Cascade approach for synthesis of R-1-phenyl ethyl acetate from acetophenone: Effect of support

Article abstract of DOI:10.1016/j.molcata.2008.01.032

One-pot synthesis of R-1-phenylethyl acetate starting from acetophenone hydrogenation was studied over supported palladium catalysts in combination with an immobilized lipase. Ethyl acetate was used as an acyl donor and the experiments were carried out under 1 bar hydrogen at 70 °C. In this work, the effect of support acidity was systematically investigated by using two mesoporous Pd-H-MCM-41 and Pd-Si-MCM-41 catalysts as hydrogenation catalysts. Furthermore their performance was compared with the performances of Pd-SiO₂ and Pd-Al₂O₃. The catalysts acidity was determined by pyridine adsorption while the metal crystallite sizes were measured by XRD technique. The initial hydrogenation rates of acetophonene decreased with increasing Pd crystallite sizes. Initially both R- and S-1-phenylethanol were formed with the same rates. After prolonged reaction times activities of both hydrogenation catalyst and lipase declined. The former catalyst deactivated due to formation of acetic acid via deacylation reaction, whereas ethanol formed as a stoichiometric product in acylation decreased the lipase activity. The formation of ethylbenzene via dehydration of an intermediate alcohol, (R,S)-1-phenylethanol was enhanced by acidic catalysts. The most selective catalyst for one-pot synthesis of R-1-phenylethylacetate was a slightly acidic Pd-Al₂O₃ in combination with lipase. A reaction network was proposed based on the kinetic data.

Full text of DOI:10.1016/j.molcata.2008.01.032

Available online at www.sciencedirect.com
Journal of Molecular Catalysis A: Chemical 285 (2008) 132–141
Cascade approach for synthesis of R-1-phenyl ethyl acetate
from acetophenone: Effect of support
a
a
a
b
¨
¨
¨
Paivi Maki-Arvela , Serap Sahin , Narendra Kumar , Teemu Heikkila ,
Vesa-Pekka Lehtob, Tapio Salmia, Dmitry Yu. Murzina,∗
a
˚
Abo Akademi University, Process Chemistry Centre, FI-20500 Turku, Finland
b University of Turku, Department of Physics, University of Turku, Finland
Received 16 November 2007; received in revised form 12 January 2008; accepted 15 January 2008
Available online 2 February 2008
Abstract
One-pot synthesis of R-1-phenylethyl acetate starting from acetophenone hydrogenation was studied over supported palladium catalysts in
combination with an immobilized lipase. Ethyl acetate was used as an acyl donor and the experiments were carried out under 1 bar hydrogen at
70 ◦C. In this work, the effect of support acidity was systematically investigated by using two mesoporous Pd–H–MCM-41 and Pd–Si–MCM-41
catalysts as hydrogenation catalysts. Furthermore their performance was compared with the performances of Pd–SiO2 and Pd–Al2O3. The catalysts
acidity was determined by pyridine adsorption while the metal crystallite sizes were measured by XRD technique. The initial hydrogenation rates
of acetophonene decreased with increasing Pd crystallite sizes. Initially both R- and S-1-phenylethanol were formed with the same rates. After
prolonged reaction times activities of both hydrogenation catalyst and lipase declined. The former catalyst deactivated due to formation of acetic acid
viadeacylationreaction, whereasethanolformed asa stoichiometric product in acylation decreased the lipase activity. The formation ofethylbenzene
via dehydration of an intermediate alcohol, (R,S)-1-phenylethanol was enhanced by acidic catalysts. The most selective catalyst for one-pot synthesis
of R-1-phenylethylacetate was a slightly acidic Pd–Al2O3 in combination with lipase. A reaction network was proposed based on the kinetic data.
© 2008 Elsevier B.V. All rights reserved.
Keywords: One-pot synthesis; Hydrogenation; Acylation; Heterogeneous catalyst
1. Introduction
reaction proceeds relatively fast [2]. If the idea is to start from
acetophenone hydrogenation in one-pot, vinyl acetate is not suit-
able acyl donor in the second step, since it is hydrogenated
instantaneously. Acylation proceeds slightly slower with ethyl
acetate, but since it is also suitable as solvent in the hydrogena-
tion step [2], it has been used in the current work. Thus, the
idea in the present work is to hydrogenate acetophenone selec-
tively to (R,S)-1-phenylethanol with molecular hydrogen over
a heterogeneous catalyst and to acylate with the enzyme the
formed R-1-phenylethanol to the corresponding ester, namely R-
1-phenyl ethyl acetate. According to our knowledge this concept
has not been investigated before. On the other hand, a continuous
hydrogenation–acylation was demonstrated in two consecutive
reactors [4] or by using molecular hydrogen and a homogeneous
metal complex as a hydrogenation and racemization catalyst,
respectively [2].
Cascade catalysis is a very intensively studied research area,
being of great interest both in academia and in industry [1].
The idea is to perform several reactions in one reactor pot, thus
saving equipment and separation costs in producing fine chem-
icals. One interesting model reaction is selective hydrogenation
of acetophenone to racemic alcohol (R,S)-1-phenylethanol. The
formed R-1-phenylethanol can be acylated with very high enan-
tiomeric excess over enzymes, like lipase to produce chiral esters
[2]. One of the simplest ester molecules is R-1-phenyl ethy-
lacetate, which can be very selectively acylated over lipases
according to Kazlauskasrule [3], sincethereexistsalargephenyl
group and a medium sized substituent, a hydroxyl group. Typ-
ically vinyl acetate has been used as an acyl donor and the
Selective hydrogenation of acetophenone is the first step to
proceed in one-pot synthesis of R-1-phenylethyl acetate. The
hydrogenation reaction gives as products a racemic mixture of
∗
Corresponding author. Tel.: +358 2 215 4985.
E-mail address: dmurzin@abo.fi (D.Yu. Murzin).
