Dynamic kinetic resolution of secondary alcohols combining enzyme-catalyzed iransesterification and zeolite-catalyzed racemization

Article abstract of DOI:10.1002/chem.200600723

Hydrophobic zeolite beta containing low concentrations of Zr or Al was found to be a good catalyst for the racemization of 1-phenylethanol. The formation of styrene as a side product could be minimized by reducing the metal concentration in the zeolite be

Full text of DOI:10.1002/chem.200600723

FULL PAPER
DOI: 10.1002/chem.200600723
Dynamic Kinetic Resolution of Secondary Alcohols Combining Enzyme-
Catalyzed Transesterification and Zeolite-Catalyzed Racemization
Yongzhong Zhu,[a] Kam-Loon Fow,[b] Gaik-Khuan Chuah,[b] and Stephan Jaenicke*[b]
Abstract: Hydrophobic zeolite beta
containing low concentrations of Zr or
Al was found to be a good catalyst for
the racemization of 1-phenylethanol.
The formation of styrene as a side
product could be minimized by reduc-
ing the metal concentration in the zeo-
lite beta. Combined with an immobi-
lized lipase from Candida antarctica,
the dynamic kinetic resolution of 1-
phenylethanol to the (R)-phenylethyl-
conducted in toluene as solvent at
608C, with higher temperatures leading
to a loss in the enantioselectivity of the
formed ester. By using high-molecular-
weight acyl-transfer reagents, such as
vinyl butyrate or vinyl octanoate, a
high enantiomeric excess of the product
esters of 92 and 98%, respectively,
could be achieved. This is attributed to
a steric effect: the bulky ester is less
able to enter the pore space of the zeo-
lite catalyst where the active sites for
racemization are localized. Close to
100% conversion of the alcohol was
achieved within 2 h. If the more
common acyl donor, isopropenyl ace-
tate, was used, the enantiomeric excess
(ee) of the formed ester was only 67%,
and the reaction was considerably
slower.
Keywords: heterogeneous catalysis ·
kinetic resolution · racemization ·
zeolites · zirconium
A
and selectivity. The reaction was best
Introduction
that it can be separated with relative ease. The other enan-
tiomer reacts much more slowly, and remains essentially un-
changed. If the enantioselective synthesis of a desired com-
pound starts with a racemic mixture, the yield will be limited
to 50%. A better yield can be achieved if the remaining off-
enantiomer can be racemized, and the resulting mixture of
enantiomers subjected to the reaction for a second time. If
the processes of racemization and kinetic transformation of
the faster-reacting enantiomer are conducted in the same
pot, the combined system is referred to as “dynamic kinetic
resolution”. This term was coined by Noyori[2] in 1989. Dy-
namic kinetic resolutions offer the prospect of up to 100%
yield of the desired stereoisomer from a racemic starting
material and are, therefore, very attractive in terms of atom
efficiency.
Enzymes are excellent stereoselective catalysts. Since
Zaks and Klibanovꢀs landmark 1984 paper[3] on enzymatic
reactions in organic solvents, their potential for organic syn-
thesis is becoming increasingly recognized. Particularly, es-
terification reactions catalyzed by lipases in nonaqueous
medium have been widely used for the kinetic resolution of
alcohols in the form of esters. To overcome the intrinsic lim-
itation of 50% yield in kinetic resolution schemes, the cata-
lytic racemization of the off-isomer has been described.[4] In-
itial work concentrated on stereolabile groups that racemize
easily. Secondary alcohols are not stereolabile, and more
Enantiomerically pure compounds are increasingly in
demand as building blocks for pharmaceutically active mole-
cules. Besides the “traditional” method of racemate resolu-
tion by salt formation, other methods are gaining in impor-
tance. Transformations starting from the chiral pool of natu-
rally occurring single enantiomers are a popular approach.
Stereoselective reactions with chiral catalysts also offer
access to enantiomerically pure products. “Chemical” cata-
lysts based on chiral ligands and (mostly) noble metal cen-
ters and biocatalysts currently share this market at a ratio of
about 2:1.[1] Chiral catalysts select only one enantiomer out
of an enantiomeric pair for chemical transformation. Kinetic
resolutions employ this to their advantage, because one
enantiomer can be selectively transformed to a chiral prod-
uct that is sufficiently different from the starting material, so
[a] Dr. Y. Zhu
Technische Universität München, Department of Chemistry
Lichtenbergstrasse 4, 85748 Garching (Germany)
[b] K.-L. Fow, Prof. G.-K. Chuah, Prof. S. Jaenicke
Department of Chemistry, National University of Singapore
3 Science Drive 3, Kent Ridge, Singapore 119543 (Singapore)
Fax : (+65)6779-1691
E-mail: chmsj@nus.edu.sg
Chem. Eur. J. 2007, 13, 541 – 547
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
541

drastic conditions are required to bring about racemization.
