Chemoselective catalysis with organosoluble lewis acidic polyoxotungstates

Article abstract of DOI:10.1002/chem.201000411

The preparation of new organosoluble Lewis acidic polyoxometalates (POMs) is reported. These complexes were prepared by the incorporation of Zr, Sc, and Y atoms into the corresponding monolacunary Dawson [P₂W ₁₇O₆₁]¹⁰- and Keggin [PW₁₁O ₃₉]^7- polyoxotungstates. The catalytic activity of these compounds was evaluated for C-C bond formation in the DielsAlder, Mannich, and Mukaiyama-type reactions. Comparisons with previously described Lewis acidic POMs are reported. Competitive reactions between imines and aldehydes or between various imines demonstrated that fine tuning of the reactivity could be reached by varying the metal atom in-corporated into the polyanionic framework. A series of experiments that employed pyridine derivatives allowed us to distinguish between the Lewis and induced Bronsted acidity of the POMs. These catalysts activate imines in a Lewis acidic way, whereas aldehydes are activated by indirect Bronsted catalysis.

Full text of DOI:10.1002/chem.201000411

DOI: 10.1002/chem.201000411
Chemoselective Catalysis with Organosoluble Lewis Acidic
Polyoxotungstates
Nathalie Duprꢀ, Pauline Rꢀmy, Kꢀvin Micoine, Cꢀcile Boglio, Serge Thorimbert,*
Emmanuel Lacꢁte,* Bernold Hasenknopf, and Max Malacria[a]
Abstract: The preparation of new orga-
nosoluble Lewis acidic polyoxometa-
lates (POMs) is reported. These com-
plexes were prepared by the incorpora-
tion of Zr, Sc, and Y atoms into the
corresponding monolacunary Dawson
[P2W17O61]10À and Keggin [PW11O39]7À
polyoxotungstates. The catalytic activi-
ty of these compounds was evaluated
reactions. Comparisons with previously
described Lewis acidic POMs are re-
ported. Competitive reactions between
imines and aldehydes or between vari-
ous imines demonstrated that fine
tuning of the reactivity could be
reached by varying the metal atom in-
corporated into the polyanionic frame-
work. A series of experiments that em-
ployed pyridine derivatives allowed us
to distinguish between the Lewis and
induced Brønsted acidity of the POMs.
These catalysts activate imines in a
Lewis acidic way, whereas aldehydes
are activated by indirect Brønsted cat-
alysis.
Keywords: aldol reaction · imines ·
Lewis acids · polyoxometalates
À
for C C bond formation in the Diels–
Alder, Mannich, and Mukaiyama-type
Introduction
grafting Lewis acidic cations onto phosphotungstic back-
bones, which allowed modulation of the catalytic activities.[6]
Besides, owing to the specific solubilities of POMs, the cata-
lysts were easily recovered. Herein, we give an overview of
our contribution to Lewis acidic POM chemistry. Where ap-
propriate, some of our previous results are rediscussed to
give a better understanding of the systems.
Since the beginning of polyoxometalate (POM) chemistry,
catalysis and POMs have been closely associated. This con-
nection is largely due to the chemical properties of those
early transition-metal–oxo complexes, that is, most common-
ly VV, MoVI, and WVI centers, which are electron acceptors
or strong Brønsted acids (in their protonated form).[1] POMs
can also form peroxo complexes and can support highly oxi-
dized metal ions or other catalytically active organometallic
groups, which increase their appeal for synthetic chemists.[2]
Yet, because of these properties, catalysis by POMs[3] has re-
mained mostly confined to oxidation and acid-catalyzed
transformations.[4]
Results and Discussion
Choice of the catalysts: We surmised that the charges and
ionic radii of the cations grafted onto the polyoxotungstic
backbone would play a primary role in the reactivity of the
POMs. Therefore, we selected a large array of Lewis acidic
cations: La3+, Eu3+, Sm3+, and Yb3+ ions were chosen as
early-, mid-, and late-lanthanide centers. The Sc3+ and Y3+
ions were logical variations outside the lanthanide group,
and Zr4+ and Hf4+ ions were chosen as representatives of
more charged and smaller Lewis acidic cations.
We have introduced Lewis acidic POMs that have opened
new options.[5] Lewis acidity was introduced by modular
[a] N. Duprꢀ, Dr. P. Rꢀmy, Dr. K. Micoine, Dr. C. Boglio,
Dr. S. Thorimbert, Dr. E. Lacꢁte, Prof. B. Hasenknopf,
Prof. M. Malacria
Because of our previous experiences with the Dawson
and Keggin polyoxotungstates,[7] we selected the a1-Dawson
lacunary platform to graft the cations onto. The inherent
chirality of this platform leaves the door open for potential
asymmetric catalysis, should there be a way to obtain an
enantiopure form.[8] Besides, the catalytically active moiety
would be embedded in the belt of the POM, thus resulting
UPMC Univ Paris 06
Institut Parisien de Chimie Molꢀculaire (UMR 7201)
C. 229, 4 place Jussieu, 75005 Paris (France)
Fax : (+33)144-27-73-60
E-mail: serge.thorimbert@upmc.fr
emmanuel.lacote@upmc.fr
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201000411.
7256
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7256 – 7264

FULL PAPER
in lower proportions of sandwich POM structures if Lewis
acidic atoms with high coordination numbers, such as lan-
thanide atoms, were used.[9,10] The reactivity of the catalyst
is certainly governed by the available coordination sites on
the metal center. To get an idea of the role of the charge of
the heteropolyanion, we also considered the Keggin lacuna-
ry phosphotungstate as a suitable platform.
The syntheses of the organo-
Polyoxometalate speciation: Larger cations have high co-
ordination numbers (typically 8 and 9). Several equilibria
occur in solution (Scheme 1). In particular, sandwich 1:2 and
dimeric 2:2 structures would preclude or decrease catalytic
activity—relative to the 1:1 complex—because they have
none or less free sites for complexation of the substrates.
