Synthesis and properties of new fluorinated polymers bearing pendant imidazole groups for fuel cell membranes operating over a broad relative humidity range

Article abstract of DOI:10.1002/pola.23763

New alternating copolymers comprising a chlorotrifluorinated backbone and midazole-terminated pendant ethylene oxide groups have been prepared with a view to their use as a component of proton-conducting membranes in polymer electrolyte fuel cells. A vinyl ether containing an imidazole (Imi) function protected by a benzyl group (BVI) was first synthesized in a three-step reaction. It was then copolymerlzed in solution with chlorotrifluoroethylene (CTFE) by conventional radical copolymerization leading to alternating poly(BVI-altCTFE) copolymers in good yields. Deprotection of the benzyl group under hydrogen produced a chlorotrifluorinated poly (Imi-alt-CTFE) copolymer. The polymer was subsequently used to form blend membranes with sulfonated poly(ether ether ke-tone) (sPEEK). The conductivity of blend membranes of poly (Imi-alt-CTFE) with sPEEK lies in the range of 4-10 mS cm^-1 at 40-70 °C and, for blend membranes rich in poly(Imi-alt-CTFE), is little dependent on relative humidity between 30 and 100%. It Is surmised that the polymer and membrane composition favor microstructural phase separation into chlorotrifluorlnated polymer backbone domains and regions in which imidazole groups are clustered.

Full text of DOI:10.1002/pola.23763

Synthesis and Properties of New Fluorinated Polymers Bearing Pendant
Imidazole Groups for Fuel Cell Membranes Operating Over a Broad
Relative Humidity Range
1
GUILLAUME FRUTSAERT,1,2 LOUIS DELON,1,2 GHISLAIN DAVID,1 BRUNO AMEDURI, DEBORAH J. JONES,2
´
2
XAVIER GLIPA,3 JACQUES ROZIERE
`
1Laboratoire Inge´nierie et Architecture Macromole´culaire, Institut Charles Gerhardt UMR CNRS 5253, Ecole Nationale Supe´rieure
de Chimie de Montpellier, 8 rue de l’Ecole Normale, 34296 Montpellier Cedex 5, France
2Laboratoire Agre´gats, Interfaces et Mate´riaux pour l’Energie, Institut Charles Gerhardt UMR CNRS 5253,
Universite´ Montpellier II, Place E. Bataillon, 34095 Montpellier Cedex 5, France
3PSA Peugeot-Citroen, Centre Technique de Carrie`res sous Poissy, 212 boulevard Pelletier, 78307 Carrie`res sous Poissy
¨
Received 4 September 2009; accepted 25 September 2009
DOI: 10.1002/pola.23763
Published online in Wiley InterScience (www.interscience.wiley.com).
ABSTRACT: New alternating copolymers comprising a chlorotri-
fluorinated backbone and imidazole-terminated pendant ethyl-
ene oxide groups have been prepared with a view to their use
as a component of proton-conducting membranes in polymer
electrolyte fuel cells. A vinyl ether containing an imidazole
(Imi) function protected by a benzyl group (BVI) was first syn-
thesized in a three-step reaction. It was then copolymerized in
solution with chlorotrifluoroethylene (CTFE) by conventional
radical copolymerization leading to alternating poly(BVI-alt-
CTFE) copolymers in good yields. Deprotection of the benzyl
group under hydrogen produced a chlorotrifluorinated poly
(Imi-alt-CTFE) copolymer. The polymer was subsequently used
to form blend membranes with sulfonated poly(ether ether ke-
tone) (sPEEK). The conductivity of blend membranes of poly
(Imi-alt-CTFE) with sPEEK lies in the range of 4–10 mS cmꢀ1 at
40–70 ꢁC and, for blend membranes rich in poly(Imi-alt-CTFE),
is little dependent on relative humidity between 30 and 100%.
It is surmised that the polymer and membrane composition
favor microstructural phase separation into chlorotrifluorinated
polymer backbone domains and regions in which imidazole
C
V
groups are clustered.
2009 Wiley Periodicals, Inc. J Polym
Sci Part A: Polym Chem 48: 223–231, 2010
KEYWORDS: alternating copolymers; blend membranes; blends;
copolymerization; fluoropolymers; imidazole; membranes;
polymer electrolyte fuel cell; sulfonated PEEK
INTRODUCTION Fluorinated polymers1–4 exhibit remarkable
properties, such as chemical, thermal, and electrical stabil-
ity,5 inertness to acids, solvents, and oils, low dielectric con-
stant, low refractive index, low or nonflammability, high re-
sistance to aging and oxidation, and low surface tension.
These polymers show low friction coefficient and relative
permittivity, and they are strongly hydrophobic. For applica-
tion in a polymer electrolyte fuel cell, they must be function-
alized,6 and the choice of the protogenic group plays an im-
portant role in the proton conductivity of fuel cell separator
materials. Up to now, the materials used for proton exchange
membranes in fuel cell applications have relied on the pres-
ence of water as proton carrier. Although practicable at
lower temperatures, water management at temperatures
higher than 100 ꢁC has a significant impact on the viability
of the membrane electrode assembly and on overall opera-
tion of the system.
should be amphoteric, exhibiting proton donor and acceptor
properties, and showing a high degree of self-dissociation. It
should also have a high dielectric constant to enhance the
charge separation and be stable under fuel cell operation
conditions. The relevance of these concepts has been ele-
gantly shown by Kreuer et al.,7 in particular, in pioneering
work on imidazole–sulfonated poly(ether ether ketone)
(sPEEK)-type ionomers.
