Photolysis of perfluoroacyl fluorides

Article abstract of DOI:10.1016/S0022-1139(99)00211-0

The photochemical decarbonylation and coupling of acyl fluorides containing perfluoroalkyl and perfluoro(oxa)alkyl chains have been re-examined. The reaction on either single acyl fluorides or on their binary mixtures has allowed the formation and isolation of macromolecular perfluoro(oxa)alkanes having interesting physical-chemical properties. These perfluoropolyethers are also useful as model compounds and standards to better characterize more complex mixtures of similar commercial products.

Full text of DOI:10.1016/S0022-1139(99)00211-0

Journal of Fluorine Chemistry 101 (2000) 117±123
Photolysis of per¯uoroacyl ¯uorides
Claudio Tonelli, Vito Tortelli*
Ausimont CRS, via S. Pietro 50, 20021 Bollate Milano, Italy
Received 21 June 1999; accepted 2 September 1999
Abstract
The photochemical decarbonylation and coupling of acyl ¯uorides containing per¯uoroalkyl and per¯uoro(oxa)alkyl chains have been re-
examined. The reaction on either single acyl ¯uorides or on their binary mixtures has allowed the formation and isolation of
macromolecular per¯uoro(oxa)alkanes having interesting physical±chemical properties. These per¯uoropolyethers are also useful as model
compounds and standards to better characterize more complex mixtures of similar commercial products. # 2000 Elsevier Science S.A. All
rights reserved.
Keywords: Photocoupling; Per¯uoroacyl ¯uoride; Per¯uoropolyether
1. Introduction
CF3COF giving a new primary radical which evolves to
stable structures by coupling reactions. The latter are formed
by the radical addition of the just formed secondary per-
¯uoroalkyl radicals to the carbonyl of a second molecule of
acyl ¯uoride. Similar structures have been obtained also
from thermal decomposition of bis-per¯uoroacyl peroxide
[6]:
The photochemical decarbonylation and coupling of
¯uorinated acyl ¯uorides has been reported previously
[1,2]: the reaction works on acyl ¯uorides with per¯uor-
oalkyl chains as well as on per¯uoropolyether acyl ¯uorides
where the chain has been formed by ring-opening polymer-
ization of C2F4 and C3F6 epoxides.
D
ꢁRfCOO†2 ! Rf Rf ‡ 2CO2
hꢀ
2Rf COF ! Rf Rf ‡ COF2 ‡ CO
where Rf is perfluoroalkyl, perfluoro(oxa)alkyl.
Such an interesting synthetic pathway is not easily applic-
able to oligomeric structures and shows an overall low
selectivity.
Linear and crosslinked polymers have been obtained
starting from bifunctional [3] and multifunctional [4] struc-
tures, respectively. Generally, the photocoupling reaction
shows high selectivity even if the productivity is strongly
limited by the poor ef®ciency of the UVactivation of the acyl
¯uoride chromophore. This is basically due to the small
fraction of useful radiation to promote the ±COF excited
state: high pressure mercury lamps have high intensity, but
over a wide spectrum of frequencies while low pressure
lamps have low intensity but at a selected frequency.
The kinetic aspects and the reaction mechanism have been
closely examined in some previous work [1,5] that has
pointed out the presence of byproducts having both lower
and higher molecular weights. The former come from a side-
reaction of the intermediate per¯uoroalkyl radical ±
CF2OCF(CF3)ꢀ , generated by photolysis of the starting acyl
¯uoride, which eliminates, at high temperature (>1008C),
Another way to synthesize similar structures is the
electrochemical Kolbe decarboxylation and dimerization
[7,8]:
2RfCOOH ! Rf Rf ‡ 2CO2 ‡ H2
The aim of the present work is the study of the photo-
chemical reactivity of the ±COF group in highly ¯uorinated
molecules to obtain new oligomeric species, thermally and
chemically stable and to get information about a favourable
post treatment process to eliminate acidic end-groups from
commercial per¯uoropolyether oils. It has been interesting
to evaluate the reactivity in the photocoupling reaction of
acyl ¯uorides having different ¯uorinated chains (per¯uor-
oalkyl, per¯uoropolyether from hexa¯uoropropeneoxide
HFPO), examining closely some aspects of the above men-
tioned studies and extending the photochemical reaction to
acyl ¯uorides having a per¯uoropolyether chain derived
from the photo-oxidation of hexa¯uoropropene [9].