Thus, the relative ratios of 1:1 versus 1:2/2:2 species should
soluble Lewis acids complexes
have been reported for lantha-
nide and hafnium cations. Simi-
lar methodologies were used
for Zr4+, Sc3+, and Y3+ ions
(see the Supporting Informa-
tion). For lanthanide POM de-
rivatives, full control of the
pH value has to be taken:
under no circumstances it
Scheme 1. Speciation equilibria of Lewis acidic polyoxotungstates.
should exceed pH 4.5 to avoid
the presence of lacunary spe-
cies. The addition of aqueous HCl (1m) was used to main-
tain the pH value of the reaction mixture under the accepta-
ble limit. The addition of excess tetrabutylammonium bro-
mide (TBABr) causes the precipitation of the TBA salts of
the POM derivatives. These complexes generally precipitate
with four water molecules bound to the metal center.
strongly impact the catalytic activity. The speciation depends
on the relative steric hindrance and electronic interactions
of the polyanionic inorganic ligands. This consideration di-
rected us to consider the a1 lacunary platform, which was
known in the lanthanide series to relatively favor 1:1 com-
plexes over sandwich 1:2 in aqueous solution.[9] In this con-
text, it is also important to note that ligands that can bind to
a POM/lanthanide complex displace the equilibria towards
the 1:1 species.[9d]
Factors that govern the reactivity of the Lewis acidic POMs:
It was expected that the polyoxometallic backbone would
play a significant role in the reactivity. We identified three
essential factors in understanding and/or predicting the be-
havior of the POMs.
Application of imines in the catalysis: We focused on the
Lewis acid-catalyzed Mannich reaction between aldimines
and silyl enol ethers as a representative benchmark reaction
for our catalysts. Acetonitrile was chosen because the TBA
salts of our complexes are soluble in polar solvents such as
acetonitrile, DMF, or dimethyl sulfoxide (DMSO). A typical
reaction was carried out in acetonitrile (0.2m) at room tem-
perature with benzylideneaniline as the electrophile, a tri-
methylsilyl enol ether derived from propiophenone, and the
substituted POM complex (Table 1). Catalyst POM/La gave
b-amino ketone 1a in 81% yield after two days. As is often
the case with lanthanide Lewis acids, the relative configura-
tion of the ketoimine was not controlled (less than 30% de).
We did not try to improve the diastereoselectivity, but con-
centrated on the efficiency of the reaction.
Charge: The inorganic ligand is anionic and loss of activity
of the cation was foreseen. Modulation could arise from
tuning of the overall cluster charge, which could be adjusted
through synthesis upon appropriate choice of both the poly-
oxotungstic framework and the embedded Lewis acidic
cation.
Ionic radii of the cations: The ionic radii of the cations that
we use vary greatly, from Sc3+ to La3+ ions.[11] Therefore,
the cation is more or less deeply inserted in the lacuna. As a
consequence, the stereoelectronic influence of the polyoxo-
metallic backbone should be more or less pronounced. In
particular, smaller cations would be less accessible than
larger ones, and thus less reactive. Additional bonding of
the metal cation to an oxygen atom of the closest phosphate
group may develop with small cations that fit better into the
lacuna. Larger cations cannot reach deeply enough to allow
this bonding. Conceivably, because the POM charge is
denser on the inside,[12] such a bonding pattern should trans-
fer additional electron density to the cation and further de-
crease its Lewis acidity. On the other hand, smaller and/or
more charged cations have higher intrinsic charge densities,
which should increase their acidity. These conflicting trends
should allow precise fine-tuning of the reactivities.[13]
The reactions were catalyzed by all the a1-[MP2W17O61]nÀ
complexes (Table 1, entries 1–8), thus confirming our initial
hypothesis that Lewis acidic POMs can be obtained through
metal substitution. When the lanthanide complexes were
used, the reactions took two days to reach completion
(Table 1, entries 1–4). The same reaction with YbACHTUNGTRENNUNG(OTf)3
(OTf=triflate) was over in only two hours. This outcome
was expected (see above) because the oxo ligands of the la-
cunary polyanion decrease the overall Lewis acidity of the
embedded cations. The activity of the yttrium complex was
slightly higher than that of the lanthanide complexes (one
day to completion; Table 1, entry 6), whereas the activity of
Chem. Eur. J. 2010, 16, 7256 – 7264
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7257

S. Thorimbert, E. Lacꢁte et al.
neutral inorganic cages with a charged core.[12a] From this
perspective, there is no real difference between the Keggin
and Dawson series.
Table 1. Evaluation of the catalysts in a Lewis acidic POM-catalyzed
Mannich reaction.
The small changes observed for the scandium complexes
may be due to the different geometries and the increased
bonding with the internal phosphate group in the Keggin
complex relative to the Dawson complex. One should
remain cautious in interpreting these results. Some elements
of rationalization have been given, but further discussion of
the mechanism is developed later. Nonetheless, it was clear
at this stage that polyoxometalates could be made Lewis
acidic and that subtle variations of the reactivities stemmed
from variations not only of the metal cation, but also of the
polyoxotungstic backbone.
Entry Catalyst
(abbreviation) Yield
[%]
1
2
3
4
5
TBA5H
TBA5H
TBA5H
TBA5H
2A
2A[a1-Sm
2A[a1-Eu(H2O)4P2W17O61]
2A[a1-Yb(H2O)4P2W17O61]
0.5A[a1-Sc-
E
A
81[a]
91[a]
71[a]
96[a]
66[b,c]
R
ACHTUNGTRENNUNG
N
G
ACHTUNGTRENNUNG
R
G
ACHTUNGTRENNUNG
TBA5.5K0.5Na0.5
(H2O)4P2W17O61]
TBA4.5K0.5 2A[a1-Y
[a1-Hf(H2O)4P2W17O61]
[a1-Zr(H2O)4P2W17O61]
(H2O)4PW11O39]
(H2O)PW11O39]a
H
G
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
6
7
8
H
C
(H2O)4P2W17O61]
A
85[b]
85[a]
TBA5K
TBA5K
U
G
ACHTUNGTRENNUNG
Various imines were tested to determine the scope of the
reaction (Table 2). The reactivity generally reflected the
electronic characteristics of the imines. Electron-donating
A
ACHTUNGTRENNUNG
84[b]
<5[d]
55[b]
9
10
TBA
TBA
CHTUNGTRENNUNG[a-YbACHTUNGTRENNUNG
C
[a] See ref. [5a]. [b] This study. [c] Six days to reach completion. [d] Di-
meric in the solid state.
