Recueil des Travaux Chimiques des Pays-Bas, 115 / 11-12, November / December 1996
Recl. Trav. Chim. Pays-Bas 115, 457-464 (1996)
SSDI 0165-0513(96)00021-6
457
0165-0513/96/11/12457-08$12.00
Lewis-acid catalysis of carbon carbon bond forming reactions in water
Jan B.F.N. Engberts * , Ben L. Feringa * , Erik Keller and Sijbren Otto
Department of Organic and Molecular Inorganic Chemistry, Nijenborgh 4, 9747 AG Groningen,
The Netherlands
(Received 7 June 1996)
Abstract. In this article, we review the recent progress that has been made in the field of Lewis-acid
catalysis of carbon carbon-bond-forming reactions in aqueous solution. Since water hampers the hard
hard interactions between the catalyst and the reactant, it often complicates catalysis. However, once
coordination has taken place, water can have beneficial effects on rates and selectivities of Lewis acid
catalysed Diels Alder reactions, aldol reactions, allylation reactions, Barbier and Mannich type
reactions as well as Michael additions.
1. Introduction
2. Lewis-acid Lewis-base interactions in water
Most synthetic reactions are performed in organic solvents.
The use of water as the reaction medium is usually avoided
since a large number of reactants decompose when brought
into contact with water. Furthermore, many organic reactants
are sparsely soluble in water.
Despite these disadvantages, water is becoming increasingly
popular as a medium for organic reactions. Being the most
abundant solvent on earth, it is very cheap, it is non-hazardous
to the environment and non-toxic. Moreover, aqueous solvents can have beneficial effects on rates and selectivities of
important organic transformations such as, for example, Diels
Alder reactions, aldol condensations and Michael additions1.
Finally, the use of water often simplifies work-up procedures.
These features, in combination with the ever more strict
legislation concerning industrial chemical processes, has led
to increased efforts to transfer reactions from organic to
aqueous solvents2. On an industrial scale, the first example of
this development is the Ruhr Chemie Rhone-Poulenc hydroformylation process3.
Also in the field of Lewis-acid catalysis water is becoming
increasingly popular, even though this solvent imposes severe
restrictions on the applicability of Lewis acids. Active Lewis
acids like BF3, TiCl4 and AlCl3 react violently with water and
cannot be used. Moreover, coordination of the Lewis acid to
the reactants is not as efficient in water as in solvents like
dichloromethane or benzene. Hence, applications of Lewisacid catalysis in water are still limited.
An extensive review of Lewis-acid catalysis in water was
published by Hay in 19874. Most of the efforts in this field are
concerned with hydrolysis reactions of coordinated reactants.
In 1993, Li reviewed carbon carbon bond forming reactions in
water5. In neither review much attention was paid to aqueous
carbon carbon bond forming reactions catalysed by Lewis
acids. This is a rapidly developing, though relatively new
field in organic chemistry and Kobayashi has already
reviewed his pioneering work on this subject6. We will now
provide a more recent overview of the progress made in this
field.
The first step in Lewis-acid catalysis is coordination of the
Lewis-acid to a Lewis basic site of a reactant. Most often this
Lewis basic site consists of one or more oxygen or nitrogen
atoms. Coordination of the Lewis acid induces a polarization
of part of the reactant molecule in conjugation with, or in the
proximity of, this site by withdrawing electron density from
the reactant to the catalyst. This activates the molecule with
respect to the second step, the actual organic transformation.
In the final step, the Lewis-acid product complex dissociates,
making the catalyst available for another cycle. The overall
catalytic efficiency is determined by the ease with which each
of these steps proceeds, as quantified by Ka, kr and Kd
(Scheme 1).
For most Lewis-acid-catalysed processes the rate constant of
the second step, kr, is not high. Values of kr for bimolecular
reactions are typically in the order of 0.01 - 10 M-1s-1. In
order to obtain an appreciable overall rate, a significant
portion of reactant molecules has to be coordinated to the
Lewis acid. Moreover, there is a qualitative, and sometimes
even quantitative7, correlation between Ka and kr, which
implies that when the Lewis acid coordinates the reactant
efficiently, it will usually also lead to efficient polarization. In
general, efficient coordination will ensure efficient catalysis
but only as long as there is no catalyst poisoning by the
product, i.e., the equilibrium constant for coordination of the
product to the catalyst, Kd, should not be too high.