Commercial perfluorosulfonic acid (PFSA) polymers, such as
R
R
R
R
R
NafionV, FlemionV, AquavionV, BAM-3GV, FumionV, Gore
R
R
R
SelectV, DowV, 3MV, or Aciplex, are reference and emerging
advanced membranes for proton exchange membrane fuel
cells, being characterized by high proton conductivity and
chemical stability.8–10 Their conductivity is dependent on
ambient relative humidity, and their glass transition temper-
ꢁ
ature limits their use to <130 C. In PFSA-type membranes,
as in other sulfonic acid-functionalized polymers, such as
PEEK, polybenzimidazole, or polysulfone,11–14 the proton is
transferred by water, and, in general, a high level of
Over the last decade, new concepts have evolved involving
alternative proton carriers. The ideal protogenic group
Correspondence to: G. David (E-mail: ghislain.david@enscm.fr)
C
V
Journal of Polymer Science: Part A: Polymer Chemistry, Vol. 48, 223–231 (2010) 2009 Wiley Periodicals, Inc.
FLUORINATED POLYMERS BEARING IMIDAZOLE GROUPS, FRUTSAERT ET AL.
223

JOURNAL OF POLYMER SCIENCE: PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
hydration is required, being around 7 (PBI-type) to 25
(PFSA-type) water molecules per sulfonic acid function. This
can lead to excessive membrane swelling at high tempera-
ture, which is detrimental to membrane mechanical proper-
ties. Furthermore, maintaining a high level of hydration at
high temperatures requires high pressure operation. Phos-
phonic acid-based systems15 may offer the possibility of
adequate proton conductivity over a broad temperature
range, although the propensity to condensation by phos-
phonic acids at high temperature is greater than that of sul-
fonic acids, ꢁand the involvement of water in proton transfer
at T < 100 C is not avoided.
m are abbreviations for singlet, doublet, triplet, quintet, and
multiplet, respectively. Coupling constants and chemical
shifts are given in Hertz (Hz) and ppm. The experimental
conditions to record 1H and 19F NMR spectra were as fol-
lows: flip angle 90ꢁ, 30ꢁ, and 30ꢁ; acquisition time: 4.5, 2.9,
and 0.7 s; pulse delay: 2, 2, and 5 s; number of scans: 16,
12,000, and 64, and a pulse width of 5 ls for 19F NMR.
Thermogravimetric analyses (TGA) were performed with a
TGA/SDTA 851 thermobalance from Mettler DAL 75965 and
Lauda RC6 CS cryostat apparatus in static air with a heating
rate of 10 ꢁC minꢀ1 from room temperature up to a maxi-
mum of 600 ꢁC. Differential scanning calorimetry measure-
ments were conducted using a TA2920 analyzer from TA
Instruments DA 73085, an RCS DA cooler and Sartorius MC5
balance. Sꢀca1ns were recorded at a heating/cooling rate of
Another approach toward high-temperature proton-conduct-
ing membranes is the substitution of water by N-hetero-
cycles,16–21 such as imidazole, pyrazole, and benzimidazole.
The proton conductivity is expected to be less dependent on
water content, because proton transfer between heterocycles
is an intrinsic property of these systems. This field was
explored by Kreuer et al., and a number of model com-
pounds were described, in which heterocycles were linked
through ethylene oxide16,17 or siloxane18 oligomer or poly-
mer backbones, and also by Jannasch and coworkers.19 How-
ever, in none of the heterocycle systems prepared to date,
nanophase separation could occur as for amphiphilic sys-
tems, partly because the concentration of heterocycle units
has been low, but also because all polymers prepared have
been built around hydrocarbon (or siloxane) backbones.
ꢁ
ꢁ
20 C min from 0 to 100 C.
Through-plane AC conductivity measurements were per-
formed by impedance spectroscopy using a HP4192 imped-
ance meter in the range 5 Hz–13 MHz using an oscillating
voltage of 100 mV. The locally designed conductivity meas-
urements cell allowed the control of temperature under full
humidification conditions. Measurements were performed on
membrane of thickness 20 lm placed between two stainless
steel electrodes of surface area 3 mm2. In-plane (four elec-
trode) resistance measurements were also made over the
ꢁ
temperature range 50–90 C and relative humidity 10–100%
values, and the conductivity calculated from the relationship
r ¼ L/RdW, where L is the distance between electrodes, d
and W are the thickness and width of the sample strip,
respectively. In all cases, R was derived from the low inter-
cept of the high-frequency semicircle on a complex imped-
ance plane with the Re (Z) axis.
To our knowledge, no fluorinated polymer containing imidaz-
ole groups has been reported in the literature. It would be
of interest to find new copolymers bearing fluorinated base
units in the backbone and imidazole side functions to ensure
thermal and mechanical stabilities and proton conductivity.
This is the purpose of this article, which describes the syn-
thesis of a novel monomer containing the imidazole function
and its radical copolymerization with a fluoro-olefin to give a
polymer with grafted imidazole along the fluoropolymer
backbone. The resulting copolymers have been processed
into blend membranes incorporating sPEEK, the properties
of which have been investigated.
Synthesis of 2-Benzyl-1-(dioxa-3,6-oct-7-ene)imidazole
Monomer (3)
Synthesis of 1-Benzylimidazole (1)
In a 1-L three-necked round-bottomed flask equipped with a
condenser and a magnetic stirrer were introduced 44.0 g
(0.65 mol) of imidazole, 100.0 g (0.78 mol) of benzyl chlo-
ride, 60.0 g (1.07 mol) of sodium hydroxide, and 700ꢁmL of
tetrahydrofuran (THF). The solution was refluxed (70 C) for
120 h, cooled to room temperature, and filtered to eliminate
the excess of sodium hydroxide. After the solvent was
removed by evaporation, the reaction mixture was diluted
with 500 mL diethyl ether, extracted with water, dried with
anhydrous Na2SO4, and finally diethyl ether was evaporated
off. The desired product was obtained with more than 95%
yield and was purified by recrystallization from toluene.
EXPERIMENTAL
Materials
Imidazole, benzyl chloride, potassium hydroxide, paraformal-
dehyde, chloroethylvinylether, and benzyltriethylammonium
chloride were provided by Aldrich, and chlorotrifluoroethy-
lene (CTFE) and 1,1,1,3,3-pentafluorobutane were a gift from
Solvay S.A. Tertiobutyl peroxypivalate was kindly given by
Akzo Nobel. sPEEK was prepared from PEEK by electrophilic
substitution in sulfuric acid under the controlled conditions
described previously.22
23
ꢁ
Tm ¼ 72.7 C.
