Oxidation of hexafluoropropylene with oxygen to hexafluoropropylene oxide

Article abstract of DOI:10.1021/op900151p

A method for pressure oxidation of hexafluoropropylene (HFP) with oxygen to hexafluoropropylene oxide (HFPO) has been presented. The oxidation was carried out in a periodically operated reactor. The influence of temperature, the nature of the solvent, and the molar ratio of HFP/O₂ on the course of oxidation was investigated during the preliminary studies. The magnitudes describing the process were the conversion of HFP, the selectivity of transformation to HFPO in relation to HFP consumed, and the yield of HFPO in relation to the initial amount of HFP. The optimisation of the technological parameters of oxidation, temperature, HFP/O₂ molar ratio, and the reaction time was performed by the application of a statistical experimental design method. The function describing the process was the yield of HFPO.

Full text of DOI:10.1021/op900151p

Organic Process Research & Development 2010, 14, 272–277
Oxidation of Hexafluoropropylene with Oxygen to Hexafluoropropylene Oxide
Agnieszka Wro´blewska,* Eugeniusz Milchert, and Egbert Meissner
Institute of Organic Chemical Technology, West Pomeranian UniVersity of Technology, Szczecin,
Pułaskiego 10, Pl 70-322 Szczecin
Abstract:
phase-transfer catalysts. However, this method is only used for
preparative purposes and on a small scale.5,8
A method for pressure oxidation of hexafluoropropylene (HFP)
with oxygen to hexafluoropropylene oxide (HFPO) has been
presented. The oxidation was carried out in a periodically operated
reactor. The influence of temperature, the nature of the solvent,
and the molar ratio of HFP/O2 on the course of oxidation was
investigated during the preliminary studies. The magnitudes
describing the process were the conversion of HFP, the
selectivity of transformation to HFPO in relation to HFP
consumed, and the yield of HFPO in relation to the initial
amount of HFP. The optimisation of the technological
parameters of oxidation, temperature, HFP/O2 molar ratio,
and the reaction time was performed by the application of
a statistical experimental design method. The function
describing the process was the yield of HFPO.
The oxidation of HFP with oxygen in the gaseous phase in
the presence of catalysts such as barium compounds, activated
silicagels, metals (Ag, V), supported silicagel, or diatomaceous
earth was also investigated.4,9
The greatest expectations to realize the commercial produc-
tion of HFPO are associated with the pressure oxidation of HFP
with oxygen in the liquid phase.10,11
The investigations on the preparation of this compound by
the method of catalyst-free oxidation using hydrogen peroxide
in the medium of a water-soluble polar solvent such as acetone
or acetonitrile are still on-going.12
The electrochemical oxidation was also the subject of
interest.1 An aqueous solution of acetonitrile or its mixture with
acetic acid and a compound to improve the conductivity
(NaClO4) was used as an electrolyte.
However, the above-mentioned methods created a
problem associated with the separation of pure HFPO from
the postreactive mixture. This resulted from the similar
boiling points of HFP and HFPO. The problem of HFPO
purification was solved by conducting the oxidation with
oxygen under supercritical carbon dioxide.13 As a result
of irradiation of the reactor by UV light, the produced
HFPO immediately underwent polymerization, and per-
fluoropolyether was formed.
