Enantioselective hydrogenation of 1-phenyl-1,2-propanedione

Article abstract of DOI:10.1006/jcat.2001.3406

Enantioselective hydrogenation of a diketone, 1-phenyl-1,2-propanedione was studied in a pressurized reactor at 5 bar and at 0-25C in different solvents: ethanol, ethyl acetate, and dichloromethane over platinum catalysts. Both in situ modification (simultaneous addition of the reagent and the modifier) and pre-modification (preadsorption of the modifier prior to the reagent) of the catalyst were investigated using cinchonidine as catalyst modifier. Racemic hydrogenation proceeded with nearly the same rate as the selective hydrogenation in the presence of the catalyst modifier. The kinetic results revealed that the hydrogenation of the carbonyl group attached to the phenyl ring was preferred, the main product being 1-hydroxy-1-phenylpropanone; the ratio between 1-hydroxy-1-phenylpropanone and 2-hydroxy-1-phenylpropanone was about 11. The most effective and enantioselective catalyst was obtained by in situ modification in dichloromethane yielding in 67 mol% of (R)-1-hydroxy-1-phenylpropanone, corresponding to the enantiomeric excess of 64%. The enantiomeric excess was independent of the reactant conversion. In the second hydrogenation step the main product among diols was (1R,2S)-1-phenyl-1,2- propanediol.

Full text of DOI:10.1006/jcat.2001.3406

Journal of Catalysis 204, 281–291 (2001)
doi:10.1006/jcat.2001.3406, available online at http://www.idealibrary.com on
Enantioselective Hydrogenation of 1-Phenyl-1,2-propanedione
�
�
�
�
�
Esa Toukoniitty, P a¨ ivi M a¨ ki-Arvela, Marek Kuzma, Alexandre Villela, Ahmad Kalantar Neyestanaki,
�
�,1
Tapio Salmi, Rainer Sj o¨ holm,† Reko Leino,‡ Ensio Laine,§ and Dmitry Yu. Murzin
�
Laboratory of Industrial Chemistry, Process Chemistry Group, †Laboratory of Organic Chemistry, and ‡Laboratory of Polymer Technology,
˚
Abo Akademi, Biskopsgatan 8, 20500 Turku, Finland; and §Department of Physics, University of Turku, Turku, Finland
Received January 31, 2001; revised August 31, 2001; accepted August 31, 2001
duces racemic mixtures of optical isomers. Heterogeneous
Enantioselective hydrogenation of a diketone, 1-phenyl-1,2- asymmetric catalysis with the participation of modifiers has
propanedione was studied in a pressurized reactor at 5 bar proven to be an efficient way of producing optically pure
�
and at 0� 25 C in different solvents: ethanol, ethyl acetate, and
chiral substances, as handling and separation is much more
simple than applying homogeneous catalysts. The role of
dichloromethane over platinum catalysts. Both in situ modification
simultaneous addition of the reagent and the modifier) and pre-
(
the modifier is to steer the adsorption of the reagent in such
a way that enantioselective hydrogenation is enabled. Het-
erogeneousenantioselectivehydrogenationof�-ketoesters
with a Pt catalyst modified by alkaloids such as cinchoni-
dine has been studied intensively during recent years (1).
However, there exist very few publications concerning the
modification (preadsorption of the modifier prior to the reagent) of
the catalyst were investigated using cinchonidine as catalyst mod-
ifier. Racemic hydrogenation proceeded with nearly the same rate
as the selective hydrogenation in the presence of the catalyst modi-
fier. The kinetic results revealed that the hydrogenation of the car-
bonyl group attached to the phenyl ring was preferred, the main
product being 1-hydroxy-1-phenylpropanone; the ratio between enantioselective hydrogenation of conjugated diketones
-hydroxy-1-phenylpropanone and 2-hydroxy-1-phenylpropanone (2–8), where an analogously modified catalyst system could
1
was about 11. The most effective and enantioselective catalyst was be used. The products, chiral �-hydroxyketones, are valu-
obtained by in situ modification in dichloromethane yielding in
able building blocks in the asymmetric synthesis of biolog-
6
7 mol% of (R)-1-hydroxy-1-phenylpropanone, corresponding to
ically active compounds (9) and they are also intermedi-
ates in the synthesis of anti-AIDS drugs (10). Also chiral
amino alcohols, which are used as vasoconstriction agents
the enantiomeric excess of 64%. The enantiomeric excess was in-
dependent of the reactant conversion. In the second hydrogena-
tion step the main product among diols was (1R,2S)-1-phenyl-1,2-
(
11), are synthesized from enantiomerically pure hydro-
xyketones. In the present study, we have chosen 1-phenyl-
,2-propanedione (A) as a model compound in enantiose-
propanediol.
�c 2001 Elsevier Science
Key Words: enantioselective hydrogenation; 1-phenyl-1,2-
propanedione; cinchonidine; solvent; modification.
1
lective hydrogenation. The reaction scheme is displayed in
Fig. 1.
Two kinds of regioisomeric intermediates, 1-hydroxy-1-
phenylpropanone and 2-hydroxy-1-phenylpropanone, may
exist in the system. The reaction proceeds further to 1-
phenyl-1,2-propanediols. The aim of this work was to in-
vestigate the enantioselective hydrogenation of 1-phenyl-1,
INTRODUCTION
One of the most important requirements for heteroge-
neous catalytic reactions in fine chemicals production is
proper selectivity, which in a broad sense should be under-
stood as chemo-, regio-, and enantioselectivity. Enantiose-
lective reactions in heterogeneous catalysis are of growing
interest, as optically pure chiral compounds are of great
importance in the areas of pharmaceuticals, agrochemicals,
flavors, and fragrances. It is well recognized that differ-
ent enantiomers can possess different physiological prop-
erties, sometimes the “wrong” enantiomer being a ballast,
sometimes a “pollutant.” Catalytic hydrogenation of ke-
tones with a prochiral center over nonchiral catalysts pro-
2
-propanedione and to maximize the selectivity of the for-
mation of (R)-1-hydroxy-1-phenylpropanone, which was
the most prominent product. The catalysts were Pt/Al2O3
and Pt/C, modified with cinchonidine.
EXPERIMENTAL
1
-Phenyl-1,2-propanedione (Acros, 20736-0050, 98%)
was hydrogenated in a pressurized reactor (Sotelem, Mi-
3
croautoclave System, volume 100 cm ). The stirring rate
was 1600 rpm. The hydrogen (AGA, 99.999%) pres-
sure was 5 bar and temperature was 0–25 C. Two differ-
�
1
To whom correspondence should be addressed. Fax: +358 2 215 4479.
E-mail: dmurzin@abo.fi.
ent Pt catalysts were used: Pt/Al2O3 (Strem Chemicals,
281
0
021-9517/01 $35.00
c 2001 Elsevier Science
All rights reserved
�

2
82
TOUKONIITTY ET AL.
Samples were withdrawn from the reactor at different
time intervals and analyzed with a gas chromatograph (GC)
Varian 3300) equipped with a chiral column (�-Dex 225;
(
length 30 m, diameter 0.25 mm, film thickness 0.25 �m).
