A study of the thermal decarboxylation of three perfluoropolyether salts

Article abstract of DOI:10.1016/S0022-1139(03)00195-7

The thermal decarboxylation of three dicarboxylic perfluoropolyether potassium salts of relatively short chain length has been investigated and the products and kinetics of the main reactions have been defined. From the rate constants and Arrhenius parameters data, the second decarboxylation appears to be quantitatively rather close to the first.

Full text of DOI:10.1016/S0022-1139(03)00195-7

Journal of Fluorine Chemistry 124 (2003) 123–130
A study of the thermal decarboxylation of three
perfluoropolyether salts
G. Marchionnia,*, S. Petriccia, G. Spataroa, G. Pezzinb
aSolvay Solexis R&T, viale Lombardia 20, 20021 Bollate, Milan, Italy
bDepartment of Physical Chemistry, University of Padova, Padova, Italy
Received 22 April 2003; received in revised form 11 July 2003; accepted 13 July 2003
Abstract
The thermal decarboxylation of three dicarboxylic perfluoropolyether potassium salts of relatively short chain length has been investigated
and the products and kinetics of the main reactions have been defined. From the rate constants and Arrhenius parameters data, the second
decarboxylation appears to be quantitatively rather close to the first.
# 2003 Elsevier B.V. All rights reserved.
Keywords: a,o-Dihydroperfluoropolyethers; Perfluoropolyether potassium salts; Thermal decarboxylation; Kinetic parameters
1. Introduction
They are models for more complex salts of higher mole-
cular weight (e.g. the H-Galden1 ZT) [1–4], whose reac-
tions are:
The thermal decarboxylation of perfluoropolyether potas-
sium dicarboxylic salts having the structure:
KOOCCF2ðOCF2ÞnðOCF2CF2ÞmOCF2COOK þ H2O
! HCF2ðOCF2ÞnðOCF2CF2ÞmOCF2H þ 2KHCO3 (2)
KOOCCF2ꢀRfꢀCF2COOK ðCCÞ
where Rf is a short moiety, when carried out in the presence
of an excess of a proton donor, such as water, proceeds
through a decarboxylation reaction that leads to the mono-
carboxylate compound (HC), and a second reaction that
gives the final a,o-hydrogenated product (HH). Schema-
tically:
Here n and m are either zero or integers. The aim of the
present work is to study the kinetics of the main thermal
reactions, to obtain relevant kinetic parameters, and to
clarify the reactions mechanisms. The literature on similar
reactions appears to be relatively scarce [5–10].
k1
2. Results and discussion
KOOCCF2RfCF2COOK ! HCF2RfCF2COOK
H2O
ðCCÞ
ðHCÞ
k2
2.1. Decarboxylation of KOOCCF2OCF2COOK (CC1)
þ KHCO3 ! HCF2RfCF2H þ KHCO3
(1)
H2O
ðHHÞ
The thermal reactions of CC1 have been investigated at
four temperatures: 140, 160, 180 and 200 8C, by determining
the chemical composition of both the liquid and gaseous
phases present in the reaction vessel at a number of selected
times (for clarity, the analytical data reported in the present
paper, refer to only one time for each temperature). The most
significant results are the final degree of conversion at the
time selected, (with respect to the initial quantity of dicar-
boxylic salt CC1), and the quantity of a,o-hydrogenated
product (HH1), both results being expressed as mol%
(Table 1). Percentage errors (E%) relative to the material
balance of the carbon and fluorine content, calculated with
In the present paper, the kinetics and the mechanisms of
the thermal decarboxylations of three perfluoropolyether
potassium salts of relatively high purity (>95% molar) are
described, their formulae are: KOOCCF2OCF2COOK
(Rf ¼ ꢀOꢀ) CC1, KOOCCF2OCF2OCF2COOK (Rf ¼
ꢀOCF2O–) CC2 and KOOCCF2OCF2CF2OCF2COOK
(Rf ¼ ꢀOCF2CF2O–) CC3.
