Convergent synthesis of fluoro and gem-difluoro compounds using trifluoromethyltrimethylsilane

Article abstract of DOI:10.1016/S0022-1139(99)00158-X

Trifluorotrimethylsilane reacts with acylsilanes to give the corresponding difluoroenoxysilanes via the Brook rearrangement of the alcoholate adducts. The difluoroenoxysilane reacts in situ with various types of electrophilic substrates, leading to gem-di

Full text of DOI:10.1016/S0022-1139(99)00158-X

Journal of Fluorine Chemistry 101 (2000) 193±198
Convergent synthesis of ¯uoro and gem-di¯uoro compounds
using tri¯uoromethyltrimethylsilane
*
Charles Portella , Thierry Brigaud, Olivier Lefebvre, Richard Plantier-Royon
Fac. des Sciences, Laboratoire ``R e actions S e lectives et Applications'', Unit e Mixte de Recherche CNRS-Universit e de Reims (UMR 6519),
BP 1039, 51687 Reims Cedex 2, France
Received 9 May 1999; received in revised form 7 June 1999; accepted 7 June 1999
Abstract
Tri¯uorotrimethylsilane reacts with acylsilanes to give the corresponding di¯uoroenoxysilanes via the Brook rearrangement of the
alcoholate adducts. The di¯uoroenoxysilane reacts in situ with various types of electrophilic substrates, leading to gem-di¯uoro
functionalized derivatives in a one-pot methodology. This paper describes reactions with Michael acceptors, prenyl, benzyl and glycosyl
donors leading to 2,2-di¯uoro-1,5-diketones, 4,4- or 6,6-di¯uorocyclohexenones, o- or p-¯uorophenols, di¯uoro analogues of terpenes, and
di¯uoro-C-glycosides. # 2000 Elsevier Science S.A. All rights reserved.
Keywords: Tri¯uoromethyltrimethylsilane; Acylsilanes; Di¯uoro compounds; Fluorophenols; Terpenes; C-glycosides
1
. Introduction
key intermediate which is isolated or reacted in situ for
various applications.
The selective synthesis of ¯uoro-substituted molecules
This versatile reaction sequence prompted us to consider
the possibility of applying a similar strategy for the synthesis
of compounds bearing a di¯uoromethylene moiety [2] or
even one ¯uorine atom in a de®nite position, via a di¯uor-
oenoxysilane. Several approaches to di¯uoroenoxysilanes
have appeared in the literature since the ®rst one reported by
Ishihara who prepared them by silylation of a zinc di¯uor-
oenolate derived from a chlorodi¯uoromethyl ketone [3].
Other approaches, based on an intermediate trialkylsilyl
tri¯uoromethyl carbinolate adduct similar to ours, used silyl
lithium reagents either to prepare a needed tri¯uoroacetyl-
silane [4] or to add to a tri¯uoromethyl ketone [5]. A
different method consists of the in situ silylation of a
di¯uoroenolate obtained by electroreduction of a tri¯uor-
omethyl ketone [6]. Our strategy needed a tri¯uoromethyl
organometallic reagent. The main group tri¯uoromethyl
derivatives being less easily available [7], we were obviously
attracted by the Ruppert discovery [8] and the further
systematic investigation by Prakash's group [9] of TFMTMS
as a convenient and effective nucleophilic tri¯uoromethyl
donor. Our initial attempts using tetrabutylammonium ¯uor-
ide as activator resulted in the formation of the aldol dimer
of a di¯uoromethyl ketone, an unexpected but interesting
result showing the probable intermediate of the desired
enoxysilane. Using the less nucleophilic tetrabutylammo-
nium di¯uorotriphenylstannate (DFTPS) [10] as activator
strongly depends on the availability of ¯uorinated reagents
and building blocks, and of course on the selectivity of the
chemical transformations. If one can combine a convenient
¯
uorinated reagent, its ef®cient conversion into a versatile
building block, and its easy elaboration towards a target
molecule, if, furthermore, the whole process may be carried
out in the same pot, then the above requirements may be
ful®lled at the same time. We propose such an approach in
which the reagent is tri¯uoromethyltrimethylsilane
TFMTMS), the building block is a di¯uoroenoxysilane,
and the key step for further elaboration is a Lewis acid
activated reaction with an electrophilic substrate.
The historical background of this methodology is sum-
marized in Scheme 1. When an acylsilane reacts with a
per¯uoroorganometallic reagent, an ef®cient and selective
conversion into various organo¯uorosilicon derivatives
occurs. The selectivity of these transformations depends
on the experimental conditions [1]. The ®rst step is a
nucleophilic addition of an F-organomagnesium or an F-
organolithium reagent, the adduct being trapped by water at
low temperature or left to rearrange (Brook rearrangement)
to give, after ¯uoride elimination, the enol silyl ether as a
(
*
Corresponding author. Fax: ‡33-(0)3-2691-3234.
E-mail address: charles.portella@univ-reims.fr (C. Portella)
0022-1139/00/$ ± see front matter # 2000 Elsevier Science S.A. All rights reserved.
PII: S 0 0 2 2 - 1 1 3 9 ( 9 9 ) 0 0 1 5 8 - X

