Synthesis and characterization of sulfonated polyimides bearing sulfonated aromatic pendant group for DMFC applications

Article abstract of DOI:10.1016/j.polymer.2010.06.005

A novel side-chain-type sulfonated aromatic diamine, 5-[1,1-bis(4-aminophenyl)-2,2,2- trifluoroethyl]-2-(4-sulfophenoxy)benzenesulfonic acid (BABSA) was synthesized and characterized. Two series of sulfonated polymides (SPI-N and SPI-B) were prepared from 1,4,5,8-naphthalene tetracarboxylic dianhydride (NTDA) or 4,4'-binaphthyl-1,1',8,8'-tetracarboxylic dianhydride (BNTDA), sulfonated diamine BABSA and various non-sulfonated aromatic diamines. The resulting sulfonated polyimide (SPI) membranes exhibited good dimensional stability with isotropic swelling of 7-22% and high thermal stability with desulfonation temperature of 283-330 °C. These membranes also displayed excellent oxidation stability and good water stability. The SPI membranes exhibited better permselectivity than Nafion 115 membrane due to their much lower methanol permeability. The ratios of proton conductivity to methanol permeability (F{cyrillic}) for the SPI membranes were almost two to three times of that for Nafion 115. The SPI-N membranes exhibited excellent conducting performance with the proton conductivity higher than Nafion 115 as the temperature over 40 °C, which attributed to their good hydrophobic/hydrophilic microphase separation structure.

Full text of DOI:10.1016/j.polymer.2010.06.005

Polymer 51 (2010) 3887e3898
Contents lists available at ScienceDirect
Polymer
journal homepage: www.elsevier.com/locate/polymer
Synthesis and characterization of sulfonated polyimides bearing sulfonated
aromatic pendant group for DMFC applications
*
Fei Sun, Taipeng Wang, Shiyong Yang, Lin Fan
Laboratory of Advanced Polymer Materials, Institute of Chemistry, Chinese Academy of Sciences, Zhongguancun, Beijing 100190, China
a r t i c l e i n f o
a b s t r a c t
Article history:
A novel side-chain-type sulfonated aromatic diamine, 5-[1,1-bis(4-aminophenyl)-2,2,2- trifluoroethyl]-2-
(4-sulfophenoxy)benzenesulfonic acid (BABSA) was synthesized and characterized. Two series of
sulfonated polymides (SPI-N and SPI-B) were prepared from 1,4,5,8-naphthalene tetracarboxylic di-
anhydride (NTDA) or 4,40-binaphthyl-1,10,8,80-tetracarboxylic dianhydride (BNTDA), sulfonated diamine
BABSA and various non-sulfonated aromatic diamines. The resulting sulfonated polyimide (SPI)
membranes exhibited good dimensional stability with isotropic swelling of 7e22% and high thermal
stability with desulfonation temperature of 283e330 ꢀC. These membranes also displayed excellent
oxidation stability and good water stability. The SPI membranes exhibited better permselectivity than
Nafion 115 membrane due to their much lower methanol permeability. The ratios of proton conductivity
to methanol permeability (V) for the SPI membranes were almost two to three times of that for Nafion
115. The SPI-N membranes exhibited excellent conducting performance with the proton conductivity
higher than Nafion 115 as the temperature over 40 ꢀC, which attributed to their good hydrophobic/
hydrophilic microphase separation structure.
Received 22 April 2010
Received in revised form
30 May 2010
Accepted 2 June 2010
Available online 9 June 2010
Keywords:
Sulfonated polyimides
Proton conductivity
Methanol permeability
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
sulfone)s [15,16], polybenzimidazoles [17] and polyimides [18e21].
They are expected to have low methanol crossover and high proton
Direct methanol fuel cell (DMFC) has been considered as the
ideal fuel cell system since it produces electric power by direct
conversion of the methanol fuel at the fuel cell anode [1,2]. The
advantages of DMFC, such as high energy density, convenient fuel
supply, quick start times, and instant refueling, make it more
attractive as mobile and portable power sources for cell phones,
notebook computers and other electronic devices [3,4]. A proton
exchange membrane with high conductivity and good stability is
one of the most important components in a DMFC system. The
perfluorosulfonic acid polymer membranes, such as Nafion
membranes, are the most successfully applied polymeric materials
for DMFC due to their high proton conductivity, good mechanical
property and excellent chemical stability [5,6]. However, their
drawbacks, including high costs, low operation temperature and
high methanol crossover, have to be overcome in commercial
applications [7e11].
conductivity. Among these sulfonated aromatic polymers, sulfonated
polyimides (SPIs) are recognized as one of the most promising alter-
native materials for fuel cell applications due to their excellent
combination properties of high thermal stability, excellent mechan-
ical strength, superior chemical resistance, good film forming ability
aswellassignificantlylow methanol permeability [22e25]. However,
common five-membered ring polyimides are generally unstable
toward acid due to the ease of hydrolysis of imide rings.
Mercier and his coworkers first successfully developed
sulfonated polyimides from a six-membered ring dianhydride
1,4,5,8-naphthalenetetracrboxylic dianhydride (NTDA), which
replaced the five-membered ring dianhydrides, a commercially
available sulfonated diamine 2,2-benzidinedisulfonic acid (BDSA)
and common non-sulfonated diamines [26,27]. They found that the
NTDA-based SPI membranes showed reasonably high performance
in a H2/O2 fuel cell system and the hydrolysis stability was greatly
improved by incorporation of naphthalenic imide structures.
However, the proton conductivities of these membranes were
rather low due to their low ion exchange capacity (IEC), which
limited their further development for applications.
In the past decades, many kinds of sulfonated aromatic polymers
have been developed as the alternative proton exchange membrane
materials, such as poly(ether ether ketone)s [12e14], poly(ether
In recent years, many kinds of sulfonated polyimides based on
NTDA and novel sulfonated aromatic diamines have been devel-
oped to improve the proton conducting performance. It is known
* Corresponding author. Tel.: þ86 10 6256 4819; fax: þ86 10 6256 9562.
E-mail address: fanlin@iccas.ac.cn (L. Fan).
0032-3861/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.polymer.2010.06.005

