Synthetic and mechanistic aspects of the reactions between bromodifluoromethyltriphenylphosphonium bromide and dibromofluoromethyltriphenylphosphonium bromide and trialkylphosphites

Article abstract of DOI:10.1016/j.jfluchem.2008.04.003

Bromofluoromethyltriphenylphosphonium bromides react with trialkylphosphites in two distinct ways. Bromodifluoromethyltriphenylphosphonium bromide undergoes a rapid exchange reaction with trialkylphosphites to give the corresponding bromodifluoromethylphosphonates in good to excellent yields. A similar exchange reaction also occurred with an analogous diethoxyphenylphosphonite to give the corresponding ethoxyphenylphosphinate. Mechanistically, the exchange process involves the formation of difluorocarbene via dissociation of the intermediate difluoromethylene ylide, capture of the difluorocarbene by the trialkylphosphite to give [ (RO)₃ over(P, +) C over(F₂, -) ], which captures bromine followed by dealkylation to the product, bromodifluoromethylphosphonate. The equilibria involved in the multi-step mechanism are all shifted to the phosphonate product by the final dealkylation step. In contrast, the dibromofluoromethyltriphenylphosphonium bromide does not under exchange reactions with trialkylphosphite. The phosphite serves as a halophilic reagent to abstract Br from the dibromofluoromethylphosphonium salt to generate the bromofluoromethylene ylide, which can easily be trapped in situ with aldehydes or ketones to give good yields of the E/Z-bromofluoroalkenes. No dissociation of the bromofluoromethylene ylide was observed.

Full text of DOI:10.1016/j.jfluchem.2008.04.003

Journal of Fluorine Chemistry 129 (2008) 583–589
Contents lists available at ScienceDirect
Journal of Fluorine Chemistry
journal homepage: www.elsevier.com/locate/fluor
Synthetic and mechanistic aspects of the reactions between
bromodifluoromethyltriphenylphosphonium bromide and
dibromofluoromethyltriphenylphosphonium bromide and trialkylphosphites
*
Richard M. Flynn, Donald J. Burton , Denise M. Wiemers
Department of Chemistry, University of Iowa, Iowa City, IA 52242, United States
Dedicated to Professor Dennis Curran.
A R T I C L E I N F O
A B S T R A C T
Article history:
Bromofluoromethyltriphenylphosphonium bromides react with trialkylphosphites in two distinct ways.
Received 12 March 2008
Received in revised form 14 April 2008
Accepted 16 April 2008
Available online 24 April 2008
Bromodifluoromethyltriphenylphosphonium bromide undergoes
a rapid exchange reaction with
trialkylphosphites to give the corresponding bromodifluoromethylphosphonates in good to excellent
yields. A similar exchange reaction also occurred with an analogous diethoxyphenylphosphonite to give
the corresponding ethoxyphenylphosphinate. Mechanistically, the exchange process involves the
formation of difluorocarbene via dissociation of the intermþediꢁate difluoromethylene ylide, capture of the
difluorocarbene by the trialkylphosphite to give ½ðROÞ3PCF2ꢀ, which captures bromine followed by
dealkylation to the product, bromodifluoromethylphosphonate. The equilibria involved in the multi-step
mechanism are all shifted to the phosphonate product by the final dealkylation step. In contrast, the
dibromofluoromethyltriphenylphosphonium bromide does not under exchange reactions with trialk-
ylphosphite. The phosphite serves as a halophilic reagent to abstract Br from the dibromofluor-
omethylphosphonium salt to generate the bromofluoromethylene ylide, which can easily be trapped in
situ with aldehydes or ketones to give good yields of the E/Z-bromofluoroalkenes. No dissociation of the
bromofluoromethylene ylide was observed.
Keywords:
Difluoromethylene ylide
Bromodifluoromethylphosphonates
Exchange reactions
Difluorocarbene
Bromodifluoromethy-
ltriphenylphosphonium salts
Dibromofluoromethyl-
triphenylphosphonium salts
Bromofluoromethylene ylide
Bromofluoroolefins
ß 2008 Elsevier B.V. All rights reserved.
1. Introduction
the difluoromethylene
The preparation of difluoromethylene ylides via the reaction of
tertiary phosphines and dibromodifluoromethane is well-
a
established reaction, and when this reaction is carried out in the
presence of an aldehyde or ketone the resultant Wittig reaction
produces the 1,1-difluoroalkene [1].
(2)
2R3P : þ CF2Br2 þ > C¼O ! > ¼CF2 þ R3PO þ R3PBr2
(1)
Mechanistic studies suggest that the intermediate phospho-
nium salt is not formed by an SN2 process but involves
difluorocarbene as an intermediate (Eq. (2)) [1,2]. Subsequent
halophilic attack on the intermediate phosphonium salt generates
ylide and the dibromophosphorane (Eq. (3))[1]. The intermediate
difluoromethylene ylide is
þ
þꢁ
* Corresponding author. Tel.: +1 319 335 1363; fax: +1 319 335 1270.
½R3PCF2BrꢀBrꢁ þ R3P : Ð ½R3PCF2ꢀ þ R3PBr2
(3)
E-mail address: donald-burton@uiowa.edu (D.J. Burton).
0022-1139/$ – see front matter ß 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.jfluchem.2008.04.003

584
R.M. Flynn et al. / Journal of Fluorine Chemistry 129 (2008) 583–589
unstable and rapidly dissociates into difluorocarbene and tertiary
phosphine (Eq. (4))[1,2]. Thus when an isolated phosphonium salt
is reacted with a different tertiary phosphine, the
þꢁ
½R3PCF2ꢀ Ð R3P : þ½: CF2ꢀ
(4)
(7)
difluorocarbene can be captured by either tertiary phosphine to
give a mixture of bromodifluoromethylphosphonium salts (Eq. (5))
[1,3]. The percentage of the two bromodifluoromethylphospho-
nium salts varies with the nucleophilicity of the tertiary
phosphines, solvent and temperature [2,3]. When a bromodifluor-
omethyltrialkylphosphonium salt is reacted with an identical
tertiary trialkylphosphine that is part of the phosphonium salt, the
2. Results and discussion
2.1. Reaction of bromodifluoromethyltriphenylphosphonium salts
with trialkylphosphites
Our initial goals in this research were to determine: (1) if
trialkylphosphites could undergo exchange reactions with bromodi-
fluoromethyltriarylphosphonium salts and would the phosphite
compete with the triarylphosphine (from dissociation of the
difluoromethylene ylide; (2) would the intermediate (from halo-
philic attack on the phosphonium) salt undergo substitution of
bromide ion to produce a mixed phosphonium–phosphonate
product.
