An approach to fluorinated surface coatings via photoinitiated cationic cross-linking of mixed epoxy and fluoroepoxy systems

Article abstract of DOI:10.1016/S0022-1139(99)00288-2

The synthesis and characterisation of four polyfluorinated monoepoxides is described. These fluoroepoxy monomers were cationically polymerised either neat or as mixtures with diglycidylether of bisphenol-A using the initiator system (CH₃CH₂)₃Al/H₂O (2/1) or the photoinitiator Ph₂I⁺, PF₆^-. Contact angle measurements indicated that, in favourable cases, the fluorinated epoxide monomer surface segregated at low loadings to give significant lowering of the surface free energy of photocured films of diglycidylether of bisphenol-A.

Full text of DOI:10.1016/S0022-1139(99)00288-2

Journal of Fluorine Chemistry 102 (2000) 269±277
An approach to ¯uorinated surface coatings via photoinitiated cationic
cross-linking of mixed epoxy and ¯uoroepoxy systems
1
*
Stephen Matuszczak , W. James Feast
Interdisciplinary Research Centre in Polymer Science and Technology, Durham University, South Road, Durham, DH1 3LE, UK
Received 9 July 1999; received in revised form 29 July 1999; accepted 18 August 1999
Dedicated to Professor Paul Tarrant on the occasion of his 85th birthday
Abstract
The synthesis and characterisation of four poly¯uorinated monoepoxides is described. These ¯uoroepoxy monomers were cationically
polymerised either neat or as mixtures with diglycidylether of bisphenol-A using the initiator system (CH CH ) Al/H O (2/1) or the
3
2 3
2
‡
�
photoinitiator Ph . Contact angle measurements indicated that, in favourable cases, the ¯uorinated epoxide monomer
surface segregated at low loadings to give signi®cant lowering of the surface free energy of photocured ®lms of diglycidylether of
2
I , PF
6
bisphenol-A. # 2000 Elsevier Science S.A. All rights reserved.
Keywords: Fluoropolymers; Mixed epoxy system; Fluoroepoxy system; Surface effect
1
. Introduction
on a cheaper bulk organic ®lm. Since photocurable systems
polymerise only on irradiation, it is possible that ®lms
formed from mixtures of conventional monomers and poly-
¯uorinated monomers could be used to make cured ®lms
having ¯uoropolymer surfaces. Success would require that
the mixture was left long enough for the ¯uorinated species
to migrate to the surface prior to irradiation and cure. This
approach, using photocurable systems offers two advan-
tages; namely, segregation takes place whilst the system
is still mobile and is expected to occur rapidly, whilst the use
of a reactive monomer means that the ¯uorinated groups are
likely to be covalently bound to the network on cure.
In this paper we report the results of attempts to obtain
¯uorinated low energy surfaces by adding small amounts of
poly¯uorinated epoxides to conventional epoxy monomers
containing photocationic initiator and exposing spread ®lms
of these mixtures to UV radiation; however, the concept is
general irrespective of cure mechanism.
Fluoropolymers offer the potential advantages of low
surface energy, good chemical resistance, low coef®cient
of friction and good thermal stability. These attractive
properties stem from the high electron af®nity, small atomic
radius and low polarisability of ¯uorine. However, the cost
of such materials restricts their applications to areas where
such properties are essential and where a comparable per-
formance cannot be achieved using less expensive polymers.
Fluoropolymers are attractive in several areas of surface
coatings technology, for example, in anti-fouling or fouling-
release and in easily cleaned coatings. Since surface proper-
ties are determined by the nature and packing of the exposed
species at the surface [1], the bulk of such coatings need not
be composed of ¯uoropolymer. Indeed the surface properties
associated with ¯uoropolymers might be achieved by having
a conventional commodity polymer with a thin surface layer
rich in per¯uorinated units. The propensity of low surface
free energy species to migrate to the air±bulk interface when
incorporated into a suitable organic medium offers a poten-
tially low cost route to the generation of a ¯uorinated surface
2
. Results and discussion
2
.1. Surface segregation of low surface energy species
*
Corresponding author. Tel.: ‡44-191-3743105;
There is good evidence for surface segregation of low
fax: ‡44-191-3744651.
surface free energy species. Zisman et al. have demonstrated
that per¯uoroalkyl acids and esters added in small amounts
to polymers such as polystyrene and poly(methyl metha-
E-mail address: w.j.feast@durham.ac.uk (W.J. Feast).
1
Present address: Sherwin-Williams Automotive Finishes Corp., 10909
South Cottage, Grove Avenue, Chicago, Illinois 60628, USA.
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 8 8 - 2

2
70
S. Matuszczak, W.J. Feast / Journal of Fluorine Chemistry 102 (2000) 269±277
crylate) segregate to the polymer surface [2,3]. It was shown
that the effectiveness of a particular additive in a particular
polymer depended on the balance between compatibility
and incompatibility of the additive and polymer. These
studies showed that small amounts of ¯uorinated additives
produced marked decreases in the surface energies and
coef®cients of friction of these polymers. Those systems
where surface segregation occurred also exhibited the prop-
erty of self-healing, in that, when damaged, the ¯uorinated
surface could be regenerated by warming the ®lm to
encourage further migration of the additive from the bulk
to the surface.
