7) although good conversions were only observed at 08C. The
possibility of using various solvents is important because
commercially available or synthetically prepared organo-
lithium reagents are often sold or prepared specifically in one
of these three solvents.[10]
We next examined the scope of organolithium reagents
(Table 2). Various alkyl lithium reagents (Table 2, entries 1–3)
could be used including tert-butyllithium, which readily added
or an aryl chloride (Table 3, entry 14). The reactivity of 3,3-
difluoropropene 4c derived from 2,2-difluoro-1-indanone is
interesting. Although its reduction using LiAlH4 proceeded
well (Table 3, entry 3), when treated with MeLi, full con-
version was observed but monofluoroalkene 5d was isolated
in low yield (35%) along with a number of unidentified side
products (Table 3, entry 4). We hypothesized that the meth-
ylene protons in 5c, which are both benzylic and allylic, can be
easily abstracted under the reaction conditions leading to
various side reactions. Based on these observations, we
envisioned an alternative route to monofluoroalkene 5d
based on the reduction of 3,3-difluoropropene 4d. As
expected, this reduction worked well and furnished 5d in
76% yield, thus demonstrating that for some problematic
substrates, alternative synthetic routes may be available
(Table 3, entry 5).
Table 2: Scope of organolithium reagents.[a]
Entry
RLi
T [8C]
Product 3
Yield [%][b]
We then studied the halogen-atom effects. We first
examined whether or not two fluorine atoms were required
for the reaction to occur by using 3-fluoropropene 6 with
nBuLi (Scheme 2).[11] The reaction of 6 proceeded smoothly
and 7 was isolated in good yield, thus indicating that the two
fluorine atoms are not essential for reactivity. We next
examined the reactivity of 3-chloro-3-fluoropropene 8 under
identical conditions. Interestingly, the monofluoroalkene 2
was isolated as the major product (65%) while the mono-
chloroalkene 9 was only isolated in 18% yield. In this case,
nucleofuge ability seems to be the controlling factor. Finally,
reaction of 10, the chloro analogue of 1, led to the isolation of
a moderate 38% yield of 9 along with a number of
unidentified products.
1
2
3
4
5
MeLi
sBuLi
tBuLi
21
0
ꢀ78
ꢀ78!21
21
3a
3b
3c
3d
3e
82
76
79
61
80
PhLi
6[c]
21
66
3 f
76
72
7[d]
3g
8
9
66
66
3h
3i
78
30
10[d]
0
3j
73
With respect to the mechanism, the reaction between
organolithium reagents and 3,3-difluoropropenes would pro-
ceed through an SN2’-type pathway where one of the fluorine
atoms would act as a leaving group (Scheme 3). Even though
fluoride is generally regarded as a poor leaving group,[2b] we
propose that in this case, the increase in nucleofuge ability
11[e]
12[d]
LiAlH4 (R=H)
LiEt3BH (R=H)
0
0
3k
3k
97
73
[a] See the Supporting Information for details of the reaction conditions.
[b] Yields after purification by flash chromatography. [c] Reaction was
performed in Et2O. [d] Performed with 3.0 equivalents of the organo-
lithium reagent [e] Performed with 2 equivalents of LiAlH4.
ꢀ
comes from a C F···Li interaction. Indeed, upon addition of
the organolithium reagent to the 3,3-difluoropropene 11, the
lithium atom[12] would interact with the carbon-bound fluo-
rine atom. The lithium atom is classified as a hard atom while
at ꢀ788C. An sp2-based organolithium could also be added
such as vinyllithium, phenyllithium, and (3-trifluoromethyl)-
phenyllithium (Table 2, entries 4–6). In the case of the lithium
reagent derived from N-methylpyrrole, a higher temperature
was required to achieve a good yield (Table 2, entry 7). Even
less-nucleophilic lithiated alkynes could react with 1 although
a higher temperature was required (Table 2, entries 8 and 9)
and led, in the case of phenylacetylene, to a low yield of the
desired product 3i. Interestingly, 2-lithio-1,3-dithiane added
smoothly to 1 in good yield (Table 2, entry 10). Finally,
reduction of 1 was possible using LiAlH4 or LiEt3BH and
furnished the same product 3k in 97% and 73% yield,
respectively (Table 2, entries 11 and 12).
ꢀ
the neutral fluorine atom in a C F bond is also a hard and
effective donor and thus, both should associate strongly. Such
chelation of the lithium atom by a carbon-bound fluorine
atom has been proposed to be a key factor in some
reactions.[4,13,14] Although double chelation as in 12a would
be energetically favored,[13b] reaction would most likely
proceed via 12b as exemplified by the successful reaction of
ꢀ
6 (Scheme 2) where only one C F···Li interaction was
possible. In addition, the fact that the fluoride would be
released as LiF would compensate for its basicity and would
also serve as a driving force for its departure.
In terms of substrate, a variety of 3,3-difluoropropenes, as
shown in Table 3, could be used. These included cyclic
(Table 3, entries 1–7) or acyclic (Table 3, entries 8–14) sub-
strates, thus allowing the preparation of structurally diverse
tri- or tetrasubstituted monofluoroalkene compounds. Nota-
bly, a number of functional groups can be tolerated and
included an alcohol (Table 3, entry 8), a carboxylic acid
(Table 3, entry 12), a Boc-protected amine (Table 3, entry 13),
Finally, to further probe the effect of the lithium, two
additional experiments were performed (Scheme 4). First, the
reaction of two n-butylmetal reagents with 1 was explored.
While nBuZnCl gave no product (0% conversion), the use of
nBuMgBr only gave 37% yield of 2 along with the starting
material 1. The reactivity observed for the different counter-
[15]
ꢀ
ion correlates with the strength of the M F bond. Indeed,
ꢀ1
ꢀ
ꢀ
the Li F bond (138 kcalmol ) is stronger than a Mg F bond
Angew. Chem. Int. Ed. 2011, 50, 11112 –11116
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
11113
Bergeron, Maxime