3888
F. Sun et al. / Polymer 51 (2010) 3887e3898
that SPIs can be classified into main-chain-type and side-chain-
type. In the former, the sulfonic acid groups are bonded directly to
the polymer backbones, while in the latter, they are attached to the
side chains of the polymer [28]. Most of the developed main-chain-
type SPI membranes displayed pretty good proton conductivities.
However, the water stability of these SPI membranes could not
satisfy the DMFC requirements. Okamoto and his coworkers
reported the side-chain-type SPIs bearing pendant sulfoalkoxy
groups based on 2,20-bis(3-sulfopropoxy)benzidine (2,20-BSPB) and
3,30-bis(3-sulfopropoxy)benzidine (3,30-BSPB) [29]. They found
that the side-chain-type SPIs exhibited higher proton conductivi-
ties than Nafion 117 and better water stability than main-chain-
type SPIs. It was confirmed that well micro-phase separated
structures played an important role in improving the proton
conductivity of the membranes. However, the proton conductivity
of the side-chain-type SPIs bearing pendant sulfoalkoxy groups
decreased largely at high temperature due to the relatively easy
cleavage of sulfopropoxy groups [30]. Aiming at developing proton
exchange membranes with good working performance at high
temperature, several attempts have been carried out on the
investigation of aromatic side-chain-type SPIs [31e34]. The SPIs
based on 3,30- and 2,20-bis(sulfophenoxy)benzidine (BSPOB) are the
typical ones. It has been reported that the BSPOB-based SPI
membranes showed high proton conductivities evenin low humidity
(50% RH) and at high temperature (120 ꢀC). They exhibited good
water stability with high mechanical strength and proton conduc-
tivity even after aging in water at 130 ꢀC for 500 h. The excellent
proton conducting performance of BSPOB-based SPI membranes is
considered to be attributed to the thermal stable wholly aromatic
structure and the side-chain-type molecular designing.
In this study, a novel side-chain-type sulfonated aromatic
diamine, 5-[1,1-bis(4-aminophenyl)-2,2,2-trifluoroethyl]-2-(4-sul-
fophenoxy) benzenesulfonic acid (BABSA), was synthesized. The
sulfonated polyimides bearing sulfonated aromatic pendant groups
were prepared from BABSA, NTDA and various non-sulfonated
aromatic diamines, i.e., 4,40-bis(4-amino-2-trifluoromethylphenoxy)
biphenyl (6FBAB), 4-bis(4-amino-2-trifluoromethylphenoxy)benzene
(6FAPB), and 4,40-bis(4-aminophenoxy)biphenyl (BAPB), respectively.
The SPI membranes are expected to have excellent proton
conductivity at high temperature combined with good water
stability and low methanol crossover. In order to get the SPI
membranes with well micro-phase separated structure, extremely
hydrophilic sulfonic acid groups were introduced to the long side
chain of the polymer, meanwhile, hydrophobic trifluoromethyl
groups were also incorporated to the polymer backbone. It is also
known that the trifluoromethyl substituents are effective in
improving the oxidation stability and the tensile properties of
polyimide membranes [35]. Binaphthyl dianhydride BNTDA with
higher electron density in carbonyl carbon atoms was proved to be
effective in enhancing the hydrolytic stability of the SPI membranes
[36]. Therefore, another series of side-chain-type sulfonated poly-
imides was also prepared from BNTDA, BABSA and non-sulfonated
diamines mentioned above and compared with the NTDA-based
SPIs. The solubility and thermal property of the side-chain-type
sulfonated polyimides were evaluated. Their mechanical property,
water and oxidative stabilities, methanol permeability as well as
proton conductivity were investigated. The relationship between
the polymer structure and their properties was also discussed.
used as received. Anhydrous lithium trifluoroacetate was prepared
in our laboratory by a reaction of lithium hydroxide and tri-
fluoroacetic acid at 5 ꢀC for 4 h, and then dried under vacuum at
130 ꢀC for 6 h 5-Bromoacenaphthene was prepared via electro-
philic substitution of acenaphthene (SigmaeAldrich) by N-bro-
mosuccinimide in N,N-dimethylformamide (DMF) at room
temperature as reported in the literature [37]. Triphenylphosphine
(PPh3) and anhydrous NiCl2 (Beihua Fine Chemicals Co., China)
were dried under vacuum. Zinc dust (Acros) was activated in
hydrochloric acid under vigorous stirring, then filtrated, washed
with diethyl ether and dried under vacuum. 1,4,5,8-Naphthalene
tetracarboxylic dianhydride (NTDA) (Beijing Multi Technology Co.,
China) was purified by vacuum sublimation. 4,40-Bis(4-amino-2-
trifluoromethylphenoxy)biphenyl (6FBAB) and 1,4-bis(4-amino-2-
trifluoromethylphenoxy)benzene (6FAPB) (Beijing POME Sci-tech
Co., China) were crystallized from ethanol. 4,40-Bis(4-amino-
phenoxy)biphenyl (BAPB) (SigmaeAldrich) was purified by
recrystallization from toluene. m-Cresol and triethylamine (TEA)
was purified by vacuum distillation and dehydrated with 4 Å
molecular sieves prior to use. Concentrated sulfuric acid (98%),
fuming sulfuric acid (SO3, 50%), and commercial available solvents,
i.e., N-methyl-2-pyrrolidone (NMP), N,N-dimethyl acetamide
(DMAc), N,N-dimethylformamide (DMF) and dimethyl sulfoxide
(DMSO) (Beihua Fine Chemicals Co., China), were used as received.
2.2. Monomers synthesis
2.2.1. 40-Phenoxyl-2,2,2-trifluoroacetophenone (PTFP)
A
mixture of anhydrous lithium trifluoroacetate (24.9 g,
0.10 mol), magnesium (2.76 g, 0.115 mol), iodine (0.02 g) and
freshly distilled tetrahydrofuran (200 mL) was placed into
a 500 mL three-necked round-bottom flask fitted with a dropping
funnel, a drying tube, and a reflux condenser. After the anhydrous
lithium trifluoroacetate was completely dissolved under stirring,
a small part of 4-bromodiphenyl ether (4.9 g, 0.02 mol) was added.
The mixture was slowly heated under stirring until the reaction
was initiated. Then, the mixture of additional 4-bromodiphenyl
ether (20 g, 0.08 mol) and freshly distilled anhydrous tetrahy-
drofuran (100 mL) was added dropwise. The reaction solution was
stirred at room temperature for 4 h, after that, it was slowly
heated to 65 ꢀC and kept refluxing for 2 h. The solution was cooled
down to room temperature, and then a mixture of concentrated
hydrochloric acid (36%, 30 mL) and distilled water (30 mL) was
slowly added with agitation. The organic phase was separated,
washed with 5% aqueous sodium bicarbonate and distilled with
water, dried with anhydrous magnesium sulfate, followed by
distilling to remove the solvent. The resulting product was purified
by distillation to give colorless liquid of PTFP (19.7 g, 74%). 1H NMR
(CDCl3, d, ppm): 8.04e8.06 (d, 2H); 7.42e7.46 (t, 1H); 7.24e7.27
(d, 2H); 7.09e7.11 (d, 2H); 7.03e7.05 (d, 2H). Elemental analysis:
Calculated for C14H9F3O2: C, 63.16%; H, 3.41%. Found: C, 63.08%;
H, 3.29%.
2.2.2. 1,1-Bis(40-aminophenyl)-1-(400-phenoxyphenyl)-2,2,2-
trifluoroethane (BAPTF)
A mixture of PTFP (20 g, 0.075 mol), aniline (75 mL), and aniline
hydrochloride (14.5 g, 0.11 mol) was heated to reflux for 24 h. After
that, the solution was cooled to room temperature and neutralized
with 10% of aqueous sodium bicarbonate. The crude product was
obtained by vacuum distillation followed by water vapor distillation
to remove excess aniline, which was thenpurified by recrystallization
2. Experimental
2.1. Materials
from ethanol to yield white crystals (24.4 g, 75%). 1H NMR (CDCl3,
d,
ppm): 7.32e7.37 (t, 2H); 7.11e7.14 (m, 3H); 7.04e7.07 (d, 2H);
6.89e6.94 (m, 6H); 6.60e6.63 (d, 4H); 3.74 (s, 4H). FTIR (KBr, cmꢁ1):
3454, 3385, 1624, 1587, 1512, 1491, 1286, 1229, 1198, 1150 827, 756,
4-Bromodiphenyl ether and magnesium turnings (Sigmae
Aldrich), as well as aniline (Beihua Fine Chemicals Co., China) were

