C
5% Pd/C
40
44
94
56
50
38
39
7
80
99
11
20
12
18
0
60
53
1
2
1
5% Pd/CaCO
5% Pd/CaCO
3
3
[
e]
[
a] Standard reaction conditions. [b] 1 h. [c] Washed with NaOH solution.
1
2
h. [d] 4 mL Et
4 h.
3
N used as solvent. 1 h. [e] Washed with NaOH solution;
the 5% Pd/CaCO catalyst washed with NaOH perfectly re-
flect the expected behavior of an ideal catalyst (compare
Table 4, entry 5 with Table 1, entry 1).
3
Figure 1. Relation between enantiomeric excess and hydrogen pressure in
racemization of (S)-1-phenylethylamine; 24 h reaction time and standard
reaction conditions.
Racemization kinetics and mechanistic proposal: The race-
mization kinetics of the can be approached in several ways.
First, one may simply assume Equation (1).
0
.01 MPa appears to be optimal. Correspondingly, the reac-
tion time was reduced from three days to one day, irrespec-
tive of the alkaline-earth support used (Table 3).
kinv
ðSÞ-amineƒ
!ƒðRÞ-amine
ð1Þ
kinv
Table 3. Activity of 5% Pd on alkaline-earth supports versus 5% Pd/C
under optimized hydrogen pressure.
[
a]
Assuming (pseudo)first-order in both amine enantiomers,
the rate equations can be integrated between time 0 and t
Entry
Catalyst
Conv.
Sel.R-amine
Sel.ETB
[%]
eeS-amine
[%]
[%]
[%]
[
Eq. (2)], where [S] and [S] are the concentrations of S en-
0 t
1
5% Pd/BaSO
4
4
56
35
56
60
58
98
40
81
91
80
67
69
2
19
8
2
25
2
1
2
[
b]
antiomer at times 0 and t.
5
% Pd/BaSO
2
3
4
5
5% Pd/CaCO
3
18
26
31
41
11
5% Pd/SrCO
5% Pd/BaCO
5% Pd/C
3
ꢀ
ꢁ
½
SŠ
0
3
ln
¼2 kinv
t
ð2Þ
6
60
2½SŠꢀ½SŠ
t
0
[
c]
5
% Pd/C
38
[
[
a] Standard reaction conditions, 0.01 MPa H
b] Reaction time 5 h. [c] Reaction time 1 h.
2
, 24 h, ETB=ethylbenzene.
As can be seen from Figure 2, the initial kinetics of race-
mization, for example, on Pd/BaSO , can be adequately
4
fitted by using Equation (2), that is, the (pseudo)first-order
assumption of Equation (1) is valid. However, when secon-
dary reactions become important, for example, at higher
conversions, or when Pd/C is used as the catalyst, a more de-
tailed kinetic scheme is necessary (Scheme 2). In this
scheme, prochiral imine 2 can be hydrogenated back with
formation of the two amine enantiomers, or it can be at-
tacked by another molecule of amine 1. Rapid elimination
Selectivity is a key issue in this racemization. The major
side product in these reactions was ethylbenzene. Data pre-
sented in Table 3 indicate that the selectivity is crucially af-
fected by the support: 5% Pd/BaSO and 5% Pd/CaCO led
4
3
to the highest selectivities for the desired R-amine. On these
catalysts the amount of secondary amine was less than 1%,
and even less than 0.5% for 5% Pd/BaSO4.
[10]
Apparently, the alkaline-earth supports preserve the pri-
mary amine much better than charcoal. Literature shows
that the hydrogenolysis of CꢀN bonds on Pd is promoted by
of NH3 from the aminal
results in a-methyl-N-(1-phen-
AHCTREUNG
ethylidene)benzylamine 3. This imine is readily hydrogenat-
ed to the two diastereoisomers of bis(1-phenylethyl)amine
(4), which were identified by GC-MS. Rate equations (3)–
(7) can be written for the reactions shown in Scheme 2.
[14]
acid sites. Charcoal usually contains phenolic and carbox-
ylic acid groups. Moreover, Pd/C is known as a hydrogenoly-
2036
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2034 – 2043

Racemization and Resolution of Benzylic Amines
FULL PAPER
Table 5. Relation between molar CO/Pd ratio and activity in racemiza-
[
a]
tion of (S)-1-phenylethylamine at 708C.
Entry Catalyst CO/Pd TOF [s ]”10
calcd for
ꢀ
1
4
ꢀ1
2
TOF [s ]”10
(
(per available
Pd)
total Pd)
1
2
3
4
5
1% Pd/BaSO
1% Pd/BaSO
5% Pd/BaSO
5% Pd/CaCO
5% Pd/C
4
4
4
(A) 0.017
7.3
10.6
13.2
17.6
32.2
4.3
5.0
6.3
3.7
4.0
(B)
0.021
0.021
0.048
0.081
3
[
a] Pd particle sizes: 9–14 nm (A), 6–10 nm (B); TOF calculated based
on conversions <10%. Reactions under standard conditions.
Figure 2. Linearization of kinetic data (standard conditions), by using
Equation (2), for 5%Pd/BaSO
amine.
