Novel reactions with the underutilized BrF3: The chemistry with nitriles

Article abstract of DOI:10.1016/S0022-1139(01)00448-1

Bromine trifluoride is a relatively new reagent in organic and fluoroorganic chemistry. Under the right conditions, it can be a useful tool and generate new types of reactions and compounds. In order to react properly, it has to be complexed by at least t

Full text of DOI:10.1016/S0022-1139(01)00448-1

Journal of Fluorine Chemistry 111 ꢀ2001) 161±165
Novel reactions with the underutilized BrF3
The chemistry with nitriles
*
Shlomo Rozen , Dalit Rechavi, Aviv Hagooly
School of Chemistry, Raymond and Beverly Sackler Faculty of Exact Sciences, Tel-Aviv University, Tel-Aviv 69978, Israel
Received 12 March 2001; accepted 28 June 2001
Abstract
Bromine tri¯uoride is a relatively new reagent in organic and ¯uoroorganic chemistry. Under the right conditions, it can be a useful tool
and generate new types of reactions and compounds. In order to react properly, it has to be complexed by at least two, close in space,
electron donating elements, after which it's nucleophilic ¯uorides can react with nearby potential electrophiles. Thus, it was found that
while BrF does not react cleanly with aliphatic nitriles, it can convert them with good yield to a tri¯uoromethyl group when a neighboring
3
3
carboxylic moiety is present. BrF attacks also the electrophilic carbonyl carbon transforming it to a di¯uoroether, but simple hydrolysis
regenerates the carboxylic group to give eventually a-tri¯uoromethyl esters ꢀe.g. 5 ! 6). In the case of the keto nitrile 14, a participation of
the internal nucleophilic oxygen in attacking the nitrile carbon was also observed. # 2001Elsevier Science B.V. All rights reserved.
Keywords: Bromine tri¯uoride; Tri¯uoromethyl; Nitriles to tri¯uoromethyls
1
. Introduction
Chlorinated and ¯uorinated solvents proved to be useful, but
with acetonitrile, we observed a total consumption of the
reagent by the solvent. It is worth noting that at high
Although, bromine tri¯uoride has been used for the
synthesis of several modern anesthetics such as des¯urane
1], and sevo¯urane [2], it is still rarely used in organic
temperatures per¯uoronitriles do react with neat BrF result-
3
[
ing in a coupling reaction forming eventually ¯uoroazaalk-
anes ꢀR CN ! R CF NˆNCF R ) [9,10]. We decided,
chemistry. In the last few years, we and others started to
unveil its synthetic potential and use it for bromination of
deactivated aromatic rings [3] transferring carbonyls to the
f
f
2
2
f
therefore, to investigate the reaction of non-¯uorinated
nitriles with the highly reactive BrF ꢀsuch an attempt has
3
imported CF group [4], oxidizing alcohols to acyl ¯uorides
2
been recorded in the literature, using anhydrous HF as a
solvent. The reaction, however, was con®ned to the volatile
aceto and propionitrile derivatives only.) [11].
[5] constructing tri¯uoromethyl ethers [6] and performing
tertiary ¯uorinations [7]. Recently, we have described a
novel C±C±N to C±N±C rearrangement using BrF [8]
3
Our ®rst experiment involved 1-adamantane carbonitrile
ꢀ1). The product, obtained in 80% yield, proved to be 1,3-
di¯uoroadamantane ꢀ2) [12]. The course of this reaction can
be explained by the ease of tertiary electrophilic ¯uorination
on this system [13], and by the stability of the adamantyl
carbocation encouraging nucleophilic displacement of the
nitrile moiety. Reacting alkyl nitriles such as tridecanenitrile
and pentadecanenitrile ꢀ3 and 4) on the other hand, resulted
in many products in small yields. The 19F NMR indicated
and it seems that the road for bromine tri¯uoride to be
included in the arsenal of useful chemical reagents is quite
promising.
