Supported N-alkylimidazole-decorated dendrons as heterogeneous catalysts for the Baylis-Hillman reaction

Article abstract of DOI:10.1039/b915577e

The Baylis-Hillman reaction between methyl vinyl ketone and aromatic aldehydes, catalyzed by N-alkylated imidazoles immobilized on polystyrene via dipolar cycloaddition, esterification or nucleophilic substitution, exhibits a remarkable positive dendritic

Full text of DOI:10.1039/b915577e

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Supported N-alkylimidazole-decorated dendrons as heterogeneous
catalysts for the Baylis–Hillman reactionw
Kerem Goren and Moshe Portnoy*
Received (in Cambridge, UK) 30th July 2009, Accepted 18th December 2009
First published as an Advance Article on the web 19th January 2010
DOI: 10.1039/b915577e
The Baylis–Hillman reaction between methyl vinyl ketone
and aromatic aldehydes, catalyzed by N-alkylated imidazoles
immobilized on polystyrene via dipolar cycloaddition, esterification
or nucleophilic substitution, exhibits a remarkable positive
dendritic effect.
enantioselective catalysis), have never been used as support-
9
tethered catalytic units.
–11
Three series of supported dendritic catalysts of first or
second generation, decorated with N-alkylated imidazoles at
the dendritic periphery, were constructed (Scheme 1). For
comparison, a non-dendritic analogue was always prepared.
Three immobilization reactions were utilized.z In the
first series the imidazoleacetic acid was used to esterify the
resin-bound alcohols. The second series of the catalysts was
formed by the nucleophilic substitution of the resin-bound
benzyl chlorides with N-(3-aminopropyl)imidazole. In the
third series propargyl ether of N-(2-hydroxyethyl)imidazole
was coupled via Sharpless–Huisgen cycloaddition to the
resin-bound azide groups.
The Baylis–Hillman reaction is a useful carbon–carbon bond
1
forming process. The popularity of this transformation
increased enormously due to its operational simplicity,
functional group tolerance and high number of reactive
functional groups in a typical reaction product. The last
property ensures the reaction usefulness for multistep organic
synthesis. One of the drawbacks of the reaction is the relatively
high loading of a catalyst, required for achieving high
yields. Under these circumstances the preparation of readily
accessible, easily removable and reusable catalysts becomes
increasingly more profitable. Immobilization of analogues of
homogeneous catalysts via covalent attachment to a support
is one of the most straightforward ways of preparing
well-defined heterogeneous catalytic systems, a task of great
A model reaction of methyl vinyl ketone with 4-nitro-
benzaldehyde was previously investigated with N-alkylated
9
d,10
imidazole catalysts in solution.
Examination of this
addition with the polystyrene-supported catalysts revealed a
dramatic influence of the solvent and the dendron generation
as well as a notable temperature dependence.y
2
economic and environmental importance.
It was already indicated in the literature that water plays a
crucial role in the acceleration of the reaction.
9b–d,12
Though some polymer-supported catalytic systems suffer
from reduced activity and selectivity, we earlier demonstrated
that dendronization of the support frequently leads to positive
dendritic effects on these catalytic properties for metal-based
Indeed,
while for the non-dendritic catalysts, low conversion was
observed in dry polar aprotic solvents THF and DMF, upon
3
catalysts. Moreover, polymer-bound dendritic catalytic
systems based on proline were recently prepared and explored
in our group, demonstrating a substantial improvement in the
4
performance upon dendronization of the support.
The aim of our recent research was the synthesis of polymer-
supported poly(aryl benzyl ether) dendrimers, decorated at the
periphery with N-alkylated imidazoles, and the exploration of
their activity, their recyclability and dendritic effects, which
they display, in the catalysis of the Baylis–Hillman reaction.
Though a few polystyrene- or silica-supported catalysts
for the Baylis–Hillman reaction have been reported,
these were based on dialkylaminopyridine, quinuclidine or
5
–7
phosphine nucleophiles.
also used, on a couple of occasions, in conjugation with a
Immobilized Brønsted acids were
8
soluble nucleophilic catalyst. Imidazole and N-alkylimidazole
derivatives that exhibited an interesting catalytic activity
in solution (particularly in aqueous solutions and for
School of Chemistry, Raymond and Beverly Sackler Faculty of Exact
Sciences, Tel Aviv University, 69978, Tel Aviv, Israel.
