Improvement of the enantioselectivity in the enantioselective hydrogenation of ethyl pyruvate by addition of achiral tertiary amines

Article abstract of DOI:10.1039/a900681h

In the asymmetric hydrogenation of ethyl pyruvate over the catalyst system cinchonidine-Pt/AI₂O₃ the enantioselectivity was significantly improved by addition of different achiral tertiary amines to the reaction mixture, however the effect appeared to be strongly solvent and concentration dependent.

Full text of DOI:10.1039/a900681h

Improvement of the enantioselectivity in the enantioselective hydrogenation of
ethyl pyruvate by addition of achiral tertiary amines
J o´ zsef L. Margitfalvi, Emilia T a´ las and Mih a´ ly Heged uˆ s
Institute of Chemistry, Chemical Research Center, Hungarian Academy of Sciences, 1025 Budapest Pusztaszeri u´ t
59-67, Hungary. E-mail: joemarg@cric.chemres.hu
Received (in Liverpool, UK) 25th January 1999, Accepted 24th February 1999
In the asymmetric hydrogenation of ethyl pyruvate over the
catalyst system cinchonidine–Pt/Al the enantioselectiv-
ity was significantly improved by addition of different
achiral tertiary amines to the reaction mixture, however the
effect appeared to be strongly solvent and concentration
dependent.
result in any decrease in the rate of hydrogenation, moreover, as
emerges from the data given in Table 1, the addition of ACTAs
led to a considerable increase in both rate constants k and k
The rate acceleration increased in the following order: TEA <
DABCO < QN.
2 3
O
1
2
.
The enantioselectivity vs. conversion dependencies show a
monotonic increase type. This form of ee–conversion depend-
ency has been observed by different authors.2
,3,7–10
In the
The hydrogenation of a-keto esters and related compounds over
1
cinchonidine–Pt/Al
O
2 3
catalyst is of great scientific interest. In
absence of ACTAs the enantioselectivity vs. conversion
dependency passes through a maximum and decreases slightly
above 0.5 conversion giving a final ee value (eeend) around 0.71.
Similar behaviour had been observed in our earlier work when
these heterogeneous catalytic asymmetric hydrogenation reac-
tions the enantioselectivity (ee) (ee = ([R] 2 [S])/([R] + [S]))
appeared to be strongly solvent dependent.2 In different
solvents the ee increased in the following order: alcohols (ee =
2
5
3,8,11
[CD]
0
< 5 3 10 M.
Analogous results were obtained
0
=
.70–0.75) < hydrocarbons (ee = 0.80–0.85) < acetic acid (ee
when TEA was added, however in this case both the eemax and
eeend values increased quite substantially (see Table 1). The
addition of quinuclidine or DABCO altered the character of the
enantioselectivity vs. conversion dependency as the monotonic
increase in ee was maintained up to 0.99 conversion. This type
2
0.92–0.95%). In all solvents the ee depended strongly on the
1,2
concentration of the modifier. This effect was studied in
detail in ref. 2, however neither the chemistry nor the surface
phenomena responsible for the observed strong solvent effect
could be explained. One of the most interesting observations is
that in acetic acid (AcOH) ten times less cinchonidine is needed
to obtain the maximum ee value than in toluene.2
Table 1 Influence of different achiral tertiary amines on the enantioselective
hydrogenation of ethyl pyruvate in the presence of the cinchonidine–Pt/
We have also shown that the addition of small amounts of
Al O catalyst system
a
2
3
3
AcOH is sufficient to increase the ee value. For instance, in
1
/min21
EtOH at [AcOH] = 5 M the ee increased to 0.91 compared to
ACTA
—
k
1
/min2
k
2
eemax
0
.73 in its absence. In toluene, under similar conditions, the ee
0.0352
0.0407
0.0886
0.1289
0.1297
0.0465
0.0676
0.1588
0.1645
0.1757
0.750 (0.714)b
0.841 (0.793)
0.915b
increased to 0.93 compared to 0.81 in its absence. Conse-
quently, the above ee values were very close to those obtained
in pure AcOH. However, it should be mentioned that the use of
AcOH requires corrosion resistant stainless steel autoclaves and
accessories, this fact strongly hinders its general use in
academic research. Therefore, any new approach resulting in an
increase in enantioselectivity in the absence of AcOH will be of
great scientific and practical interest.