2
. Results and discussion
During developing of the above mentioned chemistry, we
faced the question of ®nding suitable media for reactions of
that some of these products had a CF group adjacent to a
3
BrF . Since, it was clear that oxygenated solvents should be
3
methylene. The low yields can be attributed to competing
radical reactions on the alkyl chain. The formation of the
tri¯uoroalkyl group, however, encouraged us to continue the
work on the subject.
excluded because of their very exothermic interactions with
the reagent, we turned our attention to other possibilities.
*
From the experiments ꢀFig. 1) [8,14], we have been
learning that in order to reduce side reactions as mentioned
Corresponding author. Tel.: ‡972-3-6408378; fax: ‡972-3-6409293.
E-mail address: rozens@post.tau.ac.il ꢀS. Rozen).
0022-1139/01/$ ± see front matter # 2001Elsevier Science B.V. All rights reserved.
PII: S 0 0 2 2 - 1 1 3 9 ꢀ 0 1 ) 0 0 4 4 8 - 1

162
S. Rozen et al. / Journal of Fluorine Chemistry 111 ꢀ2001) 161±165
attacks the nitrile carbon ꢀintermediate B in Fig. 2) followed
by further attacks from ¯uoride ions on the nitrile carbon
leading eventually to ethyl 1-tri¯uoromethyl cyclohexano-
ate ꢀ6) in 35% yield. Since, however, there is an even more
electrophilic carbon around the complexation site, namely
the ester's carbonyl, the nucleophilic ¯uorides are attracted
to it as well, resulting in the overall incorporation of ®ve
¯
¯
uorine atoms to form 1-ꢀethoxydi¯uoromethylene)-1-tri-
uoromethylcyclohexane ꢀ7) in 45% yield. This major
product could be hydrolyzed in high yield to produce again
the tri¯uoromethyl ester 6 raising effectively its overall yield
to more than 70%. It should be noted that under the reaction
Fig. 1.
conditions, BrF does not react with an isolated ester group
3
such as in ethyl cyclohexanecarboxylic acid, emphasizing
the notion that two anchors for a successful reaction with this
reagent are indeed very helpful.
The mechanism presented in Fig. 2 is a schematic one. At
this stage, we cannot be sure whether one molecule of BrF3
is providing more than one ¯uorine atom, or each additional
¯uoride originates from an extra molecule of the reagent.
The second possibility seems quite likely since the reaction
above, bromine tri¯uoride has to be anchored to at least two
close electron donating groups before it can act relatively
cleanly at a desired center. In order to test this hypothesis, we
prepared 1-cyano-1-carboethoxycyclohexane ꢀ5) from ethyl
cyanoacetate and 1,5-dibromopentane [15]. This compound
possesses two electron donor elements ready to complex the
bromine atom of the reagent between them ꢀintermediate A
in Fig. 2). Once complexed, the strong nucleophilic ¯uorine
is at its best when 4±5 mole/eqv. of BrF were used. We
3
Fig. 2.

S. Rozen et al. / Journal of Fluorine Chemistry 111 ꢀ2001) 161±165
163
Fig. 3.
believe that the mechanism is mainly ionic since adding
dinitrobenzene as a radical scavenger to the reaction mixture,
did not affect the outcome by much. Another uncertainty is
the fate of the nitrogen and the carboxylic oxygen atoms.
up®eld from CFCl , which also served as an internal stan-
3
dard. IR spectra were recorded as neat ®lms, in CHCl3
solution or in KBr pellets on a Nicolet 205 FTIR spectro-
photometer. Unless otherwise indicated, mass spectra were
measured with a VG micromass 7070H instrument.