Helium was used as a carrier gas with a split ratio of
3
3. The detector (FI) and injector temperatures were 270
�
and 240 C, respectively. The temperature program of the
GC was 110 C (30 min)–15 C/min–250 C (31 min). The
GC analysis was calibrated with 2-phenyl-1,2-propanediol
�
�
�
(
Aldrich, 21376-4, 97%) and with racemic 1-hydroxy-1-
phenyl-2-propanone (57.9%) and 1-phenyl-2-hydroxy-1-
FIG. 1. Reaction scheme in the hydrogenation of 1-phenyl-1,2-
propanedione. Symbols: A, 1-phenyl-1,2-propanedione; B, (R)-1-
hydroxy-1-phenylpropanone; C, (S)-1-hydroxy-1-phenylpropanone; D,
propanone (95.9%, by NMR), which was synthesized at the
a
Laboratory of Polymer Technology, Abo Akademi (the syn-
thesis part is described in the Appendix). The internal stan-
(
R)-2-hydroxy-1-phenylpropanone; E, (S)-2-hydroxy-1-phenylpropanone;
F, (1R,2S)-1-phenyl-1,2-propanediol; G,
propanediol; H, (1S,2R)-1-phenyl-1,2-propanediol; and I, (1S,2S)- and the samples were diluted with n-hexane (Merck, 4368,
-phenyl-1,2-propanediol.
(1R,2R)-1-phenyl-1,2- dard in the GC analysis was 1-octanol (Merck, 991, 97%)
1
9
5%). The assignment of the peaks obtained by analyzing
the hydrogenation products of 1-phenyl-1,2-propanedione
in the GC was carried out by using the products of the
asymmetricsynthesisof(R)-1-hydroxy-1-phenylpropanone
7
9
8-1660) metal content 5 wt%, BET specific surface area
5 m /g, mean metal particle size 8.3 nm (XRD), disper-
2
(
(
B, Fig. 1) and (1R,2S)-1-phenyl-1,2-propanediol (F, Fig. 1)
synthesis procedures are described in the Appendix).
sion 40% (H2 chemisorption), the mean catalyst particle
size 18.2 �m (Malvern), and Pt/C (Alfa 5R18) metal con-
tent 5 wt%, BET specific surface area 988 m /g, mean metal
particle size 1.5 nm (XRD), mean catalyst particle size 15.1
2
Catalyst Characterization
�
m (Malvern). The catalysts were activated prior to the re-
Mean metal particle size was measured via X-ray diffrac-
action under hydrogen flow (100 cm /min) for 2 h at 400 C tion (XRD) carried out using a Phillips PW 1830 X-ray pow-
and cooled down to the reaction temperature. Two different der diffractometer and PW 1710 diffractometer control. A
3
�
catalyst modification procedures were tested:
Cu anode with 2 kW power was used as X-ray source.
Hydrogenchemisorptionmeasurementswereperformed
using an automatic chemisorption apparatus (Sorptomatic
(1) In situ modification: the deoxygenated solution, con-
taining the solvent, the modifier, and the substrate, was in-
jected into the reactor, where the activated catalyst was
under hydrogen and the reaction was commenced immedi-
ately. The modifier-to-catalyst mass ratio was 1 : 1. Typically
the amounts of catalyst, substrate, and modifier were 60, 75,
and 60 mg.
1
900, Carlo Erba Instruments). Adsorption isotherms were
�
recorded at 25 C and within the pressure range of 1.3–
1
�
30 mbar. The catalyst was first reduced at 400 C for 2 h
with flowing hydrogen. After the reduction, the catalyst was
�
4
�
evacuated at 10 bar for 1 h at 400 C.
The BET specific surface areas were measured with the
automatic physisorption–chemisorption apparatus (Sorp-
tomatic 1900, Carlo Erba apparatus). The catalyst was de-
(2) Premodification: after cooling the activated catalyst
3
to the reaction temperature, a solution (40 cm ) containing
an excess of the modifier (0.68 mmol) was injected into
the reactor and was stirred in the presence of air for 1 h.
Then the solvent was removed in such a way that the cata-
lyst remained covered by a thin layer of liquid, and the
reactor was flushed with hydrogen. Fresh solvent and the
substrate were injected into the reactor and the reaction
was started. Typical amounts of catalyst and substrate were
�
gassed at 300 C in vacuo prior to the surface area measure-
ment by nitrogen adsorption.
The mean particle sizes of the commercial catalysts were
measured with a Malvern Zetasizer IIc apparatus. The size
measurement is based on the scattering of He–Ne light,
which is reflected to a catalyst powder-ethanol suspension.
Interference figures of the scattered light are collected by
a Fourier lens and are reflected to the detector. The inter-
6
0 and 75 mg, respectively.
The initial concentration of 1-phenyl-1,2-propanedione ference figures are processed by the computer unit of the
3
was 0.01 mol/dm in all experiments. Three different sol- measurement apparatus to the corresponding catalyst par-
vents were studied: ethanol (Etax, Primalco, 99.5%,), ticle size distribution on the basis of light intensity variation
ethyl acetate (FF Chemicals, 99.8%), and dichloromethane of the figures.
3
(
Merck, 822271, >99%), and 0.04 mol/dm acetic acid in
The elementalanalysisof the Pt catalystswere carried out
ethanol. The mass ratio of 1-phenyl-1,2-propanedione-to- with the electron probe microanalysis (EPMA) technique
platinum was 25 in all experiments. The catalyst was modi- using a LEO S360 microscope connected to an X-ray image
fied with (-)-cinchonidine (Aldrich, C8040-7, 96%).
(IMIX) analyzer (Princeton Gamma Tech).

HYDROGENATION OF 1-PHENYL-1,2-PROPANEDIONE
283
RESULTS AND DISCUSSION
phenylpropanone in vacuo, were modeled with the Cache
program (15). Our calculations were in good agreement
with the literature data (16). According to our calculations,
the most stable conformation of the cinchonidine was anal-
ogous with the reported Open (3) conformer of the mod-
ifier (16), found previously to be the most stable among
the other conformers. The Open (3) conformer is consid-
ered to be involved in the enantiodifferentiating transition
state. The population of this conformer is thought to be the
cause of the observed solvent dependence (16) on enantios-
elective hydrogenation. The longest dimension in cinchoni-
dine is about 1.1 nm. The product molecules, 1-hydroxy-1-
phenylpropanone and 2-hydroxy-1-phenylpropanone had
their longest interatomic distances 0.71 and 0.77 nm, respec-
tively. The comparison between the catalyst pore sizes and
the molecules in the reaction indicated that the Pt/Al2O3
catalyst might be more suitable for enantioselective hydro-
genation of 1-phenyl-1,2-propanedione due to the larger
fraction of large pores more readily available for the bulky
modifier and reactant.
Choice of Catalyst, Catalyst Modification
Procedure, and Solvent
One of the most important parameters of a catalyst mod-
ified with cinchonidine might be the mean metal particle
size and the dispersion of the metal. In the hydrogenation
of ethyl pyruvate, over cinchonidine (CD) modified Pt cata-
lysts, relatively large Pt particles with dispersion lower
than 50% have yielded the best results (1). At the same
time in ethyl pyruvate hydrogenations the catalyst support
plays a minor role and good results were obtained over
silica, alumina, and active charcoal as well as with some
zeolite-supported catalysts (12, 13). For preliminary testing
in the hydrogenation of 1-phenyl-1,2-propanedione, two
different commercial catalysts, Pt/C and Pt/Al2O3, having
wt% metal loadings were chosen. Both alumina and active
charcoal-supported Pt catalyst have been active and selec-
tive in the enantioselective hydrogenation of ethyl pyruvate
12). The mean metal particle sizes for the studied Pt/C and
Pt/Al2O3 catalyst were 1.5 and 8.3 nm, respectively, accord-
ing to the XRD measurements. The BET specific surface
5
(
The preliminary hydrogenations of 1-phenyl-1,2-
�
propanedione were carried out at 25 C and 5 bar H2
in dichloromethane or ethanol (Table 1). The catalysts
were either premodified in dichloromethane with 0.68
mmol CD as in the work of Griffiths et al. (3) or modified
in situ with CD. The enantiomeric excess (ee) of (R)-
2
areas for the former and latter were 988 and 95 m /g.