1H NMR (acetone-d6, ppm) d: 7.70 (N¼¼CHAN, 1H); 7.30
(C6H5, 5H); 7.12 (m, CAHimidazole, 1H); 7.00 (m, CAHimidazole
,
Characterization
1H); 5.24 (CH2, 2H).
1H and 19F nuclear magnetic resonance (NMR) spectra were
recorded at room temperature on Bruker AC 400 instrument,
using deuterated chloroform or acetone as the solvent and
Synthesis of 1-Benzyl-2-(hydroxymethyl)imidazole (2)
A total of 50.0 g (0.312 mol) of 1-benzylimidazole (1) was
introduced in a 300 mL autoclave, with 15.0 g (1.5 mol) of
paraformaldehyde dissolved in 150 mL of 1,4-dioxane. After
1
TMS and CFCl3 as the references for H and 19F NMR nuclei.
In the figures and discussion later, the letters s, d, t, q, and
224
INTERSCIENCE.WILEY.COM/JOURNAL/JPOLA

ARTICLE
ꢁ
the autoclave was closed, the solution was heated at 135 C
oil was dissolved in few milliliters of dichloromethane, and
an orange-brown powder (yield 40 wt %) precipitated with
600 mL of diethylether.
for 20 h, and the pressure reached 7 bar. After reaction, the
solvent was removed under vacuum and the black oil
obtained was dissolved in 400 mL of dichloromethane. The
solution was washed three times with 100 mL of water, the
organic phase was dried with anhydrous Na2SO4, and
dichloromethane was removed by evaporation. The resulting
product was held under reduced pressure for 2 days and
recrystallized from ethyl acetate (100 mL) three times. This
procedure led to 1-benzyl-2-hydroxymethylimidazole (white
1H NMR (acetone-d6, CD3OD ppm) d: 7.28 (C6H5, 5H); 6.91
(NACH¼¼CHAN, 2H); 5.39 (NCH2APh, 2H); 4.52 (imidazole-
CH2AO, 2H); 4.10 (CH2ACHORACF2, 1H); 3.45–4.10 (OACH2
ACH2AO, 4H); 2.31–3.29 (CH2ACHORACF2, 2H).
19F
NMR
(acetone-d6,
CD3OD
ppm)
d:
ꢀ109
(ACF2ACFClACH2, 1F); ꢀ119 (CHOACF2A CFClA, 2F); ꢀ130
(ACF2ACFClACFClACF2, 2H). Tg ¼ 43 ꢁC. Tdec 10 ¼ 226 ꢁC.
24
ꢁ
crystals) with 50% yield. Tm ¼ 154.9 C.
1H NMR (CDCl3, ppm) d: 7.30 (C6H5, 5H); 7.12 (CAHimidazole
,
%C: 54.03 (calculated for alternating copolymer: 54.43); %F:
15.65 (calculated: 15.21); %Cl: 10.60 (calculated: 9.46); %N:
6.44 (calculated: 7.47).
1H); 7.00 (CAHimidazole, 1H); 5.24 (CH2, 2H); 4.61 (CH2OH,
2H); 6.02 (OH).
Synthesis of 2-Benzyl-1-(dioxa-3,6-oct-7-ene)imidazole (3)
In a 250-mL two-necked round-bottomed flask equipped
with a condenser and a mechanical stirrer, 7.5 g (38 mmol)
of 1-benzyl-2-(hydroxymethyl)imidazole and 22 mL of tolu-
ene were introduced. Then, 100 mL of aqueous sodium hy-
droxide (20 mol Lꢀ1) and 0.86 g (3.8 mmol) of benzyltrie-
thylammonium chloride were added. Finally, 8.92 g (74.9
mmol) of 2-chloroethylvinylether (CEVE) was added, and the
mixture was heated at 90 ꢁC with a vigorous stirring. The
disappearance of ClACH2ACH2AOACH¼¼CH2 was monitored
by 1H NMR spectroscopy over a period of 24 h. Then, the
solvent was evaporated under vacuum and the resulting
brown liquid was dissolved in 300 mL of diethylether. The
solution was washed several times with water until pH
reached 7, and the organic phase was dried with anhydrous
Na2SO4 and evaporated. A total of 9.0 g of an orange oil was
obtained corresponding to 87% yield of 2-benzyl-1-(dioxa-
3,6-oct-7-ene)imidazole (BVI).
Deprotection of Imidazole in Poly(BVI-alt-CTFE)
Copolymer by Catalytic Hydrogenation
Using the same autoclave as mentioned earlier, 100 mL of
acetic acid, 5.01 g of poly(BVI-alt-CTFE) copolymer, and 0.49 g
(3.5 mmol) of palladium dihydroxide were introduced. The
autoclave was then closed and the system was pressurized
ꢁ
with 32 bar of hydrogen. It was heated at 70 C for 72 h, and
the pressure was maintained by addition of hydrogen each day
for 6 days. The product mixture was filtered over zeolite and
the solvent was removed by evaporation. The dry solid was
then dissolved in 50 mL of an aqueous solution of Na2CO3
(5%) under reflux. The mixture was filtered, washed three
times with 50 mL of distilled water, and dried under vacuum.
Finally, the solid product was dissolved in 10 mL of methanol
and precipitated with 400 mL of diethylether. A total of 2.05 g
of an orange powder was obtained with a deprotection yield
of about 75%.
1H NMR (MeOH-d6, CD3OD ppm) d: 7.28 (C6H5); 6.91
(NACH¼¼CHAN); 5.39 (NCH2APh); 4.52 (imidazole-CH2AO);
4.10 (CH2ACHOR ACF2); 3.45–4.10 (OACH2ACH2AO);
2.31–3.29 (CH2ACHORACF2).