F. Sun et al. / Polymer 51 (2010) 3887e3898
3889
694. Elemental analysis: Calculated for C26H21F3N2O: C, 71.88%; H,
4.87%; N, 6.45%. Found: C, 71.68%; H, 4.99%; N, 6.43%.
diamine BABSA and non-sulfonated diamine, i.e., 6FBAB, 6FAPB or
BAPB, respectively. The molar ratio of sulfonated diamine to non-
sulfonated diamine was 3:1. In a typical experiment, SPI-N1, which
derived from NTDA, BABSA and 6FBAB, was prepared in the
following procedure.
2.2.3. 5-[1,1-Bis(40-aminophenyl)-2,2,2-trifluoroethyl]-2-(400-
sulfophenoxy)benzenesulfonic acid (BABSA)
To a 100 mL three-necked flask equipped with a mechanical
stirrer, BAPTF (8.68 g, 0.02 mol) was added and cooled with an ice
bath. Then concentrated sulfuric acid (8 mL) was slowly added with
stirring. After BAPTF was completely dissolved, 3.2 mL of fuming
sulfuric acid (SO3, 50%) was added dropwise. The reaction mixture
was stirred at 0 ꢀC for 0.5 h and at 40 ꢀC for 6 h, and then poured
into ice water. The resulting white precipitate was collected fol-
lowed by dissolved in sodium hydroxide solution. The solution was
concentrated to give light-brown sodium salt form product BABSA-
Na (11.5 g), which was then purified by recrystallization in ethanol
aqueous solution. The gained crystal was acidified with concen-
trated hydrochloric acid. The resulting precipitate was filtered off,
washed with water and methanol successively, and dried in vacuo
at 120 ꢀC for 10 h to give BABSA (10.82 g, 91%). 1H NMR (BABSA-Na,
To a completely dried 100 mL three-necked flask equipped with
a mechanical stirrer, a nitrogen inlet, and a thermometer, BABSA
(1.784 g, 3 mmol), m-cresol (20 mL) and TEA (1.2 mL) were fitted.
The mixture was stirred at ambient temperature under nitrogen
flow until BABSA was completely dissolved to give a homogeneous
solution. Then 6FBAB (0.504 g, 1 mmol), NTDA (1.072 g, 4 mmol)
and benzoic acid (0.95 g, 7.8 mmol) were added to the solution. The
mixture was stirred at room temperature for a few minutes and
then heated to 80 ꢀC for 4 h and 180 ꢀC for 20 h, respectively. After
the solution was cooled to 80e100 ꢀC, 20 mL of additional m-cresol
was added to dilute the highly viscous solution. The fiber-like
precipitate was obtained as the solution was poured into acetone,
and then collected by filtration, washed with acetone, and dried in
vacuum at 120 ꢀC for 12 h.
DMSO-d6,
d
, ppm): 7.74 (s, 1H); 7.54e7.56 (d, 2H); 6.89e6.91
The other sulfonated polyimides, i.e., SPI-N2 (NTDA/BABSA/
6FAPB), SPI-N3 (NTDA/BABSA/BAPB), SPI-B1 (BNTDA/BABSA/6FBAB),
SPI-B2 (BNTDA/BABSA/6FAPB) and SPI-B3 (BNTDA/BABSA/BAPB)
were synthesized by the similar procedure, except that different
kinds of dianhydrides and non-sulfonated diamines were applied.
(d, 1H); 6.85e6.87 (d, 2H); 6.73e6.75 (d, 1H); 6.66e6.68 (d, 4H);
6.49e6.51 (d, 4H); 5.16 (s, 4H). FTIR (KBr, cmꢁ1): 3452, 2926, 2624,
1632, 1589, 1514, 1485, 1250, 1196, 1158, 1124, 1094, 1030, 1007, 825,
692, 615, 565. Elemental analysis: Calculated for C26H21F3N2O7S2: C,
52.52%; H, 3.56%; N, 4.71%. Found: C, 71.68%; H, 4.99%; N, 6.43%.
Calculated for water content of 6 wt.% in the sample: C, 49.54; H,
3.96; N, 4.45. Found: C, 49.52; H, 4.04; N, 4.41.
2.3.2. Membrane preparation
Membranes of the sulfonated polyimides were prepared by
casting the m-cresol solutions of polymers in TEA salt form
(7e10 wt.%) on glass plates and drying at 120 ꢀC for 10 h. The
as-cast membranes were soaked in methanol at 60 ꢀC for 4e5 h to
remove the residual solvent followed by proton-exchange treat-
ment in 1 mol/L HCl at room temperature for 10e12 h. The proton-
exchanged membranes were thoroughly washed with water and
then dried in vacuo at 150 ꢀC for 10 h. The membrane thickness was
2.2.4. Diacenaphthene
A 250 mL three-necked round-bottomed flask equipped with
a serum cap, a gas inlet and outlet tube was filled with NiCl2 (0.39 g,
3 mmol), PPh3 (3.93 g, 15 mmol), zinc dust (11.7 g, 180 mmol) and
DMAc (50 mL). The mixture was reacted at 90 ꢀC for 0.5 h under
stirring in nitrogen atmosphere. Then a nitrogen purged solution of
5-bromoacenaphthene (10.49 g, 45 mmol) in DMAc (50 mL) was
slowly added via a syringe. The mixture was stirred at 90 ꢀC for
another 12 h, then concentrated to remove the solvent and poured
into 10% of HCl aqueous solution. The dark-gray slurry was sepa-
rated and dissolved in acetone followed by filtration to remove the
insoluble substance. The solution was dried with anhydrous
magnesium sulfate and concentrated by evaporation. The yellowish
viscous liquid was purified by silica gel column chromatography
50e60 mm.
2.4. Characterization
2.4.1. Measurements
1H NMR and 13C NMR spectra were performed on a Bruker
Avance 400 spectrometer operating at 400 MHz in CDCl3 and
DMSO-d6. FTIR spectra were recorded on a PerkineElmer 782
Fourier transform spectrophotometer. Mechanical properties were
measured on an Instron 3365 Tensile Apparatus with
80 ꢂ 10 ꢂ 0.05 mm specimens at 25 ꢀC and about 40% RH at
a crosshead speed of 2 mm/min. Differential scanning calorimetry
(DSC) was carried out using a TA Q100 instrument in nitrogen at
a heating rate of 10 ꢀC/min. Thermal gravimetric analysis (TGA) was
performed on a TA Q50 instrument in nitrogen at a heating rate of
20 ꢀC/min. Transmission electron microscopy (TEM) photographs
were taken on a JEM-1011 transmission electron microscope, using
an accelerating voltage of 100 kV. In order to stain the ionic
domains, membrane samples in proton form were converted into
Agþ form by immersing them in 0.5 M AgNO3 aqueous solution
overnight and then thoroughly rinsed with water, dried at room
temperature for 12 h before measurement.
using hexane as the eluent (3.37 g, 49.7%). 1H NMR (CDCl3,
d, ppm):
7.49e7.51 (d, 2H); 7.39e7.41 (d, 2H); 7.27e7.32 (m, 6H); 3.50 (s, 8H).
Elemental analysis: Calculated for C24H18: C, 94.08%; H, 5.92%.
Found: C, 93.95%; H, 5.89%.
2.2.5. 4,40-Binaphthyl-1,10,8,80-tetracarboxylic dianhydride
(BNTDA)
Diacenaphthene (3.06 g, 0.01 mol), sodium dichromate (26.2 g,
0.1 mol), and glacial acetic acid (450 mL) were refluxed at 120 ꢀC for
12 h. After cooling to room temperature, the solution was poured
into water (700 mL). The precipitated solid was collected and
recrystallized from DMF to give a yellow powder (2.80 g, 71%). m.p.
362 ꢀC. 1H NMR (DMSO-d6,
d, ppm): 8.70e8.72 (d, 2H); 8.59e8.61
(d, 2H); 7.98e8.00 (d, 2H); 7.79e7.83(t, 2H); 7.73e7.75 (d, 2H).
Elemental analysis: Calculated for C24H10O6: C, 73.10%; H, 2.56%.
Found: C, 73.04%; H, 2.73%. Mass spectrum (TOF, m/z, % relative
intensity): 394 (Mþ, 100).
2.4.2. Ion exchange capacity (IEC) and water uptake, dimensional
change
The ion exchange capacity (IEC) was evaluated by means of
titration. The dried SPI membranes were soaked in 15 wt.% NaCl
solution at 30 ꢀC for 48 h, then the protons released due to the
exchange reaction with Naþ ions were titrated with a 0.02 M NaOH
solution using phenolphthalein as an indicator. IEC value was
calculated from:
2.3. Polymer synthesis and membrane preparation
2.3.1. Synthesis of sulfonated polyimides
A series of sulfonated copolyimides were prepared via one-step
polycondensation from dianhydride NTDA or BNTDA, sulfonated