1. Introduction
Hexafluoropropylene oxide (HFPO) is a valuable monomer
which is subjected to oligomerization in order to prepare the
intermediates for the syntheses of highly effective nonionic,
ionic, and amphoteric surface active agents. It was found that
even small amounts of these compounds cause a significant
decrease in the surface tension of water, and for this reason
they, among others, are used as the components of the
extinguishing agents.1 The HFPO oligomers are also used as
inert solvents and hydraulic fluids.2 Higher oligomers are
important as lubricating oils with regard to high thermostabi-
lity and chemical resistance.3,4
The main reaction of oxidation of hexafluoropropylene with
oxygen both in the liquid phase as well as in the gaseous phase
can be written as follows:
Up to date, HFPO was produced by oxidation with sodium
hypochlorite,5 with oxygen under pressure6 and with the
pressureless method,4 by means of hydrogen peroxide,7 and by
the electrochemical method.1 HFPO can be obtained with high
efficiency by the epoxidation of HFP with hypochlorite using
the method of a phase-transfer catalyst in the system of two
phases, water and an organic solvent, in the presence of an
inorganic base. Quaternary ammonium salts, quaternary phos-
phonium salts, and quaternary arsenium salts, and the lipophilic
reagents complexed with a sodium cation were used as the
According to the assumed reaction mechanism, after the
addition of oxygen to the double bond, a biradical is formed
with electrons on the oxygen and carbon. Its decomposition
and rearrangement as well as the decomposition of HFPO cause
the formation of several byproducts such as trifluoroacetic acid
fluoride, carbonyl fluoride, 1-trifluoromethyl-perfluorocyclo-
propane, perfluoroethylene, and perfluorocyclopropane. Tem-
porarily, carbine is also formed. This can be illustrated by the
following chemical equations.14
* Corresponding author. Telephone: + 48 914494875. Fax + 48 914494365.
E-mail: Agnieszka.Wroblewska@zut.edu.pl.
(8) Ikeda, M.; Miura, M.; Aoshima, A. U.S. Patent 4,925,961, 1990.
(9) Zhonghi, H.; Yongming, Z.; Chehgxue, Z.; Jianhao, Q.; et al. Appl.
Catal. 2006, 303, 18.
(1) Fiters, D. W.; Griffiths, D. J.; Nash, P. Fine Technology 1969, 5 (4),
284.
(2) Millauer, H. U.S. Patent 4,014,762, 1977.
(10) U.S. Patent 1,034,492, 1964.
(3) U.S. Patent 3,660,315, 1972.
(4) U.S. Patent 3,775,439, 1973.
(11) U.S. Patent 3,536,733, 1970.
(12) U.S. Patent 3,358,003, 1967.
(5) Ikeda, M.; Miura, M.; Aoshima, A. U.S. Patent 4,902,810, 1990.
(6) Meissner, E.; Wroblewska, A. Polish J. Chem. Technol. 2007, 9, 20.
(7) Meissner, E.; Wroblewska, A. Polish J. Chem. Technol. 2006, 8, 66.
(13) W.O. Patent 28,205, 1997.
(14) Gilbert, J. R.; Slange, I. R.; Graham, E. R.; Gutman, D. J. Phys. Chem.
1976, 80, 14.
272
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development
10.1021/op900151p CCC: $40.75  2010 American Chemical Society
Published on Web 07/08/2009

valve, and a sensor for measurement of temperature within the
autoclave and on its external wall. The autoclave was purged
with nitrogen, and a volumetrically determined amount of
solvent was introduced into the autoclave. After closing the
cover, the autoclave was cooled in the cryostat to the temper-
ature of about -50 °C. Subsequently, the air was removed from
the autoclave with the oil vacuum pump. From a gas cylinder
located on the balance, a proper amount of HFP was distilled
off to the autoclave through the dosing valve. This amount of
HFP was determined by the weight. The autoclave was taken
out of the cryostat and was left to achieve ambient temperature.
A measured amount of oxygen was introduced using a special
pressure feeder. For this purpose a container for oxygen (4,
Figure 1) was filled up with carbon tetrachloride.
The oxygen supplied from gas cylinder 3 was introduced
into volumetrically calibrated container 4. This container was
previously filled up with carbon tetrachloride. After opening
valves 8 and 14, container 4 was filled up with oxygen.
Simultaneously after opening valve 9, the flow of carbon
tetrachloride took place from container 4 to container 5, which
was previously filled up with nitrogen under atmospheric
pressure. The amount of oxygen in container 4 under the
atmospheric pressure was read out on level gauge 7, after
closing valves 9 and 14 and opening valve 11. The total amount
of oxygen was introduced in three portions, at equal time
intervals. The atmospheric pressure and overpressure were
measured using mercury-filled manometer 6. After the equaliza-
tion of pressure in oxygen container 4 with atmospheric
pressure, an overpressure was generated in the nitrogen
The objective of this work was to investigate the influence
of the changes of technological parameters on the course of
direct oxidation of HFP with oxygen in the liquid phase under
pressure and to establish the most advantageous parameters of
process operation.
