Kinetic investigations of product inhibition in the amino alcohol-catalyzed asymmetric alkylation of benzaldehyde with diethylzinc

Article abstract of DOI:10.1021/ol006181r

(equation presented) In situ reaction rate measurements help to define the role of product inhibition in the asymmetric alkylation of benzaldehyde with diethylzinc using (-)-MIB as a chiral reagent. Reaction calorimetry and kinetic modeling demonstrated that the rate behavior over consecutive reactions may only be rationalized when reversible binding of the product alkoxide is taken into consideration. These results may have implications for the conversion dependence of product enantioselctivity in reactions using enantioimpure catalysts.

Full text of DOI:10.1021/ol006181r

ORGANIC
LETTERS
2
000
Vol. 2, No. 16
511-2513
Kinetic Investigations of Product
Inhibition in the Amino
2
Alcohol-Catalyzed Asymmetric
Alkylation of Benzaldehyde with
Diethylzinc
†
†
‡
Thorsten Rosner, Patrick J. Sears, William A. Nugent, and
,†
Donna G. Blackmond*
Department of Chemistry, UniVersity of Hull, Hull HU6 7RX, United Kingdom,
and DuPont Pharmaceuticals Company, Chambers Works PRF S-1,
Deepwater, New Jersey 08023
d.g.blackmond@chem.hull.ac.uk
Received June 9, 2000
ABSTRACT
In situ reaction rate measurements help to define the role of product inhibition in the asymmetric alkylation of benzaldehyde with diethylzinc
using (−)-MIB as a chiral reagent. Reaction calorimetry and kinetic modeling demonstrated that the rate behavior over consecutive reactions
may only be rationalized when reversible binding of the product alkoxide is taken into consideration. These results may have implications for
the conversion dependence of product enantioselctivity in reactions using enantioimpure catalysts.
The enantioselective addition of organozinc reagents to
aldehydes using chiral amino alcohols has received wide
attention as a model system for exploring asymmetric C-C
demonstrates how the active catalyst concentration may be
affected by a subtle interplay of reaction parameters.
One of the most striking features of this reaction is the
strong asymmetric amplification in product enantioselectivity
that is observed when enantioimpure chiral amino alcohols
are employed in the reaction. Because of the complex
reaction network and the dimer/monomer catalyst equilib-
rium, nonlinear effects in this system are more complicated
1
bond formation. Extensive mechanistic studies by Noyori
2
and co-workers utilizing DAIB 1 as the amino alcohol ligand
3
have led to a good understanding of this complex reaction.
They have shown that the reaction proceeds via in situ
formation of a monomeric zinc aminoalkoxide catalyst
species which exists in equilibrium with a noncatalytically
active dimer (Scheme 1). Their mechanistic model also
than those predicted in the ML model introduced by Kagan
2
4
and co-workers for systems in which the active catalyst is
itself a dimeric species.
Noyori and co-workers showed that computer simulations
of the overall profile of this nonlinear effect were in
†
University of Hull.
DuPont Pharmaceuticals Company.
‡
(
1) Soai, K.; Niwa, S. Chem. ReV. 1992, 92, 833.
(2) (a) Noyori, R.; Kitamura, M. Angew. Chem., Int. Ed. Engl. 1991,
3
(
1
0, 49. (b) Yamakawa, M.; Noyori, R. J. Am. Chem. Soc. 1995, 117, 6327.
c) Kitamura, M.; Suga, S.; Oka, H.; Noyori, R. J. Am. Chem. Soc. 1998,
20, 9800.
(3) Oguni, N.; Matsuda, Y.; Kaneko, T. J. Am. Chem Soc. 1988, 110,
7877-88.
1
0.1021/ol006181r CCC: $19.00 © 2000 American Chemical Society
Published on Web 07/06/2000