4
in racemization of (S)-1-phenylethyl-
d½SŠ
¼
¼
ꢀk ½SŠþ0:5k ½imineŠꢀk ½imineŠ½SŠþ0:5 k ½sec-amineŠ
1
ꢀ1
2
4
dt
ð3Þ
d½RŠ
dt
ꢀk ½RŠþ0:5k ½imineŠꢀk ½imineŠ½RŠþ0:5 k ½sec-amineŠ
1
ꢀ1
2
4
ð4Þ
ð5Þ
d½imineŠ
dt
¼
k1f½RŠþ½SŠgꢀk ½imineŠꢀk ½imineŠf½RŠþ½SŠg
ꢀ1
2
Figure 3. Relationship between Pd TOF and molar CO/Pd ratio. Data for
reactions under standard conditions.
d½sec-imineŠ
dt
Pd/BaSO (A)] and 6–10 nm [1% Pd/BaSO (B)] (Table 5,
entries 1 and 2).
¼
k ½imineŠf½RŠþ½SŠgꢀk ½sec-imineŠf½RŠþ½SŠg
4
4
2
3
ð6Þ
As can be seen from Figure 3, a fairly linear relationship
is found between the Pd TOF and the CO/Pd ratio, even for
Pd catalysts with different supports. In general, Pd/BaSO4
catalysts have lower dispersion and hence lower TOF. Dis-
persion and TOF are highest for Pd/C, while intermediate
d½sec-amineŠ
dt
ð7Þ
¼
k ½sec-imineŠꢀk ½sec-amineŠf½RŠþ½SŠg
3
4
In the steady-state approximation of constant imine con-
centration (d[imine]/dt=0), both approaches are equivalent,
values are found for Pd/CaCO . The Pd activity is primarily
3
A
C
H
T
R
E
U
N
G
determined by the number of accessible Pd atoms, that is,
by the dispersion. There is no evidence for a direct relation
between Pd dispersion and selectivity of the reaction. Equa-
tions (3)–(7) were used to fit the data points up to higher
conversions (e.g., 40%). Recall that a racemization reaction
is complete at 50% conversion of the initial enantiomer.
Figure 4 plots the experimental results collected for 5% Pd/
BaSO , 5% Pd/CaCO , and 5% Pd/C (as dots), and the
with 2kinv =k . Hence, Equation (2) remains a useful tool to
1
evaluate k when the initial data points are considered. In
1
the initial phase of the reaction, when only racemization
occurs, dehydrogenation (k ) can be considered the rate-lim-
1
iting step, since the primary imine intermediate 2 is not de-
tected at all.
Based on the values of k , calculated using Equation (2)
1
4
3
and 2kinv =k , and knowing that the typical Pd concentration
fitted curves. Excellent agreement between model and ex-
perimental data was achieved with the rate constants of
Table 6. Large values were assumed for the hydrogenations
of the primary and secondary imines, which are fast steps
1
ꢀ
4
ꢀ3
varies between 9.4”10 and 4.7”10 m, initial Pd turnover
frequencies can be obtained for the dehydrogenation of the
S-amine to the imine. For different catalysts, these values
are compared in Table 5 and Figure 3 with the amount of
CO adsorbed on Pd at room temperature. Titration of the
surface of a solid noble metal catalyst with carbon monoxide
is a well-known method to determine the number of avail-
able atoms at the surface of the clusters, and hence the dis-
persion of the catalyst. The validity of these dispersion
measurements was also confirmed by TEM imaging (not
shown). For instance, for two samples of 1% Pd/BaSO4,
CO/Pd ratios of 0.017 and 0.021 were recorded; in TEM the
same samples gave typical Pd particle sizes of 9–14 nm [1%
(k and k ). Values of k correspond to the initial Pd TOF
ꢀ
1
3
1
values, as discussed previously. Most critical for the fits are
the values of k . These values prove that the bimolecular
2
condensation, which decreases the selectivity, is most pro-
nounced on Pd/C.
Overall, the fits indicate that the reactions of Scheme 2
adequately describe the complex reaction network. For sim-
plicity, (pseudo)first-order kinetics in all reagents was as-
sumed in all reactions. This does not necessarily mean that
the surface is only partially covered with the amine reagent.
Chem. Eur. J. 2007, 13, 2034 – 2043
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2037

D. E. De Vos et al.
Table 6. Rate constants for reactions under standard conditions.
k
5% Pd/BaSO
4
5% Pd/CaCO
3
5% Pd/C
ꢀ
1
k
k
k
k
k
1
[min
]
0.0045
2
150
1
0.006
2
420
1
0.011
2
2400
1
ꢀ
1
ꢀ1 [min
]
ꢀ
1
ꢀ1
2
3
4
[m min
]
ꢀ
ꢀ
1
[min
[min
]
]
1
0.01
0.01
0.01
formation of the secondary amine and ethylbenzene was
clearly observed in the case of 5% Pd/C (Figure 4c).
Based on these kinetics, the temperature dependence of
the racemization of (S)-1-phenylethylamine was investigated
for 5% Pd/BaSO4 at three different temperatures
ꢀ
1
(
Figure 5). The activation energy of E =120ꢁ10 kJmol
a
indicates that the reaction is not subject to any diffusion lim-
itation, but under a kinetic regime.
Figure 5. Arrhenius plot for racemization of (S)-1-phenylethylamine over
5
4
% Pd/BaSO .