* Corresponding author. Tel.: þ39-02-3835-6289;
fax: þ39-02-3835-2152.
E-mail address: giuseppe.marchionni@solvay.com (G. Marchionni).
0022-1139/$ – see front matter # 2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0022-1139(03)00195-7

124
G. Marchionni et al. / Journal of Fluorine Chemistry 124 (2003) 123–130
Table 1
Degree of conversion of the dicarboxylic salt KOOCCF2OCF2COOK (CC1), yield of HCF2OCF2H (HH1), and material balance of carbon and fluorine at the
temperature and time indicated
Run
T (8C)
Time (min)
CC1 conversion (%)
HH1 yield (%)
Carbon (E%)
Fluorine (E%)
1
2
3
4
140
160
180
200
2900
1450
70
85.0
99.8
87.2
94.9
7.70
34.7
21.6
38.9
ꢀ3.3
ꢀ6.1
ꢀ2.7
ꢀ5.8
ꢀ3.2
ꢀ7.3
ꢀ2.7
ꢀ4.5
20
Table 2
Analytical data, expressed as concentration of the species (mol lꢀ1), recovered after decarboxylation reactions
Run CC1
HC1
HH1
CC2
HC2
HH2
CC3
HC3
HH3
(COOK)2 HCOOK Fꢀ
CHF3 CO
CO2
4
9
0.017 0.062 0.129
–
–
–
–
–
–
–
–
–
0.017
0.010
0.047
0.080
0.044
0.374
–
0.124 0.252
–
–
–
0.001 0.004 0.170
0.389 0.030 0.004 0.454
0.203 0.018 0.007 0.277
13
0.020 0.026 0.006
–
–
–
0.001 0.003 0.148 0.002
The initial amounts of the three bicarboxylic salts are respectively: CC1 0.332 mol lꢀ1; CC2 0.266 mol lꢀ1 and CC3 0.249 mol lꢀ1
.
respect to the stoichiometry of the processes, are also
reported, in Table 1. As an example, the concentrations of
all the species recovered after the decarboxylation reaction
relevant to run 4 are collected in Table 2.
The anionic intermediates CF2OCF2COOK and
ꢀCF2OCF2H are present in these reactions, in which degra-
dation of intermediate products also plays a role, as demon-
strated by the presence of potassium oxalates and formates
(the latter however has been experimentally monitored only
at 200 8C). CHF3 is absent in the reaction products; even
a decarboxylation reaction carried out at 200 8C in the
presence of excess fluoride ions (KF) did not result in the
production of CHF3.
From Table 1 it appears that the a,o-hydrogenated pro-
ducts yields are not very high, even when the reaction
proceeds to high degrees of conversion, and that only when
the temperature is high are the a,o-hydrogenated products
relatively abundant.
The errors in the carbon and fluorine material balance
remain almost constant (and relatively small) with increas-
ing decarboxylation temperature, a result that can be inter-
preted as an indication that the material balance is
substantially correct. At the times selected for the tests,
an increase of temperature leads to an increase of both the
final degree of conversion of CC1 and of yield of HH1, the
latter reaching moderately high values at temperatures in the
range 160–200 8C, the selected reaction time being as low as
20 min at the highest temperature explored. All the possible
salt decarboxylation reactions are given in Scheme 1, in
which the analytical results are summarized.