c
b
TEA
DABCO
c
b
QN
0.898
c
0.909b
QN
a
0
Solvent: toluene; T = 20 °C; hydrogen pressure = 50 bar; [Etpy] =
2
5
25
1
.0 M (batch No. 1), [CD]
0
= 1.2 3 10 M, [ACTA]
2 3
0
= 1.2 3 10 M.
Catalyst: 5 wt% Pt on Al
0.125 g. b Measured at the end of reaction.
amount
triethylamine, QN = quinuclidine.
O (Engelhard, E4759), dispersion = 22%,
c
=
TEA =
In this contribution we report that the addition of achiral
tertiary amines (ACTAs) to the reaction mixture containing
cinchonidine, substrate, solvent (e.g. toluene) and catalyst
significantly increases both the rate and the enantioselectivity of
the given asymmetric hydrogenation reaction. It is known that
when the hydrogenation of ethyl pyruvate is carried out in the
presence of ACTAs, e.g. triethylamine (TEA), quinuclidine
(
QN), pronounced rate acceleration is observed without asym-
4–6
metric induction. However, the rate acceleration was much
less than in the presence of cinchona alkaloids.
In this study the modifier, either cinchonidine alone or its
mixture with achiral tertiary amines, was injected by high
pressure hydrogen at t = 0 into the reactor containing the
catalyst, the substrate and toluene as solvent. Pseudo first order
rate constants describing the kinetics for the first 6–10 min (k
1
)
and 10–40 min (k ) were used to evaluate the effect of ACTAs.
2
Details of the experimental method and the kinetic approach can
be found elsewhere.3
,7,8
Reaction kinetic data and enantioselectivities obtained in a
series of experiments carried out at 20 °C either in the presence
or absence of ACTAs are summarized in Table 1, while Fig. 1
shows the corresponding enantioselectivity vs. conversion
dependencies. In these experiments the ACTAs:cinchonidine
Fig. 1 Enantioselectivity vs. conversion obtained in the presence of different
ACTAs. Experimental details and abbreviations are given in Table 1. (5)
No ACTA, (:) TEA, (X) DABCO, (-) and (8) QN.
(CD) ratio was five and almost full conversion (x < 0.99) was
achieved within 90 min. Thus, the addition of ACTAs did not
Chem. Commun., 1999, 645–646
645

View Article Online
Table 2 Influence of the concentration of quinuclidine (QN) on the
Table 3 Influence of the concentration of cinchonidine (CD) on the
enantioselective hydrogenation of ethyl pyruvate in the presence of the
enantioselective hydrogenation of ethyl pyruvate in the presence of the
cinchonidine–Pt/Al O
2 3
catalyst systema
a
cinchonidine–Pt/Al
O
2 3
catalyst system
/min2
1
k
/min21
eeb
Concentration
of [CD]/M
[QN]/M
k
1
2
k
1
/min2
1
k
2
/min21
eemaxb
0
1
6
1
6
1
.0
0.0254
0.0273
0.0601
0.0832
0.1267
0.1219
0.0065
0.0156
0.0932
0.1346
0.0371
n.m.c
0.775
0.782
0.910
0.926
0.936
0.946
2
6
6
5
5
4
.2 3 102
0.0
0.0050
0.0147
0.0273
0.0371
0.0823
0.1123
—
—
26
.0 3 10
1.4 3 10
0.0109
0.0271
0.0413
0.1270
0.216 (0.054)
0.560 (0.334)
0.646 (0.621)
0.868
2
26
.2 3 10
6.8 3 10
2
25
.0 3 10
1.2 3 10
.2 3 102
3.4 3 10
25
2
4
c
a
1.2 3 10
n.m.
