A novel synthesis of trifluoromethyl fluoroformate from trifluoromethyl hypofluorite and carbon monoxide in the presence of fluorine gas

Article abstract of DOI:10.1016/j.jfluchem.2008.06.017

Trifluoromethyl fluoroformate is prepared through a radical reaction between CF₃OF and CO initiated by elemental fluorine. The reaction could be integrated into a continuous synthesis process. This method represents a convenient synthesis of the title compound under mild condition.

Full text of DOI:10.1016/j.jfluchem.2008.06.017

Journal of Fluorine Chemistry 129 (2008) 1044–1046
Contents lists available at ScienceDirect
Journal of Fluorine Chemistry
journal homepage: www.elsevier.com/locate/fluor
A novel synthesis of trifluoromethyl fluoroformate from trifluoromethyl
hypofluorite and carbon monoxide in the presence of fluorine gas
Libin Du a, Darryl D. DesMarteau b,
*
a Department of Chemistry, The University of North Carolina, Chapel Hill, NC 27599, USA
b Department of Chemistry, Clemson University, Clemson, SC 29634, USA
A R T I C L E I N F O
A B S T R A C T
Article history:
Trifluoromethyl fluoroformate is prepared through a radical reaction between CF3OF and CO initiated by
elemental fluorine. The reaction could be integrated into a continuous synthesis process. This method
represents a convenient synthesis of the title compound under mild condition.
ß 2008 Elsevier B.V. All rights reserved.
Received 7 April 2008
Received in revised form 16 June 2008
Accepted 17 June 2008
Available online 24 June 2008
Keywords:
Trifluoromethyl fluoroformate
Synthesis
Elemental fluorine
Radical chain reaction
Continuous synthesis
1. Introduction
dehalogenation [20]. However, the current synthetic route to 1
is inefficient and the reaction conditions can easily decompose 1
The preparation of trifluoromethyl fluoroformate, CF3OC( O)F
(1), was first achieved through a photochemical reaction of CF3OF
and CO initiated by a Hg arc filtered through Pyrex glass and H2O
[1]. A pyrolytic version of the reaction was also mentioned therein,
but without any detail. A later patent described the preparation of
1 as the dimer of carbonyl fluoride, COF2 (2), using a variety of
catalysts [2]. It was also isolated as a minor product or the
rearranged product in a variety of other reactions [3–9]. The
efficient preparation of 1 is of considerable interest currently as the
key intermediate to a vinyl ether, CF3OCF2OCF CF2, whose
copolymers have been used to formulate novel fluoroelastomers
with very low glass transition temperatures [10–17]. These
properties are essential in certain areas such as aerospace as
exemplified by the discovery space shuttle disaster which was
caused by failure of elastomer seals at low temperature [17]. The
fluoroformate 1 can be easily decomposed by strong nucleophiles
such as fluoride ion [18,19], so the vinyl ether, CF3OCF2OCF CF2,
could only be prepared through photofluorination of 1 with
elemental fluorine followed by addition to CFCl CFCl and
due to the possible generation of fluoride ions [10,11,14].
Herein, we report an alternative synthetic pathway to prepare 1
through a facile radical reaction between CF3OF and CO initiated by
elemental fluorine. Compared with the previous reported pro-
cesses [1,2,10–15], this route does not require high temperatures
or UV radiation, and can occur spontaneously at or below room
temperature. The reaction is rapid and has potential for integration
into a continuous process which may be difficult for other
previously reported routes.
2. Results and discussion
In the photolysis and pyrolysis of CF3OF and CO, CF3OCO
radicals formed through combination of a CF3O radical and CO
were proposed in the chain reaction forming 1 [1]. However, the
cleavage of O-F bond requires either high temperatures or UV
irradiation, and the yields were relatively low with high
concentrations of side products such as COF2 (2) [10,11,14]. The
formation of CF3OCO radicals are also expected to compete with
the decomposition of CF3O radical into COF2 and F atoms which
may in part be responsible for the generation of COF2 as a side
product [1]. In order to carry out the reaction under milder
conditions such as room temperature, providing an easily
generated radical to initiate the fast chain reaction was considered.
* Corresponding author. Tel.: +1 864 656 1251; fax: +1 864 656 2545.
E-mail addresses: dlibin@email.unc.edu (L. Du), fluorin@clemson.edu
(D.D. DesMarteau).