3890
F. Sun et al. / Polymer 51 (2010) 3887e3898
where k is the slope of the straight-line plot of CB versus perme-
ation time, VB is the volume of solution B, L and A are the thickness
and effective area of the tested membrane.
IEC ¼ (V ꢂ NNaOH)/W
where V is the volume of NaOH consumed, NNaOH is the normality of
NaOH solution, and W is the weight of sample.
Water uptake was also measured by immersing the SPI
membranes into water at 30 ꢀC for 24 h. Then the membrane was
taken out, quickly wiped with a tissue paper and weighed on
a microbalance. Water uptake (WU) was calculated from:
3. Results and discussion
3.1. Monomers synthesis
The side-chain-type sulfonated diamine BABSA was synthesized
by a three-step reaction process as shown in Scheme 1. First, PTFP
was prepared by the Grignard reaction between anhydrous lithium
trifluoroacetate and 4-bromodiphenyl ether in the presence of
magnesium in THF. Then, the diamine BAPTF was synthesized by
the coupling of 1 equiv of PTFP with 2 equiv of aniline catalyzed by
aniline hydrochloride at about 185 ꢀC. After that, BAPTF was
sulfonated with fuming sulfuric acid (50% SO3) and precipitated as
the inner salt, which was then neutralized with sodium hydroxide
solution to get sodium salt form intermediate BABSA-Na. Finally,
the sulfonated diamine BABSA was obtained by acidification of
BABSA-Na with hydrochloric acid.
WU ¼ (Ws ꢁ Wd)/Wd ꢂ 100 wt.%
where Ws and Wd are the weights of hydrated and dried
membranes, respectively.
Dimensional change of SPI membranes was measured by
immersing the round-shaped samples (Ø40 ꢂ 0.05 mm) in water at
25 ꢀC for 4e5 h. The changes of thickness (
Dtc) and diameter (Dlc)
were calculated from:
D
tc ¼ (t ꢁ ts)/ts and Dlc ¼ (l ꢁ ls)/ls
where ts and ls are the thickness and diameter of membranes
equilibrated at 25 ꢀC and 70% RH for 24 h, respectively; t and l refer
to those of the membranes immersed in water for 4e5 h.
Fig. 1 shows the 1H NMR spectra of compounds PTFP and BAPTF,
in which all the protons in the structure could be assigned easily
according to the integral values of the intensity. In addition, the
elemental analysis values for PTFP and BAPTF were also in good
agreement with the calculated ones. The 1H NMR and 13C NMR
spectra of sodium form intermediate BABSA-Na is shown in Fig. 2.
In the 1H NMR spectrum, the signal assigned to the protons in
amino groups was observed at 5.16 ppm. The aromatic protons
H1eH5 showed the doublet peaks at 6.49e6.91 ppm. The aromatic
protons adjacent to the eSO3Na groups (H6 and H7) shifted to the
downfield because of the strong electron-withdrawing effect of the
eSO3Na groups, which were observed at 7.54e7.56 and 7.74 ppm,
respectively. In the 13C NMR spectrum, the carbon in tri-
fluoromethyl group (C6) showed the clear quartet absorption at
124e133 ppm because of the coupling interaction of the carbon
atom with fluorine atoms in the diamine molecule. All of other
carbon atoms could be clearly assigned as expected.
2.4.3. Water stability and oxidative stability
The water stability was carried out by immersing SPI
membranes into distilled water at 100 ꢀC and the stability was
evaluated by the elapsed time for the samples start to lose their
mechanical strength.
The oxidative stability of SPI membranes was determined by
immersing the samples (10 ꢂ 10 ꢂ 0.05 mm) into Fenton’s reagent
(30 ppm FeSO4 in 30% H2O2) at 30 ꢀC, which was evaluated by
recording the time when the samples began to dissolve and dis-
solved completely.
2.4.4. Proton conductivity
Proton conductivity in plane direction of SPI membrane was
determined using a four-point-probe electrochemical impedance
spectroscopy technique over the frequency from 10 Hz to 100 KHz
(Hioki 3522-50). A membrane sample (40 ꢂ 10 ꢂ 0.05 mm) and two
pieces of platinum plate electrodes were set in a Teflon cell. The
distance between the two electrodes was 2 cm. The cell was placed
in distilled water for measurement at 100% RH. The resistance value
associated with the membrane conductance was determined from
high frequency intercept of the impedance with the real axis.
The structure of BAPTF and BABSA were further confirmed by
means of FTIR spectra. As shown in Fig. 3, the characteristic
absorptions around 3450 cmꢁ1 and 1151 cmꢁ1, which due to the
Proton conductivity (s) was calculated from:
s
¼ d/(ts$ws$R)
where d is the distance between the two electrodes, ts and ws are
the thickness and width of the membrane, and R is the resistance
value measured.
2.4.5. Methanol permeability
Methanol permeability of membrane was measured using
a liquid permeation cell composed of two compartments, which
were separated by a vertical membrane. One compartment of the
cell (A) was filled with methanol aqueous solution (CA ¼ 30 or
50 wt.%), the other compartment (B) was filled with distilled water.
The tested membrane was immersed in water to get swollen
sample and then set into the measurement cell. The solutions in
both compartments were kept at 30 ꢀC and stirred continuously
during the permeation measurement. The methanol concentration
of compartment B (CB) was analyzed by a gas chromatograph
(Shimadzu, GC-2014). The methanol permeability (PM) was calcu-
lated by the following equation:
PM ¼ (k$VB$L)/(A$CA)
Scheme 1. Synthesis of side-chain-type sulfonated aromatic diamine BABSA.

F. Sun et al. / Polymer 51 (2010) 3887e3898
3891
Fig. 1. 1H NMR spectra of PTFP (a) and BAPTF (b) in CDCl3.
NeH stretching vibration of amino groups and the CeF stretching
vibration of trifluoromethyl group, respectively, were detected for
both of BAPTF and BABSA. In addition, the broad absorption band
around 1250 cmꢁ1 and the absorption at 1030 cmꢁ1, which
assigned to the stretching vibration of sulfonic acid groups, were
also observed in the FTIR spectrum of BABSA. The elementary
analysis result for BAPTF was in good agreement with the calcu-
lated one. However, the found value of elementary analysis for
BABSA was slightly different from the calculated value. This is
mainly caused by the adsorbed water in the BABSA sample (ca. 6 wt.
%). The characterization results indicated that the obtained
sulfonated diamine BABSA was pure enough to be employed for
preparation of polyimides.
dianhydride BNTDA was prepared via a two-step reaction route as
shown in Scheme 2. The diacenaphthene was firstly synthesized by
coupling of 5-bromoacenaphthene catalyzed by Ni(0) at 90 ꢀC,
which was then oxidized by sodium dichromate in glacial acetic
acid to afford BNTDA. The reaction process is very simple and the
reaction condition is moderate as comparing with that reported.
The chemical structure of BNTDA was clearly identified by means of
1H NMR spectroscopy. The elemental analysis values were in good
agreement with the calculated ones, which also provided clear
confirmation of the structure and the purity of the product. The
dinaphthalic dianhydride BNTDA is yellow powder and stable to
the moisture in air at room temperature.
The synthesis of BNTDA via a four-step reaction process had
been reported by several groups [38,39], in which 4-chloro-1,8-
naphthalix anhydride was firstly converted into ester followed by
coupling to afford the tetraester and subsequent hydrolysis and
dehydration to provide the dianhydride BNTDA. In this study, the
3.2. Polymer synthesis and characterization
A series of sulfonated copolyimides were synthesized by a one-
step method from dianhydrides NTDA or BNTDA, sulfonated
diamine BABSA and various non-sulfonated diamines, respectively,

3892
F. Sun et al. / Polymer 51 (2010) 3887e3898
Fig. 2. 1H NMR (a) and 13C NMR (b) spectra of BABSA-Na.
in m-cresol with the presence of TEA and benzoic acid as illustrated
in Scheme 3. TEA was used to liberate the protonated amino groups
for polymerization with dianhydrides and benzoic acid acted as
a catalyst. This method has been employed for preparation of many
other sulfonated polyimides in the literature [40]. The molar ratio
of sulfonated diamine BABSA to non-sulfonated diamine was 3:1,
which give a fixed sulfonation degree of 75%. Fiber-like precipitates
could be gained as the polymer solution was poured into acetone,
indicating that the polymers have relatively high molecular weight.
Fig. 4 shows the FTIR spectra of the SPI membranes. For the SPI-
N membranes derived from dianhydride NTDA, the characteristic
absorptions of naphthalenic imide rings were observed at
1715 cmꢁ1 (C]O, symmetric), 1676 cmꢁ1 (C]O, asymmetric) and
Fig. 3. FTIR spectra of BAPTF and BABSA.
Scheme 2. Synthesis of dianhydride BNTDA.

