Correlation between pKa and reactivity of quinuclidine-based catalysts in the Baylis-Hillman reaction: Discovery of quinuclidine as optimum catalyst leading to substantial enhancement of scope

Article abstract of DOI:10.1021/jo026671s

The reactivity of a variety of quinuclidine-based catalysts in the Baylis-Hillman reaction has been examined, and a straightforward correlation between the basicity of the base and reactivity has been established, without exception. The following order of

Full text of DOI:10.1021/jo026671s

Cor r ela tion betw een p Ka a n d Rea ctivity of Qu in u clid in e-Ba sed
Ca ta lysts in th e Ba ylis-Hillm a n Rea ction : Discover y of
Qu in u clid in e a s Op tim u m Ca ta lyst Lea d in g to Su bsta n tia l
En h a n cem en t of Scop e
Varinder K. Aggarwal,* Ingo Emme, and Sarah Y. Fulford
University of Bristol, School of Chemistry, Cantock’s Close, BS8 1TS Bristol, U.K.
v.aggarwal@bristol.ac.uk
Received November 6, 2002
The reactivity of a variety of quinuclidine-based catalysts in the Baylis-Hillman reaction has been
examined, and a straightforward correlation between the basicity of the base and reactivity has
been established, without exception. The following order of reactivity was established with pKa’s
of the conjugate acids (measured in water) given in parentheses: quinuclidine (11.3), 3-hydroxy-
quinuclidine (9.9), DABCO (8.7), 3-acetoxyquinuclidine (9.3), 3-chloroquinuclidine (8.9), and
quinuclidinone (7.2). The higher than expected reactivity of DABCO, based on its pKa, was analyzed
by comparing the relative basicity of DABCO and 3-acetoxyquinuclidine in DMSO. It was found
that in aprotic solvent, DABCO was 0.6 pKa units more basic than 3-acetoxyquinuclidine, thus
establishing a direct link between pKa of the amine and its reactivity. In contrast to previous
literature work that reported the contrary, quinuclidine, which has the highest pKa, was found to
be the most active catalyst. The reaction profile with quinuclidine showed significant autocatalysis,
which suggested that the presence of proton donors might further enhance rates. Thus, a series of
additives bearing polar X-H bonds were investigated and it was found that methanol, triethano-
lamine, formamide, and water all provided additional acceleration. Methanol was found to be
optimum, and the powerful combination of quinuclidine with methanol was tested with a host of
aldehydes and Michael acceptors. Not only were the reactions more efficient and faster than
previously reported, but now new substrates that were previously unreactive could be employed.
Notable examples include the use of acetylenic aldehydes and the employment of vinyl sulfones,
acrylamides, δ-lactones, and even R,â-unsaturated esters bearing a â-substituent.
SCHEME 1. Ca ta lytic Cycle of th e Ba ylis-Hillm a n
Rea ction
In tr od u ction
The Baylis-Hillman reaction1 is an exquisite reaction
as simple starting materials are converted into densely
functionalized products in a catalytic process without
generating waste or byproducts (Scheme 1).2 However,
the reaction has traditionally suffered from low reaction
rates and limited substrate scope. There has, therefore,
been considerable interest in enhancing reaction rate as
through this endeavor both practicality and scope can be
improved.
The rate-determining step (RDS) of the Baylis-Hill-
man reaction is the reaction between the ammonium
enolate 1 and the aldehyde (Scheme 1).3 Thus, increasing
the amount of the enolate or activation of the aldehyde
will result in increased rates. We first demonstrated that
the use of lanthanides resulted in modest rate enhance-
ment and that further acceleration could be achieved by
the addition of alcohol ligands, e.g., binol and triethanol-
amine.4 We believed that the primary source of accelera-
tion in these systems was due to the enhanced acidity of
the OH group of the additive as a result of metal binding,
(1) Baylis, A. B.; Hillman, M. E. D. Offenlegungsschrift 2155113,
1972; U.S. Patent 3,743,669; Chem. Abstr. 1972, 77, 34174q.
(2) For reviews, see: (a) Langer, P. Angew. Chem., Int. Ed. 2000,
39, 3049-3052. (b) Ciganek, E. Org. React. 1997, 51, 201-350. (c)
Basavaiah, D.; Rao, P. D.; Hyma, R. S. Tetrahedron 1996, 52, 8001-
8062. (d) Drewes, S. E.; Roos, G. H. P. Tetrahedron 1988, 44, 4653-
4670.
(4) (a) Aggarwal, V. K.; Tarver, G. J .; McCague, R. Chem. Commun.
1996, 2713-2714. (b) Aggarwal, V. K.; Mereu, A.; Tarver, G. J .;
McCague, R. J . Org. Chem. 1998, 63, 7183-7189. (c) For other
examples of metal-accelerated Baylis-Hillman reactions, see: Kawa-
mura, M.; Kobayashi, S. Tetrahedron Lett. 1999, 40, 1539-1542.
(3) Hill, J . S.; Isaacs, N. S. J . Phys. Org. Chem. 1990, 3, 285-288.
10.1021/jo026671s CCC: $25.00 © 2003 American Chemical Society
Published on Web 01/04/2003
692
J . Org. Chem. 2003, 68, 692-700

Quinuclidine-Based Catalysts in the Baylis-Hillman Reaction
F IGURE 1. The original catalysts used by Baylis and Hill-
man.
F IGURE 2. Previously reported activities of quinuclidine-
based catalysts in the Baylis-Hillman reaction with pKa’s.
which resulted in enhanced hydrogen-bonded activation
of the aldehyde.4b We5 and others6 have also shown that
enhanced rates could be achieved by conducting reactions
in water or formamide. In these highly polar solvents,
we believe that rate acceleration is achieved by not only
hydrogen-bonded activation of the aldehyde but also by
increasing the amount of the zwitterionic intermediate
1 by solvation.
Most methods for promoting the Baylis-Hillman reac-
tion have largely focused on activation of the aldehyde.
We have also considered methods for increasing the
amount of the ammonium enolate 1 through stabilization
of the ammonium ion. This should provide increased
concentrations of the reactive intermediate without hav-
ing a negative impact on the rate of the subsequent
reaction between the enolate and aldehyde as the enolate
is not stabilized.7 We showed that amidines8 (DBU in
particular) and guanidines8,9 provided substantial in-
creases in rate, and we believe this occurs by stabilization
of the ammonium ion through delocalization. However,
while good rates were achieved with DBU, the reaction
was limited to non-enolizable aldehydes.
