A Highly Active and Selective Catalyst System for the Baylis-Hillman Reaction

Article abstract of DOI:10.1002/anie.200352233

A key role for sulfur and the potential for N_ax→P intrabridgehead interactions are invoked to explain the impressive effect of the cocatalyst (shown in color) in the TiCl₄-mediated Baylis-Hillman reaction [Eq. (1)]. Even sterically h

Full text of DOI:10.1002/anie.200352233

Communications
Baylis–Hillman Reaction
Initially, we examined the model reaction of p-nitro-
benzaldehyde (1mmol) with comparatively unreactive cyclo-
hex-2-en-1-one (3 mmol) in CH2Cl2 (2 mL) under argon at
room temperature.[10] When 20 mol% catalyst 2a was
employed without any additives, this reaction gave only a
5% product yield (based on the aldehyde) after four days.
However, we found that the rate improved remarkably when
one equivalent of some Lewis acids was added. After
screening a variety of such acids (e.g., MgSO4, BF3·OEt2,
BCl3, SnCl4, AlCl3, Ti(OiPr)4, and TiCl4), we found that TiCl4
gave the best results. The model reaction carried out in the
presence of 20 mol% 2a and 1.0 equiv of TiCl4 gave a 95%
yield of the product in only 5 min at room temperature.[11]
Even with as little as 5 mol% 2a, a 94% yield was observed in
10 min at room temperature. Proazaphosphatrane sulfides 2b
and 2c showed slightly reduced activities over the same time
period (perhaps due to steric hindrance of the R group),
giving yields of 85 and 78%, respectively. To our knowledge,
2a/TiCl4 is the most effective catalyst system thus far reported
for BH reactions.
To shed light on the possible origin of the remarkable
activity of 2a in TiCl4-promoted BH reactions, proazaphos-
phatrane oxide 3 was first chosen for comparison. The BH
reaction catalyzed by 0.05 equiv of 3 proceeded in 48% yield
in only 10 minutes, similar to the yield realized with 1 equiv of
TiCl4 by itself (44%) over 2 h. This finding suggests a key role
for the sulfur atom in 2a in its catalytic activity. Interestingly,
reactions catalyzed by 0.05 equivalents of the analogous bases
4, lacking the bridgehead nitrogen of 2a, and 5 led to dramatic
decreases in yield (69 and 55%, respectively) after a reaction
time of 10 minutes, and the acyclic analogue 6 (0.05 equiv)
showed only moderate activity (59% yield) compared with
that of 2a over the same time period. The stronger cocatalytic
activity of 2a compared with that of 4–6 appears to be
associated with the potential for Nax!P intrabridgehead
interactions, which enhances reaction rates and the nucleo-
philicity of the sulfur atom. Although the molecular structure
of 2a reveals an Nax!P distance only 3% shorter than the
sum of van der Waals radii of the P and N atoms (3.34 Š),[12] it
is conceivable that additional shortening of this distance
occurs during the reaction.
A Highly Active and Selective Catalyst System for
the Baylis–Hillman Reaction**
Jingsong You, Juhua Xu, and John G. Verkade*
The Baylis–Hillman (BH) reaction, i.e., coupling of an
activated alkene or alkyne with an aldehyde or ketone, has
recently become a very attractive goal.[1] These reactions
usually require Lewis bases as catalysts, such as tertiary
phosphanes or tertiary amines, among which 1,4-diazabicy-
clo[2.2.2]octane (DABCO) is the most popular. However, this
approach suffers from slow reaction rates (reaction times of
weeks and even months[1,2]) which limit the scope of
substrates. Numerous chemical and physical methods have
been developed to accelerate BH reactions.[1,3] One of the
Lewis acids typically used to activate the carbonyl group[4] is
TiCl4, used with or without additives (e.g., quaternary
ammonium salts, an organic chalcogenide, phosphanes, a
diol or a bisoxazoline).[1,4,5] Although the rates of BH
reactions are improved significantly with the assistance of
TiCl4, the yields are generally only moderate, and limitations
are encountered on the structures of the Michael acceptors
and the aldehydes which undergo addition, frequently
producing complex mixtures.[5–7]
The commercially available bicyclic proazaphosphatranes
1a–1c, first synthesized in our laboratories, have attracted
considerable interest in recent years as versatile, exceedingly
strong nonionic bases and as catalysts for a variety of useful
transformations.[8] Recently, we reported that in the presence
of proazaphosphatrane 1c, activated allylic compounds such
as acrylonitrile react efficiently with aromatic aldehydes to
afford BH adducts as the only product.[9] However, when
activated alkenes (cyclohex-1-en-1-one, vinyl ketones, acryl-
ates, and allyl cyanide) were used as substrates in the presence
of 1a–1c, the BH reactions failed whether with or without a
Lewis acid (TiCl4 or AlCl3) present. Herein we report the
surprising discovery that the proazaphosphatrane sulfide 2a,
easily prepared from commercially available 1a, facilitates
unparalleled speed and selectivity in BH reactions catalyzed
by TiCl4.