1381-1169/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2008.01.032

P. Ma¨ ki-Arvela et al. / Journal of Molecular Catalysis A: Chemical 285 (2008) 132–141
133
(R,S)-1-phenylethanol. Pd supported catalysts have been very
selective to keep the phenyl ring intact in the hydrogenation
of acetophenone [5], whereas over Pt catalyst also cyclohexyl
products have been formed [6]. One potential side reaction using
Pd supported catalysts, is however, hydrogenolysis of (R,S)-1-
phenylethanol [5], which can be suppressed by adding bases
into the reaction mixture or washing the catalysts with sodium
hydroxide. Hydrogenolysis of (R,S)-1-phenylethanol is accord-
ing to literature [7] catalyzed both by Brønsted and Lewis acid
sites.
In the second step, the R-enantiomer is acylated to R-1-
phenylethyl acetate. The S-enantiomer is not reacting over
lipases [8]. Racemization of S-enantiomer has been performed
over zeolites [9,10] or by using homogeneous complexes [2].
Zeolites are not suitable to use as racemization catalysts for
S-1-phenylethanol in this work, since they promote acid cat-
alyzed dehydration of (R,S)-1-phenylethanol to styrene [5].
There might be, however, other possibilities to racemize S-1-
phenylethanol, by applying for instance vanadyl sulfate [11] and
Ru-hydroxyapatite [12], which are not very acidic and thus in
the presence of Pd and hydrogen do not catalyze dehydration of
(R,S)-1-phenylethanol.
Solution A was prepared by mixing fumed silica (Aldrich) with
distilled water under continuous stirring. Solution B was pre-
pared by adding tetramethylammonium silicate (Sachem) to
sodium silicate (Merck) and stirring for 15 min. Solution C was
prepared by dissolving tetradecyl trimethyl ammonium bromide
(Aldrich) in distilled water. Solution B was added to solu-
tion A slowly and stirred for 20 min, subsequently solution C
was introduced under vigorous stirring. After measuring pH of
the prepared gel it was introduced in a teflon cup, which was
then inserted in an autoclave. The synthesis was performed at
100 ◦C in an oven. After completion of the synthesis, the reac-
tor was quenched and the mesoporous material was filtered and
washed thoroughly with distilled water. Synthesized Si–MCM-
41 was dried at 110 ◦C and calcined at 550 ◦C. 5 wt% Pd loaded
Pd–Si–MCM-41 and Pd–H–MCM-41 catalysts were prepared
by vacuum evaporation impregnation method in a rotator evap-
orator (Buchi) using aqueous solution of palladium nitrate as
precursor for palladium. The catalysts were dried at 110 ◦C and
calcined at 400 ◦C in a muffle oven.
2.2. Catalyst characterization
The specific surface area was measured by nitrogen adsorp-
tion using Carlo Erba 1900 Instrument. The catalysts were out
gassed at 150 ◦C for 3 h prior to the surface area measurements.
The surface areas of catalysts were calculated with BET equa-
tion.
The aim in this paper was to study four different hydrogena-
tion catalysts in one-pot synthesis of R-1-phenylethyl acetate,
namelyPd–Al2O3, Pd–SiO2 aswellastwomesoporouscatalysts
being an acidic Pd–H–MCM-41 and a non-acidic Pd–Si–MCM-
41. The correlation between the support acidity and the catalytic
performance was established. Based on the kinetic data the reac-
tion network was proposed.
X-Ray powder diffraction was used for characterization of
the mesoporous phase and measuring the Pd crystallite size
with Philips PW1820 diffractometer using nickel filtered Cu
˚
K␣ (λ = 1.542 A) radiation operated at 40 kV/50 mA. The diver-
2. Experimental
gence of the primary X-ray beam was limited by an automatic
divergence slit (ADS) and a 15 mm mask. The irradiated sample
length was set at a fixed 12 mm. On the diffracted side there was
a 0.2 mm receiving slit and a 1◦ anti-scatter slit. The measured
diffractograms were analyzed using X’Pert HighScore software
(Philips, 2001) and the Powder Diffraction File (PDF) database
(PDF-2 sets 1–46, ICDD, 1996).
Temperature programmed desorption techniques was used to
study the amount of hydrogen desorbed from some catalysts.
The heating was performed under helium flow up to 700 ◦C for
60 min with the temperature ramp 10 ◦C per min (Micromerit-
ics, Autochem 2910). The catalysts were reduced at 100 ◦C for
30 min under hydrogen, thereafter they were cooled down to the
ambient temperature.
2.1. Catalyst synthesis
Five weight percent Pd–Al2O3 (UOP) and 5 wt.% Pd–SiO2
(Merck) catalysts were prepared by vacuum evaporation
impregnation method in a rotary evaporator using an aque-
ous solution of palladium nitrate (Degussa) as a precursor.
The catalysts were dried at 110 ◦C and calcined in a muffle
oven.
Synthesis of the Na/MCM-41 mesoporous molecular sieve
was carried out in 300 ml autoclave using a method described
in Refs. [13–15] with the following chemicals: fumed silica
(Aldrich), tetramethylammonium silicate (TMSiO2, Aldrich),
sodium silicate (Na2SiO3), Merck), tetradecyl trimethyl
ammonium bromide (CH3(CH2)13N(CH3)3Br) Aldrich) and
aluminium isopropoxide ([CH3)2CHO]3Al), Aldrich) as alu-
minium source. The Si/Al ratio was 20 for Na–H–MCM-41.
The synthesis was carried out at 100 ◦C. After filtering the syn-
thesized mesoporous material Na/MCM-41 was dried at 110 ◦C
and calcined at 530 ◦C. Ion-exchange of Na–MCM-41 was car-
ried out with 1 M NH4Cl for 24 h followed by washing with
distilled water, drying at 110 ◦C and calcination at 450 ◦C for
4 h.
Scanning electron microscopy pictures were taken with 360
LEO Electron Microscopy LTD equipped with a secondary and
backscattered electron detector from Na–MCM-41 for revealing
the particle morphology.
The acidity of Al2O3 (UOP) and H–MCM-41 was studied
by pyridine adsorption using infrared spectroscopy (ATI Matt-
son FTIR). Thin wafers (10–12 mg/cm2) were prepared from
support materials. The pyridine (>99.5%, a.r.) was adsorbed
at 100 ◦C for 30 min and desorbed at 200 ◦C. The spectra
were recorded at 100 ◦C with a spectral resolution of 2 cm−1
.