A number of homogeneous catalysts that racemize secon-
dary alcohols under mild reaction conditions and are poten-
tially compatible with enzymatic reactions have been de-
mization of chiral 1-phenylethanol, whereas the enzyme pro-
motes the transesterification in the organic phase. This
system appears conceptually elegant, but unfortunately, it
suffers from considerable unproductive hydrolysis of the
acyl donor, which must be continuously added during the re-
action and reaches a 12-fold stoichiometric excess by the
time the reaction is brought to completion.
Here, we report a comparison of metal-substituted beta
zeolites for the racemization of (S)-1-phenylethanol fol-
lowed by the coupling of racemization with a lipase-cata-
lyzed stereoselective transesterification in one pot
(Scheme 1). Although water can be used as the solvent for
A
aluminium propoxide, which is active in the Meerwein–
Ponndorf–Verley reduction and Oppenauer oxidation.
Better rates were obtained with homogeneous catalysts
comprising Ru complexes.[6] Although some difficulties with
these systems have been overcome, such as the need for a
strong base to form the active catalyst and the requirement
for an intermediate hydrogen acceptor, they suffer from the
common disadvantages of ho-
mogeneous catalysis, that is, dif-
ficulties in separating the prod-
ucts and regeneration of the
catalyst.
Heterogeneous catalysts offer
considerable advantages in
processing. However, few het-
erogeneous catalysts for dynam-
ic resolution have been report-
ed. Reetz and Schimossek[7]
used Pd/Cfor the dynamic res-
Scheme 1. Dynamic resolution of 1-phenylethanol.
olution of chiral amines. Acidic
ion exchangers have been used
for the racemization of optically active arylethanolamines,[8]
benzylic acids,[9] and acyloins.[10] Wuyts et al.[11] reported on
a heterogeneous catalyst based on Ru3+ exchanged into a
hydroxyapatite for the racemization of secondary alcohols
under relatively mild conditions. However, this catalyst was
not compatible with chelating groups. Even esters, which are
the product of a transesterification reaction, severely deacti-
vate the catalyst. The same group also investigated the race-
mization of 1-phenylethanol over a number of zeolites.[12]
HY and mordenites were less active than zeolite beta. The
rate of racemization increased as the Si/Al ratio increased.
This was attributed to an increase in acid strength and hy-
drophobicity that apparently favours the adsorption of phe-
nylethanol over that of water in the pore space of the cata-
lyst. However, in this study, only H-beta zeolites with a Si/
Al ratio of between 11 and 12.5 were investigated, and the
activity of material from different sources was very similar
to each other. Klomp et al.[13] found that the racemization of
1-phenyethanol occurred through elimination/addition of
the hydroxyl group rather than through a Meerwein–Ponn-
dorf–Verley–Oppenauer reduction/oxidation mechanism.
They investigated zeolite beta with varying Si/Al ratios and
established that the highest rate was obtained with a Si/Al
of 16.
the racemization reaction, it is not suitable for transesterifi-
cation. Therefore, we attempted to racemize benzylic alco-
hols in organic solvents over high-silica beta zeolites that
were doped at a low level with other metals to generate
active sites. With this class of materials, intrinsically hydro-
phobic catalysts were realized.
Results
Textural properties: Zeolite beta is a pentalsil zeolite with a
three-dimensional network of large (12-member) rings. Beta
zeolites doped with a low concentration of Sn, Ti, or Zr
were synthesized by using fluoride as a mineralizer, and
low-aluminium-containing beta zeolites were prepared
under basic conditions with tetraethylammonium hydroxide
(TEAOH) as the structure-directing agent. The physical
properties of the zeolite beta prepared are given in Table 1.
All the samples had surface areas in excess of 450 m2 gÀ1 and
micropore volumes of 0.21–0.23 cm3 gÀ1. Regardless of the
medium of synthesis, the powder X-ray diffractograms were
typical for zeolite beta with both sharp and broad reflec-
tions, indicating a certain degree of disorder in one dimen-
sion. The crystallinity of all samples synthesized in the fluo-
ride medium was higher than that of zeolite beta synthe-
sized in basic medium (Figure 1).
The main problems faced in applying heterogeneous ca-
A
and/or incompatibility of the kinetic resolution and the race-
mization procedures. In an attempt to resolve the latter
problem, Wuyts et al.[14] carried out dynamic kinetic resolu-
tion in a biphasic system, in which an acidic H-beta zeolite
catalyst suspended in the aqueous phase brings about race-
The acidic properties were determined by conducting in-
frared spectroscopy to monitor adsorption of pyridine at
room temperature and desorption at 1008C(Figure 2). Both
Sn-beta and Ti-beta show the presence of Lewis acidity with
absorption bands at 1447–1460 cmÀ1, 1488–1504 cmÀ1,
542
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 541 – 547

Kinetic Resolution of Secondary Alcohols
FULL PAPER
Table 1. Chemical and textural properties of zeolite beta racemization catalysts.
tain a much higher density of
terminal hydroxyl groups than
those prepared by using fluo-
ride.