F. Sun et al. / Polymer 51 (2010) 3887e3898
3893
vibration shifted up to 1365 cmꢁ1, respectively. The characteristic
absorption bands for sulfonic acid groups showed no significant
difference from that for SPI-N membranes. Moreover, for the SPI
membranes prepared from non-sulfonated diamines 6FBAB and
6FAPB, i.e., SPI-N1, SPI-N2, SPI-B1 and SPI-B2, the absorption
attributed to the CeF stretching vibration of trifluoromethyl group
was detected at 1054 cmꢁ1. The 1H NMR spectra of SPI-N3 and SPI-
B3 are shown in Fig. 5. All of the protons could be well assigned to
the expected polymer structure and the integration intensity ratio
of diamine moieties protons over naphthalenic or dinaphthalenic
protons was found in good accordance with the theoretical ones,
indicating the complete polycondensation.
The solubility of sulfonated polyimides in TEA salt form and in
proton form was determined by dissolving the polymer resins in
different organic solvents with 5 wt.% of solid content followed by
stirring for 24 h. As shown in Table 1, the SPIs displayed better
solubility in the TEA salt form than that in the proton form. All the
SPIs in TEA salt form could be completely dissolved in the chosen
solvents. The SPIs in proton form also could be easily dissolved in
the testing solvents except in m-cresol. The SPI-N1, SPI-N2, SPI-B2
and SPI-B3 in proton form were partially soluble in m-cresol at
room temperature, whereas, they could be dissolved after heating
at 50 ꢀC. The solubility was mainly governed by the structure of the
sulfonated diamine moieties. The good solubility of the present SPIs
was due to the synergetic effects of the bulky aromatic pendant
groups and the fluorine groups in the backbones, which disrupted
the regularity of the molecular chains and hindered the dense chain
stacking, thus increased the solubility. The polymer solutions were
very stable and no phase separation or precipitation was observed
after storage for several weeks.
3.3. Mechanical and physical properties
All the synthesized SPIs displayed pretty good film forming
abilities. Tough, flexible and transparent membranes could be
obtained by casting the m-cresol solutions of SPIs in TEA salt form
onto glass plates followed by thermal baking to remove the solvent
and subsequent proton-exchange treatment and drying. The
mechanical properties of the dry SPI membranes are listed in
Table 2. All the membranes exhibited good tensile strength and
fracture toughness. Both of SPI-N and SPI-B membranes displayed
higher Young’s modulus and tensile strength than Nafion 115,
suggesting better mechanical properties of them. As comparing the
mechanical properties of SPI-N with that of SPI-B membranes, it is
found that the former gave the tensile strengths of 69.5e74.6 MPa,
which were 17.2e22.8 MPa higher than those of the latter with
same diamine moieties. This indicates that the SPI-N membranes
were stiffer than the SPI-B membranes. The elongation at break of
SPI-N membranes was around 10% and that for SPI-B membranes
was in the range of 6.8e10.6%. Considering the membranes were
completely dried before measurement, the flexibility of the SPI
membranes was pretty good. If the membranes were partially or
fully hydrated, the flexibility must be greatly improved.
Scheme 3. Synthesis of sulfonated polyimides.
1350 cmꢁ1 (CeN, asymmetric), respectively, which were lower than
those for the phthalimide (e.g., 1780, 1720 and 1375 cmꢁ1). The
absorption assigned to the stretching vibrations of sulfonic acid
groups were found at 1094 and 1026 cmꢁ1
. For the SPI-B
membranes derived from dianhydride BNTDA, it was found that the
absorption bands of C]O stretching vibration shifted down to 1710
and 1670 cmꢁ1, while the absorption band of CeN stretching
Ion exchange capacity (IEC), water uptake and the number of
water molecules per sulfonic acid group (
l) of SPI membranes are
summarized in Table 3. The IEC values measured by titration were
in good agreement with the theoretical ones. Although these
membranes were designed with fixed sulfonation degree of 75%,
the SPI-N membranes gave the theoretical IEC values of
1.87e1.95 meq/g, which were much higher than that of SPI-B
membranes (1.61e1.67 meq/g). This is related with the difference in
molecular weight of NTDA and BNTDA.
It is well known that the water uptake mainly depends on the
IEC of the membrane.[34] For SPI membranes, higher IEC value
usually leads to the larger water uptake. As expected, the water
Fig. 4. FTIR spectra of SPI membranes.

3894
F. Sun et al. / Polymer 51 (2010) 3887e3898
Fig. 5. 1H NMR spectra of (a) SPI-N3 and (b) SPI-B3 in DMSO-d6.
uptakes of SPI-N membranes were much higher than that of SPI-B
membranes. The water uptake is also affected by the polymer
structure to a certain extent. As comparing the SPI membranes,
which have different molecular structures, it is also found that
SPI-N1 and SPI-N2, which have higher fluorine content than SPI-N3
but no significant difference in IEC values, exhibited lower water
uptakes. This could be explained by the existence of hydrophobic
CF3 groups in the polymer backbone, which resulted in the
reduction of water adsorption. However, the SPI-B1 and SPI-B2
showed the water uptakes similar to SPI-B3, although the former
have higher fluorine content. We suspect that it is related to the
twisted noncoplanar naphthalene structure in polymer backbone,
Table 1
Solubility of sulfonated polyimides in TEA salt form and in proton form.a
SPIs
Form
m-Cresol
NMP
DMAc
DMF
DMSO
SPI-N1
TEAþ
þþ
þ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
þþ
Hþ
Table 2
SPI-N2
SPI-N3
SPI-B1
SPI-B2
SPI-B3
TEAþ
Hþ
þþ
þ
Mechanical properties of SPI membranes.
Membranes
Young’s Modulus
(GPa)
Tensile Strength
(MPa)
Elongation at Break
(%)
TEAþ
Hþ
þþ
þþ
þþ
þþ
þþ
þ
TEAþ
Hþ
SPI-N1
SPI-N2
SPI-N3
SPI-B1
SPI-B2
SPI-B3
1.85
2.06
1.89
1.97
1.76
1.94
0.25
72.7
74.6
69.5
55.0
51.8
50.0
43.0
10.3
9.5
9.7
8.5
10.6
6.8
TEAþ
Hþ
TEAþ
Hþ
þþ
þ
Nafion 115
225
a
þþ: Soluble; þ: Partially soluble.