A good measure of an amine’s basicity is the pKa of its
conjugate acid. DBU has a high pKa10 but is also sterically
hindered, a feature that usually results in severely
reduced rates. The high rate increase observed with DBU
indicates that, in this specific case, the basicity of the
amine is more important than its steric hindrance.
However, in the absence of steric effects, there should
be a good correlation between reaction rate and pKa, as
amines with high pKa should give higher concentrations
of the ammonium enolate.
F IGURE 3. Catalysts investigated in the Baylis-Hillman
reaction: DABCO 4, quinuclidine 5, 3-hydroxyquinuclidine 7,
3-acetoxyquinuclidine 8, 3-chloroquinuclidine 9, and 3-quinu-
clidinone 10.
DABCO and quinuclidine13a,b and the superiority of
DABCO over quinuclidine13c were confirmed, and indeed,
for some time 3-hydroxyquinuclidine was regarded as the
optimum catalyst for the Baylis-Hillman reaction. Thus,
on the basis of the literature reports, 3-hydroxyquinu-
clidine 7 was better than DABCO 4, which was better
than quinuclidine 5 (Figure 2). However, in comparing
the three quinuclidine-based catalysts, for which steric
differences would be expected to be insignificant, the
expected correlation between pKa14 and reaction rate was
clearly not apparent. On the basis of pKa, quinuclidine
should have been the best catalyst but was reported to
be the worst. We therefore tested a broader range of
quinuclidine-based catalysts to examine whether a cor-
relation between rate and pKa existed and have reexam-
ined the activity of quinuclidine. We now report that
there is a direct correlation between pKa and reaction rate
and that quinuclidine is in fact the best catalyst to date.
This new discovery provides substantial improvements
in rate and a very substantial increase in scope of the
Baylis-Hillman reaction.
In the original paper by Baylis and Hillman, three
amine catalysts were tested, DABCO 4, quinuclidine 5,
and pyrrocoline 6, but their relative efficiencies were not
clearly stated (Figure 1).1,11
Long reaction times have been reported using DABCO,
and in the search for more active catalysts, Drewes found
that 3-hydroxyquinuclidine 7 showed considerably faster
rates.12 The superiority of 3-hydroxyquinuclidine over
Resu lts
A series of reactions were conducted between 2-pyridi-
necarboxaldehyde (1 equiv),15 methyl acrylate (1.2 equiv)
and different quinuclidine-based catalysts (Figure 3).
Reactions were conducted neat, and 5 mol % catalyst was
employed to ensure homogeneity of all of the reaction
(5) Aggarwal, V. K.; Dean, D. K.; Mereu, A.; Williams, R. J . Org.
Chem. 2002, 67, 510-514.
(6) (a) Yu, C. Z.; Liu, B.; Hu, L. Q. J . Org. Chem. 2001, 66, 5413-
5418. (b) J enner, G. High Pres. Res. 1999, 16, 243-252. (c) Auge´, J .;
Lubin, N.; Lubineau, A. Tetrahedron Lett. 1994, 35, 7947-7948.
(7) Stabilisation of the enolate would also provide increased con-
centrations of 1 but would also result in reduced reactivity of 1.
(8) Aggarwal, V. K.; Mereu, A. Chem. Commun. 1999, 2311-2312.
(9) Leadbetter has also shown that guanidines are good catalysts:
Leadbeater, N. E.; van der Pol, C. J . Chem. Soc., Perkin Trans. 1 2001,
2831-2835.
(13) (a) Marko´, I. E.; Giles, P. R.; Hindley, N. J . Tetrahedron 1997,
53, 1015-1024. (b) Bailey, M.; Marko´, I. E.; Ollis, D.; Rasmussen, P.
R. Tetrahedron Lett. 1990, 31, 4509-4512. (c) Ciganek, E. Org. React.
1997, 51, 207. In this review, personal communications between S. E.
Drewes, N. D. Emslie, and Ciganek are cited.
(14) (a) The pKa values of quinuclidine, 3-hydroxyquinuclidine,
3-acetoxyquinuclidine, 3-chloroquinuclidine, and 3-cyanoquinuclidine
were taken from: Grob, C. A. Helv. Chim. Acta 1985, 68, 882-886.
(b) The pKa values of 3-quinuclidinone and DABCO were taken from:
Hine, J .; Chen, Y.-J . J . Org. Chem. 1987, 52, 2091-2094.
(15) 2-Pyridinecarboxaldehyde was chosen as 3-hydroxyquinuclidine
showed the highest solubility in it.
(10) pKa refers to the pKa value of the conjugate acid of the amine
catalysts. The pKa of DBU is 11.3. Ho¨fle, G.; Steglich, W.; Vorbruggen,
H. Angew. Chem., Int. Ed. Engl. 1978, 17, 569-583.
(11) The catalysts appear in separate claims of the patent in the
order DABCO, pyrrocoline and quinuclidine. It is not clear if this order
reflects the order or importance of the catalysts.
(12) (a) Ameer, F.; Drewes, S. E.; Freese, S.; Kaye, P. T. Synth.
Commun. 1988, 18, 495-500. (b) Drewes, S. E.; Freese, S. D.; Emslie,
N. D.; Roos, G. H. P. Synth. Commun. 1988, 18, 1565-1572.
J . Org. Chem, Vol. 68, No. 3, 2003 693

Aggarwal et al.
F IGURE 5. Hydrogen-bonded diazabicyclo[2.2.2]octane am-
monium salt and 3-acetoxyquinuclidine ammonium salt.
which has the highest pKa, provided the fastest rate, and
quinuclidinone, which has the lowest pKa, gave the lowest
rate. We were surprised at the high reactivity of quinu-
clidine, as previous reports13 had indicated that it was a
poor catalyst. Attempts to reproduce the literature results
by using different quality catalysts (sublimed, used
directly from the commercial supplies, aged by leaving
open to air, and aged over an atmosphere of CO2) were
all unsuccessful: all forms of the catalyst were almost
equally effective. The correlation between pKa and rate
applied to all substrates except DABCO, which showed
a higher than expected rate based on pKa, but as DABCO
possesses two basic nitrogens, it was possible that its
higher effective molarity was responsible for its enhanced
rate. However, even at lower loadings (2.5 mol %)
DABCO still gave faster rates than both 3-acetoxy-
quinuclidine and 3-chloroquinuclidine.