The most frequently used solvents for Lewis-acid-catalysed
reactions are aprotic, apolar solvents like dichloromethane or
benzene. Polar and protic solvents are usually avoided. This
is due to solubility problems encountered when using
relatively apolar reagents, incompatibility of the catalyst with
these solvents and the inefficient coordination (small ka) of
the reactant to the Lewis acid in these solvents. These effects
are particularly pronounced in water.
In order to understand the way in which solvents in general,
and water in particular, affect the interactions between a
Lewis acid and a Lewis base, a thorough understanding of
these interactions is required.
458
J.B.F.N. Engberts, B.L. Feringa et al. / Lewis-acid catalysis in water
Table II Donor scales of some selected solvents.
Solvent
dichloromethane
benzene
acetonitrile
water
methanol
dimethyl sulfoxide
Scheme 1
An important approach involves the Hard-Soft-Acid-Base
(HSAB) theory, as developed by Pearson in 19638. According to this theory, Lewis acids and Lewis bases are divided
into two groups: hard acids and bases, which are usually
small, not very polarizable species with highly localized
charges, on one hand, and soft acids and bases which are
large polarizable species with delocalised charges on the
other hand. A selection of Lewis acids, ordered according to
their hardness in aqueous solution is presented in Table I. The
theory predicts high stabilities for hard-acid hard-base complexes, mainly resulting from electrostatic interactions and for
soft-acid soft-base complexes, where covalent bonding is also
important. Hard-acid soft-base and hard-base soft-acid complexes have usually small stabilities. Unfortunately, in a
quantitative sense, the predictive value of the HSAB theory is
limited.
Several other attempts have been made to define a Lewisacidity scale9. In view of the HSAB theory, the applicability
of a scale which describes Lewis acidity with only one
parameter will be unavoidable restricted to a narrow range of
structurally related Lewis bases. The use of more than one
parameter results in relationships with a more general validity10. However, a quantitative prediction of the gas-phase
stabilities of Lewis-acid Lewis-base complexes is still difficult. Hence the interpretation, not to mention the prediction,
of solvent effects on Lewis-acid Lewis-base interactions remains largely speculative. Consequently, the effects that are
considered to be important in rationalising the influence of
the solvent on Lewis-acid Lewis-base equilibria, can only be
treated in a qualitative manor.
First of all, the solvent will, in most cases, also act as a Lewis
base by coordinating to the catalyst. Analogous to the Lewisacidity scales, solvents have been ranked according to their
Lewis basicity. These scales, in as far as they try to quantify
Lewis basicity using only one parameter, also have limited
validity. An extensive discussion of the electron pair donor
ability and Lewis-basicity scales of solvents is given in Ref.
11. In Table II a number of Lewis-acidity parameters of some
relevant solvents are compared. It is apparent that aprotic and
apolar solvents coordinate relatively weakly to the catalyst,
whereas polar solvents exhibit stronger interactions. In the
latter solvents coordination of a Lewis base is less favourable
as a result of the strong solvent Lewis-acid interaction which
has to be disrupted. Steric interactions between the Lewis
base and coordinated solvent molecules are also important in
determining the stability of the complex12. Consequently,
catalysis by Lewis acids in solvents that coordinate strongly
to the Lewis acid will be less effective.
The second important influence of the solvent on Lewis-acid
Lewis-base equilibria is its interaction with the Lewis base.
Table I Classification of some selected Lewis acids according to the
HSAB theory in aqueous solution a.
Hard
H+
Fe3+
Co3+
Al3+
La3+
a
Taken from Ref. 8a.
Borderline
Fe2+
Ni2+
Cu2+
Zn2+
Soft
Cu+
Hg2+
Cd2+
DS a
6
9
12
17
18
27.5
DN b
DNbulk c
0.1
14.1
18.0
19
29.8
0.8
11.1
40.3
31.3
27
a
Donor strengths, taken from Ref. 11, based upon the solvent effect on
the symmetric stretching frequency of the soft Lewis acid HgBr2. b Gutmann’s donor number taken from Ref. 11, based upon ∆Hr for the
process of coordination of an isolated solvent molecule to the moderately
hard SbCl5 molecule in dichloroethane. c Bulk donor number calculated
as described in Ref. 45 from the solvent effect on the adsorption spectrum
of VO(acac)2.
Consequently, the Lewis acidity and, when hard Lewis bases
are concerned, the hydrogen bond donor capacity of the
solvent, are important parameters. The electron pair acceptor
capacities, quantified by the acceptor number AN, together
with the hydrogen bond donor acidities, α, of some selected
solvents are listed in Table III. Water is among the solvents
with the highest AN and, accordingly, will interact strongly
with Lewis bases. This seriously hampers the efficiency of
Lewis-acid catalysis in water.