Synthesis of the Si–MCM-41 mesoporous molecular sieve
was carried out in a 300 ml autoclave. The synthesis was per-
formed by preparing solutions named here as A, B and C.
Molar extinction coefficient for pyridine was determined by
[16].

134
P. Ma¨ ki-Arvela et al. / Journal of Molecular Catalysis A: Chemical 285 (2008) 132–141
Table 1
Catalyst characterization results
Catalyst
BET specific surface
Concentration of Brønsted
Concentration of Lewis
acid sites (␮mol/gsupport)
Average Pd crystallite
size with XRD (nm)
area (m2/gcat
)
.
acid sites (␮mol/gsupport
)
Pd–H–MCM-41
Pd–Al2O3
Pd–Si–MCM-41
Pd–SiO2
902
306
379
351
89
7
0
168
156
0
3.9
7.1
7.3
n.m.
0
0
n.m.: not measured.
2.3. Experimental setup
lowing Pd peaks were found: Pd(1 1 1), at 2( 39.962 ◦C, peak
Pd(2 0 0) at 47.208 ◦C, peak Pd(2 2 0) at 68.083 ◦C and Pd(3 1 1)
at 81.960 ◦C at 81.960. An average Pd crystallite size for this
catalyst was 3.9 nm. Analogously for Si–MCM-41 the average
Pd crystallite size was determined from XRD (Fig. 1b) being
7.3 nm. The same phases were observed for Pd–Si–MCM-41
as for Pd–H–MCM-41, but the angles were slightly shifted to
larger levels for the former catalyst. For reduced Pd–Al2O3 cat-
alyst only three forms of Pd were observed namely Pd(1 1 1),
40.086◦2(, Pd(2 0 0), 46.148◦2( and Pd(3 1 1), 82.173◦2( giving
an average Pd crystallite size of 7.1 nm. Pd particles were smaller
in the mesoporous Pd–H–MCM-41 exhibiting both a higher
concentration of Brønsted acid sites as well as 2.9-fold larger
BET surface area than that of Pd–Al2O3. This result indicated
that the larger BET surface area and the higher concentration
of Brønsted acid sites facilitated the preparation of more dis-
persed Pd particles over H–MCM-41 than over Pd–Al2O3 since
Experiments were performed in a glass reactor under atmo-
spheric hydrogen (AGA 99.999) with 295 ml/min volumetric
flow rate at 70 ◦C. The stirring speed was 370 rpm. Ethyl acetate
(Fluka, >99%) was used as a solvent and as an acyl donor. The
initial acetophenone concentration was 0.02 mol/l. The liquid
phase volume was 250 ml. The masses of both hydrogenation
catalyst and the immobilized enzyme catalysts (Novozym 435)
were 125 mg, if other amounts were not specified. The hydro-
genation catalyst was pre-reduced with flowing hydrogen at
100 ◦C for 30 min prior injection of the lipase, the reactant and
the solvent.
2.4. Product analysis
The products were analyzed by
a gas chromato-
graph equipped with a chiral column (CP Chirasil Dex
(250 ␮m × 250 ␮m × 25 m) and a flame ionization detector.
The samples were analyzed by using the following temperature
programme 100 ◦C (1 min)–0.30 ◦C/min – 130 ◦C – 15 ◦C/min
– 200 ◦C (10 min). The temperature of the injector and the split
ratio were 280 ◦C and 100:1, respectively. The GC-method was
calibrated with the following chemicals: (R)-1-phenyl ethanol
((R)-1-PE) (Sigma), (R,S)-1-phenyl ethanol (Fluka ≥ 98%) and
ethyl benzene (Fluka, >99%). The products were identified
with GC–MS. In some cases the water content in the reaction
mixture was measured with Karl Fischer titration by Hydranal
¨
(Riedel de Haen, Composite 5).
3. Results and discussion
3.1. Catalyst characterization results
The specific surface areas of the four catalysts are given in
Table 1. The largest BET specific surface area was obtained for
Pd–H–MCM-41, whereas all the other three catalysts exhibited
the BET surface areas in a range of 306–379 m2/gcat
.
The existence of mesoporous phase of H–MCM-41 was
revealed by the following characteristic peaks of MCM-41,
namely the main peak at 2.56◦, the second peak at 4.41◦, third
peak at 5.04◦ and the fourth peak at 7.1◦. In Si–MCM-41 the cor-
responding peaks were observed at 2.52◦, 4.39◦ and 5.04◦. The
fourthcharacteristicpeakcouldnotbediscernedinSi–MCM-41.
The metal particle sizes were determined by XRD for
Pd–Al2O3 and for Pd–H–MCM-41 (Table 1, Fig. 1a). For
Pd–H–MCM-41 catalyst reduced at 100 ◦C for 30 min the fol-
Fig. 1. The XRD pattern of the reduced (a) Pd–H–MCM-41 and (b)
Pd–Si–MCM-41. Reduction procedure: 100 ◦C for 30 min.

P. Ma¨ ki-Arvela et al. / Journal of Molecular Catalysis A: Chemical 285 (2008) 132–141
135
Table 2
mmol) per time unit and catalyst amount, the fractional yield
is the ratio between the formed moles of the desired product
divided by the initial amount of the reactant, acetophenone;
selectivity is a ratio between the moles of the desired product
divided by the total molar amount of products. The liquid phase
mass balance was calculated by adding the molar amounts of
acetophenone, (R,S)-1-phenylethanol, R-1-phenylethyl acetate,
ethyl benzene and styrene, water and taking into account stoi-
chiometry in ethanol formation. The mass balance was close to
100%. Only traces of acetic acid were formed.
H2 TPD results for Pd–Al2O3 and for Pd–SiO2
Catalyst
Total amount of hydrogen desorbed (mmol/gcat.
)
Tmax (◦C)
Pd–Al2O3
Pd–SiO2
0.0067
0.002
307, 367
364
The TPD was 25 ◦C – 10 ◦C/min – 700 ◦C (60 min).
