Copoly(arylene ether nitrile) and copoly(arylene ether sulfone) ionomers with pendant sulfobenzoyl groups for proton conducting fuel cell membranes

Article abstract of DOI:10.1002/pola.24486

Three series of fully aromatic ionomers with naphthalene moieties and pendant sulfobenzoyl side chains were prepared via K₂CO₃ mediated nucleophilic aromatic substitution reactions. The first series consisted of poly(arylene ether)s

Full text of DOI:10.1002/pola.24486

Copoly(arylene ether nitrile) and Copoly(arylene ether sulfone)
Ionomers with Pendant Sulfobenzoyl Groups for Proton
Conducting Fuel Cell Membranes
ELIN PERSSON JUTEMAR, PATRIC JANNASCH
Polymer and Materials Chemistry, Department of Chemistry, Lund University, P.O. Box 124, SE-221 00 Lund, Sweden
Received 30 September 2010; accepted 5 November 2010
DOI: 10.1002/pola.24486
Published online 3 December 2010 in Wiley Online Library (wileyonlinelibrary.com).
ABSTRACT: Three series of fully aromatic ionomers with naph-
thalene moieties and pendant sulfobenzoyl side chains were
prepared via K2CO3 mediated nucleophilic aromatic substitu-
tion reactions. The first series consisted of poly(arylene ether)s
prepared by polycondensations of 2,6-difluoro-20-sulfobenzo-
phenone (DFSBP) and 2,6-dihydroxynaphthalene or 2,7-di-
hydroxynaphthalene (2,7-DHN). In the second series, copoly-
(arylene ether nitrile)s with different ion-exchange capacities
(IECs) were prepared by polycondensations of DFSBP, 2,6-
difluorobenzonitrile (DFBN), and 2,7-DHN. In the third series,
bis(4-fluorophenyl)sulfone was used instead of DFBN to pre-
pare copoly(arylene ether sulfone)s. Thus, all the ionomers had
sulfonic acid units placed in stable positions close to the elec-
tron withdrawing ketone link of the side chains. Mechanically
strong proton-exchange membranes with IECs between 1.1
and 2.3 meq gꢀ1 were cast from dimethylsulfoxide solutions.
High thermal stability was indicted by high degradation
temperatures between 266 and 287 ꢁC (1 ꢁC minꢀ1 under air)
and high glass transition temperatures between 245 and
306 ꢁC, depending on the IEC. The copolymer membranes
reached proton conductivities of 0.3
S
cmꢀ1 under fully
humidified conditions. At IECs above ꢂ1.6 meq gꢀ1, the copol-
ymer membranes reached higher proton conductivities than
R
NafionV in the range between ꢀ20 and 120 ꢁC.
2010 Wiley
C
V
Periodicals, Inc. J Polym Sci Part A: Polym Chem 49: 734–745,
2011
KEYWORDS: ionomers; membranes; polyaromatics; polyconden-
sations; sulfonated polymer electrolytes; proton-exchange
membrane fuel cells
INTRODUCTION The development of polymer electrolyte
membrane fuel cells (PEMFCs) as efficient and environmen-
tally benign power sources is currently given great attention
globally because of increasing environmental concerns and a
wish to move away from the dependence of fossil fuels.1–3
Today, considerable efforts are directed to bring PEMFCs to
various markets. In this context, the development work is
primarily focused on increasing the durability and perform-
ance, reducing the cost, and to expand the operating window
of the PEMFC, for example, to higher operating temperatures.
State-of-the-art perfluorosulfonic acid (PFSA) membranes
thermal, mechanical, and chemical stability. Consequently
over the last few years, a large number of aromatic polymers
have been functionalized with sulfonic acid groups and
evaluated as PEMFC membranes.7–11
Sulfonated aromatic polymers designed for PEMFC mem-
branes usually have a balanced composition of hydrophilic
and hydrophobic segments. When these ionomer membranes
take up water, they normally phase separate into a hydro-
phobic polymer-rich phase domain and
a percolating
network of nanopores containing water. In these nanopores,
the water dissociates the acid units and functions as the pro-
ton solvent to facilitate the conduction. Although it is
obvious that the hydrophilic phase domain is important
for the proton transport, the properties of the membrane
are also highly dependent on the nature of the hydrophobic
phase domain. The hydrophobic component plays the impor-
tant role of maintaining the mechanical strength and dimen-
sional stability of the membrane during PEMFC operation.
One of the main challenges is to optimize the molecular
structure, and, hence, balance the combination of hydropho-
bic and hydrophilic segments, to obtain the overall best
membrane properties.12–14
R
such as NafionV exhibit good mechanical properties, high
oxidative stability and high proton conductivity below
4
ꢁ
90 C. However, at higher temperatures, where considerable
advantages can be reached on the PEMFC system level, the
PFSA membranes typically suffer from dehydration and loss
of proton conductivity.5 Because of these shortcomings, an
extensive worldwide research is ongoing to develop new
alternative ionomers with improved properties.6–10 Aromatic
hydrocarbon polymers, such as poly(arylene ether)s (PAEs),
poly(arylene ether sulfone)s, poly(arylene ether ketone)s,
and poly(phenylene)s are well-known for their excellent
Correspondence to: P. Jannasch (E-mail: patric.jannasch@polymat.lth.se)
C
V
Journal of Polymer Science: Part A: Polymer Chemistry, Vol. 49, 734–745 (2011) 2010 Wiley Periodicals, Inc.
734
WILEYONLINELIBRARY.COM/JOURNAL/JPOLA