We were intrigued as to the origin of the increased
reactivity of DABCO, which had a pKa of 8.5, over
3-acetoxyquinuclidine, which had a pKa of 9.3. The pKa
values of the quinuclidines have been determined in
water and we suspected that the pKa of DABCO might
be lowered as a result of hydrogen bonding of the second
nitrogen with water (Figure 5). In contrast, hydrogen
bonding between the acetate of 3-acetoxyquinuclidine and
water is likely to be much weaker, and therefore,
hydrogen bonding will not have a significant effect on
its pKa. We therefore wanted to determine the relative
basicity of DABCO and 3-acetoxyquinuclidine in aprotic
media, the medium in which our Baylis-Hillman reac-
tions were conducted. To achieve this, we measured the
equilibrium between a 1:1 mixture of DABCO‚HBF4 and
3-acetoxyquinuculidine (Scheme 2) and, as a control, a
1:1 mixture of 3-acetoxyquinuclidine‚HBF4 and DABCO
(see the Supporting Information for full details).16 Ir-
respective of which salt we started with we obtained the
same equilibrium ratio in favor of the DABCO salt (2.0:
1), thus demonstrating that DABCO was indeed a
stronger base than 3-acetoxyquinuclidine in aprotic
solvent.17 A ratio of 2.0:1 correlates to an increase in pKa
of DABCO relative to 3-acetoxyquinuclidine by 0.6.18
Using the same technique, we also established that
3-hydroxyquinuclidine was a stronger base than DABCO
F IGURE 4. Effect of different quinuclidine-based catalyst on
the rate of the Baylis-Hillman reaction. Conversion represents
the ratio of the Baylis-Hillman product: unreacted aldehyde
1
(determined by H NMR). The reaction was performed using
quinuclidine (pink asterisks, QD), 3-hydroxyquinuclidine (green
circles, 3-HQD), diazabicyclo[2.2.2]octane (dark red triangles,
DABCO), 3-acetoxyquinuclidine (blue X, 3-AcOQD), 3-chloro-
quinuclidine (yellow squares, 3-ClQD), and 3-quinuclidinone
(red diamonds, QD)O).
TABLE 1. Effect of th e Ba sicity of Qu in u clid in e-Ba sed
Ca ta lyst on th e Ra te of th e Ba ylis-Hillm a n Rea ction
ratea
(%/min)
b
catalyst
quinuclidine (QD) 5
pKa
krel
1.8
11.3c 9.0
3-hydroxyquinuclidine (3-HQD) 7
3-acetoxyquinuclidine (3-AcOQD) 8
3-chloroquinuclidine (3-ClQD) 9
8.8 × 10-1
3.1 × 10-2
8.2 × 10-3
9.9d 4.3
9.3
8.9
8.5e
6.9
0.15
0.04
1
diazabicyclo[2.2.2]octane (DABCO) 4 2.1 × 10-1
3-quinuclidinone (QDdO) 10
1.3 × 10-3
0.006
a
b
Initial rates (<20% conversion). All pKa’s were determined
in water.14 c Quinuclidine was found to be more basic than
d
3-hydroxyquinuclidine in DMSO by 0.4 pKa units. 3-Hydroxyqui-
nuclidine was found to be more basic than DABCO in DMSO by
0.3 pKa units. e DABCO was found to be more basic than 3-ac-
etoxyquinucidine in both DMSO and CD3CN by 0.6 pKa units.
(16) The equilibrium ratio was determined by NMR in CD3CN and
-
DMSO-d6. The BF4 salt was employed as it was considerably more
soluble than the Cl- salt in CD3CN. Furthermore, some hydrolysis of
the acetate occurred upon treatment of 3-acetoxyquinuclidine with
aqueous HCl. No hydrolysis occurred upon treatment of the same base
with HBF4 in Et2O.
mixtures (3-hydroxyquinuclidine was the least soluble,
and it was only possible to dissolve 5 mol % of this
catalyst in the reaction mixture). Aliquots were removed
at various time intervals, and the extent of the reaction
was determined by NMR. The results are presented in
Figure 4 and Table 1. From this simple analysis, it was
clear that there was a broad correlation between reaction
rate and pKa of the quinuclidine base. Quinuclidine,
(17) A second independent test verified our supposition that hydrogen-
bonding of the second nitrogen of DABCO was responsible for its
reduced pKa relative to 3-acetoxyquinuclidine in protic media. When
reactions were conducted in water, 3-acetoxyquinuclidine showed
similar rates to DABCO but when the DABCO concentration was
halved (to take into account the presence of two nitrogens), the reaction
using 3-acetoxyquinuclidine was faster.
694 J . Org. Chem., Vol. 68, No. 3, 2003

Quinuclidine-Based Catalysts in the Baylis-Hillman Reaction
SCHEME 2. Deter m in a tion of Rela tive Ba sicity of
DABCO a n d 3-Acetoxyqu in u clid in e
F IGURE 6. Proposed intramolecular hydrogen bonding in-
volved in Baylis-Hillman reactions using 3-hydroxyquinucli-
dine.
(1.5:1 ratio, increase in pKa by 0.3) and that quinuclidine
was a stronger base than 3-hydroxyquinuclidine (1.6:1
ratio, increase in pKa by 0.4) in aprotic media (see the
Supporting Information for full details). With these
“corrections” in the pKa’s, the correlation between pKa and
reaction rate applied to all of the quinuclidine-based
catalyst without exception. This correlation between rate
of the Baylis-Hillman reaction and pKa of the base will
be an extremely useful parameter in the design of
alternative quinuclidine-based catalysts, which can of
course be extended to all amine catalysts, including chiral
ones. However, for more general amine catalysts, rates
of reaction will depend on both the basicity of the amine
and steric hindrance around the nitrogen atom.
F IGURE 7. Reaction profiles of quinuclidine, 3-hydroxyqui-
nuclidine (pink asterisks, QD), and 3-hydroxyquinuclidine
(green circles, 3-HQD).
for reactions conducted in water 3-hydroxyquinuclidine
was still superior to DABCO,5 indicating that hydrogen
bonding was not the primary reason for enhanced rates.
From that observation and the current study, it is clear
that, rather than hydrogen bonding, it is the higher pKa
of 3-hydroxyquinuclidine relative to DABCO that is
responsible for its enhanced reactivity. The hydroxyl
group of 3-hydroxyquinuclidine does still play a role in
enhancing rates although only a minor one.