Scope of the racemization: Pd on alkaline-earth catalysts
were also tested in racemization of other optically active
amines. Table 7 lists the racemization activity of 5% Pd/
BaSO with eight substrates. Usually, reaction times of 24 h
4
were sufficient, except for (S)-1-methyl-3-phenylpropyl-
AHCTREUNG
AHCTREUNG
Figure 4. Fitting of experimental data to kinetic model in racemization of
S)-1-phenylethylamine for a) 5% Pd/BaSO , b) 5% Pd/CaCO , and
c) 5% Pd/C. Standard reaction conditions; symbols: experimental values,
were the hydrogenolysis products. The initial rates of race-
mization were studied for some substituted benzylic amines
in the same way as for (S)-1-phenylethylamine (Figure 6).
With increasing electron-donating character of the substitu-
ent, the selectivity at the same conversion clearly increases,
but the racemization rate decreases. The opposite trend,
that is, higher racemization rate for electron-rich substitu-
(
4
3
curves: fitted curves. ETB=ethylbenzene.
For the specific case of racemization, the partial coverage by
(
S)-and ( R)-amines will be proportional to the fractions of
[5]
the two enantiomers in solution, and this will lead to very
similar kinetics as in Equations (3)–(5). Importantly, the fit-
ting explains why ethylbenzene formation only becomes
considerable well after the start of racemization. This ex-
cludes the possibility that ethylbenzene could be formed in
a parallel reaction, by direct hydrogenolysis of 1-phenyl-
ents, was observed for homogeneous Ru catalysts. This
seems to indicate that different elementary steps are in-
volved in the dehydrogenation of the amine with homogene-
ous Ru or heterogeneous Pd catalysts; the mechanistic de-
tails of this process are currently under investigation. When
(S)-1-(4-chlorophenyl)ethylamine was used as substrate, a
large amount of 1-phenylethylamine was obtained by Pd-
catalyzed CꢀCl hydrogenolysis. In the racemization of the
A
C
H
T
R
E
U
N
G
ethylamine; rather it results from hydrogenolysis of the sec-
ondary product bis(1-phenylethyl)amine 4. The consecutive
2038
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2034 – 2043

Racemization and Resolution of Benzylic Amines
FULL PAPER
Table 7. Racemization activity of 5% Pd/BaSO
nes.
4
for different chiral ami-
If the imine is more stable, its concentration is higher, the
probability for bimolecular reactions, such as attack of the
amine, increases, and a lower selectivity results.
[
a]
Entry
Amine
Conv.
%]
Sel.R-amine
ee
[
[%]
[%]
As expected, substantial amounts of hydrogenolysis prod-
ucts were formed in the racemization of the secondary
amine (S)-N-methyl(1-phenylethyl)amine. This is in line
with the idea that secondary amines are more prone to hy-
drogenolysis than primary amines. When the amino group is
not in the benzylic position, but linked via an aliphatic chain
to the aromatic ring, a higher reaction temperature is re-
quired. For (S)-1-methyl-3-phenylpropylamine, racemization
1
56
50
53
81
91
2
5
3
2
3
on Pd/BaSO requires 4 d at 908C. Chiral aliphatic amines
4
[
b]
96
77
were even unreactive towards racemization under these re-
action conditions with Pd on alkaline earth supports. Clear-
ly, for racemization of chiral aliphatic amines, future work
must focus on finding better reaction conditions or other
catalysts.
4
20
68
Effect of dopants on the racemization: It is well known that
Lindlar Pd catalysts, that is, Pd/CaCO doped with various
3
5
6
46
50
81
91
19
4
loadings of Pb, display improved selectivity, for example, in
[15]
partial hydrogenation of alkynes to alkenes. Figure 7 pres-
7
8
59
57
64
62
3
[c]
10
[
a] Standard racemization conditions, 0.01 MPa
H
2
,
24 h. [b] 72 h.
[
c] 908C, four days, 0.20 mmol amine.
Figure 7. Effect of Pb loading in the racemization of (S)-1-phenylethyl-
A
H
R
U
G
3
under standard racemization conditions for
2
ents the effect of different Pb loadings on amine racemiza-
tion. Increasing Pb loadings gradually decrease the catalytic
activity, even though the Pb doping does not decrease the
number of accessible Pd atoms, as measured by CO chemi-
sorption on catalysts with 5% Pd and 0, 1, or 2.5% Pb. The
selectivity increase at higher Pb contents is merely the effect
of lower conversion.
Alternatively, addition of Pt or Bi to Pd/C catalysts was
investigated. These dopants caused a tenfold activity de-
crease in comparison with the undoped catalyst (Table 8). Pt
Figure 6. Linearization of kinetic data (standard conditions) for racemiza-
tion of (S)-1-phenylethylamine (^), (S)-1-(4-tolyl)ethylamine (
), and
S)-1-(4-methoxyphenyl)ethylamine (~).
&
[
a]
(
Table 8. Effect of dopants in 5% Pd/C versus TOF.
Catalyst
ꢀ
1
4
TOF [s ]”10
5
% Pd/C
8% Pd (+2% Pt)/C
% Pd (+2% Bi)/C
32.2
2.8
2.9
naphthylethylamines, lower selectivities were observed than
A
C
H
T
R
E
U
N
G
5
for the phenylethylamines. This might be due to increased
stabilization of the imine intermediate by the naphthyl ring.