ARTICLE
Aromatic ionomers with sulfonic acid units placed along the
polymer backbone have shown to develop quite inefficient
networks of nanopores for proton transport. Consequently,
these membranes characteristically reach lower proton con-
R
ductivity compared with NafionV.15–17 A number of different
ionomer architectures have been developed to enhance the
properties of sulfonated aromatic membranes, including con-
centrating the sulfonic acid groups to specific chain segments
in the chain.12 By using this approach, a more distinct phase
separated membrane morphology is promoted, where the
hydrophobic phase domain may efficiently restrict the water
uptake of the sulfonated hydrophilic phase domain. Thus,
enhanced properties have been reported for ionomers with
sulfonic acid groups located on side chains grafted onto the
polymer backbone, including ionomers with sulfonated aro-
matic,18–22 alifatic,23–25 and aromatic–alifatic26 side chains.
Another successful approach is to locate the sulfonic acid
groups on specific blocks in the polymer backbone, sepa-
rated by nonsulfonated blocks.27–29 These multiblock copoly-
mers have shown higher proton conductivity under reduced
relative humidity (RH) in comparison with homopolymers
and statistical copolymers. However, the ease of preparation
and the larger number of available synthetic routes of the
latter polymer classes explains why these are the membrane
polymers primarily selected for today’s PEMFCs, especially
by the industry.
SCHEME 1 Preparation of the lithium salt of DFSBP.
as water uptake, proton conductivity, and thermal character-
istics were investigated and correlated to the IEC as well as
to differences in the copolymer configuration and
functionality.
EXPERIMENTAL
Materials
DFSBP was prepared by reacting 2,6-difluorophenyllithium
with 2-sulfobenzoic acid cyclic anhydride in THF at ꢀ70 ꢁC,
according to Scheme 1. The product precipitated out in the
reactor and was recrystallized in methanol. Further details
have been described previously.30 Potassium carbonate
(Acroˆs, 99þ%) was dried at 120 ꢁC overnight before use.
2,7-DHN (TCI Europe, >99%) and 2,6-DHN (Alfa Aesar,
98%) were recrystallized from water, and bis(4-fluoro-
phenyl) sulfone (DFDPS, Alfa Aesar, 98þ%) was recrystal-
lized from toluene before being dried in vacuo at 80 ꢁC
overnight. DFBN (Acroˆs, 97%), N,N-dimethylacetamide
(DMAc, Acroˆs, 99%), toluene (Fisher Scientific, HPLC grade),
2-propanol (IPA, Fisher Scientific, HPLC grade), and di-
methylsulfoxide (DMSO, Acroˆs, 99.9þ%) were all used as
received.
We have recently reported on the synthesis and polymeriza-
tion of 2,6-difluoro-20-sulfobenzophenone (DFSBP).30 The
lithium salt of this new monomer is conveniently synthe-
sized in one pot by reacting 2,6-difluorophenyllithium with
2-sulfobenzoic acid cyclic anhydride in tetrahydrofuran (THF)
at ꢀ70 ꢁC, whereafter the product crystallizes out of solution.
More recently, we have reported on the polymerization of
DFSBP with various diols and a dithiol to yield ionomers
consisting of aromatic polymer backbones bearing pendant
sulfobenzoyl side chains.31 Among these homopolymers, a
PAE derived from DFSBP and 2,7-dihydroxynaphthalene (2,7-
DHN) showed an attractive combination of properties,
including a high thermal stability and glass transition tem-
perature (Tg). However, because of its high ion-exchange
capacity (IEC) value, the level of water uptake of membranes
based on this ionomer was too high for practical use as a
proton-exchange membrane.
Homopolymer Synthesis
Two different PAE homopolymers were synthesized by
nucleophilic aromatic substitution reactions of DFSBP with
2,7-DHN and 2,6-DHN, respectively, as outlined in Scheme 2.
The resulting polymers were designated PAE2,7 and PAE2,6,
respectively. Using the preparation of PAE2,6 as an example,
DFSBP (0.8063 g, 2.650 mmol), 2,6-DHN (0.4245 g, 2.650
mmol), and K2CO3 (0.457 g, 3.31 mmol) were added to a
mixture of DMAc (6 mL) and toluene (6 mL) in a two-necked
flask equipped with a magnetic stirrer, a nitrogen inlet, and
a Dean–Stark trap with both a condenser and an outlet fitted
with a calcium chloride filter. The reaction mixture was first
heated to 160 ꢁC for 4 h. After dehydration and removal of
the toluene, the reaction temperature was increased to
175 ꢁC, and kept at this temperature for 1.5 h (PAE2,6) or
until the polymer precipitated, which occurred after 1 h for
PAE2,7. Temperature was then lowered until the polymer
regained solubility, which occurred at 110 ꢁC and 70 ꢁC for
the PAE2,7 and PAE2,6, respectively. This reaction tempera-
ture was kept for 15 h before precipitation of the polymer in
an excess of IPA at room temperature. The precipitates of
the polymers were filtered and washed repeatedly with fresh
IPA and water. After drying, the products were redissolved in
DMAc and the same purification procedure was repeated
from a more dilute solution. Finally, the polymers were dried
In this work, various statistical copolymers bearing sulfoben-
zoyl side chains were prepared to control the IEC and vary
the polymer backbone structure. Thus, DFSBP was copoly-
merized with 2,6-difluorobenzonitrile (DFBN) and 2,7-DHN
to prepare a series of copoly(arylene ether nitrile)s (PAENs).
The introduction of highly polar nitrile groups into sulfo-
nated aromatic ionomers has previously been reported to
increase the interchain molecular interactions and decrease
the water uptake of the membranes.32–35 In a second series
of copolymers, bis(4-fluorophenyl)sulfone was used instead
of DFBN to prepare copoly(arylene ether sulfone)s (PAESs).
In addition, two PAEs were prepared by polycondensations
of DFSBP with 2,6-dihydroxynaphthalene (2,6-DHN) and
2,7-DHN, respectively. Important membrane properties, such
ꢁ
in vacuo at 80 C for 24 h.
PROTON CONDUCTING FUEL CELL MEMBRANES, JUTEMAR AND JANNASCH
735