0.870
Solvent: toluene; T = 10 °C; hydrogen pressure = 50 bar; amount of
catalyst = 0.125 g; [Etpy]
Measured at the end of reaction. c n.m. = not measurable.
2
5
a
0
= 1.0 M (batch No. 1); [CD]
0
= 1.2 3 10 M.
Solvent: toluene; T = 20 °C; hydrogen pressure = 50 bar; amount of
catalyst = 0.125 g; [Etpy]
of reaction. n.m. = not measurable.
b
= 1.0 M (batch No. 2). b Measured at the end
0
c
of enantioselectivity vs. conversion dependency was ob-
tained3
,8,11
when [CD]
> 5 3 10
25
M. There was no
0
1 2
similar changes in the reaction kinetics: (i) an increase of k , k
measurable difference between quinuclidine or DABCO, how-
ever, the phenomena was well reproducible (see Table 1 and
Fig. 1). In conclusion, results given in Table 1 and Fig. 1
precisely demonstrate that in the presence of ACTAs the final
enantioselectivity (eeend) increases from 0.71–0.72 to 0.91. This
increase in enantioselectivity is considered to be a very
pronounced selectivity improvement. Note that in toluene the
and eemax values and (ii) alteration of the form of the
enantioselectivity–conversion dependencies (see Fig. 1 in this
paper and Fig. 2 in ref. 11).
The absence of any effect of ACTAs both at high concentra-
tions of cinchonidine and in EtOH strongly resembles the
solute–solute (alkaloid–alkaloid) interactions observed in the
case of different dihydroquinines ((+)-DHQN and
highest ee value was 0.87 which was measured at [CD]
0
= 6 3
12
(
2)-DHQN). The (+)-alkaloid–(2)-alkaloid interaction was
2
4
10
M. Similar ee values were reported in ref. 2. As emerges
12
greatly reduced when an alcohol was used as solvent. We
propose that the alkaloid–alkaloid (cinchonidine–cinchonidine)
interaction is not favourable for the given catalytic reaction as it
reduces the amount of ‘free alkaloid’ required for asymmetric
induction. It should be mentioned that in crystallographic form
cinchonidine is stabilized by two hydrogen bonds between the
from data given in Fig. 1 the ability of ACTAs to increase the
enantioselectivity increased in the following order TEA <
DABCO = QN, i.e. the increase in ee is of a similar order as the
rate acceleration effect. However, no measurable effect was
induced by achiral tertiary amines when the concentration of
2
4
CD increased to 10 M.
13
OH group and the quinuclidine nitrogen, consequently
Analogous results were obtained in a series of experiments
carried out at 10 °C. In these experiments quinuclidine was used
as the ACTA and its concentration was varied. Related kinetic
data and enantioselectivities are summarized in Table 2, which
shows that the influence of ACTAs is strongly concentration
dependent. These experiments indicate that the addition of
quinuclidine (QN) in the given concentration range increases
cinchonidine exists in the form of a cyclic ‘dimer’. Based on the
above literature analogy we suggest that in the presence of
12
ACTAs a new type of solute–solute interaction (e.g., alkaloid–
ACTA interaction) appears provided the concentration of the
alkaloid is low and the solvent is not an alcohol. Due to the
above interaction the amount of ‘free alkaloid’ required for
asymmetric induction increases. Further studies are in progress
in our laboratory to elucidate the character of interactions
involved in the phenomena observed.
1 2
both the rate of the hydrogenation (rate constants k and k ) and
the enantioselectivity. In these experiments, due to the decrease
in the temperature, very high enantioselectivities were obtained.