E-mail: portnoy@post.tau.ac.il; Fax: +972-3-6409293;
Tel: +972-3-6406517
w Electronic supplementary information (ESI) available: General
experimental conditions, synthesis of A and characterization data
for the catalysts. See DOI: 10.1039/b915577e
Scheme 1 Synthesis of the catalysts. Reagents and conditions: (a)
tetramethyl-O-(1H-benzotriazol-1-yl)uronium hexafluorophosphate
i
HBTU), Pr
(
2
NEt, 1-imidazoleacetic acid, rt, DMF; (b) 1-(3-aminopropyl)-
imidazole, 50 1C, DMF; (c) A, sodium ascorbate, CuSO
4
, DMF,
50 1C.
This journal is ꢀc The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 1965–1967 | 1965

View Article Online
a
Table 1 The screening of the model reaction conditions with the
non-dendritic catalysts
Table 2 Evaluation of the dendritic effect
a
b
c
Yield (%)
Entry
Catalyst
t/h
Solvent
b
c
Yield (%)
Entry
Catalyst
t/h
Solvent
e
1
G0(A-Im)
G1(A-Im)
G2(A-Im)
G0(A-Im)
G1(A-Im)
G2(A-Im)
G0(A-Im)
G1(A-Im)
G2(A-Im)
G0(A-Im)
G1(A-Im)
G2(A-Im)
G0(E-Im)
G1(E-Im)
G2(E-Im)
G0(T-Im)
G1(T-Im)
G2(T-Im)
G1(A-Im)
G2(A-Im)
G2(A-Im)
G2(A-Im)
20
20
20
6
6
6
4
4
4
1.5
1.5
1.5
DMF
DMF
DMF
THF–H
THF–H
THF–H
27
38
91
76
95
100
34
47
91
38
62
95
31
43
59
58
74
100
32
96
99
99
d
e
1
2
3
4
5
6
7
8
G0(E-Im)
G0(E-Im)
G0(E-Im)
G0(E-Im)
G0(T-Im)
G0(T-Im)
G0(T-Im)
G0(T-Im)
G0(A-Im)
G0(A-Im)
G0(A-Im)
G0(A-Im)
G0(A-Im)
G0(A-Im)
G0(A-Im)
G0(A-Im)
G0(A-Im)
G0(A-Im)
48
48
48
48
22
22
22
22
22
22
22
22
17
17
17
17
6
DMF
THF
DMF–H
11
18
31
28
6
2
d
e
3
d
e
2
O (9 : 1)
O (9 : 1)
4
2
2
2
O (9 : 1)
O (9 : 1)
O (9 : 1)
d
e
THF–H
DMF
THF
2
5
d
e
6
d
e
7
8
DMF–H
DMF–H
DMF–H
DMF–H
DMF–H
DMF–H
DMF–H
DMF–H
DMF–H
DMF–H
DMF–H
DMF–H
DMF–H
DMF–H
DMF–H
DMF–H
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
O (9 : 1)
O (9 : 1)
O (9 : 1)
O (7 : 3)
O (7 : 3)
O (7 : 3)
O (9 : 1)
O (9 : 1)
O (9 : 1)
O (9 : 1)
O (9 : 1)
O (9 : 1)
O (9 : 1)
O (1 : 1)
O (1 : 1)
O (1 : 1)
d
e
DMF–H
2
O (9 : 1)
O (9 : 1)
58
20
23
35
80
99
66
90
99
99
69
97
8
d
e
THF–H
DMF
THF
2
9
10
d
e
9
d
e
1
1
1
1
1
1
1
1
1
0
1
2
3
4
5
6
7
8
11
12
13
14
15
16
17
18
d
d
e
e
e
e
e
e
e
DMF–H
THF–H
2
O (9 : 1)
O (9 : 1)
d
2
48
d
DMF–H
DMF–H
DMF–H
DMF–H
DMF–H
DMF–H
2
2
2
2
2
2
O (19 : 1)
O (9 : 1)
O (8 : 2)
O (1 : 1)
O (8 : 2)
O (1 : 1)
48
48
22
22
22
4
4
6
12
d
d
d
d
e,f
6
19
20
e,g
a
Reaction conditions: 1 mmol of 4-nitrobenzaldehyde, 3 mmol of
b
MVK and 10 mol% (0.1 mmol) of the catalyst in 1 mL solvent. For
e,g
2
2
1
2
e,h
c
d
e
mixtures v/v ratio. Determined by NMR. At 15 1C. At 25 1C.