(5)
Bromodifluoromethyltriphenylphosphonium bromide (1) is
known to react with potassium fluoride to produce difluorocar-
bene [8]. Thus, in an attempt to demonstrate that triethylphosphite
(2) was capable of intercepting [:CF2], a small scale exploratory
reaction between (1), KF, and (2) was carried out in triglyme at
room temperature in the presence of an equivalent amount of
tetramethylethylene (TME) (Eq. (8)). The major product was
observed to be bromodifluoromethylphosphonate (3) in 74% yield
exchange reaction is an invisible reaction and generates no new
products. However, the intermediate difluoromethylene ylide
could also attack the intermediate bromotrialkylphosphonium
cation to produce a bisphosphonium salt (Eq. (6)) [1,4]. This
bisphosphonium salt formation is unique for trialkylphosphine
derivatives. When the corresponding triarylphosphonium salts,
(
19F vs. PhCF3). Traces of the
(8)
þ
difluoromethylphosphonate, (EtO)2P(O)CF2H, as well as 1,1-
difluorotetramethylcyclopropane, were also detected. Surpris-
ingly, no 19F NMR signals were observed for the expected by-
product, Ph3PF2. Thin layer chromatography of the reaction
mixture (50% CHCl3, hexane) revealed a component in the mixture
with an identical Rf value to Ph3P. The 31P NMR spectrum of the
ðAr3PCF2BrÞBrꢁ, and trisdimethylaminophosphonium salts
þ
½ðMe2NÞ3PCF2BrꢀBrꢁ.
reaction mixture also showed a singlet,
with the reported chemical shift of Ph3P [Lit.
d
= 5.98, which agreed well
= 5.9] [9].
d
Hydrolysis of (1) is also known to proceed via a difluorocarbene
intermediate [10]. Therefore, we attempted the reaction of (1), (2)
andH2Ointriglymeatroomtemperature.Theonlyproductobserved
in the 19F NMR spectrum of the reaction mixture was CF2BrH arising
from the reaction between (1) and H2O above. Since; reactions in
triglyme were heterogenous, the above reaction was repeated in
CHCl3 in which (1) is entirely soluble. The addition of (2) to the
solution of (1) in CHCl3 caused an exothermic reaction. The addition
of H2O to this reaction mixture had no effect, since the exchange
reaction had already occurred (Eq. (9)). Since the exchange product
(3) was formed in both reactions described above in Eqs. (8) and (9)
ð6Þ
were utilized, no stable bisphosphonium salt formation
was observed [4]. When fluorotrihalomethanes, such as CFCl3
and CFBr3, are employed in reactions with trialkylphosphines, the
intermediate bisphosphonium salt can react further to give
fluorine-containing phosphoranium salts (Eq. (7)) [4,5–7]. Thus, a
wide variety of reactions are possible for the reaction of tertiary
phosphines with fluorohalomethanes.
CHCl3
H2O
ð1Þ þ ð2Þ ꢁ! ð3Þ ꢁ!no change
(9)
RT
48%ð19FÞ

R.M. Flynn et al. / Journal of Fluorine Chemistry 129 (2008) 583–589
585
Table 1
it was obvious that neither KF nor H2O actually participated in the
Exchange reaction of (1) with trialkylphosphites
exchange reaction. It was also obvious that no equilibrium
reaction, similar to that described in Eq. (5), had occurred since
no (1) was detected at the completion of these reactions.
þ
CH Cl
2
2
½Ph3PCF2BrꢀBrꢁ þ ðROÞ 3Pð2Þꢁꢁꢁ! ðROÞ2PðOÞCF2Br
ð1Þ
ð3Þ
RT
Entry
R
% Yielda
In order to determine the best reaction conditions to afford the
exchanged phosphonate (2), the reaction between (1) and (2) was
carried out in several solvents. The yields of (3) ranged from
moderate (32% CH3CN), (48% CHCl3), (54% DMF) to excellent (92%
CH2Cl2). In triglyme, the reaction was very slow in the absence of
adequate mixing. The lower yieldsþin CHCl3, CH3CN and DMF can be
ascribed to the formation of ½Ph3PCF2HꢀBrꢁ, which was observed
spectroscopically as a major by-product in these solvents and which
most likely is formed by protonation of the intermediate ylide.
In order to evaluate the potential scope of this exchange process,
(1)wasreactedwithavarietyoftrialkylphosphitesinCH2Cl2 atroom
temperature. These results are summarized in Table 1.
1
2
3
4
5
6
7
8
CH3
60b
C2H5
92 (67)
85 (68)
86
i-Pr
n-Bu
i-Bu
94
CH2ClCH2
PhP(OEt)2
Ph
61
79c
No Rx.
19F NMR yield vs. PhCF3 internal standard, values in parentheses are isolated
a
yield based on phosphite.
þ
ꢁ
b
Product reacts further to give ½Ph3PCH3ꢀ½O2PðOCH3ÞðCF2BrÞꢀ as the observed
dealkylation product.
c
Product is the phosphinate PhP(O)OEt(CF2Br).
