Additional data to the origin of rate enhancement in the enantioselective hydrogenation of activated ketones over cinchonidine modified platinum catalyst

Article abstract of DOI:10.1016/j.jcat.2009.06.009

Kinetic behaviour of different substrates such as purified ethyl pyruvate, dimer containing (20%) ethyl pyruvate, methylbenzoyl formate and ketopantolactone was investigated in both racemic and enantioselective hydrogenation over Pt/Al₂O₃ catalysts. Upon introducing the chiral modifier by injection under condition of racemic hydrogenation an immediate increase in reaction rate is observed in the case of all substrates. Consequently, significant rate enhancement (RE) was obtained in the case of all substrates. The RE increased in the following order: ketopantolactone < ethyl pyruvate < methylbenzoyl formate. This order does not follow the ability of substrates to be involved in various undesired side reactions with the formation of poisonous surface residues. Accordingly, results obtained in this study confirm that the RE must be an intrinsic feature of the asymmetric hydrogenation of activated ketones in the presence of cinchona alkaloids. However, our results also indicate that the poisoning effect by organic residues originated from ethyl pyruvate cannot be neglected.

Full text of DOI:10.1016/j.jcat.2009.06.009

Journal of Catalysis 266 (2009) 191–198
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Additional data to the origin of rate enhancement in the enantioselective
hydrogenation of activated ketones over cinchonidine modified platinum catalyst
a
b
Emília Tálas a, , József L. Margitfalvi , Orsolya Egyed
*
a Chemical Research Center, Hungarian Academy of Sciences, Institute of Nanochemistry and Catalysis, 1025 Budapest, Pusztaszeri út 59-67, Hungary
b Chemical Research Center, Hungarian Academy of Sciences, Institute of Structural Chemistry, 1025 Budapest, Pusztaszeri út 59-67, Hungary
a r t i c l e i n f o
a b s t r a c t
Article history:
Kinetic behaviour of different substrates such as purified ethyl pyruvate, dimer containing (20%) ethyl
pyruvate, methylbenzoyl formate and ketopantolactone was investigated in both racemic and enantiose-
lective hydrogenation over Pt/Al2O3 catalysts. Upon introducing the chiral modifier by injection under
condition of racemic hydrogenation an immediate increase in reaction rate is observed in the case of
all substrates. Consequently, significant rate enhancement (RE) was obtained in the case of all substrates.
The RE increased in the following order: ketopantolactone < ethyl pyruvate < methylbenzoyl formate.
This order does not follow the ability of substrates to be involved in various undesired side reactions with
the formation of poisonous surface residues. Accordingly, results obtained in this study confirm that the
RE must be an intrinsic feature of the asymmetric hydrogenation of activated ketones in the presence of
cinchona alkaloids. However, our results also indicate that the poisoning effect by organic residues orig-
inated from ethyl pyruvate cannot be neglected.
Received 20 April 2009
Revised 3 June 2009
Accepted 4 June 2009
Available online 15 July 2009
Keywords:
Enantioselective hydrogenation
Rate enhancement
Ethyl pyruvate
Condensation products
Methylbenzoyl formate
Ketopantolactone
Low substrate concentration
Pt/Al2O3
Ó 2009 Published by Elsevier Inc.
Cinchonidine
1. Introduction
is considered ligand accelerated if there is a slower unmodified
(unselective) cycle and a faster modified (selective) cycle” analo-
Heterogeneous enantioselective hydrogenation of activated ke-
tones over Pt-cinchona catalysts named as Orito’s reaction has
been widely investigated by different research groups [1–4]. The
above-mentioned reaction is very unique, because in the presence
gous to the phenomenon described previously for homogeneous
catalysis [12]. The origin of the rate enhancement was attributed
to the difference in the rate-determining step, which is the uptake
of first H atom on the modified catalyst and that of the second H
atom on the unmodified Pt sites [2]. Recently upon using a series
of single-walled carbon nanotubes (SWNTs)-supported Pt nano-
particle as catalyst with different controlled Pt loadings it has been
shown that asymmetric hydrogenation of ethyl pyruvate (EtPy)
was a ‘‘ligand-accelerated” reaction [13].
Other authors have given a completely new interpretation of
the RE phenomenon: ‘‘rate enhancement is now attributed to reac-
tion occurring at a normal rate at an enhanced number of sites, not
(as previously proposed) to a reaction occurring at an enhanced
rate at a constant number of sites [14]. It was concluded that the
‘‘rate enhancement in the presence of an alkaloid modifier is attrib-
uted to the inhibition of the pyruvate ester polymerization at the Pt
surface”. These authors suggested the formation of two types of ad-
duct: (i) a semi-ketal formed by reaction of the product and the
substrate molecule, and (ii) a dimer of the substrate formed in a
dimerization reaction with the involvement of the enol and the
keto forms of the substrate. It has been considered that the dimer
formation is an equilibrium reaction. The role of tertiary amines is
to shift the equilibrium to the formation of free substrate on the
of cinchonas a definite class of substrates, such as different a-keto
esters, shows not only high enantioselectivities, but also strong
rate enhancement (RE) [5–9]. Nowadays the origin of RE is the sub-
ject of a scientific debate. In an early work three explanations for
the rate enhancement have been suggested: (i) the substrate is
activated by the tertiary N of the quinuclidine part of cinchonas
(the rate of racemic reaction is also increased in the presence of
quinuclidine added); (ii) the H coverage is higher over the modified
catalyst than over the unmodified catalyst (H2/D2 exchange has in-
creased); and (iii) the quinoline part adsorbed generates an elec-
tronic interaction (the rate of racemic reaction has increased in
the presence of quinoline) [10].
According to ligand-acceleration model [11] chirally modified
surface sites are created by the reversible adsorption of cinchona
molecules onto Pt. The number of the total surface sites is equal
to the sum of modified (Ptm) and unmodified (Ptu) sites. ‘‘A reaction
* Corresponding author. Fax: +36 1-4281143.