Table 2. Variation of the imine in a Lewis acidic POM-catalyzed Man-
nich Reaction.
the scandium complex was much lower, thereby producing
1a in only 66% yield after six days (Table 1, entry 5). The
Sc3+ ion is much smaller than the Yb3+ ion; therefore,
deeper embedding of the cation in the lacuna and the de-
creased number of available coordination sites are the pre-
dominant factors that influence the reactivity. As anticipat-
ed, the hafnium and zirconium complexes, which include
cations with a formal charge of 4+, seem more reactive
than the corresponding lanthanide complexes (Table 1, en-
tries 7 and 8; 85 and 84% yield of the isolated products
after less than a day, respectively).
Entry
R
Ar
Catalyst
Product
Yield [%]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Ph
Ph
Ph
Ph
Ph
POM/Yb
POM/Yb
POM/Hf
POM/Yb
POM/Yb
POM/Yb
POM/Hf
POM/Hf
POM/Zr
POM/Y
POM/Yb
POM/Hf
POM/Hf
POM/Zr
POM/Zr
1a
1b
1b
1c
1d
1d
1d
1d
1d
1e
1e
1e
1e
1e
1e
96[a]
84[a]
75
o-HOC6H4
o-HOC6H4
p-MeOC6H4
Ph
Ph
Ph
81[a]
70[a]
71[b]
96[c]
81[b]
91
CO2Et
CO2Et
CO2Et
CO2Et
CO2Et
CO2Et
CO2Et
CO2Et
CO2Et
CO2Et
CO2Et
Ph
Ph
The Keggin series behaved differently (Table 1, entries 9
and 10). First of all, less than 5% conversion was observed
after two days with the Yb/Keggin complex (Table 1,
entry 9). At first glance, this outcome was surprising because
the lacunary Keggin platform is less charged (4À) and might
cause a smaller decrease in the Lewis acidic character of the
Yb substituent than the lacunary Dawson ion (7À). We at-
tributed this behavior to the persistence of the dimeric 2:2
form in solution, with a resulting decrease in the amount of
active 1:1 Yb-substituted Keggin complex. Indeed, X-ray
crystallographic studies showed that this complex is a 2:2
p-MeOC6H4
p-MeOC6H4
p-MeOC6H4
p-MeOC6H4
p-MeOC6H4
p-MeOC6H4
87
77[a]
99[c]
77[b]
91
64[b]
[a] See ref. [5a]. [b] Catalyst loading: 2 mol%. [c] See ref. [5b].
groups on the nitrogen atom decreased the rates of the reac-
tions (Table 2, entries 2–4). The reaction with para-methoxy-
phenylimine was really sluggish (6 days to reach completion;
Table 2, entry 4), less so with the ortho-hydroxyphenylimine
(3 days to reach completion; Table 2, entry 2). On the con-
trary, electron-withdrawing groups attached to the carbon
atom of the imines, such as the glyoxalate derivatives, led to
extremely fast reactions (around 1 h), even when the donor
groups were simultaneously introduced on the nitrogen
atom (Table 2, entries 5–14). This result prompted us to de-
crease the catalyst loading. Gratifyingly, the reactions of the
aniline-derived glyoxalate imine still worked with 2 mol%
of catalyst, although the reactions took up to three days to
reach completion (Table 2, entries 6, 8, 13, and 15). The Zr
and Y complexes compare well with the Hf and Yb catalysts
(Table 2, entries 10–14).
(m-H2O)2]8À dimer in the solid state,
[{YbACHTUNGTRENNUNG(H2O)ACHTUNGTRENNUNG(PW11O39)}2ACHTUNGTRENNGUN
in which two {YbPW11O39}4À units are bridged by two water
molecules.
The Sc/Keggin POM led to 1a in 55% yield after two
days (Table 1, entry 10), which is similar to the yield of the
parent Sc/Dawson complex. We think this finding may also
be explained by the POM speciation. Contrary to the Yb3+
ion, the Sc3+ ion can only have six ligands in its coordina-
tion sphere. Thus, only 1:1 complexes exist in solution, re-
gardless of the POM structure. We can gather from this set
of reactions that the POM charge has no strong influence
on reactivity. This conclusion matches the one that might
have been reached upon considering the model developed
by Day and Klemperer of the heteropolyoxotungstates as
7258
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7256 – 7264

À
C C Bond Formation with Lewis Acidic Polyoxotungstates
FULL PAPER
Variation of the nucleophile involved in the reaction was
examined next (Table 3). A simple silyl enol ether derived
from acetophenone worked equally well in less than one
POM catalysts, the Danishefsky diene typically delivered
the desired heterocycle in good yield (Table 4, entries 1–4).
The reaction was very versatile, although the Y complex
was surprisingly, yet consistently, less efficient in that case
(e.g., 2a in 67% yield after 24 h versus 88% yield in 5 h for
the Zr complex; Table 4, entries 3 and 4). Again, we seemed
to be facing very minute modulations of the reactivities of
the POMs. As previously noted, the pyruvate derivative re-
acted faster, but the yield was poor, most probably because
of the imine self-condensation (Table 4, entry 14).
We investigated the aza-Diels–Alder reaction involving
the imine as the diene and a cyclic enol ether as the dieno-
phile (Table 5). The tricyclic products that can be obtained
thanks to this reaction have three contiguous stereogenic
centers. We hoped that the increased strain during the for-
mation of cyclic molecules would lead to more diastereose-
lective reactions.
Table 3. Variation of the nucleophile in a Lewis acidic POM-catalyzed
Mannich reaction.
Entry Electrophile
Nucleophile Product
Catalyst Yield
[%]
POM/
98
1
2
Yb
POM/
Hf
92
94
3
POM/
Zr
POM/
Hf
POM/
Zr
4
5
70
63
POM/
Yb
POM/
Hf
6
7
98
90
Table 4. Aza-Diels–Alder reactions with the Danishefsky diene.