F. Sun et al. / Polymer 51 (2010) 3887e3898
3895
Table 3
Physical properties of SPI membranes.
Membranes
IEC (meq/g)
Calculated
F Content (wt.%)
Wu (wt.%)
l
Dimensional Change
tc D c
Measured
D
l
SPI-N1
SPI-N2
SPI-N3
SPI-B1
1.87
1.91
1.95
1.61
1.65
1.67
1.89
0.91
1.81
1.86
1.88
1.57
1.61
1.60
1.82
e
8.86
9.07
5.55
7.66
7.82
4.77
5.39
e
62.4
67.8
73.7
42.6
48.5
45.5
78.4
24.0
19
20
22
15
17
16
24
15
0.12
0.18
0.22
0.09
0.07
0.13
0.16
0.16
0.15
0.15
0.17
0.09
0.11
0.11
0.17
0.08
SPI-B2
SPI-B3
SPI-B30
Nafion115
which induced the decrease of the interaction of polymer chain and
the increase of inter-chain spacing. Therefore, there were more
water molecules occupied in the inter-chain spacing and adsorbed
to the sulfonic acid groups [41]. We have also prepared SPI-B30
membrane with similar structure to SPN-B3 except the sulfonation
degree increased to 87.5%. We noticed that the water uptake of SPI-
B30 (78.4%) was higher than that of SPI-N, although they have close
IEC values. It is also demonstrated that more water molecules
adsorbed to per sulfonic acid group in the SPI-B membranes.
membranes displayed much better water stability than SPI-N
membranes. This is not only due to their low IEC level but also
related to the introduction of BNTDA. It is known that electron
density of the carboxyl carbon atoms in BNTDA-based SPI was
much higher than that of NTDA-based SPI because of the lower
electron affinity of BNTDA, which restrained the nucleophilic attack
of the water molecules on the carboxyl groups. A similar result was
also reported by Ding et al. as they investigated the other BNTDA-
based sulfonated polyimides [36]. Furthermore, it is noted that
SPI-B2 membrane exhibited the best water stability, which did not
break into pieces until the membrane soaking in boiling water for
500 h. This can be explained by the flexible main chain of SPI. The
non-sulfonated diamine 6FAPB is more flexible than 6FBAB and
BAPB due to the three phenyl rings linked with ether bonds, which
providing much more flexible main chain for SPI-B2. It has been
reported that a flexible chain can undergo relaxation more easily
than the rigid ones [25]. Therefore, SPI-B2 displayed better water
stability than the others.
The number of water molecule per acid group (l) was deter-
mined from water uptake and measured IEC value by equation
l
¼ n
(H2O)/n(SOꢁ3 ) ¼ WU/18$IEC. It is found that SPI-N membranes
showed very similar
l values of 19e22, which were much higher
than that of Nafion 115 (
molecules adsorbed to per sulfonic acid groups. The SPI-B
membranes also exhibited close values of 15e17, which were
lower than SPI-N membranes because of their relatively low IEC.
But the values of SPI-B were still comparable to Nafion 115.
l
¼ 15). It is suggested that more water
l
l
The dimensional changes of the membranes equilibrated in
water at 30 ꢀC were evaluated and the results are listed in Table 3. In
general, most of the SPI membranes reported in the literature
exhibit anisotropic swelling in water [42], which is similar to the
swelling behavior observed for Nafion 115. However, both of SPI-N
and SPI-B membranes showed isotropic membrane swelling in
water because of the rigid aromatic pendant groups in the polymer
structure. It is considered that the large (or long) rigid aromatic
pendant groups make the orientation of polymer chains in plane
direction difficult, resulting in rather random orientation of poly-
mer chains [33]. Moreover, the SPI-B membranes displayed better
dimensional stability than SPI-N membranes with changes in
thickness and in plane direction no more than 13%, which is related
to the low water uptakes of SPI-B.
Membrane stability towards oxidation was examined by
observing the dissolving behavior in Fenton’s reagent at 30 ꢀC. The
oxidation stability of the membranes was characterized by elapsed
time that the membrane started to dissolve in the solution (
s1) and
dissolved completely in the solution ( 2). As shown in Table 4, all
s
SPI-N and SPI-B membranes displayed fairly good stability to
oxidation due to the wholly aromatic sulfonated diamine. The
elapsed time of the membranes started to dissolve and dissolved
completely in the solution were in the ranges of 21e48 h and
36e76 h, respectively. It is also found that SPI-N membranes
showed better stability to oxidation than the corresponding
BNTDA-based membranes. This can be interpreted by the relatively
high content of hydrophobic trifluoromethlyl groups, which could
protect the polymer main chain from being attacked by water
molecules containing highly oxidizing radical species [45]. It is
known that the fluorinated substituents in the polymer backbone
are effective in the improvement of oxidation stability of sulfonated
polyimide. However, we noted that the SPI-N1 and SPI-B1, both of
3.4. Membrane stabilities
The water stability of membranes is one of the most important
factors affecting the membrane performance. It is considered that
water stability of membranes is affected by the chemical structure,
configuration and the basicity of sulfonated diamine [43]. As shown
in Table 4, all the SPI-N and SPI-B membranes displayed reasonably
good water stability. The membranes could keep their original
shape until soaking in boiling water for 191e500 h. The good water
stability is related to the side-chain-type structure of sulfonated
diamine BABSA. The long side chain with extremely hydrophilic
sulfonic acid groups could make the water molecules mainly
gathering around the lateral chains and protect the main-chain
imide rings from hydrolysis. Moreover, the sulfonated diamine
BABSA displayed relatively high basicity because the sulfonic acid
groups appending in the side chains were separated from the
amine groups. The high basicity of the sulfonated diamine lowered
the positive electricity of carbonyl groups in imide rings, and
therefore, the hydrolytic reaction was depressed [44]. The SPI-B
Table 4
The water stability, oxidative stability and thermal stability of SPI membranes.
Membranes
Water
Stability (h)
Oxidative
Stabilitya
Thermal Stabilityb
s1 (h)
s2 (h)
Td1 (ꢀC)
Td2 (ꢀC)
SPI-N1
SPI-N2
SPI-N3
SPI-B1
SPI-B2
SPI-B3
191
237
290
264
501
468
47
37
29
48
24
21
76
47
49
61
36
44
299.1
305.0
329.2
283.9
286.4
300.7
551.8
553.8
553.9
554.3
559.6
560.2
a
s1 and s2 refer to the elapsed time that the membranes began to dissolve and
dissolved in the solution, respectively.
b
Td1 and Td2 refer to the decomposition temperature of sulfonic acid group and
the decomposition temperature of polymer backbone, respectively.