The pores in the alumina support were very large, 57%
of the pore volume located in the pores between 10 and
00 nm. The best alumina supports in the hydrogenation
of �-ketoesters also had large pores (14). However, the
1
1
-hydroxy-1-phenylpropanone (B, Fig. 1) is defined as
pores in the Pt/C catalysts were somewhat smaller, only
[B] � [C]
ee = [
� 100%.
3
5% of the pore volume located in the pores between 10
B] + [C]
and 100 nm.
The energy minimized conformations of cinchonidine, In the experiment with the premodified catalyst (see Ex-
R)-1-hydroxy-1-phenylpropanone and (R)-2-hydroxy-1- perimental) the ee of B with the Pt/C catalyst was close to
(
TABLE 1
Comparison of Pt/C and Pt/Al O Catalyst in Enantioselective Hydrogenation
2
3
�
of 1-Phenyl-1,2-propanedione at 25 C at 5 Bar H2
Initial
hydrogenation rate
�
5
2
Catalyst
Pt/C
Modification
Solvent
CH2Cl2
(10 mol/s m )
Max. ee (%)
0
Premodification in CH2Cl2
with 0.68 mmol CD
0.56
Pt/Al2O3
Pt/C
Premodification in CH2Cl2
with 0.68 mmol CD
In situ with 0.20 mmol CD
CH2Cl2
0.12
0.40
0.23
62
17
33
24
EtOH and 2 mmol
acetic acid
EtOH and 2 mmol
acetic acid
EtOH
Pt/Al2O3
In situ with 0.20 mmol CD
Pt/Al2O3
Pt/Al2O3
Pt/Al2O3
In situ with 0.20 mmol CD
In situ with 0.20 mmol CD
Premodification in CH2Cl2
with 0.68 mmol CD
0.35
0.13
0.12
a
CH2Cl2
CH2Cl2
65
b
c
13 , 62
a
Constant.
Conversion of dione 20%.
Conversion of dione 96%.
b
c

2
0
84
TOUKONIITTY ET AL.
, whereas enantiomeric excesses over 60% were obtained
with Pt/Al2O3 (Table 1). The second set of experiments was
carried out in ethanol in the presence of 0.04 M acetic acid,
using in situ modification of the catalysts with CD. Acetic
acid might protonize cinchonidine and thus have a benefi-
cial effect on the ee (17). The hydrogenation rate was higher
with the Pt/C catalyst than with Pt/Al2O3, but the ee was
lower with the former catalyst than with the latter one, i.e.,
1
7 and 33%, respectively (Table 1). There was, however, a
beneficial effect of acetic acid addition on the ee in ethanol,
the ee being 24% without acetic acid and 33% with acetic
acid.
The low enantiomeric excesses obtained over the Pt/C
catalyst can be mainly attributed to the smaller mean Pt
particle size, 1.5 vs 8.3 nm (XRD). According to literature,
the catalysts with larger mean Pt particle size give higher
enantioselectivities (18). It is also known that impurities in
the catalyst can decrease the ee as well (18). In the Pt/C
catalyst, there were more impurities than in Pt/Al2O3 ac-
cording to EPMA analysis. The most prominent impurity
in the Pt/C catalyst was sodium; however, direct attribu-
tion of catalyst behavior to Na poisoning is too premature
and more experimental data are needed to elucidate the
influence of sodium on enantioselectivity.
Regioselectivity (rs) to 1-phenyl-1-hydroxypropanone
was defined as
[
B] + [C]
rs = [
� 100%,
D] + [E]
where B, C, D, and E correspond to hydroxyketones, as
defined in Fig. 1. The regioselectivities obtained with the
Pt/C and Pt/Al2O3 catalysts were about the same, the cor-
responding values being 13 and 11, respectively. Obtained
results indicated that the large, 10-fold differences in BET
specific surface area of the catalysts had a minor effect
on the regioselectivity. For both of the catalysts, the pores
FIG. 2. (a) The yield of 1-hydroxy-1-phenylpropanone in the hydro-
genation of 1-phenyl-1,2-propanedione at 25 C in ethanol. The catalyst
wasmodified insitu (᭹) or premodified inthe presenceof air with an excess
of (-)-cinchonidine (᭢). (b) The enantiomeric excess of (R)-1-hydroxy-1-
phenylpropanone in the hydrogenation of 1-phenyl-1,2-propanedione in
�
were sufficiently large (for Pt/C and Pt/Al2O3 catalysts 36 ethanol at 25�C. The catalyst was modified in situ (᭹) or premodified in
and 57% of the pore volume located in the pores between the presence of air with excess of (-)-cinchonidine (᭢).
1
0 and 100 nm, respectively) to allow the reactant to ad-
sorb in a suitable mode to induce a high regioselectivity. of air with an excess of cinchonidine and the catalyst was
In the hydrogenation of 1-phenyl-1,2-propanedione over modified in situ in two different solvents: dichloromethane
Pt-modified MCM-41 catalysts (having well-defined pores (Table 1) and ethanol (Figs. 2a, 2b). The hydrogenation
between 2.3 and 4.1 nm), the regioselectivity remained very rates were lower with the catalyst premodified in the pres-
low (rs = 3), both in the presence and in the absence of cin- ence of air than with the catalyst modified in situ. The ee
chiondine (19). The explanation for the low regioselectivity of B was also lower with the premodified catalyst than with
was steric factors; i.e., the bulky reactant could not adsorb the catalyst modified in situ. The ee of B increased with
in an optimal mode in the narrow pores of MCM-41 and increasing conversion of dione, when using a premodified
thus lower regioselectivities were obtained. However, this catalyst, while it was constant with the catalyst modified in
was not the case with studied Pt/C and Pt/Al2O3 catalysts, situ with CD in dichloromethane as solvent. In ethanol the
which had sufficiently large pores.
ee was constant up to 95% conversion of dione, after which
The catalyst modification procedure can have an effect it increased due to the kinetic resolution (see Qualitative
on the enantioselectivity. In order to investigate the catalyst Kinetics below). According to the literature (2, 3), air could
modification procedure, the following experiments were have a beneficial effect on the enantioselectivity, particu-
carried out: the modification was carried out in the presence larly in ethanol. An explanation to our results, i.e., low ee

HYDROGENATION OF 1-PHENYL-1,2-PROPANEDIONE
285
inthereaction, couldbe a partial deactivation of the catalyst
in the presence of air and also the excess of cinchonidine,
3
0
.68 mmol in 40 cm during premodification; the amount
of cinchonidine was in fact not optimized. In these exper-
iments the solvents included small amounts of dissolved
oxygen also in cases, when the catalyst was modified in situ
and relatively high enantiomeric excesses were obtained.
In a previous communication (7), we reported enantiose-
lective hydrogenation of 1-phenyl-1,2-propanedione with
solvents distilled under argon. These results indicated that
traces of oxygen have a beneficial effect on ee.
