The Baylis-Hillman reaction in supercritical carbon dioxide: Enhanced reaction rates, unprecedented ether formation, and a novel phase-dependent 3-component coupling

Article abstract of DOI:10.1039/b111347j

The Baylis-Hillman reaction can be efficiently carried out in scCO₂ with enhanced reaction rates relative to comparable solution phase reactions; at low pressures, a novel dimerisation is observed which has led to the development of a novel one

Full text of DOI:10.1039/b111347j

The Baylis–Hillman reaction in supercritical carbon dioxide: enhanced
reaction rates, unprecedented ether formation, and a novel
phase-dependent 3-component coupling
Paul M. Rose, Anthony A. Clifford and Christopher M. Rayner*
School of Chemistry, University of Leeds, Leeds, UK LS2 9JT. E-mail: chrisr@chem.leeds.ac.uk
Received (in Cambridge, UK) 14th December 2001, Accepted 8th March 2002
First published as an Advance Article on the web 3rd April 2002
The Baylis–Hillman reaction can be efficiently carried out in
scCO2 with enhanced reaction rates relative to comparable
solution phase reactions; at low pressures, a novel dimerisa-
tion is observed which has led to the development of a novel
one-pot three component coupling reaction to form highly
functionalised ethers.
scCO2. Reactions on a 1 mmol scale in a 20 ml high pressure
view cell (approx. 0.05 M aldehyde in scCO2)15 were
homogeneous under all relevant conditions, at least during the
initial stages of the reaction. If a higher reagent concentration
was used (e.g. 3 mmol, 0.15 M aldehyde in scCO2), a liquid
layer was clearly visible at the bottom of the reactor, indicating
lack of homogeneity, and allowing the possibility that a neat
reaction may be occurring in the bottom of the reaction vessel
under an atmosphere of CO2.
In accord with observations from previous work, the reaction
proceeded best with electrophilic aldehydes.1,2 Yields could be
further enhanced by working at higher concentrations (entry 3)
whilst still retaining homogeneity, or using a larger excess of
acrylate (entry 5). It is important to note that in comparable
control experiments in conventional solvents, none of the
substrates investigated gave significant conversions ( < 5%) at
similar concentrations (0.05 M) in either toluene or acetonitrile,
thus demonstrating the unusual nature of the reaction medium
and rate accelerations possible.
We have thus demonstrated that the B–H reaction can be
efficiently carried out in scCO2 with a range of electron-
deficient aromatic aldehydes. However, in the course of this
work, we noticed, on occasions, that small quantities of by-
products were being formed. The outcomes of reactions
performed in scCO2 are known to vary with pressure, and in
some cases, very substantial beneficial effects can be obtained
by tuning the reaction medium to optimize a particular reaction
pathway.14,16,17 We thus decided to investigate the effect of
pressure on an individual reaction in more detail, the results of
which are shown in Fig. 1.
The Baylis–Hillman (B–H) reaction is a synthetically useful
carbon–carbon bond-forming reaction between an aldehyde and
an electrophilic alkene, usually in the presence of a tertiary
amine.1 One of its main attractions is the high degree of
functionality present in the products and their resultant potential
transformations. However the typical B–H reaction is notori-
ously slow in liquid solution unless particularly reactive
substrates are chosen, and often fails to reach completion due to
unfavourable equilibria.2
The reaction is usually catalysed by a tertiary amine, most
commonly 1,4-diazabicyclo[2.2.2]octane (DABCO), but also
3-hydroxyquinuclidine (HQD)3 and derivatives thereof,4 diaza-
bicyclo[5.4.0]undecene (DBU)5 and phosphines.6 The reaction
can be accelerated by Lewis acids and/or various protic
additives or solvents,7 alongside physical methods such as
microwave irradiation8 and ultrasound.9 Recent reports have
also shown that the use of ionic liquids as solvents can be
beneficial.10 The B–H reaction is known to have a very high
negative activation volume, and rates are known to be increased
by high pressures (typically 5–15 kbar)11 as can stereocontrol in
appropriate cases.12
These factors, along with the high diffusion rates, possible
clustering effects,13 and consequent rate accelerations observed
in supercritical fluids suggested to us that this reaction would be
an ideal candidate for investigation in scCO2, which is of
increasing importance as an environmentally benign reaction
medium.14
The points indicated by squares in Fig. 1 indicate the
percentage of the Baylis–Hillman product (3) in the crude
reaction mixture. At high pressures, the conversion is small, and
similar to the yield obtained in toluene solution at comparable
For our initial studies we chose to investigate DABCO (1) as
catalyst, with methyl acrylate and a variety of aromatic
aldehydes as substrates. The results of these studies are
summarised in Table 1. To ensure consistency and have most
potential for efficient reaction, we aimed for conditions where
the initial reagents would form a homogeneous solution in
Table 1 The Baylis–Hillman reaction in scCO2
Entrya
X
Conversion (%)b
1
H
< 5
2
3
4
5d
6
CF3
CN
NO2
NO2
17
40 (35, 38c)
54 (48)
64 (62)
56 (48)
Pyridin-4-al
a Reaction performed in 20 ml view cell using aldehydes (1 mmol), methyl
acrylate (2 mmol) and DABCO (1 mmol). b Isolated yields in parentheses.
c Reaction performed at double concentration. d 4 eq. methyl acrylate.
Fig. 1 Effect of pressure of scCO2 on the product distribution between the
Baylis–Hillman product (3) and dimer (4).
968
CHEM. COMMUN., 2002, 968–969
This journal is © The Royal Society of Chemistry 2002