3896
F. Sun et al. / Polymer 51 (2010) 3887e3898
which derived from non-sulfonated diamine 6FBAB, displayed
much better durability in Fenton’s regent with longer time of the
membranes started to dissolve than the others, despite their fluo-
rine contents were not as high as SPI-N2. This may be related with
the combined effects of high fluorine content and rigid biphenyl
structure of 6FBAB in the polymer backbone.
The thermal stability of SPIs was evaluated by TGA measure-
ment. As shown in Fig. 6, all the membranes exhibited three-step
weight loss. The first weight loss below 150 ꢀC is due to the evap-
oration of water absorbed in the polymers; the second one in the
range of 280e330 ꢀC is attributed to the thermal decomposition of
sulfonic acid groups; the third one up to 550 ꢀC is assigned to the
degradation of polymer main chain. The thermal stability of SPIs
mainly depends on the cleavage temperature of sulfonic acid
groups. The SPI-N membranes showed better heat resistance as
comparing with the corresponding SPI-B membranes with the
desulfonation temperature of 299e329 ꢀC, which was much higher
than the common main-chain-type sulfonated polyimides and
some side-chain-type ones [46].
Fig. 7. Temperature dependence of proton conductivity for SPI and Nafion 115
membranes at RH 100%.
than Nafion 115, therefore, the
s values for the former increased
3.5. Proton conductivity
faster than that for Nafion 115 with the temperature rising.
The proton conductivities for SPI-B membranes in water at 30 ꢀC
were relatively lower than that for SPI-N and Nafion 115
Proton conductivity is the most important factor that affects the
performance of a fuel cell system. In this study, the proton
conductivity of the SPI membranes was measured in the temper-
ature range of 30e80 ꢀC at 100 RH%. For comparison, the proton
conductivity of Nafion 115 was also determined at the same
conditions. Fig. 7 shows the temperature dependence of proton
conductivity for SPI-N, SPI-B and Nafion 115 membranes. The
membranes with the
SPI-B membranes showed lower
s
values of 0.056e0.074 S/cm. Although the
values than Nafion 115 in the
s
tested temperature range, the difference in proton conductivity
between them gradually decreased with the temperature
increasing owing to their relatively higher activation energies
(10e16 kJ/mol). The proton conductivities of SPI-B membranes at
80 ꢀC were close to that of Nafion 115 with the
s values of
proton conductivities (
are summarized in Table 5.
The values of all the membranes increased with the temper-
s
) measured at 30 ꢀC for these membranes
0.13e0.14 S/cm. The low conducting performance of SPI-B
membranes is associated with their lower IEC values. We have
investigated the proton conductivity of the above mentioned SPI-
B30 membrane, whose IEC value was close to SPI-N membranes. It is
s
ature rising. The proton conductivities of SPI-N membranes in
water at 30 ꢀC were in the range of 0.086e0.095 S/cm, which is
comparable to that of Nafion 115 (0.095 S/cm). It is noteworthy that
the SPI-N membranes exhibited better proton conductivity than
Nafion 115 as the temperature over 40 ꢀC. The proton conductivities
for SPI-N measured at 80 ꢀC increased to 0.18e0.22 S/cm, while that
for Nafion 115 was 0.15 S/cm. The high conducting performance of
SPI-N membranes can be explained from the view of activation
energy of proton conductivity. The proton conductivity of
membranes showed an Arrhenius-type temperature dependence.
found that the SPI-B30 membrane gave the
s value of 0.093 S/cm at
30 ꢀC and showed a similar increasing tendency with SPI-N3 in the
tested temperature range.
3.6. Methanol permeability
Methanol permeability is a key parameter for membranes using
in DMFC. To achieve high fuel cell performance, membranes should
have not only high proton conductivity but also low methanol
permeability. The methanol permeability (PM) measured at 30 ꢀC
with different methanol content in feed and the ratios (V) of proton
conductivity to methanol permeability of SPI and Nafion 115
membranes are summarized in Table 5.
The activation energies of proton conductivity (DEaC) for SPI-N
membranes were 11e17 kJ/mol, while the value for Nafion 115 was
8 kJ/mol. The SPI-N membranes showed higher activation energies
All the SPI membranes showed much lower methanol perme-
ability than Nafion 115. The PM values for SPI-N and SPI-B
membranes tested at 30 ꢀC with 30 wt.% of methanol content in
feed, decreased to 50% and 25% of that for Nafion 115. With an
Table 5
Proton conductivity, methanol permeability and the ratio of proton conductivity to
methanol permeability of SPI and Nafion 115 membranes.
Membranes
s
(S/cm)
PM (10ꢁ6 cm2/s)
F
(104 S cmꢁ3 s)
30%
50%
30%
50%
SPI-N1
SPI-N2
SPI-N3
SPI-B1
SPI-B2
SPI-B3
0.086
0.088
0.095
0.074
0.056
0.070
0.095
0.81
0.74
0.74
0.40
0.45
0.47
1.58
0.95
0.98
0.76
0.50
0.45
0.46
1.91
10.6
11.9
12.8
18.5
12.4
15.0
6.0
9.1
9.0
12.5
14.8
12.5
15.3
5.0
Nafion115
Fig. 6. TGA curves of SPI membranes (in nitrogen).

F. Sun et al. / Polymer 51 (2010) 3887e3898
3897
Fig. 8. TEM micrographs of SPI membranes (cross section, in Agþ form).
increase in methanol content in feed from 30 wt.% up to 50 wt.%,
the PM values for Nafion 115 increased from 1.58 ꢂ 10ꢁ6 cm2/s to
1.91 ꢂ 10ꢁ6 cm2/s. On the other hand, the PM values for SPI-N and
SPI-B membranes showed hardly dependence on the methanol
content in feed, which slightly increased for the former and kept
almost same for the latter. As comparing the PM values of SPI
membranes with different structures, it is found that SPI-B dis-
played lower methanol permeability than SPI-N. The difference in
methanol permeability between the SPI membranes can be
explained by the dissimilarity in their microstructure. It is known
that methanol mainly permeates through the hydrophilic domains
consisting of sulfonic acid groups. The Nafion membranes is known
to have well-developed hydrophobic/hydrophilic microphase
separation structure. The well connected hydrophilic domains
provide not only proton conducting channel but also methanol
permeating pathways. However, the hydrophobic/hydrophilic
microstructure in SPI membranes is less separated due to their rigid
and less hydrophobic backbone as well as the less acidic sulfonated
functional groups. Therefore, the narrower transport channels in
SPI membranes result in the lower methanol permeability
compared with Nafion membranes.
It is also found that the SPI-N1 and SPI-N2 membranes, which
derived from fluorine-containing non-sulfonated diamines 6FBAB
and 6FAPB, respectively, gave higher PM values and displayed
certain dependence on the concentration of methanol in feed than
the other SPI membranes. This is considered to be related to their
better hydrophobic/hydrophilic microphase separation structure
because of the presence of hydrophobic trifluoromethyl groups in
the polymer backbone. The good hydrophobic/hydrophilic micro-
structure played a promoting role in not only proton but also
methanol transport in membranes. The microstructures of SPI
membranes will be discussed below in more details.
The ratio V, which calculated from
s to PM, is an effective
parameter evaluating the membranes performance in a DMFC
system. All the SPI membranes displayed fairly high V values
ranging from 10.6 ꢂ 104 S cmꢁ3 s to 18.5 ꢂ104 S cmꢁ3 s as calculated
from the PM values with 30 wt.% of methanol content in feed. Their
V values were almost two to three times of the Nafion 115. More-
over, the SPI-B membranes gave the higher V values than SPI-N
membranes. There is no significant dependence of V values on the
methanol concentration in feed were observed for all the
membranes. The good permselectivity of SPI membranes is