0022-1139/$ – see front matter ß 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.jfluchem.2008.06.017

L. Du, D.D. DesMarteau / Journal of Fluorine Chemistry 129 (2008) 1044–1046
1045
Table 2
Reactions of CF3OF and CO initiated by F2 with unequal pressures in the flasks
CF3OC(O)Fa
(mmol)
Yieldb
(%)
COF2
a
Entry
CF3OF/F2 in
the small
flask
CO in the
big flask
(mmol)
mmol
atm
mmol
atm
1
2
3
10/1.5
0.57
0.52
0.52
78.5
76.9
20.4
0.96
0.94
0.25
0.9c
0.9d
0.4d
9.0
8.8
4.0
4.3
10.2/0.4
10.1/0.6
0.05
0.2
Scheme 1. The radical reaction between CF3OF and CO initiated by F2.
a
Based on the integration on 19F NMR spectra.
The yields were calculated based on the amount of CF3OF.
Contaminated with some unknown side products.
Pure.
b
c
d
Fig. 1. Static system used in the reaction between CF3OF and CO initiated by F2.
A fluoroformyl radical, FC( O) (3), generated from the facile
reaction of F2 and CO was reported to be responsible for the
efficient preparation of FC( O)OO( O)CF through a fast chain
reaction at room temperature [21]. We expected that this
fluoroformyl radical would also initiate a chain reaction between
CF3OF and CO, under milder conditions (see Scheme 1).
Fig. 2. The 1 L flow system used in the reaction of CF3OF and CO initiated by F2.
1). A high ratio of CF3OF to F2 increases the probability of 3 reacting
with CF3OF rather than F2 (Scheme 1). The effect of the amount of
N2 diluent on the reaction was not studied.
The reaction was first attempted in a closed system (see Fig. 1)
with nitrogen gas being used as a diluent in the reaction. Carbon
monoxide was added in excess compared to CF3OF in order to
facilitate the desired reaction to form 3 and 1 and minimize the
side product 2. (Caution: the reaction between fluorine and carbon
monoxide can be explosive, so it must be carried out in a controlled
manner!) A relatively high yield of 18% was obtained in a run with
a 20/3 ratio of CF3OF to F2 (entry 1 in Table 1). The amount of
product 1 was approximately equal to the amount of the fluorine
feed in entry 1, so it was initially thought that the yield was closely
related to the amount of elemental F2. Increasing the ratio of F2 to
CF3OF with the intent to improve the yield in the second run,
resulted in a violent reaction in which 1 was not even detected as a
product (entry 2). On the other hand, the lower ratio of F2 to CF3OF
in entry 3 generated a negligible amount of COF2, but gave a much
lower yield of 1.
In Table 1 the initial pressures in the two flasks were nearly the
same and the mixing of the gases on opening the connecting valve
was assumed to be slow. In Table 2 runs were conducted without
N2 diluent with a higher pressure in one of the flasks. The
assumption was that the gases would mix more rapidly. Contrary
to expectations, higher yields of 1 were not observed.
Based on the static reactions in Tables 1 and 2, a flow reaction
was carried out in a 1 l glass apparatus shown in Fig. 2. The low
conversion (3%, see Section 3) was disappointing but we expect
that this reaction could be greatly improved by reactor design and
varying reactant concentrations. The experiments described above
were run at room temperature without any mechanical mixing.
The reaction could easily become explosive as the heat dissipation
was inefficient under this condition. However, in a possible
industrial setup [18], the reaction could be run in an inert solvent
with efficient agitation. The reaction could also be run at lower
temperatures which should help to suppress the side reactions
such as the one between F2 and CO to generate COF2. Both CO and
CF3OF would be in a large excess compared with F2. In such a
system, the COF2 could be recycled to CF3OF and un-reacted CF3OF
The abstraction of a CF3O radical from CF3OF by 3 appears to be
much less energetically favored than its reaction with F atoms or
molecular fluorine. A large excess of CO is always needed to trap F
atoms before they can combine with 3 to form COF2 (see Scheme
Table 1
Reactions of CF3OF and CO initiated by F2 with equal pressures in the two flasks
a
Entry
CF3OF (mmol)
F2 (mmol)
CO (mmol)
N2 (mmol)
CF3OC(O)Fa (mmol)
Yieldb (%)
COF2 (mmol)
1
2
3
20.1
19.2
16.0
3.1
11.9
1.6
40.8
60.5
32.4
50.9
30.8
34.2
3.6
–
18.0
–
>0.7
>10.0
Trace
0.7
4.2
a
Based on the integration on 19F NMR spectra.
b
The yields were calculated based on the amount of CF3OF.