POM/
Hf
POM/
Zr
8
9
84
89
Entry
R
Ar
Catalyst
Product
Yield [%]
POM/
Hf
POM/
Zr
89[a]
84[b]
88
10
11
76
77
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
CO2Et
Ph
Ph
Ph
Ph
POM/Yb
POM/Hf
POM/Zr
POM/Y
POM/Yb
POM/Hf
POM/Zr
POM/Y
POM/Yb
POM/Hf
POM/Y
POM/Yb
POM/Y
POM/Yb
2a
2a
2a
2a
2b
2b
2b
2b
2c
2c
2c
2d
2d
2e
67
98[a]
90[b]
90
POM/
Hf
POM/
Zr
p-MeOC6H4
p-MeOC6H4
p-MeOC6H4
p-MeOC6H4
p-ClC6H4
p-ClC6H4
p-ClC6H4
o-HOC6H4
o-HOC6H4
p-MeOC6H4
12
13
70
58
81
76[a]
69[b]
68
57[a]
44
hour, whether with diphenylimine (Table 3, entries 1–3) or
the pyruvate derivative (Table 3, entries 4 and 5). The reac-
tions could be extended to thioester-derived silyl enol esters
with the same efficiency (Table 3, entries 6 and 7). The com-
mercially available (furan-2-
23
[a] See ref. [5a]. [b] See ref. [5b].
yloxy)trimethylsilane allowed
Table 5. Aza-Diels–Alder reaction using an imine as the diene.
vinylogous Mannich reactions
in less than 12 hours (Table 3,
entries 8–13). All the reactions
that led to products 1 f–k pro-
ceeded with rates and yields
comparable to those observed
for the preceding reactions and
seemed to be more sensitive to
the nature of the electrophile
and the catalyst used rather
than the nucleophile.
Entry
R
Ar
n
Catalyst
3
Yield [%][a]
d.r.[a]
1
2
3
4
5
6
7
8
9
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
CO2Et
CO2Et
CO2Et
CO2Et
Ph
Ph
Ph
Ph
1
1
2
2
1
1
1
2
1
2
1
2
POM/Yb
POM/Hf
POM/Yb
POM/Hf
POM/Yb
POM/Y
POM/Hf
POM/Yb
POM/Yb
POM/Yb
POM/Yb
POM/Yb
3a
3a
3b
3b
3c
3c
3c
3d
3e
3 f
3g
3h
67 (80)
48
60:40 (50:50)[b]
60:40
79 (89)
80:20 (50:50)[b]
–
0 [c]
o-HOC6H4
o-HOC6H4
o-HOC6H4
o-HOC6H4
Ph
63 [d] (74)
48 [d]
90:10 (55:45)[b]
80:20
87 [d]
60:40
Encouraged by these results,
we looked at other reactions of
imines (Table 4). The hetero-
39 (64)
39
<40
67 (48)
62 (<20)
80:20 (50:50)[b]
50:50[b]
10
11
12
Ph
50:50[b]
p-MeOC6H4
p-MeOC6H4
50:50 (50:50)[b]
50:50 (50:50)[b]
Diels—Alder reaction was
a
logical extension to our investi-
gation. Treated with diphenyli-
[a] Values in brackets correspond to the reaction with YbACTHNUGTRENUNG(OTf)3 as the catalyst. [b] See ref. [5a]. [c] Only deg-
mine in the presence of the radation was observed. [d] Reaction stopped after 24 hours.
Chem. Eur. J. 2010, 16, 7256 – 7264
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7259

S. Thorimbert, E. Lacꢁte et al.
Table 6. POM-catalyzed Mukaiyama–aldol reaction
The reactions between various imines and either dihydro-
furan or dihydropyran were efficiently catalyzed by
10 mol% of our Yb or Y complexes and afforded the ex-
pected cyclic adducts in acceptable-to-good yields. The
POM/Hf complex was too harsh, thus leading sometimes to
sharp drops in the yield (Table 5, entries 1–4). Highly acti-
vated aryl imines also gave moderate yields (Table 5, en-
tries 9–12). The more reactive the imine, the more degrada-
tion we observed (Table 5, compare entries 9 and 10 to en-
tries 11 and 12). However, the more active metal triflate
salts gave even lower yields or no products at all. Hence,
lowering the activity of the catalyst can prevent degradation
of the substrates.
Parallel reactions were set up to compare the activities of
the POMs for less sensitive ortho-hydroxo substrates
(Table 5, entries 5–7). All three reactions were stopped after
24 hours: the POM/Hf-catalyzed reaction reached comple-
tion (87% yield), but the conversions for the POM/Yb and
POM/Y catalysts (63 and 48% yield of the isolated prod-
ucts, respectively) were not complete. For this reaction and
substrate, POM/Y was less active than POM/Yb and POM/
Hf, which shows that we should remain cautious in issuing
absolute ranking of the reactivities of the POM complexes
(see also below).
Entry Ar
Nucleophile Catalyst
[equiv]
t
Product Yield Conv.
U
[%]
[%]
1
2
3
4
5
6
Ph
Ph
Ph
Ph
1.5
3
POM/
Hf
POM/
Hf
POM/
Zr
POM/
Zr
POM/
Hf
POM/
Zr
1 day 4a
7 h 4a
85
100
n.d.
n.d.
92
98
78
3
3 day 4a
1 day 4a
12 h 4b
12 h 4b
5
100
>95
96
p-
2
95
NO2C6H4
p-
NO2C6H4
2
89
Finally, the diastereoselectivities remained poor for all the
reactions. The use of POMs suffers some limitations as well.
In particular, no reaction has been observed so far for the
allylation of benzaldehyde, whether with allylsilanes or with
allylstannanes.
Interestingly, a sharp increase in the diastereoselectivity
(up to 90:10 d.r.) was observed when using TBA5H2[a1-
YbP2W17O61] and an ortho-hydroxy substrate (Table 5,
entry 5). The same reaction with the Yb triflate compound
was not diastereoselective. This trend was more or less ob-
served for all the ortho-hydroxy substrates and might be due
to hydrogen-bonding of the latter to the polyoxotungstic
backbone. This diasteroselectivity boost is an additional im-
portant specific feature of the POMs with regard to tradi-
tional Lewis acidic triflates, at least for Yb ions and to some
extent Y ions.
Application of aminals in the catalysis: Aminals reacted
very rapidly with silyl enol ether under POM catalysis, and
acceptable yields of the desired products were obtained
(Table 7). The reaction times (1–4 h) were generally shorter
than those of the Mannich reactions; however, the scandium
complex was an exception that required 3 days to give 79%
yield (40% yield after 4 h). As before, the diastereoselectivi-
ties were poor. A related hemiketal did not react at all and
was recovered after several days.