A closer analysis of the reaction profile of quinuclidine
revealed a sigmoidal curve with a point of inflection
instead of a parabolic curve (Figure 7). This implies
autocatalysis; i.e., the product of the reaction is enhancing
the rate. The origin of this rate acceleration is presum-
ably the hydroxyl of the product, which activates one or
both of the components as described above. In fact,
autocatalysis should be observed with other amine
catalysts that are devoid of hydroxyl groups, but as far
as we are aware, this has not been previously reported
for the Baylis-Hillman reaction. Indeed, pronounced
autocatalysis was observed with the other quinuclidine
based catalysts: 3-acetoxyquinuclidine, DABCO, and
3-chloroquinuclidine (Figure 4). 3-Hydroxyquinuclidine
showed much less pronounced autocatalytic behavior
presumably because a hydroxyl group was already present
at the start of the reaction. Thus, to enhance rates further
with quinuclidine as catalyst, hydroxyl groups needed to
be present at the start of the reaction. Thus, reactions
were conducted in the presence of protic additives:19
water, methanol, formamide and triethanolamine and the
results are presented in Figure 8. Further substantial
rate enhancements were indeed observed with all of the
additives particularly methanol, formamide and trietha-
nolamine.20 We decided to focus further efforts on metha-
nol as it is easier to evaporate and may aid solubility of
The origin of rate acceleration of 3-hydroxyquinuclidine
over DABCO was initially believed to be due to an
intramolecular hydrogen bond between the hydroxyl
group and the enolate 11,12 but subsequent modeling
studies showed substantial strain in such an inter-
mediate.13c Since then, the model has been revised and
the origin of the rate of acceleration has been ascribed
to the ability of 3-hydroxyquinuclidine to protonate the
developing zwitterionic intermediate intermolecularly 12
(Figure 6).2b,c However, our previous work showed that
(18) The increase in pKa for DABCO relative to 3-acetoxyquinucli-
dine was calculated as follows:
[B1H+][B2] h [B1][B2H+]
[B1][B2H+]
K )
[B1H+][B2]
Ka(B2)
K )
Ka(B1)
[B1][B2H+]
pKa(B2) - pKa(B1) ) -log10
[B1H+][B2]
[0.34][0.34]
pKa(B2) - pKa(B1) ) -log
10[0.17][0.17]
pKa(B2) - pKa(B1) ) 0.6
J . Org. Chem, Vol. 68, No. 3, 2003 695

Aggarwal et al.
F IGURE 9. Effect of the amounts of catalyst on the rate of
the Baylis-Hillman reaction using 3 mol % (pink circles), 10
mol % (blue asterisks), 25 mol % (red X), 50 mol % (green
triangles), 75 mol % (yellow squares), and 1 equiv (purple
diamonds) of quinuclidine.
F IGURE 8. Effect of different additives on the rate of the
Baylis-Hillman reaction using no additive (dark blue dia-
monds), 0.25 equiv of water (purple asterisk), 0.25 equiv of
formamide (blue triangles), 0.25 equiv of methanol (dark red
circles), 0.75 equiv of methanol (pink X), 1.50 equiv of methanol
(black dashes), and 0.10 equiv of triethanolamine (green
squares).
providing adducts in shorter reaction times and higher
yields than has previously been observed. p-Methoxy-
benzaldehyde (entry 4) was, as expected, found to be a
slow reacting aldehyde and gave similar results in terms
of yield and reaction time to the use of 1 equiv of 1,8-
diazabicyclo[5.4.0]undec-7-ene. The electron-deficient aro-
matic aldehydes in entries 5 and 6 were solid aldehydes
that posed additional problems as such reactions could
not be conducted neat. With these substrates a minimum
amount of dimethylformamide was added to the initial
suspension to dissolve all of the reactants and the
reactions were monitored as before. Again, higher yields
and shorter reaction times were achieved with our
improved system. Cinnamaldehyde provided another
dramatic example of the efficiency of the new conditions
(entry 9): a 62% yield after 3 h was obtained whereas
the previous highest yield was 43% after 24 h.
Acetylenic aldehydes24 could also be employed for the
first time although curiously, the methyl ether 13 was
obtained when the reaction was conducted in the pres-
ence of methanol whereas the normal Baylis-Hillman
adduct 14 was obtained just using quinuclidine (Scheme
3). Treatment of 14 under the same Baylis-Hillman
conditions provided the methyl ether 13 but only in a
low yield (40%, plus 60% starting material), suggesting
that there may be additional pathways for its formation
in the Baylis-Hillman reaction conducted in the presence
of methanol.
any solid starting materials, which can be problematic
for the reactions that are conducted neat.21 Optimization
of the amount of methanol revealed that 0.75 equiv gave
faster rates than 0.25 and 1.5 equiv. Variation in the
amount of quinuclidine revealed that higher catalyst
loading did not necessarily lead to higher rates: optimum
rates were achieved with 25, 50, and 75 mol % and so
the lower loading of quinuclidine was selected for further
study (Figure 9).
Having established the optimum conditions for the
Baylis-Hillman reaction, we explored the scope and
limitations of the reaction. Where possible, we have
compared our results with the current best results
reported. It should be noted that our conditions usually
employ 25 mol % catalyst whereas most of the literature
examples require 1 equivalent or even greater amounts
of the amine catalyst.
We initially investigated reactions of methyl acrylate
with a variety of aldehydes (Table 2). Simple aliphatic
and aromatic aldehydes worked well (entries 1-4),
(19) The use of protic additives to enhance the rates has been
described previously although not in the context of autocatalysis; see
refs 6c and 12, also: Basaviah, D.; Sarma, P. K. S. Synth Commun.
1990, 20, 1611; Bode, M. L.; Kaye, P. T. Tetrahedron Lett. 1991, 32,
5611-5614. Oishi, T.; Oguri, H.; Hirama, M. Tetrahedron: Asymmetry
1995, 6, 1241-1244.
(20) 2-Propanol and trifluoroethanol were also tested as additives,
but they showed less rate enhancement than methanol.
(21) A reviewer suggested the possibility that the amine deproto-
nated the alcohol (methanol or Baylis-Hillman product) and that the
resulting alkoxide was the true catalyst under these conditions.
However, in the attempted reaction between methyl acrylate and
benzaldehyde, sodium methoxide in methanol did not give any Baylis-
Hillman product, showing that methoxide itself is unable to catalyse
this reaction.
(22) Rosa, N. J .; Afonso, C. A. M.; Santos, A. G. Tetrahedron 2001,
19, 4189-4194.
