Reaction of tetramethylammonium fluoride with trifluoromethyltrimethylsilane

Article abstract of DOI:10.1016/S0022-1139(98)00273-5

Tetramethylammonium fluoride reacts with either triluoromethyltrimethylsilane or trimethylsilylacetonitrile in acetonitrile to form a pentacoordinate silicon species, which can act as a source of either fluoride or cyanomethyl carbanion, depending on substrate.

Full text of DOI:10.1016/S0022-1139(98)00273-5

Journal of Fluorine Chemistry 92 (1998) 123±125
Reaction of tetramethylammonium ¯uoride
with tri¯uoromethyltrimethylsilane
Dave J. Adams, James H. Clark*, Liv B. Hansen, Victoria C. Sanders, Stewart J. Tavener
Department of Chemistry, University of York, Heslington, York YO10 5DD, UK
Received 1 July 1998; accepted 30 July 1998
Abstract
Tetramethylammonium ¯uoride reacts with either triluoromethyltrimethylsilane or trimethylsilylacetonitrile in acetonitrile to form a
pentacoordinate silicon species, which can act as a source of either ¯uoride or cyanomethyl carbanion, depending on substrate. # 1998
Elsevier Science S.A. All rights reserved.
Keywords: Ruppert's reagent; Tri¯uoromethane; TMAF; Fluorodenitration
1. Introduction
addition of an equimolar quantity of Me3SiCF3 to a suspen-
sion of anhydrous tetramethylammonium ¯uoride (TMAF)
in acetonitrile caused instantaneous dissolution of all the
¯uoride salt, even at room temperature. This is remarkable
because TMAF has very limited solubility in acetonitrile,
even at elevated temperatures (although it will react with the
solvent over several days, forming the bi¯uoride salt [6]). It
seems unlikely that this increased solubility is due to simple
polarity effects because silanes are generally less polar than
acetonitrile [7]. The dissolution is accompanied by an
exotherm, and evolution of HCF3 gas (19F NMR:
The use of Ruppert's reagent (tri¯uoromethyltrimethyl-
silane, Me3SiCF3) as a convenient source of tri¯uoro-
methide has received considerable attention in recent
years for the tri¯uoromethylation of aldehydes and ketones
[1]. This reagent is normally used in THF solvent, and is
activated by use of a ¯uoride salt. The intermediate in these
reactions is assumed to be a pentacoordinate silicon species,
but this has not yet been con®rmed/identi®ed. Prakash [1]
has studied the reaction between TMAF and Ruppert's
reagent in CD3CN, but reported only the formation of
Me3SiF and CF3D.
Benzotri¯uorides are important molecules for agricultural
and pharmaceutical use due to the lipophilic and electron-
withdrawing effects of the CF3 group, but their synthesis
generally involves the use of HF or SbF3, and so there is a
need for new synthetic routes to these compounds [2,3].
2
ꢀˆ 79 ppm, d, JFH ˆ 79 Hz). 19F NMR spectroscopy
showed that the solution contains no other tri¯uoro-
methyl-containing compounds, indicating that all the [14]
CF3 is lost from the silane. Repeating the experiment in
CD3CN gave DCF3, demonstrating that the reaction
involved deprotonation of the solvent. The reaction is
noticeably slower in CD3CN, suggesting a kinetic isotope
1
effect. H NMR spectroscopy reveals peaks at 0.03 (s, 9H)
and 0.17 (s, 2H) ppm, consistent with methyl and ±CH2CN
groups attached to a silicon anion, respectively. It appears
that attack by the ¯uoride at the silicon centre releases the
strongly basic tri¯uoromethide anion (pKa of HCF3ˆ31 [8]),
which deprotonates a neighbouring acetonitrile molecule.
This in turn reattacks the silicon centre to form a penta-
coordinate complex (Scheme 1). In contrast to Prakash's
report [1], we observed no Me3SiF.
Analysis of the solution by IR shows two pairs of CN
stretches: one for the solvent (2254, 2410 cm 1), and
another for the complex (2174, 2124 cm 1). At room tem-
perature, the ¯uorine attached to the silicon cannot be
2. Results and discussion
We wanted to perform aromatic nucleophilic substitution
using Ruppert's reagent as a source of tri¯uoromethide, with
the aim of producing benzotri¯uorides [4], but aromatic
substitution reactions are best performed in solvents more
polar than THF [5]. In the course of studying the activation
of Ruppert's reagent in polar solvents, we observed that
*Corresponding author. Tel.: +44-1904-432-559; fax: +44-1904-432-
516; e-mail: jhc1@york.ac.uk
0022-1139/98/$ ± see front matter # 1998 Elsevier Science S.A. All rights reserved.
P I I : S 0 0 2 2 - 1 1 3 9 ( 9 8 ) 0 0 2 7 3 - 5