2. Experimental Section
2.1. Raw Materials. HFP was prepared by the debromi-
nation of 1,2-dibromohexafluoropropane. Oxygen was obtained
from Messer-Poland. Carbon tetrachloride, methylene chloride,
and 1,2,2-trichloro-1,1,2-trifluoroethane (CFC-113) were ob-
tained from POCh, Poland.
2.2. Oxidation Procedure. The oxidation was carried out
in the liquid phase in a steel autoclave made by Lampart
(Hungary) of 500 cm3 capacity. The autoclave content was
mixed via the swinging movements of a horizontally assembled
cradle. This cradle was electrically heated. The autoclave cover
was equipped with two metering valves, a manometer, a safety
Figure 1. Scheme of the equipment for HFP oxidation: 1-autoclave, 2-nitrogen cylinder, 3-oxygen cylinder, 4-oxygen pressure
container, 5-nitrogen pressure container, 6-mercury manometer, 7-level gauge, 9-15-valves, 16-manometer.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
273

Table 1. Influence of temperature on the conversion of HFP,
selectivity of transformation to HFPO, and the yield of
HFPOa
container 5 that exceeded the pressure in the autoclave 1 by
about 3 MPa. Nitrogen from cylinder 2 was utilized for this
purpose. Under this overpressure oxygen was dosed to the
autoclave 1. The oxidation was carried out using an initially
established temperature for the determined HFP/O2 molar ratio
at the fixed residence time.
The mass balance was performed for each synthesis
after completing the reaction. The following magnitudes
describe the process determined: conversion of HFP
(CHFP), selectivity of transformation to HFPO (SHFPO), and
the yield of HFPO (YHFPO). They are calculated according
to the following equations:
reaction
temperature of the
CHFP SHFPO YHFPO
temperature °C autoclave wall ((5 °C) mol % mol % mol %
130
140
145
150
155
160
170
150
160
165
170
175
180
185
12
32
57
81
99
99
99
83
81
77
68
62
59
50
10
26
44
55
62
58
49
a Constant parameters: HFP ) 0.766 mol; total O2 ) 9 dm3 - 3 dm3 of O2 was
introduced 3 times after 0, 60, 120 min; molar ratio HFP/O2 ) 1.91:1; solvent -
carbon tetrachloride 385 cm3; reaction time
3 h. CHFP - conversion, SHFPO -
amount of HFP consumed (mol)
initial amount of HFP (mol)
selectivity, YHFPO - yield.
CHFP
SHFPO
YHFPO
)
)
)
· 100%
· 100%
· 100%
amount of HFPO obtained (mol)
amount of HFP consumed (mol)
The liquid products of photobromination were analysed
quantitatively with the same apparatus under the following
conditions: detector FID, stainless steel packed column (3 m
× 0.4 cm) with 3% silicone oil SE-30 on Chromosorb G 80/
100 mesh. The column temperature was programmed isother-
mally at 60 °C for 4 min, followed by an increase at a rate of
8 °C/min to 170 °C, isothermally 170 °C for 10 min, injection
sample 0.1 µL.
amount of HFPO obtained (mol)
initial amount of HFP (mol)
2.3. Separation of HFPO. Due to the similar volatilities
of HFP (bp -29.4 °C) and HFPO (bp -27.4 °C) these
compounds cannot be separated by the distillation meth-
ods. With regard to close boiling points, another method
for the separation of oxidation products has been devel-
oped. The reaction products in the gaseous form were
directed to a scrubber sprinkled with a 5 wt % aqueous
solution of sodium hydroxide. During the absorption there
proceeded the hydrolysis of trifluoroacetic acid fluoride
and carbonyl fluoride, followed by the neutralization of
the products. A formed eluate was analyzed for the content
of fluoride ion. Gas that did not undergo absorption was
directed to the photobromination. As a result of HFP
photobromination there was formed 1,2-dibromohexafluo-
ropropane that remained in the reactor as a compound
having a higher boiling point, whereas the gases leaving
the reactor were subjected to the treatment in the scrubber
with an aqueous solution containing 5 wt % of NaOH and
10 wt % of NaHSO3. The treated gases were condensed
in the freezer and cooled to a temperature of -40 °C.