3.2.1.1. Initial reaction rates and the main reactions. The total
initial reaction rate for acetophenone transformation was deter-
mined by calculating the amount of acetophenone converted
at the beginning of the experiment per time and per amount
of hydrogenation catalyst. During the first 30 min of reaction
time the intermediate, (R,S)-1-phenyl ethanol was dehydrated
to styrene followed by its fast hydrogenation to ethyl benzene.
The initial dehydration rate, i.e. formation rate of styrene and
ethyl benzene was additionally calculated. The initial acylation
rate was calculated from the amount of the formed R-1-phenyl
ethylacetate per time unit and per amount of lipase. The initial
rates and acetophenone conversion after 1600 min are given in
Table 3.
Fig. 2. Hydrogen TPD from Pd–SiO2 and Pd-–Al2O3 catalysts.
the nominal metal loading was the same. Furthermore, the Pd
crystallites were the largest in Pd–Si–MCM-41 from the three-
studied catalyst, while the Pd crystallite size from Pd/SiO2 was
not determined, since its activity in the test reaction was low.
The amount of hydrogen desorbed was determined for
Pd–SiO2 and for Pd–Al2O3. The latter catalyst exhibited 3.4-
fold higher amount of desorbed hydrogen than the former one
(Table 2). Furthermore, hydrogen desorption had two maxima
in case of Pd–Al2O3, i.e. at 307 ◦C and at 367 ◦C indicating that
there are two energetically different adsorption sites available
on the catalyst surface (Fig. 2), whereas only one temperature
maximum for hydrogen desorption was seen for Pd–SiO2.
The acidity of the two supports, Al2O3 and H-MCM-41 was
determined with pyridine adsorption (Table 1). The former one
exhibited only a small concentration of Brønsted acid sites,
whereas their concentration in H–MCM-41 was 12.7-fold that
of the former one. In both support materials the concentration
of Lewis acid sites was about the same.
The
acetophenone
total
initial
decreased
transformation
as follows:
rate
Pd–H–MCM-
of
41 > Pd–Al2O3 > Pd–Si–MCM-41 > Pd–SiO2. The Pd particle
size of Pd–Al2O3 was 1.8 times larger than in Pd–H–MCM-41
resulting an initial hydrogenation rate being only 34% of
the rate obtained with the latter catalyst with smaller Pd
particles (Table 3). The third lowest activity was observed for
Pd–Si–MCM-41 exhibiting the largest Pd crystallite size of the
three studied catalysts in XRD (Table 1). This result indicates
that the smaller Pd particles enhanced the hydrogenation of
acetophenone. Furthermore, the two most active catalysts
possessing Brønsted acidity (Table 1) catalyzed both hydro-
genation and dehydration of the formed (R,S)-1-phenylethanol.
The initial dehydration rate increased linearly with increasing
Brønsted acidity.
The catalyst exhibiting the lowest activity, i.e. Pd–SiO2 was
characterized by hydrogen desorption technique and compared
with the performance and hydrogen desorption capacity of
Pd–Al2O3 (Tables 2 and 3, Fig. 2). The latter catalyst was
relatively active in the hydrogenation, while the initial hydro-
genationrateoverPd–SiO2 catalystwas17%oftherateachieved
over Pd–Al2O3 catalyst. When correlating this value to the
3.2. Kinetic results
3.2.1. Catalyst comparison results – definitions
The following definitions were used: initial hydrogenation
and acylation rates were defined as the converted amounts (in
Table 3
Kinetic results using 125 mg lipase and 125 mg hydrogenation catalyst
Catalyst
Total initial reaction
rate (mmol/
Initial hydrogenation Initial dehydration rate Ratio between the initial
Initial acylation rate
(mmol/min/genz.)
Conversion
after 1600 min
(%)
rate (mmol/
min/ghydr. cat)
(mmol/min/ghydr. cat.
)
dehydration rate to initial
hydrogenation rate
min/ghydr. cat.
)
Pd–H–MCM-41 0.44
0.35
0.12
0.07
0.02
0.085
0.004
0
0.24
0.03
0
0.04
92
47
15
7
Pd–Al2O3 0.13
Pd–Si–MCM-41 0.07
Pd–SiO2 0.02
0.002
0.005
0.004
0
0

136
P. Ma¨ ki-Arvela et al. / Journal of Molecular Catalysis A: Chemical 285 (2008) 132–141
amounts of hydrogen desorbed, it was clearly visible that the
amount of hydrogen desorbed from the former catalyst was only
30% of the value for Pd–Al2O3 indicating that there was a cor-
relation between the initial hydrogenation rate and the amount
of hydrogen desorbed form the catalyst. The temperatures at
which the maximum amounts of desorbed hydrogen were mea-
sured for these two catalysts were about the same thus indicating
that only the different amounts of hydrogen available at the cat-
alyst surfaces affected the rates, since the adsorption strengths
for hydrogen over the two catalysts should be similar.
The initial dehydration rate was the highest for the most
acidic catalyst, i.e. Pd–H–MCM-41 being 21-fold that of
the corresponding rate achieved over Pd–Al2O3 (Table 3).
The mesoporous support material, H–MCM-41, contained
89 ␮mol/g Brønsted acid sites (Table 1). Since the concentra-
tion of Lewis acid sites in Al2O3 and in H–MCM-41 was about
the same, it can be concluded that mainly Brønsted acid sites
catalyze dehydration of secondary alcohols. This result deviates
from the result of [7], in which gas-phase dehydration of 1-
phenylethanol was studied over aluminium oxides. In that work,
the main conclusion was that both Brønsted and Lewis acid sites
catalyze dehydration of 1-phenylethanol.
The initial acylation rate was the highest over the cata-
lyst exhibiting the highest initial hydrogenation rate, namely
Pd–H–MCM-41. This result can be explained by the higher
initial concentration of (R,S)-1-phenylethanol obtained over
Pd–H–MCM-41 enhancing at the same time the initial acyla-
tion rate. When (R,S)-1-phenylethanol was acylated with vinyl
acetate over native lipase in tert-butyl methyl ether at 30 ◦C
the yield of ester increased by factor 3.2 when increasing the
normalized (R,S)-1-phenylethanol concentration from 0.5 to 3.5
[17].
supported heterogeneous catalyst and the immobilized enzyme-
lipase (Scheme 1).