E-mail address: joemarg@chemres.hu (E. Tálas).
0021-9517/$ - see front matter Ó 2009 Published by Elsevier Inc.
doi:10.1016/j.jcat.2009.06.009

192
E. Tálas et al. / Journal of Catalysis 266 (2009) 191–198
surface (see Scheme 1 in Ref. [14]). Experiments with EtPy at low
concentration (10ꢀ2 M) have shown that the initial rate of racemic
and enantioselective reaction does not differ [15]. Continuous
fixed-bed reactor experiments revealed that ligand acceleration
can be linked to catalyst deactivation [15], although in the recent
studies [16,17] just the opposite was demonstrated. In Ref. [15] it
has been concluded that cinchonidine (CD) plays a double role:
(i) promoter via the inhibition of side reactions; (ii) chiral poison
at high surfaces coverage. According to this idea the ligand acceler-
ation during the hydrogenation of activated ketones over cin-
chona-modified platinum catalyst originates from the effect of
alkaloids to prevent the deactivation of catalyst induced by EtPy
[15].
The new conception discussed above is supported by several
observations from different areas of the investigation of activated
ketones. Formation of surface polymer under hydrogen poor cir-
cumstances was demonstrated by XPS, NEXAFS and STM [18].
The presence of adsorbed hydrogen suppressed the above-men-
tioned process [18]. From gas phase FTIR measurement it has be-
came clear that EtPy decomposes at Pt surface forming CO and
CxHyOz strongly bonded species [19]. Upon using ATR-IR method
it has been proved that the decomposition of EtPy is 60-fold faster
in the absence of CD than in the presence of alkaloid of 0.001 M
[20]. Hydrogenation of methyl pyruvate in gas phase resulted in
low optical yield values (651%) without rate enhancement [21].
Furthermore, several substrates hydrogenated over CD–cinchona
catalyst system give moderate RE but high ee [3].
However, other results support the idea that the rate enhance-
ment is the inherent feature of this catalytic reaction. Injection of
CD into the racemic reaction mixture in a batch reactor has led
to immediate increase in conversion [22–24]. Results obtained
upon using different modes of introduction of the substrate and
the modifier has indicated that the overall rate increase is a kinetic
phenomenon and cannot be attributed to the suppression of the
formation of oligomers [25]. In a continuous flow experiment cat-
alyst deactivation during pyruvate hydrogenation can be avoided
by proper selection of the solvent and reaction conditions, and
the introduction of the chiral modifier into the reaction mixture in-
duces considerable rate acceleration. These observations contra-
dict again the proposal that the origin of ‘‘ligand acceleration” is
the suppression of catalyst deactivation by the modifier [16]. In an-
other view it was emphasized: ‘‘using low ethyl pyruvate concen-
trations (0.01 to 0.1 mol lꢀ1) in a batch reactor, two mechanistic
cornerstones for Orito’s reaction were no longer valid – that is,
the reaction rate was not proportional to the modifier concentra-
tion, and a high ee could be obtained with no rate acceleration or
even in the presence of rate deceleration” [26]. But Baiker and
Mallat [27] have pointed out that even if the blocking of surface
sites by byproducts and impurities is negligible, an equal or higher
overall reaction rate on chirally modified Pt means intrinsic rate
acceleration itself because the Pt is highly covered by CD, conse-
quently the number of available active sites is smaller. From con-
tinuous hydrogenation experiments using methyl benzoylformate
(MBF), KPL and pyruvic aldehyde dimethyl acetal as substrates
and alternating CD, cinchonine, quinine and quinidine chiral
modifiers in time on stream experiments it has been concluded
that enantiodifferentiation and RE have the same origin, both
may be traced back to the role of the intermediate complexes of
the hydrogenation [17]. Recently the intrinsic kinetic acceleration
(e.g. the ratio of kinetic constants for enantioselective and racemic
reactions) has been formulated mathematically [28]. For the
hydrogenation of 1,2-phenylpropanedione over Pt/Al2O3 catalyst
this value has been found close to 2 [28].
The aim of this study is to get additional experimental proof re-
lated to the intrinsic character of the rate enhancement. The use of
term ‘‘intrinsic” means that we do not accept the recent views
[14,15] related to the involvement of poisonous residues formed
from the substrate in the increase of the reaction rate in enantiose-
lective hydrogenation of activated ketones compared to the rate in
racemic hydrogenation.
In this work we investigated the poisoning effect of the conden-
sation byproduct formed from EtPy on the reaction rate and the
optical yield in the enantioselective hydrogenation of EtPy. New ki-
netic treatment was applied to evaluate the effects of poisoning. In
addition the behaviour of other substrates which are unable to
form polymeric residues were also investigated. These studies
were performed especially at low initial concentrations of these
substrates. In some of these experiments transient method was ap-
plied, where the modifier was injected into the reactor under con-
dition of racemic hydrogenation. Upon using these tools our
purpose was to get further information about the character of rate
enhancement in the cinchona modified catalytic hydrogenation of
activated ketones. General reaction scheme of hydrogenation reac-
tions investigated in this work is summarized in Scheme 1.
2. Experimental
2.1. Materials
MBF and KPL substrates were purchased from Fluka and Al-
drich, respectively, and were used as received. EtPy was a Fluka
product, and was distilled under vacuum prior to use ((EtPy)1
).
Ethyl pyruvate contaminated with a dimer ((EtPy)2, dimer content
20%) was also used as a substrate. This product was obtained from
the distillation residue and was identified by 1H NMR and mass
spectrometry. The reactions were performed in dried toluene sol-
vent (Reanal). CD chiral modifier was a Fluka product and was used
as received. Commercial 5 wt% Pt/Al2O3 such as E4759 (Engelhard,
H/Pt = 0.27, CO/Pt = 0.24 [29]) and CatASiumF214 (Degussa, CO
chemisorption = 1.5 cm3 CO/g catalyst, Pt dispersion: 26.3% [30])
were used as catalysts. The first one was treated in hydrogen at
400 °C for 2 h before the catalytic reaction; the second one was
used as received.