The above results lead to the tentative conclusion that
in a basic medium HCF2OCF2H undergoes degradation
k1
KOOCCF2OCF2COOK + H2O
→
HCF2OCF2COOK + KHCO3
(HC1)
(CC1)
k2
HC1 + H2O
→
HCF2OCF2H + KHCO3
(HH1)
H2O
H2O
→ [K+ -OCF2COOK] → [FOCCOOK] → HOOCCOOK
- :CF2 - KF - HF
CC1
HC1
→ [K+ -CF2OCF2COOK]
- CO2
H2O
→ [HCF2OCF2- +K]
H2O
→ [HCF2O- +K] → [HCOF] → HCOOH
- CO2
- :CF2
K2CO3 + CO2 + H2O
CO + 2 HF
- KF
- HF
2 KHCO3
↔
→
→
:CF2 + H2O
HF + K2CO3
KF + KHCO3
KHCO3
HCF2OCF2H + H2O
→
2 CO + 4 HF
Scheme 1. Reaction mechanism for the thermal decarboxylation of salt CC1.

G. Marchionni et al. / Journal of Fluorine Chemistry 124 (2003) 123–130
125
Table 3
to CO and HF:
The kinetic constant of degradation of salt KOOCCF2OCF2COOK (CC1)
at different temperature
HCF2OCF2H þ H2Ob!ase2CO þ 4HF
a reaction apparently related to the acidity of the hydrogen
atoms of the HH1 product.
To verify the above hypothesis, the thermal stability of the
HH1 molecule has been tested separately under reaction
conditions. In particular, when 5 ml (0.22 mol) of
HCF2OCF2H were heated at 160 8C for 5 h with 0.5%
aqueous basic solutions of KHCO3, fluorides were formed
in the aqueous phase and CO and CO2 in the gas phase, 50%
of the reagent HCF2OCF2H being consumed.
Under neutral conditions the reagent HH1 is stable, as
shown by the fact that after treatment for 20 min at 200 8C
in the presence of a 10% (w/w) KF solution, it is completely
recovered.
The low yields of a,o-hydrogenated products suggest that
the reactions are rather complex. To obtain kinetic para-
meters a simplified reaction scheme cannot be used and only
the kinetic constant of the CC1 consumption can be obtained
as a function of temperature.
(3)
Temperature (8C)
ks (sꢀ1
)
140
160
180
200
1.1 Â 10ꢀ5
7.2 Â 10ꢀ5
5.1 Â 10ꢀ4
2.2 Â 10ꢀ3
times, in both the aqueous and gas phases. The final amounts
of reagent and products are summarized in Table 4 and the
reaction mechanisms are reported in Scheme 2. As example,
the concentrations of all the species recovered after dec-
arboxylation reaction relevant to run 9, are collected in
Table 2. When the conversion of the dicarboxylic salt is
high, the yield of the a,o-hydrogenated product is also high,
and the concentrations of oxalate, formate, CO and CHF3,
produced by parallel reactions are small.
In the limit of zero production of the latter compounds, i.e.
when parallel reactions are absent, the data can be treated
using a simplified kinetic model of consecutive first-order
reactions, which allows evaluation of the kinetic constants at
various temperatures.
The reaction appears to be first order from Fig. 1, where
the CC1 concentration is plotted on a log scale against time.
The linear relationships between ks at each temperature,
reported in Table 3, and 1/T (Figure not shown), give the
For the general reactions:
Arrhenius parameters of the reaction as Ea ¼ 3 (kcal molꢀ1
)
KOOCCF2ORfOCF2COOK
ðCCÞ
and A ¼ 3:6 Â 1012 sꢀ1
.
k1
k2
!HCF2ORfOCF2COOK!HCF2ORfOCF2H
2.2. Decarboxylation of KOOCCF2OCF2OCF2COOK
(CC2)
ðHCÞ
ðHHÞ
the following kinetic equations can therefore be applied:
d½CCꢁ
The reactions have been carried out at 140, 150, 160, 170
and 180 8C, the compositions being monitored, at selected
¼ ꢀk1½CCꢁ
(4)
dt
0
-2
-4
-6
-8
0
20000
40000
60000
80000
100000
time (s)
Fig. 1. KOOCCF2OCF2COOK (CC1) salt concentration vs. time at 140 8C (&), 160 8C (~), 180 8C (Â) and 200 8C (*).