a
Reaction conditions: 1 mmol of 4-nitrobenzaldehyde, 3 mmol of
MVK and 10 mol% (0.1 mmol) of the catalytic units in 1 mL
b
c
d
solvent. For mixtures v/v ratio. Determined by NMR. At 15 1C.
addition of water as a cosolvent (up to 50% v/v), the conversion
was improved, sometimes dramatically (Table 1).
e
f
At 25 1C. In the presence of 120 mg Bromo Wang polystyrene.
g
h
.5 mol% (0.025 mmol) of the catalytic units. 1 mol% (0.01 mmol)
2
of the catalytic units.
We examined zeroth to second generation catalysts of the
Gn(A-Im) series in anhydrous solvents, as well as in various
aqueous mixtures of DMF and THF (Table 2). In all solvents
a strong dendritic effect was observed. Thus, in anhydrous
DMF at 25 1C more than three-fold increase in yield was
observed comparing the second to the zeroth generation
catalysts (entries 1–3). Strong dendritic effects were also
observed in 9 : 1 mixtures of THF–water or DMF–water,
although the magnitude of the effect was very slightly reduced
The plausible cause of the dendritic effect is the assistance of
the second, nearby imidazolyl moiety in proton transfer from
the a-to-carbonyl carbon to the alkoxide oxygen in the inter-
mediate adduct, after the first imidazole unit activated the
enone, promoting the formation of the new carbon–carbon
bond (Scheme 2). Recent mechanistic studies demonstrated
that the proton shift, rather than the enolate attack on the
(
entries 4–9). An additional small reduction in the
effect magnitude was observed in 7 : 3 DMF–water mixture
entries 10–12). Remarkably, an almost quantitative yield was
1
4
aldehyde, is the rate-determining step of the reaction. The
known imidazole ability to serve as a proton shuttle can
therefore facilitate the overall reaction rate.
(
already obtained after 90 min in this case with the Gn(A-Im)
catalyst.
The effect of water is even more remarkable, when one takes
into account the lower swelling of the resins in the aqueous
solvent mixtures and the possible inaccessibility of some of the
Even more outstanding were the results of the experiments
with benzaldehyde replacing the nitrobenzaldehyde in the
reaction with methyl vinyl ketone. While at room temperature
even in a 1 : 1 DMF–water mixture the G0(A-Im) resin failed
to catalyze the reaction (o1% after 20 h), the G2(A-Im)
catalyst induced 80% yield under identical conditions!
Examination of two other series of the catalysts (Gn(E-Im)
and Gn(T-Im)) demonstrated that both are prone to the same
dendritic effect as the Gn(A-Im) series, although the catalysts
based on imidazoleacetic acid are significantly less efficient,
probably due to the electron-withdrawing effect of the carboxy-
methyl substituent (Table 2, entries 13–18). Remarkably,
perfect chemoselectivity, with the Baylis–Hillman adduct
being the only product of the reaction, was observed for all
polymer-supported catalysts as a striking difference from the
results of the model reaction with the simplest homogeneous
1
5
catalytic sites as a result. It is conceivable that the assistance
of water molecules in the proton shift step is one of its
1
6
contributions as a cosolvent. This explanation fits the larger
dendritic effect observed in anhydrous solvents. However,
in our opinion, the main effect caused by water is the
‘‘extraction’’ of the reagents from the highly polar solvent
mixture into the hydrophobic pockets of the ‘‘polystyrene
1
7,18
phase’’.
In order to prove this, we conducted an experi-
ment exactly equivalent to that in Table 2, entry 8, but with
120 mg of imidazole-lacking polystyrene beads added to the
reaction mixture (in addition to the 120 mg of the G1(A-Im)
beads). As a result, the yield of the reaction decreased to 32%
(as compared to the original 47%). This evidence strongly
supports the above-mentioned explanation of the main reason
for the water-acceleration effect.
1
3
analogue of the catalysts, N-methylimidazole.