2.2. Kinetic investigations
Hydrogenation of MBF, KPL and ethyl pyruvate (EtPy)1 and
(EtPy)2 was carried out in a 300 cm3 SS-autoclave at 20 °C. The
pressure of H2 was 50 bar in all experiments. The rate of agitation
was 500 or 1000 rpm. 1.2 ꢁ 10ꢀ5 M or 1.0 ꢁ 10ꢀ4 M was chosen for
the concentration of CD. The initial concentration of the substrates
was in the range of 2 ꢁ 10ꢀ2 to 1.0 M. Both injection and premixing
techniques were used to introduce the modifiers. In the injection
Scheme 1. Hydrogenation of activated ketones (EtPy = ethyl pyruvate, MBF = meth-
ylbenzoyl formate, KPL = ketopantolactone, (R)-PL = R-(ꢀ)-pantolactone).

E. Tálas et al. / Journal of Catalysis 266 (2009) 191–198
193
method the toluene solution of cinchonidine (5 vol.% of total liquid
volume) was used. In all experiments dried and distilled toluene
was applied and the ratio of Vinj/Vtotal was 1:20. Further details
can be found elsewhere [7,31]. Racemic reaction was performed
with premixing solvent, substrate and catalyst under inert atmo-
sphere or with injection of substrate to the slurry of solvent and
catalyst under hydrogen atmosphere. In certain experiments race-
mic reaction was started and the toluene solution of the chiral
modifier was injected into the reactor by time delay (transient
experiments). Samples were taken at different reaction times and
were analyzed by a GC using a capillary column (Supelco BETA
DEX 225) and flame ionization detector. The enantioselectivity is
expressed as ee = [R ꢀ S]/[R + S]. The eemax means the highest
enantiomeric excess value measured in a given reaction. The eeend
values were measured at the end of reaction. Calculation of ee in
the experiments of CD injection with time delay was based on
the change of products concentrations during the enantioselective
part. First-order rate constants k1 were calculated from experimen-
tal points measured in the first 3 to 10 min described earlier
[31,32]. Initial rate (r0) values were calculated from d[S + R]/dt
according to the method written in Ref. [25]. The comparison of
catalytic activities under various experimental conditions was
based on r0 and k1. The k1 values were reproducible within 10%
providing the use of the same batch of EtPy. The ee values in the
conversion – ee dependencies had a relative error in the range of
5%, but the shapes of these dependencies were completely repro-
marized in Figs. 1 and 2. In these experiments CD was injected into
the reactor with time delay. From Fig. 1A, it can be seen that upon
using substrate in the initial concentration of 1 M, the introduction
of CD during racemic hydrogenation induces instantaneous rate in-
crease and the time delay in the introduction of CD does not result
in any relative rate decrease (rrt ¼ r0;t=r0;t ¼ 0, where the t val-
ues corresponds to the time of injection of CD). By comparing Figs.
1A and 2A, it is obvious that upon injection of CD a definite instan-
taneous increase in the conversion appears at both concentrations
of (EtPy)1. Figs. 1B and 2B show that optical yield also increases
instantaneously parallel to the immediate conversion increase,
however, both the monotonic increase type behaviour and the final
ee values were maintained. Figs. 1C and 2C represent the product
formation in coordinates [R–S] vs. 2[S]. In our recent study [25]
it was shown that the slope in these dependencies corresponds
to the ke/kr ratio, where ke and kr are the first-order rate constants
in enantioselective and racemic hydrogenation reactions, respec-
tively. Fig. 1C shows that upon increasing the time between the
start of the racemic hydrogenation and the introduction of CD
the ke/kr ratios increased from 9.1 to 12.2 and 15.8, for injection
time 0, 15 and 30 min, respectively. This tendency was also ob-
served at 0.25 M of (EtPy)1 (compare Figs. 1C and 2C).
3.2. Hydrogenation of dimer containing ethyl pyruvate
In the second series of experiments dimer-containing ethyl
pyruvate (EtPy)2, originated from the distillation residue of EtPy
was used as a substrate. Based on Ref. [33] the possible struc-
ture of dimers is shown in Scheme 2. The presence of dimer
was confirmed by MS and NMR measurements. Based on NMR
analysis dimer 1a content of the substrate was in the amount
of 20 wt%.
duced. The relative rate increase is expressed by k1e/k1r and r0e
/
r0r which is the ratio of the given parameters in enantioselective
and racemic hydrogenations. The slope of dependencies in the
[R ꢀ S] vs. 2[S] coordinates gives the relative RE within the enan-
tioselective hydrogenation as the ke/kr ratio expresses the ratio of
rate constants in enantioselective and racemic hydrogenations
running parallel in the same experiment [25].
Results given in Fig. 3A clearly show that upon using the (EtPy)2
the reaction rates decreased significantly compared to (EtPy)1 (see
Fig. 1A). The decrease in the rate of racemic hydrogenation is
2.3. NMR measurements
around 11-fold (from 330 to 30 mmol ꢁ gꢀ1 hꢀ1). Consequently,
cat
All spectra were measured at 30 °C. Spectra were taken using a
Varian Unity Inova 400 spectrometer. Chemical shifts (d) are re-
ported relative to internal TMS. 512 transients were accumulated
to achieve the appropriate signal/noise ratio.