The exchange product from the reaction with trimethylpho-
sphite (entry 1) reacts further (dealkylation) with the co-product to
give the phosphonium–phosphonate salt. The products from
reaction with n-Bu, and i-Bu phosphonates (entries 4 and 5) were
identified spectroscopically by comparison with the bromodifluor-
omethylphosphonates previously prepared in this laboratoryvia the
Michaelis–Arbuzov reaction [11]. Triphenylphosphite (entry 8) did
not react with (1) at room temperature. When the triphenylpho-
sphite reaction was heated at reflux, several minor products were
observed inthe 19F NMR spectrumof the reaction mixture; however,
none of them were the desired diphenylbromodifluoromethylpho-
sphonate. A surprising result was obtained with (ClCH2CH2O)3P
(entry 6). This phosphite is inert under the Michaelis–Arbuzov
conditions [11]; however, in the exchange process, the 19F NMR
spectroscopic data indicated that the exchange reaction had
afforded the bromodifluoromethylphosphonate. When diethylphe-
nylphosphonite (entry 7) was employed in the exchange reaction, a
79% (19F NMR yield vs. PhCF3) of ethylphenyl bromodifluoromethyl-
phosphinate was produced. The spectroscopic properties of the
phosphinate were identical to a sample previously prepared from
diethylphenylphosphonite and CF2Br2 [12].
phosphite trapped intermediate (c) abstracts bromine from (b)
to afford the bromodifluoromethyltrialkoxyphosphonium bromide
(d), which on dealkylation provides the bromodifluoromethylpho-
sphonate (e). Mechanistically, the process is similar to the
exchange of the bromodifluoromethyl group between bromodi-
fluoromethyltriarylphosphonium bromide and tertiary phos-
phines (Eq. (5), Ref. [2]). However, þone major difference is that
the dealkylation reaction of ½ðROÞ3PCF2BrꢀBrꢁ is an irreversible
process and shifts all equilibria to the formation of the bromodi-
fluoromethylphosphonate. Thus, the only bromodifluoromethyl
product of the exchange process is the phosphonate.
A similar mechanism occurs with dialkoxyphenylphosphonites
to give the corresponding alkoxyphenylphosphinate. Likewise the
þ
expected reaction of ½Ph3PCF2IꢀIꢁ with (EtO)3P in CH2Cl2 gave a 60%
19F NMR yield of (EtO)2P(O)CF2I.
Ethanolreactswith(1)toextrude[:CF2](similartothereactionof
(1) with H2O [10]) and affords CF2HBr and the product of carbene
insertion, CH3CH2OCF2H in a 74/26 ratio [4]. When the reaction
between (1) and (2) is carried out in triglyme containing ethanol, the
only observed products (detected by 19F NMR) were CF2HBr (trace)
and(EtO)2P(O)CF2H(Eq. (11)).Thisresultisaclþeꢁarindicationthat(2)
has successfully trapped [:CF2] to give ½ðROÞ3PCF2ꢀ, which on
Mechanistically, the following explanation is the most plausible
to explain the exchange reaction (Eq. (10)). Halophilic attack on the
phosphonium salt by the phosphite generates the
þ
½Ph3PCF2BrꢀBrꢁ þ ðEtOÞ3P
þ EtOHtriꢁg!lyme CF2HBr þ ðEtOÞ PðOÞCF2H
(11)
2
trace
80%ð19F NMR yieldÞ
protonation and dealkylation give the reduced phosphonate [13].
An alternative mechanism for the exchange process is outlined
below (Eq. (12)). The
ð10Þ
phosphonium ylide (a) and bromotrialkoxyphosphonium bromide
(b). Dissociation of the ylide produces tertiary phosphine and
difluorocarbene, which is subsequently captured by either the
tertiary phosphine (reverse reaction) or the phosphite. The
ð12Þ

586
R.M. Flynn et al. / Journal of Fluorine Chemistry 129 (2008) 583–589
Table 2
initially formed ylide (step 1, identical to step 1 in Eq. (10)) could
conceivably react with the trialkoxyphosphonium cation to
generate a bisphosphonium salt (f) (step 2), which on dealkylation
(step 3) would afford the phosphonium phosphonate salt (g). The
last step (step 4) would be a second attack by bromide on the
difluoromethylene carbon of (g) with the elimination of triphe-
nylphosphine. This mechanistic sequence is considered unlikely
for the following reasons: (1) we have studied in detail the attack at
various sites of (RO)2P(O)CF2Br compounds [12]. The only position
that was inert to attack by a variety of reagents and nucleophiles
was the difluoromethylene carbon; (2) alþl attempts in this work to
produce a mixed salt of the type, ½Ph3PCF2PðOÞðOEtÞ2ꢀBrꢁ were
unsuccessful; and (3) Keslingþ[4] was successful in generating (in
Synthesis of bromofluoroolefins by dehalogenation of (4) with (2)
þ
CH Cl
2
2
½Ph3PCFBr2ꢀBrꢁ þ ðEtOÞ3Pð2Þ þRCðOÞR0 ꢁꢁ! RðR0ÞC ¼ CFBr
ð4Þ
R
R0
% Yielda
Z/E
Ph
CF3
78
58
75
54
55
17
61
55/45
38/62
51/49
45/55
43/57
–
Ph
CF3CF2
CF2Cl
H
Ph
Ph
Ph
CH3
–CH2CH2CH2CH2–
–(CH2)5–
–
a
Isolated yield based on the carbonyl component.
solution) the mixed salt, ½Bu3PCF2PðOÞðOEtÞ2ꢀBrꢁ, via the reaction
(Eq. (14)).
þ
between ½Bu3PCF2BrꢀBrꢁ and (EtO)3P [14]. This bis salt is very
stable and there was no evidence that it decomposed to Bu3P and
the phosphonate. Although the triphenyl analog has not yet been
prepared, it seems unlikely that its behavior would be significantly
different from the tributyl bisphosphonium phosphonate in this
respect. Thus, we feel that this alternative mechanistic interpreta-
tion is less plausible than the mechanism outlined in Eq. (10).
þ
CF3CH2OH
ð4Þ þ ð2Þ ꢁ! ½Ph3PCFHBrꢀBrꢁ
(14)
71%ð19F NMRÞ
In this case the intermediate ylide was intercepted by the
alcohol and protonated.
In order to test the generality of this route to bromofluoro
olefins, a series of aldehydes and ketones were reacted with (4) and
(2), and these results are summarized in Table 2. As noted in our
earlier work [1,18,19] the fluorine-containing ketones were the
most reactive substrates, and cyclopentanone gave mainly
protonation of the ylide via the enol of the ketone. These reactions
are clean, the solvent easily removed, and the products easily
isolated. This alternative route to bromofluoroolefins compliments
the earlier reported work [1].