Another example of surface segregation of ¯uorinated
units was provided by Hult and Ranby who used a photo-
initiator containing a per¯uoroalkyl group in free radical
photocurable systems [4]. Using ESCA, they showed that
the concentration of photoinitiator residues at the ®lm±air
interface increased as the viscosity of the coating decreased
or the interval between application and irradiation
increased. In this system the photoinitiator fragments were
also found to migrate to the substrate±®lm interface as well
as the ®lm±air interface.
hexadecane) on polytetra¯uoroethylene gave a good straight
line from which a value of 19 Æ 2 dyn cm� for g was
1
c
obtained. This compares favourably with the literature value
�
1
of 18 dyn cm [1]. The n-alkanes could not be used to
determine g for polyethylene since they wetted the sample
c
�
1
indicating that g must have a value above 27.5 dyn cm . A
c
measure of g for the polytetra¯uoroethylene sample was
c
made using a non-homologous series of liquids: water,
glycerol, formamide, methylene iodide, benzonitrile, benzyl
alcohol, n-hexadecane and n-decane. The plot of cos y
against the surface tension for this series gave a straight
�
1
line with a g value of 14 Æ 3 dyn cm , which is lower than
c
the literature value. Similarly g for the polyethylene sample
c
was measured using the above series of liquids, with the
exception of the two n-alkanes, and the plot of cosy versus
surface tension of the liquids gave a g value of
c
�
1
32 Æ 4 dyn cm which is close to the literature value [1]
�
1
of 31 dyn cm
.
Doubly distilled water and methylene iodide that had
been freshly vacuum transferred were used to estimate the
surface energy, g , of the two samples by the method of
s
Owens and Wendt [8]. The value of g obtained for poly-
s
�
1
Contact angle, ESCA and attenuated total re¯ectance IR
spectroscopy measurements have been used to demonstrate
segregation of the low energy components of polymer
mixtures of block copolymers to the air±bulk interface
tetra¯uoroethylene was 16 Æ 1 dyn cm which is lower
�
1
than the value of 19 dyn cm reported by these workers.
The value of g obtained for polyethylene was 31 Æ
s
�
1
1 dyn cm which is again slightly lower than the value
�
1
[
(
5]. The surface of a poly(dimethylsiloxane)-polystyrene
PDMS-PS) block copolymer was found to be essentially
of 33 dyn cm reported by Owens and Wendt [8].
These experiments suggested that the techniques adopted
for surface energy measurements gave reasonable agree-
ment with the literature and could be used with con®dence.
pure PDMS by contact angle and ESCA measurements [6].
Surface segregation of the polystyrene component in a PS-
poly(oxirane) block copolymer has been observed even
though the difference in surface free energy between the
two components is quite small [7]; the surface abundance of
polystyrene was shown to be dependent on the solvent used
to prepare the ®lms.
2.3. Preparation of the polyfluorinated epoxide monomers
used in this work
Poly- and per¯uorinated epoxide compounds have a quite
extensive literature. Fluorinated monoepoxides have been
used as monomers for the preparation of linear ¯uorinated
polyethers [9], and to obtain ¯uorinated surfaces via reac-
tion with suitable surface functional groups [10]. Poly¯uori-
nated diepoxide compounds have been synthesised for use
in surface coatings since they give a cross-linked network on
cure [11,12].
2.2. Determination of the surface free energy of solid
polymers
In the work reported here contact angles of probe liquids
were measured by measuring the height, h, and base, b, of a
±2 ml drop of the liquid on the surface using a microscope
equipped with a graticule. The contact angle, y, is given by:
1
The simple monoepoxide compounds 1±4 shown below
were used in the work:
�
1
y ˆ 2 tan (2h/b). For some liquids the contact angle was
found to decrease with time, probably due to evaporation.
Consequently, contact angles were measured immediately
after application of the drop or plotted against time and
extrapolated to zero time. For each surface and liquid, y was
measured several times and averaged. As a check of the
method, the surface energies of commercial samples of
polyethylene and polytetra¯uoroethylene were measured.
The polymer samples were washed with detergent and
water, rinsed thoroughly with distilled water and then left
for several days under high vacuum over P O .
2
5
A plot of cos y against the surface tension of a series of
unpuri®ed n-alkanes (heptane, octane, nonane, decane and
described in this paper.

S. Matuszczak, W.J. Feast / Journal of Fluorine Chemistry 102 (2000) 269±277
271
7,7,7,6,6,5,5,4,4,3,3,2,2-Trideca¯uoroheptyl oxirane 3
95%) was commercially available, the IR spectrum of
the material supplied indicated contamination with hydro-
xyl and carbonyl containing impurities which were removed
by vacuum distillation (Vigreux column, b.r. 72±758C, 18±
methylene iodide method. The plot of cos y against the
(
surface tension of the series of n-alkanes gave a g value of
c
�
1
19 Æ 3 dyn cm . The water/methylene iodide method
�
1
yielded a g value of 15 Æ 2 dyn cm . These measurements
s
indicated that the surface energy of ®lms of polymer 6 is
quite low being similar to that of polytetra¯uoroethylene.