there is a strong poisoning effect induced by compound 1a. The
poisoning effect can also be observed in the enantioselective
hydrogenation; however, its extent is only threefold (from 2540
to 920 mmol ꢁ gꢀ1 hꢀ1). The most important observation that the
cat
introduction of CD during racemic hydrogenation of (EtPy)2 also
resulted in instantaneous rate acceleration in all cases (see
Fig. 3A). Contrary to the results obtained in the previous series of
experiments the rrt values decreased to 0.67, 0.59 and 0.38 for
experiments with delayed injection of CD in the 15th, 30th and
90th minute of racemic hydrogenation. The latter observation
may indicate that the amount of organic residues formed from
3. Results
3.1. Hydrogenation of purified ethyl pyruvate
Results obtained in two sets of experiments by using purified
substrate (EtPy)1 in two different initial concentrations are sum-
C
B
A
1.0
0.8
0.6
0.4
0.2
0.0
1
0.8
0.6
0.4
0.2
0
1.0
k
e/kr =
15.8
9.1 12.2
0.8
0.6
0.4
0.2
0.0
0.00 0.05 0.10 0.15 0.20 0.25
2[S]corr, M
0
20 40 60 80 100
time, min
0
20 40 60 80 100
time, min
Fig. 1. A. Kinetic curves of ethyl pyruvate hydrogenation upon using purified substrate (EtPy)1 at high substrate concentration. (A) Influence of the time delay in CD injection
during racemic hydrogenation. (B) Enantiomeric excess – reaction time dependencies (corrected for product concentrations formed during time delay). (C) Kinetic
dependencies in ([R–S]) vs. 2[S] coordinates. [(EtPy)1]0 = 1.0 M (purified by distillation prior to the use), [CD] = 5 ꢁ 10ꢀ5 M, T = 20 °C, pH = 50 bar, stirring rate = 500 minꢀ1
;
2
total volume = 100 cm3, catalyst = 5% Pt/Al2O3 (E 4759, dispersion: 0.27), mcat = 0.125 g, catalyst pretreatment = 2 h in flowing H2 at 400 °C and cooled in N2. N – CD injection
at 0 min; h – CD injection at 15 min; ꢁ – CD injection at 30 min; s – no CD.

194
E. Tálas et al. / Journal of Catalysis 266 (2009) 191–198
A
B 1.0
C
0.25
0.20
0.15
0.10
0.05
0.00
0.60
0.40
0.20
0.00
ke/kr
=
9.9
8.5
0.8
0.6
0.4
0.2
0.0
0.00
0.02
0.04
0.06
0
2
4
6
0
2
4
6
8
10
time, min
time, min
2[S]corr, M
Fig. 2. Hydrogenation of purified ethyl pyruvate (EtPy)1 over cinchona–Pt/Al2O3 catalyst at medium substrate concentration. [(EtPy)1]0 = 0.25 M; mcat = 0.063 g,
catalyst = E4759; Vr = 50 cm3; Tr = 20 °C; p = 50 bar; speed of agitation = 1000 rpm, catalyst pretreatment see Fig. 1. r – no CD; h – 1 ꢁ 10ꢀ4 M CD injection at 0 min; ꢁ
H2
– CD injection at 1 min.
(EtPy)2 increases with the time delay, i.e., the duration of racemic
hydrogenation.
(EtPy)2 have almost similar ke/kr ratios (9.1 and 8.9, respectively).
This fact let us to conclude, that not dimer 1a itself but poisoning
products formed from it are responsible for catalyst deactivation.
The injection technique used in our previous work [25] gives a
possibility to vary the order of introduction of components into the
reaction mixture. Results obtained by this method on purified
(EtPy)1 at 1 M initial concentrations let us to suppose that rate
enhancement is the intrinsic behaviour of this reaction as the rates
were not really influenced by the order of introduction of interact-
ing components [25].
Fig. 3C shows that in the [R–S] vs. 2[S] dependencies upon using
(EtPy)2 the increase in the injection delay from 0 to 15, 30 and
90 min the ke/kr ratios decreased from 8.9 to 8.2, 8.0 and 7.0. It is
contrary to the observation found for (EtPy)1, the distilled sub-
strate, which shows increasing tendency (compare Figs. 1C and
3C). It has to be emphasized that the direct injection of CD (delay
time is zero) both purified (EtPy)1 and the ‘‘distillation residue”
Considering the observation in Ref. [15], i.e. at low substrate
concentration the initial rate of racemic and chirally modified reac-
tion does not differ, additional experiments were performed
decreasing the substrate concentration by one order. Results of
these measurements are summarized in Table 1. In racemic hydro-
genation performed over CatASiumF214 at low (EtPy)1 concentra-
tion the reaction rate (r0 and k1 values) was higher when the
substrate was injected into the slurry of catalyst and solvent under
H2 atmosphere than that obtained at premixing the catalyst and
the substrate (compare experiments. 1, 2 and 7, 8 in Table 1). This
finding is in good agreement with the earlier results related to the
formation of polymeric residues over Pt surface in the absence of
H2 [18] and surface poisoning by fragments of EtPy decomposition
[19].
However, in the enantioselective hydrogenation the lowest ini-
tial rates were observed when EtPy was injected. This trend was
also observed when the amount of catalyst decreased. We have cal-
culated the relative rate increase induced by CD using the r0e/r0r or
k1e/k1r ratios, where r0e and r0r or k1e and k1r are initial reaction
rates and first-order rate constants for enantioselective and race-
mic reactions. The relative rate increase was measurable in all
Scheme 2. Condensation products of ethyl pyruvate (based on Ref. [34]).
1.0
0.8
B
A
C
1.0
0.8
0.6
0.4
0.2
0.0
ke/kr= 8.9 8.2 8.0 7.0
0.8
0.6
0.4
0.2
0.0
0.6
0.4
0.2
0.0
0
50
100
150
0
50
100
150
0.00 0.02 0.04 0.06 0.08 0.10
2[S]corr, M
time, min
time, min
Fig. 3. Kinetic curves of ethyl pyruvate hydrogenation upon using dimmer contaminated substrate (EtPy)2. (A) Influence of the time delay in CD injection during racemic
hydrogenation. (B) Enantiomeric excess – reaction time dependencies (corrected for product concentrations formed during time delay). (C) Kinetic dependencies in [R–S] vs.
2[S] coordinates. [(EtPy)2]0 = 1.0 M (distillation residue containing dimer 1a in the amount of 20%), other details of reaction conditions see in Fig. 1, N – CD injection at 0 min;
h – CD injection at 15 min; ꢁ – CD injection at 30 min; } – CD injection at 90 min; s – no CD.