concentrations. As the pressure is reduced, the yield increases to
a maximum of around 50%, clearly demonstrating that this
effect is very different to the high-pressure accelerations
previously reported.11,12 An explanation similar to that used by
Leitner et al., based on ideas of concentration, can be used to
explain our observations, with a high CO2 density mimicking
more dilute solutions with consequent rate retardation.18
However, other effects may also be operating, and this is
currently under investigation.19
Scheme 1 One-pot three-component coupling by Baylis–Hillman reaction
and in situ etherification.
The additional products present in the reaction mixtures were
identified as dimers of the original Baylis–Hillman adducts,
with both possible stereoisomers, the meso and C2-symmetrical,
being formed (Table 2).20 The points indicated by triangles in
Fig. 1 show the total conversion of the aldehyde to the Baylis–
Hillman product (3) and its dimer (4). Note that the extent of
dimerisation increases as the pressure decreases, with almost all
the starting aldehyde reacting at the lowest pressures investi-
gated (cf. Table 1).
etherification, however further studies to confirm the mecha-
nism, and to expand the scope of the reaction are currently under
way.
In summary, the Baylis–Hillman reaction can be efficiently
carried out in scCO2. Enhanced reaction rates are observed
relative to comparable solution phase reactions. At low
pressures, an unprecedented dimerisation is observed which has
led to the development of a novel one-pot three component
coupling reaction to form highly functionalised ethers derived
from B–H products. The results described here provide another
important example of how the unique properties of scCO2 can
lead to the development of unprecedented reactions of im-
portance to the synthetic chemistry community.
Table 2 Formation of dimeric ethers from Baylis–Hillman products
We are very grateful to the following members of the Leeds
Cleaner Synthesis Group and their respective companies for
funding and useful discussions: Drs P. Ducouret, A. Guerreiro,
and A. Van-Sickle, Aventis; Drs M. Loft and J. Strachan,
GlaxoSmithKline; and Dr L. Harris, Pfizer Central Research.
We also thank the EPSRC and the University of Leeds for
funding.
Entrya
X
Pressure/MPa
Ratio (meso+ )c
Yield (%)
1
2
3
4
NO2
CN
NO2
CNb
13.7
13.6
8.93
9.27
58+42
69+31
58+42
63+37
82
78
52
45
b
Notes and references
1 E. Ciganek, Org. React., 1997, 51, 201; D. Basavaiah, P. D. Rao and R.
S. Hyma, Tetrahedron, 1996, 52, 8001.
2 Y. Fort, M. C. Berthe and P. Caubere, Tetrahedron, 1992, 48, 6371.
3 S. E. Drewes, S. D. Freese, N. D. Emslie and G. H. P. Roos, Synth.
Commun., 1988, 18, 1565.
a Combined isolated yield of both diastereoisomers. b Reaction carried out
over 72 h directly from aldehydes without isolation of initial B–H product.
c Assigned using chiral lanthanide shift reagents.20
4 A. G. M. Barrett, A. S. Cook and A. Kamimura, Chem. Commun., 1998,
2533.
To the best of our knowledge, such a dimerisation is totally
unprecedented in B–H chemistry. The yields for such a reaction
could be optimized by isolating the initial B–H product, and
resubjecting it to the standard reaction conditions at higher
pressure (13.7 MPa, 2000 psi), in the presence of molecular
sieves to remove the water by-product formed. Using this
procedure, good yields of dimer can be obtained (Table 2,