Catalyst
Si/Me ratio[a]
BET
Microporous
pore volume
Total
Water
lost
[%]
Si/M
Si/Al
surface area
pore volume
[m2 gÀ1
]
[cm3 gÀ1
]
[cm3gÀ1
]
Zr-beta
ZrAl-100[a]
ZrAl-25[a]
Sn-beta
Ti-beta
Al-150
Al-200
Al-300
107
104
105
125
100
–
>4000
103
27
490
510
533
500
468
454
241
260
0.23
0.23
0.21
0.23
0.22
0.22
0.10
0.06
0.27
0.31
0.24
0.31
0.25
0.27
0.18
0.10
1.7
1.8
–
Racemization of (S)-1-phenyl-
A
–
–
147
212
–
–
catalyst for the racemization of
(S)-1-phenylethanol (Table 2).
After 6 h, the enantiomeric
excess of 1-phenylethanol was
still as high as 95%, indicating
that almost no racemization oc-
0.54
5.7
5.6
–
–
–
[a] Determined by inductively coupled plasma (ICP) analysis.
curred. In comparison, Zr-beta and Ti-beta gave moderate
racemization rates. After 6 h, the enantiomeric excess (ee)
of 1-phenylethanol was 5% at 97% conversion and 22% at
79% conversion. The selectivity to racemic 1-phenylethanol
Table 2. Racemization of 1-phenylethanol by zeolite beta synthesized
with HF as mineralizer.[a]
Catalyst
Cat. wt.
[mg]
Time
[h]
Conversion
[%]
ee
[%]
Selectivity
[%]
Zr-beta
Sn-beta
Ti-beta
ZrAl-100
ZrAl-25
100
100
100
100
100
6
6
6
6
2
97
5.4
79
96
99
5
95
22
5
72
100
88
78
43
3
[a] Reaction conditions: 0.25 mmol (S)-1-phenylethanol, 5 mL toluene,
608C.
Figure 1. X-ray diffractograms of calcined a) Zr-beta, b) Sn-beta, c) Ti-
beta, and d) Al-150.
was reasonably high, 72% for Zr-beta and 88% for Ti-beta.
Styrene is the other product formed. Incorporation of a
small amount of Al into Zr-beta, as in ZrAl-100 (Si/Zr 100,
Si/Al 100) reduced the racemization selectivity slightly to
70%, whereas a higher Al content in ZrAl-25 (Si/Zr 100, Si/
Al 25) resulted in a faster reaction rate, but this was offset
by substantially more styrene formation.
The effect of the Al content in the zeolite beta on the
phenylethanol racemization was studied further to see if an
optimal balance could be reached between reaction rate and
selectivity. Zeolite beta materials synthesized by using HF as
a mineralizer have few connectivity defects, such as Si=O
À
and Si OH groups and, thus, are more hydrophobic than
Figure 2. Pyridine absorption spectra recorded after evacuation at 1008C
for 1 h for a) Sn-beta, b) Ti-beta, c) Zr-beta, and d) Al-150.
those formed under alkaline synthesis conditions. This is
confirmed by the results of thermogravimetric analysis
(TGA) that revealed a higher weight loss for Al-beta syn-
thesized by using TEAOH than for Zr-beta for which HF
was used as the mineralizer. By isotope labeling, Klomp
et al.[13] showed that surface hydroxyl groups of zeolite beta
participate in the hydration/dehydration of the alcohol.
These groups, being acidic in nature, serve as the active sites
for the racemization.
~1580 cmÀ1, and 1600–1633 cmÀ1 only. In contrast, the ab-
sorption band at 1540 cmÀ1, indicative of the pyridinium ion,
was detected in Zr-beta and Al-150. This band shows the
presence of Brønsted-acid sites in the samples.
Thermogravimetric measurements on zeolite beta synthe-
sized by using hydrogen fluoride (HF) indicated a weight
loss of 1.7–1.8 wt% over the temperature range 100–8008C.
Results of mass spectrometry show that this weight loss is
entirely due to the removal of water. Zeolite beta synthe-
sized under basic conditions lost about 5.5 wt% over the
same temperature range, indicating that these materials con-
A number of Al-beta samples synthesized in a basic
medium by using TEAOH were tested for their racemiza-
tion activity (Table 3). Al-beta with Si/Al 150 was very
active for the reaction. (S)-1-Phenylethanol was nearly com-
pletely racemized after only 30 min, whereas 6 h were re-
quired over Zr-beta.
Chem. Eur. J. 2007, 13, 541 – 547
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
543

S. Jaenicke et al.
Table 3. Racemization of 1-phenylethanol by zeolite beta prepared in
the basic medium.[a]
penyl acetate was chosen as the acyl donor because the ace-
tone that is liberated in the transesterification reaction is
easily removed from the system. Moreover, it is compatible
with the Zr-beta racemization catalyst. In experiments with-
out the racemization catalyst, the reaction comes to a com-
plete standstill at 50% conversion, and the unreactive (S)-1-
phenylethanol can be recovered in good yield and >99%
ee. Upon combining the racemization reaction with the
lipase-catalyzed reaction, the 50% conversion limit was ex-
ceeded after 2 h (Figure 4). After 48 h, as well as the acylat-
Catalyst
Cat. wt.