Chemoselectivity: To test the scope of our catalysts further,
we extended the Mannich reaction to its three-component
version (Scheme 2). Direct mixing of benzaldehyde, aniline,
and the silyl enol ether in acetonitrile in the presence of the
catalyst delivered aminoketone 1a in high yields. Water is
released upon formation of the imine, thus confirming again
the water stability of our complexes.[5b]
Interestingly, benzaldehyde did not react in the presence
of the POM/Hf in this case due, certainly, to the very rapid
formation of the imine, which consumed the aldehyde too
Application of aldehydes in the catalysis: For the Mukaiya-
ma–aldol reaction, benzaldehyde was chosen as a represen-
tative aldehyde (Table 6). It soon appeared that no reaction
took place when POM/Yb or standard heteropolyacids
(HPAs: H3PW12O40 and H4SiW12O40) were used.[14] On the
contrary, both POM/Hf and POM/Zr gave good conversions
and yields of the desired Mukaiyama–aldol products. This
outcome supported our initial assumption that metallic cat-
ions with a higher charge would lead to increased Lewis
acidity.
The POMs are air and water stable, which is an improve-
ment over the standard Hfiv and ZrIV Lewis acids generally
used in catalysis.[15] The amount of silyl enol ether strongly
influenced the efficiency of the reactions (Table 6, entries 1–
4). Indeed, the hydrated water molecules of the POM slowly
hydrolyzed the silyl enol ether. The use of an excess of nu-
cleophile (3–5 equiv) was the best compromise found. It al-
lowed fast enough reactions and reproducible results inde-
pendent of the POM dryness. Activated electrophiles led to
good yields with both POM/Hf and POM/Zr (Table 6, en-
tries 5 and 6), but aliphatic aldehydes gave low conversions.
Table 7. Reaction of aminals.
Entry
Catalyst
Yield [%]
1
2
3
4
5
POM/Yb
POM/Sc
POM/Y
POM/Hf
POM/Zr
64
79
82
78
73
7260
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7256 – 7264

À
C C Bond Formation with Lewis Acidic Polyoxotungstates
FULL PAPER
imine (see the Supporting Information). The reactions with
all the traditional Lewis acids tested were already quite che-
moselective. Nonetheless, the POMs compared very well to
more common catalysts. The yields were generally higher
and the reactions equally or more selective, especially for
the POM/Zr catalyst.
Scheme 2. Three-component Mannich reaction.
Catalyst recovery and recycling: As stated in the introduc-
tion, POMs have very narrow solubility properties, which
opened the way to catalyst recycling. The addition of a mix-
ture of acetone/ethanol/diethyl ether (1:1:20) precipitated
the POM, while keeping all the organic species in solution.
Centrifugation allowed complete recovery of the catalysts.
Given the molecular nature of the recovered material, its
purity could be assessed by IR and 31P NMR spectroscopic
analysis. The Mannich reaction could be repeated up to ten
times with no loss in yield with the POM/Yb catalyst.
We have already shown that no leaching from the POM/
Yb complex took place. Previous work indicated that the
complexation constants for the Ln complexes of POMs in
water are extremely high[9c] and are presumably higher in
less polar organic solvents.
In this study, no trace of lacunary POM was found in the
precipitated solids. Besides, the Yb3+ ion is paramagnetic
and the leached material would influence the NMR spectra
of the organic phase. As no paramagnetic shift or line
broadening was observed, it was deduced that no free Yb3+
ions remained in solution. This reasoning was further sup-
ported by the fact that the filtrate showed no catalytic activi-
ty. Thus, we excluded catalysis by decomplexed lanthanide
ions. No evidence for dissociation of the other POM com-
plexes could be found either.
quickly for the catalyst to compete. Nonetheless, this attract-
ed our attention to the potential offered by our catalysts for
chemoselective reactions. We reasoned that their lower ac-
tivity could be turned into an advantage because they would
not catalyze reactions with higher activation energies.
Weaker Lewis acids might also coordinate better Lewis
bases with increased selectivity.
First, we focused on the imine/aldehyde competition. In a
typical experiment, one equivalent of silyl enol ether was
treated simultaneously with one equivalent of diphenylimine
and one equivalent of benzaldehyde in the presence of a
catalytic amount of the Lewis acid in acetonitrile at room
temperature. The standard commercial Lewis and Brønsted
acids led to varied results (see the Supporting Informa-
tion).[16] To our pleasure, all the POM-based Lewis acids
were extremely chemoselective (Table 8). The POM com-
plexes were superior both for the metal atoms that generally
favor addition onto imines, in which case they led to in-
creased chemoselectivities, and for the more oxophilic
metals, in which case the chemoselectivities were reversed.
The complex POM/Sc proved again to be less reactive
(Table 8, entry 5).
We turned our attention to the competition between the
imines. We started with two diarylimines that differed only
in the electron density on the nitrogen atom: diphenylimine
and benzylidene(4-methoxyphenyl)amine (see the Support-
ing Information). However, the chemoselectivities were
modest in favor of diphenylimine with the POM Lewis acids
but did not change significantly with the traditional Lewis
acids. Thus, we examined more electronically different
imines, for example, a diarylimine and a pyruvate-derived
Mechanistical considerations: We wished to obtain a clear
picture of the exact mechanism of our reactions. We initially
excluded direct Brønsted catalysis by protons around the
POM. Indeed, even if the lanthanide complexes featured
two protons (e.g., TBA5H2[a1-YbP2W17O61]), the lacunary
Dawson (TBA)6H4[a1-P2W17O61] did not catalyze the Man-
nich reactions and the Hf and Zr complexes have no proton
available (e.g., TBA5K[a1-MP2W17O61]). We also excluded
catalysis by the decomplexed metal cations for the same rea-
sons as above. In the lanthanide series, the complexation
constants in water are already extremely high and are pre-
sumably higher in less polar organic solvents.[9c]
Table 8. Imine/aldehyde competition with Lewic acidic POMs.
Two possibilities should be considered. The first obvious
option was the traditional Lewis acid activation of the sub-
strates at the metal cornerstone of the POM (Scheme 3a).