When 1-phenyl-1,2-propanedione (the initial concentra-
3
tion 0.01 mol/dm ) was hydrogenated in the presence of
cinchonidine and in the absence of cinchonidine in ethanol
�
at 25 C, the initial hydrogenation rates were nearly the
same; i.e., in the absence of cinchonidine the hydrogenation
FIG. 3. The yield of 1-hydroxy-1-phenylpropanone in the hydrogena-
tion of 1-phenyl-1,2-propanedione in different solvents at 25 C. The cata-
�
5
2
rate was 0.36 � 10 mol/s m and in the presence of cin-
�
�
5
2
chonidine 0.356 � 10 mol/s m . In the case of butane-2,3- lystwaspremodifiedindichloromethanewithanexcessof(-)-cinchonidine
in the presence of air. Symbols: (᭡) ethanol, (᭿) dichloromethane, and
᭹) ethyl acetate.
dione Slipszenko et al. (4) have reported a fourfold rate
(
enhancement in the presence of a modifier. In our previ-
ous publication with higher initial reactant concentrations
3
(
0.05 mol/dm ) (6), the hydrogenation rate increased and
The initial hydrogenation rate and the ee increased with
the racemic hydrogenation was faster than the hydrogena-
increasing modifier concentration from 0.32 to 0.36 �
tion in the presence of cinchonidine, i.e., 1.03 and 0.34 �
�
5
2
1
0
mol/s m and from 20 to 24%, respectively. Here we
�
5
2
1
0
mol/s m , respectively. Due to the higher reaction
should point out that these results were not yet optimized
with respect to the amount of modifier, which will be done
for this particular reaction in the future.
rate the racemic hydrogenation was very unselective; i.e.,
cyclohexyl products were formed according to GC-MS.
From these results we can conclude that the hydrogena-
tion rate increased with increasing initial reactant concen-
tration. There was no rate enhancement compared to the
racemic hydrogenation at low initial reactant concentra-
tion, whereas at higher initial reactant concentrations in
the racemic hydrogenation the unselective hydrogenation
became more favorable. The amount of cinchonidine was
1
-Phenyl-1,2-propanedione was hydrogenated in three
different solvents: ethanol, dichloromethane, and ethyl ac-
etate (Fig. 3). In these experiments, the catalyst was mod-
ified with an excess of cinchonidine in the presence of air
(
see Experimental). The hydrogenation rates increased in
the following order: ethyl acetate < dichloromethane <
ethanol. The hydrogen solubility increases in the fol-
�
investigated at 25 C in ethanol; two different cinchonidine-
lowing order: ethanol = dichloromethane (approximated
to-catalyst mass ratios were used, 1 : 1 and 1 : 10 (Table 2).
�
from the solubility of trichloromethane at 25 C: xg
0
=
.000220(20)) < ethyl acetate (Table 3). The enantiomeric
TABLE 2
excesses increased with increasing hydrogenation rates in
ethanol, whereas in dichloromethane the ee was constant
with different initial hydrogenation rates (Table 2). One ex-
planation to this difference in the ee as a function of initial
reaction rate in ethanol and in dichloromethane can be that
dichloromethane is a rather inert solvent with respect to
the Pt surface (4), which increases the possibility of enan-
tioselective hydrogenation. Ethanol is known to be reactive
withthePtsurface(22). Whencomparingdifferentsolvents,
the highest enantiomeric excesses were obtained with the
solvents having the lowest dielectric constants (Table 1).
According to literature (23), the highest enantiomeric ex-
cesses are obtained with solvents, which have dielectric con-
stants between 2 and 10. It was put forward in (16) that
dependence of ee on solvent polarity can be explained
by concentration of cinchonidine Open(3) conformers.
The population of this conformer from density functional
The Initial Hydrogenation Rate of 1-Phenyl-1,2-propanedione
and the Enantiomeric Excess (ee) of (R)-1-Hydroxy-1-
phenylpropanone over Pt/Al O Catalyst at 5 Bar H
2
3
2
Initial
hydrogenation rate Max. ee
T
�
� 5
2
Solvent ( C) (10 mol/s m ),
(%)
Modification
EtOH
25
0.13
13
Premodification in CH2Cl2
with 0.68 mmol CD
EtOH
EtOH
0
25
0.19
0.22
13
17
In situ with 0.20 mmol CD
Premodification in EtOH
with 0.68 mmol CD
EtOH
EtOH
EtOH
CH2Cl2
CH2Cl2 15
CH2Cl2 25
0
25
25
0
0.22
0.32
0.36
0.10
0.12
0.13
17
20
24
67
63
65
In situ with 0.68 mmol CD
In situ with 0.02 mmol CD
In situ with 0.20 mmol CD
In situ with 0.20 mmol CD
In situ with 0.20 mmol CD
In situ with 0.20 mmol CD

2
86
TOUKONIITTY ET AL.
TABLE 3
The Initial Hydrogenation Rate of 1-Phenyl-1,2-propanedione
which is related to Thiele modulus �
�
�
3
1
1
and the Enantiomeric Excess (ee) of (R)-1-Hydroxy-1-
�
e
=
�
,
[1]
[2]
�
phenylpropanone in Different Solvents at 25 C
� tanh �
�
Initial hydrogenation rate
2
� 5
� 4
where
Solvent
(mol/s m ), 10
ee (%) xg, 10
ε
a
b
�
�1/2
Ethyl acetate
CH2Cl2
EtOH
0.04
0.12
0.13
13 , 43 3.43 (20) 6.0 (21)
0
�
= k �p/De
Rp,
c
d
H2
14 , 62
—
8.9 (21)
e
10
2.06 (20) 24.6 (21)
where Rp and � denote the particle radius and density, and
a
Conversion of dione 23%.
Conversion of dione 92%.
Conversion of dione 20%.
Conversion of dione 95%.
Constant.
0
k is the rate constant for spherical particles. Effective diffu-
b
c
d
e
sion coefficient DeH2 is obtained from the particle porosity
εp and tortuosity �p, and the Wilke–Chang equation (26)
gives
Note. Catalyst was premodified in dichloromethane with 0.68 mmol
CD in the presence of air. Hydrogen solubility (xg) in different solvents at
5 C and the dielectric constant (ε) of the solvent at 25 C are reported
�
�
1/2
7.4 � 10�
12
�
8M
(T/K)
�
�
2
� 1
DH2
gmol
here.
=
�
��
�
,
[3]
m2s
� 1
0.6
�
VH2
cm2
s
� 1
cP
calculations (16) follows the same trend as enantiose-
lectivity with increases of the dielectric constant; i.e., it
decreases.
From these preliminary experiments we have chosen one
polar and one nonpolar solvent for further testing in enan-
tioselective hydrogenation of 1-phenyl-1,2-propanedione.
The Pt/Al2O3 catalyst modified in situ with cinchonidine
resulted in initial reaction rates and enantiomeric excesses
higher than those of the Pt/C catalysts and was therefore
used in the further testing.
where 8 is the association factor, M the average molecular
mass, and � the dynamic viscosity.
Apparent hydrogenation rate is given as
0
0
�
R = �ek c
.
[4]
[5]
H2
By combining [1] and [4] one arrives at
(
a + 1) sinh � � � cosh � = 0
Note that cinchonidine was also hydrogenated in the
reaction mixture to dihydrocinchonidine, which was con-
firmed by GC-MS. We expected to observe the initial tran-
sient increase of ee at the beginning of the reaction due to
both the hydrogenation of the ethylene group in cinchoni-
dine (two adsorption possibilities for the modifier) and the
reaching of the adsorption equilibrium for cinchonidine,
similar to what Mallat et al. (24) observed in hydrogenation
of �-ketoesters. However, in our results the initial develop-
ment of the ee was not visible; i.e., the observed ee’s were
already from the beginning at steady state values (Fig. 2b).