entries 1 and 2). Lower, but still acceptable yields, can be
obtained directly from the original B–H precursors at lower
pressures (entries 3 and 4).
5 V. K. Aggarwal and A. Mereu, Chem. Commun., 1999, 2311.
6 S. Rafel and J. W. Leahy, J. Org. Chem., 1997, 62, 1521.
7 V. K. Aggarwal, D. K. Dean, A. Mereu and R. Williams, J. Org. Chem.,
2002, 67, 510; C. Yu, B. Liu and L. Hu, J. Org. Chem., 2001, 66,
5413.
8 M. K. Kundu, S. B. Mukherjee, N. Balu, R. Padmakumar and S. V. Bhat,
Synlett, 1994, 444.
9 G. H. P. Roos and P. Rampersadh, Synth. Commun., 1993, 23, 1261.
10 J. N. Rosa, C. A. M. Afonso and A. G. Santos, Tetrahedron, 2001, 57,
4189.
A closer analysis of the phase behaviour of the reaction in
Fig. 1 suggests the origin of the ether formation. In general, at
pressures around 11.0 MPa and below, as the reaction proceeds,
the initial B–H product begins to separate as an oil at the bottom
of the reactor. This is also the pressure below which significant
dimerisation is observed. This suggests that the dimerisation
reaction is occurring in the liquid layer at the bottom of the
reactor, which may also help to explain why it has not been
observed in previous studies of the B–H reaction. It is in effect,
a solventless reaction occurring under an atmosphere of
scCO2.21
A potentially more useful variant of this reaction would be if
a different alcohol could be used in the etherification step for
form an unsymmetrical ether. This would represent a novel one-
pot 3-component coupling protocol by tandem B–H reaction
and subsequent etherification. Our preliminary studies indicate
that this process is indeed feasible (Scheme 1).
11 J. S. Hill and N. S. Isaacs, J. Chem. Res., 1988, 330.
12 T. Oishi, H. Oguri and M. Hirama, Tetrahedron Asymm., 1995, 6,
1241.
13 J. F. Brennecke and J. E. Chateauneuf, Chem. Rev., 1999, 99, 433.
14 R. S. Oakes, A. A. Clifford and C. M. Rayner, J. Chem. Soc., Perkin
Trans. 1, 2001, 917; See also P. G. Jessop and W. Leitner, Chemical
Synthesis using Supercritical Fluids, Wiley-VCH,Weinheim, 1999.
15 As with all reactions under high pressure, appropriate safety precautions
must be taken. See reference 14 for further information.
16 See for example: R. S. Oakes, A. A. Clifford, K. D. Bartle, M. Thornton-
Pett and C. M. Rayner, Chem. Commun., 1999, 247.
17 N. Shezad, A. A. Clifford and C. M. Rayner, Tetrahedron Lett., 2001,
42, 323.
18 A. Fürstner, L. Ackermann, K. Beck, H. Hori, D. Koch, K. Langemann,
M. Liebl, C. Six and W. Leitner, J. Am. Chem. Soc., 2001, 123, 9000.
19 K. A. Lee, P. M. Rose, C. M. Rayner and A. A. Clifford, unpublished
results.
20 The two diastereoisomers could be separated by column chromatog-
raphy. In the presence of europium tris[3-(heptafluoropropylhydroxy-
methylene)-(+)-camphorate], the 300 MHz 1H NMR spectrum of the
C2-symmetric isomer in CDCl3 showed two clear singlets for the ester
methyl groups, whereas no such splitting could be induced on the meso
isomer.
Interestingly, the reaction works well from the initial B–H
precursors, and is particularly efficient with p-nitrobenzylalco-
hol, which we believe is due to its relative insolubility in scCO2
compared to benzyl alcohol, which is consistent with the neat
reaction explanation. Preliminary investigations suggest that
both DABCO and high pressure CO2 are required for efficient
21 G. Rothenberg, A. P. Downie, C. L. Raston and J. L. Scott, J. Am. Chem.
Soc., 2001, 123, 8701.
CHEM. COMMUN., 2002, 968–969
969