1046
L. Du, D.D. DesMarteau / Journal of Fluorine Chemistry 129 (2008) 1044–1046
would be easily recycled, so the single-cycle yield might not be
critical. An optimized system of this type would require consider-
able engineering. We made no attempt in this study to optimize
this process but simply wanted to demonstrate that a flow system
with potential for scale-up was viable.
CF3OC( O)F was 18.4%. The characteristic properties and spectral
values of CF3OC( O)F agreed with the literature [1,8]. 19F NMR
(CDCl3):
1F, CF).
d
= À60.5 ppm (d, 3F, 3JF–F = 9.9 Hz, CF3),
= À13.5 ppm (q,
d
In summary, the preparation of CF3OC( O)F has been achieved
at room temperature through a radical reaction between CF3OF
and CO initiated by the presence of elemental fluorine. This
reaction has potential for scale-up in a continuous flow system.
3.3.2. Preparation of CF3OC( O)F in the flow system
The reaction was carried out in a flow system (Fig. 2). Helium
and CO gas cylinders were attached to the system through two
calibrated Gilmont gas flow meters. A passivated Monel bomb
reactor containing 61 mmol CF3OF and 10.2 mmol of F2 was also
attached to system through a digital Hastings Fluorine mass
flowmeter. The gas outlet was attached to an oil bubbler containing
Krytox1 fluids through a collection trap maintained at À196 8C.
First, the system was purged with helium gas at 54 ml/min for
50 min followed by CO at 20 ml/min for another 30 min. The bomb
reactor containing the mixture of CF3OF and F2 was then slightly
opened and the flow rate was adjusted to 15 ml/min initially with
the helium flow being reduced to 40 ml/min simultaneously. The
flow rate of the mixture of CF3OF and F2 decreased slowly as the
pressure dropped in the cylinder and was adjusted over time to
maintain the flow at 10–30 ml/min during the addition. The flow of
CF3OF and F2 from the bomb reactor continued for 2.5 h. The CO
was then stopped after 5 min and the helium flow was continued
for 45 min. The contents of the 196 8C trap were separated by trap-
to-trap distillation using two traps at À135 and À196 8C,
respectively. About 45.6 mmol CF3OF contaminated with a little
COF2 was collected in À196 8C trap, and 12.4 mmol of the mixture
of CF3OF and CF3OC( O)F were collected in À135 8C trap. The
mixture in the À135 8C trap was then separated using À130 and
À196 8C traps. Pure CF3OC( O)F (1.8 mmol) was collected in the
À130 8C trap. The isolated yield of CF3OC( O)F was approximately
3%.
3. General experimental procedures
3.1. Instruments
19F NMR spectra were recorded on a JOEL FT/NMR at 282.8 MHz.
The samples were sealed in a 4-mm glass tube with CCl4 as solvent
and CFCl3 as reference. The 4-mm tube was inserted into a
standard 5-mm NMR tube for measurement. The lock D2O was
added to the outer tube. Infrared spectra were recorded on a
PerkinElmer spectrum 2000 FTIR. The sample was contained in a
10 cm glass cell fitted with a Kontes glass-Teflon valve and AgCl
windows held in place with Halocarbon 1500 wax.
3.2. Apparatus and reagents
The reactions were carried out either in a static (Fig. 1) or a flow
systems (Fig. 2). The closed system was comprised of two bulbs
sized at 2000 and 500 ml, respectively, connected by a Kontes
glass-Teflon valve. The reactor was dried at 100 8C in an oven
followed by an overnight evacuation for 10–12 h at 22 8C. The flow
reactor was first dried at 100 8C in an oven followed by purging
with helium gas for 50 min at 22 8C.
The preparation of CF3OF was carried out according to the
standard literature procedure [22]. Fluorine gas was purchased
from Air Products and Chemicals and passed through a NaF
scrubber before use. (Caution: fluorine and CF3OF are extremely
reactive and toxic and must be handled with care!) Phosgene, CO
and CsF were used as received. Carbonyl fluoride was prepared
from phosgene by reaction with NaF in CH3CN at 40–50 8C.
Acknowledgement
We gratefully acknowledge the financial support of this
research by Solvay Solexis.