2.2. Reaction of dibromofluoromethyltriphenylphosphonium salts
with trialkylphosphites
þ
þ
The hydrolysis of ½Ph3PCF2BrꢀBrꢁ and ½Ph3PCFBr2ꢀBrꢁ with H2O
are significantly different. Hydrolysis of the bromodifluoromethyl
salt in the presence of a radioactive isotope of bromide or sodium
iodide gave unequivocal evidence that the mechanism for this
reaction proceeds through a difluorocarbene intermediate [10]. On
the other hand, hydrolysis of the dibromofluoromethyl salt in the
presence of a radioactive isotope of bromide gave evidence that the
mechanism of this reaction proceeds via the dibromofluoromethide
ion and not via a bromofluorocarbene intermediate [15]. Likewise,
we have observed that difluoromethylene ylides readily dissociate
into tertiary phosphine and difluorocarbene (Eq. (4), [2]), whereas
the corresponding bromofluoromethylene ylide is more stable, and
we have not observed any dissociation in our detailed investigations
of this ylide. Thus, it was of interest to us to determine if a similar
exchange reaction occurs when dibromofluoromethylphosphonium
3. Experimental
3.1. General experimental procedures
The 19F NMR spectra were recorded on a Varian HA-100
Spectrometer operated at 94.075 MHz in the HR (non-lock) mode.
The 19F NMR spectra were obtained as 10–15% (w/v) solutions of a
pure compound in CDCl3 or as aliquots of reaction mixtures vs. an
internal standard of CFCl3. Chemical shifts are reported in ppm
upfield of CFCl3. Spectra were internally calibrated by the audio
side band technique. Quantitative measurements were carried out
by integration relative to an internal standard (C6H5CF3). Routine
1H NMR spectra were obtained as 10–15% (w/v) solutions of a pure
compound in CDCl3 or, occasionally, as aliquots of reaction
mixtures vs. an internal standard of TMS on a Varian A-60
analytical NMR spectrometer. Chemical shift values are reported in
ppm downfield of TMS. 13C NMR spectra were obtained as 15% (w/
v) solutions in a suitable deuterated solvent (usually CDCl3), which
also provided the lock signal, on a Bruker HX-90E Spectrometer
operated at 22.63 MHz vs. TMS as internal standard. In all cases,
broad band proton decoupling was employed to simplify the
spectra. Chemical shift values are reported in ppm downfield from
TMS. 31P NMR spectra were obtained as 15% (w/v) solutions of a
pure compound in a suitable deuterated solvent (usually CDCl3)
which also provided the lock signal or approximately 50/50
mixtures of a reaction mixture. The 31P NMR Spectra were obtained
on a Bruker HX-90E Spectrometer operated at 36.435 MHz vs. 85%
H3PO4 as the external standard. In most cases, broad band proton
decoupling was employed to simplify the spectra. When P–H
coupling information was desired, the instrument was operated in
the gated decoupled mode which retained nearly the full NOE and
still yielded accurate coupling data. Chemical shift values are
reported in ppm vs. H3PO4. Mass spectra were obtained on a
Hitachi-PerkinElmer RMU-6E Mass Spectrometer operated at
70 eV. Analytical GLC were carried out on either an F&M Dual
salts were reacted with trialkylphosphites.
þ
When ½Ph3PCFBr2ꢀBrꢁ (4) was treated with triethylphosphite
(2) in CH2Cl2, the phosphonium salt was rapidly consumed but
only traces of fluorine containing products were detected by 19F
NMR analysis of the reaction mixture. When the same reaction was
carried out in the presence of trifluoroacetophenone, (E)- and (Z)-
2-phenyl-1-bromotetrafluoropropene were formed in excellent
isolated yield (Eq. (13)) [16]. When the reaction between (4) and
(2) was
ð13Þ
carried out in the presence of TME, no bromofluorocyclopropane
[17] was detected by 19F NMR analysis of the reaction mixture.
When (4) and (2) were reacted in the presence of CF3CH2OH, a 71%
(
19F NMR yield) of the reduced phosphonium salt was produced

R.M. Flynn et al. / Journal of Fluorine Chemistry 129 (2008) 583–589
587
Column Gas Chromatograph Model 700, or when it was desired to
collect a sample for mass spectral or C, H, N analysis on an F&M
Dual Column Research Chromatograph Model 5750. Both instru-
ments were equipped with TCD. The columns utilized in this work
were as follows: column A was a 10 ft ꢂ 0.25 in. column packed
with 20% (w/w) Carbowax 20 M on 80–100 mesh Chromosorb P.
Column B was a 10 ft ꢂ 0.25 in. column packed with 20% (w/w) SE-
30 on 80–100 mesh Chromosorb P. Column C was a 6 ft ꢂ 0.25 in.
column packed with 5% (w/w) SE-30 on 80–100 mesh Chromosorb
P. Infrared spectra were obtained as thin films between NaCl plates
in the case of liquid samples or as KBr pellets for solid samples on a
Beckman IR-20A Infrared Spectrometer. C, H, N analyses were
CH2Cl2. After the salt completely dissolved, (EtO)3P (4.15 g,
0.025 mol) was slowly syringed into the homogeneous solution.
A very slight exotherm occurred and the solution assumed a light
yellow–brown color. The reaction was complete after 10 min as
shown by the total absence of the salt in the 19F NMR spectrum of
the reaction solution. The solution was distilled to remove most
of the CH2Cl2 and ethyl bromide. The 1H NMR of the distilled
material showed a triplet,
d = 1.62 and quartet, d = 3.42 for the
EtBr, and these peaks were enhanced upon the addition of an
authentic sample of EtBr to the NMR tube. The remaining
solution was then distilled through a short path distillation
apparatus to yield 4.45 g (67%) of (EtO)2P(O)CF2Br (bp 67.5–
69.58/0.6 mm Hg). The residue remaining in the flask was
recrystallized from methanol (35 ml) to yield 4.2 g (65%) of a
solid identified as triphenylphosphine (mp 79–808), mixed mp
with authentic Ph3P, 79–81 8C. The infrared spectrum of this solid
material was identical to the reported spectrum for Ph3P.
