Asymmetric lewis acid catalysis in water: α-amino acids as effective ligands in aqueous biphasic catalytic michael additions

Article abstract of DOI:10.1002/ejoc.200800922

This article explores the potential of native α-amino acids as chiral ligands in aqueous asymmetric Lewis acid catalysis, employing the C-C bond forming Michael addition as a model reaction. Some insights are provided regarding the details of Yb(OTf)

Full text of DOI:10.1002/ejoc.200800922

FULL PAPER
DOI: 10.1002/ejoc.200800922
Asymmetric Lewis Acid Catalysis in Water: α-Amino Acids as Effective
Ligands in Aqueous Biphasic Catalytic Michael Additions
Karolina Aplander,[a] Rui Ding,[a] Mikhail Krasavin,[b] U. Marcus Lindström,*[b] and
Johan Wennerberg*[c]
Keywords: Homogeneous catalysis / Asymmetric catalysis / Water-tolerant Lewis acids / Michael addition / Recyclable
catalyst / Reaction mechanisms / Amino acids
This article explores the potential of native α-amino acids as
chiral ligands in aqueous asymmetric Lewis acid catalysis,
employing the C–C bond forming Michael addition as a
model reaction. Some insights are provided regarding the
applicable to a wider range of donors and acceptors than pre-
viously demonstrated. Importantly, it was also demonstrated
that the α-amino acid complexed ytterbium catalyst might
have potential for large-scale applications as it displays not
details of Yb(OTf)3/α-amino acid-catalyzed Michael ad- only large ligand accelerations, but also good solubility and
ditions in water through new kinetic data as well as studies
on how both yield and selectivity are influenced by variations
in metal/ligand ratio, pH, temperature, and structure of the
α-amino acid. Through this investigation it was found that
reaction conditions that require only 5 mol-% of the Lewis
acid, provides enantiomeric excesses of up to 79% and is
stability in water. It can be recycled multiple times without
appreciable loss of activity. The result is a promising example
of a water-compatible chiral Lewis acid.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2009)
Introduction
has been made. A recent review by Li on C–C bond-form-
ing reactions in aqueous media, many of which are metal-
mediated, listed nearly 1000 references.[9]
The use of homogeneous metal-based catalysts to pro-
mote chemical reactions in terms of both rate and selecti-
vity is of great importance in organic synthesis. For exam-
ple, complexation of transition metals with chiral ligands
has been a successful strategy to achieve various catalytic
enantioselective reactions that have found wide application
in fine chemicals production.[1–3] The vast majority of these
methods were developed for use under anhydrous condi-
tions because of the notorious lability of organometallics
towards water. In recent years, however, the surge of interest
in the use of water as an economical and environmentally
benign reaction medium for organic reactions[4–8] has called
for the development of transition-metal catalysts that dis-
play solubility, stability, and efficacy in water. It is clear that
for the full potential of water as reaction solvent to be real-
ized aqueous metal catalysts are required that can at least
match the efficacy of traditional metal catalysts developed
for non-aqueous media. Accordingly, research efforts in this
area have intensified in the last decade and some progress
Lewis acid catalysis is one of the most useful strategies
for enhancing the rate and selectivity of a synthetic reaction,
so that it may be carried out under milder conditions and
with fewer by-products. In addition, the coordination prop-
erties of Lewis acidic metals make them suitable for use in
asymmetric catalysis through the complexation of chiral, ba-
sic ligands to the central metal ion. A significant step for-
ward in the area of aqueous synthesis and catalysis would
be the development of Lewis acid catalysts that present high
stability, activity, and selectivity in an aqueous environment.
However, making the solvent switch to water for Lewis acid
catalyzed reactions is a formidable challenge. The first re-
ports on the use of water-compatible Lewis acids, employing
lanthanide triflates, were reported by Kobayashi and co-
workers in 1991.[10] It is, however, only in the most recent
decade that the rational design of water-compatible Lewis
acid catalysts has been properly addressed. The seminal
work is a paper by Kobayashi and co-workers from 1998
where a large number of metal salts were tested for their
efficiency in promoting aqueous aldol reactions.[11] A basic
understanding of which characteristics of a metal salt that
are important for Lewis acid activity in water, such as having
a hydrolysis constant (Kh) within a certain range and a high
water-exchange rate constant (WERC), crystallized from
this study. The first enantioselective Lewis acid catalyzed re-
action in water was disclosed as late as in 1998.[12,13]
[a] Division of Organic Chemistry, Lund University,
P. O. Box 124, 22100 Lund, Sweden
[b] Department of Chemistry, McGill University,
Montréal, Quebec, Canada H3A 2K6
[c] DuPont Chemoswed,
P. O. Box 839, 20180 Malmö, Sweden
Fax: +46-40-186805
E-mail: johan.wennerberg@swe.dupont.com
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
810
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 810–821

Asymmetric Lewis Acid Catalysis in Water
A few years ago, we began an investigation aimed at de- for their ability to catalyze the Michael addition reaction
veloping more efficient Lewis acid catalysts for synthesis in between ethyl acetoacetate, 1, and methyl vinyl ketone
water. The aim was to study the influence of various ligands (MVK) to give 2 using TMEDA as ligand (Scheme 1).
on the catalytic activity of various metal salts in water.
Having established the potential of Yb(OTf)3 in ligand-
While ligand complexation to a Lewis acid metal in organic assisted aqueous Michael additions a study of the effect of
media is commonly associated with reaction rate decelera- α-amino acids as ligands in these reactions were carried out.
tion, we reasoned that the aqueous medium might be par- With respect to their natural abundance and role in biocata-
ticularly amenable to ligand-accelerated catalysis.[14] As- lysis, as integral parts of enzymes, the use of native α-amino
suming that tightly coordinated water ligands become more acids as chiral ligands in catalysis has been relatively little
labile upon complexation of an organic ligand, this should explored.[26] In recent years, α-amino acids have been
lead to a higher rate of exchange with substrate molecules studied as organocatalysts and applied as such in a variety
and consequently to higher reaction rates. By the same ra- of reactions in organic as well as aqueous media.[27] On
tionale, ligand complexation in water could also lead to the other hand, we find it surprising that metal-amino acid
more stable catalysts by reducing the rate of decomposition complexes, although physically well-characterized through
through hydrolysis. A simple illustration of this concept is a number of studies,[28] have found so little application in
outlined in Figure 1. Indeed, positive ligand effects such as the area of organic synthesis. A reason for this may be the
increased hydrolytic stability[15] and rate accelerations[16] are limited solubility of such complexes in the organic solvents
commonly noted in papers on aqueous Lewis acid catalysis, traditionally used for synthetic applications.
although quantitative data have rarely been given.
In a recent communication, we disclosed our initial re-
sults on the use of α-amino acids as ligands in aqueous
biphasic Lewis acid catalyzed Michael additions.[29] Re-
markably, in the presence of -alanine, the second-order
rate constant for the ligand-assisted reaction between 1 and
MVK was increased by a factor of 138 compared to the
ligand-free reaction. That is seven times the ligand effect we
observed for the same reaction when TMEDA was used as
ligand. It is to the best of our knowledge one of the strong-
est ligand effects reported for any Lewis acid catalyzed reac-
tion. In addition, the initial enantioselectivities we ob-
tained, albeit modest (up to 53%) were the highest observed
for any reaction using natural α-amino acids as Lewis acid
ligands in water. Herein we provide further insight into the
details of Yb(OTf)3/α-amino acid-catalyzed Michael ad-
ditions in water through additional kinetic data as well as
studies on how both yield and selectivity are influenced by
variations in pH, temperature, and type of α-amino acid.
Figure 1. Schematic illustration of the potential influence of a li-
gand on the activity and hydrolytic stability of a Lewis acidic metal
in aqueous media.
As a model reaction for studying ligand effects in water
the Michael addition reaction, which is one of the most
important carbon–carbon bond-forming reactions in or-
ganic synthesis, was chosen. While a great deal of success
has been achieved with catalytic Michael additions in or-
ganic solvents,[17–19] progress towards efficient Lewis acid
catalyzed Michael additions in water has been slow. Until
very recently aqueous catalytic Michael additions required
long reaction times and a large excess of acceptor in order
to get high yields of adduct.[20–23] Recent significant ad-
Results and Discussion
Determination of Reaction Order and Rate-Determining
Step
We previously established the reaction between 1 and
vances have come from Kaneda and co-workers on solid MVK catalyzed by 10 mol-% Yb(OTf)3 and 12 mol-% -
phase Lewis acid catalysis[24] and from our own research. alanine as following typical second-order kinetics for the
In 2006, we reported our first results on ligand-accelerated first 2 h. Thus, we used the initial (2 h) rate constants to
aqueous biphasic catalysis of C–C bond forming Michael determine the rate difference between the ligand-assisted
additions.[25] In this study we screened some metal triflates and ligand-free reaction. This was reasonable because after
Scheme 1. Ligand acceleration in Lewis acid catalyzed aqueous biphasic Michael addition.
Eur. J. Org. Chem. 2009, 810–821
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
811