Thirdly, the intramolecular association of a solvent affects the
Lewis-acid Lewis-base equilibrium13. Upon complexation, one
ore more solvent molecules that were initially coordinated to
the Lewis acid or the Lewis base, are liberated into the bulk
liquid phase, which is entropically favourable. This effect is
more pronounced in aprotic solvents than in protic solvents,
which usually have a higher cohesive energy density. The less
favourable entropy change in protic solvents is somewhat
counteracted by the more favourable enthalpy change upon
release of a coordinated solvent molecule into the bulk liquid,
resulting from the newly formed hydrogen bonds.
Finally, the solvent also interacts with sites of the Lewis acid
and the Lewis base that are not directly involved in mutual
coordination, thereby altering the electronic properties of the
complex. For example, delocalization of charges into the
surrounding solvent molecules causes ions in solution to be
softer than in the gas phase12. Again, water is particularly
effective since it can act as an efficient electron-pair acceptor
as well as donor.
In summary, we conclude that complexation of a hard Lewis
acid to a hard Lewis base is seriously hampered by water14.
Since water interacts through either oxygen or hydrogen
atoms, it can be classified as a hard solvent (Table IV). This
implies that it will interfere most strongly with the interactions
between hard Lewis acids and hard Lewis bases. Soft soft
interactions, on the other hand, will be much less affected8b.
However, this is of limited practical use in Lewis-acid
catalysis, since the majority of reactants contain electronegative and hard elements as Lewis-basic sites.
Thermodynamic analysis clearly shows a difference in the
way water affects hard-hard interactions, compared to soft-soft
Table III Acceptor numbers (AN) and hydrogen bond donor acidities
(α) of some selected solvents.
Solvent
dichloromethane
benzene
acetonitrile
water
methanol
dimethyl sulfoxide
AN a
20.4
8.2
19.3
54.8
41.3
19.3
αb
0.00
(0.30)
0.19
1.17
0.93
0.00
a
Taken from Ref. 46, based on the 31P-NMR chemical shift of triethylphosphine oxide in the respective pure solvent. b Taken from Ref. 11b
based on the solvatochromic shift of a pyridinium-N-phenoxide betaine
dye.
Recueil des Travaux Chimiques des Pays-Bas, 115 / 11-12, November / December 1996
459
Table IV Softness of some selected solvents according to the µ-scale.
Solvent
acetonitrile
water
methanol
dimethyl sulfoxide
µa
0.35
0.00 b
0.02
0.22
Taken from Ref. 47, resulting from a comparison of ∆Gtro(water →
solvent) for Ag+ with the mean of ∆Gtro(water → solvent) for Na+ and
K+. b By definition.
a
interactions. In water, hard hard interactions are usually endothermic and occur only as a result of a gain in entropy,
originating from a liberation of water molecules from the
hydration shells of the Lewis acid and the ligand. By contrast,
soft-soft interactions are mainly enthalpic in origin and are
characterized by a negative change in entropy15.
Equilibrium constants for coordination of hard monodentate
reactants to a Lewis acid catalyst in water are generally small.
This equilibrium constant has actually been determined in
only a few cases. Chin et al.16 reported an equilibrium
constant for coordination of the amide carbonyl group of
formylmorpholine (2) to a CoIII catalyst (1) of 0.4 ± 0.1 M-1
(Scheme 2). The same catalyst coordinates acetonitrile in
water with an equilibrium constant of 2.5 M-1 17.
Larger equilibrium constants can be achieved when reagents
are used that contain two (or more) Lewis-basic sites in a
suitable orientation to form a chelate with the catalyst18. In
general, complexes of bidentate ligands are more stable than
expected on the basis of the stability of complexes of the
corresponding monodentate ligands. This observation is commonly referred to as the chelate effect. Although its exact
origin is still under debate19, the entropy gain upon release of
the second coordinated solvent molecule into the bulk phase
constitutes part of the effect. This part of the chelate effect is
less pronounced in highly ordered solvents, such as water,
than in less structured solvents. Unfortunately, detailed studies of the effect of the solvent on the chelate effect have not
been performed.
The equilibrium constants for bidentate coordination of reactants to Lewis acids in water are typically two to three orders
of magnitude larger than those for monodentate binding.
Examples can be found in the section dealing with the DielsAlder reaction. For this reason, the majority of Lewis-acidcatalysed reactions in water involve bidentate reactants.