124
D.J. Adams et al. / Journal of Fluorine Chemistry 92 (1998) 123±125
Scheme 1.
observed by 19F NMR spectroscopy, presumably due to
rapid exchange between silicon centres: the pentacoordinate
silicon has a vacant coordination site, so this will be facile.
On cooling to 208C, a broad peak appeared at 60 ppm.
The Si±F bond could be seen at room temperature by
solution IR (950 cm 1), as this technique has a much faster
timescale than NMR. The identity of the complex was
con®rmed by reacting authentic Me3SiCH2CN (from
Aldrich) with TMAF in acetonitrile, which produced a
Fig. 1. Evolution of HCF3 formed by reaction of Me3SiCF3 with
acetonitrile, catalysed by the pentacoordinate silicon complex.
1
homogeneous solution with a H NMR spectrum identical
to that of the complex prepared from Me3SiCF3. In the
absence of ¯uoride and moisture, the silane is stable in
acetonitrile for at least 24 h.
potent soluble ¯uoride source. No ±CH2CN containing
products (other than Me3SiCH2CN) were observed. How-
ever, no reactions occurred for the more demanding 1,3-
dinitrobenzene and 3,5-dinitrobenzonitrile.
For successful preparation of the complex, it is necessary
to use anhydrous TMAF. Replacement of anhydrous TMAF
with TMAFÁH2O did not lead to dissolution of the ¯uoride
salt, although the silane is hydrolysed to form hexamethyl-
disilylether. Addition of Me3SiCF3 to a suspension of dry
KF in acetonitrile also caused dissolution of the ¯uoride salt.
However, in this case, no complex was formed. Instead,
¯uorotrimethylsilane and HCF3 were formed, presumably
along with KCH2CN. Evidently a larger cation is necessary
to stabilise the pentacoordinate silicon complex.
Addition of excess Me3SiCF3 to the complex prepared
from TMAF in solution causes evolution of stoichiometric
[14] quantities of HCF3, and formation of Me3SiCH2CN
(Fig. 1). The complex is therefore an active source of
¯uoride, and there is facile transfer of the ¯uorine in the
complex to the more electropositive silicon centre of the
Me3SiCF3. The complex is, in effect, catalysing the reaction
between the silane and acetonitrile.
The complex reacts with benzophenone by nucleophilic
attack at the carbonyl carbon, followed by elimination,
forming Ph2C=CHCN (12% yield). The reaction between
Me3SiCH2CN and ketones in the presence of ¯uoride is
known to give this product [9], although this was previously
believed to follow a mechanism in which ¯uoride attack at
the silicon centre formed Me3SiF and [CH2CN] . Our work
suggests that the reactive species in this reaction may also be
a pentacoordinate silicon containing species. Reaction of
the complex with 4-nitrobenzophenone, which is activated
towards ¯uorodenitration and attack at the carbonyl
carbon, showed exclusive cyanomethylation at the carbonyl
carbon, indicating that this is far more facile than ¯uoro-
denitration.
Addition of CFCl3 to the solution as an internal reference
for 19F NMR spectroscopy caused a white solid to preci-
pitate. This solid was soluble in water and gave a positive
silver nitrate test, indicating the presence of chloride ion,
3. Experimental
All chemicals and solvents were purchased from Aldrich,
except Me3SiCF3 and TMAF which were purchased from
Apollo Scienti®c. All were used without further puri®ca-
tion, except for KF, which was dried at 3008C for at least
24 h prior to use, and TMAF, which was dried by the method
of Christe et al. [10].
Gas chromatography was carried out on a packed HP5
column in a Hewlett Packard HP6890 gas chromatography.
GC±MS spectra were obtained on a DB5 capillary column
in a Varian 3400 CX gas chromatograph interfaced to a
Finnigan Mat Magnum mass spectrometer. Solution state
NMR spectra were obtained on a JEOL 270 EX270 spectro-
meter (operating at 254 MHz for 19F, referenced to HCF3).
‡
and IR con®rmed the identity of the action as Me4N . This
shows that a halogen exchange reaction is occurring
between the complex and CFCl3, although the identity of
the halomethane product was not con®rmed.
The complex was found to be an active source of
[CH2CN] , as well as ¯uoride, and reacted with a number
of model substrates. Reaction with 2-chloro-6-nitrobenzo-
nitrile at room temperature gave the ¯uorodenitration pro-
duct, 2-chloro-6-¯uorobenzonitrile as the only product
(40% conversion): this is remarkable as ¯uorodenitration
reactions usually require higher temperatures and more
polar solvents [5]. This indicates that the complex is a