This can be explained by the excess of cinchonidine in the
hydrogenation.
with
0
2
R R � �
p
p
p
a =
,
[6]
�
3DH2 εpxH ctot
2
�
where x is the hydrogen mole fraction in solution.
Results of the calculations (Table 4) showed that the in-
fluence of internal mass transfer was practically negligible.
H2
TABLE 4
Evaluation of the Role of Intraparticular Diffusion
�
T
P
R
Rp
�
p
25 C
Qualitative Kinetics
5 bar
0.0068 mol(kg s)
9.1 �m
3
0.4
0
� 1
The influence of external diffusion was determined by
following the published procedure (25), which was previ-
ously applied to liquid-phase hydrogenation reactions. Due
to rather high values of the reactants’ diffusion coefficient,
the small particle size of the catalyst, and vigorous stirring
εp
�
p
1000
17152
ctot
0.13021 � 10 m2 s� 1
�
8
(and thus the high value of the specific mixing power) the
DH
2
�
x
0.0002
0.996
0.94
effect of external diffusion on the reaction rate was negligi-
ble. The influence of internal diffusion for porous materials
was estimated via calculation of the effectiveness factor �e
for pseudo-first-order reactions with respect to hydrogen,
H2
�
�
e
Note. ctot = �mix/M.

HYDROGENATION OF 1-PHENYL-1,2-PROPANEDIONE
287
Hydrogenation reaction of compounds containing two
functional groups can be described involving a “rateau”
(
rake) scheme (30)
A
B, C, D, E
m
m
[
7]
+
H2
+H2
Aads → (B, C, D, E)ads
→
F, G, H, I.
(
+H2)
(+H2)
For a similar consecutive reaction A
→ B
→ F it
was shown, that it can occur via the mechanism (31, 32)
(
1)
N(2)
0
0
1
N(1)0
N(2)0
N
1
1
0
1
0
1
2
3
4
5
.A + Z � AZ
.AZ + H2 → BZ
.BZ + H2 → FZ
.BZ � B + Z
1
1
0
1
0
1
1
1
0
1
[
8]
� 1
FIG. 4. Hydrogenation kinetics of 1-phenyl-1,2-propanedione in
dichloromethane at 25 C. Catalyst: Pt/Al2O3 modified in situ with (-)-
cinchonidine. Symbols: (᭜) 1-phenyl-1,2-propanedione, (᭹) 1-hydroxy-1-
phenylpropanone, (᭢) 2-hydroxy-1-phenylpropanone, and (᭡) 1-phenyl-
,2-propanediol.
.FZ � F + Z
1
�
(
(
1)
2)
N
N
A + H2 = B
B + H = F
2
1
0
0
(
(
1)
2)
N
N
A + H2 = B
A + 2 H2 = F.
Typical hydrogenation kinetics of 1-phenyl-1,2-propane-
dione is displayed in Fig. 4. In this experiment, 1-phenyl Here Z is a site on the surface of the catalyst and AZ, BZ,
�
1
,2-propanedione was hydrogenated at 25 C in dichloro- and FZ are adsorbed species. On the right side of the equa-
methane. The catalyst was Pt/Al2O3 modified in situ with tions for the elementary steps their stoichiometric numbers
cinchonidine. The most important product was 1-hydroxy- along different stoichiometric pathways (routes) are pre-
1
-phenylpropanone with the maximum yield of 82%. sented. According to the Horiuti–Temkin rule (32, 33) the
The ratio between 1-hydroxy-1-phenylpropanone (B + C, basic set of pathways in scheme [8] should contain two path-
(
1)
Fig. 1) and 2-hydroxy-1-phenylpropanone (D + E, Fig. 1) ways. Two sets of pathways are shown in the scheme N ,
0
0
(2)
(2)
(1)
was 11 and that ratio was constant with increasing conver-
N
and N , N . The final equations of the first set de-
sion of dione. Hence the C=O group, which is in the 1- scribe the formation of products B and F as sequential, and
position, can be hydrogenated more readily than the C=O the second set describes it as parallel. However, if a steady-
group in the 2-position, close to the methyl group. To an- state or quasi-steady-state reaction is under examination,
swer the question of regioselectivity ab initio calculations both descriptions are equivalent in the same way as previ-
for 1-phenyl-1,2-propanedione were performed and the de- ously demonstrated for a similar mechanism (30, 31).
tailsofcalculationswillbe reportedina separatepaper (27).
Therefore, at least based on present steady-state data,
It follows from the minimum energy calculations (27) that we cannot differentiate between two possible modes of
the carbonyl group and the phenyl ring are coplanar; sim- chinconidine influence: direction toward formation of a R-
ilar results were reported in (28) for ab initio calculations center in the first reduction step, but toward an S-center
of �-ketoesters with the phenyl ring adjacent to the C=O in the second during one residence on the surface or via
bond, e.g., ethyl phenylglyoxylate. At the present moment readsorpion of R-hydroxyketone. More experiments in a
wecanonlyspeculatethat1-phenyl-1,2-propanedioneisad- continuous fixed bed reactor under transient conditions as
sorbed preferentially via both the phenyl ring and the C=O well as quantum chemical calculations are in progress for
bond, thus providing high regioselectivity to 1-hydroxy-1- further elucidation of the reaction mechanism.
phenylpropanone.
The hydrogenation kinetics of 1-phenyl-1,2-propane-
It is clear from Fig. 4 that diols start to appear in the dione in ethanol were studied at two different tempera-
�
product mixture right from the very beginning of the reac- tures, 0 and 25 C. The catalyst was modified in situ in these
�
tion. The main product among diols was (1R,2S)-1-phenyl- experiments. The reaction proceeded at highest rate at 25 C
�
5
2
1
,2-propanediol (F, Fig. 1), but not (1R,2R)-1-phenyl-1,2- (0.356 � 10 mol/s m ), and at that temperature the max-
propanediol (G, Fig. 1). Such a result means that cinchoni- imum ee was also slightly higher than that at a lower tem-
dine directs the reaction toward formation of a R-center in perature, about 24% (Table 2). The ee in ethanol was about
the first carbonyl reduction, but toward an S-center in the 12 up to 95% conversion of dione and after that started
second carbonyl reduction step.
to increase due to the kinetic resolution, which means

2
88
TOUKONIITTY ET AL.
to diols was much slower in the case of dichloromethane;
i.e., the yield of diols was 46% in ethanol and 10% in
dichloromethane.