*Corresponding author.
0022-1139/00/$ ± see front matter # 2000 Elsevier Science S.A. All rights reserved.
PII: S 0 0 2 2 - 1 1 3 9 ( 9 9 ) 0 0 2 1 1 - 0

118
C. Tonelli, V. Tortelli / Journal of Fluorine Chemistry 101 (2000) 117±123
Last, but not least, the reaction products, both symmetric
0
(Rf ±Rf) and asymmetric (Rf±Rf), if a two component
2.1. Photocoupling of HFPO oligomers
mixture was irradiated, have been isolated and characterized
by spectroscopic and physical±chemical methods. The data
obtained on these model compounds have been useful for
better understanding the properties of similar but more
complex mixtures of industrial relevance.
2.1.1. Preparation of perfluoro-5,6-dimethyl-4,7-dioxa-
decane (PFPE22)
In the photochemical unit (reactor of 0.1 l, optical path
0.7 cm, external distillation apparatus of 0.5 l capacity),
400 g (1.19 mol) of per¯uoro-2-methyl-3-oxa-hexanoyl
¯uoride (HFPO dimer, 2a) were charged. The temperature
during irradiation was maintained at 208C in the reactor and
708C in the distillation pot to allow a constant recycle of the
acyl ¯uoride (Method B). After 50 h GLC analysis showed
that 20% of a new high-boiling product and 80% of starting
reagent were present in the reactor while in the distillation
pot the product was 82%. The whole solution (360 g)
was collected and fractionated to give 197 g of coupled
product and 15 g of higher molecular weight residue. 19F
NMR analysis con®rms that the main product was
[CF3CF2CF2OCF(CF3)±]2, yield: 69.0%.
2. Experimental
The acyl ¯uorides used in the present work are:
CF3O‰CF2CFꢁCF3†OŠmCF2COF; m ˆ 2:9 ꢁ1†
C3F7O‰CFꢁCF3†CF2OŠnCFꢁCF3†COF;
n ˆ 0; 1; 2 ꢁ2a; b; c†
CF3ꢁCF2†6COF ꢁ3†
They were synthesized following literature methods:
by careful fractional distillation of the mixture derived
from photooxidation of hexafluoropropene, after thermal
treatment 1 [9], by anionic oligomerization of HFPO 2a,b,c
[10] and by fluorination of perfluorooctanoic acid 3
[11].
2.1.2. Preparation of perfluoro-5,8,9,12-tetramethyl-
4,7,10,13-tetraoxa-hexadecane (PFPE23)
In a photochemical unit of 0.3 l (Method A) were charged
430 g (0.86 mol) of per¯uoro-2,5-dimethyl-3,6-dioxa-non-
anoyl ¯uoride (HFPO trimer, 2b). After 50 h of irradiation,
402 g of crude mixture were obtained. From GLC, the
composition is: 48% starting, 46% main product, 2% of
lower molecular weight product and 4% of higher molecular
weight product. After work-up as before, 205 g of 99.3%
pure (by GLC) coupling product were obtained whose 19F
NMR is in agreement with the structure [CF3CF2CF2-
OCF(CF3)CF2OCF(CF3)±]2, yield: 55%.
The photocoupling reactions were performed in a photo-
chemical unit which includes an immersion high-pressure
Hanau mercury lamp (150 W), quartz jacketed and cooled
by distilled water or CF2ClCFCl2, and an outer Pyrex
cylindrical reactor. This reactor was provided with magnetic
stirring, cooling system (dry-ice/isopropanol), thermocou-
ple and inlet tube for withdrawing samples (Method A).
Sometimes the reaction was performed by recycling the
volatile acyl ¯uorides after distillation of a part of the
reacting mixture (Method B). The latter experiments
increased the productivity of the reaction, photolyzing only
the pure starting materials whereas the higher-boiling reac-
tion products are isolated as the distillation residue. The
reactions were monitored by GLC analyses until the con-
version reached the desired value. The reaction product
(when conversion was quantitative) was washed with aqu-
eous NaOH solution, the organic layer washed twice with
water, dried over sodium sulphate, ®ltered and distilled in a
30-plates column. When conversion was not complete, the
crude mixture was directly distilled.