Repetition of this reaction on a smaller scale afforded a yield
carried out by service personnel of this department on
a
PerkinElmer 240 (automated) Elemental Analyzer. All melting
points were obtained in capillaries on a Thomas–Hoover Unimelt
apparatus and are uncorrected. All bp were determined during
fractional distillation and are uncorrected.
Potassium fluoride (J.T. Baker, anhydrous) was dried at 250 8C/
0.5 mm Hg for 18 h and stored in a desiccator. Alternatively, KF was
placed in an evaporating dish and heated to ꢃ450 8C with a Bunsen
burner. After breaking up of the initial crust which formed, the
finely divided salt was heated for 1 h at 450 8C and transferred to a
weighed container while still hot. Triglyme (Ansul Ether 161)
and other glymes were purified by distillation from a sodium
(
19F NMR vs. PhCF3) of 92%. The 19F, 31P, and 13C NMR data for the
(EtO)2P(O)CF2Br prepared in this reaction was identical to the
data for (EtO)2P(O)CF2Br prepared via the Michaelis–Arbuzov
method [11,12,22].
3.4. Preparation of [(CH3)2CHO)]2P(O)CF2Br
˚
benzophenone ketyl and stored in brown bottles over 4 A
þ
molecular sieves. Diethyl ether was dried and stored over sodium
wire in a brown bottle until used. Methylene chloride was refluxed
overnight over P2O5, distilled at atmospheric pressure and stored
Similar to the procedure described in Section 3.3, ½Ph3PCF2Brꢀ
Brꢁ (14.2 g, 0.030 mol), 50 ml dry CH2Cl2 and 6.45 g [(CH3)2CHO)3
P] (0.03 mol) gave (after distillation) 6.0 g (68% of [(CH3)2CHO)]2-
P(O)CF2Br (bp 758/2 mm Hg). An aliquot of the reaction mixture
showed an 85% 19F NMR yield (vs. PhCF3) of the bromodifluor-
omethyldiisopropylphosphonate [23].
˚
in a brown bottle over 4 A molecular sieves.
Trichlorofluoromethane (Freon 11) and dibromodifluoro-
methane (Freon 12B2) were obtained commercially and used as
obtained. Trifluoroethanol was obtained commercially and stored
þ
over 4 A molecular sieves. Trimethyl, triethyl, tributyl and
3.5. Small scale exchange reaction between ½Ph3PCF2BrꢀBrꢁ and
˚
triisobutyl phosphites were obtained from commercial vendors,
distilled from sodium metal under vacuum and stored in brown
(RO)3P (Table 1)
˚
bottles over 4 A molecular sieves. Triisopropyl, triphenyl, and tris-
The small-scale exchange reactions cited in Table 1 were all
NMR scale reactions and were carried out in an identical manner.
The reaction with trimethyl phosphite will be describedþin detail as
an illustrative example. The phosphonium salt, ½Ph3PCF2BrꢀBrꢁ
(0.38 g, 0.0008 mol) was added to an NMR tube, then 0.5 ml dry
CH2Cl2 was added to dissolve the phosphonium salt. To this
solution was added (MeO)3P (0.1 g, 0.0008 mol). The addition of
the phosphite caused precipitation of the salt which then
rapidly dissolved to give a clear, yellow–brown solution. Benzotri-
fluoride (0.04 g, 0.00027 mol) was added and the 19F NMR
2-chloroethyl phosphites were obtained commercially and used as
obtained. Diethylphenylphosphonite was prepared from Et3N,
EtOH, and phenyldichlorophosphine in 59% yield (bp 70–738/
0.3 mm Hg) by the literature procedure. Tribromofluoromethane
was prepared by the method of Birchall and Haszeldine [20] and
þ
stored in a refrigeratþor over copper wire.þThe salts, ½Ph3PCF2BrꢀBrꢁ,
þ
½Ph3PCFBr2ꢀBrꢁ, ½Ph3PCF2HꢀBrꢁ, and ½Ph3PCF2IꢀIꢁ were prepared by
previously described methods from this laboratory [1]. The
phosphonates employed for comparison purposes (spectroscopi-
cally) were prepared by the previously described method from this
laboratory [11].
2
spectrum recorded. 19F NMR (ppm): d = ꢁ55.3 (d, JPF = 67 Hz),
ꢀ
ꢁ
þ
60% 19F NMR yield of Ph3PCH3 ½O2PðOMeÞCF2Brꢀ. The identity
ꢁ
3.2. Solvent survey for the exchange reaction between (EtO)3P
and ½Ph3PCF2BrꢀBrꢁ
of this compound was confirmed by spiking the sample with a
solution of the phosphonium–phosphonate salt in CH2Cl2 prepared
by the reaction between (MeO)2P(O)CF2Br [11] and Ph3P. No new
peaks were detected.
þ
In these reactions, the phosphonium salt was placed into a
weighed NMR tube, the appropriate solvent added and
a
þ
3.6. Reaction of ½Ph3PCF2BrꢀBrꢁ and (ClCH2CH2O)3P
stoichiometric amount of triethyl phosphite added via a syringe.
The reaction with the phosphite was instantaneous in all solvents,
except triglyme. The yields of the exchanged product, (EtO)2-
P(O)CF2Br were determined by 19F NMR vs. PhCF3 internal
standard. Results: (solvent, % yield (EtO)2P(O)CF2Br: CHCl3 (48%),
CH3CN (32%), DMF (54%), CH2Cl2 (92%) [21], TG (slow).
A 50 mlþflask fitted with a condenser and glass tee was charged
with ½Ph3PCF2BrꢀBrꢁ (2.7 g, 0.0057 mol), 20 ml CH2Cl2 and
(ClCH2CH2O)3P (1.54 g, 0.0057 mol). The salt was soluble in CH2Cl2,
and the addition of the phosphite had no visible effect. The solution
was then heated to reflux overnight; then PhCF3 (0.27 g, 0.0019 mol)
was added as an internal standard. In addition to a small amount of
reduced phosphonium salt, the only other fluorine-containing
product detected in the 19F NMR spectrum of the reaction mixture
þ
3.3. Large scale exchange reaction between ½Ph3PCF2BrꢀBrꢁ and
(EtO)3P
was a doublet at
d
= ꢁ62.2 (2JPF = 96 Hz) in 61% (19F) yield consistent
Into an apparatus which was equipped for simple distillation
þ
was charged ½Ph3PCF2BrꢀBrꢁ (11.8 g, 0.025 mol) and 30 ml dry
with the formation of (ClCH2CH2O)2P(O)CF2Br.