126
G. Marchionni et al. / Journal of Fluorine Chemistry 124 (2003) 123–130
Table 4
Table 5
Degree of conversion of the dicarboxylic salt KOOCCF2OCF2OCF2COOK
(CC2), yield of HCF2OCF2OCF2H (HH2), and material balance of carbon
and fluorine
Constant of first and second decarboxylation reaction of salt KOOCC-
F2OCF2OCF2COOK (CC2) at various temperatures
Temperature (8C)
k1 (sꢀ1
)
k2 (sꢀ1
)
Run
T
Time
CC2
HH2
Carbon Fluorine
140
150
160
170
180
7.5 Â 10ꢀ5
2.5 Â 10ꢀ4
8.5 Â 10ꢀ4
1.9 Â 10ꢀ3
5.9 Â 10ꢀ3
1.0 Â 10ꢀ4
5.1 Â 10ꢀ4
1.3 Â 10ꢀ3
1.9 Â 10ꢀ3
5.7 Â 10ꢀ3
(8C) (min)
conversion (%) yield (%) (E%)
(E%)
5
6
7
8
9
140
150
160
170
180
980
160
130
60
99.0
91.7
98.9
99.8
99.5
65.3
59.2
67.3
67.4
63.7
ꢀ7.2
ꢀ3.5
ꢀ3.0
ꢀ6.8
ꢀ7.0
ꢀ5.9
ꢀ4.8
ꢀ3.3
ꢀ4.8
ꢀ2.1
40
By applying Eq. (7) to the experimental data the first
decarboxylation reaction constant kl is obtained. Succes-
sively a procedure is applied that optimizes the k2 value by
minimizing the sum of the quadratic differences between the
HC2 and HH2 values obtained from Eqs. (9) and (10) and the
experimental values. The results are collected in Table 5.
The experimental CC2, HC2 and HH2 concentrations,
plotted as a function of time, and those calculated, derived
from Eqs. (7), (9) and (10), in which the optimized values of
the k1 and k2 are inserted, compare satisfactorily for the CC2
and HC2 values, but not for HH2: this result is due to the fact
that the simplified model used does not take into account
parallel degradation reactions. As example, in Fig. 2 calcu-
lated and experimental kinetic data relatively to run 7 are
reported.
d½HCꢁ
dt
¼ k1½CCꢁ ꢀ k2½HCꢁ
¼ k2½HCꢁ
(5)
(6)
d½HHꢁ
dt
The concentration of CC follows a first-order decay equa-
tion:
½CCꢁ ¼ ½CCꢁ0eꢀk t
(where [CC]0 is the initial CC concentration).
When, at time zero, the concentration of HC and HH is
zero, the differential equation becomes:
(7)
1
d½HCꢁ
¼ k1½CCꢁ0ðeꢀk tÞ ꢀ k2½HCꢁ
(8)
1
dt
From a suitable plot, (Fig. 3) the Arrhenius parameters Ea
and A have been obtained.
For the first decarboxylation reaction:
and therefore:
k1½CCꢁ0
½HCꢁ ¼
ðeꢀk1t ꢀ eꢀk t
Þ
(9)
k2 ꢀ k1
From the material balance one obtains, in the absence of
2
KOOCCF2OCF2OCF2COOK
ðCC2Þ
k1
parallel reactions:
ꢀ
þ H2O!HCF2OCF2OCF2COOK þ KHCO3
(11)
ꢁ
ðHC2Þ
k2eꢀk t ꢀ k1eꢀk t
k2 ꢀ k1
1
2
½HHꢁ ¼ ½CCꢁ0 1 ꢀ
(10)
Eað1Þ ¼ 40 (kcal molꢀ1); Að1Þ ¼ 1:1 Â 1017 sꢀ1
.