The results of the reaction of BrF with ®ve-membered
3
ring derivatives were not substantially different than the
reaction with the six-membered 5. In order to avoid volatility
problems, however, we have prepared the trichloroethyl
ester of 1-cyano-1-carboxycyclopentane ꢀ8). Its room tem-
3.1. Preparation and handling BrF3
Although commercially available, we prepare our own
BrF by simply passing 0.58 mole of pure ¯uorine through
perature reaction with BrF proceeded smoothly and in a few
3
3
minutes, we obtained two products, the major one being the
penta¯uoro derivative 9 in 65% yield while the minor was
identi®ed as the 2,2,2-trichloroethyl ester of 1-tri¯uoro-
methyl-1-carboxycyclopentane ꢀ10) in 25% yield. The
volatility issue was less of a problem with 1-cyano-1-ethyl-
carboxylate-2-methylcyclopentane ꢀ11) which gave mainly
0.2 mole of bromine placed in a copper reactor at 0 to
‡108C. When no excess of bromine is present, the BrF
3
obtained is an inde®nitely stable pale yellow liquid and has a
density of 2.5 with an mp of 7±98C [17]. At that temperature,
the higher oxidation state derivative ꢀBrF ) will not form to
5
any appreciable amount [18], although, we always use a
small excess of bromine thereby keeping the reagent from
disproportionation to BrF . This is also responsible for the
1-di¯uoromethylene-1-tri¯uoromethyl-2-methylcyclopentyl
ethyl ether ꢀ12) in 65% yield without affecting the tertiary
5
hydrogen which in some cases can be replaced by ¯uorine
when reacting with BrF [16]. Here too, the di¯uoromethy-
reddish coloration of the reagent. We store the reagent in
te¯on containers for long periods. BrF is a strong oxidizer
3
3
lene group could be hydrolyzed quantitatively to the corre-
sponding ethyl ester 13 converting in effect the cyano group
to a tri¯uoromethyl one in higher than 60% yields.
An interesting case supporting the need for two complex-
and tends to react very exothermically with water and
oxygenated organic solvents. The work with BrF should
be conducted in a well ventilated area and caution and
common sense be exercised.
3
ing sites for a meaningful reaction with BrF is further
3
demonstrated by 2-ꢀ2-cyanoethyl)-cyclohexanone ꢀ14).
While no reaction was observed between BrF and either
3.2. General procedure for the reactions with BrF3
3
cyclohexanone or 2-cyanoethylcyclohexane, the combina-
tion of the two potentially reactive sites and the geometry of
The nitriles ꢀ2 mmol) were dissolved in 30±60 ml dry
CFCl
substrate contains also an ester group) of BrF
3
and cooled to 08C. About 6 mmol ꢀ8±10, if the
3
1
4, allowed under comparable conditions, a very distinctive
was dissolved
reaction where the nucleophilic ¯uorides of the reagent were
able to attack the electrophilic carbons of both the nitrile and
the carbonyl. The ®nal nucleophilic attack, however, was
performed by the carbonylic oxygen atom to form trans-
in 30 ml of the same solvent, cooled to 08C and added
dropwise to the nitrile solution over a period of 10±15 min.
The reaction mixture was then washed with aqueous Na
until colorless, neutralized with aqueous NaHCO , extracted
with CH Cl , dried over MgSO and evaporated. The oily
products were puri®ed by column chromatography with
various ratios of pentane/Et O serving as eluent.
SO
2
3
3
2
,2,9-tri¯uoro-1-oxadecaline ꢀ15). The trans-con®guration
2
2
4
is evident from examining the 19F NMR spectrum of the
bridgehead ¯uorine atom, which appears as a broad triplet at
À124.77 ppm with a coupling constant of 30 Hz, typical of
two axial±axial HF interactions ꢀFig. 3).
2
3.3. The reaction of 1-cyano-1-carboethoxycyclohexane
ꢀ5) with BrF3
3
. Experimental
The nitrile 5 ꢀ0.363 g, 2 mmol) was reacted with BrF
3
according the above procedure. Two main products were
isolated by chromatography. The less polar one proved to be
1-ꢀethoxydi¯uoromethylene)-1-tri¯uoromethylcyclohexane
1H NMR spectra were recorded with Bruker AC-200 and
AM-360 WB spectrometers, with CDCl as solvent and
3
Me Si as an internal standard. The 19F NMR spectra were
measured at 338.8 MHz and are reported in parts per million
1
ꢀ7) in 45% yield. H NMR: 3.95 ꢀ2H, q, J ˆ 7 Hz), 1.26 ppm
4
1
9
4
ꢀ3H, t, J ˆ 7 Hz); F NMR: À69.3 ꢀ3F, t, JꢀFF† ˆ 12 Hz),

164
S. Rozen et al. / Journal of Fluorine Chemistry 111 ꢀ2001) 161±165
À78.87 ppm ꢀ2F, q, JꢀFF† ˆ 12 Hz); 13C NMR: 126.76 ꢀq,
J ˆ 284 Hz), 124.72 ꢀt, J ˆ 268 Hz), 59.51, 49.39 ꢀsextet,
J ˆ 25 Hz), 24.78, 24.23, 20.85, 14.83 ppm. Anal. Calcd.