19F NMR spectra were recorded on a Varian 200 MHz
spectrometer as neat sample and employing CFCl3 as
internal standard. IR spectra were recorded on a Nicolet
205 X FTIR instrument, UV spectra on a Beckman UV 5290
instrument. GLC analyses were performed with a HRGC
533 Carlo Erba instrument equipped with thermoconduc-
tivity detectors (4 m columns packed with 10% Fomblin
YR1 on Chromosorb W HP (60±80 mesh)). Differential
scanning calorimetry (DSC) was carried out using a Perkin
Elmer DSC 2 Calorimeter; calorimetric traces have been
obtained at a 108C/min heating rate.
2.1.3. Preparation of perfluoro-5,8,11,12,15,18-
hexamethyl-4,7,10,13,16,19-hexaoxa-docosane (PFPE24)
Following the same procedure as above (Method A) 400 g
(0.6 mol) of HFPO tetramer 2c were irradiated for 50 h.
After work-up 168 g of coupled product were isolated,
whose structure by 19F NMR was {CF3CF2CF2O[CF(CF3)-
CF2O]2CF(CF3)±}2, yield: 45.0%.
2.2. Photocoupling of perfluoropolyether acyl
fluorides
2.2.1. Photocoupling of CF3O[CF2CF(CF3)O]mCF2COF
In a photochemical reactor of 0.1 l capacity (Method A),
50 g of the acyl ¯uoride 1 were irradiated for 16 h. Con-
version was quantitative (by GLC) and, after work-up, 42 g
of {CF3O[CF2CF(CF3)O]mCF2±}2 were isolated.
2.3. Photocoupling of binary mixtures of acyl fluorides
2.3.1. HFPO dimer and perfluorooctanoyl fluoride
In a 0.1 l photochemical reactor, 42 g (0.1 mol) of
C7F15COF (3) and 34 g (0.1 mol) of n-C3F7OCF(CF3)COF
(2a) were charged. After 8 h of irradiation, the crude

C. Tonelli, V. Tortelli / Journal of Fluorine Chemistry 101 (2000) 117±123
119
mixture was distilled. The volatile fractions were essentially
3. Results and discussion
the starting acyl ¯uorides; then the symmetric coupling
product from HFPO dimer distilled, followed by the asym-
metric coupling product CF3CF2CF2OCF(CF3)CF2(CF2)5-
CF3. The residue of distillation was a white solid: the
major component (>80%) was identi®ed as CF3(CF2)12CF3
by GLC and 19F NMR comparison with an authentic
sample.
3.1. Syntheses
The linear per¯uoropolyethers (PFPE) described in this
work are obtained by photodimerization of acyl ¯uorides,
usually with good selectivity. The reaction mechanism
involves the excitation, by absorption of a photon, of a
carbonyl group (n±ꢁ* transition); the photoexcited acyl
¯uoride molecule cleaves to give a per¯uoro(oxa)alkyl
radical and a ¯uoroformyl one. The former reacts giving,
by a coupling reaction, a per¯uoro(oxa) alkane, the latter
couples giving oxalyl ¯uoride which, in turn, dispropor-
tionates to CO and COF2. The presence of by-products
resulting from side-reactions of the intermediate radicals is
strongly limited if the temperature is lower than 508C. The
whole reaction pathway is summarized in Scheme 1 (for the
per¯uoropolyether acyl ¯uoride 1).