k1
KOOCCF2OCF2OCF2COOK + H2O
→
HCF2OCF2OCF2COOK + KHCO3
(CC2)
(HC2)
k2
HC2 + H2O
→ HCF2OCF2OCF2H + KHCO3
(HH2)
H2O
2 H2O
→ HOOCCOOK + 3 HF + CO2
CC2
HC2
→ [K+ -CF2OCF2OCF2COOK]
→ [K+ -OCF2OCF2COOK]
- :CF2
→ [FOCOCF2COOK]
-KF
- CO2
H2O
→ [HCF2OCF2OCF2- +K]
- CO2
2 H2O
→ HCOOH + 3HF + CO2
→ [HCF2OCF2O- +K]
- :CF2
→ [HCF2OCOF]
- KF
2 KHCO3
↔
K2CO3 + CO2 + H2O
CO + 2 HF
:CF2 + H2O
:CF2 + HF
HF + K2CO3
→
→
→
CHF3
KF + KHCO3
Scheme 2. Reaction mechanism for the thermal decarboxylation of salt CC2.

G. Marchionni et al. / Journal of Fluorine Chemistry 124 (2003) 123–130
127
0,3
0,2
0,1
0,0
0
2000
4000
6000
8000
10000
time (s)
Fig. 2. Decarboxylation of CC2 salt at 160 8C (run 7): experimental data relating to CC2 (^), HC2 (Â) and HH2 (~) species, in comparison with calculated
trends (continuous lines). See text for discussion of the above data.
-4,0
2.3. Decarboxylation of KOOCCF2OCF2CF2OCF2COOK
(CC3)
The reactions have been carried out at 140, 150, 160 and
-6,0
170 8C, the compositions being monitored, at selected time
intervals, in both the aqueous and gas phases.
In Table 6, the relevant data are collected, for each tem-
perature, as well as the material balance of carbon and
fluorine, expressed as percentage errors E%. As an example,
the concentrations of all the species recovered after decarbox-
-8,0
ylation reaction relevant to run 13, are collected in Table 2.
The material balances of fluorine and carbon are affected
by errors larger than those of CC1 and CC2. In the tem-
perature range explored, the a,o-hydrogenated products
yields are close to 60%. The reaction mechanisms are shown
in Scheme 3; it is interesting to point out the presence of the
CC1, HC1 and HH1 species, whose formation has been
described.
-10,0
2,1
2,2
2,3
2,4
2,5
103/T (K-1)
Fig. 3. Arrhenius plot of kinetic constant k1 for the thermal decarboxyla-
tion of salt CC2.
For the second decarboxylation reaction:
When the simplified kinetic model of consecutive reac-
tions is applied to these data, using Eqs. (7), (9) and (10), and
the results are optimized as described previously, one
obtains the kinetic constants collected in Table 7.
As already reported for CC2, the experimental CC3, HC3
and HH3 concentrations, plotted as a function of time, and
HCF2OCF2OCF2COOK
ðHC2Þ
k2
þ H2O!HCF2OCF2OCF2H þ KHCO3
(12)
ðHH2Þ
Eað2Þ ¼ 35 (kcal molꢀ1); Að2Þ ¼ 3:8 Â 1014 sꢀ1
.