[mg]
Time
[h]
Conversion
[%]
ee
[%]
Selectivity
[%]
Al-12.5[b]
Al-75[c]
Al-150
10
10
100
100
10
2
10
10
0.25
1
99
97
98
97
98
92
95
0
2
2
3
4
2
9
5
100
79
86
61
60
84
90
84
–
0.5
0.5
0.5
1.5
1
Al-150[d]
Al-150[d]
Al-150[d]
Al-200
Al-300
6
[a] Reaction conditions: 0.25 mmol (S)-1-phenylethanol, 5 mL toluene,
608C. [b] Zeolyst CP814E. [c] Zeolyst CP811E-150. [d] Zeolyst CP811C-
300.
However, the selectivity to racemic 1-phenylethanol was
only 60%, with the dehydration to styrene as a major side
reaction. The activity of Al-150 was very similar to that of a
commercial Al-zeolite beta with the same Si/Al ratio (Zeo-
lyst CP811C-300). Reduction in the amount of Al-beta used
in the racemization improved the selectivity. By using 2 mg
instead of 100 mg, the selectivity to racemic 1-phenylethanol
could be increased to 90% at a threefold increase in the re-
action time (Figure 3). A further decrease in the aluminium
Figure 4. Dynamic resolution of 1-phenylethanol over Zr-beta. Left-hand
&
^
y axis: ( ) conversion of 1-phenyethanol, ( ) selectivity to 1-phenylethyl
~
*
acetate, and ( ) selectivity to styrene (%). Right-hand y axis: ( ) ee of
1-phenylethyl acetate (%). Reaction conditions: 608C, 1 mmol phenyle-
thanol, 2 mmol isoproprenyl acetate, 30 mg of Novozym 435, 0.4 g Zr-
beta.
ed product, 1-phenylethyl acetate, two byproducts were also
detected; one is styrene and the other is acetophenone. The
selectivity to styrene was about 10%. Acetophenone is
formed through the Oppenauer oxidation of 1-phenyletha-
nol, in which acetone, the leaving group of isopropenyl ace-
tate, is the oxidizing agent. However, although Bäckvall
et al.[6b] reported that with dimeric Ru(II) complex as race-
mization catalyst, 28% of 1-phenylethanol was converted to
acetophenone, in our experiments over Zr-beta, the selectiv-
ity to acetophenone never exceeded 1%. This shows that
racemization of (S)-1-phenylethanol over Zr-beta occurs by
means of an elimination/hydration pathway rather than in-
volving an Oppenauer oxidation.
Figure 3. Enantiomeric excess (filled symbols) and selectivity (open sym-
&
bols) (shown in %) of 1-phenylethanol upon racemization by ( , &)
~
~
*
*
) 2 mg of Al-zeolite beta (Si/Al 150). Re-
100 mg, (
,
) 10 mg, and (
,
action conditions: 0.25 mmol (S)-1-phenylethanol, 5 mL toluene, 608C.
content as realized in the catalyst Si/Al 200 did not signifi-
cantly improve the selectivity, but a longer reaction time
was needed. For Si/Al 300, no conversion was observed
even after 6 h. The results show that Zr-beta, Ti-beta, and
Al-beta with low metal content are suitable candidates for
application in a dynamic kinetic resolution.
As the reaction progressed, the ee of the product, 1-phe-
nylethyl acetate, decreased from almost 100% at up to 50%
conversion to about 83% at >90% conversion. This reduc-
tion in ee could be due to several reasons: 1) the selectivity
of the kinetic resolution is too low so that (S)-1-phenyletha-
nol is acylated as well; 2) Zr-beta catalyzes nonselective
transesterification; or 3) Zr-beta racemizes the acylated
product, (R)-1-phenylethyl acetate. The first mechanism can
be ruled out, as experiments showed that Novozym 435 cata-
lyzes the transesterification of (R)-1-phenylethanol at a very
much higher rate than for the (S)-isomer. However, Zr-beta
can indeed catalyze the acylation of 1-phenylethanol, albeit
at a very low reaction rate. In a separate experiment, (R)-1-
Dynamic kinetic resolution
Zr-beta: The enzyme used for the transesterification of 1-
phenylethanol and isopropenyl acetate was lipase B from
Candida antarctica, supported on an acrylic resin (commer-
cially available under the trade name Novozym 435). Isopro-
544
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 541 – 547

Kinetic Resolution of Secondary Alcohols
FULL PAPER
Table 5. Effect of temperature on the dynamic resolution of 1-phenyle-
thanol.[a]
phenylethyl acetate was slowly racemized by Zr-beta.
Hence, the drop in ee of the product especially after long re-
action times can be attributed to both the Zr-beta catalyzed
transesterification of 1-phenylethanol and the racemization
of 1-phenylethyl acetate.
Entry
Temperature
[8C]
Time
[h]
Conv.