[a] Standard racemization conditions.
Chem. Eur. J. 2007, 13, 2034 – 2043
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2039

D. E. De Vos et al.
slightly enhanced the selectivity in (R)-amine, while doping
with Bi led to an almost unselective catalyst.
considered ([R]=0.01–0.08m), the acylation is pseudo-first-
order in [R]. In terms of the Michaelis–Menten formalism,
this implies that K >[R], or the enzyme is used at substrate
M
Kinetic resolution of benzylic amines: For resolution of the
amines, Candida antarctica lipase B (CalB) was used in im-
mobilized form. In commercial Novozyme 435 this lipase is
entrapped in an acrylic matrix. While the nature of the car-
rier has been shown to affect the enantioselectivity of reso-
lution reactions, for example, in the esterification of 4-meth-
yloctanoic acid with EtOH, such effects are not expected
for acyl transfer reactions such as transesterification or
amide formation using an ester as an acylating agent. A typ-
ical time profile of the acylation of racemic 1-phenylethyl-
concentrations well below those of its maximal rate.
Combined bio- and chemocatalytic dynamic kinetic resolu-
tion: The dynamic kinetic resolution (DKR) of racemic
amines was performed in a one-pot process by adding an
immobilized lipase and an acyl donor to the reaction sus-
pension (Table 9). Generally, the selectivity for by-products
[16]
[a]
Table 9. DKR of 1-phenylethylamine.
Entry
Catalyst
Conv.
[%]
Sel.R-amide
Sel.ETB
[%]
eeR-amide
A
C
H
T
R
E
U
N
G
amine by isopropyl acetate is shown in Figure 8.
[%]
[%]
1
2
3
4
5
5% Pd/BaSO
5% Pd/CaCO
4
89
89
83
89
100
90
84
85
75
30
10
16
15
9
>99
>99
>99
>99
>99
3
5% Pd/SrCO
5% Pd/BaCO
5% Pd/C
3
3
25
[
a] Standard DKR conditions, 0.01 MPa H
2
, 24 h, 0.35 mmol ethyl ace-
tate.
in the DKR of 1-phenylethylamine is equal to, or even
lower than, that in the racemization, since the amine is acy-
lated to the amide before it can be irreversibly lost by hy-
drogenolysis. In the coupled reactions, the relative selectivi-
ties of Pd/C and Pd supported on alkaline earth compounds
remained similar, with Pd/C still producing large amounts of
coupling products and ethylbenzene. Pd on BaSO proved
4
Figure 8. Conversion profile for kinetic resolution of 1-phenylethylamine
with 100 mg Novozyme 435 and 0.35 mmol iPrOAc at 708C (0.33 mmol
of substrate in 4 mL of toluene).
again to be the most selective catalyst. When Pd on an alka-
line earth support is used as the racemization catalyst, the
amide yield is always far above the 50% limit and the prod-
uct ee exceeds 99% in all cases. This proves that there are
no parallel unselective routes to the amide. While it was
previously shown that washing the racemization catalyst
with NaOH increased the selectivity of racemization
(Table 4), the NaOH-treated catalysts were not used further
in DKR because traces of NaOH decreased the enzyme ac-
tivity.
A first observation is that the (S)-amine is not acylated at
all by the enzyme. At an amine conversion very close to
5
0%, the product ee (ee ) is superior to 99.5%, while the ee
P
of the residual amine substrate (ee ) is 99%. Based on the
s
[17]
approach of Rakels et al.,
it can be calculated that the
enantioselectivity of the enzyme is very high, with an E
value theoretically exceeding 2000. This high enantioselec-
tivity is in agreement with earlier work on CalB-catalyzed
acylation of aryl ethyl amines, in which E values larger than
In a DKR process, the rates of racemization and kinetic
resolution should be of the same order of magnitude. If the
racemization is too slow, eeS quickly rises, and this might
entail enzyme-catalyzed formation of the wrong product
isomer, particularly with less enantioselective enzymes. The
theoretical framework of DKR has been outlined in detail
[3]
1
00 have frequently been observed. Secondly, Figure 8
allows the acylation rate to be evaluated under conditions
similar to those for the dynamic kinetic resolution. At 708C
in toluene, and at 0.083m of racemic amine, the specific ac-
tivity of the enzyme, as calculated from the initial rates, is
[18]
by Kitamura et al. In their model, it is assumed that not
only the racemization, but also the resolution, is first-order
in the substrate. As proven previously, the latter assumption
can reasonably be made for the Novozyme 435-catalyzed
acylation of 1-phenylethylamine. Hence, kinetic data of the
DKR can be modeled by replacing rate equation (4) by
Equation (8) and by adding Equation (9).
0
.038 mmol substrate converted per minute and per milli-
ꢀ
1
gram immobilized enzyme, or 0.038 IUmg . Figure 8 shows
that the acylation rate decreases as the reaction progresses.