2.3.2. Perfluoropolyether acyl fluoride and HFPO
tetramer
In the same reactor used before, equimolar quantities
(55 mmol) of acyl ¯uoride 1 and HFPO tetramer 2c were
irradiated for 16 h until complete conversion. After usual
work-up and distillation, the three main isolated products
were:
The selectivity of these reactions is quite high. For
symmetric photocouplings, the selectivity is over 85%
and the main by-products are branched structures of higher
molecular weight (5±12%), while lower molecular weight
molecules are less (<4%). The former come from the attack
of per¯uoro(oxa)alkyl radical on the COF group of a second
molecule, the latter arise from the homolytic scission of the
®rst formed secondary radical with elimination of CF3COF
or COF2 and subsequent coupling of the so generated new
fCF3O‰CF2CFꢁCF3†OŠmCF2 g2
fCF3CF2CF2O‰CFꢁCF3†CF2OŠ2CFꢁCF3† g2
fCF3CF2CF2O‰CFꢁCF3†CF2OŠ2CFꢁCF3†CF2
‰OCFꢁCF3†CF2ŠmOCF3
All the results are listed in Table 1. The 19F NMR spectra
are presented in Table 2.
Table 1
Photocoupling of fluorinated acyl fluorides
Reagents
Products
Time (h)
50
Conversion (%)
74.1
Selectivity (%)
93
50
50
16
52.0
53.2
100
88
85
90
25a
50
8
53.0
25
25a
16
100
50
25
a Minor amounts of by-products have not been considered.
b Rf: CF3O[CF2CF(CF3)O]m, Rf: n-C3F7O[CF(CF3)CF2O]2.

120
C. Tonelli, V. Tortelli / Journal of Fluorine Chemistry 101 (2000) 117±123
Table 2
19F NMR spectra of photocoupling productsa
Compound
ꢂ (ppm)
b
(a) 81.5; (b) 129.2; (c) 79.9 (JFF ˆ 127 Hz); (d) 139.2, 140.6; (e) 77.4, 78.4
(a) 81.5; (b) 129.2; (c) 79.9 (JFF ˆ 127 Hz); (d) 144.1; (e) 79.6; (f) 80.9 (JFF ˆ 122 Hz); (g)
139.8, 140.4; (h) 77.5, 78.1
(a) 81.5; (b) 129.2; (c) 79.9; (d) 143.9; (e) 79.6; (f) 80; (g) 139.7; (h) 77.7, 78.1
(a) 59.2; (b) 82.8; (c) 144.7; (d) 87.8
c
(a) 81.5; (b) 129.2; (c) 79.9 (JFF ˆ 127 Hz); (d) 139.7; (e) 78.0; (f, g) 119.7/ 122.1; (h)
125.6; (i) 81.7
(a) 81.5; (b) 129.2; (c) 79.5/ 80.5; (d) 144.0/ 144.4; (e) 142.8; (f) 78.7; (g) 59.4
a Reference CFCl3, neat samples.
b From Ref. [6].
c Some structural irregularities are present due to the starting acyl fluoride [12].
Scheme 1.
radical. When equimolar mixtures of two different acyl
¯uorides are reacted, a statistical distribution of coupling
products is found: two symmetric per¯uoro(oxa)alkanes
(50%, 1:1 ratio) and the asymmetric derivative (50%). In
this case, the presence of minor by-products has not been
evaluated.
The yields are satisfactory even if the mercury lamp gives
only a very low amount of emission below 254 nm, i.e. in
the range of absorption of the carbonyl group (see Table 4)
[13]. We determined also a product medium quantum yield
of the photocoupling dimerization of the oligomers of
HFPO. A good evaluation of the rate of ±COF consumption

C. Tonelli, V. Tortelli / Journal of Fluorine Chemistry 101 (2000) 117±123
121
Table 3
Table 4
Rf±COF consumption at Èi ˆ 0.103
" values of RfCOF as a function of wave length
a
Time (s)
Exp. conc.
1
Calc. conc.
1
(mol l )
1
I/I0
wave length (nm)
(")
(mol l
)
238
248
254
265
270
275
280
289
297
302
313
334
336
405
24.30
7.30
4.00
0.89
0.67
0.66
0.65
0.75
0.79
0.80
0.73
0.35
0.02
<0.01
0
9900
3.40
3.34
3.29
3.24
3.20
3.17
3.15
3.11
3.06
3.02
2.97
2.91
2.88
2.79
2.69
±
0.943
0.941
0.938
0.936
0.935
0.933
0.932
0.931
0.928
0.926
0.924
0.921
0.919
0.914
0.908
3.35
3.30
3.26
3.21
3.18
3.16
3.11
3.06
3.01
2.96
2.91
2.85
2.77
2.69
21 900
31 920
42 600
47 100
52 500
61 200
75 900
86 700
97 500
108 300
122 400
140 400
158 100
a I0: intensity of the incident light; I: intensity of the transmitted light.