for C H F O: C, 48.78; H, 6.14. Found: C, 48.92; H,
4
14.39, 14.32 ppm; MS: regular measurements such as EI and
CI with commercial instruments fail to show any molecular
ion peak. Repeating the measurement with an unique super-
sonic GC±MS, using electron ionization of the sample
compound while it is vibrationally cooled in a supersonic
molecular beam, reveals the molecular ion peak at
1
0 15 5
6.00%. The more polar fraction proved to be ethyl 1-
tri¯uoromethyl cyclohexanoate ꢀ6) in 35% yield.
À1
1
‡
IR ˆ 1734 cm ; H NMR: 4.24 ꢀ2H, q, J ˆ 7 Hz), 2.3
m=e ˆ 246 ꢀM) . The vibrational cooling considerably
1
9
ꢀ
2H, m), 1.9±1.25 ꢀ8H, m), 1.3 ppm ꢀ3H, t, J ˆ 7 Hz);
F
enhances the relative abundance of the molecular ion
[20,21]. Anal. Calcd. for C H F O: C, 48.78; H, 6.14.
13
NMR: À75.4 ppm ꢀs); C NMR: 168.65, 125.77 ꢀq,
J ˆ 282 Hz), 61.56, 53.0 ꢀq, J ˆ 24 Hz), 27.29, 24.70,
1
0 15 5
Found: C, 48.59; H, 5.82%. This compound was accom-
panied by 10% of ethyl 1-tri¯uoromethyl-2-methyl cyclo-
pentanoate ꢀ13) whose overall yield was effectively
increased to higher than 70% by treating the reaction
mixture with acidic ꢀ10% aqueous HCl or HF) hot ꢀ708C)
2
H, 6.74. Found: C, 53.57; H, 6.65%. Compound 7 was
2.16, 13.82 ppm. Anal. Calcd. for C H F O : C, 53.57;
10 15 3 2
dissolved in acidic ꢀ10% aqueous HCl or HF) EtOH/H O
2
and heated to 708C for 7 h. The tri¯uoromethyl derivative 6
was obtained in higher than 80% yield.
EtOH/H O for 10 h quantitatively hydrolyzing the penta-
2
¯
ꢀmixture of diastereoisomers). IR: 1731 cm ; H NMR:
uoro derivative 12 to the tri¯uoromethyl ethyl ester 13
À1
1
3.4. The reaction of trichloroethyl ester of 1-cyano-1-
carboxycyclopentane ꢀ8) with BrF3
4.33 ꢀ2H, q, J ˆ 7 Hz), 2.75±2.28 ꢀ2H, m), 2.21ꢀ 1H , m),
2.1±1.5 ꢀ4H, m), 1.41 and 1.39 ꢀfor both diastereoisomers)
Compound 8 was prepared by reacting 2,2,2-trichlor-
oethyl cyanoacetate [19] with 1,4-dibromobutane according
to the procedure outlined in [15]. It was reacted with BrF3
according to the above procedure. Two main products were
isolated by chromatography. The less polar one proved to be
ꢀ3H, t, J ˆ 7 Hz), 1.28 ꢀ3H, dq, J ˆ 7 Hz;
1
19
4
J ꢀHF† ˆ 2 Hz), 1.16 ppm ꢀ3H, d, J ˆ 7 Hz); F NMR:
2
1
3
À65.96 ꢀbs), À71.75 ppm ꢀs); C NMR ꢀeach pair repre-
sents the mixture of the diastereoisomers): 170.79, 169.88;
127.11 ꢀq, J ˆ 279 Hz), 126.55 ꢀq, J ˆ 281Hz); 6 1. 60,
61.10; 60.3 ꢀm); 41.93, 40.78; 34.81, 34.03; 32.23, 31.62;
30.18, 29.72; 24.61, 22.99; 15.65, 15.11; 13.99,13.81 ppm;
1-ꢀ2,2,2-trichloroethoxydi¯uoromethylene)-1-tri¯uorome-
1
thylcyclopentane ꢀ9) obtained in 65% yield. H NMR: 4.45
2H, s), 2.2±1.9 ꢀ4H, m), 1.9±1.6 ppm ꢀ4H, m); 19F NMR:
MS ꢀCI) m/e: 225 ꢀM ‡ 1) Anal. Calcd. for C H F O : C,
‡
ꢀ
1
0 15 3 2
4
À72.3 ꢀ3F, t, JꢀFF† ˆ 10 Hz), À80.0 ppm ꢀ2F, q,
53.57; H, 6.74. Found: C, 53.24; H, 6.67%.