1
94
C. Portella et al. / Journal of Fluorine Chemistry 101 (2000) 193±198
Lewis acid in order to obtain, if possible in catalytic
conditions, interesting yields for the overall one-pot pro-
cess. Although we have studied the Mukaiyama type aldol
reaction [11,12], the remaining part of this report will be
devoted to the applications of this methodology to Michael
acceptors, prenyl and benzyl, or glycosyl donors as electro-
philic substrates.
2. Synthesis of 2,2-difluoro-1,5-diketones,
gem-difluorocyclohexenones, and fluorophenols
The reaction path leading to the title compounds is
depicted in Scheme 3. Ytterbium tri¯ate as Lewis acid gave
the best results, working in catalytic amount. Overall yields
from the starting acylsilanes were satisfactory for prepara-
tive purpose, with aliphatic and aromatic acylsilanes, and
with unsubstituted (methyl vinyl ketone) as well as b-
substituted (cyclohexenone) enones [13].
The 2,2-di¯uoro-1,5-diketones obtained were submitted
to basic annulation conditions (Scheme 4). The regiochem-
istry of the cyclization of the diketone derived from ben-
zoylsilane was not problematic, owing to a unique
enolizable side in the molecule, and 4,4-di¯uoro-3-phe-
nyl-cyclohex-2-enone was cleanly obtained by the catalytic
basic treatment. Despite the presence of two enolizable
sites, the diketone derived from an aliphatic acylsilane
cyclized regiospeci®cally as well, but gave a 6,6-di¯uor-
ocyclohex-2-enone. Therefore, an interesting feature of
these a,a-di¯uorodiketones is the regiospeci®c formation
of the enolate of the di¯uoroketone moiety [13].
Another interesting extension of this chemistry is the
synthesis of ¯uoroaromatic compounds by a simple elim-
ination of hydrogen ¯uoride, followed by the aromatization
of the molecule. When the di¯uoro diketones were treated
by an excess of methanolic potassium hydroxide, the cor-
responding o- or p-¯uorophenol was directly obtained [13].
These examples show the feasibility of a methodology
which allows, from TFMTMS, an acylsilane and an enone,
Scheme 1.
allowed the reaction to stop at the right stage and to
quantitatively convert the acylsilane into the corresponding
di¯uoroenoxysilane (Scheme 2) [11]. The ®rst advantage of
this sequence is that the reaction was completed by using a
catalytic amount of DFTPS, because of the ¯uoride elim-
ination which follows the Brook rearrangement. The second
one is the high ef®ciency of the reaction in various solvents,
in particular in methylene chloride, a non-basic solvent ideal
to perform Lewis acid catalyzed reactions in the same pot.
Hence, we had in hand an overall process which starts with a
chain reaction between the acylsilane and TFMTMS,
initiated by a catalytic amount of ¯uoride, leading to the
di¯uoroenoxysilane which reacts in situ with an electro-
philic reagent to give an elaborated functionalized gem-
di¯uoro compound.
From the investigations carried out so far, it appeared that
each kind of electrophile needed a careful choice of the
Scheme 2.