1H NMR (acetone-d6, ppm) d: 7.42 (C6H5, 5H); 7.12
(CH¼¼CHANC,1H); 6.92 (NBz-CH¼¼CH, 1H); 6.49 (dd, CH2CHO,
1H); 5.30 (CH2, 2H); 4.61 (CH2O, 2H); 4.18 (CH2CHO, 1H);
3.99 (CH2CHO, 1H); 3.79 (CH2CHOCH2CH2OA, 2H); 3.67
(CH2CHOCH2CH2OA, 2H).
19F NMR (MeOH-d6, CD3OD ppm) d: ꢀ109 (ACF2ACFClACH2,
1F);
ꢀ119
(CHOACF2A
CFClA,
2F);
ꢀ130
(ACF2ACFClACFClACF2, 2H). Tg 38 ꢁC.
Synthesis of Poly(BVI-alt-CTFE) Copolymer
The radical copolymerization of BVI with CTFE was carried
out in a 0.1-L HC-276 Hastelloy autoclave equipped with a
rupture disk, inlet and outlet valves, a manometer, a ther-
mometer, and a magnetic stirrer. This vessel was left closed
for 20 min and purged with 15 bar of nitrogen pressure to
prevent any leak, before purging through two nitrogen-vac-
uum cycles to limit the presence of oxygen. A total of 10.0 g
(38.7 mmol) of liquid monomer BVI, 0.210 g (1.6 mmol) of
tert-butylperoxypivalate, and 60 mL of 1,1,1,3,3-pentafluoro-
butane were introduced into the autoclave. Next, the auto-
clave was cooled to ꢀ80 ꢁC in an acetone/liquid nitrogen
bath and 10.1 g (86.7 mmol) of CTFE was added. The auto-
Membrane Preparation
Poly(Imi-alt-CTFE) copolymer (30 wt %) was added to N-
methylpyrrolidinone (NMP) giving a suspension, to which
aqueous lithium chloride (3 mol Lꢀ1) was added. The same
preparation was carried out with sPEEK (15 wt %), and the
copolymer suspension was slowly added in sPEEK suspen-
sion ꢁunder strong stirring. The suspensions were heated at
120 C to yield a dark brown solution that was treated with
ultrasound for 1 h to increase the melting of polymer chains.
The solution was filtered to separate any insoluble material,
and water was evaporated by heating under vacuum. It was
cast on to a heated glass plate with a doctor blade and
transferred to a vacuum oven, where solvents were removed
by heating at 110 ꢁC for 10 h. The resultant film was
removed from the glass support either by immersion in
water or in aqueous hydrochloric acid (1 mol Lꢀ1), when the
acid form membrane could be recovered directly. Any
ꢁ
clave was then heated to 75 C and the pressure reached 15
bar. After 14 h of reaction, the pressure dropped to 12 bar,
and the autoclave was cooledꢁ to room temperature.
Unreacted CTFE was released at 0 C, and it was determined
by weighing that 49% had been converted. The solvent was
evaporated under vacuum, and the resulting black viscous
FLUORINATED POLYMERS BEARING IMIDAZOLE GROUPS, FRUTSAERT ET AL.
225

JOURNAL OF POLYMER SCIENCE: PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
SCHEME 1 Synthesis of poly(BVI-
alt-CTFE) copolymer by direct
copolymerization and imidazole
deprotection. PTC, CEVE, TBPPI,
and CTFE represent phase trans-
fer catalysis, 2-chloroethyl vinyl
ether, tertiobutyl peroxypivalate,
and chlorotrifluoroethylene, res-
pectively
residual solvent and lithium ions [13] were eliminated by
stitution of 2 with CEVE. This reaction was carried out by
phase-transfer catalysis and led to desired monomer 3 in
87% yield. This monomer was also characterized by 1H
NMR, the spectrum of which (Fig. 1) shows the vinyl protons
centered at 4.04 ppm for terminal CH2 and 6.49 ppm for the
adjacent CH proton (CH2¼¼CH), the methylene groups adja-
cent to the oxygen atoms giving signals at 3.79 and 3.67
ppm. The overall yield of 3 was 44% relative to starting
imidazole.
washing the membrane in HCl (1 mol Lꢀ1) at 100 C for 2 h,
ꢁ
ꢁ
then in deionized water at 100 C for 2 h.
RESULTS AND DISCUSSION
The poly(BVI-alt-CTFE) copolymer was prepared according
to a five-step reaction (Scheme 1). This novel copolymer first
requires the synthesis of the nonfluorinated imidazole-termi-
nated vinylic ether monomer (three steps), that is then
directly copolymerized with CTFE. The deprotection of imid-
azole groups by catalytic hydrogenation in a final step leads
to the desired fluorinated copolymers bearing pendant imid-
azole functions.
Direct Radical Copolymerization of BVI with CTFE
The proton conductivity being expected here, as in other
polymer electrolytes, to depend on the number of charge
carriers, it was an important aim here to incorporate a suffi-
cient number of imidazole-containing groups within the final
copolymer. It was known from previous work that the radi-
cal copolymerization of vinyl ether (electron-donating mono-
mer) with CTFE (electron-withdrawing monomer) leads to
alternating copolymers.4 Radical copolymerization of 3 with
CTFE was carried out using various sets of experimental con-
ditions. These reactions were performed in solution with dif-
ferent initial molar ratios C0 {C0 ¼ [In]0/([CTFE]0 þ [BVI]0)}.
The results are summarized in Table 1. In initial experi-
ments, reaction was carried out with a C0 of 1% using differ-
ent solvents. This first set of experiments clearly demon-
strated that C4F5H5 only is appropriate as solvent and good
yields (about 50%) were achieved. Indeed, experiments were
repeated in acetonitrile instead but only low yields could be
reached. This may be explained by higher solubility of the
fluorinated monomer as well as the corresponding copoly-
mers. Indeed, the concentration of CTFE incorporated into
the copolymers always remained higher than that of BVI.