[%][b]
Yield
[%][c]
ee
[%][d]
1
2
3
4
50
60
70
80
24
48
24
48
6
24
6
24
84
89
89
93
89
95
87
95
75
79
79
82
76
82
82
81
84.9
84.2
84.8
83.1
78.4
79.4
76.8
74.9
Effect of solvents: The dynamic kinetic resolution was evalu-
ated in several organic solvents of different polarity. The
best results were obtained by using toluene as solvent, for
which an ee value of 83% at 93% conversion was achieved
after 48 h (Table 4). The nonpolar solvent, n-hexane, gave
[a] Reaction conditions: 608C, 1 mmol phenylethanol, 2 mmol isopro-
prenyl acetate, 30 mg Novozym 435, 0.4 g Zr-beta. [b] Conversion of race-
mic 1-phenylethanol. [c] Yield of 1-phenylethyl acetate. [d] ee of (R)-1-
phenylethyl acetate.
Table 4. Effect of solvents on the dynamic kinetic resolution of 1-phenyl-
ethanol over Zr-beta.[a]
Entry
Solvent
Time
[h]
Conv.
[%][b]
Yield
[%][c]
ee
[%][d]
1
2
3
4
5
toluene
6
48
6
48
6
48
6
48
6
78
93
63
84
47
67
58
69
43
64
71
82
59
76
40
53
53
60
38
46
87.5
83.1
94.3
81.6
98.4
97.0
96.8
92.3
>99.5
>99.5
n-hexane
1,4-dioxane
isopropyl ether
acetonitrile
48
[a] Reaction conditions: 608C, 1 mmol phenylethanol, 2 mmol isopro-
prenyl acetate, 30 mg Novozym 435, 0.4 g Zr-beta. [b] Conversion of race-
mic 1-phenylethanol. [c] Yield of 1-phenylethyl acetate. [d] ee of (R)-1-
phenylethyl acetate.
Figure 5. Dynamic resolution of 1-phenylethanol at different tempera-
~
&
*
^
tures: ( ) 508C , ( ) 608C , ( ) 708C, and ( ) 808C.
results that were almost as good; 84% conversion and
81.6% ee after the same reaction time. A high ee was also
obtained in both 1,4-dioxane and isopropyl ether, although
the conversion and yield were low due to a slow racemiza-
tion rate. In the polar solvent, acetonitrile, no racemization
took place and the conversion did not increase above 50%.
Presumably, the Lewis-basic acetonitrile blocks the Zr-beta
sites required for the racemization reaction.
increased to 808C. Hence, 608Cwas established to be the
optimum reaction temperature for this reaction system.
Optimization of the racemization catalyst and the acyl
donor: Because the racemization over Al-beta (Si/Al 150)
was faster than over Zr-beta, the transesterification of 1-
phenylethanol and isopropenyl acetate in the presence of
lipase was carried out by using this catalyst. The racemiza-
tion catalyst was added after 2 h, when the conversion had
reached 50%. The transesterification reaction resumed after
addition of the racemization catalyst, and after 17.5 h, the
conversion had reached 97% (Figure 6). However, more
styrene was formed, 20%, in contrast to the 10% formed
over Zr-beta. In addition, racemization of the ester, 1-phe-
nylethyl acetate, occurred to a greater extent so that at the
end of the reaction, the enantiomeric excess was only 67%.
Hence, despite the faster racemization rate over Al-beta rel-
ative to Zr-beta, Al-beta appeared not to be an effective
catalyst for the dynamic kinetic resolution of 1-phenyetha-
nol if isopropenyl acetate was used as the acyl-transfer re-
agent. We reasoned that the active sites that catalyzed the
racemization of the alcohol were the same as those capable
of bringing about racemization of the ester. If the size of the
ester molecule is increased, it should be possible to prevent
its diffusion into the pore space where the active sites are lo-
Effect of temperature: Temperature is another important pa-
rameter that must be optimized in performing a dynamic
resolution. The racemization is usually faster at increased
temperatures. However, it is frequently observed that the
enzyme loses selectivity or even undergoes denaturation at
higher temperatures. Moreover, additional side reactions
may occur. To find the optimum reaction temperature for
the kinetic resolution of 1-phenylethanol catalyzed by Zr-
beta and Novozym, the reaction was studied at temperatures
ranging from 50 to 808C(Table 5). At 50 8C, the conversion
of 1-phenylethanol remained below 90% even after 48 h.
The rate of reaction was fastest at 808C, with more than
95% of 1-phenylethanol converted after 24 h. However, at
the elevated temperature, the ee of the 1-phenylethanol ace-
tate decreased to 75%. In Figure 5, the ee of (R)-1-phenyl-
A
apparent that the enantioselectivity is somewhat higher at
508Cand 60 8C, and decreases notably if the temperature is
Chem. Eur. J. 2007, 13, 541 – 547
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
545

S. Jaenicke et al.
Conclusion
The dynamic kinetic resolution of secondary alcohols was
achieved in a single nonaqueous liquid phase in the pres-
ence of two heterogeneous catalysts. An immobilized lipase
(C. antarctica lipase B, Novozym 435) catalyzes the stereose-
lective transesterification of the secondary alcohol with acyl-
transfer reagents, such as isopropenyl acetate, vinyl acetate,
vinyl butyrate, or vinyl octanoate, under conditions that are
compatible with the catalytic racemization of the secondary
alcohol over zeolite beta. Pure Si-beta or substitution with
Sn gives catalysts that have almost no activity for the race-
mization reaction. Hydrophobic Zr-beta prepared by a hy-
drothermal synthesis in fluoride medium or Al-beta with a
high Si content are good racemization catalysts. To create
the active sites, Zr and Al have to be incorporated into the
framework at a doping level of about 1%. Styrene is formed
as the major side product. Racemization catalysts with
higher Al content lead to increased styrene formation.