Hydrogenation of acetophenone over Pd leads to racemic
(R,S)-1-phenylethanol and is followed by the consecutive acyla-
tionofR-1-phenylethanoltoR-1-phenylethylacetateoverlipase.
The stoichiometric product of the acylation reaction, ethanol
can deactivate enzyme being very hydrophilic [8]. The main
side reaction is dehydration of (R,S)-1-phenylethanol to styrene,
which occurs on the supports, possessing acidic properties. Due
to the availability of hydrogen the subsequent hydrogenation of
styrene over Pd catalyst to ethylbenzene is feasible and in fact
proceeds very fast, since the amounts of styrene in the reaction
mixtures were below 1%.
In order to achieve higher yields of the desired product
racemization of S-1-phenylethanol to (R,S)-1-phenylethanol is
required. It is known from the literature, that racemization is typ-
ically carried out over zeolites [9,10], vanadyl sulfate [11] and
Ru-hydroxyapatite [12]. It should be kept in mind, however, that
water is formed as a side product in reaction 4 of Scheme 1. Pres-
ence of water enhances hydrolysis of R-1-phenyl ethyl acetate
over lipase to produce R-1-phenylethanol and acetic acid (step 3
in Scheme 1). This reaction can be also catalyzed by acidic mate-
rials [18] indicating that formation of water due to dehydration
is not beneficial in the system under investigation, since it facil-
itates the hydrolysis of the desired product catalyzed either by
enzyme or by the acidic catalyst. It means, in essence, that zeo-
lites cannot be used as catalysts for racemization reaction (step
6 in Scheme 1) and some other non-acidic catalysts are needed
forthecatalyticcascadecombininghydrogenation, racemization
and acylation reactions avoiding, however, dehydration.
3.2.1.2. Conversions after reaction times. The con-
versions over four different catalysts after prolonged
reaction times decreased as follows: Pd–H–MCM-
41 > Pd–Al2O3 > Pd–Si–MCM-41 > Pd–SiO2 (Table 3, Fig. 3).
Experimental observations, i.e. formation of styrene, ethyl
benzene and acetic acid (see below), allow to present a reac-
tion network for the complex cascade system, containing the
Scheme 1. Reaction scheme for cascade one-pot synthesis of R-1-phenylethyl acetate starting from acetophenone hydrogenation. Reaction steps: (1) hydrogenation
of acetophenone to (R,S)-1-phenylethanol; (2) acylation of R-1-phenylethanol to R-1-phenylethyl acetate over immobilized lipase and formation of ethanol; (3)
deacylation of R-1-phenylethyl acetate in the presence of water and formation of acetic acid; (4) dehydration of (R,S)-1-phenylethanol in the presence of an acidic
catalyst; (5) hydrogenation of styrene to phenylethanol; (6) racemization of S-1-phenylethanol over a racemization catalyst.

P. Ma¨ ki-Arvela et al. / Journal of Molecular Catalysis A: Chemical 285 (2008) 132–141
137
Fig. 3. Conversion of acetophenone as a function of time in one-pot synthesis
of R-1-phenyl ethylacetate using lipase with (ꢀ) Pd–Al2O3; (ꢁ) Pd–SiO2; (ꢂ)
Pd–H–MCM-41;and(᭹)Pd–Si–MCM-41. Conditions:1 barhydrogenat70 ◦C.
The deactivation of Pd catalysts was clearly visible after
prolonged reaction times. There might be several possible
reasons for Pd catalysts deactivation, like formation of ethanol
as a stoichiometric product in the acylation step, formation of
acetic acid in the hydrolysis of the desired product, R-1-phenyl
ethylacetate (see below), effect of lipase or sintering and/or
leaching of Pd. The effect of ethanol was studied by carrying
out the hydrogenation of acetophenone over Pd–Al2O3 catalyst
without lipase in ethanol. The reaction rate was high and the
final conversion was 92% after 1660 min indicating that ethanol
is not the reason for deactivation of Pd–Al2O3. Analogous
results have been obtained in acetophenone hydrogenation over
Pd–Al2O3 in ethanol at 130 ◦C under 60 bar hydrogen [19].
Additionally, when acetophenone was hydrogenated under the
same reaction conditions using 62.5 mg Pd–Al2O3 catalyst
without lipase in ethyl acetate the hydrogenation activity was
close to zero after 1200 min reaction time corresponding to
conversion of 13%. If both Pd–Al2O3 and lipase were used an
analogous deactivation of the hydrogenation catalyst occurred.
These results indicated that deactivation of the hydrogenation
catalyst occurred also in the absence of lipase, therefore leach-
ing of enzyme, which was not studied in the current work, was
not the main reason for catalyst deactivation. When comparing,
however, the hydrogenation rates in the presence and in the
absence of lipase it could be observed that slightly lower rates
were achieved in the presence of lipase giving the conversion
of 10% after 1200 min. In other words, it can be concluded that
the hydrogenation catalyst and the enzyme interfered with each
other. Acetic acid is most probably the reason for deactivation
of Pd catalyst, although Pd sintering could not be excluded,
since in both Pd–SiO2 and Pd–MCM-41 catalysts sintering of
Pd was observed [20], when the catalysts were used in liquid
phase hydrogenation of 1-hexene in ethanol at 25 ◦C at 1 bar.
In this work [20], Pd nitrate was used as the metal precursor
and incipient wetness impregnation was used for introducing
Pd. Thereafter, the catalysts were dried and calcined in air at
500 ◦C for 2 h. In the current work, the catalysts were prepared
from palladium nitrate by using vacuum evaporation method,
followed by drying and calcination at 400 ◦C.
Fig. 4. (a) Fractional yield of R-1-phenyl ethanol; (b) S-1-phenylethanol
using lipase with (ꢀ) Pd–Al2O3; (ꢁ) Pd–SiO2; (ꢂ) Pd–H–MCM-41; and (᭹)
Pd–Si–MCM-41. Conditions: 1 bar hydrogen at 70 ◦C.