1
966 | Chem. Commun., 2010, 46, 1965–1967
This journal is ꢀc The Royal Society of Chemistry 2010

View Article Online
The resin was washed with DMF–water, DMF, THF–water, THF,
DCM and then dried under vacuum.
y General procedure for the Baylis–Hillman reaction: The catalytic resin
(
1
0.1 mmol of N-alkylated imidazole units, 0.1 equiv.) was added to
mL of solvent. Methyl vinyl ketone (3 mmol, 3 equiv.) and the
aldehyde (1 mmol, 1 equiv.) were added and the suspension was mixed
at room temperature. The resin was filtered out and washed with ethyl
acetate. Water (10 mL) and saturated aqueous NH
10 mL) were added to the combined filtrate. The mixture was
extracted with ethyl acetate (3 ꢁ 10 mL). The organic phase was dried
on MgSO and the solvent was evaporated to yield the crude material
for analysis.
4
Cl solution
(
4
1
(a) G. Masson, C. Housseman and J. Zhu, Angew. Chem., Int. Ed.,
007, 46, 4614; (b) D. Basavaiah, K. V. Rao and R. Reddy, Chem.
Soc. Rev., 2007, 36, 1581.
2
2
3
(a) M. Benaglia, A. Puglisi and F. Cozzi, Chem. Rev., 2003, 103,
3401; (b) F. Cozzi, Adv. Synth. Catal., 2006, 348, 1367.
(a) A. Dahan and M. Portnoy, Chem. Commun., 2002, 2700;
(
(
b) A. Dahan and M. Portnoy, Org. Lett., 2003, 5, 1197;
c) A. Dahan and M. Portnoy, J. Am. Chem. Soc., 2007, 129, 5860.
4
5
(a) T. Kehat and M. Portnoy, Chem. Commun., 2007, 2823;
b) K. Goren, T. Kehat and M. Portnoy, Adv. Synth. Catal.,
009, 351, 59.
(a) J. W. Huang and M. Shi, Adv. Synth. Catal., 2003, 345, 953;
b) A. Corma, H. Garcıa and A. Leyva, Chem. Commun., 2003,
(
2
Scheme 2 Proposed explanation of the dendritic effect.
(
´
Recycling experiments showed that there is a gradual
deterioration in the catalyst activity upon recycling. On the
other hand, the use of the optimized conditions and second
generation catalyst permits a significant reduction in the
catalyst loading, required to achieve high yield (compare
Table 2, entries 20–22 vs. Table 2, entry 9 and Table 1,
entries 18 and 14, respectively). Though these results do not
eliminate the need for improving the recyclability of the
catalysts, they reduce to some extent the incentive to recycle.
In conclusion, we applied new modes of immobilization of
N-alkylated imidazole-based catalysts to the regular and
dendronized polymeric supports. We demonstrated the
remarkable positive influence of the dendritic spacers and of
water on the yield of the Baylis–Hillman reaction with these
catalysts. Additional experiments, aimed at the understanding
of the mechanism of the dendritic effect are underway.
2806; (c) H. T. Chen, S. Huh, J. W. Wiench, M. Pruski and
V. S.-Y. Lin, J. Am. Chem. Soc., 2005, 127, 13305; (d) B. Zhao,
X. Jiang, D. Li, X. Jiang, T. G. O Lenick, B. Li and C. Y. Li,
J. Polym. Sci., Part A: Polym. Chem., 2008, 46, 3438.
S. Luo, X. Zheng, H. Xu, X. Mi, L. Zhang and J. P. Cheng, Adv.
Synth. Catal., 2007, 349, 2431.
C. K. W. Kwong, R. Huang, M. Zhang, M. Shi and P. H. Toy,
Chem.–Eur. J., 2007, 13, 2369.
0
6
7
8
X. Wang, P. Han, X. Qiu, X. Ji and L. Gao, Catal. Lett., 2008, 124,
418.
9 (a) M. Shi, J. K. Jiang and C. Q. Li, Tetrahedron Lett., 2002, 43,
1
7
¨
27; (b) R. Gatri and M. M. El Gaıed, Tetrahedron Lett., 2002, 43,
835; (c) H. J. Davies, A. M. Ruda and N. C. O. Tomkinson,
Tetrahedron Lett., 2007, 48, 1461; (d) K. Asano and S. Matsubara,
Synlett, 2009, 1, 35.