In the relevant concentration domain of the acylating agent
isopropyl acetate (0.02–0.2m), the initial acylation rate is
hardly dependent on the concentration of the acylating
agent. Hence, the decrease in acylation rate during the ki-
netic resolution is ascribed to the order of the enzymatic re-
action in the amine substrate. In a first-order linearization
of the data in Figure 8, a straight line is obtained for ln[R]
versus time. This indicates that in the concentration domain
d½RŠ
¼
ꢀk ½RŠþ0:5 k ½imineŠꢀk ½imineŠ½RŠ
1
ꢀ1
2
dt
ð8Þ
þ0:5 k ½sec-amineŠꢀk ½RŠ
4
R
2040
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2034 – 2043

Racemization and Resolution of Benzylic Amines
FULL PAPER
d½R-amideŠ
[3]
ð9Þ
with C=C bonds, such as vinyl acetate, are unsuitable be-
¼
kR½RŠ
dt
cause they are hydrogenated on the Pd catalysts. Moreover,
the alcohol that is released should be less nucleophilic than
the amine, since it might otherwise cause undesired reac-
Herein [R] refers to the concentration of R-amine sub-
strate, and [R-amide] is the concentration of enantiopure
amide product. An exemplary fit is shown in Figure 9. De-
[
19]
tions with the amide reaction product.
Testing various
acylating agents for DKR of 1-phenylethylamine showed
that conversions and selectivities were highest with EtOAc,
iPrOAc, and ethyl methoxyacetate. The performances of the
last two acylating agents are even slightly better than those
of EtOAc. A superior selectivity for (R)-amide (98%) is ob-
tained with methyl decanoate, but the reaction is slower
(
Table 10). With methyl formate, a lower ee was obtained,
because of uncatalyzed formation of the (S)-amide.
As expected, the hydrogen pressure has a subtle influence
on the DKR via the racemization. For 5% Pd/BaSO4 in
DKR of 1-phenylethylamine (Table 11), a maximum yield of
chiral amide is found at an H pressure of 0.02MPa. Increas-
2
ing the reaction time from 24 to 72 h slightly increases the
conversion while preserving the same selectivity. The activi-
ty of 5% Pd/BaSO in the DKR of other amines was also
4
tested (Table 12); DKR was successful with at least seven
substrates. The enzyme was not able to acylate N-methyl(1-
phenylethyl)amine. Like in the case of racemization, benzyl-
ic amines with electron-donating substituents are the best
Figure 9. Fitting of experimental data to kinetic model in DKR of 1-phe-
nylethylamine over 5% Pd/BaSO
dard DKR conditions.
4
and 100 mg of Novozyme 435. Stan-
[
a]
4 3
Table 10. Influence of acyl donor type in DKR of 1-phenylethylamine on 5% Pd/BaSO and 5% Pd/CaCO .
spite the assumptions and sim-
plifications made, such as the Entry Catalyst
Acyl donor, mmol
Conv. [%] Sel.R-amide [%] Sel.ETB [%] eeR-amide [%]
pseudo-first-order dependence
of acylation on (R)-amine, the
concentrations of (R)-amide,
total amine, (S)-amine, and (R)-
amine can be fitted adequately.
Exactly the same constants were
employed as in Figure 4a and
1
2
3
4
5
6
7
8
9
1
5% Pd/BaSO
5% Pd/BaSO
5% Pd/BaSO
5% Pd/BaSO
5% Pd/BaSO
5% Pd/BaSO
4
4
4
4
4
4
methyl formate, 0.83
MeOAc, 0.41
EtOAc, 0.35
iPrOAc, 0.35
ethyl methoxyacetate, 0.35 91
94
72
89
91
57
97
90
94
96
93
98
90
84
87
0
3
10
6
4
6
2
9
16
7
14
99
>99
>99
99
methyl butyrate, 0.34
67
72
85
89
93
97
5% Pd/BaSO4 methyl decanoate, 0.34
>99
>99
>99
>99
5% Pd/CaCO
5% Pd/CaCO
5% Pd/CaCO
3
3
3
MeOAc, 0.41
EtOAc, 0.35
iPrOAc, 0.35
Table 6, except for the k value.
2
0
The use of the same k and k
1
ꢀ1
2
[a] Standard DKR conditions, 0.01 MPa H , 24 h.
values means that the activity of
the racemization catalyst is un-
affected by the presence of the
ꢀ
1
ꢀ1
2
Table 11. Influence of H pressure and time in DKR of 1-phenethyl-
enzyme. For k , a higher value (330m min ) was employed
2
[
a]
A
H
R
U
G
4
amine on 5%Pd/BaSO .
to obtain a good fit. This is not surprising, as even a minute
amount of undesired hydrolysis of the acyl donor results in
the presence of acetic acid, which could favor acid-catalyzed
H
p 2
Conv.
[%]
Sel.R-amide
[%]
Sel.ETB
[%]
eeR-amide
A
H
R
U
G
[%]
[
b]
0
0.02
.01
95
98
99
89
90
90
85
90
10
10
15
10
>99
>99
>99
>99
bimolecular condensation of 1 and 2. The value for k of
R
[
b]
ꢀ
1
0
(
.009 min
0.0045 min ) which described the rate-limiting step of the
racemization. In classical DKR kinetics as described by Ki-
is of comparable magnitude to that of k1
[b]
0
0
.2
.01
ꢀ
1
[c]
[
a] Standard DKR conditions, 0.35 mmol acylating agent, EtOAc.