3. The Diels-Alder reaction
The rate of the Diels-Alder reaction was traditionally considered to be insensitive to the solvent. However, in 1980,
Breslow showed that this reaction can be accelerated considerably when performed in water20. The endo-exo selectivity
also benefits from the use of this solvent. The preference for
the endo cycloadduct usually observed, becomes more pronounced in water. A large number of papers have appeared
discussing the origin of these remarkable observations21. Evidence has been presented that there are two effects causing
this aqueous rate and selectivity enhancement: enforced hydrophobic interactions and hydrogen bonding to the
activating group of the dienophile21.
Lewis acids are also known to increase the rate and selectivity
Scheme 3
Table V Second-order rate constants for the uncatalysed and the
Cu2 + -catalysed Diels-Alder reaction between 4a and 5 in water and
acetonitrile at 25°C23.
2nd-order rate
constant
(M-1s-1)
uncatalysed
CH3CN
1.40 · 10-5
H2O
4.02 · 10-3
catalysed
0.01M Cu(NO3)2 in CH3CN
2.21
0.01M Cu(NO3)2 in H2O
3.25
Medium
of Diels-Alder reactions22. The combination of Lewis acid
catalysis and water was not, however, studied until 1995,
when it was shown that the Diels-Alder reaction between
3-(4-nitrophenyl)-1-(pyrid-2-yl)-2-propen-1-one (4a) and cyclopentadiene (5) was subject to Lewis-acid catalysis in water
(Scheme 3)23. The rate of this reaction has been determined in
several solvents (Table V). It is clear that very large accelerations can be achieved by the combined use of a Lewis acid as
a catalyst and water as the solvent. However, the large rate
enhancement in going from acetonitrile to water for the
uncatalysed reaction is strongly reduced in the case of the
CuII -catalysed reaction. The selectivity of the catalysed reaction is slightly decreased in water (Table VI). The catalytic
efficiency of the reaction is determined by the equilibrium
constant (Ka) for binding of the dienophile to the catalyst and
by the rate constant (kr) of the subsequent reaction with 5
(Scheme 1). For several metal ions these parameters have
been determined (Table VII). There is clearly no quantitative
correlation between Ka and kr in this case, although the
empirical Irving Williams order24 (CoII < NiII < CuII » ZnII) is
obeyed for both series.
The bidentate character of the dienophile 4 is essential in
order to obtain efficient catalysis. For the reaction of comTable VI Solvent effect on the endo-exo selectivity (%endo%exo) of the
uncatalysed and Cu2 + -ion-catalysed Diels-Alder reaction between 4b
and 5 at 25°C23.
Solvent
acetonitrile
ethanol
water
Uncatalysed
67 33
77 23
84 16
Cu(NO3)2 , 10 mM
94 6
96 4
93 7
Table VII Equilibrium constants for complexation of 4b to different
Lewis acids (Ka) and second-order rate constants for the reaction of
these complexes with 5 (kr) in water at 2M ionic strength (KNO3) at
25°C.
Lewis acid
Co2+
Ni2+
Cu2+
Zn2+
Scheme 2
Relative
rate
constant
1
278
158000
232000
Ka (M-1)
1.17 · 102
6.86 · 102
1.16 · 103
7.28 · 101
kr (M-1 · s-1)
8.40 · 10-2
9.46 · 10-2
2.56
1.18 · 10-1
460
J.B.F.N. Engberts, B.L. Feringa et al. / Lewis-acid catalysis in water
Scheme 4
Scheme 5
Scheme 7
reported in the early seventies. This titanium-tetrachloridecatalysed cross-aldol reaction28 proceeds in a regioselective
manner, giving high yields. These transformations are normally carried out under strictly non-aqueous conditions to
prevent decomposition of the catalyst and hydrolysis of the
silyl enol ethers used in the reaction.
Kobayashi29 reported the first Lewis-acid-catalysed cross-aldol reaction in aqueous solution. He examined the effects of
lanthanide triflates in the reaction of several silyl enol ethers
with a commercial formaldehyde solution (Scheme 7). The
same reaction in the absence of catalyst was previously
studied by Lubineau30, who reported the water-promoted
aldol reaction of silyl enol ethers with aldehydes. However,
the scope and yields were not satisfactory. When the reaction
was performed in the presence of a Lewis-acid catalyst, in
most cases the silyl enol ethers reacted smoothly with formaldehyde to give the corresponding hydroxymethylated product
Table VIII Cross-aldol reaction in aqueous formaldehyde / THF (1/9)
solution.
Scheme 6 a Yield without catalyst
pound 6 (Scheme 4) with cyclopentadiene, no catalysis is
observed, because 6 is not capable of forming a chelate with
the Lewis acid.