However, whatever the experimental parameters used, alter-
nating copolymers were formed having near to 50% of BVI
units. However, a few excess in CTFE conversion was
observed, characterized by CTFE-CTFE diades at the 19F
NMR analysis (signal centered at 127 ppm). Entry 4 con-
firms this observation and also shows that the use of a
greater amount of BVI entails a drop in monomer conver-
sion. The system pressure also seems to be an important pa-
rameter too, Entry 2 is the experiment using the most quan-
tity of fluorinated gas (higher pressure), which results to a
Synthesis of 2-Benzyl-1-(dioxa-3,6-oct-7-ene)imidazole
Monomer 3
The novel imidazole-containing vinyl-ether 3 was synthe-
sized in a three-step reaction (Scheme 1). Imidazole was ini-
tially protected with a benzyl group, then condensed with
paraformaldehyde, and the resulting hydroxymethyl-
protected imidazole submitted to nucleophilic substitution
with commercially available CEVE. The first step of protec-
tion of the imidazole group led to quantitative yield of prod-
uct 1, which was purified by recrystallization from toluene.
Protection of imidazole was required to avoid any grafting
onto nitrogen during the substitution step and then any
transfer to the acid proton during radical copolymerization.
1
The H NMR spectrum of compound 1 shows a characteristic
singlet centered at 5.24 ppm assigned to the methylene of
the benzyl group. A multiplet centered at 7.30 ppm corre-
sponds to the aromatic protons of the phenylene group.
Finally, two signals attributed to the imidazole protons are
distinguished at 7.00 and 7.12 ppm.
The second step is the condensation25 of molecule 1 with a
large excess of paraformaldehyde. This hydroxymethylation
leads to compound 2 bearing a hydroxymethyl group on the
position adjacent to the two imidazole nitrogen atoms. Com-
pound 2 was purified by recrystallization from ethyl acetate
in 50% yield. The 1H NMR spectrum was similar to that of
intermediate 1 with an additional signal at 4.61 ppm
assigned to CH2AOH. The last step was the nucleophilic sub-
226
INTERSCIENCE.WILEY.COM/JOURNAL/JPOLA

ARTICLE
FIGURE 1 1H NMR spectrum of 1-benzyl-2-(vinyloxyethyloxy-
methyl)imidazole (BVI) in acetone-d6 (presence of residual
alcohol).
FIGURE 2 1H NMR spectrum of poly(BVI-alt-CTFE) copolymer
in deuterated DMF.
mass yield higher than all the other experiments. The 1H
NMR spectrum of the sample of entry 1 is given in Figure 2.
The copolymers precipitate during the course of copolymer-
ization, which may not be favorable to reach high-molecular-
weight polymer. The determination of polymer molecular
weight was impeded by the insolubility of the resulting co-
polymer in any organic solvent commonly used for GPC anal-
ysis (DMF, THF) probably due to a chemical affinity between
the polymer and the chromatographic column. A low molecu-
lar weight would seem to be indicated by the Tg value of
about 40 ꢁC, because high-molecular-weight fluoropolymers
usually exhibit higher Tg values.4 TGA (Fig. 3) shows that
weight loss on heatiꢁng in air occurs in two main steps. The
first, at around 230 C, is characteristic of the degradation of
the heterocycle BVI units. After a weight loss of around 50%,
a plateau is observed between 350 and 500 ꢁC, and there-
after the fluorinated units begin to degrade.
of an a-aminated acid) and needs more drastic conditions.
The literature26–28 describes different methods for removal
of benzyl groups supported on the nitrogen atom in position
1 of imidazole. The most used method is catalytic hydrogen-
ation in the presence of a palladium catalyst. Different palla-
dium-based catalysts can be used for this reaction with vari-
ous efficiencies: Pd/C (5%) < Pd/C (10%) < Pd(OH)2/C
(15–20%). Various parameters influence the result of the
reaction such as temperature or hydrogen pressure. This
reaction is often carried out at room temperature for ali-
phatic amines but, generally, higher temperature is required
in the case of heterocycles (50–80 ꢁC). (The hydrogen pres-
sure is most of the time higher than 30 bar for nitrogen-con-
taining heterocycles). The nature of the solvent is also cru-
cial, alcohols being good solvent media for hydrogenation,
whereas acidification of the medium (HCl or CH3COOH) is
known to favor benzyl group reduction. A series of experi-
ments were performed so as to identify the conditions allow-
ing the highest yield of deprotected copolymer (Table 2).
Experiments 5–8 have clearly failed to deprotect the imidaz-
ole group, at all reaction temperatures, solvents, and hydro-
gen partial pressures used. Methanol is inappropriate as a
Deprotection of the Imidazole Group
Removal of the protecting benzyl group is required to regen-
erate the imide group on imidazole. Deprotection by hydro-
genation is more difficult on a nitrogen-containing hetero-
cycle than on an aliphatic amine (e.g., on the amine function
TABLE 1 Radical Copolymerization of BVI (0.5 mol L21) with Chlorotrifluorethylene at 75 8C Initiated by
tert-Butylperoxypivalate (In)
d
e
e
Mole Percent of BVI
in the Feed (%)
Mass
Yieldb (%)
BVI Unit in
Copolymerc (%)
Tg
T10
T50
a
Entry
C0 (%)
(ꢁC)
(ꢁC)
(ꢁC)
1
2
3
4
a
30
20
30
50
1.0
2.0
0.5
1.0
45
50
40
31
48
47
48
47
43
42
39
42
226
217
228
224
362
345
359
347
d
C0 ¼ [In]0/([BVI]0 þ [CTFE]0).
Tg: glass transition determined by DSC.
T10 and T50: 10 and 50% decomposition, respectively, determined by
TGA, under air.
b
c
e
m(copolymer)expt/m(copolymer)calculated max.
Calculated from elemental analyses using nitrogen values.
FLUORINATED POLYMERS BEARING IMIDAZOLE GROUPS, FRUTSAERT ET AL.
227

JOURNAL OF POLYMER SCIENCE: PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
Finally, assessment of the solubility of the copolymers in var-
ious solvents showed that the unprotected copolymers are
soluble in solvents such as DMF, DMSO, or NMP after addi-
tion of a lithium salt.