(
yielding 0.05-0.07 M solution) into a 2 mL mixture of 1
M diethylzinc in toluene containing a catalytic amount of
-)-MIB at 25 °C. Catalyst amounts ranging from 15 to 30
Scheme 1. Reaction Network
(
µmol were employed. Heat flow data may be acquired at
the rate of 1 datum point/6 s, providing the equivalent of
3
00 separate initial rate measurements at different substrate
concentrations in a half-hour experiment. The reaction heat
flow also allows determination of the molar heat of reaction,
∆Hrxn, from the area under the heat flow curve divided by
the number of moles of substrate reacted. This was found to
be 44 ( 4 kcal/mol. Conversion of substrate may be obtained
from the partial heat flow at any point during the reaction.
Comparison of conversion obtained in this way with that
obtained from GC analysis of samples extracted from the
mixture verified that the calorimetric measurement provided
a valid measure of the rate of formation of the alkoxide
product. Benzyl alcohol was the only side product and was
observed in low concentrations (<1% of product). Enantio-
selectivity was found to be 96-98% for all reactions.7
The reaction heat flow rose rapidly to a maximum with
injection of benzaldehyde and then decreased smoothly with
time as the reaction proceeded to completion. The decreasing
rate as a function of time is attributed to the positive order
kinetics in both substrates but also to the presence of a
catalyst-bound product which builds up over the course of
the reaction and effectively removes active species from the
catalytic cycle.
2
c
agreement with experimental trends. Quantitative validation
of the kinetic model is hindered, however, by the difficulty
in obtaining extensive and quantitative experimental kinetic
data covering a wide range of substrate concentrations. For
example, a discrepancy between the predicted and observed
influence of conversion on product enantioselectivity was
attributed to product inhibition, but this could not be
confirmed experimentally.
When diethylzinc is used in excess, further reactions may
be monitored in the same reaction vial by injecting additional
aliquots of benzaldehyde. Each subsequent reaction is thus
carried out with increasing amounts of catalyst-bound product
present in the mixture. A four-part reaction of this type is
shown in Figure 1. By comparing the heat flow profiles from
the reaction of the first and subsequent aliquots of benz-
We report here in situ experimental kinetic studies of the
addition of diethylzinc to benzaldehyde using 2, (2S)-(-)-
3
-exo-(N-morpholino)isoborneol, (-)-MIB, which was re-
5
cently shown by Nugent to be a practical and efficient chiral
amino alcohol for the alkylation of aldehydes. We show that
the combination of detailed and accurate experimental
reaction rate measurement and kinetic modeling can suc-
cessfully account for the effects of the presence of the product
in reactions using enantiopure catalysts. In particular, a “one-
pot” monitoring of consecutive reactions allows us to
decouple product inhibition from the primary reaction
kinetics. These results may lead to a quantification of the
influence of the product alkoxide in the asymmetric ampli-
fication observed when the reaction is carried out using
enantioimpure chiral reagents. The importance of reaction
rate measurements in conjunction with enantioselectivity in
the use of nonlinear effects as a mechanistic tool was recently
highlighted by Blackmond.6
Reaction rate data for this study were obtained by
monitoring of the heat flow resulting from formation of the
alkoxide product as a function of time, which provides a
sensitive and accurate measure of reaction progress. The
reaction was carried out in an Omnical CRC90 reaction
calorimeter by injecting an aliquot of neat benzaldehyde
Figure 1. Reaction rate (from heat flow) as a function of time for
a four-part reaction in which subsequent aliquots of benzaldehyde
were injected into a reaction vial containing an excess of diethylzinc
(0.05-0.07 M benzaldehyde injected into ca. 1 M diethylzinc in
toluene, 0.014 M (-)-MIB).
2512
Org. Lett., Vol. 2, No. 16, 2000