Table 6
Degree of conversion of the dicarboxylic salt KOOCCF2OCF2CF2OCF2COOK (CC3), yield of HCF2OCF2CF2OCF2H (HH3), and material balance of carbon
and fluorine
Run
T (8C)
Time (min)
CC3 conversion (%)
HH3 yield (%)
Carbon (E%)
Fluorine (E%)
10
11
12
13
140
150
160
170
500
120
100
40
98.0
95.6
99.8
99.7
55.7
52.7
60.4
59.6
ꢀ6.9
ꢀ4.7
ꢀ7.5
ꢀ6.1
ꢀ15.1
ꢀ9.4
ꢀ11.0
ꢀ11.8

128
G. Marchionni et al. / Journal of Fluorine Chemistry 124 (2003) 123–130
k1
KOOCCF2OCF2CF2OCF2COOK
+
H2O
→
HCF2OCF2CF2OCF2COOK
+
KHCO3
(CC3)
(HC3)
k2
HC3 + H2O
→
HCF2OCF2CF2OCF2H
+
KHCO3
(HH3)
H2O
→
- CO2
CC3
[K+ -CF2OCF2CF2OCF2COOK]
→ [K+ -OCF2CF2OCF2COOK]
- :CF2
→ [FOCCF2OCF2COOK]
- KF
H2O
→
- HF
K2CO3
→
- KHCO3
HOOCCF2OCF2COOK
H2O
KOOCCF2OCF2COOK
(CC1)
HC3
→
[HCF2OCF2CF2OCF2- +K ]
→
[HCF2OCF2CF2O- +K]
→ [HCF2OCF2COF]
- KF
- CO2
- :CF2
H2O
→
K2CO3
→
HCF2OCF2COOH
HCF2OCF2COOK
-HF
- KHCO3
(HC1)
2 KHCO3
↔
→
→
→
K2CO3 + CO2 + H2O
CO + 2 HF
CHF3
:CF2 + H2O
:CF2 + HF
HF + K2CO3
KF + KHCO3
Scheme 3. Reaction mechanism for the thermal decarboxylation of salt CC3.
Table 7
Constant of first and second decarboxylation reaction of KOOCC-
F2OCF2CF2OCF2COOK (CC3) at various temperatures
those calculated, derived from Eqs. (7), (9) and (10) (in
which the optimized values of the kl and k2 are inserted),
compare satisfactorily for CC3 and HC3, but not for HH3, a
result which is due to the fact that the simplified model used
does not take into account parallel degradation reactions.
From an appropriate plot of the above data (Fig. 4), the
Arrhenius parameters Ea and A have been obtained.
For the first decarboxylation reaction:
Temperature (8C)
k1 (sꢀ1
)
k2 (sꢀ1
)
140
150
160
170
1.3 Â 10ꢀ4
4.3 Â 10ꢀ4
1.1 Â 10ꢀ3
3.1 Â 10ꢀ3
1.0 Â 10ꢀ4
3.6 Â 10ꢀ4
9.5 Â 10ꢀ4
1.8 Â 10ꢀ3
KOOCCF2OCF2CF2OCF2COOK
ðCC3Þ
k1
þ H2O!HCF2OCF2CF2OCF2COOK þ KHCO3 (13)
ðHC3Þ
-4,0
-6,0
-8,0
Eað1Þ ¼ 38 (kcal molꢀ1); Að1Þ ¼ 2:1 Â 1016 sꢀ1
.
For the second decarboxylation reaction:
HCF2OCF2CF2OCF2COOK
ðHC3Þ
k2
þ H2O!HCF2OCF2CF2OCF2H þ KHCO3
(14)
ðHH3Þ
Eað2Þ ¼ 35 (kcal molꢀ1); Að2Þ ¼ 4:7 Â 1014 sꢀ1
.
-10,0
2,2
3. Conclusions
2,3
2,4
2,5
103/T (K-1)
The present paper summarizes the main features of the
thermal decarboxylation reactions of representative dicar-
boxylic potassium salts having rather short perfluorinated
Fig. 4. Arrhenius plot of kinetic constant k2 for the thermal decarboxyla-
tion of salt CC3.

G. Marchionni et al. / Journal of Fluorine Chemistry 124 (2003) 123–130
129
Table 8
chains, that represent a simplified model of longer species,
precursor in the production of the H-Galden1 ZT com-
pounds. Mechanisms are proposed for the principal two-step
ionic decarboxylations, as well as for some secondary
degradation processes taking place.
A parallel process has been found to give, in the CC2 and
CC3 reactions, the intermediate species difluorocarbene
:CF2, together with oxalate and formate and the short chain
CC1 compound, respectively. The overall reactions leading
to the final a,o-hydrogenated products are characterized, for
the CC2 and CC3 salts, by yields close to 60–65%, which do
not appear to be very temperature dependent in the range
explored.