C. Portella et al. / Journal of Fluorine Chemistry 101 (2000) 193±198
195
Scheme 3.
the regiospeci®c construction of multisubstituted di¯uoro-
cyclohexenones and ¯uorophenols.
The results summarized in Scheme 6 show the versatility
of the method for the synthesis of di¯uoro analogues of
sesquiterpenes including an aryl moiety. In the synthesis of
di¯uoro ar-turmerone, the aryl moiety comes from a ben-
zylation of the di¯uoroenoxysilane derived from an acylsi-
lane bringing the prenylic moiety. On the other hand, in the
synthesis of di¯uorodihydro-ar-curcumene, an aroylsilane
was converted into the corresponding di¯uoroenoxysilane
which reacted with the prenyl donor [15].
An unexpected and interesting result was observed
when we attempted to synthesize a di¯uoro analogue of
farnesol by geranylation. Instead of the farnesol deri-
vative, the only product obtained resulted from the rear-
ranged cyclized intermediate carbocation (Scheme 7) [16].
Furthermore, a long cascade rearrangement was involved,
leading to a compound which can be considered as a
piperitol like derivative, rather than the terpineol like deri-
vatives, compounds generally observed in such rearrange-
ments [17].
3
. Synthesis of difluoroanalogues of terpenes
The introduction of ¯uorine in a terpene molecule can
modify the physico±chemical properties of the molecule
itself or of intermediates, most often carbocations, involved
in their interactions with biological receptors [14]. Our
methodology might contribute favorably to synthetic meth-
ods leading to terpene analogues selectively gem-di¯uoro
substituted.
After several attempts, preparative conditions were found
for the in situ prenylation of the di¯uoroenoxysilane. The
reaction works with different types of acylsilanes [15], but
the most interesting one is acetyltrimethylsilane which leads
to the di¯uoro analogue of the key intermediate ketone
towards the synthesis of monoterpenes. The best results
were obtained with prenyl benzoate and trimethylsilyl tri-
These examples demonstrate that the strategy using
TFMTMS as ¯uorinated source is valuable in the synthesis
of di¯uoro analogues of various types of terpene deri-
vatives.
¯
ate as the Lewis acid catalyst. Unfortunately, the reaction
gave a dif®cult to separate 90/10 mixture of regioisomers,
with the expected one as the major component (Scheme 5).
One or two further steps carried out on the mixture led to
simple gem-di¯uoro mono terpenes. Di¯uoro linalool was
prepared as a pure compound by simple addition of vinyl
magnesium bromide and chromatographic separation of the
regioisomers. A Horner±Wadsworth±Emmons reaction fol-
4. Synthesis of difluoro C-glycosides
A preliminary observation of reaction with acetals [11]
prompted us to study the trapping of the di¯uoroenoxysi-
lanes with glycosyl donors giving access to a di¯uoro C-
glycoside as glycoside mimics. A completely different
lowed by LiAlH reduction led to an analogue of di¯uor-
4
ogeraniol we have not yet managed to separate from its
regioisomer [16].
Scheme 4.

1
96
C. Portella et al. / Journal of Fluorine Chemistry 101 (2000) 193±198
Scheme 5.
approach towards this type of compound was investigated
by Motherwell's group [18±20]. Our ®rst attempt with a
glycosyl donor bearing an ester group at C-2 normally
imposing a C-glycosylation via the b-face failed, leading
to coupling with the non-anomeric ketal carbocation, what-
ever the nature (acetic or benzoic) of the ester (Scheme 8).
Reaction with the anomeric carbon was restored as soon as
the C-2 position was deoxygenated (Scheme 9) [21].
An effective coupling was obtained using tri-o-acetyl D-
glucal as electrophilic substrate. Activation of the reaction
by boron tri¯uoride etherate mildly gave the corresponding
C-glycoside with a high overall yield (Scheme 10) [22]. The
Scheme 6.
Scheme 7.