The dynamic kinetic resolution of secondary alcohols was
best carried out in a nonpolar medium, preferably toluene
or n-hexane, rather than in a polar solvent. The optimum
temperature was 608C. Apart from racemizing the reacted
alcohol, the zeolite catalyst also caused a loss of the enantio-
meric purity by racemization of the product ester. However,
this could be substantially suppressed by using larger acyl-
transfer reagents, such as vinyl butyrate or vinyl octanoate.
The resulting bulky ester is not able to enter the active sites
located in the pore channels of the zeolite catalyst and,
therefore, does not undergo subsequent racemization. In
this way, the (R)-ester can be prepared at >85% yield and
>97% ee.
Figure 6. Dynamic resolution of 1-phenylethanol by using Al-150. Left-
&
^
hand y axis: ( ) conversion of 1-phenyethanol, ( ) selectivity to 1-phe-
~
nylethyl acetate, and ( ) selectivity to styrene (%). Right-hand y axis:
*
(
) ee of 1-phenylethyl acetate (%). Reaction conditions: 1 mmol phe-
nylethanol, 1.5 mmol isopropenyl acetate, 30 mg enzyme, 50 mg Al-150,
5 mL toluene, 608C.
calized, and thereby improve the enantiomeric excess of the
final product.
To test this hypothesis, we carried out the transesterifica-
tion reactions with the bulkier acyl donors, vinyl butyrate
and vinyl octanoate (Figure 7). The racemization catalyst,
Experimental Section
Synthesis of Sn-, Zr-, and Ti-beta zeolites: Beta zeolites containing Sn
and Zr were synthesized in a fluoride-assisted hydrothermal synthesis
with the addition of seeds from dealuminated zeolite beta. Typically, tet-
raethylorthosilicate (TEOS) was hydrolyzed in a solution of tetraeth-
Figure 7. Dynamic resolution in the transesterification between 1-phenyl-
ethanol and various esters. The y axis represents selectivity (empty bars),
yield (hatched bars), and ee (black bars) of the ester (%). Reaction con-
ditions: 1.0 mmol racemic 1-phenylethanol, 1.5 mmol acyl-transfer re-
agent, 30 mg Novozym 435, 50 mg zeolite Al-150, 5 mL toluene, 608C.
A
SnCl4·5H2O or ZrOCl2·8H2O in water was added, and the mixture was
stirred until all ethanol formed by hydrolysis of the TEOS had evaporat-
ed.[15,16] Hydrogen fluoride and an aqueous suspension of seeds of dealu-
minated nanocrystalline zeolite beta (ca. 50 nm in diameter, as deter-
mined by TEM) were added. Crystallization was carried out in a static
Teflon-lined stainless-steel autoclave at 1408Cfor 25 days. The solid
product was filtered, washed with deionized water, dried at 1008C, and
activated at 5808Cfor 4 h. Two Al-containing Zr-beta samples, ZrAl-100
(Si/Zrꢀ104, Si/Alꢀ103) and ZrAl-25 (Si/Zrꢀ105, Si/Alꢀ27) were syn-
thesized in an analogous way to the synthesis of Al-free Zr-beta, except
for the addition of the required amounts of ZrOCl2·8H2O and Al-
Al-150, was added after 50% conversion was reached. The
rate of transesterification observed for the vinyl esters was
faster than that for isopropenyl acetate, despite the bigger
butyrate or octanoate group. The conversion approached
100% within 150 min for both acyl donors, and the selectivi-
ty towards ester formation was 85–88%. The only byproduct
was styrene. The enantiomeric excess of the ester formed,
phenylethyl butyrate or phenylethyl octanoate, was 92 and
98%, respectively. This is substantially higher than for the
(R)-1-phenylacetate formed with vinyl acetate (65% ee) or
isopropenyl acetate (68% ee).
A
Ti-beta (Si/Tiꢀ100) was prepared by following the procedure described
by Blasco et al.[17] For a typical synthesis, titanium ethoxide (TEOT)
(0.12 mL) was added to TEOS (10.42 g), followed by the dropwise addi-
tion of a 40% TEAOH solution (10.31 g). To facilitate the dissolution of
TEOT, H2O2 (30% w/w; 1.93 mL) was added. The resulting mixture was
stirred until the ethanol formed upon hydrolysis of TEOS and TEOT
had completely evaporated. Hydrogen fluoride (HF) (Aldrich, 48%,
1.014 mL) was added to the cloudy solution, whereupon a thick paste
546
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 541 – 547

Kinetic Resolution of Secondary Alcohols
FULL PAPER
formed. Dealuminated nanocrystalline zeolite beta seeds (~50 nm,
0.105 g) in water (2.4 mL) were added. The crystallization was carried
out in a Teflon-lined stainless-steel autoclave at 1408Cfor 10 days.
washing with 0.6m NaHCO3 solution (3”20 mL). The organic layer was
dried with anhydrous Na2SO4 and styrene and hexane were removed by
vacuum distillation (1 mbar at 508Cfor 2 h). After work-up, essentially
pure (R)-1-phenylethyl octanoate was obtained with an isolated yield of
72% and 97.7% ee (average from two runs).