U. M. Lindström, J. Wennerberg et al.
FULL PAPER
2 h we already had 57% conversion of starting material into Lewis Acid Catalyzed vs. Organocatalytic Mechanism
adduct 2. In the present study, to determine the reaction
Considering recent work on metal-free reactions using
order with respect to each of the reactants, we performed
amino acids as organocatalysts,[27] any suspicion that an or-
the reaction under pseudo-first-order conditions for the nu-
ganocatalytic enamine-based mechanism was a major con-
cleophiles, 1, and MVK. As can be seen in Figure 2, when
tributor to the results was investigated. To determine this
plotting ln1/[1] or ln1/[MVK] against the reaction time
some experiments under “enforced” organocatalytic condi-
straight lines typical of first-order rates are obtained.
tions were performed. Enamine adduct 3 (Dane’s salt) of
acetylacetone (4), and alanine were prepared according to
a literature procedure.[31] Subjecting 3 to a reaction with the
Michael acceptor 2-cyclohexen-1-one (5) (Scheme 2), in the
absence or presence of a Brønsted acid, afforded poor yields
(23% and 33%) and ee values (7% and 11%) of the adduct
6. These results, together with other supporting observa-
tions throughout our studies (see below) made us confident
that an organocatalytic enamine mechanism was not signifi-
cantly involved in our reactions.
Scheme 2. Aqueous Michael addition under “enforced” organocat-
alytic conditions.
Figure 2. Graph illustrating first-order behavior with respect to
both donor and acceptor.
Proposed Mechanism/Influence of the Basic Amino Group
on Rate and Selectivity
The large difference in observed rates, kobsd.[1]/
kobsd.[MVK] = 14, did not allow us to make an unambigu-
To successfully arrive at more efficient reaction condi-
ous assignment of these experiments as first order. Being tions an improved understanding of the large ligand effect
run at the same concentrations we had expected the two and the significant level of stereocontrol that we observe
experiments to have similar rates if they were of true first was required. On the basis of our own observations and
order, but a large excess of donor (1/MVK, 10:1) had a extensive literature precedence,[22,30,32] a tentative mecha-
relatively modest effect on the overall reaction rate, while nism for the reaction (Scheme 3) is proposed. It suggests
an excess of acceptor (MVK/1, 10:1) had a much greater that when Yb(OTf)3 is dissolved in aqueous medium, water
influence on the reaction rate. When MVK was used in will become the major ligand and the counterion (OTf–)
large excess, the first step of the reaction (formation of a will exist in solution. Then the deprotonated amino acid
metal-diketonato complex) became saturated, meaning that first displaces one of the water ligands to give A. These
all activated 1 reacted instantly with MVK and the reaction types of 1:2 complexes between lanthanides and α-amino
order of 1 went from first order towards zero order. The acids in water are well studied and have been subjected to
fact that kobsd.[1] is much larger than kobsd.[MVK] indicates extensive review.[28] Upon addition of the reactants another
that the rate-limiting step is not formation of the metal- water ligand is displaced by a bidentate β-keto enolate to
diketonato complex, but the step involving formation of the give complex B, which is planar and may be stabilized by
new C–C bond. Recent mechanistic studies on transition- π-delocalization. Through ligand exchange with water, the
metal-catalyzed Michael additions of 1,3-dicarbonyl com- enone will then coordinate to a vacant site on the metal to
pounds to enones in organic solvents have shown initial re- form the reactive complex C. Subsequent bond formation
action rates to be almost independent of the substrate con- affords the metal-Michael adduct complex, D, and finally
centrations (zero order in both donor and acceptor).[30] ligand exchange with a new donor molecule releases the
This difference in kinetic behavior between aqueous and or- product.
ganic media may be ascribed to the fact that metal com-
We believe that the origin of the enantioselectivity and
plexes are more labile in water and thus more liable to take the large rate acceleration is the effective coordination of a
part in concentration-dependent equilibria. For example, as chiral dibasic ligand (the deprotonated amino acid) to the
shown by Kobayashi,[11] the activation of a carbonyl oxygen Lewis acidic metal. It has been shown that negatively
by a metal in aqueous media is limited by the concentra- charged ligands can significantly enhance the reactivity of
tion-dependent rate of exchange with coordinated water li- a metal bound β-keto enolate towards electrophiles, pre-
gands on the metal.
sumably by weakening the covalent character of its M–O
812
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 810–821

Asymmetric Lewis Acid Catalysis in Water
Scheme 3. Proposed mechanism for the Yb(OTf)3/α-amino acid catalyzed aqueous Michael addition.
bond.[30a] With common counterion ligands (e.g. Cl–) this
leads to fast protonation of the enolate by the acidic reac-
tion mixture and a decrease in reaction rate. In the present
system, however, the dibasic character of the α-amino acid
ligand may give a kinetic advantage to the reaction between
the β-keto enolate and a coordinated electrophilic Michael
acceptor. According to this mechanistic proposal, we
should find the reaction to be highly sensitive to variations
in pH. To obtain support for this hypothesis we performed
two lines of experiments. First, we monitored the depen-
dence of enantioselectivity on the pH of the Yb(OTf)3/-
alanine-catalyzed reaction between acetylacetone (4), and
2-cyclohexen-1-one (5). Second, we varied the pKa of the
amino group by adding one or two N-methyl groups to -
alanine and observed the impact of these changes on both
yield and ee of 6. In addition, to include an example of a
monobasic ligand, an experiment was performed with -
Figure 3. Graph illustrating the observed influence of pH on the
enantiomeric excess of Michael adduct (R)-6.
lactic acid in place of -alanine. We found that when using
alanine as ligand, the ee of 6 varied significantly with the
pH of the catalyst solution, but interestingly the optimal
pH range within which to obtain the best enantioselectivity change in the reaction mechanism with a change in pH.
was quite narrow (Figure 3). The results from the Michael reaction between 4 and 5 with
At pH 6.7, an ee of 38% was observed while at pH 6.1 structurally varied ligands are collected in Table 1.
the ee dropped to 26% and at pH 5.8 it was only 17%. As expected, the reaction with deprotonated lactic acid
Above pH 7, the ee plunged to low levels (pH 7.7 = 10% (Entry 1) afforded a much lower conversion (12%) and ee
ee, pH 8.5 = 5% ee), most likely due to increased impor- (6%) of 6 than the reference reaction with -alanine (Entry
tance of the base-catalyzed pathway that leads to racemic 2, 93% yield, 57% ee), clearly showing the importance of
product. The rapid drop in ee below pH 4 can be ascribed the basic amino group for the rate effects and selectivities
to increased protonation of the carboxylate ion leading to observed. Somewhat surprisingly, however, the reactions
weaker metal ion complexation. We argue that the optimal with the ligands N-methyl--alanine (Entry 3), or N,N-di-
pH around 6.7 is high enough for nitrogen–metal coordina- methyl--alanine (Entry 4), also led to much lower yields
tion to effectively compete with N-protonation, but low (20% and 18%, respectively) and ee values (19% and 5%,
enough to avoid the non-selective, base-catalyzed pathway. respectively). This indicated that, while the basicity of the
Another prominent feature of the ee vs. pH graph in Fig- amino group is important, making the nitrogen too basic
ure 3 is that two maxima are observed. This suggests a might have a negative influence on both rate and selectivity.
Eur. J. Org. Chem. 2009, 810–821
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
813