[18]
tamura et al., one obtains a ratio k /k =0.25. With such
inv
R
[b] 72 h. [c] 24 h.
a ratio, ee is expected to decrease only to a limited extent
P
during DKR, especially if the enzyme is highly enantioselec-
tive.
substrates; they exhibit excellent amide selectivity with high
product ee and high yield. The conditions employed in the
DKR process, such as type and amount of acylating donor
and reaction time, depend on the nature of the amine. As in
racemization, lower yields were obtained with 1-naphthyle-
thylamine. For resolving 1-methyl-3-phenylpropylamine,
Additional improvements can be made by varying the
type of acylating agent. Even if the literature abounds with
reports regarding kinetic resolution of optically active
amines using different acylating agents, some of these are in-
compatible with the racemization process. Acylating agents
Chem. Eur. J. 2007, 13, 2034 – 2043
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2041

D. E. De Vos et al.
[
a]
Table 12. DKR of benzylic amines with Pd on alkaline-earth supports.
Entry
Amine
Time
h]
Conv.
[%]
Sel.R-amide
[%]
eeR-amide
[%]
[
[
[
b]
c]
1
24
48
48
91
92
90
94
97
98
>99
>99
>99
2
3
[
[
[
d]
e]
e]
Figure 10. Reuse of the enzyme and 5% Pd/BaSO
1
0
4
catalysts in DKR of
-phenylethylamine. Standard DKR conditions, 0.01 MPa H , 24 h, and
.35 mmol acylating agent.
2
4
5
48
48
64
89
87
87
>99
>99
Conclusion
In summary, Pd immobilized on supports such as BaSO4,
CaCO , SrCO , or BaCO is an efficient heterogeneous cata-
3
3
3
lyst for the racemization of benzylic amines. The racemiza-
tion proceeds via dehydrogenation and hydrogenation, and
is first-order in the amine substrate. The by-products arise
via formation of a secondary amine and its subsequent hy-
drogenolysis. Side product formation is strongly suppressed
when BaSO or CaCO is used as support for Pd. The race-
[
[
c]
f]
6
7
72
96
84
98
90
91
>99
>99
4
3
mization can be combined with enzymatic kinetic resolution
in a one-pot process leading to enantiomerically pure
amides from racemic amines with good yields. Particulary
[
a] Standard DKR conditions, 0.01 MPa H
iPrOAc. [c] 0.35 mmol EtOAc. [d] 0.60 mmol EtOAc, [e] 0.60 mmol
iPrOAc. [f] 5% Pd/CaCO , 908C, 0.33 mmol ethyl methoxyacetate, 15 mg
Na CO , 15 mg Novozyme 435.
2 4
, 5% Pd/BaSO . [b] 0.33 mmol
3
Pd on BaSO showed a satisfactory activity combined with
4
2
3
high selectivity in the racemization and DKR of benzylic
amines. The catalytic system formed by the chemocatalyst
and the immobilized enzyme can be recovered and reused
several times without any loss of activity. Although very
good results were obtained for benzylic chiral amines, the
design of a better heterogeneous chemocatalyst for the
DKR of saturated amines still represents a future objective.
special reactions conditions were employed, since at 708C
only kinetic resolution and no racemization was observed.
Better results (98% conversion, 91% amine selectivity,
9
0% ee) were obtained by raising the temperature to 908C,
extending the reaction time to four days, and adding less
enzyme and some Na CO , which seems to prevent uncata-
2
3
lyzed formation of S-amide. Especially this reaction shows
that side product formation is limited in a DKR compared
to racemization (compare Table 12, entry 7 with Table 7,
entry 8).
Experimental Section
Materials: All reactants were obtained from commercial sources and
used as received.
To evaluate the heterogeneity of the system and to in-
crease the productivity of the catalysts, reuse of both cata-
lysts was attempted. In the DKR of 1-phenylethylamine a
sample was taken for analysis after the first run and the
liquid was removed by centrifugation. The physical mixture
of the two solid catalysts was washed several times with sol-
vent, dried under nitrogen flow, and a fresh solution of sub-
strate (4 mL toluene and 0.33 mmol of 1-phenethylamine)
was added together with the acylating agent. After three
consecutive runs no appreciable loss of conversion was ob-
served and the selectivity remained higher than 90%, with
an ee of the resulting (R)-amide higher than 99%
Catalysts: Literature procedures were followed for the preparation of
[20a]
[20a]
[20b]
5
% Pd/BaSO
cedures (B), Pd was reduced with an alkaline solution of formaldehyde
at a formaldehyde/Pd molar ratio of 10/1, and PdCl was used as metal
source. 1% Pd/BaSO (B) was prepared using the same source of metal
and the same molar formaldehyde/Pd ratio as for 5% samples. For the
preparation of 1% Pd/BaSO (A), Pd(OAc) was used as metal source,
with a formaldehyde/Pd ratio of 50/1. 5% Pd/CaCO , 5% Pd/CaCO (+
4
,
5% Pd/BaCO
3
,
and 5% Pd/SrCO
3
.