E. Tálas et al. / Journal of Catalysis 266 (2009) 191–198
195
Table 1
Hydrogenation of (EtPy)1 over CatASiumF214 Pt/Al2O3 catalysts. Influence of the mode of introduction of reaction components.
No.
[CD] (10ꢀ5 M)
Mode of introduction
r0r (M minꢀ1
)
r0e (M minꢀ1
)
r0e/r0r
k1r (minꢀ1
)
k1e (minꢀ1
)
k1e/k1r
eemax
eeend
convend
1
2
3
4
5
6
–
–
–
0.0152
0.0271
–
–
–
–
0.1166
0.2488
–
–
–
–
–
–
–
–
0.988
0.953
0.951
0.997
0.965
0.954
0.943
0.920
0.964
0.994
0.910
0.947
(EtPy)1inj
premix
CD inj
(EtPy)1inj
CD inj at 5 min
–
(EtPy)1inj
Premix
1.2
1.2
1.2
1.2
–
0.0455
0.0580
0.0411
–
–
–
0.0262
0.0312
0.0148
–
3.0
3.8
1.5
0.4958
0.5423
0.3880
0.3161
–
4.3
4.7
1.6
2.1
0.765
0.730
0.702
0.779
–
0.756
0.718
0.679
0.779
–
–
–
0.0169
0.0074
0.0119
–
–
–
0.1488
0.0567
0.1127
–
–
–
7a,b
8a,b
9a,c
10a,c
11a
12a,c
–
–
–
–
1.2
1.2
1.2
1.2
3.5
4.2
1.2
0.1759
0.1775
0.1491
0.0973
3.1
3.1
1.3
1.8
0.801
0.766
0.724
0.763
0.795
0.754
0.714
0.741
CD inj
(EtPy)1inj
CD inj at 5 min
0.0076
0.0530
(EtPy)1 = ethyl pyruvate purified, CD = cinchonidine, inj = injection [(EtPy)1]0 = 0.1 M; solvent = toluene, mcat = 40 mg; catalyst = CatASiumF214 (Degussa); Vr = 40 cm3;
Tr = 20 °C; p = 50 bar; tr = 30 min; stirring rate = 1000 minꢀ1; r0 from d[S + R]/dt [25]; r-racemic; e-enantioselective.
H2
a
mcat = 20 mg.
tr = 90 min.
tr = 60 min.
b
c
cases. Similar trends were observed at both series of experiments
using different amounts of catalyst. The highest rate increase was
observed when CD was injected into the reaction mixture.
i.e., MBF is not able to form condensation products. In spite of this
fact the rate enhancement appeared instantaneously. These results
show again that the rate enhancement must be considered as an
inherent property of Orito’s reaction. Tendencies of [R–S] vs. 2[S]
dependencies obtained upon using MBF substrate (Fig. 4C) are very
similar to those of (EtPy)1 (Fig. 1C and 2C. Table 2 shows the kinetic
results of MBF hydrogenation upon the variation of initial concen-
tration of the substrate in the range of 0.02 to 0.5 M. It is notewor-
thy that the ee values slightly increase with increasing the
concentration of substrate till 0.25 M. A similar tendency has been
observed for EtPy [34]. Data in Table 2 show that the initial rate of
3.3. Hydrogenation of methylbenzoyl formate
Results obtained in the hydrogenation of MBF are summarized
in Fig. 4 and Table 2. Introduction of CD into the racemic reaction
led to immediate increase both in the conversion (Fig. 4A) and the
enantioselectivity (Fig. 4B) similarly to (EtPy)1 although this
increase is smaller. Note the different characters of substrates,
A 0.80
B 1.00
C
0.20
k /k =
ke/kr
=
13.6
15.6
21.3
16.8
e
r
0.80
0.60
0.40
0.20
0.00
0.60
0.40
0.20
0.00
0.15
0.10
0.05
0.00
0
20
40
60
80
0
20
40
60
80
0.00 0.01 0.02 0.03 0.04
time, min
time, min
2[S]corr, M
Fig. 4. Hydrogenation of methylbenzoyl formate (MBF) over cinchona–Pt/Al2O3 catalyst. (A) Influence of the time delay in CD injection during racemic hydrogenation. (B)
Enantiomeric excess – reaction time dependencies (corrected for product concentrations formed during time delay). (C) Kinetic dependencies in [R–S] vs. 2[S] coordinates.
[MBF]0 = 0.25 M; mcat = 0.063 g, catalyst = E4759; Vr = 50 cm3; Tr = 20 °C; pH2 = 50 bar; speed of agitation = 1000 rpm, catalyst pretreatment see Fig. 1, r – no CD; M –
1 ꢁ 10ꢀ4 M CD injection at 0 min; ꢁ – CD injection at 9 min.
Table 2
Hydrogenation of methylbenzoyl formate over cinchona-Pt/Al2O3 catalyst system. Influence of substrate concentration.
a
No.
[MBF]0 (M)
[CD]0 (10ꢀ4 M)
r0 (M minꢀ1
)
r0e/r0r
k1 (minꢀ1
)
k1e/k1r
eemax
eeend
tr (min)
convend
1
2
3
4
5
6
7
8
9
10
0.02
0.05
0.10
0.25
0.50
0.02
0.05
0.10
0.25
0.50
–
–
–
–
–
1
1
1
1
1
0.0021
0.0033
0.0040
0.0042
0.0046
0.0040
0.0078
0.0161
0.0211
0.0255
0.1347
0.0770
0.0508
0.0122
0.0073
0.2547
0.1642
0.1405
0.0624
0.0542
–
–
–
–
–
–
–
–
180
180
180
240
240
120
90
120
120
240
0.990
0.986
0.906
0.961
0.716
0.990
0.990
0.978
0.918
0.989
–
–
1.9
2.4
4.0
5.0
5.5
1.9
2.1
2.8
5.1
7.4
0.768
0.838
0.888
0.919
0.898
0.768
0.838
0.875
0.904
0.895
MBF = methylbenzoyl formate; mcat = 0.063 g, catalyst = E4759; Vr = 50 cm3; Tr = 20 °C; p = 50 bar; stirring rate = 1000 minꢀ1; r – racemic; e – enantioselective.