588
R.M. Flynn et al. / Journal of Fluorine Chemistry 129 (2008) 583–589
þ
þ
3.7. Reaction of ½Ph3PCF2BrꢀBrꢁ, (EtO)3P and CH3CH2OH in triglyme
3.8.3. Reaction of ½Ph3PCFBr2ꢀBrꢁ, (EtO)3P and C6H5CHO
þ
In a similar manner (cf. 3.8) ½Ph3PCFBr2ꢀBrꢁ (19.9 g, 0.037 mol),
(EtO)3P (6.23 g, 0.037 mol) and C6H5CHO (3.18 g, 0.03 mol) in
CH2Cl2 were reacted. The 19F NMR spectrum after the addition of
the phosphite still showed some unreacted phosphonium salt, so
additional (EtO)3P (0.96 g, 0.006 mol) was added slowly via syringe
to the reaction mixture. The phosphonium salt was completely
consumed. The wash procedure in this case consisted of NaHSO3
(40%, 2ꢂ 30 ml) and H2O (2ꢂ 30 ml). The yield of (E)- and (Z)-1-
bromo-1-fluoro-2-phenylethene was 3.25 g (54%) (bp 78–828/
þ
To a 5 mm NMR tube was added ½Ph3PCF2BrꢀBrꢁ (0.12 g,
0.00026 mol), triglyme (0.5 ml), (EtO)3P (0.0042 g, 0.00026 mol)
and CH3CH2OH (0.034 g, 0.00077 mol). The NMR tube was allowed
to stand, with occasional shaking, for 2 days, at the end of which
time the solution was homogenous. PhCF3 (0.025 g, 0.00017 mol)
was added and the 19F NMR spectrum showed a signal at
d
= ꢁ136.4 (dd) for (EtO)2P(O)CF2H in 80% 19F NMR yield; identical
to the 19F NMR spectrum of (EtO)2P(O)CF2H previously prepared in
this laboratory [12].
8 mm Hg) Z/E = 45/55. 19F NMR (CDCl3) (ppm):
d
= ꢁ65.9 (d,
3
3JHF = 15 Hz, (Z)-F);
(CDCl3) (ppm):
3JHF = 15 Hz, (Z)-H);
d
= ꢁ68.4 (d, JHF = 33 Hz, (E)-F); 1H NMR
þ
3
3.8. General procedure for the reaction of ½Ph3PCFBr2ꢀBrꢁ and
d
= 5.97 (d, JHF = 34 Hz, (E)-H);
d
= 6.66 (d,
(EtO)3P in presence of an aldehyde or ketone [24]
d = 7.2–7.6 (m). This spectral data was in
good agreement with the data for this compound reported by
Vander Haar [18,19].
A 100 ml, 3-necked round bottom flask was equipped with a
septum, constant pressure addition funnel and a glass tee leading to
þ
source of nitrogen and a mineral oil bubbler. The flask was charged
3.8.4. Reaction of ½Ph3PCFBr2ꢀBrꢁ, (EtO)3P and CH3C(O)C6H5
þ
þ
with ½Ph3PCFBr2ꢀBrꢁ (15.9 g, 0.03 mol), dry CH2Cl2 (45 ml) and
trifluoroacetophenone(5.22 g,0.03 mol).Thesuspensionwascooled
in an ice bath; (EtO)3P (4.98 g, 003 mol) in 20 ml CH2Cl2 was then
added dropwise slowly via the constant addition funnel. After the
addition of (EtO)3P was completed, the homogeneous solution was
examined by 19F NMR and found to contain some unreacted ketone.
Therefore, an additional 4 g (0.0075 mol) of the phosphonium salt
wasaddedtothesolutionandadditional(EtO)3P(1.25 g,0.0075 mol)
inCH2Cl2 (5 ml)addeddropwisetoentirelyconsumetheketone. The
solvent (CH2Cl2) was then distilled at atmospheric pressure. When
most of the CH2Cl2 had been removed, water (20 ml) was added to
the remaining reaction mixture, and the residue steam distilled;
approximately 100 ml of distillate was collected. The aqueous layer
was separated and extracted with 50 ml of ether. The organic layers
were combined, washed with 5% NaOH (2ꢂ 15 ml) and H2O (2ꢂ
50 ml) and dried over anhydrous MgSO4. The ether was removed via
rotary evaporation at reduced pressure and the residue distilled
through a short path distillation apparatus to yield 6.3 g (78%, based
on ketone) of (E)- and (Z)-1-bromo-2-phenyl-F-propene (bp 83.58/
In a similar manner (cf. 3.8) ½Ph3PCFBr2ꢀBrꢁ (19.9 g, 0.037 mol),
(EtO)3P (6.23 g, 0.037 mol), and CH3C(O)C6H5 (3.6 g, 0.03 mol)
afforded a mixture of the (E)- and (Z)-isomer of (E)- and (Z)-1-
bromo-1-fluoro-2-phenylpropene (3.53 g, 55%) and unreacted
acetophenone (0.6 g, 17%); bp 55–648/1.3 mm Hg). The wash
procedure in this case consisted only of H2O (3ꢂ 50 ml) Z/E = 43/
57. 19F NMR (CDCl3) (ppm):
d
= ꢁ75.2 (overlapping quartets,
4
4JHF = 3 Hz, (Z)-F, JHF = 4 Hz, (E)-F); 1H NMR (CDCl3) (ppm):
4
d
= 2.06 (
d
,
4JHF = 4 Hz, (Z)-CH3),
d
= 2.06 (d, JHF = 3 Hz, (E)-CH3),
d
= 7.36 (s),
d
= 2.57 (CH3, acetophenone). These values are in good
agreement with those reported by Vander Haar [18,19].