3.2.1.3. Product distribution and selectivity. The primary prod-
uct in acetophenone hydrogenation was (R,S)-1-phenylethanol
and its formation rate decreased as follows: Pd–H–MCM-
41 > Pd–Al2O3 > Pd–Si–MCM-41 > Pd–SiO2 (Fig. 4a and b).
The hydrogenation of acetophenone was racemic, since the
formation rates for R- and S-1-phenylethanol were initially
the same, as expected. The corresponding maximum fractional
yields of R-1-phenylethanol were 22%, 14%, 6% and 1% there-
after this enantiomer reacted further either to R-1-phenylethyl
acetate or via dehydration over acidic sites first to styrene and
furthermore via very fast hydrogenation to ethylbenzene, as
only traces of styrene could be observed in the reaction mix-
ture. The S-enantiomer reacted further only to styrene and to
ethylbenzene, since the enantiomeric excess of R-1-phenyl ethy-
lacetate was close to 100%. Over Pd catalysts supported on
non-acidic SiO2 and Si–MCM-41 as well as over Al2O3 the frac-
tional yields of ethyl benzene was below 1% (Table 4), whereas
over Pd–H–MCM-41 the fractional yields of ethylbenzene was
46% after 1600 min (Table 4). The fractional yields of S-1-
phenylethanol remained constant after reaching its maximum
with exception of Pd–H–MCM-41, over which the dehydration
took place. Since the amount of S-1-phenylethanol remained
constant over non-acidic Pd supported catalysts, it can be con-
cluded that Pd did not catalyze racemization.
The acylation of R-1-phenylethanol was initially the fastest
over Pd–H–MCM-41 most probably due to concentration effect.

138
P. Ma¨ ki-Arvela et al. / Journal of Molecular Catalysis A: Chemical 285 (2008) 132–141
Table 4
S-1-phenylethanol was not acylated (Table 4). When plotting
selectivity of the desired product as a function of acetophenone
conversion, the consecutive reaction pathway is clearly visible
for R-1-phenylethyl acetate (Fig. 5b).
Selectivity to the desired product and the fractional yield of the side product,
ethyl benzene using 125 mg lipase and 125 mg hydrogenation catalyst
Catalyst
Selectivity to
R-1-PEAC (%)
Fractional yield of ethyl
benzene after 1600 min (%)
The dehydration of (R,S)-1-phenylethanol occurs on the
acidic sites of the catalysts. Thus, nearly no dehydration was
observed over Pd–SiO2 and over Pd–Si–MCM-41 (Table 3).
The dehydration rate was constant over Pd–H–MCM-41 up to
91 min, afterwards declining by 44%, which was due to lowered
concentration of (R,S)-1-phenylethanol. The fractional yield of
ethyl benzene as a function of R-1-phenylethyl acetate showed
that lipase was active in combination with heterogeneous cata-
lysts after prolonged reaction times in spite of the large amounts
of ethyl benzene present in the reaction mixture, thus indicat-
ing that its presence is not inhibiting the acylation reaction. The
enzymatic activity was, however, lowered at prolonged reaction
times, since ethanol is formed as a stoichiometric product in the
acylation leading to deteriorating activity, as it is known that
lipases are more active in hydrophobic solvents [21]. Another
reason for the decline of the acylation rate is the formation of
water in dehydration of (R,S)-1-phenylethanol. As an example
when using 125 mg of Pd–H–MCM-41 the water content of the
reaction mixture was 0.4 wt.%, with corresponding fractional
yield of ethyl benzene formed after 1700 min being 50%. On
the other hand, the presence of trace amounts of water is crucial
to guarantee the enzymatic activity, since in hydrophilic sol-
vents the enzyme deactivation occurs due to stripping off the
bound water in enzyme. The role of the bound water is thought
to preserve the enzyme conformation flexibility [22].
Pd–H–MCM-41
Pd–Al2O3
Pd–Si–MCM-41
Pd–SiO2
7a, 20 (93)b
16a, 47 (46)b
0a, 49c
46
<1
<1
<1
47d
a
40% conversion.
Selectivity after 1700 min, conversion in parenthesis in %.
15% conversion.
7% conversion.
b
c
d
The acylation rate over this catalyst, however, decreased to the
level of 18% from the initial one after 30 min of reaction time
(Fig. 5a), while the rate for styrene formation and its hydrogena-
tion to ethylbenzene was very high and remained constant up to
91 min. On the other hand, the acylation rate over mildly acidic
Pd–Al2O3 remained constant until 393 min, declining thereafter.
Thus, the highest fractional yields of the desired product, R-
1-phenylethyl acetate were achieved over this catalyst, being
22% after 1850 min. The third best catalyst for acylation of R-
1-phenylethanol was Pd–Si–MCM-41 yielding to 8% fractional
yield of the desired product. The selectivities to R-1-phenylethyl
acetate were close to 50%, i.e. to a maximum value, since
3.3. The effect of hydrogenation catalyst amounts
In the catalyst screening experiments, the same amounts of
hydrogenation catalyst and enzyme were used. It turned out that
complete conversion of acetophenone was not achieved. The
logical idea was to study the influence of the doubled amount of
the hydrogenation catalyst in the acylation and dehydration step
in order to get a deeper understanding of the reaction mecha-
nism. A non-acidic Pd–SiO2 and slightly acidic Pd–Al2O3 were
tested. The initial rates over Pd–SiO2 and Pd–Al2O3 catalysts
using 125 mg of lipase and 250 mg of hydrogenation catalyst
are given in Table 5. When duplicating the amount of the hydro-
genation catalyst the initial hydrogenation rate over Pd–Al2O3
enhanced only by factor 1.3, whereas over Pd–SiO2 the initial
rate became slightly more than 2-fold. This result indicates that
over the former catalyst there is an interference between the
lipase and the hydrogenation catalyst, which is in agreement
with experiments showing that faster hydrogenation rates were
achieved over Pd–Al2O3 in the absence of lipase compared to the
case, when both Pd–Al2O3 and lipase were introduced into the
reaction mixture. The interactions between lipase and Pd–SiO2
are probably less strong than between Pd–Al2O3 and lipase. The
initial acylation rates were also enhanced when duplicating the
amount of hydrogenation catalyst (Tables 4 and 5), but the initial
dehydration rate remained unchanged with the double amount
of Pd–Al2O3, which was somewhat unexpected. It should, how-
ever, be pointed out the dehydration is a consecutive reaction,
Fig. 5. (a) Fractional yield of R-1-phenylethyl acetate; (b) selectivity of R-1-
phenylethyl acetate as a function of acetophenone conversion using lipase with
(ꢀ) Pd–Al2O3; (ꢁ) Pd–SiO2; (ꢂ) Pd–H–MCM-41; and (᭹) Pd–Si–MCM-41.