Synthesis of Al-beta zeolites: Beta zeolites with Si/Al 75–200 were pre-
pared according to reference [18]. Typically, aluminium metal was dis-
solved in TEAOH (40 wt%) followed by the addition of deionized water
and fumed silica (Aerosil 200). The mixture was stirred for 2 h before
being placed in a Teflon-lined stainless-steel autoclave and kept at 1408C
for 3 days under autogenous pressure. The molar composition of the final
Acknowledgements
gel mixture was Al2O3:xSiO2:(0.26x+1)TEA2O:15xH2O, for which x
ranged from 400 to 150. The samples are denoted Al-n, for which n
A
Financial support for this work from the National University of Singa-
pore grants R-143–000–188–112 and R-143–000–201–112 is gratefully ac-
knowledged. We thank Mr. Lester Poon, undergraduate exchange student
from the University of British Columbia, for his help with the scale-up
experiment.
stands for the Si/Al ratio (75–200).
Catalyst characterization: The surface area and porosity properties were
determined by using nitrogen adsorption (Micromeritics Tristar). Prior to
measurement, the samples were thoroughly degassed under a nitrogen
flow for 4 h. The formation of the crystalline phase was measured by
using powder X-ray diffractometry (Siemens D5005 with Cu anode and
variable slits). To determine the hydrophobicity of the samples in terms
of amount of water lost, the weight loss was monitored by thermogravi-
metric analysis (TGA) in a Dupont SDT 2960 analyzer. About 5–10 mg
of sample was first dried in a flow of dry nitrogen (100 mLminÀ1) before
[1] a) Developments in Global Chiral Technologies, Frost & Sullivan,
May 2003, cited in A. M. Rouhi, Chem. Eng. News 2004, 82, 47–62;
b) SusChem—A European Technology Platform for Sustainable
Chemistry: The Vision for 2025 and Beyond, March 2005; available
at http://www.suschem.org/media.php?mId=2170 (accessed July
2006).
[2] a) R. Noyori, T. Ikeda, T. Ohkuma, M. Widhalm, M. Kitamura, H.
Takaya, S. Akutagawa, N. Sayo, T. Saito, T. Taketomi, H. Kumo-
bayashi, J. Am. Chem. Soc. 1989, 111, 9134–9135; b) M. Kitamura,
M. Tokunaga, R. Noyori, J. Am. Chem. Soc. 1993, 115, 144–152.
[3] A. Zaks, A. Klibanov, Science 1984, 224, 1249–1251.
[4] a) S. Akai, K. Tanimoto, Y. Kita, Angew. Chem. 2004, 116, 1431–
1434; Angew. Chem. Int. Ed. 2004, 43, 1407–1410; ; b) O. Pꢂmies, J.-
E Bäckvall, Curr. Opin. Biotechnol. 2003, 14, 407–413; c) J. J. Venit,
G. D. Madding (Bristol Myers Squibb Company),WO-02/28809,
2002; d) S. F. Mayer, A. Steinreiber, R. V. A. Orru, K. Faber, J. Org.
Chem. 2002, 67, 9115–9121; e) K. Faber, Chem. Eur. J. 2001, 7,
5005–5010; f) U. T. Strauss, U. Felfer, K. Faber, Tetrahedron: Asym-
metry 1999, 10, 107–117.
[5] a) P. M. Dinh, J. A. Horwarth, A. R. Hudnott, J. M. J. Williams, W.
Harris, Tetrahedron Lett. 1996, 37, 7623–7626; b) T. H. Riermeier, P.
Gross, A. Monsees, M. Hoff, H. Trauthwein, Tetrahedron Lett. 2005,
46, 3403–3406.
[6] a) J. H. Koh, H. M. Jung, M.-J. Kim, J. Park, Tetrahedron Lett. 1999,
40, 6281–6284; b) B. A. Persson, A. L. E. Larsson, M. Le Ray, J.-E.
Bäckvall, J. Am. Chem. Soc. 1999, 121, 1645–1650; c) A. Dijksman,
J. M. Elzinga, Y.-X. Li, I. W. E. Arends, R. A. Sheldon, Tetrahedron:
Asymmetry 2002, 13, 879–884.
[7] M. T. Reetz, K. Schimossek, Chimia 1996, 50, 668–669.
[8] S. Y. Kalliney, M. V. Ruggeri (Schering-Plough Corporation), WO-
91/08196, 1991.