aldehyde, we are able to observe the influence of a known
amount of product in the reaction mixture.
The magnitudes of the equilibrium constants determined
by the two models are also of interest. When the influence
To quantify the influence of the presence of product on
the reaction rate, we employed a kinetic model derived from
the mechanism shown in Scheme 1, based on that proposed
by Noyori. We have modified this model to allow for the
possibility that the product may not readily dissociate, thus
inhibiting further reaction by occupying catalytic sites. The
modified reaction rate expression is presented in eq 1. The
of product is not included (K
attempts to compensate by predicting a smaller value of
assoc. Further, the dimer-monomer equilibrium is predicted
p
set equal to zero), the model
K
to be driven further toward the dimer species (larger value
for the dimerization constant Kdimer) in the absence of product
inhibition. The dimer-monomer equilibrium position has
important implications for the asymmetric amplification
observed when enantioimpure catalysts are employed, and
an inaccurate value of Kdimer could lead to an incorrect
prediction of the equilibrium position for the heterochiral
dimer which is formed under enantioimpure conditions. Thus,
a failure to take the role of product into consideration may
lead to erroneous conclusions about the relative concentra-
tions of catalytic species in the reaction mixture.
In most cases one seeks to avoid complications due to the
influence of product by carrying out experimental kinetic
studies under initial reaction conditions. However, study of
the initial reaction period neglects valuable fundamental
information about the reaction which may be important in
practical applications where the reaction cannot be limited
to initial conditions. The monitoring of catalyst behavior over
the entire course of the reaction helps to broaden our
understanding of complex reaction networks. Coupled with
kinetic modeling, the multiple reaction experimental protocol
described in this paper provides a useful approach to
complicating issues such as product inhibition. The ability
to predict rate and selectivity as a function of reaction
progress greatly expands the synthetic utility of complex
reactions.
influence of product is accounted for in the term K
the denominator, where K represents the product binding
constant and [P] is the product concentration at any time
p
[P] in
p
during the reaction. This term approaches zero (K ) 0) if
p
product inhibition is considered to be negligible.
Table 1 presents the values of the kinetic and equilibrium
constants determined by fitting the four-reaction sequence
Table 1. Kinetic and Thermodynamic Parameters Determined
from Kinetic Modeling of the Data Shown in Figure 1 for the
Reaction Network Shown in Scheme 1
without product
inhibition
with product
inhibition
constant
krls (min-1
)
2820
45.4
1380
116
36.9
344
Kassoc (M-2)
Acknowledgment. D.G.B. gratefully acknowledges a
grant from Pfizer Central Research supporting this research
and a grant from the EPSRC for purchase of the reaction
calorimetry system.
KP (M-1)
Kdimer (M-2)
978
to eq 1. The model fits are shown in Figure 1, with the solid
line representing the model including product inhibition and
the dashed line representing the best fit which could be
obtained without including the influence of product in the
OL006181R
(4) Puchot, C.; Samuel, O.; Du n˜ ach, E.; Zhao, S.; Agami, C.; Kagan, H.
B.; Nonlinear Effects in Asymmetric Synthesis. Examples in Asymmetric
Oxidations and Aldolization Reactions. J. Am. Chem. Soc. 1986, 108, 2353-
5
7.
(
model (K ) 0). It is clear that the rate is indeed inhibited
P
5) MIB has been introduced as a crystalline and air-stable analogue of
as product builds up in the system; after ca. 10 turnovers,
the kinetic model can no longer provide an accurate
description of the reaction rate unless binding of the product
to the catalyst is taken into consideration.
DAIB. Samples of MIB used in this study were crystallized from hexanes
prior to use. Nugent, W. A. Chem. Commun. 1999, 1369.
(
6) (a) Blackmond, D. G., J. Am. Chem. Soc. 1997, 119, 12934-39. (b)
Blackmond, D. G. J. Am. Chem. Soc. 1998, 120, 13349-53
(7) Rosner, T. Doctoral Thesis, Ruhr-Universit a¨ t Bochum, 2000.
Org. Lett., Vol. 2, No. 16, 2000
2513