U. M. Lindström, J. Wennerberg et al.
FULL PAPER
Table 1. Effects of varying the structure of the chiral ligand.
proved much and even falls again at a ligand/metal ratio of
3.0. The trend for the enantiomeric excess follows the trend
in yield up to a ligand/metal ratio of 1.2.
Entry Ligand
Yield (%)[a]
ee (%)[b]
1 equiv. NaOH
1
2
3
4
-lactic acid
-alanine
N-methyl--alanine
N,N-dimethyl--alanine
12
93
20
18
6
57
19
5
Figure 4. Graph demonstrating the influence of ligand concentra-
tion on yield and ee with 10 mol-% Yb(OTf)3. Reactions were per-
formed at 60 °C for 18 h with 4 (1 mmol), 5 (3 mmol), -alanine
(0.02–0.30 mmol) and Yb(OTf)3 (0.10 mol) in water (1 mL) at
pH 6.70.
pH-control (6.70)
5
6
7
-alanine
N-methyl--alanine
N,N-dimethyl--alanine
93
25
31
57
21
25
No significant changes in enantioselectivities were ob-
served in the reactions with higher ligand/metal ratios. As
one of our goals of the present investigation was to signifi-
cantly reduce the amount of Lewis acid required, we contin-
ued by studying the reaction between 4 and 5 in the pres-
ence of only 5 mol-% of Yb(OTf)3 and 12 mol-% ligand
(Table 2).
[a] Not isolated. Determined by NMR spectroscopy. [b] Deter-
mined by chiral HPLC.
In an attempt to evaluate the steric influence of the extra
N-methyl groups on rates and selectivities, we performed
additional experiments where the pH was adjusted (pH =
6.70) to compensate for differences in the pKa of the amino
groups (Entries 6 and 7). Slightly higher yields (25% and
31% for N-methyl- and N,N-dimethyl--alanine, respec-
tively) were observed compared to the reactions without
pH-adjustment. The ee values also increased after adjusting
the pH to 6.70. As expected, the effect was significantly
larger with N,N-dimethyl--alanine. The results obtained
after pH-adjustment suggested that the electronic influence
of the basic nitrogen was largely offset by steric effects im-
posed by the methyl groups.
Table 2. Influence of metal/ligand ratio on yield and ee in the reac-
tion between 4 and 5 to give (R)-6.[a]
Entry
Yb(OTf)3 [mol-%] -Ala [mol-%] T [°C] Yield[b] [%] ee[c] [%]
1
2
3
4
5
6
7
8
10
5
5
5
5
5
5
5
12
12
10
11
12
13
14
15
60
60
70
70
70
70
70
70
93
46
98
57
57
60
63
63
63
64
64
97
100
100
100
100
[a] All reactions were run twice and listed values are averages. Reac-
tions were performed with 4 (1 mmol), 5 (3 mmol), in water (1 mL)
at pH 6.70 for 18 h. [b] Not isolated. Determined by NMR spec-
troscopy. [c] Determined by chiral HPLC.
Effect of Ligand/Metal Ratio on Yield and
Enantioselectivity
Initially, we found that there was a significant drop in
yield from 93% to 46% upon reducing the loading of Lewis
In order to get a clearer picture of how yields and selec- acid from 10 mol-% to 5 mol-% (Entries 1 and 2). The
tivities were influenced by variations in ligand/metal ratio, enantioselectivity, however, did not change. In order to
we performed the reaction between 4 and 5 with varying compensate for the reduced yield we increased the reaction
amounts of -alanine and Yb(OTf)3. Because there is a temperature to 70 °C. Gratifyingly, under otherwise iden-
good correlation between yield and conversion, the latter tical conditions, we now observed quantitative conversion
was used to analyze the reactions. First, we ran the reac- of 4 into 6 (Entry 5). Then, by varying the amount of ligand
tions with 2–30 mol-% of -alanine while keeping the from 10–15 mol-% (ligand/metal ratio 2–3, Entries 3–8) we
amount of Yb(OTf)3 constant at 10 mol-%. These results were able to confirm that a ligand/metal ratio of 2.4 that
are illustrated in Figure 4. As can be seen, the conversion we had arrived at earlier (see above) was optimal in the
of the Michael adduct 6 increases steadily up to a ligand/ range studied. Higher ratios did not give any significant im-
metal ratio of about 2.4, after which the yield is not im- provement (Entries 6–8). The ratio may be explained due
814
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 810–821

Asymmetric Lewis Acid Catalysis in Water
to the dynamic character of the catalytic process i.e. the Table 4. Influence of α-amino acid structure and temperature on
enantioselectivity in the Michael addition reaction between 4 and
excess of ligand is necessary to keep the system saturated
and thus more efficient. Interestingly, slightly higher ee val-
ues were obtained in the reactions run at 70 °C than in the
reactions run at 60 °C. This unusual reversed temperature
effect will be addressed further below.
5.[a]
Entry
Amino acid
Yield (%)[b]
ee (%, S)[c]
1
2
3
4
5
6
7
-Ala
-Val
-Ile
-Phe
-Pro
-Trp
-Met
46
72
78
97
60
100
96
57 (R)
63
61
50
70 (R)
56
43
Effect of the Reaction Temperature on Yield and Selectivity
In order to investigate how the ee values and yields were
changed with reaction time, eight Lewis acid catalyzed reac-
tions between 4 and 5 were performed (Table 3).
average: 78
average: 57
8
9
10
11
12
13
-Gln
-Ser
-Asn
-Cys
-Tyr
-Thr
67
53
79
24
49
46
35
24
39
17
Table 3. Investigation of ee values and yields with reaction time of
the Yb(OTf)3/-alanine-catalyzed reaction between 4 and 5.[a]
46
36
average: 51
61
average: 35
14
15
-Asp
-Glu
8
83
14
average: 11
27
average: 72
94
16
17
-Lys
-His
100
20
average: 97
average: 24
Time [h]
Yield (%)
ee (%)
[a] All reactions were run at least twice and listed values are
averages. Reactions were performed at 60 °C for 24 h with 4
(1 mmol), 5 (3 mmol), amino acid (0.12 mmol) and Yb(OTf)3
(0.05 mol) in water (1 mL) at pH 6.70. [b] Not isolated. Determined
by NMR spectroscopy. [c] Determined by chiral HPLC.
1
2
4
8
12
16
20
24
8
47
49
47
51
45
50
49
52
14
23
30
40
52
67
92
Somewhat surprisingly, considering that alanine is the
sterically least demanding of all the chiral α-amino acids,
only three α-amino acids were able to induce a higher ee
than alanine in this reaction, namely valine (63%, Entry
2), isoleucine (61%, Entry 3) and proline (70%, Entry 5).
Unfortunately, none of these also provided high conversion.
It is important to note that in this study, the use of -amino
acids exclusively led to (S)-6 and that the use of -amino
acids afforded the enantiomeric (R)-6. While the yields and
selectivities observed in this study vary significantly among
all of the ligands tested, it is very interesting to note that if
the α-amino acids are listed according to the character of
their side-chain, one finds distinct groupings in terms of
yields and ee values. As for the yields, the following ranking
of side-chain character is observed, with average yields for
the groups in parenthesis: basic (97%)Ͼnon-polar
(78%)Ͼacidic (72%)Ͼpolar (51%). In terms of enantio-
selectivities, the order is different (average ee in parenthe-
sis): non-polar (57%) Ͼ polar (35%) Ͼ basic (Ͼ24%) acidic
(11%) side-chains. Correlating α-amino acid structure with
reactivity/selectivity of the catalyst will be important in the
design of synthetic amino acid derivatives with optimized
ligand properties.
[a] All experiments were performed twice and reported values are
averages.
The data in Table 3 demonstrates that the enantiomeric
excesses were not changed to any larger extent during the
reaction. Even after 1 h when only 8% conversion was
achieved the ee was 47% compared to 52% ee after 24 h
and almost complete conversion (92%). This confirmed
that the enantiomeric excess was obtained immediately and
no racemization occurred during the reaction.
Influence of Various α-Amino Acids on Yield and
Enantioselectivity
While we have previously screened a number of α-amino
acids for their ability to induce asymmetry in the studied
reaction, these first experiments used 20 mol-% Yb(OTf)3
and 24 mol-% of amino acid.[12] In our current investigation
we decided to screen a much wider selection of α-amino
acids for their efficacy as chiral ligands, and to employ our
optimized catalytic conditions 5 mol-% Yb(OTf)3 and
12 mol-% α-amino acid.
The results are collected in Table 4. Gratifyingly, out of
17 reactions only two, those with threonine and cysteine,
afforded yields (36% and 24%, respectively, Entries 11 and
13) that were lower than observed in the reaction with ala-
Influence of Temperature on Enantioselectivity
That the enantioselectivity so clearly changes with the
nine (46%, Entry 1). Two of the reactions, using tryptophan polarity of the amino acid side-chain serves as an indication
and histidine, even gave complete conversion of starting that solvation effects are important in the enantiodifferenti-
material into adduct (Entries 6 and 17).
ating step of the reaction. If this is true, the Gibbs acti-
Eur. J. Org. Chem. 2009, 810–821
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
815