A peculiar consecutive reaction characterizes the degra-
dation of the short chain CC1 salt; the expected decarbox-
ylation product HH1 is unstable under the reaction
conditions and degrades further to CO and fluorides.
For the CC1 compounds the temperature dependence of
the kinetic constant has been obtained for the salt consump-
tion. For the CC2 and CC3 salts the kinetic model of first-
order consecutive reactions was used, from which both the
kinetic constants of the first and the second decarboxylation
as well as the relative Arrhenius parameters have been
derived. Small differences in the reaction rates are present,
at a given temperature, for CC2 and CC3; moreover, the
kinetic constants of the first and second decarboxylation are
rather close, particularly for the CC3 compound. Even when
two carboxylic groups are as close as in the short molecules
investigated here, there is little mutual influence, and there-
fore any influence will presumably be absent in compounds
having longer chains.
Quantitative analysis of dicarboxylic potassium salts
Product
MW
Molar (%)
KOOCCF2OCF2COOK (CC1)
282
348
398
99.0
96.3
99.4
KOOCCF2OCF2OCF2COOK (CC2)
KOOCCF2OCF2CF2OCF2COOK (CC3)
with ethanol. Details of the processes are described else-
where [11]. Pure salts have been prepared from a selective
fractionation of diethyl esters followed by an aqueous
hydrolysis with quantitative amounts of KOH, a suitable
distillation of the ethanol–water azeotrope, and a final
dehydratation in an oven at 90 8C. In Table 8, molecular
weight and quantitative analysis of the dicarboxylic salts are
reported. In order to determinate the purity of the above
compounds, 19F NMR quantitative analysis on aqueous
solutions containing 10% by weight of salt, were performed.
4.2. Experimental apparatus of thermal decarboxylation
Due to the impossibility of extracting directly, at selected
times, successive portions of the reaction mixture from the
decarboxylation vessel, in order to follow the reaction
kinetics, suitable parallel batch experiments were carried
out on 10 g solutions, each containing 10% by weight of salt,
and the relative aqueous and gas phases were analysed. For
each decarboxylation test 10 steel AISI 316 cylindrical
20 ml flask were used. They were thermostatted in a bath
oil at the selected temperature. Rapid immersion, the extrac-
tion and the forced cooling of the steel vessels was provided
by a suitable mechanical apparatus.
4. Experimental
4.3. Analytical section
4.1. Preparation of dicarboxylic potassium salts
In order to extract completely the gaseous products in the
vapor phase, the decarboxylation vessels were heated at
50 8C and then connected to a volumetric flask under
vacuum. The gas phases were then analysed by gas chro-
matography using a gas chromatograph, type CE 8000 Top
The a,o-diethylic esters CH3CH2OOCCF2–Rf–CF2COO-
CH2CH3 were obtained from oxidative polymerization of
tetrafluoroethylene with oxygen using UV irradiation, fol-
lowed by catalytic reduction and successive esterification
Table 9
Chemical shifts and coupling constants for 19F NMR signals of carboxylic salts
Product
Chemical shift (ppm)a
Coupling const. (Hz)
1
2
3
4
1–2
2–3
3–4
F–H
1
KOOCCF2OCF2COOK ðCC1Þ
ꢀ78.0 (s)
ꢀ78.9 (t)
ꢀ52.2 (quint)
ꢀ79.9 (t)
ꢀ78.4 (t)
ꢀ78.5 (t)
1
2
KOOCCF2OCF2H ðHC1Þ
ꢀ85.4 (dt)
ꢀ79.9 (t)
ꢀ54.3 (tt)
ꢀ89.4 (t)
ꢀ89.3 (t)
4.9
9.8
69.7
69.7
69.7
1
2
KOOCCF2OCF2OCF2COOK ðCC2Þ
1
2
3
KOOCCF2OCF2OCF2H ðHC2Þ
ꢀ86.9 (dt)
ꢀ89.9 (t)
9.8
4.9
1
2
KOOCCF2OCF2CF2OCF2COOK ðCC3Þ
11.4
11.4
1
2
3
4
KOOCCF2OCF2CF2OCF2H ðHC3Þ
ꢀ86.0 (dt)
4.9
a With reference to CFCl3.