Properties of Poly(Imi-alt-CTFE)sPEEK Blend Membranes
Charge defects are needed for proton conduction in any sys-
tem based on nitrogen-containing heterocycles. In this case,
an increase in charge carrier concentration was brought
about by blending the above poly(Imi-alt-CTFE) copolymer
with sPEEK.22 Membrane compositions were chosen to have
different proportions of poly(Imi-alt-CTFE) copolymer to
sPEEK, one membrane series being richer in poly(Imi-alt-
CTFE), a second richer in sPEEK. Mass ratios were, respec-
tively, 70:30 and 30:70 in membranes ImiCTFE-sPEEK-73
and ImiCTFE-sPEEK-37, corresponding to mole ratios of
imidazole to sulfonic acid groups of 3.00 and 0.77, respec-
tively (concentrations of sulfonic acid groups of 0.52 and
0.91 meq gꢀ1).
FIGURE 3 Thermogravimetric analysis of protected (a) and
deprotected (b) copolymer poly(Imi-alt-CTFE) and of blend
membrane ImiCTFE-sPEEK-73 (c). Heating rate 10 ꢁC minꢀ1
in air.
solvent, with a poor yield, even after 2 days using a Pd/C
catalyst. When the reaction time and the pressure were
increased and acetic acid used as a solvent (Experiment 9),
deprotection occurred in 50% yield. The use of Pd(OH)2 at
higher hydrogen pressure and higher temperature (Experi-
ment 10) led to 74% deprotection of imidazole, as deter-
TGA of the ImiCTFE-sPEEK-73 blend membrane (Fig. 3)
shows that there is no loss of weight in air until >250 ꢁC.
This is higher than for the poly(Imi-alt-CTFE) copolymer, and
the thermal stability has thus been improved by blending,
probably due to proton transfer from the sulfonic acid
groups to imidazole and concomitant ionic crosslinking. Ob-
servation by SEM coupled with EDX analysis indicates the
membrane composition to be homogenous with true blend
formation between the s-PEEK and poly(Imi-alt-CTFE)
copolymer components (Fig. 4).
1
mined from the H NMR analysis of the final product. Depro-
tection yield was calculated using integration as follows:
I5:39
1 ꢀ
;
I4:52
The proton conductivity at 65 ꢁC of the ImiCTFE-sPEEK-73
and ImiCTFE-sPEEK-37 membranes is shown in Figure 5 in
the relative humidity range 10–90%. Interestingly, the two
membranes show rather distinct behavior. The conductivity
of membrane ImiCTFE-sPEEK-37 (higher proportion of
sPEEK) has a strong dependence on RH, as is observed for
where I5.39 and I4.52 represent the integrals of the signals
centered at 5.39 and 4.52 ppm in the 1H NMR spectra,
respectively. The resonance corresponding to the methylene
protons of the benzyl group centered at 5.39 ppm disap-
peared, whereas a second signal appears at 4.52 ppm corre-
sponding to the methylene protons adjacent to the imidazole
function. The resulting copolymer (5) is characterized by a
chlorotrifluorinated backbone, in which CTFE units alternate
with a pendant imidazole group linked to the backbone
through a flexible ethylene oxide spacer; this copolymer is
referred to as poly(Imi-alt-CTFE).
dense membranes of sPEEK alone.22 At 65 C, the conductiv-
ꢁ
ity value increases by six orders of magnitude from 5 ꢂ
10ꢀ5 mS cmꢀ1 at 10% RH to 15 mS cmꢀ1 at 90% RH. In
contrast, the conductivity of membrane ImiCTFE-sPEEK-73
(higher proportion of ImiCTFE) is significantly less
TABLE 2 Catalytic Removal of Benzyl Groups on Poly(BVI-alt-CTFE) Copolymer under H2 Pressure
Temp.
(ꢁC)
Time
(h)
Yield
(%)
Exp.
Catalyst
P (bar)
Solvent
5
6
Pd/C (10%)
55
60
70
70
70
70
90
90
70
1
MeOH
40
48
–
Pd/C (10%)
10
MeOH
–
7
Pd/C (10%)
20
MeOH
48
–
8
Pd/C (10%)
25–30
34
MeOH
48
–
9
Pd/C (10%)
CH3CO2H
CH3CO2H
CH3CO2H
CH3CO2H
CH3CO2H
120
120
120
120
120
50
74
75
77
75
10
11
12
13
Pd(OH)2 (10%)
Pd(OH)2 (10%)
Pd(OH)2 (40%)
Pd(OH)2 (40%)
34
31
34
34
228
INTERSCIENCE.WILEY.COM/JOURNAL/JPOLA

ARTICLE
FIGURE 4 SEM micrograph of the
surface (a) and edge (b) of blend
membrane ImiCTFEsPEEK73 and
atomic ratio of fluorine to sulfur
from EDX analysis across the mem-
brane thickness (c). [Color figure
can be viewed in the online issue,
which is now available at www.
interscience.wiley.com.]
dependent on the relative humidity, and the values at low
RH are higher. Thus, the conductivity at 30% RH is 4 mS
cmꢀ1 and at 50–90% RH is in the range 8–10 mS cmꢀ1. This
is a central result, because it responds at least in part to the
application-driven need for proton-conducting materials hav-
ing low dependence of the conductivity on the degree of am-
bient humidity. The results can be directly related to the pro-
portion of pendant imidazole functionalities in the
membranes, which determines the dominant proton-conduc-
tion mechanism occurring in each sample. Although a mecha-
nistic study is not the objective of this work, it can be
expected that in ImiCTFE-sPEEK-73 (higher concentration of
imidazole) a mechanism that does not depend on membrane
water content is more prevailing than in ImiCTFE-sPEEK-37,
where water-assisted proton transport is predominant.
in Figure 6. At 25 and 100% RH, the conductivity is about
0.5 and 6ꢁ mS cmꢀ1 respectively, between 50 and 70 ꢁC.