Recently Wang and co-workers25 claimed catalysis of a hetero
Diels-Alder reaction involving a monodentate in-situ generated dienophile. In 1985, Grieco had already reported that
the Diels-Alder reaction between dienes and simple inactivated iminium salts, generated in-situ from formaldehyde and
amine hydrochlorides under Mannich-like conditions, proceeds smoothly without catalyst (Scheme 5)26. The scope of
this reaction could be extended to the synthesis of optically
pure heterocycles by employing chiral amines and amino
acids27. Unfortunately, the synthetic applications of the uncatalysed reaction are limited to either formaldehyde or activated aldehydes, higher aldehydes showing much lower reactivity. However, the work of Wang shows that, by using
lanthanideIII triflates a [Ln(OTf)3] as Lewis-acid catalysts, the
yield of the aza Diels-Alder adducts of higher aldehydes and
cyclopentadiene can be increased considerably (Scheme 6).
Moreover, the yield of the cycloadducts of the aza Diels-Alder reaction between L-phenylalanine methyl ester, formaldehyde and several dienes could be increased considerably
when Nd(OTf)3 was used as Lewis acid. However, the reaction of higher aldehydes with dienes other than cyclopentadiene was not successful either in the absence or in the presence
of the lanthanide triflates. The authors did not suggest a
mechanism by which the Lewis acid affects the two-step
process.
4. The aldol reaction
The Lewis-acid-catalysed aldol reaction of silyl enol ethers,
commonly known as the Mukayama aldol reaction, was first
a
trifluoromethanesulfonate, CF3SO3-
Table IX Cross-aldol reaction of various aldehydes.
Recueil des Travaux Chimiques des Pays-Bas, 115 / 11-12, November / December 1996
in high yields. The reactions were most effectively carried out
in a 1/9 commercial aqueous formaldehyde/THF mixture
under the influence of 5-10 mol% of Yb(OTf)3. Some typical
examples are listed in Table VIII. After completion of the
reaction, nearly 100% of the Yb(OTf)3 catalyst is recovered
from the aqueous layer and can be re-used quite easily31.
Various aldehydes could be used to obtain the cross-aldol
product in high yield as a mixture of syn and anti isomers32.
For example, the water-soluble aldehydes, such as acetaldehyde, acrolein and chloroacetaldehyde can be reacted with
silyl enol ethers to give the corresponding cross-aldols. Benzaldehyde, substituted benzaldehydes and higher aldehydes
gave good results in this reaction. Even salicylaldehyde and
2-pyridinecarboxaldehyde have been employed successfully
(Table IX). This is remarkable because the free hydroxyl
group, present in the former compound, and the pyridine
nitrogen, present in the latter, can give problems due to
undesired coordination to the Lewis acid in organic solvents.
The amount of water in the reaction mixture is crucial. The
best results were obtained when the organic solvent contained
10-20% (v/v) of water. When the amount of water increased,
the yield of aldol product decreased. This drop in yield is
caused by the competitive hydrolysis of the silyl enol ether,
which precedes the desired aldol reaction.
5. Allylation reactions
The synthesis of homoallylic alcohols via a Lewis-acid-catalysed reaction of organometallic reagents with a carbonyl
compound has been widely studied33. The first example of
such a Lewis-acid-catalysed allylation reaction in aqueous
medium was again reported by Kobayashi34. In a smooth
reaction under the influence of 5 mol% of Sc(OTf)3, tetraallyltin was reacted with several ketones and aldehydes (Table
X). The reactions were carried out in water/THF,
water/ethanol or water/acetonitrile mixtures (1/9) providing
high yields of the corresponding homoallylic alcohol. Furthermore the reaction of unprotected sugars gave the desired
product directly, again in high yield. Unprotected phenols
such as salicylaldehyde can also be used in this reaction.
Some typical examples are listed in Table X. The much
Table X Allylation of various aldehydes.
461
Scheme 8
Scheme 9
Table XI Indium-promoted allylation.
MLn
Yb(OTf)3
allylSiMe3 /SnCl4
Solvent (time)
H2O (10 h)
DMF/H2O = 4/1 (2 h)
DMF/H2O = 4/1 (1 h)
CH2Cl2, -78°C
9a/9b (yield)
41/59 (66%)
17/83 (82%)
6/94 (88%)
88/12 (76%)
cheaper Yb(OTf)3 is also effective and the Lewis acid can be
re-used without loss of activity35.