Glass slides were coated using a solution of polymer 7 in
per¯uoro-2-butyl THF. The plot of cos y against the surface
1
9 mm Hg).
Compounds 1 and 2 were made by reaction of the
appropriate poly¯uoroalcohol with epichlorohydrin in the
presence of base [13]. Compound 4 was prepared by the
method of Brace [14] which involves the free radical
reaction of allyl acetate and per¯uoroethyl iodide in the
presence of a small amount of azobisisobutyronitrile.
tension of the series of n-alkanes gave a g value of
c
�
1
13 Æ 3 dyn cm whilst the water/methylene iodide method
�
1
gave a value of 12 Æ 1 dyn cm , these are in reasonable
agreement and are somewhat lower than for polymer 6,
The intermediate 5 is then treated with sodium hydroxide
to generate the required:
which is attributed to the ±CF group and higher ¯uorine
3
content.
product.
2.5. Addition of the poly¯uorinated epoxides to
conventional epoxy resins
2.4. Characterisation and surface energy measurements of
polymers prepared by the conventional cationic
polymerisation of monomers 1 and 3
Films of 100 mm nominal thickness were prepared on
polyethylene using a draw-down block or in the case of low
viscosity mixtures by ®lling a channel of 100 mm depth. The
®lms contained approximately 2% of diphenyliodonium
hexa¯uorophosphate as photoinitiator [16] and 1% w/w
acetone with the remainder comprising, in most cases,
diglycidylether of bisphenol A (DGEBA) or a mixture of
DGEBA and poly¯uorinated epoxides. The ®lms were
irradiated using two 1.8 kW medium pressure mercury
lamps in a Primarc Mini-Cure Apparatus and a total expo-
sure of ca. 10 s and then stored for 24 h prior to making
contact angle measurements. Unless otherwise stated the
measurements were carried out on the free surface (top
surface) using the water/methylene iodide method to give gs
values.
Compounds 1 and 3 were polymerised using AlEt /0.5
3
H O initiation (see Section 3 and earlier work [15]) to
2
give:
2.5.1. Effect of a low concentration of 1 on the surface
energy of films
Glass slides cleaned with nitric acid, rinsed thoroughly
with distilled water and dried were coated using a solution
of polymer 6 in methanol. After allowing the solvent to
evaporate the coated slides were stored under high vacuum
for several days and surface energy measurements were
carried out using the series of n-alkanes and the water/
The values of g for both the top surface and the surface
s
adjacent to the substrate (bottom surface) were determined
for ®lms containing 0, 2 and 5% w/w 1. The ®lms were
irradiated either immediately after application or after
standing for 4 h. The values obtained for g are recorded
s
in Table 1.

2
72
S. Matuszczak, W.J. Feast / Journal of Fluorine Chemistry 102 (2000) 269±277
Table 1
s
Values of g for DGEBA films containing 0, 2 and 5% w/w of 1
�
1
%
w/w 1
Time left prior
to cure (h)
g (dyn cm )
s
Top
Bottom
0
2
0
48 Æ 1
42 Æ 1
0
4
43 Æ 2
43 Æ 1
43 Æ 2
43 Æ 1
5
0
4
41 Æ 1
39 Æ 1
45 Æ 1
38 Æ 1
The value of g for the bottom surface of the ®lm contain-
s
ing 1 appeared to be signi®cantly lower than that for the top
surface, the value for the top surface decreased on addition
of 2% w/w 1 and showed a further slight decrease when 5%
s
Fig. 1. Decrease of g as the concentration of 1 increases (D, films
irradiated immediately; *, films left for 4 h).
w/w 1 was used. There was no signi®cant decrease in g for
s
the top surfaces of the ®lms containing 1 which were left
prior to irradiation. The fairly consistent values of g for the
bottom surface of ®lms containing 1 indicate that it did not
segregate to the ®lm±substrate interface. If segregation of 1
The surface segregation of 1 in an epoxy novolac system
was also investigated with similar results.
It appears that complete surface segregation of 1 did not
s
to the top surface was complete the value of g would be
occur although the decrease in the values of g indicated
s
s
expected to approach that of the polymer 6, i.e. less than
�
some surface enrichment in 1 and the presence of small
amounts of 1 increased the lacquers' ability to wet the
substrate. The limited surface segregation of 1 could be a
result of its compatibility with the resin.
1
2
0 dyn cm . Measurement of g for the substrate surface
s
adjacent to the ®lms showed that it was not signi®cantly
different from that for a pristine sample.