4
JꢀFF† ˆ 10 Hz); Anal. Calcd. for C H Cl F O: C,
9
10
3 5
32.22; H, 3.0. Found: C, 32.52; H, 3.04%. The more polar
fraction proved to be the 2,2,2-trichloroethyl ester of 1-
3.6. The reaction of 2-ꢀ2-cyanoethyl)-cyclohexanone
ꢀ14) with BrF3
tri¯uoromethyl-1-carboxycyclopentane ꢀ10) isolated in 25%
1
yield. H NMR: 4.81ꢀ2H, s), 2.45±2.22 ꢀ4H, m), 2.22±
.00 ppm ꢀ4H, m); 19F NMR: À72.2 ppm ꢀs).
About 10 mmol of 14 were reacted with two-fold excess
of BrF ꢀ20 mmol) forming trans-2,2,9-tri¯uoro-1-oxadeca-
2
3
line ꢀ15) in 70% yield. H NMR: 2.5±1.2 ppm ꢀm); 19F
1
3
.5. The reaction of 1-cyano-1-ethylcarboxylate-2-
methylcyclopentane ꢀ11) with BrF3
NMR: À62.02 ꢀ1F, dddd, J ˆ 156 Hz; J ˆ 27 Hz;
1
2
J ˆ 11 Hz; J ˆ 6 Hz); À65.25 ꢀ1F, br. d, J ˆ 156 Hz)
3
4
13
À124.77 ppm ꢀ1F, br. t, J ˆ 30 Hz); C NMR: 123.23
ꢀdd, J ˆ 264 Hz; J ˆ 245 Hz), 114.09 ꢀdd, J ˆ 230 Hz;
Compound 11 ꢀas a diastereoisomeric mixture) was made
according to the procedure outlined in [15], using 1,4-
1
2
1
J ˆ 5 Hz), 40.45 ꢀCH, d, J ˆ 27 Hz), 34.93 ꢀCH , d,
2
1 2
2
2
À1
1
dibromo-2-methylbutane, IR: 1734, 2243 cm ; H NMR
each peak appears as two almost superimposed peaks for
J ˆ 24 Hz), 31.02 ꢀCH , dd, J ˆ 32 Hz; J ˆ 27 Hz),
2 2 2
ꢀ
24.77 ꢀCH , s), 21.96 ꢀCH , s), 28.32 ꢀCH , s), 21.29 ppm
ꢀCH , d, J ˆ 7 Hz); MS m/e: 194 ꢀM) , 174 ꢀM±HF) .
‡
‡
the two diastereoisomers): 4.25 ꢀ2H, q, J ˆ 7 Hz), 1.35 ꢀ3H,
2
t, J ˆ 7 Hz), 1.19 and 1.05 ꢀ3H, d, J ˆ 7 Hz); MS ꢀCI) m/e:
‡
1
82 ꢀM ‡ 1) . The main product of its reaction with BrF
3
was identi®ed as a diastereoisomeric mixture of 1-di¯uor-
omethylene-1-tri¯uoromethyl-2-methylcyclopentyl ethyl
ether ꢀ12) in 65% yield. 1H NMR: 3.95 ꢀ2H, q,
J ˆ 7 Hz), 2.41±1.35 ꢀ7H, m) 1.27 ppm ꢀ3H, t,
Acknowledgements
This research was supported by the USA±Israel Bina-
tional Science Foundation ꢀBSF), Jerusalem, Israel.