D.J. Adams et al. / Journal of Fluorine Chemistry 92 (1998) 123±125
125
IR spectra were recorded on a Bruker Equinox 55 FT-IR
spectrometer.
3.3. Reaction of the pentacoordinated complex with
Me3SiCF3
3.1. Reaction between TMAF and Me3SiCF3
TMAF (0.11 g, 1.18 mmol, anhydrous) was added to
acetonitrile (6 ml, anhydrous) under argon in a 25 ml ¯ask.
Me3SiCF3 (175 ml, 1.18 mmol) was added. This was con-
nected to a pressure gauge (accuracy of 1 Torr) and the
system sealed under argon. The pressure was read and a
known amount of Me3SiCF3 added. The pressure was again
read and more Me3SiCF3 added.
TMAF (0.22 g, 2.3 mmol, anhydrous) was placed in a
50 ml round-bottomed ¯ask under an argon atmosphere.
Acetonitrile (10 ml, anhydrous) was added and the solution
stirred. Using a microsyringe, Me3SiCF3 (0.34 ml,
2.3 mmol) was added. For the IR analysis, only 1.2 ml
acetonitrile was used.
1H-NMR (CD3CN): d 0.03 (s), 0.17 (s) ppm; 19F-NMR
(CD3CN, 208C): d 60 ppm: IR (CH3CN solution, omit-
ting solvent peaks) ꢁ max (cm 1): 3009 (s), 2964 (s) (C±H
stretching); 2174 (s), 2124 (s) (C±N stretching); 1590 (s),
1490 (s) (C±H bending); 947 (s, Si±F stretching).
Acknowledgements
We thank the EPSRC Clean Technology Unit for funding
through a ROPA award (SJT) and the Royal Academy of
Engineering/EPSRC for a fellowship (to JHC). Thanks to
Barbara Chamberlain for running the variable temperature
19F NMR study.
3.2. Reaction of complex with model substrates
For the reaction of the pentacoordinated complex with
various substrates, a bulk solution was formed via the
addition of TMAF (0.11 g, 1.18 mmol, anhydrous) in a
25 ml ¯ask to acetonitrile (6 ml, anhydrous) under argon.
Me3SiCF3 (175 ml, 1.18 mmol) was then added. Two ml of
the above solution was then added to 2-chloro-6-nitroben-
zonitrile (0.0716 g, 0.39 mmol) and biphenyl (0.020 g,
0.13 mmol) in acetonitrile (3 ml, anhydrous) under argon
at room temperature. The reaction was followed by GC. The
reactions with benzophenone and 4-nitrobenzophenone
were carried out similarly.
References
[1] G.K.S. Prakash, A.K. Yudin, Chem. Rev. 97 (1997) 757.
[2] M.A. McClinton, D. McClinton, Tetrahedron 48 (1992) 6555.
[3] J.H. Clark, D. Wails, T.W. Bastock, Aromatic Fluorination, CRC
Press, Boca Raton, 1996.
[4] D.J. Adams, J.H. Clark, L.B. Hansen, V.C. Sanders, S.J. Tavener,
Perkin Trans. 1 (1998), 3081.
[5] J.H. Clark, D.J. Macquarrie, J. Fluorine Chem. 35 (1987)
591.
‡
[6] K.O. Christe, W.W. Wilson, J. Fluorine Chem. 47 (1990)
117.
2-Chloro-6-¯uorobenzonitrile: MS (m/z, %) M 155: 155
(100), 120 (18), 100 (12), 82 (8), 93 (5), 75 (4).
[7] C. Reichardt, Solvents and Solvent Effects in Organic Chemistry,
2nd ed., VCH, Weinheim, 1988.
‡
3,3-Diphenyl-2-propenenitrile: MS (m/z, %) M 205: 205
(100), 165 (42), 178 (28), 190 (18), 51 (21), 76 (16), 88 (14),
102 (14).
3-Phenyl-3-nitrophenyl-2-propenenitrile: MS (m/z, %)
M 250: 250 (100), 203 (77), 176 (36), 165 (16), 155
(14), 77 (12), 220 (9), 151 (7).
[8] S. Andreades, J. Am. Chem. Soc. 86 (1964) 2003.
[9] C. Palomo, J.M. Aizpurua, M.C. Lopez, B. Lecea, Perkin Trans. 1
(1989) 1692.
[10] K.O. Christe, W.W. Wilson, R.O. Wilson, R. Bau, J. Feng, J. Am.
Chem. Soc. 112 (1990) 7619.
‡