The series of n-alkanes were found to wet the top surfaces
of the ®lms totally consistent with g greater than
2.5.2. Effect of increasing the concentration of 1 on the
surface energy of films
c
�
1
2
7.5 dyn cm . An estimate of g for the top surfaces of
c
the ®lms was made using a non-homologous series of
liquids (water, glycerol, methylene iodide, benzonitrile,
benzyl alcohol, glycol and formamide). This gave a value
Films of lacquers containing up to 60% w/w 1 were
prepared and left for 0 or 4 h. prior to irradiation, ®lms
containing >80% of 1 failed to cure.
�
1
of 44 Æ 3 dyn cm for the ®lm containing 1 and values of
Fig. 1 shows a plot of g as a function of the mol% of 1 for
s
�
1
3
left for 0 and 4 h, respectively, prior to irradiation.
9 Æ 2 and 38 Æ 1 dyn cm for the ®lms containing 5% 1
®lms cured immediately after application. The value of gs
decreases as the concentration of 1 rises to 20 mol% and
then shows a less pronounced decrease as the concentration
In an attempt to aid the surface segregation of 1 by
reducing the viscosity of the DGEBA system using the
reactive diluent butane-1,4-diglycidylether (BDGE), ®lms
were prepared from DGEBA containing 15% BDGE and
is increased further. The values of g for ®lms left for 4 h
s
prior to cure were slightly higher than the values exhibited
by ®lms cured immediately after application. For a statis-
tical mixture of base epoxy resin and ¯uorinated additive
5
% w/w 1. The values of g after cure were similar to those
s
obtained previously and there was no evidence of enhanced
surface segregation of 1. Increasing the temperature at
which ®lms were stored prior to cure failed to increase
the surface segregation since the effectiveness with which
the lacquers wet the polyethylene was decreased. For exam-
ple, ®lms prepared from DGEBA containing 2 and 5% of 1
could only be left for 8 and 16 min, respectively, at a
temperature of 588C, before they started breaking up into
one might expect a linear decrease in g from the value of the
s
pure epoxy to that of the pure ¯uoroepoxy polymer, the
marked divergence from such a straight line can be taken as
an indicator of surface segregation.
These results con®rm that although complete surface
segregation did not occur, a signi®cant reduction of gs
was achieved with increasing concentration of 1. ESCA
analysis was carried out on ®lms containing 1.7, 8.6 and
35.6 mol% 1 and cured immediately after application. Fig. 2
shows the C1s core level spectra, similar spectra were
obtained for the corresponding ®lms left for 4 h prior to
irradiation.
droplets and the g values for these ®lms were the same as
s
those for control ®lms left at room temperature. It was
considered that the use of a high surface energy substrate,
steel, might encourage surface segregation of 1. However,
®
containing 5% w/w of 1 and left for 0 or 12 h prior to
lms prepared on mild steel panels from DGEBA resin
The two component peaks at 288.15 and 292.00 eV are
attributable to CF ±CH ±O and ±CF -moieties, respec-
tively, and as expected, increase in intensity as the con-
2
2
2
irradiation showed values of g close to those reported
s
above.
centration of 1 increased. The F1s signal was also found to

S. Matuszczak, W.J. Feast / Journal of Fluorine Chemistry 102 (2000) 269±277
273
Fig. 3. Decrease of the contact angle of DGEBA/1 mixtures on
polyethylene as the concentration of 1 increases.
does not progress further to give a F/C ratio of ꢀ1.2,
corresponding to a surface composed of 1, is unclear.
Comparison of the experimental F/C ratios determined
Fig. 2. C1s core level spectra as a function of the concentration of 1 (a,
.7; b, 8.6; c, 35.6 mol%).
from the C1s core level spectra and F /C intensity ratios
1s 1s
1
shows that the latter gives a higher F/C ratio than the former.
A plausible explanation for this observation, bearing in
mind the shallower sampling depth for ¯uorine than for
carbon, is that the concentration of 1 increases towards the
surface which provides further evidence for surface enrich-
ment with 1.
increase in intensity as the concentration of 1 increased and
�
no signal due to PF6 could be detected.
The F/C atom ratio was determined from both the C core
1
s
level spectra and the F /C intensity ratio and compared
s
1
1s
with the expected value from the lacquer composition. Table
records the results as a function of the concentration of 1
Fig. 3 shows a plot of the contact angle of DGEBA/1
mixtures on polyethylene against the mol% of 1. The
contact angle, y, decreases and hence the ability of the
lacquer to wet polyethylene increases as the concentration
of 1 increases.
2
and the interval between application and irradiation.
As expected the experimental F/C ratios increase with
increasing concentration of 1. Both the experimental ratios
however, are signi®cantly greater than the expected value
for each concentration of 1 based on uniform ®lm composi-
tion. Although the magnitude of the difference increases as
the concentration of 1 increases, the proportionate increase
in ¯uorine at the surface diminishes. Since leaving ®lms for
2.5.3. Effect of low concentrations of 3 alone or with 1 on
the surface energy of films
Mixtures containing 2 or 5% w/w 3 were cloudy.
Attempts to prepare ®lms from the mixtures were unsuc-
cessful because of the incompatibility of 3 with DGEBA.