JOURNAL OF POLYMER SCIENCE: PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
SCHEME 2 Synthetic pathway to the PAE homopolymers via potassium carbonate mediated nucleophilic aromatic substitution
reactions.
Copolymer Synthesis
2.812 mmol), DFBN (0.0978 g, 0.703 mmol), 2,7-DHN
(0.5631 g, 3.515 mmol), and K2CO3 (0.607 g, 4.39 mmol)
were added to a mixture of DMAc (8 mL) and toluene
(8 mL) in a two-necked flask equipped with a magnetic
stirrer, a nitrogen inlet, and a Dean–Stark trap with both a
condenser and an outlet fitted with a calcium chloride filter.
The reaction mixture was first heated to 160 ꢁC for 4 h.
After dehydration and removal of the toluene, the reaction
Eight different PAEN and PAES copolymers were synthesized
by nucleophilic aromatic substitution reactions involving
DFSBP, as outlined in Scheme 3. The copolymers were desig-
nated PAENa and PAESa, where a denoted the molar per-
centage of the sulfonated DFSBP monomer in total feed of
the difluoro monomers. In a typical procedure, using the
preparation of PAEN80 as an example, DFSBP (0.8556 g,
SCHEME 3 Synthetic pathway to the various PAEN and PAES copolymers via potassium carbonate mediated nucleophilic aromatic
substitution reactions. The parameter a designates the value of [DFSBP]/([DFSBP] þ [DFBN]) ꢃ 100% or [DFSBP]/([DFSBP] þ
[DFDPS]) ꢃ 100%, respectively, used in the polycondensations.
736
WILEYONLINELIBRARY.COM/JOURNAL/JPOLA

ARTICLE
temperature was increased to 175 ꢁC and kept at this tem-
perature for 20 h (PAEN60, PAEN40, PAEN20, PAES49, and
PAES27) or until the copolymer precipitated, which occurred
after 1.5, 3, and 2 h for PAEN80, PAES85 and PAES68,
respectively. For the latter three ionomers, the reaction tem-
perature was subsequently lowered until the polymer
regained solubility, which occurred at 120, 140, and 155 ꢁC,
respectively. This reaction temperature was kept for 15 h
before precipitation of the copolymers in an excess of IPA
at room temperature. The copolymers were then purified
using the same procedure as described above for the
homopolymers.
Water Uptake and IEC Measurements
The water uptake (wwater) was measured under immersed
conditions after equilibration in distilled water for at least
48 h, and at 98% RH after storage in a sealed vessel with a
saturated aqueous solution of CuSO4ꢄ5H2O. To obtain the wet
weight (Wwet), the excess water was gently removed with
tissue paper before weighing the swollen membranes. The
dry wꢁeight (Wdry) was obtained after drying under vacuum
at 80 C overnight. The water uptake was then calculated as
follows:
À
Áꢀ
wwater ¼ Wwet ꢀ Wdry Wdry ꢃ 100%
(1)
Structural Characterization
The IEC was measured by titration of acidic membranes.
Protonated membranes were soaked in an aqueous 2 M NaCl
solution for at least 72 h. The solutions were titrated with a
0.01-M KOH solution, using phenolphthalein as indicator.
1H NMR data were collected using a Bruker DRX400 spec-
trometer. Spectra were recorded at 400.13 MHz and the
chemical shifts are reported relative to DMSO-d6 (d
2.50 ppm). Fourier transform infrared (FTIR) analysis was
carried out on a Bruker IFS 66 spectrometer. The polymers
were ground together with KBr before tablets were pressed.
Spectra of the samples were recorded between 400 and
4000 cmꢀ1 at a resolution of 4 cmꢀ1 using 128 cumulative
scans. The intrinsic viscosity ([g]) of the polymers was
measured by using an Ostwald capillary viscometer in a
thermostated water bath at 25 ꢁC. The samples were
dissolved in a 0.05-M solution of LiBr in DMSO and analyzed
The state of the water in the membranes was investigated
via DSC by observing the endothermic peaks associated with
water melting. The membranes were first allowed to equili-
brate at room temperature in distilled water for at least
24 h. Excess water was carefully removed with tissue paper
before placing the samples in a sealed aluminum container.
In the DSC exꢁperiment, the samples were first cooled from
ꢁ
25 C to ꢀ60 C, kept isothermally for 3 min at this temper-
ꢀ1
ꢁ
ꢁ
in the concentration range 1.0–12.2 g Lꢀ1
.
ature, before heating to 25 C. The scan rate was 5 C min
in all cases. The amount of freezing water was calculated by
integrating the peak of the melt endotherm and comparing
Ionomer Membrane Preparation
this value with the heat of fusion of pure ice, 334 J gꢀ1 36
By
.
Membranes of the ionomers were cast in their potassium
salt form from 7 wt % solutions in DMSO. All solutions were
passed through 0.45 lm porous polytetrafluoroethylene
filters befoꢁre membrane casting in Petri dishes under N2
combining the calculated amount of freezing water and the
gravimetrically determined total water absorption, the total
number of water molecules per sulfonic acid group (k) and
the freezing number of water molecules per sulfonic acid
group (kfreezing) were determined.
ꢁ
flow at 90 C for 24 h, followed by drying in vacuo at 80 C
for 24 h. Membranes with a thickness of 70–140 lm were
ion-exchanged to the protonated form by immersion in 1 M
aqueous HCl for at least 24 h, followed by leaching with
distilled water for 2 days during which the water was
exchanged several times.
Conductivity Measurements
Proton conductivity was evaluated by electrochemical imped-
ance spectroscopy (EIS) using a Novocontrol high-resolution
dielectric analyzer V 1.01S equipped with a Novocontrol
temperature system. Impedance data were collected using a
two-electrode cell in the frequency range of 10ꢀ1–107 Hz at
a voltage amplitude of 50 mV and was then analyzed using
Thermal Characterizations
A Q1000 calorimeter from TA Instruments was used to carry
out the differential scanning calorimetry (DSC) investiga-
tions. During the DSC experiment, the polymers were first
heated to 10 ꢁC below degradation temperature (Td). The
samples were then cooled to 50 ꢁC, followed by a heating
scan to 380 ꢁC. All heating and cooling rates were kept at
10 ꢁC minꢀ1. Tgs were taken as the midpoint of the transi-
tion recorded during the second heating scan.
R
the software WinDetaV from Novocontrol. The proton con-
ductivity data of all the membranes were recorded under
immersed conditions during heating from ꢀ20 to 120 ꢁC,
cooling to ꢀ20 ꢁC, and finally heating to 120 ꢁC. The
reported data were collected during the second heating scan.
For the PAE2,7 membrane, measurements were also per-
formed at 100% RH with the membrane preequilibrated
at 98% RH. The sample was placed together with a small
droplet of pure water in the sealed measurement cell.
The thermal stability was evaluated by thermogravimetric
analysis (TGA) using a Q500 analyzer from TA Instruments.
The samples were analyzed, both in the protonated form and
the sodium salt form, under N2 during heating from 50 to
600 ꢁC at 10 ꢁC minꢀ1 and under air during heating from
50 to 600 ꢁC at 1 ꢁC minꢀ1. Before the heating scan, the
samples were predried under N2 at 150 ꢁC for 10 min to
remove water. The Td was taken as the temperature at which
the polymer had lost 5% of its original weight during the
heating.
RESULTS AND DISCUSSION
Homopolymer Preparation and Characterization
We have previously described the synthesis of DFSBP and
demonstrated the efficiency of this monomer in polycon-
densations to prepare sulfonated aromatic homopolymers
with various backbone structures.30,31 In particular, the
PROTON CONDUCTING FUEL CELL MEMBRANES, JUTEMAR AND JANNASCH
737