(23) Kundu, M. K.; Mukherjee, S. B.; Balu, N.; Padmakumar, R.;
Bhat, S. V. Synlett 1994, 444.
(24) Activated acetylenic fluoro ketones have been employed in the
Baylis-Hillman reaction: Venkat-Ram-Reddy, M.; Rudd, M. T.; Veer-
araghavan-Ramachandran, P. J . Org. Chem. 2002, 67, 5382-5385.
696 J . Org. Chem., Vol. 68, No. 3, 2003

Quinuclidine-Based Catalysts in the Baylis-Hillman Reaction
TABLE 2. Ba ylis-Hillm a n Rea ction of Ald eh yd es w ith Meth yl Acr yla tea
a
Conditions employed: (A) aldehyde (1 equiv), methyl acrylate (1.2 equiv), quinuclidine (0.25 equiv), methanol (0.75 equiv); (B) aldehyde
(1 equiv), methyl acrylate (1.2 equiv), quinuclidine (1 equiv), methanol (0.75 equiv): (C) aldehyde (1 equiv), methyl acrylate (1.2 equiv),
quinuclidine (0.5 equiv), methanol (0.75 equiv), DMF (minimum amount). Literature conditions: (a) 1:3:1 aldehyde/acrylate/DABCO,
1,4-dioxane/H2O (1:1, v/v), rt; (b) 1:1:1 aldehyde/acrylate/DBU, rt; (c) 1:1.1:1 aldehyde/acrylate/DABCO, rt, 100 µL/mmol [bmim][PF6]; (d)
1:10:0.5 aldehyde/acrylate/DABCO, microwave (domestic oven, 650 W).
SCHEME 3. Ba ylis-Hillm a n Rea ction of
P r op a r gylic Ald eh yd es
Acrylamides are very poor Michael acceptors and until
earlier this year30 had not succumbed to the Baylis-
Hillman reaction. The recently reported conditions em-
ployed DABCO (1 equiv) in water and moderate to high
yields were achieved, but only with highly reactive
aldehydes. We found that using our conditions but with
additional methanol or DMF to dissolve the acrylamide,
faster reactions were achieved (Table 4, entries 1-3).
Furthermore, less reactive electrophiles could now be
employed (e.g. benzaldehyde, entry 4) whereas in the
(25) Hill, J . S.; Isaacs, N. S. Tetrahedron Lett. 1986, 27, 5007-5010.
(26) Reddy, L. R.; Rao, K. R. Org. Prep. Proced. Int. 2000, 32, 185-
188.
Reactions with acrylonitriles were also tested and
again higher yields at shorter reaction times were
achieved (Table 3). Among others, previous conditions
have employed high pressure to achieve high yields after
very short reaction times, but our simple chemical system
was once again found to be highly effective for all
substrates tested.
(27) (a) Kundu, M. K.; Bhat, S. V. Synth. Comm. 1999, 29, 93-101.
(b) Kundu, M. K.; Sundar, N.; Kumar, S. K.; Bhat, S. V.; Biswas, S.;
Valecha, N. Bioorg. Med. Chem. Lett. 1999, 9, 731-736.
(28) Hill, J . S.; Issacs, N. S. J . Chem. Res., Miniprint 1988, 10,
2641-2676.
(29) Basavaiah, D.; Gowriswari, V. V. L. Synth. Commun. 1987, 17,
587-592.
(30) Yu, C.; Hu, L. J . Org. Chem. 2002, 67, 219-223.
J . Org. Chem, Vol. 68, No. 3, 2003 697

Aggarwal et al.
TABLE 3. Ba ylis-Hillm a n Rea ction of Ald eh yd es w ith Acr ylon itr ilea
a
Conditions employed: (A) aldehyde (1 equiv), acrylonitrile (1.2 equiv), quinuclidine (0.25 equiv), methanol (0.75 equiv). Literature
conditions: (a) 1:1:1 aldehyde/acrylonitrile/DABCO, 5 kbar; (b) 1:1:1 aldehyde/acrylonitrile/DABCO, â-cyclodextrin complex, solid-state
conditions; (c) 1:1.2:1 aldehyde/acrylontirile/3-HQD, H2NCHO (5 equiv), Yb(OTf)3 (0.05 equiv); (d) no condition or yield given; (e) 1:1:0.06
aldehyde/acrylonitrile/DABCO; (f) 1:1.5:0.15 aldehyde/acrylonitrile/DABCO. *After 20 min rapid decomposition was observed.
TABLE 4. Ba ylis-Hillm a n Rea ction of Ald eh yd es w ith Acr yla m id ea
a
Conditions employed: (a) aldehyde (1 equiv), acrylamide (1 equiv), quinuclidine (0.5 equiv) in methanol; (b) aldehyde (1 equiv),
acrylamide (1 equiv), quinuclidine (1 equiv) in methanol. Literature conditions: 1:1:1 DABCO/aldehyde/acrylamide, 5 mL of water, rt; (c)
compared with 3-hydroxy-2-methylene-3-(5-hydroxymethylfuran-2-yl)propionamide.
previous system they were inert. Other slow-reacting
substrates have also been examined (vinyl sulfones), and
once again, our new catalyst system was found to be
markedly superior to existing processes (Table 5).31 This
is most dramatically illustrated in the reaction between
phenyl vinyl sulfone and benzaldehyde. Using DABCO,
57% yield was achieved after 21 days, but under our
conditions we obtained 63% yield after just 1 day (entry
2).
We were even able to extend the Baylis-Hillman
reaction to a new class of Michael acceptors, R,â-unsatur-
ated δ-lactones, which gave good yields of the corre-
sponding adducts with aromatic aldehydes (Table 6).33
Finally the power of the new methodology is illustrated
with methyl crotonate, a substrate which only partici-
pates in the Baylis-Hillman reaction under microwave
irradiation,23 and is even unreactive at 10 kbar pres-
sure.25 Using quinuclidine and methanol, at ambient
pressure, methyl crotonate reacted with 2-pyridinecar-
boxaldehyde, furnishing the Baylis-Hillman adduct in
25% yield after 21 days (Table 6). We are currently
exploring the combination of the newly discovered chemi-
cal method for accelerating the Baylis-Hillman reaction
with known physical methods (high pressure/microwave
(33) A chiral R,â-unsaturated γ-lactone has recently been used in
the Baylis-Hillman reaction. Standard conditions were ineffective, but
the method of Hu (ref 6a) was successfully employed. A mixture of R,
γ, and R + γ “aldol” products was obtained in nearly racemic form
from enantiomerically pure lactone (15% ee measured by rotation),
which shows that a significant pathway is via enolization by DABCO
rather than conjugate addition: Franck, X.; Figade`re, B. Tetrahedron
Lett. 2002, 43, 1449-1451.