In these pro-
2
4
4
A
H
R
U
G
2
3
3
2.5% Pb), 5% Pd/CaCO
were gifts from Johnson Matthey; 8% Pd/C (+2% Pt), 5% Pd/CaCO
+5% Pb) were purchased from Heraeus; all catalysts were used without
3
(+1% Pb), 5% Pd/C, 5% Pd/C (+2% Bi)
3
(
any further pretreatment. Candida antarctica lipase B immobilized in
acrylic resin (Novozyme 435) was purchased from Aldrich.
Racemization reactions: Reactions were performed in 10 mL stainless
steel autoclaves at 708C under a hydrogen pressure of 0.01–0.2 MPa. To
(
Figure 10). This result proves the robustness of the system.
2042
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2034 – 2043

Racemization and Resolution of Benzylic Amines
FULL PAPER
easily obtain H
2
pressures below 0.1 MPa, 5% H
2
in N
2
was used as reac-
Chem. Rev. 2003, 103, 3247–3261; c) N. J. Turner, Curr. Opin. Chem.
Biol. 2004, 8, 114–119; d) J. H. Choi, Y. H. Kim, S. H. Nam, S. T.
Shin, M. J. Kim, J. Park, Angew. Chem. 2002, 114, 2479–2482;
Angew. Chem. Int. Ed. 2002, 41, 2373–2376; e) M.-J. Kim, Y. Ahn, J.
Park, Curr. Opin. Biotechnol. 2002, 13, 578–587; f) S. Wuyts, D. E.
De Vos, F. Verpoort, D. Depla, R. De Gryse, P. A. Jacobs, J. Catal.
tive gas. For the standard racemization 0.33 mmol of (S)-1-phenylethyla-
mine, 4 mL of toluene, and 40 mg of catalyst were used.
Dynamic kinetic resolution: Reactions were performed under similar
conditions with 0.33 mmol of racemic 1-phenylethylamine, 4 mL toluene,
1
00 mg of immobilized Candida Antarctica lipase B (Novozyme 435) as
2
003, 219, 417–424; g) S. Wuyts, K. De Temmerman, D. E. De Vos,
resolution catalyst, 40 mg of racemization catalyst, and 0.35 mmol of acy-
lating agent. At the end of the reaction the autoclave was cooled to room
temperature, the catalyst was separated by centrifugation and a sample
was taken for further analysis.
P. A. Jacobs, Chem. Commun. 2003, 1928–1929; h) K. Yamaguchi,
T. Koike, M. Kotani, M. Matsushita, S, Shinachi, N. Mizuno, Chem.
Eur. J. 2005, 11, 6574–6582; i) D. Klomp, T. Maschmeyer, U. Hane-
feld, J. A. Peters, Chem. Eur. J. 2004, 10, 2088–2093; j) D. Klomp,
K. Djanashvili, N. C. Svennum, N. Chantapariyavat, C. S. Wong, F.
Vilela, T. Maschmeyer, J. A. Peters, U. Hanefeld, Org. Biomol.
Chem. 2005, 3, 483–489; k) L. Veum, L. T. Kanerva, P. J. Halling, T.
Mashmeyer, U. Hanefeld, Adv. Synth. Catal. 2005, 347, 1015–1021.
5] O. Pami›s, A. H. Ell, J. S. M. Samec, N. Hermanns, J. E. Bäckvall,
Tetrahedron Lett. 2002, 43, 4699–4702.
Instrumentation: XRD measurements were made on a Stoe StudiP dif-
fractometer with CuKa1 radiation (l=1.54 Š). For all catalysts, the XRD
pattern was the same as that of the pure supports; in the case of BaSO
traces of Ba(OH) ·8H O were also observed. No traces of PdO or PdCl
were detected. HRTEM and EDX measurements were respectively per-
formed with a JEOL 4000EX and a Philips CM20 operated at 400 kV.
The EDX measurement confirmed for all samples that Pd catalysts did
not contain any traces of Cl ions. Pd particle size was measured by aver-
aging over 100 individual particles. ICP measurements were made on a
Jobin Yvon Ultima instrument. The metal loadings found for Pd/BaSO
4
2
2
2
[
ꢀ
[6] J. Paetzold, J. E. Bäckvall, J. Am. Chem. Soc. 2005, 127, 17620–
7621.
7] a) F. Funke, S. Liang, A. Kramer, R. Stürmer, A. Hçhn, US
576795, 2003; [Chem. Abs. 2003, 137, 48866]; b) H. Riechers, J.
Simon, A. Hçhn, A. Kramer, F. Funke, W. Siegel, C. Nübling, US
160178, 2000, ; [Chem. Abstr. 2000, 132, 308056].
1
[
4
,
6
Pd/SrCO , and Pd/BaCO were 4.7, 5, and 5 wt%, respectively. All cata-
3
3
lysts are denoted as 5% Pd/support. Yields and enantiomeric purities of
substrate and reaction products were determined by GC (HP 6890) on a
CP-CHIRASIL-DEX CB chiral column (25 m) with FID detector and
tetradecane as internal standard. The temperature program was as fol-
lows: 25 min at 708C, heating at 158Cmin to 1508C, and finally 10 min
at 1508C. Under these conditions, typical retention times are: ethylben-
zene, 2.07 min; (R)-1-phenylethylamine, 13.35 min; (S)-1-phenylethyla-
mine, 14.3 min; (S)-N-1-phenylethyl acetamide, 30.5 min; (R)-N-1-phe-
nylethyl acetamide, 30.9 min. Ethylbenzene and the secondary amines
were identified by using a GC-MS Agilent 6890-N with an Agilent 5973-
MSD on a silica column HP-5MS (30 m). MS data for bis(1-phenethyl)a-
mine (4): m/z (%): 225 (1), 210 (43), 120 (8), 106 (75), 105 (100), 79 (16),
6
[
[
8] M. T. Reetz, K. Schimossek, Chimia 1996, 50, 668–669.