The product after the condensation contained at least 99.8
wt % of HFPO. The analyses were carried out with a GC
method according to the procedure presented below.
2.4. Analytical Methods. The fluoride ions were deter-
mined potentiometrically in the eluate after the absorption
of gaseous products of oxidation. A Radelkis OP-108
apparatus with a fluoride electrode Orion 9403 was used
for measurements.
The mass balance of experiments was performed on the basis
of these analyses.
3. Results and Discussion
3.1. Preliminary Investigations. The influences of tem-
perature, the nature of the solvent, and the molar ratio of HFP
to oxygen on the HFP conversion, the selectivity of transforma-
tion to HFPO, and the yield of HFPO have been determined
during the preliminary investigations. The results of studies on
the effect of temperature are presented in Table 1.
The results presented in Table 1 indicate that an increase
in the reaction temperature in the range of 130-170 °C
increases the conversion of HFP from 12 mol % at a temper-
ature 130 °C to practically the complete conversion in the
temperature range 155-170 °C. An inverse tendency of the
changes in the regarded temperature range takes place in
the case of the selectivity of conversion to HFPO. The highest
selectivities at a level of 80 mol % are achieved in the range of
temperatures 130-140 °C; subsequently, the selectivities
decrease to 50 mol % at a temperature of 170 °C. Hence, the
highest yield is a choice between a high selectivity of
transformation to HFPO and a high conversion of HFP. It was
found that, under the specified technological conditions, a high
selectivity of transformation to HFPO in relation to HFP
consumed corresponds to a lower conversion of HFP and vice
versa.
The studies on the effect of the nature of the solvent (Table
2) under the remaining conditions being comparable indicate
that the highest conversion of HFP, selectivity of transformation
to HFPO, and the yield of HFPO is achieved with carbon
tetrachloride.
Table 3 presents data on the influence of the molar ratio of
HFP to oxygen on the selectivity and conversion of HFP. It
results from these data that high selectivity of transformation
to HFPO in relation to HFP consumed can be obtained in case
The quantitative analyses of HFPO and the gaseous products
of HFP oxidation were performed chromatographically on a
Chrom 5 apparatus equipped with a thermal conductivity
detector (TCD). The concentrations of the following compounds
in this gas were determined: HFP, HFPO, perfluoroethylene,
perfluorocyclopropane, trifluoroacetic acid fluoride. A stainless
steel column 2.5 m × 0.4 cm packed with silica gel ASTM-E
80/100 mesh was used. The remaining conditions of detection
were as follows: carrier gas H2 ) 20 cm3/min, oven temperature
) 150 °C, injection sample ) 100 µL.
274
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

Table 2. Influence of solvent on the conversion of HFP, selectivity of transformation to HFPO, and the yield of HFPOa
solventb
cm3
temperature °C
HFP mol
oxygen dm3
CHFP mol %
SHFPO mol %
YHFPO mol %
CCl4
CH2Cl2
CFC-113
380
380
380
150
155
152
0.766
0.766
0.766
9
9
9
85
64
76
64
72
58
55
46
44
a Constant parameters: molar ratio HFP/O2 ) 1.91:1, 3 dm3 of oxygen was introduced three times: after 0, 60, and 120 min. b The amount of solvent was constant in all
experiments.