Conditions: 1 bar hydrogen at 70 ◦C.

P. Ma¨ ki-Arvela et al. / Journal of Molecular Catalysis A: Chemical 285 (2008) 132–141
139
Table 5
Kinetic results using 125 mg lipase and 250 mg hydrogenation catalyst
Catalyst
Total initial reaction
rate (mmol/
Initial hydrogenation
rate (mmol/
Initial dehydration
rate (mmol/
min/ghydr. cat.)
Ratio between the initial
dehydration rate to initial
hydrogenation rate
Initial acylation rate
Conversion
after 1600 min
(%)
(mmol/min/genz.
)
min/ghydr. cat.
)
min/ghydr. cat
)
Pd–Al2O3 0.16
Pd–SiO2 0.05
0.156
0.05
0.004
0
0.03
0
0.004
0.03
97
17
which is clearly visible in the formation of ethyl benzene after
prolonged reaction times (see below).
a slight decrease in this rate was observed. The presence of extra
water formed in the dehydration of (R,S)-1-phenylethanol is not
beneficial in the reaction system containing the lipase and an
ester, since hydrolysis of ester is catalyzed by lipase forming a
carboxylic acid [23]. Formation of propionic acid was confirmed
in the acylation of (R,S)-1-phenylethanol with vinyl propionate
30 ◦C with lipase [23]. In the current work, the presence of traces
of acetic acid has been verified by GC–MS. Thus, it can be con-
cluded that double amount of Brønsted acid leads to extensive
dehydration of (R,S)-1-phenylethanol (see below).
The selectivities to R-1-phenylethylacetate as a function of
conversion using different amounts of Pd–Al2O3 and Pd–SiO2
lipase pairs are shown in Fig. 8b. It can be stated that the
acylation of R-1-phenylethanol proceeded after prolonged reac-
tion times in all studied systems, with except of the case of
250 mg of Pd–Al2O3 and lipase (Fig. 8b), when the main prod-
The conversion of acetophenone became 2.1- and 2.4-fold,
respectively, when duplicating the amounts of Pd–Al2O3 or
Pd–SiO2. The reaction rate over Pd–Al2O3 remained constant
up to 90 min of reaction time, declining thereafter. All the
hydrogenation catalysts, however, deactivated, since complete
conversion of acetophenone was not achieved (Fig. 6). The rea-
son for deactivation of Pd catalysts might be formation of acetic
due to lipase catalyzed hydrolysis of R-1-phenyl ethyl acetate
(see below) [23].
Racemic formation of (R,S)-1-phenylethanol was observed
as stated above and the fractional yields of R- and S-1-
phenylethanol as a function of reaction time are shown in Fig. 7a
and b. The maximum fractional yield of R-1-phenylethanol was
38% after 150 min in case of 250 mg of Pd–Al2O3 and 125 mg of
lipase (Fig. 7a), whereas with the half amount of the hydrogena-
tioncatalystthecorrespondinglevelwas14%after330 min. This
result indicated that it is possible to increase the fractional yield
ofR-1-phenylethanolbyincreasingtheamountofhydrogenation
catalyst, i.e. by duplicating the amount of hydrogenation cata-
lyst the fractional yield of R-1-phenylethanol increased by factor
2.7. This result would suggest that the acylation reaction should
alsobeenhanced. Theresultswere, however, notaspromising, as
could be anticipated. Initially the acylation of R-1-phenylethanol
proceeded at a constant rate up to 400 min of reaction time
with 250 mg Pd–Al2O3 and lipase (Fig. 8a) corresponding to
the conversion level of 97%. Thereafter, the dehydration of
(R,S)-1-phenylethanol became the most prominent reaction. The
dehydration rate remained constant up to 1650 min, after which
Fig. 6. Conversion of acetophenone using lipase with (᭹) 250 mg Pd–Al2O3;
(ꢀ) 125 mg Pd–Al2O3; (ꢂ) 250 mg Pd–SiO2; and (ꢃ) 125 mg Pd–SiO2. Con-
ditions: 1 bar hydrogen at 70 ◦C.
Fig. 7. (a) Fractional yield of R-1-phenyl ethanol; (b) S-1-phenyl ethanol
using lipase with (᭹) 250 mg Pd–Al2O3; (ꢀ) 125 mg Pd–Al2O3; (ꢂ) 250 mg
Pd–SiO2; and (ꢃ) 125 mg Pd–SiO2. Conditions: 1 bar hydrogen at 70 ◦C.

140
P. Ma¨ ki-Arvela et al. / Journal of Molecular Catalysis A: Chemical 285 (2008) 132–141
tion rates decreased with increasing metal particle sizes as
follows: Pd–H–MCM-41 > Pd–Al2O3 > Pd–Si–MCM-41. The
lowest initial hydrogenation rate was observed over Pd/SiO2.
The most promising catalyst pair for the one-pot synthesis of
R-1-phenylethylacetate was Pd–Al2O3–lipase (Novozym 435)
giving under optimal conditions maximally 45% selectivity to
R-1-phenylethyl acetate at 47% conversion. The acidic cata-
lysts, Pd–H–MCM-41 and Pd–Al2O3 promoted dehydration of
the intermediate product (R,S)-1-phenylethanol and due to water
formation also hydrolysis of R-1-phenylethyloacetate, catalyzed
either by the lipase or the acidic support.
The observed deactivation of the hydrogenation catalyst
could be attributed to the presence of acetic acid being formed in
the deacylation step, which in turn was promoted by the presence
of water formed in the dehydration of (R,S)-1-phenylethanol.