The first mechanism suggested for BH reactions mediated
by a chalcogenide/metal halide catalyst system involved
attack of the sulfur atom from R2S on
the activated alkene.[6a] However, it is
currently believed that these reactions
proceed by attack of a halide ion
released from a Lewis acid by R2S.[7b]
Supporting evidence for this was the
unsuccessful use of BF3·OEt2 as the
Lewis acid, from which release of a
halide ion is far less likely.[6] Very recently a BH reaction of
methyl vinyl ketone with aldehydes mediated by tetrahydro-
thiophene/BF3·OEt2 was developed which gave moderate
product yields, indicating the unambiguous requirement of
the sulfide for attack on the activated alkene.[13] In accord
with these observations, the use of 1.0 equiv of BF3·OEt2 in
the reaction of cyclohex-2-en-1-one with p-nitrobenzaldehyde
in the presence of 2a (0.2 equiv) at room temperature gave a
[*] Prof. J. G. Verkade, Dr. J. You, Dr. J. Xu
Department of Chemistry
Iowa State University
Ames, IA 50011 (USA)
Fax: (+1)515-294-0105
E-mail: jverkade@iastate.edu
[**] The authors are grateful to the National Science Foundation for
grant support.
5054
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200352233
Angew. Chem. Int. Ed. 2003, 42, 5054 –5056

Angewandte
Chemie
50% yield of the BH adduct in four hours, while no desired
adduct was isolated in the absence of 2a.[14] It is reasonable to
suggest that coordination of BF3 to the carbonyl oxygen of the
alkene activating group is followed
our catalytic system provided very good yields (entries 5–7).
Noncyclic Michael-type acceptors were also examined
under our aforementioned conditions (Table 2). Methyl vinyl
by attack of the sulfur atom of 2a
on the carbon b to the carbon of the
activating groups of the alkene,
with a concomitant Nax!P intra-
bridgehead interaction that elec-
tronically enriches the carbon a to
the activating group[15] for nucleo-
philic attack on the aldehyde
carbon. The low yield (53%) of
BH product achieved in 10 min in
the reaction of p-nitrobenzalde-
hyde with 3-methylcyclohex-2-ene-
1-one in the presence of TiCl4
(1equiv) and 0.05 equiv of 2a is
also in agreement with attack of 2a
(although impeded by the methyl
group) at the 3-position of the 3-
substituted enone.[16]
BH reactions mediated by 2a/
TiCl4 tolerate a relatively wide
scope of Michael acceptors and
aldehydes at room temperature.
As is evident in Table 1, cyclic
enones (which are less reactive
than acyclic analogues) were effi-
ciently coupled with a wide variety
of aldehydes to give BH adducts.
Additionally, reactions of electron-
Table 1: Baylis–Hillman reactions of aldehydes with cyclic enones catalyzed by 2a/TiCl4.[a]
Entry
Aldehyde
Enone
Product
t [min]
Yield [%][b]
94
1
10
2
3
10
10
92
91
4
5
6
7
10
10
30
30
94
88
81
89
8
10
88
9
10
10
91
90
deficient
aromatic
aldehydes
afforded excellent yields in very
short times (Table 1, entries 1–4).
We were also delighted to find that
electron-neutral aromatic alde-
hydes such as p-methylbenzalde-
hyde and benzaldehyde gave BH
adducts with cyclohex-2-en-1-one
in 91% and 90% yields, respec-
tively, in only 10 min (entries 9 and
10). Surprisingly, the reaction of
the notoriously recalcitrant hin-
dered and deactivated o-anisalde-
hyde with cyclo-hex-2-en-1-one
proceeded in 88% yield (entry 8).
In sharp contrast, BH reactions of
inactive aromatic aldehydes such
as this are sluggish with other
reported catalyst systems, and the
related reactions with cyclic enones
are generally limited to the use of
aromatic aldehydes with strong
10
[a] The reactions were carried out with 1.0 mmol of aldehyde and 3.0 mmol of enone in the presence of
2a (0.05 mmol) and TiCl4 (1.0 mmol) in CH2Cl2 under Ar at room temperature. [b] Yield of isolated
product based on aldehydes.