Table 2 shows the influence of ACTAs is already very
pronounced at ACTA:CD = 0.5. The increase in the QN
The partial financial support of OTKA (grants T023317,
T025732) is greatly acknowledged.
concentration resulted in a further increase in the k
eemax values. The increase in the rate constant k and eemax is
gradual and levels off at ACTA:CD = 5–10, while the rate
constant k passes through a maximum. The enantioselectivities
obtained in this series of experiments (eemax = 0.93–0.94) are
the highest values that have ever been obtained in this reaction
in the absence of AcOH.
1 2
, k and
1
Notes and references
2
1
H. U. Blaser, H. P. Jalett, M. M u¨ ller and M. Studer, Catal. Today, 1997,
7, 441 and refs. cited therein.
3
2
3
4
H. U. Blaser, M. Garland and H. P. Jallett, J. Catal., 1993, 144, 569.
J. L. Margitfalvi and M. Heged uˆ s, J. Mol. Catal. A, 1996, 107, 281.
J. L. Margitfalvi and M. Heged uˆ s, J. Catal., 1995, 156, 175.
When EtOH was used as solvent no increase in ee was
2
5
24
observed in the 10 –10 M concentration range of CD and at
ACTA:CD = 1–5. No effect was observed in other alcohols,
such as methanol or propanol. All of these results strongly
indicate that the recognized effect induced by ACTAs depends
on (i) the type of solvent used, (ii) the concentration of achiral
tertiary amines, and (iii) the concentration of cinchonidine. Note
that the observed effect appears at a very low concentration of
5 H. U. Blaser, H. P. Jalett, D. M. Monti, J. F. Reber and J. T. Wehrli, Stud.
Surf. Sci Catal., 1988, 41, 153.
6 G. Bond, P. A. Meheux, A. Ibbotson and P. B. Wells, Catal. Today,
1
991, 10, 371.
7
J. L. Margitfalvi, B. Minder, E. T a´ las, L. Botz and A. Baiker, in New
Frontiers in Catalysis, ed. L. Guczi, F. Solymosi and P. T e´ t e´ nyi, Proc.
1
0th Int. Cong. Catal., Budapest, July 1992, Elsevier, Amsterdam, 1993,
p. 2471.
2
5
cinchonidine (1.2 3 10 M), i.e. in the concentration range
8
J. L. Margitfalvi, M. Heged uˆ s and E. Tfirst, Stud. Surf. Sci. Catal. (11th
International Congress on Catalysis), 1996, 101, 241.
characteristic of enzyme catalytic reactions.
The observed increase of both the rate of hydrogenation and
the enantioselectivity upon addition of ACTAs strongly re-
9 R. A. Augustine and S. K. Tanielyan, J. Mol. Catal. A: Chem., 1996,
112, 93.
10 J. Wang, Y. Sun, C. LeBlond, R. N. Landau and D. G. Blackmond,
J. Catal., 1996, 161, 752.
sembles the influence of the initial concentration of cinchoni-
dine on the kinetics and enantioselectivity.2
,3,8,11
Results
1
1 J. L. Margitfalvi, E. Tfirst, M. Heged uˆ s and E. T a´ las, Catalysis of
Organic Reactions, ed. Frank E. Herkes, Marcel Dekker, New York,
obtained in this study indicated that addition of ACTAs would
increase the amount of cinchonidine involved in the enantio-
selective hydrogenation. The above suggestion is strongly
supported by results attained at difference concentrations of
CD, but in the absence of ACTA. These results are summarized
in Table 3. As emerges from the data in Tables 2 and 3, the
increase in the concentration of both CD and ACTA resulted in
1
998, vol. 75, p. 531.
1
2 T. Williams, R. G. Pitcher, P. Bommer, J. Gutzwiller and M. Uskovic,
J. Am. Chem. Soc., 1969, 91, 1870.
1
3 B. J. Oleksyn, Acta Crystallogr., Sect. B, 1982, 38, 1832.
Communication 9/00681H
646
Chem. Commun., 1999, 645–646