H2
a
r0 from d[S + R]/dt [25].

196
E. Tálas et al. / Journal of Catalysis 266 (2009) 191–198
B
A
0.030
0.020
0.010
0.000
500
y = 4.875x + 208.741
R2 = 0.995
400
300
200
100
0
y = 4.472x + 28.471
R2 = 0.992
20
40
0
0.2
0.4
0.6
0
60
1/[MBF]0, M-1
[MBF]0, M
Fig. 5. Initial rate-substrate concentration dependences (r0 from d[S + R]/dt [27]) MBF = methylbenzoyl formate, mcat = 0.063 g, catalyst = E4759; Vr = 50 cm3; Tr = 20 °C;
pH = 50 bar; speed of agitation = 1000 rpm, catalyst pretreatment see Fig. 1 r – no CD; h – 1 ꢁ 10ꢀ4 M CD.
2
racemic and enantioselective reaction at very low initial concen-
tration of the substrate is different. This observation is in contrast
to the results obtained in EtPy in Ref. [15], whereas r0 of racemic
and enantioselective hydrogenation was equal in the concentration
range of 0.01 to 0.1 M.
The influence of initial concentration on the r0 values of MBF
hydrogenation is depicted in Fig. 5. Initial rates vs. initial concen-
tration of MBF (Fig. 5A) resulted in the same saturation type of
dependencies for both racemic and enantioselective reactions.
The dependencies in co-ordinates reciprocal rates (1/r0) vs. reci-
procal concentration (1/M0) are demonstrated in Fig. 5B. This type
of linear dependencies is characteristic for substrates having reac-
tion orders between one and zero.
Similarly to MBF substrate KPL is not capable to form condensa-
tion product. For this reason it seemed to be worthwhile to com-
pare the behaviour of all three substrates (EtPy)1, MBF and KPL
under identical reaction conditions. Results of this series of exper-
iments are collected in Table 3. All three substrates showed a sig-
nificant rate enhancement. The analysis of r0e/r0a and k1e/k1r ratios
shows the following trend in the rate enhancement MBF >
(EtPy)1 > KPL, however, the above-mentioned order does not corre-
late with the ability of these compounds for the decomposition at
the catalyst surface.
As emerges from Figs. 1A and 2A at [(EtPy)1]0 = 1.0 and 0.25 M the
calculated initial rates at the moment of injection, within the
experimental error, were not altered. The relative rate increases
(k1e/k1r) are 12 and 3 at [(EtPy)1]0 = 1.0 and 0.25, respectively.
Figs. 1C and 2C show that upon increasing the time between the
start of the racemic hydrogenation and the introduction of CD the
ke/kr ratios increased slightly. Based on this observation it is as-
sumed that organic residues formed during racemic hydrogenation
have a positive effect on the enantioselectivity; they suppress the
rate of racemic hydrogenation without substantial influence on
the overall rate.
In these experiments the delay in the injection of CD had no
influence on either the final ee values or the form of ee-conversion
dependencies. All these findings indicate that racemic products
formed up to 10% conversion have no measurable influence on
the enantioselective hydrogenation of EtPy; consequently the size
of free Pt surface available for enantioselective hydrogenation is
not altered during racemic hydrogenation. When similar series of
experiments were performed in ethanol at [(EtPy)1]0 = 1 M in the
above conversion range, a slight decrease in the initial rates was
observed [22]. The relative rate decrease, rr22 (rrt ¼ r0;t=r0;t ¼ 0,
where the t values correspond to the time of injection of CD),
was in the range of 0.88. Further increase in the time of injection
resulted in more pronounced rate decrease ((rr80 = 0.35) see data
given in Table 1 in Ref. [22]). In ethanol it is assumed that the
poisoning effect is induced by the semi-ketal formed during
racemic hydrogenation. The semi-ketal formation in alcoholic
solvents and possible negative effect has been evidenced and
discussed in different studies [35–37].
4. Discussion
In this study four different substrates (EtPy)1, (EtPy)2, MBF and
KPL were hydrogenated under condition of racemic and enantiose-
lective hydrogenation. In all enantioselective hydrogenations CD
was used as a modifier. The aim of this study was to collect new
data supporting the view that RE is the intrinsic feature of Orito’s
reaction.
Both static and transient kinetic experiments were performed
to find comprehensive evidences for the validity the rate enhance-
ment phenomena. Upon using distilled substrate (EtPy)1 the injec-
tion of CD during racemic hydrogenation resulted in instantaneous
rate acceleration in the whole concentration range of the substrate.
Upon using (EtPy)2 contrary to results obtained for (EtPy)1 the
rrt values decreased to 0.67, 0.59 and 0.38 for experiments with
delayed injection of CD during racemic hydrogenation (15th,
30th and 90th min) (see Fig. 3A). Parallel to that the slopes in the
[R–S] vs. 2S coordinates decreased also with the time delay. This
decrease is quite substantial (from 8.9 to 7.0). These data may
indicate that in the racemic hydrogenation the amount of poison-
ous organic residues formed from (EtPy)2 increases with the time
Table 3
Hydrogenation of different substrates over cinchona-Pt/Al2O3 catalyst system.
No.