In the second path, a water molecule might bind to the
metal center and become activated toward deprotonation
(Scheme 3b).[17] We label this pathway as indirect Brønsted
catalysis because it is triggered by the initial coordination of
water to the Lewis acidic POM. Indeed, thermogravimetric
analysis (TGA) shows water associated to the POMs, so
water molecules are always present in the catalytic reaction
mixtures.
Entry
Catalyst
1a/4a
Yield [%]
1
2
3
4
5
6
7
8
POM/La[a]
POM/Sm[a]
POM/Eu[a]
POM/Yb[a]
POM/Sc
100:0
100:0
100:0
100:0
100:0
98:2
82
97
96
97
55
84
86
87
POM/Hf[b]
POM/Zr
POM/Y
100:0
100:0
[a] See ref. [5a]. [b] See ref. [5b].
Chem. Eur. J. 2010, 16, 7256 – 7264
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7261

S. Thorimbert, E. Lacꢁte et al.
Table 9. Inhibition of the Mannich reaction.
Entry
Catalyst
Additive
Time
[days]
Yield
[%]
AHCTUNGTRENNUNG
1
2
3
4
5
6
7
8
9
POM/Yb
POM/Yb
POM/Yb
POM/Yb
POM/Hf
POM/Hf
POM/Hf
POM/Hf
POM/Zr
POM/Zr
POM/Zr
–
2
2
2
8
1
1
1
5
1
1
1
78
77
0
2,6-di-tert-butylpyridine
pyridine
pyridine
–
2,6-di-tert-butylpyridine
pyridine
pyridine
–
16
92
94
37
95
73
68
14
Scheme 3. Possible pathways for catalysis by Lewis acidic POMs.
10
11
2,6-di-tert-butylpyridine
pyridine
In both cases, catalysis would originate from Lewis acidity.
Yet, discrimination between the two hypotheses was re-
quired to predict reactivities and devise potential further ap-
plications in asymmetric catalysis.
Mizuno and co-workers reported a way of discriminating
between Lewis and Brønsted acid catalysis by protonated
Lewis acidic g-Keggin dialuminum- and dihafnium-substitut-
ed silicotungstates.[5c,d] It appeared to us that such a method
could also apply to the problem in hand, even if no protons
of the POMs were considered. Thus, we ran two sets of reac-
tion in the presence or absence of pyridine and 2,6-di-tert-
butylpyridine. Both bases are strong enough to capture pro-
tons and should inhibit any Brønsted acid-catalyzed reac-
tion. On the other hand, pyridine also binds to metal cen-
ters, thus inhibiting Lewis acid-catalyzed reactions. Howev-
er, 2,6-di-tert-butylpyridine is highly hindered and should co-
ordinate less to the Lewis acid corners located in our bulky
POMs, and its effect on Lewis acid catalysis is expected to
be weak.[18]
dine. This outcome supports our conclusion that dimeriza-
tion of the HfIV and ZrIV a1-Dawson polyoxometalates
under our reaction conditions is negligible and that the reac-
tions are blocked by the additives not by an increase in
pH value.
The Mukaiyama–aldol reaction was examined (Table 10),
and again the substrate (benzaldehyde) was mixed with
enough silyl enol ether (5 equiv) to ensure smooth reactions.
The POM/Yb complex was omitted because it did not cata-
Table 10. Inhibition of the Mukaiyama–aldol reaction.
To test the Mannich reaction, we selected diphenylamine,
which was treated with an excess of silyl enol ether
(5 equiv) in the presence of 20 mol% of several POM cata-
lysts with or without the pyridine derivatives (Table 9). The
excess silyl enol ether ensured a rapid enough conversion
and avoided artifacts that arise from hydrolysis of the re-
agents. The reactions were stopped after one (Hf, Zr) or
two days (Yb). The 2,6-di-tert-butylpyridine did not influ-
ence any of the reaction outcomes, regardless of the metal
center (Table 9, entries 2, 6, and 10). From the opposite
view, pyridine significantly decreased the overall rate of all
the reactions (Table 9, entries 3, 7, and 11), but without to-
tally inhibiting the catalyst (Table 9, entries 4 and 8). Thus,
the Mannich reaction proceeds through substrate complexa-
tion.
Entry
Catalyst
Additive
Time
[days]
Yield
[%]
AHCTUNGTRENNUNG
1
2
3
4
5
6
POM/Hf
POM/Hf
POM/Hf
POM/Zr
POM/Zr
POM/Zr
–
1
1
1
1
1
1
100
6
5
92
0
16
2,6-di-tert-butylpyridine
pyridine
–
2,6-di-tert-butylpyridine
pyridine
lyze the reactions with aldehydes. Nearly complete inhibi-
tion of the reaction was observed in all the cases in which
an additive was included. This finding tends to prove that
the reactions follow the indirect pathway and is in good
agreement with our DFT calculations, which indicated an
easy deprotonation of a complexed water molecule in the
POM/Hf complexes, but not in the POM/Yb catalysts.[20]
Thus, we conclude that both the POM/Hf and POM/Zr com-
plexes act as indirect Brønsted catalysts in the aldol reac-
tion.
It has been reported that a2-[HfP2W17O61]6À dimerizes
with increasing pH value.[19] As stated above, such dimers
are expected to be less active Lewis acids because of the
fewer available coordination sites. In our case, the addition
of the basic additive 2,6-di-tert-butylpyridine did not impact
reactivity. Furthermore, in the case of the Hf catalyst,
31P NMR spectra and mass spectra of the resulting solution
were not modified upon the addition of 2,6-di-tert-butylpyri-
To sum up, we propose that all the Mannich reactions
proceed through the direct Lewis acid pathway, regardless
of the POM complexes used, whereas the POM/Hf and
7262
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7256 – 7264

À
C C Bond Formation with Lewis Acidic Polyoxotungstates
FULL PAPER
Nsouli, M. Carraro, A. Sartorel, G. Scorrano, H. Oelrich, L. Walder,
M. Bonchio, U. Kortz, Inorg. Chem. 2010, 49, 7–9.
[3] See a recent special issue of J. Mol. Catal. A 2007, 262
POM/Zr complexes catalyze the Mukaiyama–aldol reactions
through an indirect Brønsted mechanism. This mechanistic
difference may also account for the high chemoselectivity,
which constitutes the most important feature of our Lewis
acidic POMs.
[4] For recent use as Brønsted acids, see: a) S. Luo, J. Li, H. Xu, L.