0 (a) J. E. Imbriglio, M. M. Vasbinder and S. J. Miller, Org. Lett.,
1
2
003, 5, 3741; (b) C. E. Aroyan, M. M. Vasbinder and S. J. Miller,
Org. Lett., 2005, 7, 3849; (c) M. M. Vasbinder, J. E. Imbriglio and
S. J. Miller, Tetrahedron, 2006, 62, 11450.
11 For stoichiometric use of imidazole in aqueous solutions see:
a) S. Luo, B. Zhang, J. He, A. Janczuk, P. G. Wang and
(
This research was supported by The Israel Science Foundation
Grant No. 960/06). We thank Prof. Scott J. Miller for fruitful
J. P. Cheng, Tetrahedron Lett., 2002, 43, 7369; (b) S. Luo,
P. G. Wang and J. P. Cheng, J. Org. Chem., 2004, 69, 555;
(c) P. Narender, B. Gangadasu, M. Ravinder, U. Srinivas,
G. Y. S. K. Swamy, K. Ravikumar and V. J. Rao, Tetrahedron,
(
discussion.
2006, 62, 954.
2 R. O. M. A. de Souza, V. L. P. Pereira, P. M. Esteves and
Notes and references
1
z General procedure for the 1,3-dipolar cycloaddition: 1-(2-(prop-2-
ynyloxy)ethyl)-1H-imidazole (A, 5 equiv. per azidobenzyl unit),
sodium ascorbate (1 equiv. per azidobenzyl unit, 1 M in DMF) and
copper(II) sulfate pentahydrate (0.25 equiv. per azidobenzyl unit) were
M. L. A. A. Vasconcellos, Tetrahedron Lett., 2008, 49, 5902.
13 In this homogeneously catalyzed model reaction, both in
anhydrous and aqueous DMF, the Baylis–Hillman adduct
constitutes less than 50% of a complex product mixture.
added to a suspension of polystyrene azidobenzyl-terminated resin
(
to 50 1C overnight. The resin was washed with DMF–water, DMF,
THF–water, THF, DCM and then dried under vacuum.
General procedure for the coupling reaction: 1-imidazoleacetic acid
14 (a) V. K. Aggarwal, S. Y. Fulford and G. C. Lloyd-Jones, Angew.
Chem., Int. Ed., 2005, 44, 1706; (b) K. E. Price, S. J. Broadwater,
H. M. Jung and D. T. McQuade, Org. Lett., 2005, 7, 147;
(c) K. E. Price, S. J. Broadwater, B. J. Walker and
D. T. McQuade, J. Org. Chem., 2005, 70, 3980.
4
1 equiv.) in DMF (10 mL per 1 g resin). The suspension was heated
(4 equiv. per hydroxybenzyl unit), HBTU (4 equiv. per hydroxybenzyl
unit) and Pr
2
15 For instance, the volume of Gn(A-Im) in 9 : 1 and 1 : 1 DMF–H O
i
2
NEt (8 equiv. per hydroxybenzyl unit) were added to a
suspension of polystyrene hydroxybenzyl-terminated resin (1 equiv.) in
solutions is only 67 and 53%, respectively, of that in pure DMF.
16 D. Roy and R. B. Sunoj, Chem.–Eur. J., 2008, 14, 10530.
17 M. Gruttadauria, F. Giacalone and R. Noto, Adv. Synth. Catal.,
2009, 351, 33.
18 For ‘‘on water’’ reactions see, for instance: (a) ref. 9d; (b) N. Shapiro
and A. Vigalok, Angew. Chem., Int. Ed., 2008, 47, 2849.
19 (a) A. Dahan and M. Portnoy, Macromolecules, 2003, 36, 1034;
(b) A. Mansour, T. Kehat and M. Portnoy, Org. Biomol. Chem.,
2008, 6, 3382.
19
DMF (10 mL per 1 g resin). The suspension was mixed for 3 h. The
resin was washed with DMF–water, DMF, THF–water, THF, DCM
and dried under vacuum.
General procedure for the halide substitution: 1-(3-aminopropyl)-
imidazole (5 equiv. per halobenzyl unit) was added to a suspension
of polystyrene halobenzyl-terminated resin (1 equiv.) in DMF
1
9b
(
10 mL per 1 g resin). The suspension was heated to 50 1C for 24 h.
This journal is ꢀc The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 1965–1967 | 1967