Substrate
[CD] (10ꢀ4 M)
r0r (M minꢀ1
)
r0e (M minꢀ1
)
r0e/r0r
k1r (minꢀ1
)
k1e (minꢀ1
)
k1e/k1r
eemax
eeend
tr (min)
convend
1
2
4
5
7
8
(EtPy)1
(EtPy)1
MBF
MBF
KPL
–
1
–
1
–
1
0.0225
–
0.0042
–
0.0155
–
–
0.0869
–
0.0122
–
0.0528
–
–
–
–
60
60
240
120
180
180
0.976
0.978
0.961
0.918
0.938
0.879
0.0585
–
0.0211
–
2.6
5.0
1.8
0.1803
–
0.0624
–
2.1
5.1
1.8
0.878
–
0.919
–
0.860
–
0.904
–
KPL
0.0283
0.0966
0.674
0.630
(EtPy)1 = ethyl pyruvate purified, MBF = methylbenzoyl formate, KPL = ketopantolactone, [Substrate]0 = 0.25 M; reaction time = 120 min; mcat = 0.063 g, catalyst = E4759;
Vr = 50 cm3; Tr = 20 °C; p = 50 bar; stirring rate = 1000 minꢀ1; r0 from d[S + R]/dt [25]; r-racemic; e-enantioselective.
H2

E. Tálas et al. / Journal of Catalysis 266 (2009) 191–198
197
delay. This finding is an indication that in the enantioselective
hydrogenation of (EtPy)2 the component 1a can be considered as
a definite source of poison. The longer the reaction time in racemic
hydrogenation the higher the extents of rate decrease. This obser-
vation strongly resembles our earlier results obtained in ethanol
solvent upon using freshly distilled substrate [22].
trend in the extent of RE might indicate that the formation of [sub-
strate–modifier] complex and its concentration either at the Pt sur-
face or in the liquid phase controls all the key events, i.e., both the
enantio-differentiation and the RE.
Finally let us discuss the behaviour of three substrates under
identical condition (Table 3). The results show that all these sub-
strates have RE. The extent of RE increases in the following order
KPL < EtPy < MBF. This order does not follow the ability of sub-
strates to be involved in various undesired side reactions with
the formation of poisonous surface residues. Consequently, the
RE cannot be attributed to the poisoning effect.
The analysis of results shown in Figs. 1–3 indicates that the
intrinsic RE cannot be questioned in the enantioselective hydroge-
nation of (EtPy)1. Results obtained in the hydrogenation of (EtPy)2
are even more convincing as in the presence of around 20 wt% high
molecular weight organic compound the rate enhancement was
instantaneous, although the initial rates both in racemic and enan-
tioselective hydrogenation decreased substantially (see Fig. 3). This
rate decrease is attributed to the presence of compound 1a.
Please note that in all experiments in Figs. 1–3 the conversion in
the moment of CD injection was not higher than 10%. It can be con-
cluded that the instantaneous rate increase observed in all exper-
iments is in a strong contradiction to the statement given in Ref.
[14], namely ‘‘rate enhancement is now attributed to reaction
occurring at a normal rate at an enhanced number of sites, not
(as previously proposed) to a reaction occurring at an enhanced
rate at a constant number of sites”. It is hardly to suggest that
the addition of 5 ꢁ 10ꢀ5 M modifier will compete with 0.2 M high
molecular weigh product and can remove their adsorbed forms
instantaneously from the Pt surface.
On the other hand our results indicate also that the poisoning
effect by organic residues originated form EtPy cannot be ne-
glected. In the experiments at [(EtPy)1]0 = 0.1 M presented in Table
1 the variation of the mode of introduction of reaction components
was investigated. In racemic hydrogenation higher reaction rates
were obtained upon injection than upon premixing of the substrate
in both sets of experiments (compare exps. 1 vs. 2 and 7 vs. 8.). This
behaviour can be attributed to the decomposition of EtPy in the
period of premixing. Data given in Table 1 show that in enantiose-
lective hydrogenation the relative rate increase expressed by ratios
r0e/r0r and k1e/k1r is quite substantial even at [(EtPy)1]0 = 0.1 M.
Even at this low concentration of substrate the conversion in-
creased instantaneously after the introduction of CD during race-
mic reaction (see experiments 6 and 12). This fact provided
further proof related to the origin of RE phenomena.
5. Conclusions
In this study it has been shown that the introduction of cincho-
nidine during racemic hydrogenation of different activated ketones
induces instantaneous rate increase. The extent of rate enhance-
ment depends on the type of substrate and experimental condition.
Pronounced rate enhancement was observed upon using both puri-
fied (EtPy)1 and (EtPy)2 containing 20 wt% high molecular weight
condensation product. Higher substrate concentration favours pro-
nounced rate acceleration, however, in the case of methylbenzoyl
formate higher reactions rates were observed in enantioselective
hydrogenation in the whole concentration range. The increase in
the extent of RE does not follow the ability of substrates to form
poisonous byproducts. The observed RE in Orito’s reaction may
be related to the formation of [substrate–modifier] complex in-
volved both in the rate acceleration and the enantiodifferentiation.
All results obtained in this study unambiguously demonstrated
that the RE is an intrinsic feature of Orito’s reaction and cannot
be attributed to poisoning effects. However, our results also show
that poisoning of the Pt is unavoidable during the hydrogenation of
activated ketones especially EtPy, which can decompose at the Pt
surface and form various condensation products in the liquid
phase. The ability of cinchona alkaloids or other tertiary amines
to clean the catalyst surface has only inferior contribution to the
rate enhancement.
Acknowledgments
It is interesting to mention that in enantioselective hydrogena-
tion in both sets of experiments the rate order was as follows:
CDinj > premix > (EtPy)1inj. This is an opposite trend compared to
results obtained in racemic hydrogenation. This observation might
have two possible explanations: (i) the lowest rate at (EtPy)1inj can
be attributed to the loss of CD over hydrogen covered pure Pt sites.
This side reaction has great importance only at low concentration
of CD [38]; the fact that the lowest ee values are obtained when
EtPy is injected (see experiments 5 and 11) strongly supports this
view; (ii) in addition, these results might also indicate that the CD
injected into the reaction mixture counterbalance the poisonous
effect induced by the decomposition of EtPy on the Pt surface.
Experiments with MBF substrate provided further evidence for
the intrinsic character of the RE phenomena. The extent of RE in-
The authors thank Engelhard Inc. and Degussa for providing Pt/
Al2O3 catalyst as a gift. MS measurements for Dr. Ágnes Gömöry
are greatly acknowledged.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jcat.2009.06.009.