(31) Under standard conditions, employing methanol as additive,
an inseparable side product was obtained. Side products were also
observed when DMF was used, but these were minimised when dioxane
was utilised.
(32) Auvray, P.; Knochel, P.; Normant, J . F. Tetrahedron 1998, 44,
6095-6106.
698 J . Org. Chem., Vol. 68, No. 3, 2003

Quinuclidine-Based Catalysts in the Baylis-Hillman Reaction
TABLE 5. Ba ylis-Hillm a n Rea ction of Ald eh yd es w ith
similar reactivity. It is believed that the second nitrogen
of DABCO is involved in hydrogen bonding in H2O, which
results in a lower than expected pKa. The origin of the
rate acceleration of 3-hydroxyquinuclidine over DABCO
had previously been ascribed to hydrogen bonding but it
is now believed that its higher pKa is the primary factor
for its increased rate. Not only can the correlation
between pKa and rate be used to predict the reactivity of
other potential catalysts but it can also be incorporated
into the design of new chiral catalysts for the asymmetric
Baylis-Hillman reaction. In this analysis, the parent
compound quinuclidine, which had the highest pKa of all
the quinuclidines and had previously been reported to
be a poor catalyst, was reevaluated and found to be the
best catalyst to date.
P h en yl Vin yl Su lfon ea
The reactions of all the quinuclidine-based catalysts
devoid of hydroxyl groups showed significant autocataly-
sis. The origin of the autocatalysis was the hydroxyl
group of the Baylis-Hillman product, which promoted
the reaction through hydrogen bonding. This observation
led to the addition of hydrogen bond donors to further
enhance the rate of the reaction. A number of additives
were effective (methanol, formamide, triethanolamine,
water) and methanol was found to be optimum. This new
combination of quinuclidine and methanol is the most
effective system to date for catalyzing the Baylis-
Hillman reaction and was found to be general for a broad
range of substrates. Reaction of a variety of aldehydes
with methyl acrylate and acrylonitrile was found to give
adducts in higher yield and shorter reaction times than
any previous catalytic system. The same was observed
with acrylamides, and vinyl sulfones. Furthermore, the
new system allowed previously unreactive partners to
couple (benzaldehyde with acrylamide) and allowed new
Michael acceptors to be employed (R,â-unsaturated δ-lac-
tones and crotonates) for the first time. No doubt this
new system will find broad application in synthesis.
a
Conditions employed: (E) aldehyde (2.5 equiv), phenyl vinyl
sulfone (1 equiv), quinuclidine (0.25 equiv); (F) aldehyde (1 equiv),
phenyl vinyl sulfone (1.2 equiv), quinuclidine (0.25 equiv), dioxane
(minimum amount). Literature conditions: (b) phenyl vinyl sulfone
(1 equiv), aldehyde (large excess), DABCO (0.1 equiv).
TABLE 6. Ba ylis-Hillm a n Rea ction of Ald eh yd es w ith
5,6-d ih yd r o-2H-p yr a n -2-on e a n d Meth yl Cr oton a tea
Exp er im en ta l Section
Gen er a l Meth od s. Chromatography: Flash chromatogra-
phy was performed on silica gel (400-630 mesh). TLC was
performed on aluminum-backed silica plates precoated with
silica (0.2 mm), which were developed using standard visual-
izing agents: UV fluorescence (254 and 366 nm), ∆, PMA/∆,
potassium permanganate. Infrared spectra: Only selected
absorbances (νmax) are reported. 1H NMR spectra: These were
recorded at 400 MHz. Chemical shifts (δH) are quoted in parts
per million (ppm), referenced to the appropriate residual
solvent peak. Coupling constants (J ) are reported to the
nearest 0.1 Hz. 13C NMR spectra: These were recorded at 100
MHz. Chemical shifts (δC) are quoted in ppm referenced to the
appropriate solvent peak. Mass spectra: Only the molecular
ions (M+) and major peaks are reported with intensities quoted
as percentages of the base peak. Solvents and reagents:
Commercial grade solvents were dried and purified by stan-
dard procedures as specified in Purification of Laboratory
Chemicals, 3rd ed. (Perrin, D. D.; Armarego, W. L. F. Perga-
mon Press: New York, 1988). Amine catalysts were obtained
from commercial sources and were sublimed prior to use,
although it was later found that they could be employed
without further purification. All aldehydes were either distilled
or sublimed before use. Acrylonitrile and methyl crotonate
were distilled before use. Methyl acrylate, phenyl vinyl sulfone,
5,6-dihydro-2H-pyran-2-one, and acylamide were obtained
from commercial sources and used directly.
a
Conditions employed: (B) aldehyde (1 equiv), 5,6-dihydro-2H-
pyran-2-one (1.2 equiv), quinuclidine (1 equiv), methanol (0.75
equiv); (D) aldehyde (1 equiv), 5,6-dihydro-2H-pyran-2-one (1.2
equiv), quinuclidine (0.25 equiv). Literature conditions: (b) no
previous optimum results, new compounds.
irradiation) to reduce reaction times with these especially
difficult substrates.
In summary, we have found that there is a direct
correlation between pKa and activity of quinuclidine-
based catalysts in the Baylis-Hillman reaction: the
higher the pKa, the faster the rate. Presumably, high pKa
provides enhanced stabilization of the intermediate am-
monium enolate, resulting in its increased concentration
without compromising its reactivity, which in turn leads
to faster rates. Even though the pKa’s are measured in
H2O and the Baylis-Hillman reactions are conducted in
aprotic media, the correlation nevertheless applied to all
quinuclidine-based catalyst except for DABCO. Analysis
of the pKa of DABCO (8.7 in H2O) revealed that while it
was indeed a weaker base than 3-acetoxyquinuclidine
(9.3 in H2O) in H2O, in DMSO the situation was reversed.
This correlated with reactivity: DABCO was superior to
3-acetoxyquinuclidine in aprotic media but in H2O showed
Gen er a l P r oced u r e for th e Ba ylis-Hillm a n Rea ction .
To a stirred mixture of the substrate aldehyde (1.0 mmol) and
J . Org. Chem, Vol. 68, No. 3, 2003 699

Aggarwal et al.