Even if this is not a correct kinetic approach, we believe that
the data are note worthy because of the same experimental
conditions. They strongly suggest that the in¯uence of the
per¯uoro(oxa)alkyl chain on the rate of photocoupling
reaction of acyl ¯uorides is negligible. This result is in
agreement with the fact that all acyl ¯uorides have practi-
cally the same ꢃ in the range 238±350 nm [13] (typical
values are reported in Table 4). The above observations
together with the statistical distribution of the products in the
photocoupling of two different acyl ¯uorides (see Table 1)
support that quenching phenomena (Eq. (1)) are more
important than `in cage recombination' (Eq. (2)) to explain
the rather low quantum yield.
The 19F NMR spectra are useful for the elucidation of the
new structures (see Table 2). For the PFPE22±24 series, at
least two pairs of diastereoisomers (meso and dl) could be
present if we look only at the two secondary carbon atoms of
the newly formed carbon±carbon bond. The two isomers are
always strictly equimolar (from integration of the two
signals relative to the CF±CF ¯uorine chemical shifts) for
all the three dimers, independently of the HFPO-based chain
length; this evidence suggests that the secondary per¯uori-
nated radicals, generated in the ®rst step of the photo
reaction, have no steric requirement as well as no preferred
conformation leading to a very fast (diffusion controlled)
coupling.
has been obtained using the following equation:
"
#
X
cr"i
dCOF=dt ˆ I0
Fi fiꢁ1 e
† =V;
i
where I0 is the intensity of the incident light (E s 1); fi
the photon emission fraction at i line; "i the molar absorption
coefficient at i line (l mol 1 cm 1); V the reagents' volume
(0.337 l); c the concentration of absorbing species (mol l 1);
r the absorbing path length (0.7 cm); Èi the efficiency of the
photon at i line.
The best ®tting between theoretical and experimental
concentration vs time is achieved with Èi ˆ 0.103 (average
value over all the range of absorption) as shown in Table 3.
This value is in very good agreement with literature data
[14].
In the last column of Table 3, the calculated fractions of
the absorbed radiation by the mass of reaction are reported.
It appears that the light intensity is absorbed at over 90%,
while the quantum yield is much lower; this means that there
are signi®cant quenching phenomena (Eq. (1)) or `in cage
recombination' of primary radicals (Eq. (2)) which are
faster than the cleavage of the excited carbonyl groups or
`out of cage' diffusion:
3.2. Properties
Ã
‰RfOCF2COFŠ
!
quenching RfOCF2COF ‡ Energy
(1)
Cage reaction RfOCF2COF
Ã
Some of the main physical properties of PFPE22±24 are
reported in Table 5. In the same table other per¯uoropo-
lyethers of similar molecular weight and structure have been
characterized for comparison: the per¯uorooxetanes synthe-
sized by photocyclodimerization of the HFPO oligomers
with hexa¯uoropropene (PFOE2±4) [13] and the per¯uoro-
ethers obtained previously by ¯uorination of the HFPO
trimer and tetramer (PFPE3±4) with elemental ¯uorine [15].
From the data of Table 5, we observed the expected
increase of the boiling temperature Tb with increasing
‰RfOCF2COFŠ ! ‰RfOCF2 ÁÁ COFŠ
!
(2)
The data for the photocoupling of binary mixtures of acyl
¯uorides (Table 1) shows that the rate of reaction is similar.
We have irradiated separately two samples of 1 and 2c (very
similar molecular weight and density), at an initial concen-
tration of about 2.6 mol l 1, to con®rm this behaviour. After
8 h, conversion was 60% and after 16 h conversion was
quantitative (by GLC and 19F NMR analyses) in both cases.