Table 2: Baylis–Hillman reactions of aromatic aldehydes with acyclic Michael-type acceptors catalyzed
by 2a/TiCl4.[a] EWG=electron-withdrawing group.
Entry
Aldehyde
EWG
t [min]
Yield [%][b]
1
2
3
4
5
6
7
8
9
4-O2NC6H4CHO
C6H5CHO
4-O2NC6H4CHO
4-O2NC6H4CHO
4-ClC6H4CHO
C6H5CHO
4-CH3OC6H4CHO
C6H5CHO
4-O2NC6H4CHO
COCH3
COCH3
COOC2H5
COOCH3
COOCH3
COOCH3
COOCH3
CN
5
5
92
85
92
92
92
88
87
95
88
10
10
10
10
10
20
10
electron-withdrawing
groups.[6]
CN
Whereas BH reactions with
simple aliphatic aldehydes using
TiCl4/Lewis base copromoters
often give low product yields,[6]
[a] The reactions were carried out with 1.0 mmol of aldehyde and 3.0 mmol of alkene in the presence of
2a (0.05 mmol) and TiCl4 (1.0 mmol) in CH2Cl2 under Ar at room temperature. [b] Yield of isolated
product based on aldehyde.
Angew. Chem. Int. Ed. 2003, 42, 5054 –5056
www.angewandte.org
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5055

Communications
Iwamura, S. Watanabe, O. Muraoka, G. Tanabe, Tetrahedron
2000, 56, 4725 – 4731; e) T. Kataoka, H. Kinoshita, S. Kinoshita,
T. Iwamura, S. Watanabe, Angew. Chem. 2000, 112, 2448 – 2450;
Angew. Chem. Int. Ed. 2000, 39, 2358 – 2360.
ketone gave exclusively BH products in high yields (entries 1
and 2), and side products were not observed by H NMR
1
spectroscopy. This result remarkably contrasts previous find-
ings that reactions of a,b-unsaturated acyclic ketones with
aldehydes mediated by metal and non-metal halides resulted
in complex reaction mixtures.[7] Acrylates are normally very
unreactive in Lewis acid catalyzed BH reactions, and very low
to zero yields of coupling products are obtained.[6b,7c,7d]
Interestingly, our new methodology works very well for
both electron-poor and electron-rich aromatic aldehydes,
giving excellent yields of Baylis–Hillman products in only
10 min (entries 3–7). When acrylonitrile was employed,
reactions with aromatic aldehydes also gave high product
yields (entries 8 and 9).
[7] a) S. Uehira, Z. Han, H. Shinokubo, K. Oshima, Org. Lett. 1999,
1, 1383 – 1385; b) M. Shi, J.-K. Jiang, Tetrahedron 2000, 56, 4793 –
4797; c) M. Shi, J.-K. Jiang, Y.-S. , Feng, Org. Lett. 2000, 2, 2397 –
2400; d) M. Shi, Y.-S. Feng, J. Org. Chem. 2001, 66, 406 – 411;
e) G. Li, J. Gao, H.-X. Wei, M. Enright, Org. Lett. 2000, 2, 61 7 –
620.
[8] a) J. G. Verkade, Top. Curr. Chem. 2003, 233, 1– 44, and
references therein; b) P. B. Kisanga, J. G. Verkade, Tetrahedron
Reports, 2003, in press.
[9] P. B. Kisanga, J. G. Verkade, J. Org. Chem. 2002, 67, 426 – 430.
[10] An excess of a,b-unsaturated ketone required for complete
consumption of aldehyde and optimized yields.
[11] Typical procedure for the BH reaction: To a solution of aldehyde
(1.0 mmol), activated alkene (3.0 mmol), and 2a (0.05 mmol) in
anhydrous dichloromethane (2.0 mL) was added TiCl4
(1.0 mmol) under an Ar atmosphere at room temperature. The
mixture was stirred for 10 min and then it was quenched with
saturated aqueous NaHCO3 (5.0 mL). The inorganic precipitate
was filtered through Celite. The organic phase was dried over
Na2SO4 and concentrated in vacuo to give the crude product,
which was purified by column chromatography on silica gel
(ethyl acetate/hexane 1:3).
[12] S. K. Xi, H. Schmidt, C. Lensink, S. Kim, D. Wintergrass, L. M.