Reducing the amount of 3 to 0.6% w/w resulted in a clear
lacquer from which clear ®lms were prepared, but these
4 h does not result in a signi®cant change in the F/C ratio, it
can be concluded that there is a limited but rapid surface
segregation of 1. Why the surface segregation is limited and
®
lms broke up into droplets after ꢀ20 min. The value of g
s
�
1
(40 Æ 1 dyn cm ) for such ®lms left 20 min before irradia-
tion was slightly lower than the value (45 dyn cm ) for a
Table 2
� 1
F/C ratios from the C1s core level spectra (a) and F1s/C1s intensity ratios
®lm cured immediately after application suggesting that
some surface segregation occurred. To overcome the incom-
(
b) as a function of the concentration of 1
Mol% of 1
Time left prior
to cure (h)
F/C ratio
a)
Expected
F/C ratio
patibility of 3 with DGEBA mixtures containing both 3 and
1
were prepared in the hope that the latter would act as a
(
(b)
wetting agent. Mixtures containing 5% 1 with 1 or 2% w/w 3
gave good clear ®lms. Increasing the concentration of 3
relative to 1 resulted in the problem of incompatibility
reappearing. Films of the above composition were prepared
1
8
.7
.6
0
4
0.034
0.038
0.050
0.043
0.010
0.053
0.265
0
4
0.110
0.094
0.137
0.136
and cured and g measured as a function of the substrate and
s
3
5.6
0
4
0.340
0.340
0.433
0.435
interval between application and irradiation, the results are
recorded in Table 3and show that surface segregation of low

2
74
S. Matuszczak, W.J. Feast / Journal of Fluorine Chemistry 102 (2000) 269±277
Table 3
values for films of DGEBA containing both 3 and 1
weight loss amounts to ꢀ10% of the initial weight of the
lm.
Thus, the volatility of 2 is responsible for its lack
of surface effect, the loss was probably exacerbated by
the heat emitted by the 1.8 kW UV source used in the
cure.
g
s
®
�
1
)
%
w/w 3 % w/w 1 Substrate
Time left prior
to cure (h)
g
s
(dyn cm
1
5
Polyethylene
0
8
37 Æ 2
31 Æ 2
32 Æ 3
7
4
4
2
2.5.5. Effect of low concentrations of 4 alone or with 1 on
the surface energy of films
2
2
5
5
Polyethylene
Steel
0
8
32 Æ 1
25 Æ 1
Mixtures of DGEBA containing 2 and 5% w/w 4 were
prepared and found to be clear, ®lms were left for 0 or
0
8
34 Æ 1
26 Æ 2
3
0 min prior to irradiation. The values of g for these
s
cured ®lms did not vary signi®cantly and were similar
to that of a ®lm containing no ¯uorinated additive. Films
containing 2% 4 with 5% w/w 1 left for 0, 30 min and
surface energy material occurred and that the initial and
nal value of g was dependent on the concentration of 3 but
®
4 h prior to irradiation showed values of g which did not
s
s
not the substrate. However, the ®nal values of g obtained do
not reach those of either polymer 6 or 7 so segregation to the
surface was incomplete.
vary signi®cantly and were comparable to that of a ®lm
containing 5% w/w 1 alone; thus, the presence of 4 has
no effect on the surface energy of the ®lms. As in the case
of 2 this lack of effect may be attributed to the volatility
of 4.
s
2.5.4. Effect of low concentrations of 2 alone or with 1 on
the surface energy of films
In view of the partially successful attempts to obtain a
poly¯uorinated surface using 1 and 3, compounds 2 and 4
were tested to see if these smaller species would show a
greater propensity for surface segregation.
Films containing 2 and 5% w/w of 2 were prepared, but
broke up into droplets after only 5±10 min. Unlike the
systems discussed above, a ®lm cured immediately after
3. Experimental
3.1. 7,7,6,6,5,5,4,4,3,3,2,2-Dodecafluoroheptyloxymethyl
oxirane 1
7,7,6,6,5,5,4,4,3,3,2,2-Dodeca¯uoroheptanol (164.8 g,
0.50 mol), epichlorohydrin (ECH) (231.5 g, 2.50 mol),
sodium hydroxide (20.2 g, 0.51 mol) and a catalytic amount
of water were placed in a three-necked RB-¯ask carrying
thermometer, stirrer and condenser and swept by nitrogen.