Zhang, J.-P. Cheng, Org. Lett. 2007, 9, 3675–3678; b) J. Li, S. Hu, S.
Luo, J.-P. Cheng, Eur. J. Org. Chem. 2009, 132–140. For recent use
in oxidation see: c) A. E. Kuznetsov, Y. V. Geletii, C. L. Hill, K. Mo-
rokuma, D. G. Musaev, Inorg. Chem. 2009, 48, 1871–1878; d) A. Sar-
torel, M. Carraro, A. Bagno, G. Scorrano, M. Bonchio, Angew.
Chem. 2007, 119, 3319–3322; Angew. Chem. Int. Ed. 2007, 46, 3255–
3258; e) A. Sartorel, P. Miro, E. Salvadori, S. Romain, M. Carraro,
G. Scorrano, M. Di Valentin, A. Llobet, C. Bo, M. Bonchio, J. Am.
Chem. Soc. 2009, 131, 16051–16053.
[5] a) C. Boglio, G. Lemiꢃre, B. Hasenknopf, S. Thorimbert, E. Lacꢁte,
M. Malacria, Angew. Chem. 2006, 118, 3402–3405; Angew. Chem.
Int. Ed. 2006, 45, 3324–3327; b) C. Boglio, K. Micoine, P. Rꢀmy, B.
Hasenknopf, S. Thorimbert, E. Lacꢁte, M. Malacria, C. Afonso, J.-C.
Tabet, Chem. Eur. J. 2007, 13, 5426–5432; c) Y. Kikukawa, S. Yama-
guchi, Y. Nakagawa, K. Uehara, S. Uchida, K. Yamaguchi, N.
Mizuno, J. Am. Chem. Soc. 2008, 130, 15872–15878; d) Y. Kikuka-
wa, S. Yamaguchi, K. Tsuchida, Y. Nakagawa, K. Uehara, K. Yama-
guchi, N. Mizuno, J. Am. Chem. Soc. 2008, 130, 5472–5478; Lewis
acid properties are not limited to catalysis: a recent report examined
Lewis acid–base equilibria involving tin-substituted polyoxotungs-
tates, see: e) I. Bar-Nahum, J. Ettedgui, L. Konstantinovski, V.
Kogan, R. Neumann, Inorg. Chem. 2007, 46, 5798–5804.
Conclusion
Organosoluble a1-substituted Dawson polyoxotungstates in-
corporating Ln, Sc, Y, Hf, or Zr metal cations are air- and
water-stable Lewis acids and were used in several organic
reactions. Grafting Lewis acidity onto the inorganic back-
bone of POMs led to catalysts with high chemoselectivities,
thus favoring activation of imines over aldehydes. The poly-
anionic framework also enhanced diastereoselectivities in
some cases, thus indicating that secondary interactions be-
tween the substrates and the POM skeleton take place.[21]
The Lewis acidity of the catalysts allows direct complexation
of organic substrates and exaltation of the acidity of water
molecules coordinated to the catalytic cornerstone; further-
more, both pathways are accessible, depending on the sub-
strates. These findings provide new and growing opportuni-
ties for POM catalysis.[5] Future developments will focus on
new Lewis acidic POM structures, investigation of the influ-
ence of the central group,[22] and extension to asymmetric
catalysis with chiral structures.
[6] C. Boglio, G. Lenoble, C. Duhayon, B. Hasenknopf, R. Thouvenot,
C. Zhang, R. C. Howell, B. P. Burton-Pye, L. C. Francesconi, E.
Lacꢁte, S. Thorimbert, M. Malacria, C. Afonso, J.-C. Tabet, Inorg.
Chem. 2006, 45, 1389–1398.
[7] a) S. Bareyt, S. Piligkos, B. Hasenknopf, P. Gouzerh, E. Lacꢁte, S.
Thorimbert, M. Malacria, Angew. Chem. 2003, 115, 3526–3528;
Angew. Chem. Int. Ed. 2003, 42, 3404–3406; b) S. Bareyt, S. Piligkos,
B. Hasenknopf, P. Gouzerh, E. Lacꢁte, S. Thorimbert, M. Malacria,
J. Am. Chem. Soc. 2005, 127, 6788–6794.
[8] a) For a review on chiral POMs, see: B. Hasenknopf, K. Micoine, E.
Lacꢁte, S. Thorimbert, M. Malacria, R. Thouvenot, Eur. J. Inorg.
Chem. 2008, 5001–5013; b) for a recent kinetic resolution of a
hybrid a1-Dawson complex, see: K. Micoine, B. Hasenknopf, S.
Thorimbert, E. Lacꢁte, M. Malacria, Angew. Chem. 2009, 121,
3518–3520; Angew. Chem. Int. Ed. 2009, 48, 3466–3468.
[9] a) M. Sadakane, M. H. Dickman, M. T. Pope, Inorg. Chem. 2001, 40,
2715–2719; b) Q.-H. Luo, R. C. Howell, M. Dankova, J. Bartis,
C. W. Williams, W. D. Horrocks, Jr., V. G. Young, A. L. Rheingold,
L. C. Francesconi, M. R. Antonio, Inorg. Chem. 2001, 40, 1894–
1901; c) C. Zhang, R. C. Howell, Q.-H. Luo, H. L. Fieselmann, L. J.
Todaro, L. C. Francesconi, Inorg. Chem. 2005, 44, 3569–3578; d) C.
Boglio, B. Hasenknopf, G. Lenoble, P. Rꢀmy, P. Gouzerh, S. Thorim-
bert, E. Lacꢁte, M. Malacria, R. Thouvenot, Chem. Eur. J. 2008, 14,
1532–1540.
[10] For selected examples of Zr and Hf complexes, see: a) X. Fang,
T. M. Anderson, Y. Hou, C. L. Hill, Chem. Commun. 2005, 5044–
5046; b) X. Fang, T. M. Anderson, C. L. Hill, Angew. Chem. 2005,
117, 3606–3610; Angew. Chem. Int. Ed. 2005, 44, 3540–3544;
c) C. N. Kato, A. Shinohara, K. Hayashi, K. Nomiya, Inorg. Chem.
2006, 45, 8108–8119; d) B. S. Bassil, S. S. Mal, M. H. Dickman, U.