122
C. Tonelli, V. Tortelli / Journal of Fluorine Chemistry 101 (2000) 117±123
Table 5
Physical properties of perfluoropolyethers
Properties
PFPE22
PFPE23
PFPE24
PFOE2
PFOE3
PFOE4
PFPE3
PFPE4
Molecular weight
Boiling point (8C)
Pour-point (8C)
Kinematic viscosity (cSt, 208C)
Density (g cm 3, 208C)
Vapour pressure (mmHg, 208C)
Surface tension (dynes cm 1, 208C)
570
134
107
1.2
902
210
81
1234
272
67
482
115
±
648
162
91
814
201
85
4.6
±
451
96
617
148
±
112
0.6
5.0
1.83
0.1
19
14.3
1.85
±
1.0
1.75
24.0
16
2.1
1.79
4.0
19
1.3
1.74
±
1.76
6.8
1.68
36.0
14
±
17
20
19
16
molecular weight (MW). The dependence of the Tb on the
structure was described for linear hydrogenated and per-
¯uorinated paraf®ns as well as for linear per¯uoropo-
lyethers by Marchionni et al. [16]. The behaviour of the
structures of the present work (PFPE22±24) compared with
PFOE2±4 and PFPE3±4 con®rms what was highlighted in that
work. In particular, the presence of oxygen atoms along the
backbone strongly decreases the intermolecular attractive
forces and this effect increases with oxygen content. More-
over, our results suggest that the presence of branching in
the molecule (±CF3 groups) does not signi®cantly affect the
Tb, the boiling point being actually determined by the
molecular attractive forces within the saturated liquid and
the vapor phase. As a consequence Tb is determined, for a
per¯uoropolyether structure, only by the MW and the oxy-
gen/carbon ratio. This is clearly shown in Fig. 1, where the
dependence of Tb against the number of backbone chain
atoms N for three different series of per¯uorinated mole-
cules is reported. It appears that all the products with
branching along the main chain (±CF3 and even oxetane
substituents) lie on the same line as linear per¯uoropo-
lyethers having similar O/C ratios, while the corresponding
per¯uoroparaf®ns show signi®cantly higher boiling points.
Some interesting remarks can be made on the dependence of
the density (or its inverse speci®c volume) on the MW and
the structure. For a monodisperse polymer, the molar
volume is de®ned by the relation:
V ˆ nV0 ‡ 2Ve;
where n ˆ M/M0 (M0 is the molar weight of the repeating
unit of molar volume V0) and Ve is the excess molar volume
due to the chain-end.
As a consequence, the speci®c volume v can be repre-
sented as follows:
v ˆ V=M ˆ V0=M0 ‡ 2Ve=M:
For a polydisperse polymer, n and M are the average
number degree of polymerization and the molecular weight,
respectively.
The speci®c volume at 293 K as a function of the
reciprocal average number molecular weight Mn for Fom-
blin Y1 polyethers [17] and for the PFPE22±24 are plotted
in Fig. 2. Both series of products give a good linearity with
experimental data. The following equations are derived:
Fig. 1. Boiling point of perfluoroparaffins (&), linear perfluoropolyethers
(^) [16], and branched perfluoropolyethers (*) vs the number of
backbone chain atoms (N).
Fig. 2. Specific volume of Fomblin Y1 polyethers (&) [17], and
PFPE22±24 series (*) vs the reciprocal of the average number of molecular
weight (1/Mn).

C. Tonelli, V. Tortelli / Journal of Fluorine Chemistry 101 (2000) 117±123
123
Table 6
Molar volume and excess molar volume as a function of PFPEs structure
1
)
1
1
1
)
1
Sample
V0/M0 (ml g
M0 (g mol
)
V0 (ml mol
)
Ve (ml mol
Ve/V0
VCFꢁCF † (ml mol
)
3
Fomblin Y
PFPE22±24 series
0.5205
0.5161
161
166
83.8
85.7
10.0
14.7
0.12
0.17
48.3
50.1
V ˆ 0:5205 ‡ 19:89=M ꢁR2 ˆ 0:989†; Fomblin Y1;
V ˆ 0:5161 ‡ 29:36=M ꢁR2 ˆ 0:989†; PFPE22
:
24
Assuming the literature value for the CF2 contribution to
V0 (27 ml/mol) [17,18] and oxygen (8.5 ml/mol) [19], it is
possible to derive the ±CF(CF3) contribution to V0 and the Ve
value. The data are summarized in Table 6. Actually the two
series are quite similar. However, the PFPEs22±24 present a
higher contribution to the real macromolecule volume from
chain-ends in terms of absolute numerical values (14.7 vs
10.0 ml/mol) as well as the fraction to the volume of the
basic repeating unit (17% vs 12%). This is in agreement
with the different chain-ends of the two series: Fomblin Y1
polyethers have ±OCF3 and ±OCF2CF3 as chain-ends while
PFPE22±24 have the longer ±OCF2CF2CF3 end.