CONCLUSIONS
Hydrogenation of 1-phenyl-1,2-propanedione was inves-
�
tigated in a batch reactor at 5 bar and at 0–25 C in different
solvents over Pt/C and Pt/Al2O3 catalysts. The hydrogena-
tion experiments were carried out under kinetic regime,
i.e., in the absence of both external and internal diffusion
limitations. In the presence of cinchonidine no rate en-
hancement was observed; the racemic hydrogenation pro-
ceeded with nearly the same rate as the selective hydro-
genation in the presence of the catalyst modifier. In situ
catalyst modification resulted in both a reaction rate and
enantioselectivity higher than those in catalyst premod-
ification. The large amount of the modifier and partial
catalyst deactivation due to the exposure of air during
the premodification were the probable causes for the low
enantioselectivity and reaction rate observed over the
premodified catalyst. The main product in the first hy-
drogenation step was 1-hydroxy-1-phenylpropanone; the
ratio between 1-hydroxy-1-phenylpropanone and 2-
hydroxy-1-phenylpropanone was about 11. Preferentially
(
R)-1-hydroxy-1-phenylpropanone was formed, while in
the second hydrogenation step the main product among
diols was (1R,2S)-1-phenyl-1,2-propanediol. Hydrogena-
tion of 1-phenyl-1,2-propanedione over the Pt/Al2O3 cata-
lyst showed that it is possible to obtain a high ee
(
65%) of (R)-1-hydroxy-1-phenylpropanone provided that
dichloromethane is used as a solvent and the cata-
FIG. 5. (a) The yield of 1-hydroxy-1-phenylpropanone in the hy- lyst is preactivated with H2 and modified in situ with
�
drogenation of 1-phenyl-1,2-propanedione at three temperatures: 0 C
cinchonidine. The ee was temperature independent at
�
�
(
᭿), 15 C (᭡), and 25 C (᭹) in dichloromethane. Catalyst: Pt/Al2O3
�
0
–25 C. The highest enantioselectivities in the hydrogena-
modified in situ with (-)-cinchonidine. (b) The enantiomeric excess of
tion of 1-phenyl-1,2-propanedione were obtained for (R)-
1-hydroxy-1-phenylpropanone with the Pt/Al O catalyst
(
R)-1-hydroxy-1-phenylpropanone in the hydrogenation of 1-phenyl-
�
�
�
1
,2-propanedione at three temperatures: 0 C (᭿), 15 C (᭡), and 25 C
2
3
(
(
᭹) in dichloromethane. Catalyst: Pt/Al2O3 modified in situ with with metal particles of 8.3 nm (by XRD). The pores were
-)-cinchonidine.
sufficiently large for the reagents and the modifier to diffuse
into them. Contrary to the Pt/Al2O3 catalyst, the ee was rel-
atively low over the Pt/C catalyst, which can be attributed
that (S)-1-hydroxy-1-phenylpropanone reacted faster to
diols than its R-enantiomer. The effect of temperature
in the case of dichloromethane as solvent in the enan-
tioselective hydrogenation of 1-phenyl-1,2-propanedione
is shown in Figs. 5a and 5b and in Table 2. The initial
hydrogenation rate in dichloromethane was lower than that
in ethanol. This difference in the initial hydrogenation rate
can be due to the inert dichloromethane; i.e., the cata-
lyst surface is more efficiently modified by cinchonidine in
dichloromethane than in ethanol. The ee was about 65%,
and it was independent of dione conversion and temper-
ature (Fig. 5b). There was no kinetic resolution observed
in dichloromethane as solvent. This could be explained by
to the much smaller metal particle size of the Pt/C catalyst.
In a forthcoming paper optimization of the modifier
amount and an investigation of the reaction kinetics will
be reported. Further studies to quantify the influence on
reaction parameters on regioselectivity and enantioselec-
tivity are in progress.
APPENDIX
Asymmetric Synthesis of
(
R)-1-Hydroxy-1-phenylpropanone
The asymmetric synthesis of (R)-1-hydroxy-1-phenyl-
the fact that the consecutive reaction from hydroxyketones propanone (B, Fig. 1) was performed according to a method

HYDROGENATION OF 1-PHENYL-1,2-PROPANEDIONE
289
(
m-C-arom.);138.0(C1-arom.), and207.2(CO). Thesignals
were assigned by the use of tables (34).
Asymmetric Synthesis of
1R,2S)-1-Phenyl-1,2-propanediol
FIG. 6. Reaction scheme in the asymmetric synthesis of (R)-1-
hydroxy-1-phenylpropanone.
(
Procedure
based on the Grignard addition of methylmagnesium io-
dide to the protected (R)-(+)-mandelonitrile ((R)-(+)-2-
phenyl-2-hydroxyethanenitrile) as depicted in Fig. 6 (33).
It is known that the procedure would afford mainly the R
isomer, since the starting compound for the synthesis was
The asymmetric synthesis of (1R,2S)-1-phenyl-1,2-
propanediol (F, Fig. 1) was performed by the LiAlH re-
4
duction of the carbonyl group of the �-ketols obtained in
the previously described synthesis, Fig. 7 (35). Because of
the starting material used, it is known that the synthesis
would afford all the four possible isomers. The determina-
tion of their structures was based on a combination of the
GC equipped with a chiral column, molecular modeling,
(
R)-(+)-mandelonitrile.
Protection of the Hydroxyl Group
of the (R)-(+)-Mandelonitrile
1
and H NMR.
In a dry round bottom flask, 5.3 g (78 mmol) of imidazole
In a dry flask, under argon atmosphere, approximately
.20 g (32 mmol) of LiAlH4 and tetrahydrofuran (THF)
3
and 75 cm of distilled dimethylformamide (DMF) were
1
�
3
added. After cooling to 0 C, 7.2 cm (6.2 g, 57 mmol) of
trimethylsilylchloride (TMSCl) was added under stirring.
After 15 minutes of stirring, 5.044 g (38 mmol) of (R)-
were added. The glassware for the reactions was dried by
heating under vacuum. The diethyl ether and the THF used
were purified by distillation over Na/benzophenone under
(
+)-mandelonitrile (Aldrich) was added. The mixture was
�
argon atmosphere. After cooling to approximately –90 C, a
stirred for 1.45 h at room temperature. The reaction was car-
ried out under argon atmosphere. An amount of 150 cm3
of water was added and the products were extracted with
solution of 1.69 g of the oil obtained in the previous synthe-
sis in THF was slowly added. The mixture was stirred and
slowly heated to room temperature for 25 h. The flask con-
tent was poured into another flask with water under stirring
and acidified with dilute hydrochloric acid. The products
3
three 50 cm portions of diethyl ether. The organic layer
was dried over sodium sulfate and then the solvent was
removed. This procedure afforded a light brown oil. The
oil was analyzed by GC (Varian 3300, DB-1 column, 20 m,
3
were extracted with three approximately 50 cm portions of
diethyl ether. The organic layer was washed with water and
then dried over sodium sulfate. The solvent was removed.
To remove all the THF and other volatile substances, the
mixturewasheatedundervacuum. Thisprocedureafforded
0
.53 mm i.d.).
Synthesis of the �-Ketol
In a dry round bottom flask, supplied with a reflux con- 1.26 g of a light brown oil. This oil was analyzed by NMR
denser and a dropping funnel, under argon atmosphere, (JEOL JNM-A500, 500 MHz), GC (Varian 3300, column
3
1
5 cm of a 3 M solution (44 mmol) of methylmagnesium DB-1, 20 m, 0.53 mm i.d.), and GC-MS (HP 5973, column
3
iodide in diethyl ether (Aldrich) and 85 cm of diethyl ether HP-1, 15 m, 0.2 mm i.d., 0.25 �m film thickness, MS pa-
were added. A solution of the oil obtained in the first step rameters: E.I. at 70 eV, scan range 35 to 600 a.m.u.). The
3
in 50 cm of diethyl ether was slowly added through the spectral data were in agreement with the molecular struc-
dropping funnel. The mixture was stirred and refluxed for ture of the target compound. Pair of diastereomers (1R,2S)
1
6
h. The mixture was added to a flask containing 120 g of and (1S,2R)-1-phenyl-1,2-propanediol: H-NMR (CDCl3)
3
ice and 5 cm of concentrated sulfuric acid. The resulting � (ppm) : 1.05 (CH3, d, J = 6.4 Hz, 3H); 2.60–3.20 (OH,
mixture was stirred overnight. The aqueous layer was ex- br, 1H); 3.97 (CH3-CHOH-CHOHC6H5, qd, J = 6.4 and
3
tracted with three 80 cm portions of diethyl ether. The or-
ganic layers were combined and dried over sodium sulfate.