Lipase activity decreased after prolonged reaction times due to
accumulation of the stoichiometric product ethanol in the reac-
tion mixture. A reaction network was proposed based on the
catalytic data.
One of the future directions in developing this type of cas-
cade reaction is to apply non-acidic hydrogenation catalysts in
order to avoid water formation and at the same time deacylation.
Additionally, application of bases to neutralize acetic acid might
increase the catalyst stability. Furthermore, a promising, albeit
challenging, option is to combine hydrogenation–acylation and
racemization through utilization of non-acidic racemization
catalysts.
Fig. 8. (a) Fractional yield of R-1-phenylethyl acetate as a function of time and
(b) selectivity to R-1-pehynlethyl acetate as a function of conversion using lipase
with (᭹) 250 mg Pd–Al2O3; (ꢀ) 125 mg Pd–Al2O3; (ꢂ) 250 mg Pd–SiO2; and
(ꢃ) 125 mg Pd–SiO2. Conditions: 1 bar hydrogen at 70 ◦C.
Acknowledgements
˚
This work is part of the activities at the Abo Akademi Pro-
uct was styrene being hydrogenated fast to ethyl benzene with
its fractional yield of 38% after 1700 min. When instead using a
non-acidic 250 mg Pd–SiO2 and lipase only traces of ethyl ben-
zene were formed. The selectivities to R-1-phenyl ethyl acetate
after 1700 min were over Pd–Al2O3 and Pd–SiO2 lipase pairs
24% and 49% with conversions of 98% and 18%, respectively.
Themainconclusionisthatthedesired50%fractionalyieldof R-
1-phenylethylacetate could not be achieved over the most active
catalyst pair, Pd–Al2O3-lipase due to extensive dehydration of
(R,S)-1-phenylethanol over Brønsted acid sites. The possibility
of using larger amount of the hydrogenation catalyst cannot be
consideredasaviableoptionsincetheinitialhydrogenationrates
were not correlating with the catalyst bulk density, the reason
being either mass transfer limitations or detrimental interactions
with the enzyme.
˚
cess Chemistry Centre (AA-PCC) within the Finnish Centre of
Excellence Programme (2000–2011) appointed by the Academy
of Finland.
References
[1] A. Bruggink, R. Schoevaart, T. Kieboom, Org. Proc. Res. Dev. 7 (2003)
622–640.
[2] H.M. Jung, J.H. Koh, M.-J. Kim, J. Park, Org. Lett. 2 (2000) 2487–2490.
[3] U. Hanefeld, Org. Biomol. Chem. 1 (2003) 2405–2415.
[4] H.R. Hobbs, B. Kondor, P. Stephenson, R.A. Sheldon, N.R. Thomas, M.
Poliakoff, Green Chem. 8 (2006) 816.
[5] A. Drelinkiewicza, A. Waksmundzka, W. Makoski, J.W. Sobczak, A. Kro´l,
A. Zie˛ba, Catal. Lett. 94 (2004) 143–156.
[6] F. Gao, A.D. Allian, H. Zhang, S. Cheng, M. Garland, J. Catal. 241 (2006)
189–199.
[7] R.G. Romanova, A.A. Lamberov, I.G. Shmelev, Kin. Catal. 45 (3) (2004)
422–428.
[8] K. Faber, Biotransformations in Organic Chemistry, Springer, Berlin, 1992,
IX, 319.
4. Conclusions
Four different hydrogenation catalysts together with an
immobilized lipase were investigated in one-pot synthesis
of R-1-phenylethylacetate starting from acetophenone hydro-
genation. The experiments were performed in ethyl acetate
as an acyl donor using molecular hydrogen at 70 ◦C. The
concentrations of the Brønsted and Lewis acid sites of
the supports were determined with pyridine adsorption and
sizes of Pd crystallites by XRD. The initial hydrogena-
[9] S. Wuyts, K. De Temmerman, D.E. De Vos, P.A. Jacobs, Chem. Comm.
(2003) 1928–1929.
[10] S. Wuyts, K. De Temmerman, D.E. De Vos, P.A. Jacobs, Chem. Eur. J. 11
(2005) 386–397.
[11] S. Wuyts, J. Wahlen, P.A. Jacobs, D.E. De Vos, Green Chem. 9 (2007)
1104–1108.
[12] S. Wuyts, D.E. De Vos, F. Verpoort, D. Depla, R. De Gryse, P.A. Jacobs,
J. Catal. 219 (2003) 417–424.
[13] J. S. Beck, US Patent 5057296 (1991).

P. Ma¨ ki-Arvela et al. / Journal of Molecular Catalysis A: Chemical 285 (2008) 132–141
141
ˇ
ˇ
[14] C. T. Kresge, M.E. Leonowicz, W.J. Roth, J.C. Vartuli, US Patent 5098 684
(1992).
[19] M. Bejblova´, P. Za´mostny´, L. Cerveny´, J. Cejka, Collect. Czech. Chem.
Commun. 68 (2003) 1969–1984.
[15] A. Bernas, N. Kumar, P. Ma¨ki-Arvela, N.V. Kulkova, B. Holmbom, T.
Salmi, D.Yu. Murzin, Appl. Catal. A: Gen. 245 (2003) 257–275.
[16] C.A. Emeis, J. Catal. 141 (1993) 347–354.
[17] Y. Ikeda, R. Nozaki, Y. Kurokawa, J. Chem. Technol. Biotechnol. 77 (2001)
86–91.
[20] J. Panpranot, K. Pattamakomsan, P. Praserthdam, React. Kinet. Catal. Lett.
86 (1) (2005) 141–147.
[21] E. Cernia, C. Palocci, S. Soro, Chem. Phys. Lipids 93 (1998) 157–168.
[22] L.S. Chua, M.R. Sarmidi, J. Mol. Catal. B: Enzym. 38 (2006) 551–556.
[23] F.J. Herna´ndez-Ferna´ndez, A.P. de los R´ıos, F. Toma´s-Alonso, D. Go´mez,
M. Rubio, G. V´ıllora, Chem. Eng. Proc. 46 (2007) 818–824.
[18] A. Hilton, R.R. Myers, J. Am. Chem. Soc. 64 (1942) 2362–2365.