Table 3. Influence of molar ratio HFP/O2 on the conversion
Table 4. Real and normalized values of the technological
of HFP, selectivity of transformation to HFPO, and the yield
parameters
of HFPOa
real factors
HFP/O2,
normalized temperature,
reaction
HFP/O2
CHFP
SHFPO
YHFPO
HFP mol oxygen dm3 mol/mol mol % mol % mol %
level
factor Xi
x1 °C
x2 mol/mol time, x3 h
lower
basic
-1.682
0
150
165
180
0.4
1.7
3.0
1
3
5
0.766
0.766
0.766
0.766
0.766
14
11
9
0.62:1
0.78:1
0.96:1
1.24:1
1.74:1
99
99
86
82
62
32
36
62
65
72
32
36
54
53
45
higher
1.682
7
The design matrix utilized for the experiment realization, the experimentally
obtained values of HFPO yields, and the results of statistical analysis were compiled
in Table 5.
5
a Constant parameters: temperature 155°C, carbon tetrachloride 385 cm3,
reaction time 3 h, oxygen was introduced in 3 portions, after 0, 60, 120 min.
Table 5. Design matrix, the yield of HFPO, and results of
statistical analysisa
of excessive HFP in relation to oxygen. However, this causes
a decrease of the conversion of HFP as well as the yield of the
HFPO.
X1 temperature X2 HFP/O2 X3 reaction time YHFPO mol %
1
-1.682
+1.682
-1.682
+1.682
-1.682
+1.682
-1.682
+1.682
0
-1
-1
+1
+1
-1
-1
+1
+1
0
-1
-1
-1
-1
+1
+1
+1
+1
0
16
18
25
28
29
28
35
32
44
44
43
2
3
3.2. Optimization of Oxidation Process of HFP. The
following ranges of changes in the parameters describing
the oxidation process of HFP were established on the basis
of the preliminary studies: temperature 150-180 °C, molar
ratio of HFP/O2 0.4-3.0 mol/mol, and the reaction time
1-5 h. The amount of carbon tetrachloride used as a
solvent was equal to 400 cm3 in all the experiments. The
optimization was performed for the function of HFPO
yield. The optimization of the technological parameters
of HFP oxidation was performed according to the math-
ematical method of experiment design.15,16 The plan was
realized for three technological parameters (input variables
x1-x3): x1 ) temperature, x2 ) molar ratio of HFP/O2, x3
) the reaction time and for two levels. The number of
possible combinations for the above-mentioned system
amounted to 23 ) 8. The total number of experiments
amounted to 11, whereas there were 3 in the plan centre.
The real and normalized values of the technological
parameters (input variables) at levels resulting from
experimental design are presented in Table 4.
4
5
6
7
8
9
10
11
0
0
0
0
0
0
a R2 ) 99.83, S2
) 1.35.
) 1.8218, S2
) 18200, frepeat ) 2, F ) 1.000989, ∆zmax
repeat
adeq
(regression equations) containing the linear and square com-
ponents and the double products (interactions):
YHFPO ) -893.579 + 10.834x1 + 12.12x2 + 17.93x3 +
0.00705x1x2 - 0.0437x1x3 - 0.361x2x3 - 0.0325x21 -
2.797x22 - 2.31x23
The coefficients of the regression equation for the
normalized input variables (parameters) were determined
by the least-squares method with the application of matrix
calculations. After the determination of the function
approximating the experimental results, a verification of
the adequacy of the equation based on the Fisher-Snedecor
test on significance level R ) 0.05 was proved as F <
F(R). This function is adequate because it has a lower
value than that read out from F(R) tables. The values of
R2, the coefficient of multiple correlation, S2adeq, the
variance of adequacy, S2repeat, the variance of inaccuracy,
frepeat, the freedom degrees number of variance of inac-
curacy, F, the value of the quotient of the test of
Fisher-Snedecor, and ∆zmax, the maximum error of
approximation, are given in Table 5.
Experimental design and the calculation results were
performed by computer, applying the program Cadex:
Esdet 2.2.