130
G. Marchionni et al. / Journal of Fluorine Chemistry 124 (2003) 123–130
with TCD detector. In order to analyse the hydroderivates
HH1, HH2 and HH3 a copper packed column was used
(diameter 4 mm, length 2 m) filled with Fomblin1 M10 in
ratio 20/100 on Chromosorb P 60–80 mesh. In the analysis
for CO2, CO and CHF3 a steel packed column was used
(diameter 1/8 in., length 3 m), filled with 100/120 Carbo-
sieve S-II (Supelco). Quantitative analysis were performed
using suitable gas chromatographic response factors, pre-
viously determined analyzing the pure compounds.
Acknowledgements
The authors wish to thank Dr. E. Barchiesi and Dr. G.
Geniram for their help in the analytical characterizations and
Solvay Solexis for permission to publish this work.
References
Ion chromatography was used for the analysis of oxalates,
formates, fluorides and mono- and dicarboxylic fluorinated
compounds in the aqueous phases. All solutions, diluted by a
factor of about 100, were analysed using a IONPAC AS11
column with a gradient of NaOH (from 5 to 100 mM). In the
analysis of fluorides, oxalates, formates and dicarboxylic
salts, the calibration was performed using aqueous solutions
containing known amounts of pure compounds; decarboxy-
lated solutions previously analysed by 19F NMR were used
in order to determine the amounts of monocarboxylic salts.
19F NMR spectroscopy was used for quantitative analysis
of carboxylic salts: the spectra of the aqueous solutions
were recorded using a Varian Inova 400 spectrometer, at
376 MHz. The quantitative analysis were performed using
an internal capillary tube containing a known amount of a
fluorinated standard. In Table 9, all chemical shifts and
coupling constants for the mono- and dicarboxylic salts
examined are collected.
[1] M. Visca, R. Silvani, G. Marchionni, Chemtech 27 (2) (1997) 33–37.
[2] G. Marchionni, G. Spataro, E. Strepparola, US Patent 5,969,192
(1999).
[3] G. Marchionni, R. Silvani, G. Fontana, G. Malinverno, M. Visca, J.
Fluorine Chem. 95 (1999) 41–50.
[4] G. Marchionni, M. Bassi, G. Fontana, P. Maccone, G. Ajroldi, J.
Fluorine Chem. 98 (1999) 41–54.
[5] J.D. LaZerte, L.J. Hals, T.S. Reid, G.H. Smith, J. Am. Chem. Soc. 75
(1953) 4525–4528.
[6] R.E. Banks, E.D. Burling, B.A. Dodd, K. Mullen, J. Chem. Soc. (C)
(1969) 1706–1709.
[7] G.P. Gambaretto, G. Troilo, M. Napoli, C. Fraccaro, La Chimica e
l’Industria 53 (1971) 1033–1038.
[8] V.A. Gubanov, I.A. Zevakin, V.E. Gurari, A.V. Tumanova, I.M.
Dolgopol’skii, R.M. Ryazanova, D.S. Rondarev, Zh. Org. Khim. 17
(8) (1981) 1587–1590.
[9] P.R. Resnick, A.C. Sievert, US Patent 5,410,095 (1995).
[10] J.L. Howell, M.A. Hofmann, A. Waterfeld, A.M. Sipyagin, C.M.
Friesen, J.S. Thrasher, J. Fluorine Chem. 89 (1998) 131–135.
[11] C. Tonelli, P. Gavezotti, E. Strepparola, J. Fluorine Chem. 95 (1999)
51–70.