hazardous) to synthesize. The problem with the use of sodium
chlorodifluoroacetate derives from the high temperature of
its use and because it must be used in large excess (11 equiv
in the case given) to obtain decent conversion of most alkene
substrates. The use of HFPO (a gas) requires both high
temperatures and an autoclave (or sealed tube) environment.
Thus, in evaluating all difluorocarbene reagents currently
available, there obviously remains a need for one that can
add in decent yields to modestly electron deficient alkenes,
such as allylic ethers or terminal alkenes. Addition of
difluorocarbene to highly electron deficient alkenes, such as
acrylic esters, is unprecedented.
Because of our long standing interest in the reactivity of
difluorocyclopropanes, we have always been alert regarding
possible new methods for adding difluorocarbene to alkenes,
and in the course of the development of FSO2CF2COOCH3
as a source of trifluoromethyl copper (Scheme 2), we became
Scheme 4. Synthesis of Difluorocyclopropanes
with unprecedented efficiency, as shown in Scheme 4. These
yields have not been optimized. For example, use of 2 equiv
of TFDA led to an increase in yield to 89% for allyl benzoate
1b.
Although at this point no single “recipe” has been found
that is optimal for all potential alkene substrates, nevertheless,
with a minimum of effort at optimization, we were able to
find satisfactory conditions to difluorocyclopropanate in good
to excellent yield virtually every alkene substrate that we
examined. Although, in this preliminary study, most reactions
were run on a very small (1.6 mmol) scale, the reaction with
n-butyl acrylate 1e, the least reactive of substrates tested,
was also carried out on a 5 g (3.9 mmol) scale at 130 °C,
using 25 mg of NaF as initiator, 16 g (1.6 equiv) of TFDA,
and 3.6 g (1 equiv) of toluene as solvent, to obtain, after
distillation, 6.1 g (89%) of n-butyl 2,2-difluorocyclopropane-
carboxylate.
Scheme 2. Trifluoromethyl Copper Reagent
aware of the potential to modify this reaction in order to
create a new difluorocarbene reagent.10,11 The idea was to
delete the copper and minimize the fluoride ion concentration
by making the reaction a chain reaction, catalytic in fluoride.
The ultimately designed difluorocarbene precursor was
trimethylsilyl fluorosulfonyldifluoroacetate (TFDA),12 which
at moderate temperature, under N2, was added slowly to the
mixture of initiator, olefin, and solvent as shown in Schemes
3 and 4 below.13
Experiments directed at optimization indicated that the
choice of (a) fluoride source, (b) temperature, (c) solvent,
and (d) rate of addition of the TFDA can strongly affect the
yields of reactions with individual substrates. In general, NaF
appears to be the best source of fluoride ion (perhaps because
of its relative insolubility); a temperature of at least 90 °C
is required for efficient reaction with the alkenes studied
(perhaps because of the relatively high activation barrier for
CF2: addition); little or no solvent should be used (neat
reactions are favored for benzoate esters of alkenols, with 1
equiv of methyl benzoate or toluene being very beneficial
for other substrates);14 the optimal addition rate appears to
be ∼0.8 equiv of TFDA per hour.
Scheme 3. Difluorocarbene Formation/Reaction Process
We have reported in this Letter a novel, highly reactive
difluorocarbene reagent, which has been demonstrated to be
effective in difluorocyclopropanating even the most highly
electrophilic alkenes, such as acrylic esters, in excellent
yields. This new methodology should open the door to the
simple synthesis of a wide range of new geminal difluoro-
cyclopropane derivatives that previously were not readily
accessible.
In this preliminary study, a total of five representative
olefins, known to be reluctant substrates with difluorocar-
bene, were examined. The reactions proceeded cleanly and
Acknowledgment. Support of this research in part by the
National Science Foundation is gratefully acknowledged.
(10) Chen, Q. Y.; Wu, S. W. J. Org. Chem. 1989, 54, 3023-3027.
(11) Duan, J.; Dolbier, W. R., Jr.; Chen, Q.-Y. J. Org. Chem. 1998, 63,
9486.
(12) TFDA can be readily synthesized in a yield of 78% by simply adding
3.6 equiv of trimethylsilyl chloride to fluorosulfonyldifluoroacetic acid at
0 °C, stirring overnight, and distilling the product: bp 62-63 °C at 27
mm; 1H NMR, δ 0.40 ppm (s); 13C NMR, δ 155.13 (t, J ) 27.0 Hz), 112.22
(dt, J ) 31.5, 299.0 Hz), -1.05 ppm (s); 19F NMR, δ 40.58 (1F, s), -103.74
ppm (2F, s); HRMS (CI), C5H10O4SSiF3, calcd 251.0012, found 251.0015.
Although neither TFDA nor fluorosulfonyldifluoroacetic acid is commonly
available commercially at this time, they are in principle inexpensive, and
both will certainly become available when the demand for them becomes
obvious. Contact the authors regarding current sources.
OL0055622
(13) A critical aspect of the reaction is that the addition of TFDA be
very slow, first to minimize the concentration of CF2:, but also to control
the inevitable foaming which results from evolution of CO2 and SO2 in the
reaction. Use of a syringe pump (with a Teflon needle) for the small-scale
reactions reported in this Letter proved absolutely necessary.
(14) For some reason, the presence of an aromatic ring, either in the
substrate (as in benzoate esters) or in the small amount of added solvent,
enhances the efficiency of CF2: addition tremendously.
564
Org. Lett., Vol. 2, No. 4, 2000