Carbonyl and olefin reactivities for the Baylis–Hillman reaction of
fluorocarbonyls
P. Veeraraghavan Ramachandran,* M. Venkat Ram Reddy and Michael T. Rudd
H. C. Brown and R. B. Wetherill Laboratories of Chemistry, Purdue University, West Lafayette, Indiana 47907-1393,
USA. E-mail: chandran@purdue.edu
Received (in Corvallis, OR, USA) 10th November 2000, Accepted 6th March 2001
First published as an Advance Article on the web 4th April 2001
The product formation and yields for the Baylis–Hillman
reaction of fluorine-containing carbonyl compounds depend
on a balance between the reactivities of the carbonyl and
olefin partners.
fluoral was obtained. Olefin 1d did not yield any BH product 3d
at room or lower temperature.
Reaction of 2,2,3,3,4,4,4-heptafluorobutanal (4), the fluorin-
ated homolog of 2, showed identical reaction patterns with
slightly improved yields of the products. Thus, while 1a and 1b
provided 50 and 70% yield, respectively, of products at 225 °C,
1c provided 22% yield of the product 5c and 1d failed to provide
any product at rt.
The dissimilarity in product yields from activated olefins of
differing reactivity captivated us. We considered a less reactive
perfluorinated aldehyde that does not undergo polymerization
in the presence of a 3°-amine and tested pentafluorobenzalde-
hyde (6) with 1a–d. While the reactions of 1b–d were complete
under neat condition at rt within 2–4 d, 1a polymerized.
However, it reacted in THF at 0 °C within 15 min providing
95% yield of the product [eqn. (3)].
It is well established that the biological properties of medicinal
compounds can often be influenced by fluorine substitution.1
The physical properties of several electronic and optical devices
also depend immensely on the structure of fluoroorganic
molecules.2 Fluorine substitution provides organic chemists
with an opportunity to study an extreme case of electronic effect
in reactions.1,2 As part of our ongoing projects in fluoroorganic
chemistry,3 we examined the Baylis–Hillman (BH) reaction4 of
activated olefins with fluoro-aldehydes and -ketones in the
(1)
presence of 10 mol% of 1,4-diazabicyclo[2.2.2]octane
(DABCO). Herein we report a fascinating relation between the
reactivities of the carbonyl and olefin partners for the BH
reaction of fluorocarbonyls; a match providing optimum yields
of the products, whereas a mismatch resulting in the decomposi-
tion or side reaction of the faster reacting partner.
(3)
Acrolein (1a), methyl vinyl ketone (1b), ethyl acrylate (1c)
and acrylonitrile (1d) were the olefins chosen for the reaction.
Initially we studied the reaction of fluoral (2) with 1a–d. Upon
mixing 2 with 1a at rt under neat conditions, in the presence of
10% DABCO, polymerization of both reactants occurred. Both
of these are known to polymerize in the presence of amines
[eqn. (1)].5,6 We then carried out the reaction in THF and
obtained a very low yield of the expected product along with the
polymerized starting materials. With the hope of arresting the
polymerization, we lowered the reaction temperature to 225 °C
and obtained a 40% yield of the product, 3a [eqn. (2)]. However,
Ordinary ketones undergo Baylis-Hillman reaction occasion-
ally under high pressures4 and activated carbonyls, such as a-
keto esters and hexafluoroacetone undergo relatively fast
reaction.7,8 1,1,1-Trifluoroacetone (8) is known to trimerize in
the presence of amines.9 However, our partial success with 2
persuaded us to carry out its reaction at low temperatures. In
fact, we obtained only 10–12% yield of the products 9a with 1a
in THF at 225 °C, and 9d with 1d at rt. Olefins 1b and 1c failed
to provide any product with 8. We isolated a polymeric material
in both of these cases.
To avoid the polymerization initiated by abstraction of the a-
hydrogen atom, we focused our attention on aromatic tri-
fluoromethyl ketones. The treatment of 2,2,2-trifluoroacetophe-
none (10) with two equiv. of 1a, under neat conditions, at rt did
not provide any product. Decreasing the reaction temperature to
225 °C yielded 15% of acrolein dimer along with its polymer.
Olefin 1b also did not provide any of the expected BH products
at rt, although we obtained a 30% yield of the dimer. Lowering
the temperature resulted only in the suppression of the
dimerization. In contrast, a slow reaction (7 d) between 10 and
1c resulted in 70% yield of the expected allylic alcohol 11c. The
reaction with 1d was faster, complete within 24 h, and provided
94% yield of the product 11d [eqn. (4)].
(2)
we could not suppress the polymerization completely. Further
lowering the temperature had a deleterious effect since
polymerization of both reactants was faster than the BH reaction
at this temperature.
Reaction of 2 with 1b provided the product 3b in 35% yield
under neat conditions, at rt, 1 h. Surprisingly, the yield in THF
at 225 °C was 65%! [eqn. (2)]. However, the reaction of 1c
provided only a 20% yield of the product 3c at rt under neat
conditions [eqn. (2)]. Decreasing the reaction temperature
suppressed the BH reaction completely and only the polymer of
2-(Trifluoroacetyl)thiophene (12) provided similar results.
On testing with 1a–d, it underwent reaction only with 1c within
7 d providing the product 13c in 65% yield, and with 1d, within
24 h, providing the product allylic alcohol 13d in 82% yield
[eqn. (4)].
DOI: 10.1039/b009177b
Chem. Commun., 2001, 757–758
This journal is © The Royal Society of Chemistry 2001
757
Ramachandran