JOURNAL OF POLYMER SCIENCE: PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
TABLE 1 Homopolymer and Copolymer Synthesis
all aromatic shifts between d ¼ 6.4–7.9 ppm. The signal
arising from the proton next to the sulfonic acid group
(denoted a) was in both cases found between d ¼ 7.6–
7.7 ppm, whereas the protons ortho-to-ether on the DFSBP
residues (denoted e) gave rise to signals between d ¼ 6.4–
6.5 ppm. The signals originating from the protons of the
Monomer Feeda (mmol)
Polymer
DHNb
DFSBP
DFBN
DFDPS
[g]c (dL gꢀ1
)
PAE2,7
PAE2,6
PAEN80
PAEN60
PAEN40
PAES85
PAES68
PAES49
a
1.826
2.650
3.515
3.641
4.100
3.091
3.245
3.227
1.826
2.650
2.812
2.185
1.640
2.628
2.206
1.581
–
–
0.84
0.70
0.66
0.43
0.69
0.57
0.46
0.32
naphthalene moieties (denoted
g and h) were shifted
–
–
upward in the case of PAE2,6 as compared to PAE2,7. The
integrals of the peaks were in excellent agreement with the
respective ionomer structures. Both the homopolymers were
found to be soluble in polar aprotic solvents like DMAc,
DMSO, N-methylpyrrolidone (NMP), and N,N-dimethylforma-
mide (DMF) (Table 2).
0.703
–
1.456
–
2.460
–
–
–
–
0.464
1.038
1.646
The thermal characterizations as well as the water uptake
and proton conductivity measurements were conducted on
membranes cast from DMSO solutions. As expected, the use
of 2,6-DHN gave a more stiff backbone polymer, and PAE2,6
Monomers fed together with a 25% excess of potassium carbonate to
yield 20 wt % solutions in DMAc.
b
2,6-DHN used for PAE2,6 and 2,7-DHN for the remaining polymers.
c
Measured in solutions of 0.05 M LiBr in DMSO.
ꢁ
showedꢁa Tg at 334 C in comparison with the Tg of PAE2,7
at 300 C. In addition, the thermal stability was found to be
higher for the PAE2,6 membrane as compared to the PAE2,7-
membrane (Table 3). The water uptake of the PAE2,6 and
PAE2,7 membranes were 103 and 89%, respectively, under
immersed conditions at 20 ꢁC. However, the proton con-
ductivity of the two membranes was quite similar at 0.15 S
cmꢀ1 under immersed conditions at the same temperature.
On the whole, there were seemingly no significant advan-
tages of using 2,6-DHN, and thus 2,7-DHN was used in the
subsequent preparation of the copolymers.
polymerization of DFSBP and 2,7-DHN produced an ionomer
with a high intrinsic viscosity, high thermal stability, and
high level of proton conductivity.31 This ionomer evidently
also had a very high Tg as no glass transition was detected
in the temperature range up to the onset of degradation at
400 ꢁC. The primary focus of this study was the synthesis
and study of copolymers and membranes based on DFSBP.
However, initially we were interested to investigate if an ion-
omer with a stiffer configuration of the naphthalene moieties
in the backbone polymer would give significantly improved
properties. Thus, an ionomer was prepared by polycondensa-
tion of DFSBP and 2,6-DHN.
Copolymer Preparation and Characterization
Four PAEN copolymers and four PAES copolymers with dif-
ferent IECs were prepared on the basis of DFSBP and 2,7-
In the preparation of the homopolymers, DFSBP was charged
in equimolar amounts to the respective DHN together with a
25% excess of potassium carbonate (Scheme 2 and Table 1).
During the 4-h dehydration step, the di-phenolate salts
slowly precipitated. However, in the preparation of PAE2,7,
they regained solubility when the toluene was boiled off dur-
ꢁ
ing heating to 175 C. At this temperature, the reaction solu-
tion increased in viscosity and the polymerization was con-
tinued until the polymer lost solubility, which occurred after
1 h. The reaction temperature was then decreased to a tem-
perature at which the polymer regained solubility, and was
then maintained at this temperature for 15 h before isolation
of the polymer. On the other hand, PAE2,6 remained insolu-
ble as the toluene was boiled off during heating to 175 ꢁC.
ꢁ
However, during slow cooling to 70 C, the ionomer regained
solubility, and this temperature was kept for 18 h before
isolation of the product.
The intrinsic viscosities of PAE2,7 and PAE2,6 were found to
be 0.84 and 0.70 dL gꢀ1, respectively (Table 1). Notably, the
intrinsic viscosity was found to be high for PAE2,6, despite
the precipitation during the reaction. This may indicate that
the polymerization proceeded beyond the point of precipita-
tion or that the ionomer had reached a high molecular
weight already before the precipitation.
Figure 1 shows the 1H NMR spectra of the purified homo-
polymers. As expected, the spectra were very similar, with
FIGURE 1 1H NMR spectra of the homopolymers (a) PAE2,7,
and (b) PAE2,6 recorded using DMSO-d6 solutions.
738
WILEYONLINELIBRARY.COM/JOURNAL/JPOLA