Jacobson, J. G. Verkade, Inorg. Chem. 1990, 29, 2214 – 2220.
[13] L. M. Walsh, C. L. Winn, J. M. Goodman, Tetrahedron Lett.
2002, 43, 8219 – 8222.
[14] Under the same conditions the reaction of methyl vinyl ketone
with p-nitrobenzaldehyde gave the BH adduct in 75% yield.
[15] Attempts to detect this intermediate by NMR spectroscopy have
thus far failed.
In summary, we have developed a highly active and
selective catalyst system (2a/TiCl4) for the BH reaction,
which to our knowledge is the most generally effective one so
far described. Our protocol is applicable to activated alkenes
such as enones (including less reactive b-substituted deriva-
tives), acrylonitrile, and acrylates. We believe that Nax!P
intrabridgehead interaction in 2a may play an important role
in stabilizing a 2a/activated alkene/TiCl4 intermediate in the
reaction pathway. Further development of this methodology,
including its application to asymmetric reactions using chiral
proazaphosphatrane sulfides, is in progress.
Received: June 26, 2003 [Z52233]
Keywords: Baylis–Hillman reaction · heterocycles · Lewis bases ·
.
titanium
[16] Electron-rich aromatic aldehydes such as anisaldehyde do not
provide BH adducts when
a more sterically hindered 3-
substituted cyclohex-2-en-1-one is used.
[1] For recent reviews see: a) D. Basavaiah, A. J. Rao, T. Satyanar-
ayana, Chem. Rev. 2003, 103, 811 – 891, and references therein;
b) P. Langer, Angew. Chem. 2000, 112, 3177 – 3180; Angew.
Chem. Int. Ed. 2000, 39, 3049 – 3052.
[2] a) S. Matsumoto, Y. Okubo, K. Mikami, J. Am. Chem. Soc. 1998,
120, 4015 – 4016; b) Y. Fort, M. C. Berthe, P. Caubere, Tetrahe-
dron 1992, 48, 6371– 6384.
[3] a) C. Yu, B. Liu, L. Hu, J. Org. Chem. 2001, 66, 5413 – 5418;
b) W.-D. Lee, K.-S. Yang, K. Chen, Chem. Commun. 2001, 1 61 2 –
1613; c) J. Cai, Z. Zhou, G. Zhao, C. Tang, Org. Lett. 2002, 4,
4723 – 4725; d) V. K. Aggarwal, D. K. Dean, A. Mereu, R.
Williams, J. Org. Chem. 2002, 67, 510 – 514; e) C. Yu, L. Hu, J.
Org. Chem. 2002, 67, 219 – 223; f) V. K. Aggarwal, I. Emme, S. Y.
Fulford, J. Org. Chem. 2003, 68, 692 – 700.
[4] a) W. Pei, H.-X. Wei, G. Li, Chem. Commun. 2002, 1856 – 1857;
b) W. Pei, H.-X. Wei, G. Li, Chem. Commun. 2002, 2412 – 2413;
c) D. Balan, H. Adolfsson, J. Org. Chem. 2002, 67, 2329 – 2334;
d) V. K. Aggarwal, A. Mereu, G. J. Tarver, R. McCague, J. Org.
Chem. 1998, 63, 7183 – 7189; e) K. Yang, W. Lee, J. Pan, K. Chen,
J. Org. Chem. 2003, 68, 915 – 919; f) T. Iwamura, M. Fujita, T.
Kawakita, S. Kinoshita, S. Watanabe, T. Kataoka, Tetrahedron
2001, 57, 8455 – 8462; g) M. Shi, C. Wang, Helv. Chim. Acta 2002,
85, 841– 846.
[5] a) G. Li, H.-X. Wei, J. J. Gao, T. D. Caputo, Tetrahedron Lett.
2000, 41, 1– 5; b) A. Patra, S. Batra, B. S. Joshi, R. Roy, B.
Kundu, A. P. Bhaduri, J. Org. Chem. 2002, 67, 5783 – 5788.
[6] a) T. Kataoka, T. Iwama, S. Tsujiyama, Chem. Commun. 1998,
197 – 198; b) T. Kataoka, T. Iwama, S. Tsujiyama, T. Iwamura, S.
Watanabe, Tetrahedron 1998, 54, 11813 – 11824; c) T. Iwama, H.
Kinoshita, T. Kataoka, Tetrahedron Lett. 1999, 40, 3741– 3744;
d) T. Kataoka, H. Kinoshita, T. Iwama, S. Tsujiyama, T.
5056
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 5054 –5056