U. M. Lindström, J. Wennerberg et al.
FULL PAPER
vation energy for this step is expected to have a large en- data show that the diastereomeric transition state that leads
tropic contribution because solvation effects are usually to the major enantiomer has a lower Gibbs’ free energy of
considered to be entropy-driven.[33] To find support for this, activation because of an unusually large entropic factor.
we returned to the discovery that the enantioselectivity in- This indicates that a difference in solvation of the dia-
creases with temperature. Reversed temperature effects in stereomeric transition states is a crucial factor in determin-
enantioselective reactions, while unusual, are not entirely ing the stereoselectivity of this reaction. Further investiga-
unknown, and have been explained by analyzing the en- tion of the reversed temperature effect will be reported in
thalpy and entropy factors of the reaction [Equation due course.
(1)].[34–37] Given that the observed effects are linear, the
Drawing from the positive results obtained by increasing
differential Eyring equation [Equation (2)] can be used to the temperature in the reaction using -alanine as ligand,
obtain the differential activation enthalpy (∆∆H‡R–S) and we then proceeded to test the reaction with -proline as
entropy (∆∆S‡R–S) values.[38]
ligand at 90 °C as that had provided the highest ee (70%)
at 60 °C. Fortunately, again a significant temperature effect
was observed and (R)-6 was obtained in quantitative yield
and 79% ee after 10 h at 90 °C. As far as we know, this is
the highest ee value reported for a metal-catalyzed synthetic
reaction that uses a native α-amino acid as ligand. As ex-
pected for enantiomeric ligands of equal optical purity, re-
placing -proline with -proline afforded the opposite en-
antiomer (S)-6 with the same ee (79%).
ln(kR/kS) = ln[(100 + % ee)/(100 – % ee)]
(1)
(2)
ln(kR/kS) = ∆∆S‡R–S/R – ∆∆H‡R–S/r.t.
Figure 5 shows the Eyring plot, i.e., the temperature ef-
fect on enantioselectivity, for the 5 mol-% Yb(OTf)3/
12 mol-% -alanine-catalyzed reaction between 4 and 5. As
can be seen, the ee increases in a linear fashion from 25%
ee at 30 °C up to 67% ee at 90 °C, a difference of 42% ee!
Yb(OTf)3/L-Proline as Catalyst in Aqueous Michael
Addition Reactions Between Various Donors and Acceptors
Having arrived at the conclusion that a combination of
5 mol-% Yb(OTf)3 and 12 mol-% -proline provides both
high yield and enantioselectivity, we proceeded to apply
these conditions in reactions between various donors and
acceptors. These results are listed in Table 5. Michael ac-
ceptors with 1,2-substituted double bonds are more chal-
lenging acceptors than 1,1- or 1-substituted acceptors. Until
our recent discovery of the Yb(OTf)3/amino acid catalysts
no examples of this type had been reported using homogen-
eous Lewis acid catalysis in water. Under the present condi-
tions mostly good yields were obtained with such acceptors
in acceptable reaction times. We were also able to decrease
Figure 5. Eyring-plot for the Yb(OTf)3/-alanine-catalyzed
Michael reaction between 4 and 5.
Considering the known susceptibility of many Lewis the amount of acceptor needed. 1.1 Equivalents were suf-
acidic transition metals to hydrolysis in water, it is an unex- ficient although the reaction is faster with 3 equiv. of the
pected, and quite remarkable, observation that the optimal acceptor. This was also evidenced in the kinetic studies. In
reaction temperature for a Lewis acid catalyzed reaction in the reaction between ethyl acetoacetate (1), and 2-cy-
water is close to the boiling point of water! Using the data clohexen-1-one (5), an isolated yield of 77% of the expected
shown in Figure 5 it was then possible to extract the adduct with an ee of 71% was obtained after 12 h (Entry
differential activation enthalpy (∆∆H‡R–S) and entropy 1). Higher yields may be obtained if the reaction times are
(∆∆S‡R–S) values from Equation (2). Not surprisingly, we extended, such as in the addition of acetylacetone (4) to 5,
found
a
large entropy contribution (∆∆S‡
=
which provided the Michael adduct in both high yield
R–S
59.7 JK–1 mol–1) to the differential Gibbs energy of acti- (94%) and ee (79%) (Entry 2). The addition of acetyl-
vation, which serves to explain the selectivity enhancement acetone to 2-cyclopenten-1-one (7), gave a moderate yield
with increasing temperature. The differential enthalpy term (61%) and ee (49%) (Entry 3).
was found to be of the same sign (∆∆H‡
=
Benzalacetone (8), is a synthetically useful, but less reac-
R–S
16.5 kJmol–1). These data provided some preliminary in- tive Michael acceptor. Under optimized conditions it
sight into the origin of the relationship between enantio- underwent a slow Michael addition with acetylacetone to
selectivity and reaction temperature. Because a linear rela- afford the corresponding adduct in 70% yield and 19% ee
tionship was found over the temperature range studied, the after 24 h, along with recovered starting material (Entry 4).
enantioselectivities observed are likely the results of a single Nitroalkenes may also act as acceptors. β-Nitrostyrene (9),
enantiodifferentiating mechanism. Assuming that the two reacted with acetylacetone to afford the Michael product in
reaction pathways leading to the two possible enantiomers 29% isolated yield, and a similar ee (17%) as the adduct
proceed through the same type of reaction mechanism, our with benzalacetone (Entry 5). The low yield was explained
816
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 810–821