The solvent was removed. This procedure afforded 2.32 g
of a light brown oil. This oil was analyzed by NMR and GC
1
(
Varian 3300, DB-1 column, 20 m, 0.53 mm i.d.). The H
NMRspectralcharacteristics(JEOLJNM-A500, 500MHz)
were in good agreement with literature data, whereas the
1
3
C NMR data (JEOL JNM-LA400, 400 MHz) have not
1
been reported previously. H NMR (CDCl3) �: 2.05 (CH3,
s, 3H); 4.25–4.35 (OH, br, 1H); 5.07 (CH, s, 1H), and 7.29–
13
7
8
.38 (arom.-H, m, 5H). C NMR (CDCl3) �: 25.3 (CH3);
0.2 (CHOH); 127.4 (o-C-arom.); 128.8 (p-C-arom.); 129.1 phenyl-1,2-propanonedione.
FIG. 7. Reaction scheme in the asymmetric synthesis of (1R,2S)-1-

2
4
90
TOUKONIITTY ET AL.
.2 Hz, 1H); 4.64 (CH3-CHOH-CHOHC6H5, d, J = 4.1 trum of the first pair should be smaller than that of
Hz, 1H); and 7.20–7.40 (arom.-H, m, 5H). Pair of diastere- the other pair, according to the Karplus equation (36).
omers (1R,2R) and (1S,2S)-1-phenyl-1,2-propanediol: 1.03 This conclusion was supported by the H NMR data on
1
(
(
4
7
CH3, d, J = 6.3 Hz, 3H); 2.60–3.20 (OH, br, 1H); 3.83 (1R,2R)-1-phenylpropane-1,2-diol reported by Sharpless
CH3-CHOH-CHOHC6H5, dq, J = 7.5 and 6.3 Hz, 1H); and co-workers (38). Considering this information,
.32 (CH3-CHOH-CHOHC6H5, d, J = 7.5 Hz, 1H); and it was possible to find out, by integration of the signals of
1
.20–7.40 (arom.-H, m, 5H).
the protons in the H NMR spectrum, that the ratio be-
The minimum energy conformations for the two pairs of tween the (1R,2S) and (1S,2R) and the (1R,2R) and (1S,2S)
diastereomers ((1R,2S), (1S,2R)) and ((1R,2R), (1S,2S)) was approximately 76:24. This ratio was in good agreement
were determined by molecular mechanics calculations us- with the ratio found by the authors of the reference (38)
ing Cache programme (15) and are presented in Fig. 8 (80:20). So, it was known that the isomer (1R,2S) would be
for the major (1-hydroxy) product. Based on the obtained that obtained in larger amounts, followed by the (1R,2R)
data, the hydrogen atoms attached to C1 and C2 in the isomer and after these would come the (1S,2R) isomer, fol-
minimum energy conformation of the major pair of di- lowed by the (1S,2S) isomer. With this knowledge it was
astereomers are in a gauche relationship. On the other possible to assign the four peaks due to the diols obtained
hand, the same hydrogens in the minimum energy confor- by the analysis of the asymmetric synthesis product of the
mation of the minor pair of diastereomers have a trans (1R,2S)-1-phenyl-1,2-propanediol. By this analysis it was
relationship to each other. Thus, the coupling constant also possible to determine the ratio between both pairs
1
(
J) between these two hydrogens in the H NMR spec- of diastereomers, which was found to be approximately
7
5 : 25.
Synthesis of Racemic 1-Hydroxy-1-phenylpropanone
The synthesis of the racemic mixture of the 1-hydroxy-
-phenylpropanone was carried out by the method used
1
in the synthesis of (R)-1-hydroxy-1-phenylpropanone (B,
Fig. 1)bystartingfromtheracemicmandelonitrile(Aldrich,
technical grade) in order to make the calibration curve for
gaschromatographic analysis. It was found out that some
impurities in the racemic mandelonitrile gave acidic char-
acteristics to this material. Hence, it was purified using an
aqueous solution of sodium hydrogen carbonate. Then the
protection of the hydroxyl group and the Grignard reac-
tion were carried out as described previously. The proce-
1
dure afforded a light brown oil, which was analyzed by H
NMR (JEOL NMR-LA400, 400 MHz), GC (Varian 3300,
DB-1 column, 20 m, 0.53 mm i.d.), and GC-MS (HP 5890
series II, column DB-1, 25 m, 0.32 mm i.d., 0.25 �m film
thickness, MS parameters: HP 5971-A, E.I. at 70 eV). The
purity of the product, 1-hydroxy-1-phenylpropanone was
1
5
7.9% as determined by H NMR analysis. 1-Hydroxy-1-
1
phenylpropanone H-NMR (CDCl3) � 2.09 (CH3, s, 3H);
.35 (OH, d, 1H); 5.10 (CH, d, 1H); and 7.27–7.43 (Ar.-
H, m, 5H). 2-Hydroxy-1-phenylpropanone � 1.5 (CH3, d,
H); 3.84 (OH, d, 1H); 5.13–5.20 (CH, m, 1H); 7.49–7.54
4
3
(
(
Ar.-H, m, 2H); 7.61–7.66 (Ar-H, m, 1H); and 7.92–7.95
Ar-H, m, 2H). The purity of the product was relatively low
due to the technical quality of the starting material, man-
delonitrile. During this purification, the other �-hydroxy-
ketone (2-hydroxy-1-phenylpropanone) was formed. We
believe that it happened via the �-ketol rearrangement
since the mixture possibly contained small amounts of
acid from the previous step. The purity of this com-
FIG. 8. The energy minimized conformations of (a) (1R,2S)-1-
phenyl-1,2-propanediol and (b) (1R,2R)-1-phenyl-1,2-propanediol mod-
1
pound was also determined by quantitative H NMR to be
eled with Cache program (15).
29.7%.

HYDROGENATION OF 1-PHENYL-1,2-PROPANEDIONE
291
Synthesis of Racemic 2-Hydroxy-1-phenylpropanone
7. Toukoniitty, E., M a¨ ki-Arvela, P., Villela, A., Kalantar Neyestanaki, A.,
Salmi, T., Leino, R., Sj o¨ holm, R., Laine, E., V a¨ yrynen, J., Ollonqvist, J.,
and Kooyman, P., Catal. Today 60, 175 (2000).
General Considerations
8
. Toukoniitty, E., M a¨ ki-Arvela, P., Kalantar Neyestanaki, A., Salmi,
T., Villela, A., Leino, R., Sj o¨ holm, R., Laine, E., V a¨ yrynen, J., and
Ollonqvist, T., Stud. Surf. Sci. Catal. 130, 3363, 2000.
All operations with organometallic reagents were car-
ried out in an argon atmosphere. Tetrahydrofuran was
dried and distilled from Na/benzophenone prior to use.
9. Adam, W., Diaz, M., Fell, R., and Saha-M o¨ ller, C., Tetrahedron
Asymmetry 7, 2207 (1996).
3
n-Butyllithium (Acros, 2.5 mol/dm solution in hexane)
1
1
1
1
1
0. Gala, D., DiBenedetto, D., Clark, J., Murphy, B., Schumacher, D.,
and Steinman, M., Tetrahedron Lett. 37, 611 (1996).