ARTICLE
TABLE 2 Solubility of Sulfonated Polymers (d 5 Solubility
decreased with increasing IEC values of the copolymer.
Notably, the copolymers with an IEC of 1.55 meq gꢀ1, or
lower, remained in solution throughout the polymeriza-
tions and were kept at 175 ꢁC for 20 h before isolation of
the products. The intrinsic viscosities of the copolymers
were measured in DMSO solutions and ranged from 0.32
to 0.69 dL gꢀ1 (Table 1). The lower intrinsic viscosities of
the copolymers, in comparison with the homopolymers,
were most likely due to a lower solubility in DMSO of the
former.
Parameter [cal1/2cm23/2])
DMSO
DMF
NMP
DMAc
Polymer
(d ¼ 13.0)
(d ¼ 12.1)
(d ¼ 11.2)
(d ¼ 11.1)
PAE2,6
s
s
s
s
s
p
s
s
s
s
s
s
s
s
p
p
s
s
s
p
s
s
s
p
p
p
s
s
s
p
s
s
s
s
p
p
s
s
p
p
PAE2,7
PAEN80
PAEN60
PAEN40
PAEN20
PAES85
PAES68
PAES49
PAES27
Figure 2 shows the 1H NMR spectra of the purified copoly-
mers. In comparison with the PAE homopolymers, the spec-
tra of the PAEN copolymers expectedly contained a larger
number of peaks that were broader and partly overlapped.
All shifts were found between d ¼ 6.4–8.2 ppm. The protons
ortho-to-ether on the DFSBP residues (denoted a) gave rise
to signals between d ¼ 6.5–6.6 ppm and were found to
increase with an increasing feed ratio of DFSBP. In parallel,
the shift from the protons placed ortho-to-ether on the
DFBN residue at d ¼ 6.7–6.9 (denoted b) decreased in inten-
sity. The ratios of the integrals from these two peaks were
25:75, 41:59, and 60:40, respectively, which was in agree-
ment with the feed ratio of the monomers. For the PAES
copolymers, all shifts were found between d ¼ 6.4–8.2 ppm.
The signal originating from the protons ortho-to-ether on the
DFSBP residues (denoted c) was found between d ¼ 6.5–
6.6 ppm, and increased with increasing feed ratio of DFSBP.
The shift originating from the protons ortho-to-sulfone in the
DFDPS residue (denoted d) was found between d ¼ 7.9–
8.0 ppm and decreased in intensity with increasing feed
ratios of DFSBP. Unfortunately, this peak was overlapped by
adjacent peaks, which prevented quantification by NMR.
s, completely soluble; p, partially soluble.
DHN as shown in Scheme 3. Copolymers containing polar
nitrile groups have previously been reported to reduce the
water uptake of proton-exchange membranes.32 Therefore,
DFBN was chosen as one of the comonomers in this study.
The properties of the resulting copolymers were compared
with corresponding copolymers prepared with DFDPS as
comonomer. The IEC value was controlled by varying the
feed ratio of DFSBP:DFBN and DFSBP:DFDPS for the PAEN
an PAES copolymers, respectively. The monomers were
charged in equimolar amounts, according to Table 1,
together with a 25% excess of potassium carbonate. During
the 4-h dehydration step, the reactants slowly precipitated,
but regained solubility when the toluene was boiled off
during heating to 175 ꢁC. At this temperature, the reaction
solutions increased in viscosity and the polymerizations
were continued until the polymer lost solubility after 1.5 to
3 h. The reaction temperature was then decreased to a tem-
perature at which the polymers regained solubility, possibly
because the solubility of the salts decreased, which in turn
increased the solubility of the polymers. This temperature
FTIR spectra of the copolymers are depicted in Figure 3
together with the spectrum of the PAE2,7 homopolymer. For
the PAEN copolymers, the incorporation of DFBN was
confirmed by the appearance of a vibrational band at
2229 cmꢀ1 arising from the nitrile triple bond stretch. As
expected, this band increased in intensity with an increase in
TABLE 3 Sulfonated Ionomer Membrane Data
Sodium Salt Form
Acid Form
98% RH
wwater
Immersed
b
c
b
c
IECa
(meq gꢀ1
Tg
(ꢁC)
Td (ꢁC)
Td (ꢁC)
Td (ꢁC)
Td (ꢁC)
wwater
Membrane
)
Under N2
Under air
Under N2
Under air
(%)
k
(%)
k
kfreezing
PAE2,6
PAE2,7
PAEN80
PAEN60
PAEN40
PAES85
PAES68
PAES49
a
2.29 (2.38)
2.28 (2.38)
1.97 (2.07)
1.55 (1.69)
1.16 (1.24)
1.91 (2.06)
1.58 (1.68)
1.13 (1.24)
334
300
306
266
254
300
276
245
412
386
400
385
395
393
395
397
372
339
361
342
356
354
352
364
298
291
294
303
303
297
298
310
c
270
261
266
278
284
269
276
287
51
44
36
24
19
37
25
19
12
11
10
9
103
89
54
43
21
61
43
27
25
22
15
15
10
18
15
13
5
4
0
0
0
1
0
0
9
11
9
9
Measured by titration; theoretical values within parenthesis.
Measured at 1 ꢁC minꢀ1
.
Measured at 10 ꢁC minꢀ1
.
b
PROTON CONDUCTING FUEL CELL MEMBRANES, JUTEMAR AND JANNASCH
739