6. Barbier-type reactions
The use of indium in aqueous solution has been reported by
Li and co-workers5,36 as a new tool in organometallic chemistry. Indium can be readily used to promote the allylation of
aldehydes and ketones in water at room temperature without
the need for an inert atmosphere (Scheme 8). In fact, protection of hydroxyl groups is not required and acid-sensitive
groups such as an acetal functionality are left intact during
allylation.
The use of a Lewis-acid promotor can have a large effect on
the stereoselectivity of the reaction37. When Yb(OTf)3 was
used as a Lewis acid38 in the allylation of aldehyde 8
(Scheme 9), performed in water using DMF as a cosolvent,
the diastereoselectivity was much higher than in the absence
of Lewis acid (Table XI). Furthermore, the reaction time
could be reduced to one hour. Compared to the reaction under
aprotic conditions, using allyltrimethylsilane instead of allylbromide/indium, almost complete reversal of diastereoselectivity was found.
In the indium-mediated coupling reaction of ethyl (E)-4bromoprop-2-enoate (10) with carbonyl compounds in water,
providing β-hydroxy esters 11 (Scheme 10), the presence of
La(OTf)338 also increases the rate and diastereoselectivity of
the reaction39, as can be seen from Table XII. The authors
explain the high anti selectivity by assuming a cyclic transi-
Scheme 10
Table XII Indium-promoted allylation of benzaldehyde with γ-substituted allyl bromides.
a
Isolated after acylation.
γ-substituent
MLn
Solvent
Anti / syn (yield)
CO2Et
CO2Et
CO2Et
cis-CO2Et
cis-CO2Et
La(OTf)3
La(OTf)3
La(OTf)3
La(OTf)3
H2 O
H2O
DMF
H2O
DMF
86/14 (59%)
90/10 (99%)
86/14 (92%)
99/1 (90%)
86/14 90%
462
J.B.F.N. Engberts, B.L. Feringa et al. / Lewis-acid catalysis in water
Scheme 14
Scheme 15
Scheme 11
Scheme 12
tion state as is shown in Scheme 11. The role of the Lewis
acid has not been further elucidated so far.
7. Mannich-type reaction
For the synthesis of β-amino ketones the Mannich reaction is
one of the most attractive processes. A new type of Lewisacid-catalysed Mannich reaction can be carried out conveniently in aqueous solution. An aldehyde, an amine and a
vinyl ether are reacted in THF/water (9/1) mixtures, in the
presence of 10 mol% of Yb(OTf)3, to give the β-amino ketone
in 55-100% yield (Scheme 12)40. Commercially available
aqueous formaldehyde and chloroacetaldehyde solutions can
be used directly. Phenyloxoacetaat monohydrate, methyl
oxoacetaat, aliphatic and α,β-unsaturated aldehydes also gave
modest to high yields. The reaction is proposed to proceed via
a mechanism involving imine formation and successive addition of a vinyl ether in aqueous solution (Scheme 13). Again,
the catalyst could be recovered after the reaction was completed and re-used with comparable results. The exact role of
the Lewis acid remains unclear.
also successful for other 1,3-diketones, the reaction of β-keto
esters under nearly neutral conditions did not give satisfactory
results43. However, in the presence of a catalytic amount of
Yb(OTf)3 a highly efficient reaction was observed for a large
range of β-keto esters with various β-unsubstituted enones44.
In fact, nearly quantitative yields of 1,4 adducts were obtained
when the reaction was performed in water using 10 mol% of
Lewis acid catalyst (Scheme 15). Results obtained using
several Michael donors are listed in Table XIII. The scope of
the Michael acceptors has been investigated to some extent. In
sharp contrast to the high yields found for α,β-unsaturated
enones bearing different substituents at the α' position, no
reaction was observed between β-keto esters and α,β-unsaturated enones containing substituents at the β position
(Scheme 16). This lack of adduct formation might be atTable XIII Michael addition of various β-ketoesters catalysed by
Yb(OTf)3.
8. Michael additions
Conjugate addition reactions of carbon nucleophiles are among
the most widely used methods for carbon-carbon-bond formation and various stereoselective catalytic 1,4 additions have
been described recently41. The use of water as a solvent in the
Michael addition of 1,3-diketones has been reported occasionally42. When water was used as the solvent for the reaction of
dione 12 with but-3-en-2-one the yield and purity of triketone
13 improved considerably (Scheme 14). This reaction apparently proceeds under slightly acidic conditions due to the
enolate structure of the dione. Although this procedure was
Scheme 13
a
Catalyst Yb(OTf)3, solvent H2O.