References
[1] A. Baiker, H.U. Blaser, in: G. Ertl, H. Knözinger, J. Weitkamp (Eds.), Handbook of
Heterogeneous Catalysis, Wiley-VCH, Weinheim, 1997, p. 2422.
[2] M. Studer, H.U. Blaser, C. Exner, Adv. Synth. Catal. 345 (2003) 45. and
references cited therein.
[3] D.Yu. Murzin, P. Maki-Arvela, E. Toukoniitty, T. Salmi, Catal. Rev. Sci. Eng. 47
(2005) 175.
[4] M. Bartók, Curr. Org. Chem. 10 (2006) 1533.
[5] H.U. Blaser, H.P. Jalett, D.M. Monti, J.F. Reber, J.T. Wehrli, Stud. Surf. Sci. Catal.
41 (1988) 153.
creases upon increasing the concentration of MBF (compare r0e
/
r0r and k1e/k1r ratios in Table 2). Results shown in Fig. 5B unambig-
uously demonstrate that upon using MBF enantioselective and
racemic hydrogenations must have different r0 even at extremely
low initial concentration of this substrate. These results can be
considered as an additional proof that the RE is the intrinsic prop-
erty of Orito’s reaction and the attribution of the observed rate in-
crease to surface cleaning as the only reason for RE may not be
valid.
[6] A. Baiker, J. Mol. Catal. A. 115 (1997) 473.
[7] J.L. Margitfalvi, M. Hegedüs, J. Mol. Catal. A. 107 (1996) 281.
[8] K. Szöri, B. Török, K. Felföldi, M. Bartók, in: F.E. Ford (Ed.), Chemical Industries,
(Catalysis of Organic Reactions), Marcel Dekker, vol. 85, 2000, p. 489.
[9] R.L. Augustine, S.K. Tanielyan, L.K. Doyle, Tetrahedron Asymmetry 4 (1993)
1803.
[10] G. Bond, P.A. Meheux, A. Ibbotson, P.B. Wells, Catal Today 10 (1991) 371.
[11] M. Garland, H.U. Blaser, J. Am. Chem. Soc. 112 (1990) 7048.
[12] E. Jacobsen, I. Marko, M.B. France, J.S. Svendsen, K.B. Sharpless, J. Am. Chem.
Soc. 111 (1989) 737.
Results given in Table 2 also show a slight dependence of eemax
and eeend values of the initial concentration of the substrate. Sim-
ilar dependence was found for EtPy [34]. This finding and the same

198
E. Tálas et al. / Journal of Catalysis 266 (2009) 191–198
[13] L. Xing, F. Du, J.J. Liang, Y.S. Chen, Q.L. Zhou, J. Mol. Catal. A. 276 (2007) 191.
[14] D.J. Jenkins, A.M.S. Alabdulrahman, G.A. Attard, K.G. Griffin, P. Johnston, P.B.
Wells, J. Catal. 234 (2005) 230.
[26] E. Toukoniitty, D.Yu. Murzin, J. Catal. 251 (2007) 244.
[27] T. Mallat, A. Baiker, J. Catal. 251 (2007) 246.
[28] D.Yu. Murzin, J. Mol. Catal. A. 289 (2008) 91.
[15] E. Toukoniitty, D.Yu. Murzin, J. Catal. 241 (2006) 96.
[29] H.U. Blaser, H.P. Jalett, D.M. Monti, Appl. Catal. 52 (1989) 19.
[30] R. Hartung, J.G.E. Krauter, S. Seebald, P. Kukula, M. Studer, <http://
www.degussa-catalysts.com/asp/documents/79_Document.pdf>.
[31] J.L. Margitfalvi, E. Tálas, E. Tfirst, C.V. Kumar, A. Gergely, Appl. Catal. A 191
(2000) 177.
[16] D.M. Meier, D. Ferri, T. Mallat, A. Baiker, J. Catal. 248 (2007) 68.
}
[17] Gy. Szöllosi, Sz. Cserényi, F. Fülöp, M. Bartók, J. Catal. 260 (2008) 245.
[18] M. Bonello, J.F. Williams, A.K. Santra, R.M. Lambert, J. Phys. Chem. B 104 (2000)
9696.
[19] T. Mallát, Z. Bodnár, B. Minder, K. Borszéky, A. Baiker, J. Catal. 168 (1997) 183.
[20] D. Ferri, T. Bürgi, A. Baiker, J. Phys. Chem. B 108 (2004) 14384.
[21] M.V. Arx, N. Dummer, D.J. Willock, S.H. Taylor, R.P. K Wells, P.B. Wells, G.J.
Hutchings, Chem. Commun. (2003) 1926.
[32] J.L. Margitfalvi, E. Tfirst, J. Mol. Catal. A. 139 (1999) 81.
[33] D. Ferri, T. Bürgi, K. Borszeky, T. Mallat, A. Baiker, J. Catal. 193 (2000)
139.
[34] H.U. Blaser, H.P. Jalett, M. Garland, M. Studer, H. Thies, A. Wirth-Tijani, J. Catal.
173 (1998) 282.
[22] J.L. Margitfalvi, B. Minder, E. Tálas, L. Botz, A. Baiker, Sud. Surf. Sci. Catal. 75
(1993) 2471.
[35] B. Minder, T. Mallat, P. Skrabal, A. Baiker, Catal. Lett. 29 (1994) 115.
}
}
[23] J.L. Margitfalvi, M. Hegedus, J. Catal. 156 (1995) 175.
[36] J.L. Margitfalvi, M. Hegedus, E. Tfirst, Tetrahedron Asymmetry 7 (1996) 571.
[24] J.U. Köhler, J.S. Bradley, Langmuir 14 (1998) 2730.
[37] T. Mallat, A. Baiker, J. Catal. 176 (1998) 271.
[25] J.L. Margitfalvi, E. Tálas, E. Tfirst, Top. Catal. 39 (2006) 77.
[38] E. Tálas, J.L. Margitfalvi, Catal. Commun. 9 (2008) 984.