The reaction mixture was stirred at 708C for 5 h, cooled, the
white solid ®ltered off and the excess ECH removed under
rotary evaporation. The impure product was distilled
through a Vigreux column under reduced pressure and
the fraction with b.r. 108±1128C at 9±11 mm Hg was
redistilled using a Fischer±Spaltrohr concentric tube dis-
tillation apparatus. The major fraction (83.5 g, 82±838C,
1.1 mm Hg) was shown to be a single component by GC
analysis. This material was transferred to a storage vessel
under vacuum using grease-free apparatus and identi®ed
as 7,7,6,6,5,5,4,4,3,3,2,2-dodeca¯uoroheptyloxymethylox-
irane 1, found C, 30.4; H, 2.4; F, 56.1% C H F O requires
application showed a value of g similar to those of a ®lm
containing no ¯uorinated additive. Films containing 1% 2
with 5% 1 were prepared and left for 0, 10, 20 min and 4 h
s
before being irradiated. The values of g for the ®lms did not
s
vary signi®cantly and were similar to that of a ®lm contain-
ing 5% w/w 1; thus, the presence of 2 had no effect on
the surface energy of ®lms. The possibility that the volatility
of 2 might result in rapid loss from ®lms was tested when
a ®lm containing 10% 2 with 5% w/w 1 was prepared and
its weight followed with time. As Fig. 4 shows, there is
a rapid weight loss over the ®rst 10±20 min amounting
to ꢀ6% of the initial weight of the ®lm. After 22 h the
10 8 12 2
C, 30.9; H, 2.1; F, 58.8, 43% yield. IR absorptions (assign-
ments) at 3070±3010 (oxirane C±H stretch), 2940±2860
(
aliphatic C±H stretch), 1485±1405 (aliphatic C±H bend),
1
210±1155 (C±H stretch ‡ aliphatic C±O±C asymmetric
1
stretch) and 905 (oxirane ring stretch). The H-NMR spec-
trum provided good con®rmatory structural evidence
and the resonances and assignments are shown in Table
4
ments.
. The assignments were con®rmed by decoupling experi-
13
The C-NMR spectrum was fully consistent with the
assigned structure with peaks at 43.32(s), 50.21(s) (oxirane
Fig. 4. Weight loss of a film containing 10% 3 with 5% w/w 1 as a
function of time.

S. Matuszczak, W.J. Feast / Journal of Fluorine Chemistry 102 (2000) 269±277
275
Table 4
1
Assignments of the H spectrum of 1
Hydrogen
Shift (ppm)
Splitting
Integration
H
H
a
d
d
A
ˆ 2.58
ˆ 2.77
Ab
2
a� b
q
1
1
3
b� c
3
a� c
b
B
J
ˆ 4.9 Hz, J
ˆ 4.2 Hz, J ˆ 2.6 Hz
H
c
3.13
Multiplet
1
H
H
d
e
or H
or H
e
d
d
A
ˆ 3.48
ˆ 3.93
Ab
2
d� e
q
3
3
d
B
J
ˆ 11.9 Hz, J
d=e� c
ˆ 6.0 Hz, J
e=d� c
ˆ 2.5 Hz
1
H
f
4.00
Multiplet
Triplet of triplets: J2
3; (2H
f
‡ Hd/e
)
3
H
g
6.02
ˆ 51.9 Hz, J
ˆ 5.2 Hz
1
Hg� F
Hg� F
C), 68.07 (t, J2 ˆ 25.6, CF ±CH ±O), 73.12(s), 107.67 (tt,
3040±3000 (oxirane C±H stretch), 2920 (aliphatic C±H
stretch), 1450±1350 (aliphatic C±H bend), 1200±1100
(C±F stretch ‡ aliphatic C±O±C asymmetric stretch) and
CF
2
2
1
CF
2
CF
J
ˆ 254.6, J ˆ 31.5, HCF CF and overlapping multi-
2 4
2
2
plets between 104.51 and 118.68, (CF ) .
�
1
1
9
00 cm (oxirane ring stretch); the H-NMR resonances
3
.2. 3,3,3,2,2-Pentafluoropropyloxymethyl oxirane 2
and assignments are shown in Table 5 and were con®rmed
by decoupling experiments.
The C-NMR spectrum displayed resonances at 42.83(s)
13
3,3,3,2,2-Penta¯uoropropanol (112.8 g, 0.75 mol), ECH
352.8 g, 3.81 mol), sodium hydroxide (32.0 g, 0.80 mol)