JOURNAL OF POLYMER SCIENCE: PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
FIGURE 2 1H NMR spectra of the PAEN and PAES copolymers recorded using DMSO-d6 solutions.
DFBN contents. In parallel, the vibrational bands at 1680
cmꢀ1 and 1088 cmꢀ1, arising from the carbonyl stretch and
the S¼¼O stretch in the sulfonic acid groups, respectively,
were found to decrease in intensity. In the spectra of the
PAES copolymers, the incorporation of DFDPS was confirmed
mer showed a very low water uptake, below 16% under
immersed conditions at room temperature.
Thermal Properties
The ionomer structure and the IEC value had large impact
on the Tg, as seen in Table 3. The Tgs were only measured
for the membranes in the sodium salt form, and no glass
transitions were detected for the membranes in the acid
form. As already mentioned, the homopolymer membranes
cast from DMSO showed high Tgs. Previously, we have
reported that no Tg was detected for PAE2,7 cast from NMP
in the temperature range up to the onset of degradation at
400 ꢁC.31 These results indicated differences in the mem-
brane properties depending on the solvent used in the
casting process. For both PAEN and PAES copolymer series,
the Tg was found to increase with increasing IECs. This was
expected because of the increasing ionic contents, which
increased the intermolecular interactions, and hence
decreased the segmental mobility leading to higher Tgs.
by the appearance of a vibrational band at 1105 cmꢀ1
,
arising from the S¼¼O stretch of the sulfone link. The inten-
sity of this vibrational band was found to increase with an
increase of the DFDPS feed, whereas the intensity of the
vibrational bands from the carbonyl stretch and the S¼¼O
stretch in the sulfonic acid groups decreased, just as
observed for the PAEN copolymers.
The solubility of the purified copolymers was investigated in
DMAc, DMSO, NMP, and DMF, and the results are shown in
Table 2. In general, the solubility was found to increase with
increasing IEC of the ionomers in these polar aprotic
solvents with high dielectric constants, presumably because
of increasing polarity of the ionomers. All but one of the
ionomers (PAEN20) were soluble in DMSO at room tempera-
ture. Consequently DMSO was chosen for the subsequent
membrane preparation. At this point, PAEN20 and PAES27
copolymers were not investigated further because the former
copolymer lacked solubility in DMSO and the latter copoly-
The thermal stability of the ionomer membranes both in the
sodium salt and in the acid form was evaluated by TGA
measurements. The measurements under air at 1 ꢁC minꢀ1
were undertaken to study the stability under oxidative
740
WILEYONLINELIBRARY.COM/JOURNAL/JPOLA

ARTICLE
observed to increase from 260 to 290 ꢁC with decreasing
IECs. In addition, at temperatures between Td and 400 ꢁC,
the TGA traces were shifted toward increasing temperatures
with decreasing IEC values. Comparing copolymers with
comparable IEC values, the Tds were very similar for the
PAEN and PAES copolymers, indicating the same degradation
mechanisms. As seen in Table 3, all but one ionomer
followed the trends described above, with the sodium salt
FIGURE 3 FTIR spectra of the PAEN and PAES copolymers.
conditionsꢀ, 1whereas the measurements under nitrogen at
ꢁ
10 C min were performed to study the degradation under
less drastic conditions. Figure 4 shows the TGA traces of the
membranes in the acid form when heated under air. Of the
two homopolymers, PAE2,6 was found to have the highest Td
value. For both the copolymer series, the Td values were
FIGURE 4 TGA traces of the ionomer membranes in the acid
form recorded under air at 1 ꢁC minꢀ1
.
PROTON CONDUCTING FUEL CELL MEMBRANES, JUTEMAR AND JANNASCH
741

JOURNAL OF POLYMER SCIENCE: PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
PAES membranes. As seen in Figure 5(b), the k-value seem-
ingly varied less with the IEC than the water uptake, particu-
larly at lower IEC values.
In Figure 6(a), the water uptake is shown for the mem-
branes under immersed conditions. As previously discussed,
the water uptake of the PAE2,6 homopolymer membrane
was higher (103%) than that of the PAE2,7 homopolymer
membrane (89%). Moreover, PAE2,7 membranes cast from
NMP have previously shown to take up excessive amounts of
water, 627% under immersed conditions.31 These results
indicated that the choice of solvent for the membrane casting
had a very strong influence on the water uptake, which is in
agreement with observations previously reported by Guiver
and coworkers.37 The water uptake was efficiently restricted
at lower IECs and followed similar trends as observed at
98% RH. However, under immersed conditions the water
uptake was more distinctly dependent on the IEC. Again, as
was already seen at 98% RH, the PAEN and the PAES mem-
branes showed a very similar water uptake as function of
IEC, with the PAEN membranes showing slightly higher
values. However, the reductions in the water uptake
FIGURE 5 Water uptake (a) and the corresponding k-values (b)
of the ionomer membranes as a function of IEC after equilibra-
tion at RH 98% at 25 ꢁC.
form of PAEN60 showing lower Td values than expected. As
anticipated, the stability was in all cases higher under nitro-
gen than under air, and the ionomers in the acid form
showed lower values of Td than those in the sodium salt
form.
Water Uptake Characteristics
The water uptake of proton-exchange membranes is highly
dependent on the IEC and has a great influence on the
proton conductivity. However, at high water contents the me-
chanical properties typically deteriorate because of the high
degree of swelling. Consequently, the membrane properties
should be tuned, so that the water uptake is controlled and
kept at a moderate level.
Figure 5(a,b) shows the water uptake data and k-values (i.e.,
the number of water molecules per sulfonic acid group),
respectively, for the different ionomer membranes as a func-
tion of IEC at 98% RH. As expected for the copolymers,
decreasing IECs led to a decreasing water uptake. The water
uptake of the PAES and PAEN membranes at a given IEC was
quite similar with only a slightly higher water uptake of the
FIGURE 6 Water uptake (a) and the corresponding k-values (b)
of the ionomer membranes as a function of IEC after immer-
sion in water at 25 ꢁC.
742
WILEYONLINELIBRARY.COM/JOURNAL/JPOLA