þ
3.8.5. Reaction of ½Ph3PCFBr2ꢀBrꢁ, (EtO)3P and cyclopentanone
þ
In a similar manner (cf. 3.8) ½Ph3PCFBr2ꢀBrꢁ (19.9 g, 0.037 mol),
(EtO)3P (6.23 g, 0.037 mol), and cyclopentanone (2.52 g, 0.03 mol)
were reacted in CH2Cl2. After complete addition of the phosphite
some phosphonium salt still remained, so additional phosphite
(0.97 g, 0.006 mol) was added to the reaction mixture to
completely consume the phosphonium salt. The wash procedure
in this case consisted of NaHSO3 (40%, 3ꢂ 35 ml) and water. The
yield of 1-bromo-1-fluoromethylenecyclopentane was 0.9 g (17%)
42 mm Hg) Z/E = 55/45, 19F NMR (CDCl3) (ppm):
4JFF = 13 Hz, Z vinyl F);
= ꢁ58.3 (m, E vinyl F); = ꢁ59.8 (overlap of
CF3 groups) 1H NMR (CDCl3) ppm: = 7.38 (m). The 19F NMR spectra
d
= ꢁ55.7 (q,
d
d
d
bp 37.5–38.58/8 mm Hg. 19F NMR (CDCl3) (ppm):
d
= ꢁ78.3 (bs), lit.
= 1.5 and 2.0–2.5
= 26.1
= 124.8 (d,
= 123.9 (d, JCF = 13 Hz, C = CFBr); IR:
m (C C), lit. 5.98 m [19]. This olefin is unstable and
were in agreement with those reported by Vander Haar [18,19].
d
= ꢁ74.9 (m) [19]. 1H NMR (CDCl3) (ppm):
d
(complex multiplets in 1:1 ratio; 13C NMR (CDCl3) (ppm):
d
þ
3.8.1. Reaction of ½Ph3PCFBr2ꢀBrꢁ, (EtO)3P and CF3CF2C(O)C6H5
(s), 27.4 (s), 29.2 (s), 31.6 (s)—ring carbon and d
þ
2
In a similar manner (cf. 3.8) ½Ph3PCFBr2ꢀBrꢁ (19.9 g, 0.037 mol),
(EtO)3P (6.23 g, 0.037 mol) and CF3CF2C(O)C6H5 (5.88 g, 0.03 mol)
afforded 5.6 g (58% of (E)- and (Z)-1-bromo-2-phenyl-F-1-butene
(bp 55–598/5.5 mm Hg) Z/E = 38/62. 19F NMR (CDCl3) (ppm):
1JCF = 312 Hz, CFBr),
5.91
d
m
m
decomposes on standing.
þ
4
5
d
= ꢁ54.6 (tq, JFF = 27 Hz, JFF = 11 Hz, (E)-vinyl F);
d
= ꢁ83.9 (dt,
3.8.6. Reaction of ½Ph3PCFBr2ꢀBrꢁ, (EtþO)3P and cyclohexanone
In a similar manner (cf. 3.8) ½Ph3PCFBr2ꢀBrꢁ (26.6 g, 0.05 mol),
(EtO)3P (8.31 g, 0.05 mol) and cyclohexanone (3.92 g, 0.04 mol)
afforded 4.7 g (61%) of 1-bromo-1-fluoromethylenecyclohexane
(bp 61.5–63.58/12 mm Hg). The wash procedure consisted of
NaHSO3 (40% 3ꢂ 35 ml) and water. 19F NMR (CDCl3) (ppm):
3
4
5JFF = 12 Hz, JFF = 3 Hz, (E)-CF3);
d
= ꢁ109.6 (dq, JFF = 27 Hz,
3JFF = 3 Hz, (E)-CF2);
d
= ꢁ46.0 (t, JFF = 8 Hz, (Z)-vinyl F);
4
3
4
d
= ꢁ82.5 (t, JFF = 3 Hz, (Z)-CF3);
d
= ꢁ111.0 (dq, JFF = 8 Hz,
3JFF = 3 Hz, (Z)-CF2). This 19F NMR data for these isomers was in
good agreement to that reported by Vander Haar [18,19].
d
= ꢁ82.6 (bs); 1H NMR (CDCl3) (ppm):
d = 1.4–1.6 and 1.9–2.4
þ
3.8.2. Reaction of ½Ph3PCFBr2ꢀBrꢁ, (EtO)3P and CF2ClC(O)C6H5
(broad singlets, ratio 6:4); IR: 5.97 mm (C C), mass spectrum (m/e,
þ
In a similar manner (cf. 3.8) ½Ph3PCFBr2ꢀBrꢁ (19.9 g, 0.037 mol),
(EtO)3P (6.23 g, 0.037 mol) and CF2ClC(O)C6H5 (5.72 g, 0.03 mol)
afforded 6.4 g (75%) of (E)- and (Z)-1-bromo-2-phenyl-3-chloro-F-
propene (bp 66–708/2.2 mm Hg) Z/E = 51/49. The wash procedure
relative intensity): 194 (22), 192 (26), 153 (26), 151 (29), 113 (53),
94 (43), 68 (100), 41 (32), 39 (48).
4. Conclusions
in this case consisted of 5% NaOH (3ꢂ 40 ml) and H2O (3ꢂ 50 ml).
4
19F NMR (CDCl3) (ppm):
d
= ꢁ46.5 (d, JFF = 29 Hz, (E)-CF2Cl);
Bromofluoromethylphenylphosphonium halides react with
trialkylphosphites in two distinct ways. Bromodifluoromethyltri-
phenylphosphonium halides undergo a rapid exchange reaction
with trialkylphosphites to give the corresponding bromodifluor-
omethylphosphonates. A similar exchange reaction also occurred
4
4
d
= ꢁ46.7 (d, JFF = 11 Hz, (Z)-CF2Cl);
d
= ꢁ57.4 (t, JFF = 29 Hz, (E)-
4
vinyl F);
(ppm):
d
= ꢁ57.4 (t, JFF = 11 Hz, (Z)-vinyl F); 1H NMR (CDCl3)
d = 7.41 (s). This spectral data was in close agreement with
the data for these isomers reported by Vander Haar [18,19].