Above 70 C, however, the conductivity takes a sharp down-
turn in both cases. These observations were reproducible on
membranes incorporating poly (Imi-alt-CTFE) copolymer
prepared from different batches. As the drop in conductivity
above 70 ꢁC is observed at both low and high relative
humidities, it is considered not to be due to any influence of
water-assisted proton transport, that is, to membrane dehy-
dration. A more likely explanation is that a change in mem-
ꢁ
brane microstructure occurs above 70 C, which is unfavora-
ble to proton transfer and transport. A strong thermal event
ꢁ
was indeed detected for ImiCTFE-s-PEEK73 at 68 C (Fig. 7).
Application of the Fox-Flory equation29 (1/Tg ¼ w1/Tg1
þ
w2/Tg2) gives a calculated Tg value for the blend of 70.6 ꢁC.
On the basis of these observations, we conclude that ꢁmicro-
structure of ImiCTFE-sPEEK-73 is modified above 70 C and
that this new structure creates a barrier to proton transport.
The conductivity of membrane ImiCTFE-sPEEK-73 in the
temperature range 40–90 ꢁC at 25 and 100% RH is shown
FIGURE 5 Dependence of proton conductivity at 65 ꢁC of blend
membranes ImiCTFEsPEEK73 (~) and ImiCTFEsPEEK37 (n) on
relative humidity.
FIGURE 6 Temperature dependence of the proton conductivity
of blend membrane ImiCTFEsPEEK73 at 100 (~) and 25% (~)
relative humidity.
FLUORINATED POLYMERS BEARING IMIDAZOLE GROUPS, FRUTSAERT ET AL.
229

JOURNAL OF POLYMER SCIENCE: PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
FIGURE 7 Differential scanning
calorimetry of blend membrane
ImiCTFEsPEEK73. [Color figure
can be viewed in the online issue,
which is now available at
www.interscience.wiley.com.]
Although no direct information is available, some suggestions
can be put forward to interpret these observations. The dif-
ferent characters of the CTFE backbone of the polymer and
the imidazole-terminated and hydrocarbon-based ethylene
oxide pendant chains are expected to lead to a phase-sepa-
rated microstructure within the cast membranes, in which
imidazole groups will be clustered together into imidazole-
rich domains. The sPEEK component of the blend will also
be associated with these domains, both through nonspecific
interactions between the hydrocarbon chains, but more par-
ticularly through hydrogen bonding between the sulfonic
acid and imidazole groups. This configuration is clearly
higher polymer molecular weight could help to maintain the
low-temperature membrane microstructure and endeavors
continue in this direction.
The recent literature provides several examples of imidazole-
or benzimidazole-functionalized species bound, for example,
to polyoxoethylene spacers16,17 or as functional units on a
hydrocarbon or siloxane polymer chain.18 Although the Imi-
POE-Imi systems have related proton-conduction properties,
such ‘‘model compounds’’ of short chain length have no film-
forming ability. The propensity of the imidazole and sulfonic
acid groups in the present ImiCTFE-sPEEK system to cluster
into confined domains is a property of the membrane, and in
this sense, the properties exhibited by model (molecular) com-
pounds cannot be directly related to those developed by a
polymer membrane system. In other recent work, polymeric
benzimidazole- or imidazole-functionalized polymers have been
developed, the conductivities of which have been described as
in the range of 10ꢀ7 S cmꢀ1 when measured under anhydrous
conditions.19 In comparison with these studies, we have
increased the density of heterocycle moieties in the polymer
chains, while using a flexible pendant group, both factors tend-
ing to encourage clustering of imidazole units in the mem-
brane. In addition, we have introduced an acidic polymer as
proton ‘‘dopant’’ to promote a higher degree of dissociation.
Unlike the phosphoric acid doping of polybenzimidazole,30 or
the polyphosphoric acid doping of imidazole-functionalized
poly(diphenyl phenylene) oxide,31 the sPEEK used in the blend
membranes of this work does not suffer any risk of elution
under conditions when liquid water is present.
ꢁ
favorable to proton transfer and transport. Above 70 C, the
loss in conductivity can be interpreted in terms of a ther-
mally induced rearrangement within the membrane that is
possibly due to disruption of the hydrogen bonded network
as a result of flow of the fluorinated polymer backbone.
Whatever ꢁthe physical origin of the loss of conductivity
above 70 C, it is a reversible phenomenon, because the low
temperature conductivities are almost completely recovered
when the membrane is cooled, Figure 8. It is possible that a
CONCLUSIONS
The poly(Imi-alt-CTFE) copolymers prepared here for the first
time associate a CTFE backbone with pendant imidazole groups
linked through ethylene oxide groups. This composition favors
microstructural phase separation of the polymer backbone
regions giving chemical stability on the one hand, and pendant
group regions providing proton-conduction properties on the
other. It is expected that PEO and imidazole groups form clusters
FIGURE 8 Temperature dependence (60–120 ꢁC) of conductivity
of blend membrane ImiCTFEsPEEK73 at 80% relative humidity
showing reversibility of thermally induced phenomenon at
70 ꢁC.
R
in the same way as the sulfonic acid groups in PFSA NafionV-type
membranes. No weight loss in TGA was observed below 230 ꢁC
for the poly(Imi-alt-CTFE) copolymers and the Tg is about 40 ꢁC.
230
INTERSCIENCE.WILEY.COM/JOURNAL/JPOLA

ARTICLE
14 Hickner, M. A.; Ghassemi, H.; Kim, Y. S.; Einsla, B. R.; McGrath, J. E.
Chem Rev 2004, 104, 4587–4611.
Polymer blend membranes of poly(Imi-alt-CTFE) and sPEEK can
be formulated in various compositions, and all compositions
have high proton conductivity above 70% RH. At lower RH val-
ues, only those membranes with higher poly(Imi-alt-CTFE) con-
tent retain a conductivity of 4–10 mS cmꢀ1, whereas membranes
with a high concentration of sPEEK have lower conductivity at
low RH, even though the concentration of proton charge carriers
is higher. The conductivity decreases above 70 ꢁC, and it is
believed that this is due to a thermally induced structural
rearrangement.