Asymmetric Lewis Acid Catalysis in Water
Table 5. Aqueous biphasic Michael additions catalyzed by Yb- at 60 °C (Entry 7).[22] This could be a useful Entry into α-
(OTf)3/-Pro.[a]
alkylated amino acids and other nitrogen containing com-
pounds. In the reactions between MVK and the five- and
six-membered cyclic donors 12 and 13 relatively short reac-
tion times at 60 °C were required to obtain excellent iso-
lated yields (95 and 99%, respectively, Entries 8 and 9). Un-
fortunately, and somewhat surprisingly, we obtained both
of these products in racemic or near-racemic form (Ͻ5%
ee). It has been reported that some quaternary carbon cen-
ters formed by Michael addition of dicarbonyls may be sus-
ceptible to racemization.[39] To examine whether the lack of
enantioselectivity in this reaction was due to racemization
after initial asymmetric induction by the chiral catalyst, we
performed the addition between 13 and MVK using an-
other procedure known to be enantioselective.[40]
The isolated enantiomerically enriched addition product
was then subjected to prolonged heating (18 h, 90 °C) in the
presence of 5 mol-% Yb(OTf)3 and 12 mol-% -proline. No
significant racemization was observed, which demonstrated
that the addition product was stable under the present con-
ditions and that the low ee was likely a direct result of low
asymmetric induction in the reaction. In opposite to the
previous examples (except Entry 7) the stereogenic center is
created in the donor in these cases. In contrast to these re-
sults, Kobayashi have previously reported asymmetric AgI-
catalyzed Michael additions in water using similar do-
nors.[41] They highlighted the importance of the structure
of the ester moiety for achieving high ee. In the reaction
between the tert-butyl ester derivative 14 and MVK they
had obtained as high as 83% ee. To see if a change in the
ester structure would have a beneficial effect on ee with our
catalyst, we subjected 14 to our catalytic conditions. Unfor-
tunately, no significant ee was observed (Entry 10). At this
point we have no further explanation to the lack of asym-
metric induction with the Yb(OTf)3/-proline catalyst in
these reactions. Last to be examined were some more steri-
[a] Conditions: 1 mmol donor, 1.1 mmol acceptor, 0.05 mmol
cally demanding acceptors. As seen in Table 5 (Entry 11
and 12) neither mesityl oxide (15), nor 3-methyl-2-cy-
clohexen-1-one (16), underwent any reaction, not even un-
der enforced conditions. High temperature and high catalyst
loading together with prolonged reaction times gave no
conversion of the starting material. By using basic condi-
tions these substrates undergoes Michael reactions
smoothly. Obviously the explanation should be sought in
the mechanism. In our proposed mechanism both acceptor
and donor are coordinated to the same metal center and if
the donor is di-substituted at the 3-position this may create
Yb(OTf)3, 0.12 mmol -proline, 0.12 mmol NaOH, 1 mL of water,
90 °C. All experiments were run at least twice and listed yields and
ee values are averages. [b] Isolated yields after flash column
chromatography. [c] Determined by chiral HPLC except Entry 8
and 9, which were determined by chiral GC. [d] Low yield because
the addition product undergoes further reactions. [e] An exact value
could not be obtained because of overlapping peaks. [f] Run as for
[a] but with at 60 °C. [g] Run as for [a] but with 3.0 mmol acceptor.
[h] Conditions: 1 mmol donor, 3.0 mmol acceptor, 0.20 mmol
Yb(OTf)3, 0.24 mmol -alanine, 0.24 mmol NaOH, 1 mL of water,
90 °C.
by the initial adduct undergoing further reactions. In this a steric hindrance that prevent the bond-forming step.
study we also found that the scope of donors can be ex-
panded beyond the typical 1,3-dicarbonyl donors, which
significantly increases the synthetic potential. First, the
Stability and Leaching of the Aqueous Yb(OTf)3/
Catalyst
L-Proline
structurally rigid donor 10 undergoes addition to 5 in a
74% yield (Entry 6). Obviously, because of its geometry this
donor cannot coordinate both of its carbonyl oxygen atoms
As demonstrated in preliminary communication, the
to the metal simultaneously, which may serve to explain the aqueous phase containing the Yb(OTf)3/amino acid com-
low ee observed (Ͻ10%). As an example of a reaction with plex can be recycled a number of times without appreciable
a non-1,3-dicarbonyl donor, methyl nitroacetate (11), adds loss of activity. In other words, it appears as if the ligand
efficiently to MVK in 87% yield and 59% ee after only 4 h stabilizes the complex towards hydrolytic decomposition
Eur. J. Org. Chem. 2009, 810–821
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
817

U. M. Lindström, J. Wennerberg et al.
FULL PAPER
into catalytically inactive and water-insoluble ytterbium hy-
droxide. These characteristics are highly desirable, if not re-
quired, for aqueous biphasic catalysis to become more gen-
erally useful in a large-scale applications. In order to quan-
tify these practically useful ligand effects, we performed
some simple but clarifying experiments. First, an aqueous
catalytic solution of 5 mol-% Yb(OTf)3 and 12 mol-% -
proline was prepared, but rather than immediately adding
the reactants to the mixture, the catalyst solution was set
aside at room temperature for 7 d. After this time, visual
inspection of the solution did not reveal any precipitation
of water-insoluble ytterbium hydroxide. The reactants (4
and 5) were then added and the reaction was performed as
before. Quite remarkably, the yield and enantiomeric excess
of 6 was identical compared to the experiment where the
reactants were added to the reaction immediately after mix-
ing Yb(OTf)3 and -proline. Second, an aqueous solution
of Yb(OTf)3 without amino acid was prepared for direct
comparison. After 7 d at room temperature a significant
amount of precipitation was formed, which upon quantifi-
cation revealed that at least 30% of the original Yb(OTf)3
had been converted into insoluble and catalytically inactive
ytterbium hydroxide. This is compelling evidence of the sta-
bilizing influence of the α-amino acid ligand on the Lewis
acid. The Yb(OTf)3 solutions in the presence and absence
of ligand are pictured in Figure 6.
Conclusions
In the present work, we have been able to further evalu-
ate the potential of ligand-assisted Lewis acid catalysis in
water through an elaborate study on the use of α-amino
acids as ligands in Yb(OTf)3-catalyzed aqueous biphasic
Michael addition reactions. Several developments and dis-
coveries were made during the course of our study.
Through an extensive series of experiments involving op-
timization of metal/ligand ratios and reaction temperature
we were able to reduce the amount of Yb(OTf)3 used from
our original 10–20 mol-% to only 5 mol-%. The rate-acce-
lerating effect on Yb(OTf)3-catalyzed Michael additions ini-
tially observed with alanine was shown to be a general ef-
fect associated with α-amino acids as it was found also with
all of the other 16 α-amino acids tested in this study, albeit
to varying degrees. After screening a range of α-amino acids
we were able to relate the structure of a ligand to its influ-
ence on activity and selectivity. Some further insight into
the ligand effects on rate and selectivity through studies of
the influence of pH and temperature on the reactions was
obtained. An unexpected discovery from this study was the
large increase in enantioselectivity with increasing tempera-
ture. An Eyring plot showed that the unusual effect displays
a linear behavior in the temperature range between 10–
90 °C. Calculation of the differential activation enthalpy
and entropy for this reaction revealed a large entropic con-
tribution to the overall Gibbs energy of activation for the
enantiodifferentiating step. This result, in combination with
the large dependence of yields and selectivities on the po-
larity of the amino acid side-chain, suggests a very strong
influence of solvation effects in determining the outcome of
these reactions. From a practical point of view we devel-
oped a process, which is faster, more general, and require
less amounts of reactants and catalyst than previously re-
ported.[22,23] Although, the reactions in this article are per-
formed in a small scale including extractive work-up and
chromatographic purification there is a potential to avoid
these techniques when applied at larger scale. The work pre-
sented herein is an example of the academic relevance in
exploring water as a reaction medium as its distinctive and
unique solvent properties may lead to the discovery of new
types of reactions and thus to expand the scope of organic
synthesis. Our search for better enantioselectivities and ap-
plications will be reported in due course.
Figure 6. Aqueous solutions of Yb(OTf)3 without added ligand
(left tube) and with added -proline (right tube).
Finally, we were interested in determining the extent of
metal leaching in these reactions. The fact that we were able
to recycle the aqueous catalytic phase multiple times with-
out any observable decrease in activity suggested that no
major decomposition of the metal catalyst was taking place
between cycles. In order to quantify the amount of metal
leaching from the aqueous phase, we performed the reac-
tion between 12 and MVK with 1 gram of donor and ana-
lyzed the organic product phase for inorganic residues. It
was found that the product phase, which spontaneously
separated from the aqueous phase upon termination of stir-
ring and after cooling, contained as little as 0.2% inorganic
residues. This result indicated that the aqueous catalytic
phase, in theory, can be recycled tens of times without seri-
ous loss of activity.
Experimental Section
General Experimental Details: Deionized water was used in all reac-
tions. All reagents and ligands are commercially available com-
pounds, or simple derivatives thereof, and used as received from the
supplier. Conversions were determined by 1H NMR at 400 MHz in
CDCl3 or [D6]acetone. Enantiomeric excesses were determined by
HPLC with chiral column Chiralpak AD-H, n-hexane/2-propanol
unless otherwise stated. Purification of products was performed by
flash column chromatography on normal phase silica gel, 35–70 µ,
1
60 Å. All products are known compounds and gave consistent H
and 13C NMR spectroscopic data. For documentation of purity
818
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 810–821