[9] D. W. House (UOP), US-5476964, 1995.
[10] P. ꢃdman, L. A. Wessjohann, U. Bornscheuer, J. Org. Chem. 2005,
70, 9551–9555.
[11] S. Wuyts, D. E. De Vos, F. Verpoort, D. Depla, R. D. Gryse, P. A.
Jacobs, J. Catal. 2003, 219, 417–424.
[12] S. Wuyts, K. De Temmerman, D. E. De Vos, P. A. Jacobs, Chem.
Eur. J. 2005, 11, 386–397.
[13] D. Klomp, T. Maschmeyer, U. Hanefeld, J. A. Peters, Chem. Eur. J,
2004, 10, 2008–2093.
[14] S. Wuyts, K. De Temmerman, D. De Vos, P. Jacobs, Chem. Commun.
2003, 1928–1929.
increasing the temperature to 8008Cat 10 8CminÀ1
.
IR spectroscopy of adsorbed pyridine was employed to determine the
type of acidic sites present on the zeolites. The IR spectra were recorded
by using a Bruker Equinox 55 with a resolution of 2 cmÀ1. Typically, sam-
ples were pressed into self-supported wafers of 8–10 mg. The wafer was
mounted inside a Pyrex IR cell with NaCl windows and dried under
vacuum (~10À3 mbar) for 2 h at 3008C. After cooling to RT, a back-
ground spectrum was recorded. Subsequently, the sample was exposed to
pyridine at 22 mbar for 15 min. The pyridine was pumped off and the
sample was evacuated under vacuum (~10À3 mbar) for 1 h before an IR
spectrum was recorded. The sample was then heated to 1008Cfor 1 h
under vacuum to desorb loosely bound pyridine, cooled to RT, and an-
other IR spectrum was recorded. A further measurement was taken after
desorption at 2008C.
Racemization of (S)-1-phenylethanol: The racemization of (S)-1-phenyle-
thanol was carried out in a two-necked round-bottomed flask equipped
with septum port, reflux condenser, and a guard tube. Typically, (S)-1-
phenylethanol (0.25 mmol; ee>99.5%) was dissolved in toluene (5 mL)
and heated to 608C. Catalyst (100 mg) was added to the reaction mixture
(time=0) Progress of the reaction and enantiomeric excess were moni-
tored by performing gas chromatography (column: Supelco Beta Dex 325
(250 mm”0.25 mm”25 m); flame ionization detector). Only the two iso-
mers of 1-phenylethanol, and the byproduct, styrene, were observed.
Dynamic kinetic resolution of 1-phenylethanol: The zeolite catalyst (50–
400 mg) was added to a reaction mixture of 1 mmol racemic (R,S)-1-phe-
nylethanol in solvent (5 mL) and Novozym 435 (30 mg). The reaction
slurry was stirred and heated to the desired temperature, normally 608C,
although temperatures of 50–808Cwere also studied. At time t=0, iso-
propenyl acetate (2 mmol) was added. The progress of the reaction was
monitored by performing gas chromatography using a Supelco Beta
Dex 325 chiral capillary column. The identity of the products was verified
by comparing the retention times and GC/MS spectra with authentic
samples. As well as isopropenyl acetate, other esters, such as vinyl ace-
tate, vinyl butyrate, and vinyl octanoate, were also used as acyl-transfer
reagents. A Supelco Beta Dex 120 column was used for separating the
enantiomers of phenyl butyrate and phenyl octanoate. Close to 100%
conversion was reached after about 2 h of reaction at 608C.
Scale-up to a semipreparative scale: 1-Phenylethanol (1.22 g, 10 mmol)
and vinyl octanoate (1.70 g, 10 mmol) were dissolved in toluene (5 mL),
and Novozyme 435 (150 mg) was added after the temperature had
reached 608C. After 2 h, zeolite beta Al-150 (250 mg) was added to the
reaction mixture and the reaction was allowed to proceed for another
4 h. At completion of reaction, the solid catalysts were removed by filtra-
tion, and toluene was distilled off under reduced pressure. The product
containing mainly 1-phenylethyl octanoate and side products, such as
styrene and octanoic acid (due to the hydrolysis of vinyl octanoate), were
redissolved in hexane (20 mL) and the octanoic acid was removed by
[15] A. Corma, M. E. Domine, S. Valencia, J. Catal. 2003, 215, 294–304.
[16] Y.-Z. Zhu, G. K. Chuah, S. Jaenicke, J. Catal. 2004, 227, 1–10.
[17] T. Blasco, M. A. Camblor, A. Corma, P. Esteve, J. M. Guil, A. Marti-
nez, J. A. Perdigꢄn-Melꢄn, S. Valencia, J. Phys. Chem. B 1998, 102,
75–88.
[18] M. A. Camblor, A. Corma, A. Mifud, J. Pꢅrez-Pariente, S. Valencia,
Stud. Surf. Sci. Catal. 1997, 105, 341–348.
Received: May 24, 2006
Published online: September 27, 2006
Chem. Eur. J. 2007, 13, 541 – 547
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
547