R.M. Flynn et al. / Journal of Fluorine Chemistry 129 (2008) 583–589
589
References
with an analogous dialkoxyphenylphosphonite. Mechanistically,
the exchange process involves formation of difluorocarbene via
[1] D.J. Burton, Z.-Y. Yang, W. Qiu, Chem. Rev. 96 (1996) 1641–1715.
[2] D.J. Burton, D.G. Naae, R.M. Flynn, B.E. Smart, D.R. Brittelli, J. Org. Chem. 48 (1983)
3616–3618.
dissociation of the intermediate difluoromethylene ylide, caþpture
ꢁ
of the difluorocarbene by the trialkylphosphite to give ½ðROÞ3PCF2ꢀ,
which captures bromine followed by dealkylation to produce the
bromodifluoromethylphosphonate. The equilibria involved in the
multi-step mechanism are all shifted to the phosphonate product
by the final dealkylation step. This method provides a rapid clean
entry to bromodifluoromethylphosphonates or phosphinates from
a common intermediate. The yields are equal or better than the
Michaelis–Arbuzov methodology [18] and the reaction is espe-
cially useful for volatile phosphites due to the mild reaction
conditions (RT). In contrast to the bromodifluoromethylpho-
sphonium salt, dibromofluoromethyltriphenylphosphonium bro-
mide does not undergo exchange reactions with
trialkylphosphites. The phosphite serves as a halophilic reagent
to abstract Br from the dibromofluoromethylphosphonium salt to
generate the bromofluoromethylene ylide, which can easily be
trapped with aldehhydes or ketones in situ to give E/Z-bromo-
fluoroalkenes. This approach gives yields as good or better than
those from R3P/CFBr3 or metal dehalogenation of dibromofluor-
omethylphosphonium salts [1,18,19]. It also avoids the formation
of R3PBr2 as a by-product, which can react with aldehyde
substrates to decrease the yield of the bromofluoroolefin. It also
works well with ketones, such as C6H5C(O)CF2Cl, which reacts
readily with Zn or Zn/Cu, again lowering the yield of the olefinic
product.
[3] D.G. Naae, Ph.D. Thesis, University of Iowa, 1972.
[4] H.S. Kesling III, Ph.D. Thesis, University of Iowa, 1975.
[5] D.J. Burton, J. Fluorine Chem. 23 (1983) 339–357.
[6] D.G. Cox, D.J. Burton, J. Org. Chem. 53 (1988) 366–374.
[7] D.G. Cox, N. Gurusamy, D.J. Burton, J. Am. Chem. Soc. 107 (1985) 2811–
2812.
[8] D.J. Burton, D.G. Naae, J. Am. Chem. Soc. 95 (1973) 8467–8468.
[9] V. Mark, C.H. Dungan, M.M. Crutchfield, J.R. Van Wazer, Top. Phosphorus Chem. 5
(1967) 227.
[10] R.M. Flynn, R.G. Manning, R.M. Kessler, D.J. Burton, S.W. Hansen, J. Fluorine Chem.
18 (1981) 525–531.
[11] D.J. Burton, R.M. Flynn, J. Fluorine Chem. 10 (1977) 329–332.
[12] R.M. Flynn, Ph.D. Thesis, University of Iowa, 1979, pp. 78–79.
[13] (EtO)2P(O)CF2Br does not react with ethanol to give (EtO)2P(O)CF2H.
[14] Similar to the reaction outlined in Eq. (6), the trialkylphosphonium salts appear to
be unique in the formation of bis analogs.
[15] D.J. Burton, R.M. Flynn, R.G. Manning, R.M. Kessler, J. Fluorine Chem. 21 (1982)
371–376.
[16] When (1) and (2) were reacted with trifluoroacetophenone in CH2Cl2, a low yield
of 2-phenyl-F-propene (ꢃ15%) was detected by 19F NMR analysis of the reaction
mixture and the reaction was not clean. Thus, this route is not a good synthetic
route to difluoromethylene alkenes.
[17] D.J. Burton, J.L. Hahnfeld, J. Org. Chem. 42 (1977) 828–831.
[18] R.W. Vander Haar, D.J. Burton, D.G. Naae, J. Fluorine Chem. 1 (1971/1972) 381–
383.
[19] R.W. Vander Haar, Ph.D. Thesis, University of Iowa, 1973.
[20] J.M. Birchall, R.N. Haszeldine, J. Chem. Soc. 13 (1959).
[21] This yield was reported as a preliminary result in Ref. [2].
[22] Spectroscopic data for (CH3CH2O)2P(O)CF2Br: 19F NMR (CDCl3) (ppm): d = ꢁ61.9
(d, 2JPF = 92 Hz); 31P NMR (CDCl3) (ppm): d = 1.2 (t, 2JPF = 93 Hz); 13C NMR (CDCl3)
1
1
(ppm): d = 116.8 (td, JCF = 330 Hz, JCP = 238 Hz).
[23] Spectroscopic data for [(CH3)2CHO]2P(O)CF2Br: 19F NMR (CDCl3) (ppm): d = ꢁ62.5
(d, 2JPF = 93 Hz); 31P NMR (CDCl3) (ppm): d = 2.6 (t, 2JPF = 93 Hz); 13C NMR (CDCl3)
1
1
Acknowledgements
(ppm): d = 117.8.(td, JCF = 329 Hz, JCP = 237 Hz).
[24] For the use of JCF , [(E) and (Z)] and JHF [(E) and (Z)] to assign the configuration of
;F
fluoroolefins see3, D.J. Burton, H.C. Krutzsch, J. Org. Chem. 35 (1970) 2125–2130, X.
Zhang, L. Lu, D.J. Burton, Collect. Czech. Chem. Commun. 67 (2002) 1247–1261
and J.W. Emsley, L. Phillips, V. Wray, Fluorine Coupling Constants, Pergamon, NY,
1977.
We gratefully appreciate the financial support from the Air
Force Office of Scientific Research and the National Science
Foundation.