ARTICLE
FIGURE 7 Arrhenius proton conductivity plots for (a) the PAE homopolymers, (b) PAE2,7 cast from different solvents, (c) the PAEN
copolymers, and (d) the PAES copolymers. The data were measured under fully immersed conditions, if not otherwise stated. The
R
V
corresponding data of Nafion 117 has been included for comparison.
was lower than expected, as compared to the reductions of
10–15% previously reported by Sumner et al.32 from a study
of sulfonated poly(arylene ether nitrile) copolymers. A possi-
ble explanation for this might be the quite low concen-
trations of dissimilar segments in the copolymers in the
freezable water molecules per sulfonic acid group (kfreezing)
is presented in Table 3 and shows that all the studied ion-
omers contained no, or only very little, freezable water.
Thus, PAE2,7 and PAE2,6 contained 4 and 5 freezable water
molecules per sulfonic acid group, respectively. These num-
present case. In contrast to the data at 98% RH, the k-values
increased with an increase in IEC under immersed condi-
tions, as seen in Figure 6(b).
bers were expectedly lower than the value of kfreezing
¼
87 previously measured for the highly swollen PAE2,7
membrane cast from NMP.31 Of the copolymers, only mem-
brane PAES85 contained small amounts of freezable water
(kfreezing ¼ 1).
On the basis of DSC measurements, the amount of freezable
water in the membranes was determined after immersion.
The local environment of water in the membrane can be
identified from the temperature at which water in the mem-
brane freezes.38 Nonfreezable water strongly interacts with
sulfonic acid groups, whereas freezable water is ‘‘free’’ and is
not intimately bound to the sulfonic acid groups. Under
hydrated conditions, the tightly bound nonfreezable water
has a critical influence on the depression of the Tg, which
indirectly affects the proton conductivity.38,39 The number of
Proton Conductivity
The proton conductivity was measured by EIS during heating
ꢁ
ꢁ
from ꢀ20 C and 120 C. Figure 7(a) shows the proton con-
ductivity of PAE2,6 and PAE2,7 membranes measured under
immersed conditions. At subzero temperatures, the PAE2,6
membrane had a lower proton conductivity, which may be
explained by its higher water uptake and higher amount of
freezable water.38 Above 0 ꢁC, the proton conductivity of
PROTON CONDUCTING FUEL CELL MEMBRANES, JUTEMAR AND JANNASCH
743

JOURNAL OF POLYMER SCIENCE: PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
both membranes reached similar values, despite their differ-
ences in water uptake. This may be explained by dilution
effects in membrane PAE2,6, which had a negative influence
on the conductivity. The proton conductivity data of PAE2,7
membranes cast from NMP and DMSO, respectively, under
100% RH are depicted in Figure 7(b). Measurements under
fully immersed conditions were not possible for the PAE2,7
membrane cast from NMP because of its excessive water
uptake. As seen, the proton conductivity of the membrane
cast from DMSO was nearly an order of magnitude higher
than for the membrane cast from NMP, indicating that the
solvent used for the membrane casting had a profound influ-
ence on the proton conductivity. Similar solvent effects have
previously been reported from studies of sulfonated poly
(ether ether ketone)s.40 These solvent effects have been
explained by differences in solvent–ionomer interaction dur-
ing the membrane casting, which affect the conformation
and interactions of the polymer chains in the membrane, and
hence the proton conductivity.41
ductivity. Both copolymer series had a high thermal stabiliꢀty1
with values of Tds between 266 and 287 ꢁC (1 ꢁC min
under air) and Tgs between 245 and 306 ꢁC, depending on
IEC. At IECs above ꢂ1.6 meq gꢀ1, the copolymer membranes
R
reached higher proton conductivities than NafionV between
ꢀ20 and 120 ꢁC under immersed conditions. Despite differ-
ent structures and functionalities of the copolymer back-
bones, the PAES and the PAEN membrane series showed sur-
prisingly similar thermal and water uptake properties, and
proton conductivity at a given IEC. One possible explanation
for this may be that the concentration of dissimilar segments
in the two copolymer series was too low. Thus, the presence
of the strongly polar nitrile groups in the PAEN copolymers
did not lead to a significantly decreased water uptake in
relation to the PAES copolymers. Future work will focus on
evaluating the hydrolytical stability, as well as measuring the
proton conductivity of the membranes under reduced RH.
The authors thank the Swedish Foundation for Strategic Envi-
ronmental Research, MISTRA, for financial support.
Overall, the copolymer membranes reached high proton con-
ductivities, up to 0.3 S cmꢀ1 under fully humidified condi-
tions at 80 C [Fig. 7(c,d)]. At IECs above ꢂ1.5 meq gꢀ1, the
ꢁ
REFERENCES AND NOTES
copolymer membranes reached higher proton conductivities
1 de Bruijn, F. Green Chem 2005, 7, 132–150.
R
than NafionV in the range between ꢀ20 and 120 ꢁC. As
2 Zhang, J.; Xie, Z.; Zhang, J.; Tang, Y.; Song, C.; Navessin, T.;
Shi, Z.; Song, D.; Wang, H.; Wilkinson, D. P.; Liu, Z.-S.; Hold-
croft, S. J Power Sources 2006, 160, 872–891.
expected, the proton conductivity increased with increasing
IEC. However, within the range between 20 and 80 ꢁC, the
proton conductivity of PAES68 exceeded just above that of
the PAES85 [Fig. 7(d)]. Despite the rather similar water
uptake, the membranes in the PAEN series reached a higher
proton conductivity than the membranes of the PAES series
when compared at IEC values above 1.5 meq gꢀ1. However,
at an IEC of 1.1 meq gꢀ1 the proton conductivity was higher
for the PAES49 compared with the PAEN40. At subzero
temperatures, all membranes had proton conductivities
3 Hickner, M. A. Mater Today 2010, 13, 34–41.
4 Mauritz, K. A.; Moore, R. B. Chem Rev 2004, 104, 4535–4585.
5 Casciola, M.; Alberti, G.; Sganappa, M.; Narducci, R. J Power
Sources 2006, 162, 141–145.
6 Rusanov, A. L.; Kostoglodov, P. V.; Abadie, M. J. M.; Voyteku-
nas, V. Y.; Likhachev, D. Y. Adv Polym Sci 2008, 216, 125–155.
R
exceeding that of NafionV, which showed a sharp increase
7 Rozie`re, J.; Jones, D. J. Annu Rev Mater Res 2003, 33,
ꢁ
503–555.
between ꢀ20 to 20 C, presumably because of the melting of
freezable water.
8 Maier, G.; Meier-Haack, J. Adv Polym Sci 2008, 216, 1–62.
9 Higashihara, T.; Matsumoto, K.; Ueda, M. Polymer 2009, 50,
CONCLUSIONS
5341–5357.
Fully aromatic PAES and PAEN copolymers, as well as PAE
homopolymers, with naphthalene moieties and pendant sul-
fobenzoyl side chains were successfully prepared via potas-
sium carbonate mediated nucleophilic aromatic substitution
reactions using DFSBP. Consequently, all the polymers had
the sulfonic acid units placed in hydrolytically stable posi-
tions close to the electron withdrawing ketone links of the
side chains. The PAE2,6 homopolymer, derived from 2,6-
DHN, showed a higher thermal stability and Tg in compari-
son with the PAE2,7 homopolymer prepared with 2,7-DHN.
Measurements on membranes cast from DMSO indicated a
considerably higher water uptake for membranes based on
the former polymer. Still, the proton conductivity reached by
the two membranes was quite similar. Comparison of PAE2,7
membranes cast from DMSO and NMP, respectively, indicated
that the nature of the solvent strongly influenced the water
uptake and the proton conductivity of the membranes, with
the membranes cast from DMSO reaching the highest con-
10 Iojoiu, C.; Sanchez, J.-Y. High Perform Polym 2009, 21,
673–692.
11 Thomas, O. D.; Peckham, T. J.; Thanganathan, U.; Yang, Y.;
Holdcroft, S.
3640–3650.
J Polym Sci Part A: Polym Chem 2010, 48,
12 Yang. Y.; Holdcroft. S. Fuel Cells 2005, 5, 171–186.
13 Bae, B.; Miyatake, K.; Watanabe, M. Macromolecules 2009,
42, 1873–1880.
14 Hickner, M. A.; Ghassemi, H.; Kim, Y. S.; Einsla, B. R.;
McGrath, J. E. Chem Rev 2004, 104, 4587–4612.
15 Luu, D. X.; Cho, E.-B.; Han, O. H.; Kim, D. J Phys Chem B
2009, 113, 10072–10076.
16 Park, H. B.; Shin, H.-S.; Lee, Y. M.; Rhim, J.-W. J Memb Sci
2005, 247, 103–110.
17 Dyck, A.; Fritsch, D.; Nunes, S. P. J Appl Polym Sci 2002, 86,
2820–2827.
744
WILEYONLINELIBRARY.COM/JOURNAL/JPOLA