Asymmetric Lewis Acid Catalysis in Water
1H NMR spectrum for each of the products is included in the
a
product was determined by chiral HPLC (n-hexane/2-propanol,
Supporting Information
87:13).
Michael Addition with
D-Lactic Acid and N-Mono- and: N,N-Di-
Determination of Reaction Order in Acceptor and Donor: See Fig-
ure 2. NaOH (0.06 , 0.12 equiv.) was added to a mixture of -
alanine (5.3 mg, 0.06 mmol) in a Radley tube and the mixture was
stirred for 15 min at room temperature and then ytterbium tri-
fluoromethanesulfonate (31.0 mg, 0.05 mmol) was added. Ethyl
acetoacetate (63 µL, 0.5 mmol) and methyl vinyl ketone (406 µL,
5.0 mmol) were then added. The reaction tubes were capped and
the reaction mixtures were stirred vigorously for 30, 60, 90 and
120 min, respectively at 25 Ϯ0.5 °C. The reaction mixture was then
diluted with water and extracted with ethyl acetate (10 mL). The
organic phase was dried with MgSO4, filtered and concentrated.
methylated Alanine as Ligands: See Table 1. a) No pH Adjustment:
The ligand (0.12 mmol) was stirred in NaOH (0.2 , 0.12 equiv.) in
a Radley tube for 15 min at room temperature and then ytterbium
trifluoromethanesulfonate (0.10 mmol) was added. Acetylacetone
(1.0 mmol) and 2-cyclohexen-1-one (3.0 mmol) were then added.
The reaction tube was capped, and the mixture was stirred vigor-
ously at 60 °C for 24 h. The reaction mixture was then diluted with
water and extracted with ethyl acetate (6 mL). The organic phase
was dried with MgSO4, filtered and concentrated to give 3-(3-oxo-
cyclohexyl)pentane-2,4-dione. The conversion and enantiomeric
excess of the crude product were determined by 1H NMR (CDCl3)
and chiral HPLC (n-hexane/2-propanol, 87:13), respectively. b)
With pH Adjustment: These experiments were performed in an
identical fashion except the pH was adjusted to 6.70 before ad-
dition of the reactants.
1
The crude mixture was analyzed by H NMR ([D6]acetone), which
afforded the ratios of 1 to ethyl 2-ethanoyl-5-oxohexanoate (2). 10
= 100%, ln[1]0 = 1.39, 130 = 15%, ln[1]30 = 3.28, 160 = 1%,
ln[1]60 = 5.99, 190 = 0%, 1120 = 0%. The rate constant for MVK
was obtained in an identical way by using ethyl acetoacetate
(633 µL, 5.0 mmol) and methyl vinyl ketone (40 µL, 0.5 mmol). The
Influence of Metal/Ligand Ratio on Yield and ee: See Table 2. To a
mixture of -alanine (0.10 mmol–0.15 mmol) and ytterbium tri-
fluoromethanesulfonate (0.05 or 0.10 mmol) in deionized water
(1.0 mL) was added NaOH (0.2 ) until the solution reached
pH 6.70. The reaction mixtures were stirred in Radley tubes at
room temperature for 15 min. Acetylacetone (1.0 mmol) and 2-cy-
clohexen-1-one (3.0 mmol) were then added. The reaction tubes
were capped, and the mixtures were stirred vigorously at 60 °C or
70 °C for 18 h. The reaction mixture was then diluted with water
and extracted with ethyl acetate (6 mL). The organic phase was
dried with MgSO4, filtered and concentrated to give 3-(3-oxocy-
clohexyl)pentane-2,4-dione. The conversion and enantiomeric ex-
cess of the crude product were determined by 1H NMR (CDCl3)
and chiral HPLC (n-hexane/2-propanol, 87:13), respectively.
1
crude mixture was analyzed by H NMR ([D6]acetone), which af-
forded the ratios of MVK to ethyl 2-ethanoyl-5-oxohexanoate 2.
MVK0 = 100%, ln[MVK]0 = 1.39, MVK30 = 91%, ln[MVK]30
=
1.48, MVK60 87%, ln[MVK]60 1.53, MVK90 80%,
=
=
=
ln[MVK]90 = 1.61, MVK120 = 75%, ln[MVK]90 = 1.67. When plot-
ting ln[1] and ln[MVK] against time, a straight line, typical for a
first-order reaction, is obtained.
Reaction Under “Enforced” Organocatalytic Conditions: See
Scheme 2.
3-(3-Oxocyclohexyl)pentane-2,4-dione (6): A) Without Acetic Acid:
Compound 3[36] (50.2 mg, 0.24 mmol) was stirred in H2O (1 mL)
and then acetylacetone (78 µL, 0.76 mmol) and 2-cyclohexene-1-
one (291 µL, 3.0 mmol) were added. The reaction mixture was
stirred vigorously for 24 h at 60 °C. The reaction was diluted with
water and extracted with ethyl acetate (10 mL). The organic phase
was dried with MgSO4, filtered and concentrated. The crude mix-
ture was analyzed for conversion by 1H NMR (CDCl3) and for
enantiomeric excess by chiral HPLC (2-propanol/hexane, 1:6.5),
which afforded 23% conversion and 7% ee (S). B) With Acetic
Acid: Compound 3[28] (50.2 mg, 0.24 mmol) was stirred in H2O
(1 mL) and then acetic acid (19.5 µL, 0.34 mmol), acetylacetone
(78 µL, 0.76 mmol) and 2-cyclohexene-1-one (291 µL, 3.0 mmol)
were added. The reaction mixture was stirred vigorously for 24 h
at 60 °C. The reaction was diluted with water and extracted with
ethyl acetate (10 mL). The organic phase was dried with MgSO4,
filtered and concentrated. The crude mixture was analyzed for con-
version by 1H NMR (CDCl3) and for enantiomeric excess by chiral
HPLC (n-hexane/2-propanol, 6.5:1), which afforded 33% conver-
sion and 11% ee (S).
Change in ee Values and Yields with Reaction Time: See Table 3.
NaOH (0.2 ) was added to a mixture of -alanine (0.12 mmol)
and ytterbium trifluoromethanesulfonate (0.05 mmol) in deionized
water (1.0 mL) until the solution reached pH 6.70. Acetylacetone
(1.0 mmol) and 2-cyclohexen-1-one (3.0 mmol) were then added.
The reaction tube was capped and the reaction was stirred vigor-
ously for 1–24 h at 60 °C. The reaction mixture was then diluted
with water and extracted with ethyl acetate (10 mL). The organic
phase was dried with MgSO4, filtered and concentrated to give 3-
(3-oxocyclohexyl)pentane-2,4-dione. The conversion and enantio-
meric excess of the crude product were determined by 1H NMR
(CDCl3) and chiral HPLC (n-hexanes/2-propanol, 87:13), respec-
tively.
Procedure for Screening Ligand/Metal Ratios: See Figure 4. NaOH
(0.2 ) was added to
a mixture of -alanine (0.02 mmol–
0.30 mmol) and ytterbium trifluoromethanesulfonate (0.10 mmol)
in deionized water (1.0 mL) until the solution reached pH 6.70. The
reaction mixtures were stirred in Radley tubes at room temperature
for 15 min. Acetylacetone (1.0 mmol) and 2-cyclohexen-1-one
(3.0 mmol) were then added. The reaction tubes were capped, and
the mixtures were stirred vigorously at 60 °C for 18 h. The reaction
mixture was then diluted with water and extracted with ethyl ace-
tate (6 mL). The organic phase was dried with MgSO4, filtered and
concentrated to give 3-(3-oxocyclohexyl)pentane-2,4-dione. The
conversion and enantiomeric excess of the crude product were de-
Influence of pH on Enantioselectivity: See Figure 3. Deionized water
(2 mL) was added to a mixture of amino acid (0.24 mmol) and
ytterbium trifluoromethanesulfonate (0.20 mmol). The reaction
mixtures were stirred in Radley tubes for 15 min at room tempera-
ture. The pH of the solutions was measured to 5.16 and one tube
was kept at this pH. The pH of the other ten tubes was adjusted
by either adding HCl (1.0 ) or NaOH (0.2 ) giving pH 3.05, 3.9,
5.49, 5.81, 6.11, 6.53, 6.71, 6.78, 7.66 and 8.49. Acetylacetone
(1 mmol) and 2-cyclohexen-1-one (3 mmol) were then added. The
reaction tubes were capped, and the mixtures were stirred vigor-
ously at 40 °C for 22 h. The reaction mixture was then diluted with
water and extracted with ethyl acetate (10 mL). The organic phase
was dried with MgSO4, filtered and concentrated to give 3-(3-oxo-
cyclohexyl)pentane-2,4-dione. The enantiomeric excess of the crude
1
termined by H NMR (CDCl3) and chiral HPLC (n-hexane/2-pro-
panol, 87:13), respectively.
Yb(OTf)3-Catalyzed Michael Additions Using Various Amino Acids
as Ligands: See Table 4. NaOH (0.2 ) was added to a mixture of
amino acid (0.12 mmol) and ytterbium trifluoromethanesulfonate
Eur. J. Org. Chem. 2009, 810–821
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
819