Michael adduct.
b
Isolated yields >98%.
c
Mono
Recueil des Travaux Chimiques des Pays-Bas, 115 / 11-12, November / December 1996
5
6
7
8a
b
c
9
b
10 a
b
c
d
11 a
b
12
13
14
15 a
b
16
17
Scheme 16
18
19
tributed to steric hindrance, or an unfavourable equilibrium
resulting mainly in retro-Michael reaction in water.
20
21
b
9. Outlook
c
As a result of the growing need for clean chemistry, in
particular using catalytic processes and less organic solvents,
Lewis-acid catalysis in water will be a field of considerable
interest in the future. Catalysis by soft Lewis acids in water is
generally easier than catalysis by hard catalysts, as encountered in most carbon-carbon bond forming reactions. In the
slip stream of the success of the Ruhr Chemie Rhone Poulenc
process, more breakthroughs in the former field can be expected in the coming years, whereas catalysis by hard Lewis
acids in aqueous media will remain troublesome for the time
being. The use of substrates with non-chelating functionalities
and the exploration of bimetallic Lewis-acid catalysts in
aqueous solution are further challenges for the near future.
d
22
23 a
b
24
25
26 a
b
27 a
b
c
28 a
b
29
Acknowledgements
30
31
We gratefully acknowledge financial support from the Research School “Netherlands Institute for Catalysis Research”
(NIOK).
32 a
b
33
34
References
35
36 a
1
2
b
c
d
3
4
For review see: A. Lubineau, J. Augé and Y. Queneau, Synthesis 741
(1994).
Recent reviews: a P. Kalck and F. Monteil, Adv. Organometal.
Chem. 34, 219 (1992);
W.A. Herrmann and C.W. Kohlpainter, Angew., Int. Ed. Engl. 32,
1524 (1993);
J. Haggins, Chem. Eng. News 28 (1994);
D.M. Roundhill, Adv. Organometal. Chem. 38, 155 (1995).
E.G. Kuntz, Chemtech 570 (1987).
R.W. Hay, “Comprehensive Coordination Chemistry”, Chapter 61.4,
G. Wilkinson, R.D. Gillard and J.A. McCleverty, eds., Pergamon
Press, Oxford, 1987.
b
c
37
38
39
40
41
463
C.J. Li, Chem. Rev. 93, 2023 (1993).
S. Kobayashi, Synlett. 689 (1994).
R. Nakon and P.R. Rechani and R.J. Angelici, J. Am. Chem. Soc.
96, 2117 (1974).
R.G. Pearson, J. Am. Chem. Soc. 85, 3533 (1963);
R.G. Pearson, J. Chem. Educ. 64, 561 (1987);
R.G. Pearson, Coord. Chem. Rev. 100, 403 (1990).
See for example: a P. Laszlo and M. Teston, J. Am. Chem. Soc. 112,
8750 (1990);
Y. Luo and S. W. Benson, Inorg. Chem. 30, 1676 (1991) and references cited therein.
R.S. Drago and B.B. Wayland, J. Am. Chem. Soc. 87, 3571 (1965);
R.S. Drago, Structure and Bonding 15, 73 (1973);
R.S. Drago, J. Chem. Educ. 51, 300 (1974);
R.S. Drago, L.B. Parr and C.S. Chamberlain, J. Am Chem. Soc. 99,
3203 (1977).
M. Sandström, I. Persson and P. Persson, Acta Chem. Scand. 44,
653 (1990);
C. Reichardt, “Solvents and Solvent Effects in Organic Chemistry”,
VCH, Weinheim, 1988.
R.D. Hancock and A.E. Martell, Adv. Inorg. Chem. 42, 89 (1995).
I. Persson, Pure Appl. Chem. 58, 1153 (1986).
S. Ahrland, Pure Appl. Chem. 51, 2019 (1979).
S. Ahrland, Helv. Chim. Acta 50, 306 (1967);
S. Ahrland, Structure and Bonding 15, 167 (1973).
B.J. Takasaki, J.H. Kim, E. Rubin and J. Chin, J. Am. Chem. Soc.
115, 1157 (1993).
J.H. Kim, J. Britten and J. Chin, J. Am. Chem. Soc. 115, 3618
(1993).
For an extensive review of the factors determining the affinity of
ligands for Lewis acids, see: R.D. Hancock and A.E. Martell, Chem.
Rev. 89, 1875 (1989).
F.R. Hartley, C. Burgess, and R.M. Alcock, “Solution Equilibria”,
Ellis Horwood, Chichester, 1980, p 267.