ARTICLE
18 Lafitte, B.; Karlsson, L. E.; Jannasch, P. Macromol Rapid
31 Jutemar, E. P.; Jannasch, P. ACS Appl Mater Interfaces,
Commun 2002, 23, 896–900.
In press.
19 Lafitte, B.; Puchner, M.; Jannasch, P. Macromol Rapid Com-
32 Sumner, M. J.; Harrison, W. L.; Weyers, R. M.; Kim, Y. S.;
McGrath, J. E.; Riffle, J. S.; Brink, A.; Brink, M. H. J Memb Sci
2004, 239, 199–211.
mun 2005, 26, 1464–1468.
20 Lafitte, B.; Jannasch, P. Adv Funct Mater 2007, 17, 2823–2834.
21 Wang, L.; Xiang, S.; Zhu, G. Chem Lett 2009, 38, 1004–1005.
22 Jutemar, E. P.; Jannasch, P. J Memb Sci 2010, 351, 87–95.
23 Karlsson, L. E.; Jannasch, P. J Memb Sci 2004, 230, 61–70.
33 Gao, Y.; Robertson, G. P.; Guiver, M. D.; Wang, G.; Jian, X.;
Mikhailenko, S. D.; Li, X.; Kaliaguine, S. J Memb Sci 2006, 278,
26–34.
34 Gao, Y.; Robertson, G. P.; Guiver, M. D.; Mikhailenko, S. D.;
Li, X.; Kaliaguine, S. Macromolecules 2005, 38, 3237–3245.
24 Zhu, J.; Shao, K.; Zhang, G.; Zhao, C.; Zhang, Y.; Li, H.;
Han, M.; Lin, H.; Xu, D.; Yu, H.; Na, H. Polymer 2010, 51,
3047–3053.
35 Gao, Y.; Robertson, G. P.; Kim, D.-S.; Guiver, M. D.; Mikhai-
lenko, S. D.; Li, X.; Kaliaguine, S. Macromolecules 2007, 40,
1512–1520.
25 Zhang, Y.; Zhang, G.; Wan, Y.; Zhao, C.; Shao, K.; Li, H.;
Han, M.; Zhu, J.; Xu, S.; Liu, Z.; Na, H. J Polym Sci Part A:
Polym Chem 2010, 48, 5824–5832.
36 Nakamura, K.; Hatakeyama, T.; Hatakeyama, H. Polymer
1983, 24, 871–876.
26 Pang, J. H.; Zhang, H. B.; Li, X. F.; Wang, L. F.; Liu, B. J.;
37 Xing, P.; Robertson, G. P.; Guiver, M. D.; Mikhailenko, S. D.;
Jiang, Z. H. J Memb Sci 2008, 318, 271–279.
Wang, K.; Kaliaguine, S. J Memb Sci 2004, 229, 95–106.
27 Bae, B.; Yoda, T.; Miyatake, K.; Uchida, H.; Watanabe, M.
38 Siu, A.; Schmeisser, J.; Holdcroft, S. J Phys Chem B 2006,
Angew Chem Int Ed 2010, 49, 317–320.
110, 6072–6080.
28 Roy, A.; Yu, X.; Dunn, S.; McGrath, J. E. J Memb Sci 2009,
39 Kim, Y. S.; Dong, L.; Hickner, M. A.; Glass, T. E.; McGrath, J.
327, 118–124.
E. Macromolecules 2003, 36, 6281–6285.
29 Nakabayashi, K.; Higashihara, T.; Ueda, M. J Polym Sci Part
40 Ye, G.; Mills, C. M.; Goward, G. R. J Membrane Sci 2008,
A: Polym Chem 2010, 48, 2757–2764.
319, 238–243.
30 Jutemar, E. P.; Takamuku, S.; Jannasch, P. Macromol Rapid
41 Robertson, G. P.; Mikhailenko, S. D.; Wang, K.; Xing, P.;
Commun 2010, 31, 1348–1353.
Guiver, M. D.; Kaliaguine, S. J Memb Sci 2003, 219, 113–121.
PROTON CONDUCTING FUEL CELL MEMBRANES, JUTEMAR AND JANNASCH
745