U. M. Lindström, J. Wennerberg et al.
FULL PAPER
(0.05 mmol) in deionized water (1.0 mL) until the solution reached
pH 6.70. Acetylacetone (1.0 mmol) and 2-cyclohexen-1-one
(3.0 mmol) were then added. The reaction tube was capped and the
reaction was stirred vigorously for 24 h at 60 °C or for 10 h at
90 °C. The reaction mixture was then diluted with water and ex-
tracted with ethyl acetate (10 mL). The organic phase was dried
with MgSO4, filtered and concentrated to give 3-(3-oxocyclohexyl)-
pentane-2,4-dione. The conversion and enantiomeric excess of the
crude product were determined by 1H NMR (CDCl3) and chiral
HPLC (n-hexanes/2-propanol, 87:13), respectively.
1.0 mmol) and trans-β-nitrostyrene (164 mg, 1.1 mmol) were re-
acted at 90 °C for 14 h. Flash chromatography on silica gel (dichlo-
romethane) afforded the product (72 mg, 29%, 17% ee).
3-Hydroxy-2-(3-oxocyclohexyl)cyclohex-2-enone: See Table 5, Entry
6.[47] Following the general procedure, 1,3-cyclo-hexanedione
(112 mg, 1.0 mmol) and 2-cyclohexen-1-one (106 mL, 1.1 mmol)
were reacted at 90 °C for 14 h. Flash chromatography on silica gel
(n-hexane/ethyl acetate, 1:1 + 2% triethylamine) afforded the prod-
uct (154 mg, 74%, Ͻ10% ee).
Methyl 2-Nitro-5-oxohexanoate: See Table 5, Entry 7.[48] Following
the general procedure, methyl nitroethanoate (92 µL, 1.0 mmol)
and methyl vinyl ketone (89 µL, 1.1 mmol) were reacted at 60 °C
for 4 h. Flash column chromatography on silica gel (n-heptane/
ethyl acetate, 4:1) afforded the product (164.6 mg, 87%, 59% ee).
Procedure for the “Eyring Plot Reactions”: See Figure 5. NaOH
(0.2 ) was added to a mixture of amino acid (0.12 mmol) and
ytterbium trifluoromethanesulfonate (0.05 mmol) in deionized
water (1.0 mL) until the solution reached pH 6.70. Acetylacetone
(1.0 mmol) and 2-cyclohexen-1-one (3.0 mmol) were then added.
The reaction tube was capped and the reaction was stirred vigor-
ously for 24 h at 30 °C, 50 °C, 70 °C or 90 °C. The reaction mixture
was then diluted with water and extracted with ethyl acetate
(10 mL). The organic phase was dried with MgSO4, filtered and
concentrated to give 3-(3-oxocyclohexyl)pentane-2,4-dione. The en-
antiomeric excess of the crude product was determined by chiral
HPLC (n-hexane/2-propanol, 87:13).
Ethyl 2-Oxo-1-(3-oxobutyl)cyclohexanecarboxylate: See Table 5,
Entry 8.[49] Following the general procedure, ethyl cyclohexanone-
2-carboxylate (160 µL, 1.0 mmol) and methyl vinyl ketone (89 µL,
1.1 mmol) were reacted at 60 °C for 6 h. Flash column chromatog-
raphy on silica gel (n-heptane/ethyl acetate, 2:1 + 2% triethylamine)
afforded the product (228.5 mg, 95%, Ͻ2% ee).
Ethyl 2-Oxo-1-(3-oxobutyl)cyclopentanecarboxylate: See Table 5,
Entry 9.[49] Following the general procedure, ethyl 2-oxocyclopen-
tanecarboxylate (145 µL, 1.0 mmol) and methyl vinyl ketone
(89 µL, 1.1 mmol) were reacted at 60 °C for 5 h. Flash column
chromatography on silica gel (n-heptane/ethyl acetate, 2:1 + 2%
triethylamine) afforded the product (224.3 mg, 99%, Ͻ2% ee).
Details for Michael Reactions Between Various Donors and Ac-
ceptors: See Table 5.
General Procedure: -proline (0.12 mmol) was stirred in NaOH
(0.030 , 1 mL) in a Radley tube for 15 min at room temperature
and then ytterbium trifluoromethanesulfonate (0.05 mmol) was
added. Donor (1.0 mmol) and acceptor (1.1 mmol) were then
added. The reaction tube was capped, and the mixture was stirred
vigorously at 60 °C or 90 °C for 4 h to 24 h. The reaction mixture
was then diluted with water and extracted with ethyl acetate
(6 mL). The organic phase was dried with MgSO4, filtered and con-
centrated. Flash column chromatography of the residue on silica
gel afforded the Michael adduct. The enantiomeric excess was de-
termined by chiral HPLC (n-hexane/2-propanol, 87:13).
tert-Butyl
2-Oxo-1-(3-oxobutyl)cyclopentanecarboxylate:
See
Table 5, Entry 10.[50] Following the general procedure, tert-butyl 2-
oxocyclopentanecarboxylate (220 mg, 1.2 mmol) and methyl vinyl
ketone (295 µL, 3.0 mmol) were reacted at 90 °C for 10 h. Flash
chromatography on silica gel (n-hexane/ethyl acetate, 4:1) afforded
the product (256 mg, 84%, Ͻ5% ee). HPLC [1:99 n-hexane/2-pro-
panol, 0.44 mL/min, 280 nm, tR(1) = 31.3 min, tR(2) = 33.7 min,
broad overlapping peaks].
Determination of Inorganic Residues in the Product Phase of the
Michael Addition Between 12 and MVK: The product phase
(1.0319 g) from the reaction between 12 and MVK, performed ac-
cording to the procedure described above, was placed in a crucible
that previously had been heated at 600 °C for 30 min and then
cooled in a dessicator. Concentrated sulfuric acid (1 mL, p.a.
quality) was added to the sample. The mixture was heated gently
on a hot plate until the sample was thoroughly charred. The sample
was then heated at 600 °C overnight. The crucible was cooled in a
dessicator and then weighed. The mass of residual inorganics
was 0.00204 g, which corresponded to 0.2% of the total sample
weight.
Ethyl (S)-3-Oxo-2-(3-oxocyclohexyl)butanoate: See Table 5, Entry
1.[42] Following the general procedure, ethyl acetoacetone (127 µL,
1.0 mmol) and 2-cyclohexen-1-one (106 µL, 1.1 mmol) were reacted
at 90 °C for 12 h. Flash column chromatography on silica gel (n-
heptane/ethyl acetate, 2:1 + 2% triethylamine) afforded the product
(174 mg, 77%, 70% ee).
(S)-3-(3-Oxocyclohexyl)pentane-2,4-dione: See Table 5, Entry 2.[43]
Following the general procedure, acetylacetone (103 µL, 1.0 mmol)
and 2-cyclohexen-1-one (106 µL, 1.1 mmol) were reacted at 90 °C
for 21 h. Flash column chromatography on silica gel (n-heptane/
ethyl acetate, 2:1 + 2% triethylamine) afforded the product
(184 mg, 94%, 79% ee).
Supporting Information (see also the footnote on the first page of
this article): 1H NMR spectra of all products. Determination of
absolute configuration of 6.
(S)-3-(3-Oxocyclopentyl)pentane-2,4-dione: See Table 5, Entry 3.[44]
Following the general procedure, acetylacetone (103 mL, 1.0 mmol)
and 2-cyclopenten-1-one (92 µL, 1.1 mmol) were reacted at 90 °C
for 21 h. Flash chromatography on silica gel (n-hexane/ethyl ace-
tate, 3:2 + 2% triethylamine) afforded the product (111 mg, 61%,
49% ee).
Acknowledgments
This work was supported by a McGill University start-up grant,
National Sciences and Engineering Research Council of Canada
(NSERC) and the Crafoord Foundation. R. D. acknowledges a
post-doc scholarship from the Carl Trygger Foundation. We are
grateful to the group of Prof. M. Damha for unlimited access to
an HPLC.
3-Acetyl-4-phenylheptane-2,6-dione: See Table 5, Entry 4.[45] Fol-
lowing the general procedure, acetylacetone (103 µL, 1.0 mmol)
and benzalacetone (161 mg, 1.1 mmol) were reacted at 90 °C for
24 h. Flash column chromatography on silica gel (n-heptane/ethyl
acetate, 2:1) afforded the product (172 mg, 70%, 19% ee).
(S)-3-(2-Nitro-1-phenylethyl)pentane-2,4-dione: See Table 5, Entry
5.[46] Following the general procedure, acetylacetone (103 µL,
[1] Catalytic Asymmetric Synthesis, 2nd ed. (Ed.: I. Ojima), Wiley-
VCH, New York, 2000.
820
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 810–821