1. Subramanian, P., Chatterjee, S., and Bhatia, M., J. Chem. Tech. Bio-
technol. 39, 215 (1987).
2. Blaser, H. U., Jalett, H. P., Monti, D. M., Reber, J. F., and Wehrli,
J. T., Stud. Surf. Sci. Catal. 41, 153 (1988).
1
was used as received. Products were confirmed by H
and 13C NMR spectra recorded in CDCl3 solution using
a JEOL JNM-LA400 spectrometer and referenced against
the residual protons of the deuterated solvent.
3. Reschetilowski, W., B o¨ hmer, U., and Wiehl, J., Stud. Surf. Sci. Catal.
Preparation of 2-Hydroxy-1-phenylpropanone
8
4, 2021 (1994).
4. Blaser, H., Jalett, H., Monti, D., Baiker, A., and Wehrli, J., in
Structure-Activity and Selectivity Relationship in Heterogeneous
(
D + E, Fig. 1)
-Phenyl-1,3-dithiane was prepared from benzaldehyde
Baker) and propane-1,3-dithiol (Acros) as described pre-
viously (38). The racemic 2-hydroxy-1-phenylpropanone
“
2
Catalysis” (R. Grasselli and A. Sleight, Eds.), p. 147. Elsevier Science,
Amsterdam, 1991.
5. Cache Work System, Version 3.2, 1999.
(
1
(
1
D + E, Fig. 1) was prepared by lithiation of 2-phenyl- 16. B u¨ rgi, T., and Baiker A., J. Am. Chem. Soc. 120, 12920 (1998).
17. Minder, B., Mallat, T., Skrabal, P., and Baiker, A., Catal. Lett. 29, 115
,3-dithiane with n-butyllithium and the subsequent re-
(
1994).
action with acetaldehyde followed by treatment with
HgCl2/CaCO3 in aqueous MeOH (35). Distillation gave
fairly pure 2-hydroxy-1-phenylpropanone as a yellow oil
1
1
8. Baiker, A., J. Mol. Catal. A: Chem. 115, 473 (1997).
9. Toukoniitty, E., Sev cˇ ´ı kov a´ , B., Kumar, N., M a¨ ki-Arvela, P., Salmi,
ˇ
T., V a¨ yrynen, J., Ollonqvist, T., Laine, E., Kooyman, P. J., and Murzin,
D. Yu., Stud. Surf. Sci. Catal. 135, 23 (2000).
0. Fogg, P., and Gerrard, W., “Solubility of Gases in Liquids,” p. 307.
Wiley, Chichester, 1991.
�
(
bp. 84–86 C/0.5 mbar) that was purified by crystallization
�
2
2
2
from diethyl ether at � 30 C. The purity of 2-hydroxy-1-
phenylpropanone was determined to be 95.8 % by quanti-
1. Reichardt, C., in “Solvents and Solvent Effects in Organic Chemistry,”
1
1
tative H NMR. H NMR (CDCl3, �) : 1.42 (d, CH3, J =
2nd ed., p. 408. VCH, New York, 1990.
7
.0 Hz, 3H); 3.80 (d, -OH-, J = 6.3 Hz, 1H); 5.14 (dq,
2. Sexton, B., Rendulic, K., and Hughes, A., Surf. Sci. 121, 181 (1982).
-
3
6
CO-CH-, J = 7.0 and 6.3 Hz, 1H); 7.61–7.56 (m, Ar–H, 23. Blaser, H.-U., Jalett, H.-P., M u¨ ller, M., and Studer, M., Catal. Today
1
3
3
7, 441 (1997).
4. Mallat, T., Bodnar, Z., Minder, B., Borzeky, K., and Baiker, A., J. Catal.
68, 183 (1997).
H); 7.92–7.88 (m, Ar–H, 2H). C NMR (CDCl3, �): 22.22;
2
9.25; 128.59; 128.81; 133.91; 202.31.
1
2
2
5. Murzin, D. Yu., and Kul’kova, N. V., Sov. Chem. Indus. 635 (1992).
6. Reid, R. C., Prausnitz, J. M., and Poling, B. E., “The Properties of
Gases and Liquids,” 4th ed. McGraw–Hill, New York, 1988.
ACKNOWLEDGMENTS
˚
27. Toukoniitty, E., M a¨ ki-Arvela, P., Nieminen, V., Hotokka, M.,
P a¨ iv a¨ rinta, J., Salmi, T., and Murzin, D. Yu., submitted for
publication.
This work is part of the activities at the Abo Akademi Process Chem-
istry Group within the Finnish Centre of Excellence Programme (2000–
005) by the Academy of Finland. The authors express their gratitude to
2
2
2
3
3
8. Ferri, D., B u¨ rgi, T., and Baiker, A., J. Chem. Soc. Perkin Trans. 2, 221
2000).
9. Phillipson, J. J., Wells, P. B., and Wilson, G. P., J. Chem. Soc. A 135
1969).
0. Murzin, D. Yu., Kul’kova, N. V., and Temkin, M. I., Kinet. Katal. 31,
83 (1990).
1. Murzin, D. Yu., React. Kinet. Catal. Lett. 58, 65 (1996).
Mr. Markku Reunanen from GC-MS analysis, to Mr. Patrick Eriksson for
the particle size measurements, and to Mr. Clifford Ekholm for the EPMA
analysis. The project was economically supported by the Academy of
Finland. An unknown reviewer is gratefully acknowledged for many sug-
gestions, which helped to improve the original manuscript.
(
(
9
REFERENCES
32. Horiuti, J., J. Res. Inst. Catal. Hokkaido Univ. 5, 1 (1957).
3. Temkin, M. I., Adv. Catal. 28, 173 (1979).
3
1
2
. Baiker, A., J. Mol. Catal. A: Chem. 163, 205 (1997).
. Vermeer, W., Fulford, A., Johnston, P., and Wells, P., J. Chem. Soc.,
Chem. Commun. 1053 (1993).
. Griffiths, S., Johnston, P., Vermeer, W., and Wells, P., J. Chem. Soc.,
Chem. Commun. 2431 (1994).
. Slipszenko, J., Griffiths, P., Johnston, P., Simons, K., Vermeer, W., and
Wells, P., J. Catal. 179, 267 (1998).
. M a¨ ki-Arvela, P., Kuzma, M., Sj o¨ holm, R., Leino, R., and Salmi, T., in
34. Brussee, J., Roos, E. C., and Van Der Gen, A., Tetrahedron Lett. 29,
4485 (1988).
35. Hase, T., Tables for Organic Spectrometry, number 893. Otatieto,
Espoo, 1992.
36. Bowlus, S. B., and Katzenellenbogen, J. A., J. Org. Chem. 39, 3309
(1974).
37. Jackman, L. M., and Sternhell, S., “Applications of Nuclear Magnetic
Resonance Spectroscopy in Organic Chemistry,” p. 280. Pergamon
Press, Oxford, 1969.
3
4
5
“
9th International Symposium on Relations between Homogeneous
and Heterogeneous Catalysis, Book of Abstracts,” P23, 1998.
. Toukoniitty, E., M a¨ ki-Arvela, P., W a¨ rn a˚ , J., and Salmi, T., Catal. Today
38. Norrby, P.-O., Becker, H., and Sharpless, K. B., J. Am. Chem. Soc. 118,
35 (1996).
6
6
6, 411 (2001).
39. Seebach, D., and Corey, E., J. Org. Chem. 40, 231 (1975).