Kortz, H. Oelrich, L. Walder, J. Am. Chem. Soc. 2008, 130, 6696–
6697; e) Y. Saku, Y. Sakai, A. Shinohara, K. Hayashi, S. Yoshida,
C. N. Kato, K. Yoza, K. Nomiya, Dalton Trans. 2009, 805–813; f) for
a chiral Hf sandwich-type POM, see: Y. Hou, X. Fang, C. L. Hill,
Chem. Eur. J. 2007, 13, 9442–9447.
Acknowledgements
We thank the CNRS, IUF, UPMC, le ministꢃre de l’ꢀducation nationale,
de l’enseignement supꢀrieur et de la recherche, and ANR (grant
JC05 41806 to E.L., S.T., and B.H.) for financial support of this study.
FR2769 is acknowledged for technical assistance. J.-C. Tabet and C.
Afonso are warmly thanked for discussions about MS determination.
[1] a) “Polyoxo anions: Synthesis and Structure”: M. T. Pope in Com-
prehensive Coordination Chemistry II, Vol. 4 (Ed.: A. G. Wedd),
Elsevier, Oxford, 2004, pp. 635–678; b) “Polyoxometalates: Reactiv-
ity”: C. L. Hill in Comprehensive Coordination Chemistry II, Vol. 4
(Ed.: A. G. Wedd), Elsevier, Oxford, 2004, pp. 679–759; c) M.
Misono, Catal. Today 2005, 100, 95–100; d) “Applications of Poly-
oxometalates in Homogeneous Catalysis”: R. Neumann in Polyoxo-
metalate Molecular Science Vol. 98 (Eds.: J. J. Borras-Almenar, E.
Coronado, A. Mꢄller, M. T. Pope), Kluwer, Dordrecht, 2003,
pp. 327–349; e) “Heterogeneous Catalysis by Heteropoly Com-
pounds”: I. V. Kozhevnikov in Polyoxometalate Molecular Science
Vol. 98 (Eds.: J. J. Borras-Almenar, E. Coronado, A. Mꢄller, M. T.
Pope), Kluwer, Dordrecht, 2003, pp. 351–380; f) L. E. Briand, G. T.
Baronetti, H. J. Thomas, Appl. Catal. A 2003, 256, 37–50.
[2] a) W. Adam, P. L. Alsters, R. Neumann, C. R. Saha-Moeller, D. See-
bach, A. K. Beck, R. Zhang, J. Org. Chem. 2003, 68, 8222–8231;
b) W. Adam, P. L. Alsters, R. Neumann, C. R. Saha-Moeller, D. See-
bach, A. K. Beck, R. Zhang, Org. Lett. 2003, 5, 725–728; c) T. M.
Anderson, W. A. Neiwert, M. L. Kirk, P. M. B. Piccoli, A. J. Schultz,
T. F. Koetzle, D. G. Musaev, K. Morokuma, R. Cao, C. L. Hill, Sci-
ence 2004, 306, 2074–2077; d) Y. Nakagawa, K. Kamata, M. Kotani,
K. Yamaguchi, N. Mizuno, Angew. Chem. 2005, 117, 5266–5271;
Angew. Chem. Int. Ed. 2005, 44, 5136–5141; e) S. S. Mal, N. H.
[11] R. D. Shannon, Acta Crystallogr. Sect. A 1976, 32, 751–767.
[12] a) V. W. Day, W. G. Klemperer, Science 1985, 228, 533–541; b) J. M.
Maestre, X. Lopez, C. Bo, J. M. Poblet, N. Casan-Pastor, J. Am.
Chem. Soc. 2001, 123, 3749–3758.
[13] S. Kobayashi, T. Busujima, S. Nagayama, Chem. Eur. J. 2000, 6,
3491–3494.
[14] For a recent review on the use of POMs as Brønsted acids, see: T.
Ueda, H. Kotsuki, Heterocycles 2008, 76, 72–97.
Chem. Eur. J. 2010, 16, 7256 – 7264
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7263

S. Thorimbert, E. Lacꢁte et al.
[15] a) S. Kobayashi, T. Ogino, H. Shimizu, S. Ishikawa, T. Hamada, K.
Manabe, Org. Lett. 2005, 7, 4729–4731; b) K. Saruhashi, S. Kobaya-
shi, J. Am. Chem. Soc. 2006, 128, 11232–11235; for a more general
review, see: c) S. Kobayashi, C. Ogawa, Chem. Eur. J. 2006, 12,
5954–5960.
[19] a) M. N. Sokolov, N. V. Izarova, E. V. Peresypkina, D. A. Mainichev,
V. P. Fedin, Inorg. Chim. Acta 2009, 362, 3756–3762; b) Y. Saku, Y.
Sakai, K. Nomiya, Inorg. Chim. Acta 2010, 363, 967–974.
[20] E. Derat, E. Lacꢁte, B. Hasenknopf, S. Thorimbert, M. Malacria, J.
Phys. Chem. A 2008, 112, 13002–13005.
[16] a) S. Kobayashi, S. Nagayama, J. Am. Chem. Soc. 1997, 119, 10049–
10053; b) S. Kobayashi, Chem. Rev. 2002, 102, 2227–2302.
[17] a) H. Yamamoto, Proc. Jpn. Acad. Ser. B 2008, 84, 134–146; b) for
recent reports, see: K. Shimizu, K. Niimi, A. Statsuma, Catal.
Commun. 2008, 9, 980–983; c) X. Zhang, J. Li, Y. Chen, J. Wang, L.
Feng, X. Wang, Energy Fuels 2009, 23, 4640–4646.
[21] G. Lenoble, B. Hasenknopf, R Thouvenot, J. Am. Chem. Soc. 2006,
128, 5735–5744.
[22] D.-L. Long, Y.-F. Song, E. F. Wilson, P. Kçgerler, S.-X. Guo, A. M.
Bond, J. S. J. Hargreaves, L. Cronin, Angew. Chem. 2008, 120, 4456–
4459; Angew. Chem. Int. Ed. 2008, 47, 4384–4387.
[18] a) H. C. Brown, R. B. Johannesen, J. Am. Chem. Soc. 1953, 75, 16–
20; b) A. Corma, Chem. Rev. 1995, 95, 559–614.
Received: February 16, 2010
Published online: May 7, 2010
7264
www.chemeurj.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7256 – 7264