2
CF
(
and a catalytic amount of water were mixed as described
and 49.87(s) (oxirane carbons), 67.38 (t, J ˆ 2.64 Hz,
1
CF
3
CF
CF CH O), 72.88(s), 112.94 (tq, J ˆ 254.2 Hz, J
ˆ
ˆ
2
2
1
CF
2
CF
previously. After stirring for 8 h at 708C, the white solid was
®
nitrogen (Fischer±Spaltrohr apparatus) to give a colourless
37.1 Hz CF ) and 118.61 (qt, J ˆ 285.3 Hz, C
2
ltered off and the product recovered by distillation under
35.0 Hz, CF3).
liquid b.r 144.5±144.78C as a single component (GC)
3.3. 3,3,3,2,2-Pentafluoropropyl oxirane 4
(
was identi®ed as 3,3,3,2,2-penta¯uoropropyloxymethylox-
51.0 g, 33%). After transfer under vacuum the material
Freshly recrystallised azobisisobuteronitrile (4.1 g,
0.025 mol) and allyl acetate (75.8 g, 0.76 mol), distilled
under nitrogen and stored over molecular sieve prior to
irane 2, found C, 35.2; H, 3.1; F, 45.6% C H F O requires
6
7
5 2
C, 35.0; H, 3.4; F, 46.1%; IR absorptions (assignment) at
Table 5
Assignment of the H resonances in the spectrum of 2
1
Proton
Shift (ppm)
Splitting
Integration
2
a� b
3
b� c
H
H
a
d
d
A
ˆ 2.59
ˆ 2.78
Ab
3
a� c
q
: J
ˆ 4.9 Hz, J
ˆ 4.2 Hz
1
1
b
B
J
ˆ 2.7 Hz
H
c
3.4
Multiplet
1
1
2
d� e
3
H
H
d
or H
or H
e
d
d
A
ˆ 3.47
ˆ 3.94
Ab
3
q
: J
ˆ 11.9 Hz, J
ˆ 6.0 Hz
d=e� c
e
d
B
J
e=d� c
ˆ 2.6 Hz
H
f
3.96
Multiplet
3; (2H
f
‡ Hf/e
)

2
76
S. Matuszczak, W.J. Feast / Journal of Fluorine Chemistry 102 (2000) 269±277
use, were placed in a steel autoclave (1.11). Per¯uoroethyl
iodide (185.4 g, 0.75 mol) was transferred into the cooled
vessel under vacuum which was then heated slowly using a
furnace in an isolation cell. The temperature of the mixture
was monitored by a thermocouple placed in a well in the
bottom of the vessel. As the temperature of the mixture
approached 50±608C an exotherm occurred peaking at
3.4. Polymerisation of 1
Compound 1 (6.6 g, 0.017 mol) was treated with a solu-
3
tion of AlEt /0.5 H O in hexane (ca. 2.4 M in Al) (1.7 cm ,
3
2
[monomer]/[AlEt ], 20/1). The mixture was stirred for 6 h at
3
3
808C. Methanol (40 cm ) was added to quench the reaction.
The solution was ®ltered and chloroform added to the
®ltrate to precipitate the polymer. The polymer was dis-
solved in methanol and reprecipitated a further two times
and then dried under high vacuum. Polymer 6 was obtained
as a viscous oil, found C, 30.7; H, 2.2; F, 59.2% C H F O
ꢀ
858C. The reaction mixture was then maintained at
778C for 5 h. After allowing the vessel to cool the clear,
ꢀ
pale green liquid reaction product was ®ltered to remove
insoluble byproducts and distilled, (Vigreux column) under
reduced pressure to give a main fraction b.r. 82±858C (6±
1
0 8 12
requires C, 30.9; H, 2.1; F, 58.8%, 4.4g (66%). The IR
spectrum of the product polymer was similar to that of the
monomer with signi®cant differences consistent with it
being a polyether with hydroxyl end groups. Namely, the
appearance of a weak, broad, hydroxyl absorption at
7 mm Hg), (194.2 g, 75%). GC-mass spectroscopy analysis
indicated that this material was mostly the expected inter-
mediate (see Section 2). The intermediate (193.2 g,
3
.56 mol) was dissolved in diethyl ether (150 cm ) in a
3
00 cm RB-three-neck ¯ask equipped with a mechanical
0
5
� 1
ꢀ3440 cm and disappearance of absorptions attributable
�
1
to oxirane C±H (3070±3010 cm ) and oxirane ring stretch-
stirrer/nitrogen inlet, condenser and thermometer. A small
portion of the sodium hydroxide (total used 42.9 g, 1.1 mol)
was added and the mixture gently warmed to initiate reac-
tion. Further portions of the sodium hydroxide were added
at intervals keeping the ether re¯uxing. After addition of
all the sodium hydroxide the mixture was re¯uxed for a
further 7 h.
The mixture was ®ltered to remove the white precipitate
that had formed, distillation under nitrogen (Fischer±Spal-
trohr apparatus) gave 3,3,3,2,2-penta¯uoropropyloxirane 4
�
1
ing (905 cm ) modes. Absorptions attributable to aliphatic
�
1
C±H stretching (2920±2880 cm ) increased in relative
�
1
intensity as did the absorption in the 1240±1080 cm
region attributed to aliphatic C±O±C. GPC (PL gel columns
5
(10 , 10 and 500 A), THF, RI detection and polystyrene
3
Ê
5
calibration) gave M ˆ 1.3  10 with M /M ˆ 2.4).
n
w
n
3.5. Polymerisation of 3
(
C H F O requires C, 34.1; H, 2.8; F, 54.0%. IR spectrum
20.9 g b.r. 67±698C); Found: C, 34.6; H, 3.1; F, 53.3%
Compound 3 (8.0 g, 0.021 mol) was polymerised using
the AlEt /0.5 H O initiator solution as described above
(2.7 cm , [monomer]/[AlEt
red for 5 h at 858C. Methanol (40 cm ) was then added to
quench reaction and Polymer 7 precipitated. After washing
with methanol product was puri®ed by reprecipitation from
per¯uoro-2-butyl THF using carbon tetrachloride as the
non-solvent to give Polymer 7 as a yellow viscous oil
4.04 g (51%) after drying under vacuum. Found C, 28.5;
4
5 5
3
2
3
absorption (assignment) 3060±3010 (oxirane C±H stretch),
930 (aliphatic C±H stretch), 1480±1315 (aliphatic C±H
3
] ˆ 16). The mixture was stir-
3
2
bend), 1200 (C±F stretch) and 905 cm
�
1
(oxirane ring
1
stretch); the H-NMR spectral parameters and assignments
are shown in Table 6.