3898
F. Sun et al. / Polymer 51 (2010) 3887e3898
attributed to their much smaller methanol permeabilities as
compared with Nafion 115.
same sulfonation degree. The proton conductivities of SPI
membranes were increased with temperature. The SPI-N
membranes exhibited higher proton conductivity than Nafion 115
3.7. Membrane morphology
as the temperature over 40 ꢀC and their
s values increased to
0.18e0.22 S/cm at 80 ꢀC. The highly conducting performance of
SPI-N membranes is attributed to their good hydrophobic/hydro-
philic microphase separation structure.
It is considered that there is a close contact between the micro-
phase separation structure and the proton conductivity of the
membranes. In order to get a comprehensive understanding of the
proton conductivity behavior of membranes, we tried to investigate
the microstructureof the SPI membranes byTEM analysis. Thecross-
section micrographs of SPI-N and SPI-B are shown in Fig. 8, in which
the dark regions represent hydrophilic ionic domains and the bright
regions refer tohydrophobic moieties. It is found that both SPI-Nand
SPI-B membranes exhibited clear microphase separation structures.
In the TEM micrographs of SPI-B membranes, a large amount of
bigger ionic clusters (6e8 nm) and a certain amount of smaller ionic
clusters (2e3 nm) were observed combined with medium size
clusters. This is attributed to the twisted non-coplanar naphthalene
structure in polymer backbone, which is favorable for the aggre-
gation of hydrophilic sulfonic acid groups to form larger and smaller
ionic clusters. On the contrary, the TEM micrographs of SPI-N
membranes showed the spherical ionic clusters in a uniform
distribution with the average size of 4e5 nm. The clusters were
close to each other, which is favorable for water keeping and proton
transport. This is the reason why SPI-N membranes exhibited better
proton conductivities and high water uptakes than SPI-B
membranes, although they were synthesized with the same sulfo-
nation degree. Moreover, the hydrophilic domains in SPI-N1 and
SPI-N2 seemed smaller and with better connectivity as comparing
with SPI-N3. This is considered to be related to the hydrophobic
trifluoromethyl groups in the polymer backbone, which is helpful to
the formation of microphase separation structure. From the TEM
observation combined with the proton conductivities of SPI
membranes, it can be concluded that the good microphase sepa-
ration structure is contribute to their better proton-conducting
performance. The well-balanced properties of SPI membranes are
promising candidates as proton exchange membrane for DMFC
applications. The DMFC performance of these SPI membranes will
be investigated in the future work and reported later.
References
[1] Hogarth MP, Hards GA. Platinum Met-Rev 1996;40(4):150e9.
[2] Zhang Y, Wan Y, Zhao C, Shao K, Zhang G, Li H, et al. Polymer 2009;50
(19):4471e8.
[3] Zhai F, Guo X, Fang J, Xu H. J Membr Sci 2007;296:102e9.
[4] Okamoto K, Yin Y, Yamada O, Islam MN, Honda T, Mishima T, et al. J Membr Sci
2005;258(1e2):115e22.
[5] Mauritz KA, Moore RB. Chem Rev 2004;104(10):4535e86.
[6] Asano N, Aoki M, Suzuki S, Miyatake K, Uchida H, Watanabe M. J Am Chem Soc
2006;128(5):1762e9.
[7] Heinzel A, Barragán VM. J Power Sources 1999;84(1):70e4.
[8] Kreuer KD. J Membr Sci 2001;185(1):29e39.
[9] Gurau B, Smotkin ES. J Power Sources 2002;112(2):339e52.
[10] Hobson LJ, Nakano Y, Ozu H, Hayase S. J Power Sources 2002;104(1):79e84.
[11] DeLuca NW, Elabd YA. J Polym Sci Part A Polym Phys 2006;44:2201e25.
[12] Ye YS, Yen YC, Cheng CC, Chen WY, Tsai LT, Chang FC. Polymer 2009;50
(14):3196e203.
[13] Jones DJ, Roziere J. J Membr Sci 2001;185(1):41e58.
[14] Chu PP, Wu CS, Liu PC, Wang TH, Pan JP. Polymer 2010;51(6):1386e94.
[15] Wang F, Hickner M, Kim YS, Zawodzinski TA, McGrath JE. J Membr Sci
2002;197(1e2):231e42.
[16] Miyatake K, Chikashige Y, Higuchi E, Watanabe M.
J Am Chem Soc
2007;129:3879e87.
[17] Staiti P, Lufrano F, Aricò AS, Passalacqua E, Antonucci V. J Membr Sci 2001;188
(1):71e8.
[18] Song JM, Miyatake K, Uchida H, Watanabe M. Electrochim Acta
2006;51:4497e504.
[19] Zou L, Roddecha S, Anthamatten M. Polymer 2009;50(14):3136e44.
[20] Somboonsub B, Invernale MA, Thongyai S, Praserthdam P, Scola DA,
Sotzing GA. Polymer 2010;51(6):1231e6.
[21] Li N, Liu J, Cui Z, Zhang S, Xing W. Polymer 2009;50(19):4505e11.
[22] Genies C, Mercier R, Sillion B, Cornet N, Gebel G, Pineri M. Polymer 2001;42
(2):359e73.
[23] Guo X, Fang J, Watari T, Tanaka K, Kita H, Okamoto K. Macromolecules
2002;35(17):6707e13.
[24] Meyer G, Perrot C, Gebel G, Gonon L, Morlat S, Gardette J-L. Polymer 2006;47
(14):5003e11.
[25] Fang J, Guo X, Harada S, Watari T, Tanaka K, Kita H, et al. Macromolecules
2002;35:9022e8.
[26] Vallejo E, Pourcelly G, Gavach C, Mercier R, Pineri M. J Membr Sci 1999;160
(1):127e37.
4. Conclusions
[27] Genies C, Mercier R, Sillion B, Petiaud R, Cornet N, Gebel G, et al. Polymer
2001;42(12):5097e105.
[28] Yin Y, Yamada O, Tanaka K, Okamoto K. Polym J 2006;38(3):197e219.
A novel side-chain-type sulfonated aromatic diamine, 5-[1,1-bis
(4-aminophenyl)-2,2,2- trifluoroethyl]-2-(4-sulfophenoxy)benze-
nesulfonic acid (BABSA), was synthesized and characterized. Two
series of sulfonated polymides (SPI-N and SPI-B) were synthesized
from six-membered dianhydrides NTDA or BNTDA, sulfonated
aromatic diamine BABSA and various non-sulfonated aromatic
diamines. These membranes displayed good solubility in many
common organic solvents even in proton form because of the
synergetic effects of the bulky aromatic pendant groups and the
fluorinated substituents in the polymer structure. The flexible and
strong sulfonated polyimide membranes SPI-N and SPI-B exhibited
good dimensional stability with isotropic swelling less than 22%
and high thermal stability with the desulfonation temperature of
283e330 ꢀC. Their water stability is greatly affected by the flexi-
bility of polymer main chain. These membranes also displayed
excellent oxidation stability, which could be improved by the
incorporation of trifluoromethyl groups in polymers backbone. All
the SPI membranes have better permselectivity with the ratios of
proton conductivity to methanol permeability (V) of nearly two to
three times of that for Nafion 115. The SPI-N membranes based on
NTDA showed much better proton conductivity and slightly higher
methanol permeability than SPI-B membranes based on BNTDA
because of the higher IEC values of the former, despite they have
[29] Yin Y, Fang J, Watari T, Tanaka K, Kita H, Okamoto K.
J Mater Chem
2004;14:1062e70.
[30] Hu Z, Yin Y, Kita H, Okamoto K, Suto Y, Wang H, et al. Polymer
2007;48:1962e71.
[31] Li N, Cui Z, Zhang S, Xing W. J Membr Sci 2007;295:148e58.
[32] Sutou Y, Yin Y, Hu Z, Chen S, Kita H, Okamoto K, et al. J Polym Sci Part A Polym
Chem 2009;47(5):1463e77.
[33] Hu Z, Yin Y, Chen S, Yamada O, Tanaka K, Kita H, et al. J Polym Sci Part A Polym
Chem 2006;44:2862e72.
[34] Chen K, Chen X, Yaguchi K, Endo N, Higa M, Okamoto K. Polymer 2009;50
(2):510e8.
[35] Miyatake K, Zhou H, Matsuo T, Uchid H, Watanabe M. Macromolecules
2004;37:4961e6.
[36] Yan J, Liu C, Wang Z, Xing W, Ding M. Polymer 2007;48(21):6210e4.
[37] Mitchell RH, Lai Y-H, Williams RV. J Org Chem 1979;44(25):4733e5.
[38] Li N, Cui Z, Zhang S, Xing W. Polymer 2007;48:7255e63.
[39] Cao J, Wang Z. J Polym Sci Part A Polym Chem 1995;33:1627e35.
[40] Savard O, Peckham TJ, Yang Y, Holdcroft S. Polymer 2008;49(23):4949e59.
[41] Li N, Cui Z, Zhang S, Li S, Zhang F. J Power Sources 2007;172:511e9.
[42] Yin Y, Yamada O, Suto Y, Mishima T, Tanaka K, Kita H, et al. J Polym Sci Part A
Polym Chem 2005;43:1545e53.
[43] Fang J, Guo X, Xu H, Okamoto K. J Power Sources 2006;159(1):4e11.
[44] Chen S, Yin Y, Kita H, Okamoto K. J Polym Sci Part A Polym Chem 2007;45
(13):2797e811.
[45] Qiu Z, Wu S, Li Z, Zhang S, Xing W, Liu C. Macromolecules 2006;39
(19):6425e32.
[46] Watari T, Fang J, Tanaka K, Kita H, Okamotoa K, Hirano T. J Membr Sci
2004;230:111e20.