15 Lafitte, B.; Jannasch, P. J Polym Sci Part A: Polym Chem 2006, 45,
269–283.
16 Kreuer, K. D.; Fuchs, A.; Ise, M.; Spaeth, M.; Maier, J. Electrochim
Acta 1998, 43, 1281–1288.
17 Kreuer, K. D. Solid State Ionics 2001, 145, 85–93.
18 Herz, H. G.; Kreuer, K. D.; Maier, J.; Scharfenberger, G.; Schuster, M.
F. H.; Meyer, W. H. Electrochim Acta 2003, 48, 2165–2171.
The authors acknowledge the award of doctoral funding (to G.
Frutsaert) by the CNRS Direction aux Entreprises and Peugeot-
Citroe¨n S.A.
19 (a) Persson, J. C.; Jannasch, P. Macromolecules 2005, 38,
3283–3289; (b) Persson, J. C.; Jannasch, P. Solid State Ionics 2006,
177, 653–660; (c) Persson, J. C.; Josefsson, K.; Jannasch, P. Polymer
2006, 47, 991–1002.
REFERENCES AND NOTES
20 Li, S.; Zhou, Z.; Liu, M.; Li, W.; Ukai, J.; Hase, K.; Nakanishi, M.
Electrochim Acta 2006, 51, 1351–1358.
1 Wall, L. A. Fluoropolymers; Wiley: New York, 1972.
2 Feiring, A. E. In Organofluorine Chemistry: Principles and Commer-
cial Applications; Banks, R. E.; Smart, B. E.; Tatlow, J. C., Eds.; Plenum
Press: New York, 1994; Vol. 15, pp 339–372.
21 Woudenberg, R. C.; Yavuzcetin, O.; Teominen, M. T.; Coughlin, E. B. Solid
State Ionics 2007, 178, 1135–1141.
22 Bauer, B.; Jones, D. J.; Rozie`re, J.; Tchicaya, L.; Alberti, G.; Massi-
nelli, L.; Casciola, M.; Peraio, A.; Besse, S.; Ramunni, E. J New Mater
Electrochem Appl 2000, 3, 93–99.
3 Scheirs, J. Modern Fluoropolymers; Wiley: New York, 1997; pp
1–69.
4 Ame´duri, B.; Boutevin, B. Well-Architectured Fluoropolymers: Syn-
23 Owen, C. P.; Dhananni, S.; Patel, C. H.; Shahid, I.; Amhed, S. Bioorg
Med Chem Lett 2006, 16, 4011–4015.
thesis, Properties and Application; Elsevier: Amsterdam, 2004.
5 Hougham, G.; Davidson, T.; Cassidy, P.; Johns, K. Fluoropolymers 2: Proper-
24 Grimmett, M. R.; Richard, E. L. Aust J Chem 1965, 18, 1855–1861.
ties; Kluwer: New York, 1999.
25 Gamez, P.; Simons, P.; Aromi, G.; Driessen, W. L.; Challa, G.; Reedijk, J.
Appl Catal A 2001, 214, 187–192.
6 (a) Rozie`re, J.; Jones, D. J. Annu Rev Mater Res 2003, 33, 503–
555; (b) Jones, D. J.; Rozie`re, J. Adv Polym Sci 2008, 215, 219–245.
26 Felix, A. M.; Heimer, E. P.; Lambros, T. J.; Tzougraki, C.; Meienhofer, J. J Org
Chem 1978, 43, 4194–4202.
7 (a) Kreuer, K. D.; Dippel, T.; Maier, J. Proc Electrochem Soc 1995,
95, 241–252; (b) Scharfenberger, G.; Meyer, W. H.; Wegner, G.; Schus-
ter, M.; Kreuer, K. D.; Maier, J. Fuel Cells 2006, 6, 237–250.
27 Marciniak, G.; Decolin, D.; Leclerc, G.; Decker, N.; Schwartz, J. J Med
Chem 1988, 31, 2289–2294.
8 Doyle, M.; Rajandran, G. In Handbook of Fuel Cells—Fundamentals, Tech-
nology and Applications; Vielstich, W.; Hubert, A.; Gasteiger, M.; Lamm, A.,
Eds.; Wiley: New York, 2003; Vol. 3, Chapter 3, pp 351–395.
28 Jones, R. G. J Am Chem Soc 1949, 71, 383–391.
29 Flory, P. J.; Fox, T. G. J Polym Sci 1950, 5, 745–752.
9 Arcella, V.; Ghielmi, A.; Tommasi, G. Ann N Y Acad Sci 2003, 984,
226–244.
30 (a) Wainright, J. S.; Wang, J.-T.; Savinell, R. F.; Litt, M.; Moaddel, H.;
Rogers, C. Proc Electrochem Soc 1994, 94, 255–261; (b) Xiao, L.;
Zhang, H.; Scanlon, E.; Ramanathan, L. S.; Cho, E.-W.; Rogers, D.; Apple,
T.; Benicewicz, B. C. Chem Mater 2005, 17, 5328–5333; (c) Glipa, X.;
Bonnet, B.; Mula, B.; Jones, D. J.; Rozie`re, J. J Mater Chem 1999, 9,
3045–3050.
10 Souzy, R.; Ame´duri, B. Prog Polym Sci 2005, 30, 644–687.
11 Jones, D. J.; Rozie`re, J. J Membr Sci 2001, 185, 41–58.
12 Glipa, X.; El Haddad, M.; Jones, D.J.; Rozie`re, J. Solid State Ionics
1997, 97, 323–330.
31 Ionescu, D.; de Heer, M.; Jannsen, G.; Wakker, A. Available at:
13 Pe´ron, J.; Ruiz, E.; Jones, D. J.; Rozie`re, J. J Membr Sci 2008, 314,
247–256.
http://www.ecn.nl/fileadmin/ecn/units/h2sf/pdf/PosterLuzern2005.
pdf.
FLUORINATED POLYMERS BEARING IMIDAZOLE GROUPS, FRUTSAERT ET AL.
231