13
The C NMR spectrum showed resonances at 35.1 (t,
2
CF
3
CF
J
ˆ 21.8 Hz, CF CH ), 44.71 (t, J ˆ 5.3 Hz,), 45.62(s),
2
2
1
1
15.05 (tq, J1 ˆ 252.5 Hz, J ˆ 38.6 Hz, CF ) and
2
CF
H, 1.1; F, 65.3% C
65.7%.
9
H
5
F
13O requires C, 28.7; H, 1.3; F,
CF
2
19.07 (qt, J1 ˆ 284.8 Hz, J ˆ 35.6 Hz, CF ).
2
CF
CF
3
Table 6
1
Assignments of the H NMR spectrum of 4
Proton
Shift (ppm)
2.29
Splitting
Integration
2
H
a
Multiplet
2
b� c
3
c� d
H
H
b
d
d
A
ˆ 2.56
ˆ 2.85
Ab : J
q
ˆ 4.8 Hz, J
ˆ 4.4 Hz
1
1
3
c
B
J
b� d
ˆ 2.5 Hz
H
d
3.18
Multiplet
1

S. Matuszczak, W.J. Feast / Journal of Fluorine Chemistry 102 (2000) 269±277
277
The IR spectrum of Polymer 7 was broadly similar to that of
the monomer with similar differences to those found for
Polymer 6 consistent with formation of the polyether.
Polymer 7 could not be analysed using GPC as it was not
soluble in any appropriate solvent.
neering Research Council together with Camrex (Holdings)
Ltd. for a studentship (SM) and Drs. H.S. Munro and J.G.
Eaves for useful discussions and help in obtaining the ESCA
spectra. This work is abstracted, in part, from the Ph.D.
Thesis of SM, Durham University, 1987.
4
. Conclusions
References
The concept of producing a ¯uorinated polymer surface
[1] W.A. Zisman, Adv. Chem. Ser. 43 (1964) 1.
[
[
[
2] R.C. Bowers, N.L. Jarvis, W.A. Zisman, Ind. Eng. Chem. Prod. Res.
Dev. 4 (1965) 86.
on a cheaper photocured commodity polymer via surface
segregation prior to cure has been demonstrated. Contact
angle and ESCA measurements indicate that, although
surface segregation does occur in favourable cases, in the
systems studied here it was never complete in as much as,
although a signi®cant lowering of surface free energy was
achieved, the measured value never reached that displayed
by the pure ¯uorinated homopolymer. It is also established
that ¯uorination of a monomer is not suf®cient to ensure
surface segregation and indicates that an interplay of struc-
ture, volatility and compatibility with the commodity poly-
mer substrate determines the outcome of this approach to
low cost ¯uorocarbon surface effects.
3] N.L. Jarvis, R.B. Fox, W.A. Zisman, Adv. Chem. Ser. 43 (1964)
317.
4] A. Hult, B. Ranby, A.C.S. Polym. Preprints 25 (1984) 329.
[5] G.L. Gaines, Macromolecules 14 (1981) 208.
[
6] D.T. Clark, J. Peeling, J.M. O'Malley, J. Polym. Sci., Polym. Chem.
Ed. 14 (1976) 543.
[
[
[
7] H.R. Thomas, J.J. O'Malley, Macromolecules 12 (1979) 323.
8] D.K. Owens, R.C. Wendt, J. Appl. Polym. Sci. 13 (1969) 1741.
9] P. Tarrant, G.G. Allison, K.P. Barthold, Fluor. Chem. Rev. 5 (1971)
77.
[
[
10] S.A. Reines, U.S. Patent 3,707,483 (1972).
11] J.G. O'Rear, J.R. Griffith, A.C.S. Div. Org. Coat. Plast. Tech. Lett.
3
3 (1973) 657.
12] J.G. O'Rear, J.R. Griffith, S.A. Reines, J. Paint Tech. 43 (1971)
13.
[
[
1
13] F.R. Damont, L.M. Sharp, M. Schonhorn, J. Polym. Sci., Polymer
Lett. B3 (1963) 1021.
14] N.O. Brace, J. Org. Chem. 27 (1962) 3033.
Acknowledgements
[
[
[
15] W.J. Feast, S. Matuszczak, Eur. Polym. J. 25 (1989) 813±823.
16] J.V. Crivello, J.H.W. Lam, J. Polym. Sci., Polym. Symp. 56 (1976)
It is a pleasure for one of us (WJF) to acknowledge the
generous friendship and encouragement given by Paul
Tarrant over many years. We thank the Science and Engi-
383.