Asymmetric Lewis Acid Catalysis in Water
[2] Comprehensive Asymmetric Catalysis (Eds.: E. N. Jacobsen, A.
Pflatz, H. Yamamoto), Springer, Berlin, 1999.
[3] Special issue on asymmetric catalysis: Proc. Natl. Acad. Sci.
USA 2002, 101, 5347–5850.
[4] Organic Reactions in Water (Ed.: U. M. Lindström), Blackwell
Publishing, Oxford, 2007.
[5] C. J. Li, T. H. Chan in Comprehensive Organic Reactions in
Aqueous Media, Wiley, 2007.
[6] H. C. Hailes, Org. Proc. Res. Dev. 2007, 11, 114.
[7] C. J. Li, L. Chen, Chem. Soc. Rev. 2006, 35, 68.
[8] Aqueous-Phase Organometallic Catalysis (Eds.: B. Cornils,
W. A. Herrmann), Wiley-VCH, Weinheim, 2004.
[9] C. J. Li, Chem. Rev. 2005, 105, 3095.
[10] S. Kobayashi, Chem. Lett. 1991, 2187.
[11] S. Kobayashi, S. Nagayama, T. Busujima, J. Am. Chem. Soc.
1998, 120, 8287.
[12] S. Otto, G. Boccaletti, J. B. F. N. Engberts, J. Am. Chem. Soc.
1998, 120, 4238.
[13] S. Otto, J. F. B. N. Engberts, J. Am. Chem. Soc. 1999, 121, 6798.
[14] For a review of ligand accelerated catalysis, see: D. J. Berris-
ford, C. Bolm, K. B. Sharpless, Angew. Chem. Int. Ed. Engl.
1995, 34, 1059.
[15] For an example, see: S. Kobayashi, T. Ogino, H. Shimizu, S.
Ishikawa, T. Hamada, K. Manabe, Org. Lett. 2005, 7, 4729.
[16] For some examples, see: a) Z. Bourhani, A. V. Malkov, Chem.
Commun. 2005, 4592; b) S. Kobayashi, N. Aoyama, K. Man-
abe, Chirality 2003, 15, 124; c) S. Kobayashi, N. Aoyama, K.
Manabe, Synlett 2002, 483; d) See ref. 21.
Ed. 2004, 43, 2152; b) A. Córdova, H. Sundén, M. Engqvist,
I. Ibrahem, J. Casas, J. Am. Chem. Soc. 2004, 126, 8914; c)
A. B. Northrup, D. W. C. MacMillan, J. Am. Chem. Soc. 2002,
124, 6798; d) K. Sakthivel, W. Notz, T. Bui, C. F. Barbas, J.
Am. Chem. Soc. 2001, 123, 5260.
For a review, see: C. Kremer, J. Torres, S. Domínguez, A. Med-
eros, Coord. Chem. Rev. 2005, 249, 567.
K. Aplander, R. Ding, U. M. Lindström, J. Wennerberg, S.
Schultz, Angew. Chem. Int. Ed. 2007, 46, 4543.
a) S. Pelzer, T. Kauf, C. van Wüllen, J. Christoffers, J. Or-
ganomet. Chem. 2003, 684, 308; b) J. Christoffers, Eur. J. Org.
Chem. 1998, 1259.
G. Gal, J. M. Chemerda, D. F. Reinhold, R. M. Purick, J. Org.
Chem. 1977, 42, 142.
J. Comelles, M. Moreno-Mañas, E. Pérez, A. Roglans, R. M.
Sebastián, A. Vallribera, J. Org. Chem. 2004, 69, 6834.
G. M. Whitesides, E. E. Simanek, J. P. Mathias, C. T. Seto,
D. N. Chin, M. Mammen, D. M. Gordon, Acc. Chem. Res.
1995, 28, 37.
R. Saito, S. Naruse, K. Takano, K. Fukuda, A. Katoh, Y. In-
oue, Org. Lett. 2006, 8, 2067, and references therein.
M. P. Sibi, U. Gorikunti, M. Liu, Tetrahedron 2002, 58, 8357.
D. Sinou, Tetrahedron Lett. 1981, 22, 2987.
K. Harada, T. Yoshida, J. Org. Chem. 1972, 37, 4366.
H. Buschmann, H.-D. Scharf, N. Hoffmann, P. Esser, Angew.
Chem. Int. Ed. Engl. 1991, 30, 477.
K. Tan, R. Alvarez, M. Nour, C. Cavé, A. Chiaroni, C. Riche,
J. D’Angelo, Tetrahedron Lett. 2001, 42, 5021.
J. Christoffers, U. Rößler, T. Werner, Eur. J. Org. Chem. 2000,
701.
S. Kobayashi, K. Kakumoto, Y. Mori, K. Manabe, Isr. J.
Chem. 2001, 41, 247.
K. Majima, R. Takita, A. Okada, T. Ohshima, M. Shibasaki,
J. Am. Chem. Soc. 2003, 125, 15837.
A. Soriente, R. Arienzo, M. De Rosa, L. Palombi, A. Spinella,
A. Scettri, Green Chem. 1999, 1, 157.
T. Ikariya, H. Wang, M. Watanabe, K. Murata, J. Organomet.
Chem. 2004, 689, 1377.
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[17] H. Li, Y. Wang, L. Tang, F. Wu, X. Liu, C. Guo, B. M. Fox-
man, L. Deng, Angew. Chem. Int. Ed. 2005, 44, 105.
[18] J. Christoffers, A. Baro, Angew. Chem. Int. Ed. 2003, 42, 1688.
[19] N. Krause, A. Hoffmann-Röder, Synthesis 2001, 171.
[20] S. Shirakawa, S. Kobayashi, Synlett 2006, 1410.
[21] Y. Mori, K. Kakumoto, K. Manabe, S. Kobayashi, Tetrahedron
Lett. 2000, 41, 3107.
[22] E. Keller, B. L. Feringa, Synlett 1997, 842.
[23] E. Keller, B. L. Feringa, Tetrahedron Lett. 1996, 37, 1879.
[24] T. Kawabata, M. Kato, T. Mizugaki, K. Ebitani, K. Kaneda,
Chem. Eur. J. 2005, 11, 288.
[25] R. Ding, K. Katebzadeh, L. Roman, K.-E. Bergquist, U. M.
Lindström, J. Org. Chem. 2006, 71, 352.
P. P. Baruah, A. Boruah, D. Prajapati, J. S. Sandhu, Ind. J.
Chem., Sect. B 1998, 37, 425.
H. Brunner, B. Kimel, Monatshefte Chem. 1996, 127, 1063.
K. Conrow, J. Org. Chem. 1966, 31, 1050.
W. J. Thompson, C. A. Buhr, J. Org. Chem. 1983, 48, 2769.
J. Christoffers, J. Chem. Soc., Perkin Trans. 1 1997, 3141.
Y. Hamashima, D. Hotta, M. Sodeoka, J. Am. Chem. Soc.
2002, 124, 11240.
[26] For some examples, see: a) Ref.[21]; b) J. Kofoed, J.-L. Re-
ymond, T. Darbre, Org. Biomol. Chem. 2005, 3, 1850–1855; c)
J. Gyarmati, C. Hajdu, Z. Dinya, K. Micskei, C. Zucchi, G.
Pályi, J. Organomet. Chem. 1999, 586, 106; d) T. Darbre, M.
Machuqueiro, Chem. Commun. 2003, 1090–1091; e) K. Ra-
jender Reddy, C. V. Rajasekhar, G. Gopi Krishna, Synth. Com-
mun. 2007, 37, 1971–1976.
[46]
[47]
[48]
[49]
[50]
[27] For some prominent examples, see: a) A. B. Northrup, I. K.
Mangion, F. Hettche, D. W. C. MacMillan, Angew. Chem. Int.
Received: September 22, 2008
Published Online: January 2, 2009
Eur. J. Org. Chem. 2009, 810–821
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
821