9] Y. K. Choi, M. J. Kim, Y. Ahn, M.-J. Kim, Org. Lett. 2001, 3, 4099–
4101.
10] S. I. Murahashi, N. Yoshimura, T. Tsumiyama, T. Kojima, J. Am.
Chem. Soc. 1983, 105, 5002–5011.
ꢀ
1
[
[
11] a) G. C. Bond, A. F. Rawle, J. Mol. Catal. B J. Mol. Catal. A 1996,
1
09, 261–271; b) S. S. Nikam, B. E. Komberg, D. R. Johnson, A. M.
Doherty, Tetrahedron Lett. 1995, 36, 197–200; c) S. Bailey, F. King
in Fine Chemicals through Heterogeneous Catalysis (Eds.: R. A.
Sheldon, H. van Bekkum), Wiley-VCH, 2001, pp. 351–362.
[
[
[
12] B. Tçrçk, G. K. Surya Prakash, Adv. Synth. Catal. 2003, 345, 165–
7
7 (28), 51 (6).
1
68.
13] A. Parvulescu, D. De Vos, P. Jacobs, Chem. Commun. 2005, 42,
307–5309.
CO chemisorption measurements: CO chemisorption measurements
were made using an Omnistar TM mass spectrometer coupled to a Pfeiff-
er Vacuum pump and an oven with automatic temperature control. The
5
14] a) M. Kanai, K. Ueda, M. Yasumoto, Y. Kuriyama, K. Inomiya, T.
ꢀ
1
pretreatment procedure comprised heating in 10 mLmin
H
2
flow at
Ootsuka, Y. Katsuhara, K. Higashiyama, A. Ishii, J. Fluorine Chem.
ꢀ
1
1
08Cmin up to 708C, 1 h under flowing H at 708C, cooling to 258C
2
2
005, 126, 377–383; b) M. Kanai, M. Yasumoto, Y. Kuriyama, K. In-
omiya, Y. Katsuhara, K. Higashiyama, A. Ishii, Org. Lett. 2003, 5,
007–1010.
ꢀ
1
under He flow (10 mLmin ), and 30 min at 258C. The measurements
were made at room temperature by giving 5mL pulses of CO at regular
time intervals.
1
[
[
15] a) H. Lindlar, Helv. Chim. Acta 1952, 35, 446; b) J. G. Ulan, E. Kuo,
W. F. Maier, R. S. Rai, G. Thomas, J. Org. Chem. 1987, 52, 3126–
3
132.
16] N. W. J. T. Heinsman, C. G. P. H. Schroꢁn, A. van der Padt, M. C. R.
Acknowledgements
Franssen, R. M. Boom, K. Van’t Riet, Tetrahedron: Asymmetry
2
003, 14, 2699–2704.
We are grateful to the Bilateral program Flanders-Romania BIL 04/42,
to the IAP program Supramolecular Chemistry and Catalysis of the Bel-
gian Federal Government, to F.W.O. Vlaanderen for a grant, and to Cost
action D24.
[
17] J. L.L. Rakels, A. J.J. Straathof, J. J. Heijnen, Enzyme Microb. Tech-
nol. 1997, 21, 559–571.
18] M. Kitamura, M. Tokunaga, R. Noyori, Tetrahedron 1993, 49, 1853–
[
1
860.
[
[
19] U. Hanefeld, Org. Biomol. Chem. 2003, 1, 2405–2415.
20] a) R. Mozingo, Organic Syntheses, Coll. Vol. 3, 1946, p. 685; b) W. S.
Johnson, E. R. Rogier, J. Szmuszkovicz, H. I. Hadler, J. Ackerman,
B. K. Bhattacharyya, B. M. Bloom, L. Stalmann, R. A. Clement, B.
Bannister, H. Wynberg, J. Am. Chem. Soc. 1956, 78, 6280.
[
[
1] M. Breuer, K. Dietrich, T. Habicher, B. Hauer, M. Keßeler, R.
Stürmer, T. Zelinski, Angew. Chem. 2004, 116, 806–843; Angew.
Chem. Int. Ed. 2004, 43, 788–824.
2] K. Faber, Biotransformations in Organic Chemistry, 3rd ed., Spring-
er, Heidelberg, 1997.
[
[
3] F. van Rantwijk, R. A. Sheldon, Tetrahedron 2004, 60, 501–519.
4] a) B. A. Persson, A. L. E. Larsson, M. Le Ray, J. E. Bäckvall, J. Am.
Chem. Soc. 1999, 121, 1645–1650; b) O. Pami›s, J. E. Bäckvall,
Received: June 23, 2006
Published online: December 7, 2006
Chem. Eur. J. 2007, 13, 2034 – 2043
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2043