C. Portella et al. / Journal of Fluorine Chemistry 101 (2000) 193±198
197
Scheme 8.
Scheme 9.
Scheme 10.
5
. Conclusion
Tri¯uoromethyltrimethylsilane, a known convenient and
effective reagent for nucleophilic tri¯uoromethylation of
ketones, must also be considered as a reagent of choice for
the synthesis of various functionalized ¯uoro and di¯uoro
compounds. This method, based on the use of acylsilanes as
starting materials and their ability to rearrange after tri-
Scheme 11.
¯uoromethylation, is advantageous owing to its one-pot and
usefulness of the method for the synthesis of di¯uoro-C-
disaccharides type products is also exempli®ed in Scheme
convergent character, as summarized in Scheme 11.
TFMTMS acts in these transformations as an equivalent
of a di¯uoromethylene dianion. Furthermore, the method is
very versatile owing to the various structure possibilities for
acylsilanes and for electrophiles. We can expect various
applications of the reactions reported here, as well as further
extension of the methodology towards other types of elec-
trophiles. Some of them are currently under investigation.
1
0: reaction of an acylsilane synthesized from D-xylose [23]
with TFMTMS, according to our standard conditions, gave
the corresponding sugar derived di¯uoroenoxysilane which
was trapped by the boron tri¯uoride activated D-ribosyl
donor [22]. The obtained product may be considered as a
di¯uoro analogue of a disaccharide.

1
98
C. Portella et al. / Journal of Fluorine Chemistry 101 (2000) 193±198
[
9] G.K.S. Prakash, A.K. Yudin, Chem. Rev. 97 (1997) 757.
Acknowledgements
[
[
10] M. Gingras, Tetrahedron Lett. 32 (1991) 7381.
11] T. Brigaud, P. Doussot, C. Portella, J. Chem. Soc., Chem. Commun.
The authors thank the CNRS and the MENRT for their
nancial support.
(
1994) 2117.
®
[
[
[
[
[
[
12] T. Brigaud, D. Saleur, J.-P. Bouillon, C. Portella, Synlett. (1999) 432.
13] O. Lefebvre, T. Brigaud, C. Portella, Tetrahedron 54 (1998) 5939.
14] J.-M. Dolence, C.D. Poulter, Tetrahedron 52 (1996) 119.
15] O. Lefebvre, T. Brigaud, C. Portella, Tetrahedron 55 (1999) 7233.
16] O. Lefebvre, T. Brigaud, C. Portella, in press.
References
17] W.F. Erman (Ed.), Chemistry of the Monoterpenes. Part A, Marcel
Dekker, New York, 1985, p. 628.
[
[
[
[
[
1] P. Doussot, C. Portella, J. Org. Chem. 58 (1993) 6675.
2] M.J. Tozer, T. Herpin, Tetrahedron 52 (1996) 8619.
[18] W.B. Motherwell, B.C. Ross, M.J. Tozer, Synlett, 1989, 68.
[19] T.F. Herpin, W.B. Motherwell, M.J. Tozer, Tetrahedron: Asym. 5
(1994) 2269.
3] M. Yamana, T. Ishihara, T. Ando, Tetrahedron Lett. 24 (1983) 507.
4] F. Jin, B. Jiang, Y. Xu, Tetrahedron Lett. 33 (1992) 1221.
5] I. Fleming, R.S. Roberts, S.C.J. Smith, J. Chem. Soc., Perkin Trans.
[20] T.F. Herpin, W.B. Motherwell, J.-M. Weibel, J. Chem. Soc., Chem.
Commun. (1997) 923.
1
(1998) 1215.
6] K. Uneyama, K. Maeda, T. Kato, T. Katagiri, Tetrahedron Lett. 39
1998) 3741.
[
[21] H. Berber, T. Brigaud, O. Lefebvre, R. Plantier-Royon, C. Portella,
submitted.
(
[
[
7] D.J. Burton, Z.-Y. Yang, Tetrahedron 48 (1992) 189.
[22] T. Brigaud, O. Lefebvre, R. Plantier-Royon, C. Portella, Tetrahedron
Lett. 37 (1996) 6115.
8] I. Ruppert, K. Schlich, W. Volbach, Tetrahedron Lett. 25 (1984)
2
195.
[23] R. Plantier-Royon, C. Portella, Tetrahedron Lett. 37 (1996) 6113.