Michael acceptor (1.2 mmol) were added quinuclidine (0.25
mmol) and methanol (0.75 mmol). The homogeneous reaction
mixture was stirred at ambient temperature, and the reaction
progress was monitored by NMR. Upon completion or as
indicated, the reaction mixture was purified by flash column
chromatography on silica gel, eluting with diethyl ether and
petroleum ether to give the desired product.
3-Tr iisopr opylsilyleth yn yl-3-m eth oxy-2-m eth ylen epr o-
p ion ic Acid Meth yl Ester (13). A mixture of triisopropylsi-
lanylpropynal (127 mg, 0.603 mmol), methyl acrylate (62.3 mg,
0.724 mmol), quinuclidine (17 mg, 0.15 mmol), and methanol
(19.6 mg, 0.603 mmol) was stirred for 30 min at room
temperature. Column chromatography on silica (Et2O/hexane
1:2) gave 131 mg (70%) of 13 as a colorless oil (Rf ) 0.8, Et2O/
hexane 1:2): IR (neat) 2943, 2866, 1729, 1640, cm-1; 1H NMR
(CDCl3) δ 1.05-1.12 (m, 21 H), 3.44 (s, 3 H), 3.79 (s, 3 H),
5.07 (s, 1 H), 6.23 (s, 1 H), 6.38 (s, 1 H); 1H NMR (C6D6) δ
0.98-1.18 (m, 21 H), 3.31 (s, 3 H), 3.34 (s, 3 H), 5.17 (s, 1 H),
6.23 (s, 1 H), 6.29 (s, 1 H); 13C NMR (CDCl3) δ 11.2, 18.6, 52.1,
55.9, 69.3, 89.3, 102.9, 127.6, 138.0, 165.9; 13C NMR (C6D6) δ
11.3, 18.6, 51.2, 55.9, 69.6, 88.3, 104.4, 126.1, 138.9, 165.3; MS
m/z 295 (4) [M+ - CH3], 279 (20), 267 (38), 133 (100), 105 (83).
Anal. Calcd for C17H30O3Si: C, 65.76; H, 9.74; O, 15.46; Si,
9.05. Found: C, 65.84; H, 10.10.
mmol), isobutyraldehyde (540 mg, 7.49 mmol), and quinucli-
dine (83 mg, 0.75 mmol) was stirred for 4 days at room
temperature. Column chromatography on silica (Et2O/CH2Cl2/
hexane 10:70:30) gave 245 mg (40%) of the product as a
colorless oil (Rf ) 0.3, Et2O/CH2Cl2/hexane, 10:70:30): IR (neat)
1
3502, 2965, 1388, 1367, 1291 cm-1; H NMR (CDCl3) δ 0.76
(d, J ) 6.8 Hz, 3 H), 0.89 (d, J ) 6.8 Hz, 3 H), 1.97 (dqq, J )
6.5 Hz, 6.5 Hz, 6.5 Hz, 1 H), 2.32 (br s, 1 H), 4.08 (d, J ) 6.4
Hz, 1 H), 6.07 (s, 1 H), 6.46 (s, 1 H), 7.53-7.59 (m, 2 H), 7.62-
7.67 (m, 1H), 7.87-7.92 (m, 2 H); 13C NMR (CDCl3) δ 16.7,
19.5, 32.1, 74.4, 125.7, 128.3, 129.3, 133.7, 139.3, 152.5; MS
(CI) m/z 241 (32) [MH+], 223 (100), 197 (20), 143 (58); HRMS
(CI) calcd for C12H17O3S 241.0894, found 241.0898. Anal. Calcd
for C12H16O3S: C, 59.98; H, 6.71; O, 19.97; S, 13.34. Found:
C, 59.69; H, 6.87.
3-(H yd r oxyp h e n ylm e t h yl)-5,6-d ih yd r op yr a n -2-on e
(Ta ble 6, En tr y 1). A mixture of 5,6-dihydro-2H-pyran-2-one
(293 mg, 2.98 mmol), benzaldehyde (264 mg, 2.49 mmol), and
quinuclidine (69 mg, 0.62 mmol) was stirred for 3 days at room
temperature. Column chromatography on silica (Et2O/CH2Cl2/
hexane 8:50:30) gave 340 mg (70%) of the product as a colorless
oil (Rf ) 0.5. Et2O): IR (neat) 3412, 2902, 1693 cm-1; 1H NMR
(CDCl3) δ 2.46 (m, 2 H), 3.55 (d, J ) 4.8 Hz, 1 H), 4.29-4.40
(m, 2 H), 5.57 (d, J ) 4.8 Hz, 1 H), 6.69 (t, J ) 4.4 Hz, 1 H),
7.28-7.45 (m, 5 H); 13C NMR (CDCl3) δ 24.2, 66.3, 72.3, 126.7,
127.9, 128.5, 135.0, 140.8, 164.8; MS (EI) m/z 204 (52) [M+],
186 (32), 127 (43), 105 (100), 77 (74); HRMS (EI) calcd for
3-Hyd r oxy-2-m eth ylen e-5-tr iisop r op ylsila n ylh ex-1-en -
4-yn oic Acid Meth yl Ester (14). A mixture of methyl acrylate
(166 mg, 1.94 mmol), triisopropylsilanylpropynal (336 mg, 1.61
mmol), and quinuclidine (45 mg, 0.40 mmol) was stirred for
30 min at room temperature. Column chromatography on silica
(hexane/Et2O 4:1) gave 273 mg (55%) of 14 as a colorless oil
(Rf ) 0.40, hexane/Et2O 1:1): IR (neat) 3464, 2944, 2866, 1714
C
12H12O3 204.0782, found 204.0786.
2-(Hyd r oxyp yr id in -2-ylm eth yl)bu t-2-en oic Acid Meth -
yl Ester (Ta ble 6, En tr y 3). A mixture of methyl crotonate
(126 mg, 1.25 mmol), pyridine-2-carboxaldehyde (112 mg, 1.04
mmol), and quinuclidine (116 mg, 1.04 mmol) was stirred for
3 weeks at room temperature. Column chromatography on
silica (hexane/EtOAc 7:3) gave 54 mg (25%) of the two isomers
of the product in a 63:37 ratio as colorless oil (Rf ) 0.50, 0.55
(hexane/EtOAc, 1:1), which could be partly separated using
column chromatography on silica: IR (neat) 3388, 3059, 2951,
1708 cm-1; MS (EI) m/z 207 (21) [M+], 190 (100), 108 (55), 78
(63); HRMS (EI) calcd for C11H13NO3 207.0891, found 207.0895.