References
[1] P.J. Aymonino, Chem. Commun. 1 (1965) 241.
3.3. Typical procedure
[2] B.C. Anderson, G.R. Morlock, US Patent 3 226 418 (1965).
[3] D. Sianesi, G. Bernardi, G. Moggi, FR Patent 1 545 639 (1968).
[4] D. Sianesi, G. Bernardi, G. Moggi, US Patent 3 721 696 (1973).
[5] F.A. Hohorst, D.D. DesMarteau, L.R. Anderson, D.E. Gould, W.B. Fox, J. Am. Chem.
Soc. 95 (1973) 3866–3869.
[6] M. dos Santos Afonso, R.M. Romano, C.O. Della Vedova, J. Czarnowski, Phys. Chem.
Chem. Phys. 2 (2000) 1393–1399.
[7] K. Takahashi, Y. Matsumi, T.J. Wallington, M.D. Hurley, Chem. Phys. Lett. 352
(2002) 202–208.
[8] T. Johnston, J. Heicklen, W. Stuckey, Can. J. Chem. 46 (1968) 332–334.
[9] M. Mashino, M. Kawasaki, T.J. Wallington, M.D. Hurley, J. Phys. Chem. A 104
(2000) 2925–2930.
[10] M. Apostolo, F. Triulzi, V. Tortelli, M. Galimberti, EP 1 621 551 (2006).
[11] M. Apostolo, F. Triulzi, V. Tortelli, M. Galimberti, EP 1 621 558 (2006).
[12] M. Albano, M. Stanga, F. Triulzi, WO 2 007 085 545 (2007).
[13] F. Triulzi, M. Albano, M. Stanga, WO 2 007 082 867 (2007).
[14] F. Triulzi, M. Apostolo, V. Tortelli, M. Galimberti, EP 1 621 559 (2006).
[15] M. Stanga, F. Triulzi, M. Albano, WO 2 007 085 546 (2007).
[16] V. Tortelli, P. Calini, US Patent 2 004 186 324 (2004).
[17] J. Scheirs, Structure/property considerations for fluoropolymers and fluoroelas-
tomers to avoid in-service failure, in: J. Scheirs (Ed.), Modern Fluoropolymers:
High Performance Polymers for Diverse Applications, John Wiley & Sons, Chiche-
ster/New York/Weinheim/Brisbane/Singapore/Toronto, 1997, pp. 1–71.
[18] L. Du, PhD Thesis, Clemson University, 2006.
[19] I. Tari, D.D. DesMarteau, J. Org. Chem. 45 (7) (1980) 1214–1217.
[20] G. Marchionni, U. De Patto, G. Spataro, V. Tortelli, J. Fluorine Chem. 119 (2003) 83.
[21] A. Russo, D.D. DesMarteau, Inorg. Chem. 34 (1995) 6221–6222.
[22] J.K. Ruff, A.R. Pitochelli, M. Lustig, J. Am. Chem. Soc. 88 (1966) 4531–4532.
3.3.1. Preparation of CF3OC( O)F in the closed system
In a typical reaction, 40.8 mmol CO and 50.9 mmol N2 were first
added to 2 L bulb of the closed system (Fig. 1). A premixed sample
of 20.1 mmol CF3OF and 3.1 mmol F2 was then added to the smaller
0.5 L bulb. (Caution: the reaction between fluorine and carbon
monoxide can be explosive and must be carefully controlled!) The
inside pressure of the 2 L bulb was calculated to be 1.14 atm vs.
1.16 atm in the 0.5 L bulb. With the reactor at 22 8C, the valve
connecting the two bulbs was opened slowly and the rector was
kept at room temperature for 48 h behind a shield. The contents of
the reactor bulbs were then collected through traps at À135, À146
and À196 8C. CF3OC( O)F (2.1 mmol) was collected in the À135 8C
trap. The rest of the chemicals collected in the other two traps were
CF3OF contaminated with CF3OC( O)F and COF2. The latter was
then passed through three traps at À130, À145 and À196 8C. A ca.
1.6 mmol of CF3OC( O)F contaminated with a little COF2 was
collected in the À130 8C trap. Carbonyl fluoride contaminated with
CF3OC( O)F (0.7 mmol) was collected in À145 8C trap and
16.1 mmol of CF3OF contaminated with a small amount of COF2
was collected in the À196 8C trap. The overall isolated yield of