Fig. 3. Pour-point of PFPE22±24, PFOE3±4 and PFPE3 series vs the
reciprocal of the average number of molecular weight (1/Mn).
[3] R.A. Mitsch, P.H. Ogden, A.H. Stoskopf, J. Org. Chem. 35 (1970)
2816.
The freeze temperature (pour point) is always very low
for all per¯uoropolyethers. The freeze temperatures mea-
sured for the per¯uoropolyethers of this work (Table 5) are
related to the molecular weight and this relationship can be
somewhat rationalized [20] by the following equation:
[4] J.L. Zollinger, J.R. Throckmorton, S.T. Ting, R.A. Mitsch, J.
Macromol. Sci., Chem. A 37 (1969) 1443.
[5] H.S. Eleuterio, J. Macromol. Sci., Chem. A 66 (1972) 1027.
[6] Z. Chengxue, Z. Renmo, P. Heqi, J. Xiangshan, Q. Yangling, W.
Chengjiu, J. Xikui, J. Org. Chem. 47 (1982) 2009.
[7] F.L.M. Pattison, J.B. Stothers, R.G. Woolford, J. Am. Chem. Soc. 78
(1956) 2255.
Tf ˆ A K=Mn:
[8] P. Bickle, W. Schwertfeger, Ger. Offen. 3 828 848, 1990.
[9] D. Sianesi, A. Pasetti, R. Fontanelli, G.C. Bernardi, G. Caporiccio,
Chim. Ind. (Milan) 55 (1973) 208.
A good ®t has been obtained as shown in Fig. 3.
The generally accepted assumption that Tg decreases with
decreasing Mn comes from the fact that each chain-end
contributes an excess molar free volume; this is also true for
the PFPEs of this work.
These properties together with some others typical of the
highly ¯uorinated structure (chemical and thermal stabili-
ties, low dielectric constants, etc) make the PFPEs inves-
tigated in the present work interesting candidates for special
electronic applications.
[10] H. Millauer, W. Schwertfeger, G. Siegemund, Angew. Chem., Int.
Ed. Engl. 24 (1985) 161.
[11] D.G. Cox, L.G. Sprague, D.J. Burton, J. Fluor. Chem. 23 (1983) 383.
[12] M. Pianca, N. Del Fanti, E. Barchiesi, G. Marchionni, Chimicaoggi
(1995) 29.
Á
[13] G. Bargigia, C. Tonelli, M. Tato, J. Fluor. Chem. 36 (1987) 449.
[14] N.N. Kravcenko, Vysokomol. Soedin. B 23 (1981) 180.
[15] G.E. Gerhardt, R.J. Lagow, J. Chem. Soc., Perkin 1 (1981) 1321.
[16] G. Marchionni, G. Ajroldi, M. Righetti, G. Pezzin, Macromolecules
26 (1993) 1751.
[17] G. Marchionni, G. Ajroldi, G. Pezzin, Eur. Pol. J. 2412 (1988) 1211.
[18] G. Marchionni, G. Ajroldi, P. Cinquina, E. Tampellini, G. Pezzin,
Polym. Eng. Sci. 30 (1990) 829.
References
[19] D.W. Van Krevelen, Properties of Polymers, Elsevier, Amsterdam,
1976.
[1] J.F. Harris Jr., J. Org. Chem. 30 (1965) 2182.
[2] A.S. Millian Jr., US Patent 3 214 478, 1965.
[20] G. Marchionni, R. Silvani, G. Fontana, G. Malinverno, M. Visca,
J. Fluor. Chem., in press.