Anal. Calcd for C11H13NO3: C, 63.76; H, 6.32; N, 6.76, O, 23.16.
Found: C, 63.53; H, 6.54; N, 6.62.
Major isomer: 1H NMR (CDCl3) δ 2.05 (d, J ) 7.3 Hz, 3 H),
3.69 (s, 3 H), 4.90 (br s, 1 H), 5.47 (br s, 1 H), 6.42 (q, J ) 7.3
Hz, 1 H), 7.19-7.25 (m, 1 H), 7.34 (d, J ) 7.8 Hz, 1 H), 7.68
(dt, J ) 7.8, J ) 1.5, 1 H), 8.55 (d, J ) 4.9 Hz, 1H); 13C NMR
(CDCl3) δ 15.8, 51.3, 74.4, 121.0, 122.5, 134.4, 136.9, 140.3,
148.0, 160.1, 167.4.
1
cm-1; H NMR (CDCl3) δ 1.07 (s, 21 H) 2.98 (br s, 1 H), 3.81
(s, 3 H), 5.25 (s, 1 H), 6.16 (d, J ) 0.7 Hz, 1 H), 6.31 (s, 1 H);
13C NMR (CDCl3) δ 11.1, 18.6, 52.2, 62.7, 88.2, 104.9, 126.7,
139.5, 166.4; MS (FAB) m/z 295 [M+] (15), 279 (100), 253 (70),
237 (10). Anal. Calcd for C16H28O3Si: C, 64.82; H, 9.52; O,
16.19; Si, 9.47. Found: C, 64.72; H, 9.31.
2-(F u r a n -2-ylh yd r oxym eth yl)a cr ylon itr ile (Ta ble 3,
en tr y 3). A solution of furfuraldehyde (340 mg, 3.54 mmol),
acrylonitrile (225 mg, 4.25 mmol), quinuclidine (98.4 mg, 0.88
mmol), and methanol (85.0 mg, 2.65 mmol) was stirred for 20
min at room temperature. Column chromatography on silica
(hexane/Et2O 2:1) gave 531 mg (78%) of the product as an
orange liquid (Rf ) 0.3): IR (neat) 3418, 2230, 1625, 1502, 1397
cm-1; 1H NMR (CDCl3) δ 2.45 (br s, 1H), 5.37 (s, 1H), 6.15 (d,
J ) 1.2 Hz, 1H), 6.20 (d, J ) 1.2 Hz, 1H), 6.40 (dd, J ) 3.1 Hz,
J ) 1.7 Hz, 1H), 6.44 (dd, J ) 3.1 Hz, J ) 0.9 Hz, 1H), 7.44
(dd, J ) 1.7 Hz, J ) 0.9 Hz, 1H); 13C NMR (CDCl3) δ 67.9,
108.8, 110.8, 116.6, 123.5, 131.2, 143.5, 151.6; MS m/z 149 (40)
[M+], 132 (90), 97 (100), 53 (9); HRMS (EI) calcd for C8H7NO2
149.0476, found 149.0477.
Minor isomer: 1H NMR (CDCl3) δ 2.00 (d, J ) 7.3 Hz, 3 H),
3.65 (s, 3 H), 4.98 (br s, 1 H), 5.79 (br s, 1 H), 7.15 (q, J ) 7.3
Hz, 1 H), 7.17-7.21 (m, 1 H), 7.37 (d, J ) 8.0 Hz, 1 H), 7.68
(dt, J ) 8.0 Hz, J ) 2.0 Hz, 1 H), 8.54 (d, J ) 4.9 Hz, 1 H); 13
C
3-F u r a n -2-yl-3-h yd r oxy-2-m e t h yle n e p r op ion a m id e
(Ta ble 4, En tr y 2). A solution of acryl amide (230 mg, 3.23
mmol), furfuraldehyde (311 mg, 3.23 mmol), and quinuclidine
(150 mg, 1.62 mmol) in methanol (250 µL) was stirred for 5 h
at room temperature. Column chromatography on silica
(EtOAc to EtOAc/MeOH 10:1) gave 355 mg (66%) of the
product as a colorless solid (Rf ) 0.45, EtOAc; mp 83-84 °C):
IR (neat) 3375, 3188, 1660, 1630, 1604, 1433, 1142, 1011, 950,
NMR (CDCl3) δ 14.5, 51.7, 69.0, 120.1, 122.1, 133.6, 136.8,
141.8, 148.1, 161.0, 167.4.
Ack n ow led gm en t. We thank the German Academic
Exchange Service (DAAD), Bristol University, EPSRC,
and ICI for support. We thank Dr. Richard Williams
for carrying out the initial studies and Professor Roger
Alder for helpful discussions.
734 cm-1 1H NMR (DMSO-d6) δ 5.49 (d, J ) 5.8 Hz, 1 H),
;
5.65 (s, 1 H), 5.74 (d, J ) 5.8 Hz, 1 H), 5.86 (s, 1 H), 6.13 (d,
J ) 3.1 Hz, 1 H), 6.36 (dd, J ) 3.1 Hz, J ) 1.8 Hz, 1 H), 7.04
(br s, 1 H), 7.53 (br s, 1 H), 7.56 (d, J ) 1.8 Hz, 1 H); 13C NMR
(DMSO-d6) δ 65.4, 106.9, 110.8, 118.8, 142.6, 145.6, 156.4,
169.1; MS m/z 167 (23) [M+], 150 (24), 122 (100), 95 (79), 68
(41). Anal. Calcd for C8H9NO3: C, 57.48; H, 5.43; N, 8.38; O,
28.71. Found: C, 57.33; H, 5.63; N, 8.22.
Su p p or tin g In for m a tion Ava ila ble: 1H NMR spectra for
all compounds, reaction profiles for the quinuclidine-based
1
catalysts, and H NMR spectra showing the equilibrium ratio
of the HBF4 salts used to determine the relative basicity of
the quinuclidine-based catalysts in aprotic media. This
material is available free of charge via the Internet at
http://pubs.acs.org.
2-(4-Meth yl-3-h ydr oxypen ten yl) P h en yl Su lfon e (Table
5, En tr y 1). A mixture of phenyl vinyl sulfone (504 mg, 3.00
J O026671S
700 J . Org. Chem., Vol. 68, No. 3, 2003