The influence of independent parameters: temperature, HFP/
O2 molar ratio, reaction time on the response function, HFPO
yield, was presented in the form of second-order polynomial
(15) Montgomery, D. C. Design and Analysis of Experiments; John Wiley
& Sons: New York, 1976.
(16) Polan´ski, Z.; Goˇrecka-Polan´ska, R. Cadex:Esdet 2.2 Program, Planning
and Statistical Analysis of Experimental Investigations Applying
Determined Statistical Analysis; Department of Technical Development
and Implementation: Krako´w, Poland, 1992.
The obtained equation of HFPO yield is adequate and has a
high value of the multiple correlation coefficient R2 ) 99.83.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
275

Figure 2. Influence of temperature and reaction time on the
yield of HFPO, HFP/O2 ) 1.7:1.
Figure 3. Influence of HFP/O2 molar ratio and the reaction
time on the yield of HFPO, temperature 165 °C.
The yields calculated from this equation correlate well with
values obtained experimentally for the particular points of
rotatable uniform design.
The course of variations in the function of HFPO yield
(YHFPO) at the constant molar ratio of HFP/O2 ) 1.7:1
during the changes of temperature, and time is shown in
Figure 2.
Figure 2 shows that the highest yield of HFPO on a
level of YHFPO ) 43 mol % is achieved during the changes
of temperature in the range 160-170 °C for the reaction
times 3.1-4.5 h. The prolongation of the reaction time
over 4.5 h does not cause significant changes in the value
of the yield function of HFPO. A decrease of temperature
below 160 °C or its increase above 170 °C causes a
decrease of value of this function by a 12 mol %. This
results from the rate of the epoxidation reaction being
lower when the temperature is below 160 °C. However,
above 170 °C the reactions lead to the formation of
byproducts, including the reaction associated with the
decomposition of HFPO to trifluoroacetic acid fluoride.
It was found in several experiments carried out at temper-
atures above 170 °C that the epoxidation process pro-
ceeded rapidly and even explosively. This was visualized
by a rapid increase of both pressure and temperature. In
this case, the major product of reaction was carbonyl
fluoride together with a small amount of trifluoroacetic
acid fluoride.
Figure 4. Influence of temperature and HFP/O2 molar ratio
on the yield of HFPO, reaction time 3 h.
HFPO yield is shown to undergo decreases over the entire
examined range of molar ratios.
It results from a combined effect of the molar ratio HFP/O2
and temperature, that after the reaction time of 3 h (Figure 4)
the highest yields of HFPO are achieved in the range of
temperature 160-170 °C.
A combined effect of the molar ratio HFP/O2 and the
reaction time (Figure 3) at the temperature of 165 °C indicates
that the highest yield can be achieved during the changes of
the molar ratio in the range of 1.6-2.6, after reaction time of
3.1-4.5 h.
The temperature is here an essential technological parameter;
its lowering or increasing by 5 °C beyond the optimal range
causes a decrease of HFPO yield of the order of a few or a
dozen percentage points.
A decrease of the molar ratio HFP/O2 below 1.6, while the
reaction times were shorter than 3 h, causes a significant
decrease of HFPO yield. For the reaction times below 3 h the
Conclusions
The performed studies on the influence of temperature, the
molar ratio of HFP/O2, and the reaction time showed that the
276
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

HFPO yield achieves the highest value of 43 mol % at
temperature of 165 °C, at the molar ratio of HFP/O2 ) 2.2,
after the reaction time of 3.1 h, in the medium of carbon
tetrachloride as a solvent. The same value of the yield is
obtained when the molar ratio of HFP/O2 is lowered to a value
of 1.6, while the remaining parameters maintain their values.
The highest yield of HFPO can be also achieved by lowering
the temperature to 160 °C. However, this requires maintain-
ing the HFP/O2 molar ratio of about 2.2 mol/mol and the
reaction time 3 h. The presented method enables to run the
process on a small scale. In the case of technical scale a
continuous dosage of oxygen is required. This will prevent
rapidly growing pressure and temperature and exclude the
possibility of explosion.
Received for review June 15, 2009.
OP900151P
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
277