D.C Rideout and R. Breslow, J. Am. Chem. Soc. 102, 7871 (1980).
See: a W. Blokzijl, M.J. Blandamer, and J.B.F.N. Engberts, J. Am.
Chem. Soc. 113, 4241 (1991);
S. Otto, W. Blokzijl and J.B.F.N. Engberts, J. Org. Chem. 59, 5372
(1994);
J.B.F.N. Engberts, Pure Appl. Chem. 67, 823 (1995);
J.W. Wijnen, S. Zavarise and J.B.F.N. Engberts, J. Org. Chem. 61,
2001 (1996) and references therein.
U. Pindur, G. Lutz and C. Otto, Chem. Rev. 93, 741 (1993).
S. Otto and J.B.F.N. Engberts, Tetrahedron Lett. 36, 2645 (1995);
S. Otto, F. Bertoncin and J.B.F.N. Engberts, J. Am. Chem. Soc.,
118, 7702 (1996).
H. Irving and R.J.P. Williams, J. Chem. Soc. 3192 (1953).
L. Yu, D. Chen and P.G. Wang, Tetrahedron Lett. 37, 2169 (1996).
S.D. Larsen and P.A. Grieco, J. Am. Chem. Soc. 107, 1768 (1985);
P.A. Grieco and S.D. Larsen, Org. Synth. 68, 206 (1990).
P.A. Grieco and A. Bahsas, J. Org. Chem. 52, 5746 (1987);
H. Waldman, Angew. Chem., Int. Ed. Engl. 27, 274 (1988);
H. Waldman and M. Braun, Liebigs Ann. Chem. 1045 (1991).
T. Mukaiyama, K. Narasaka and K. Banno, Chem. Lett. 1011
(1973);
T. Mukayama, K. Banno and K. Narasaka, J. Am. Chem. Soc. 96,
7503 (1974).
S. Kobayashi, Chem. Lett. 2087 (1991).
A. Lubineau and E. Meyer, Tetrahedron 44, 6065 (1988).
S. Kobayashi, I. Hachiya and Y. Yamanoi, Bull. Chem. Soc. Jpn. 67,
2342 (1994).
S. Kobayashi and T. Hachiya, Tetrahedron Lett. 33, 1625 (1992);
S. Kobayashi and T. Hachiya, J. Org. Chem. 59, 3590 (1994).
For review see: Y. Yamamoto and N. Asoa, Chem. Rev. 93, 2207
(1993).
I. Hachiya and S. Kobayashi, J. Org. Chem. 58, 6958 (1993).
S. Kobayashi, I. Hachiya and Y. Yamanoi, Bull. Chem. Soc. Jpn. 67,
2342 (1994).
C.J. Li, Tetrahedron Lett. 32, 7017 (1991);
T.H. Chan, C.J. Li, M.C. Lee and Z.Y. Wei, Can. J. Chem 72, 1181
(1994);
P. Cintas, Synlett 1087 (1995).
R. Wang, G.-M. Lim, C.H. Tan, B.-K. Lim, K.-Y. Sim, and T.-P. Loh,
Tetrahedron Asymmetry 6, 1825 (1995).
Stoichiometric amounts of Lewis acid were used in this reaction.
S.-C.H. Diana, K.-Y. Sim and T.-P. Loh, Synlett 263 (1996).
S. Kobayashi and H.J. Ishitani, J. Chem. Soc., Chem. Commun.
1379 (1995).
B.L. Feringa and A.H.M. de Vries, “Advances in Catalytic Processes; Asymmetric Chemical Tranformations” M.P. Doyle. ed., Jai
Press Inc., London, Vol 1, 151 (1995).
464
42 a
b
43
44
45
J.B.F.N. Engberts, B.L. Feringa et al. / Lewis-acid catalysis in water
U. Eder, G. Saber, R. Wiechert, Angew. Chem., Int. Ed. Engl. 10,
469 (1971);
Z.G. Hajos and R.J. Parrish, J. Org. Chem. 39, 1612 (1974).
E. Keller and B.L. Feringa, unpublished results.
E. Keller and B.L. Feringa, Tetrahedron Lett. 37, 1879 (1996).
Y. Marcus, J. Solution Chem. 13, 599 (1984).
46 a
b
47
U. Mayer, V. Gutmann and W. Gerger, Monatshefte Chem. 106,
1235 (1975);
V. Gutmann, “The Donor Acceptor Approach to Molecular Interactions”, Plenum Press, N.Y., 1978.
Y. Marcus, J. Phys. Chem. 91, 4422 (1987).