19
J ˆ 7 Hz), 1.19 ꢀ3H, d, J ˆ 8 Hz); F NMR ꢀone diaster-
4
eoisomer): À66.36 ꢀ3F, t, JꢀFF† ˆ 11 Hz), À80.32 ppm
ꢀ
2F, m); ꢀthe second diastereoisomer): À71.86 ꢀ3F, t,
References
4
13
JꢀFF† ˆ 11 Hz), À74.7 ppm ꢀ2F, m); C NMR ꢀfor the
mixture of both diastereoisomers): 126.84 ꢀq, J ˆ 280 Hz),
[
1] L.A. Rozov, C. Huang, D.F. Halpern, G.G. Vernice, US Patent
,283,372 ꢀ1994).
[2] D.F. Halpern, M.L. Robin, US Patent 4,996,371 ꢀ1991).
1
26.36 ꢀq, J ˆ 266 Hz), 125.04 ꢀt, J ˆ 280 Hz), 124.37 ꢀt,
5
J ˆ 266 Hz), 58.93 ꢀm) 58.77, 39.69, 35.13, 30.73, 24.08,

S. Rozen et al. / Journal of Fluorine Chemistry 111 ꢀ2001) 161±165
165
[
[
3] S. Rozen, O. Lerman, J. Org. Chem. 58 ꢀ1993) 239.
[13] S. Rozen, C. Gal, J. Org. Chem. 53 ꢀ1988) 2803.
4] S. Rozen, E. Mishani, A. Bar-Haim, J. Org. Chem. 59 ꢀ1994)
[14] I. Ben-David, D. Rechavi, E. Mishani, S. Rozen, J. Fluor. Chem. 97
ꢀ1999) 75.
2918.
[
[
5] S. Rozen, I. Ben-David, J. Fluor. Chem. 76 ꢀ1996) 145.
[15] N.S. Zefirov, T.S. Kuznetsova, S.I. Kozhushkov, L.S. Surmina, Z.A.
Rashchupkina, J. Org. Chem. USSR ꢀEnglish translation) 19 ꢀ1983)
474.
6] I. Ben-David, D. Rechavi, E. Mishani, S. Rozen, J. Fluor. Chem. 97
ꢀ1999) 75.
[
7] L.S. Boguslavskaya, A.V. Kartashov, N.N. Chuvatkin, Russian Chem.
Rev. 59 ꢀ1990) 501 ꢀtranslated from Uspekhi Khimii 59 ꢀ1990) 865).
8] S. Rozen, I. Ben-David, J. Org. Chem. 66 ꢀ2001) 496.
[16] L.S. Boguslavskaya, A.V. Kartashov, N.N. Chuvatkin, J. Org. Chem.
USSR ꢀEnglish translation) 25 ꢀ1989) 1835.
[
[
[17] J.H. Simons, Inorg. Synth. 3 ꢀ1950) 184.
[18] L. Stein, J. Am. Chem. Soc. 81 ꢀ1959) 1269.
[19] J.F. Carson, Synthesis ꢀ1979) 24.
9] W.J. Chambers, C.W. Tullock, D.D. Coffman, J. Am. Chem. Soc. 81
ꢀ1959) 1269.
[
[
[
10] C.B. Sokolov, C.A. Mazalov, Russian Patent 172,769.
11] T.E. Stevens, J. Org. Chem. 26 ꢀ1961) 1627.
[20] S. Dagan, A. Amirav, J. Am. Soc., Mass. Spectrom. 6 ꢀ1995) 120.
[21] A. Gordin, N. Tzanani, A. Amirav, Rapid. Com. Mass Spectrom. 15
ꢀ2001) 811.
12] W. Adcock, G.B. Kok, J. Org. Chem. 52 ꢀ1987) 356.