'Click'-functionalization of poly(sulfone)s and a study of their utilities as proton conductive membranes in direct methanol fuel cells

Article abstract of DOI:10.1016/j.polymer.2010.09.041

Using the copper-catalyzed 1,3-dipolar "click" cycloaddition reaction, poly(sulfone)s containing pendant azide moieties were functionalized with various quantities of sodium 3-(prop-2-ynyloxy)propane-1-sulfonate and crosslinked with 1,7-octadiyne. The deg

Full text of DOI:10.1016/j.polymer.2010.09.041

Polymer 51 (2010) 5352e5358
Contents lists available at ScienceDirect
Polymer
journal homepage: www.elsevier.com/locate/polymer
‘Click’-functionalization of poly(sulfone)s and a study of their utilities as proton
conductive membranes in direct methanol fuel cells
a
b
b
b
a,
*
Brent C. Norris , Wen Li , Eungje Lee , Arumugam Manthiram , Christopher W. Bielawski
a
Department of Chemistry and Biochemistry, The University of Texas at Austin, 1 University Station, A5300, Austin, TX 78712, USA
Materials Science and Engineering Program, The University of Texas at Austin, Austin, TX 78712, USA
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 28 June 2010
Received in revised form
Using the copper-catalyzed 1,3-dipolar “click” cycloaddition reaction, poly(sulfone)s containing pendant
azide moieties were functionalized with various quantities of sodium 3-(prop-2-ynyloxy)propane-1-
sulfonate and crosslinked with 1,7-octadiyne. The degrees of sulfonation and crosslinking were
systematically varied by changing the ratios of the aforementioned reagents. The polymers were cast into
membranes, acidified, and then tested for proton conductivity, methanol permeability, and membrane-
electrode assembly (MEA) performance. The membranes showed a reduction in methanol permeability
with increasing concentration of crosslinker and exhibited performance on par with direct methanol fuel
cells containing Nafion-based membranes.
12 September 2010
Accepted 15 September 2010
Available online 8 October 2010
Keywords:
Polysulfone
Fuel cells
Ó 2010 Elsevier Ltd. All rights reserved.
Click chemistry
1. Introduction
To enhance conductivity, various N-heterocycles, such as imid-
azole or triazole, may be added to the aforementioned membranes
Direct methanol fuel cells (DMFC) provide a convenient source
of power because they do not require recharging with an electrical
outlet and, unlike hydrogen fuel cells, they use a relatively less
volatile liquid fuel that is easy to store and transport [1,2]. Unfor-
tunately, current DMFC technology is hampered by the sluggish
methanol oxidation kinetics and the crossover of methanol fuel
from the anode to the cathode through the membrane, which leads
to fuel loss, cathode catalyst poisoning, and cell voltage drop [3].
The most common electrolyteused in fuel cells, includingDMFCs,
is a sulfonated fluoropolymer called Nafion. This polymer is known
to exhibit high proton conductivities (100e150 mS/cm; typical
[12]. The N-heterocycles are believed to function as proton carriers
and have been shown to improve the conductivities of membranes
which contain them under a range of humidities [13]. However, the
added N-heterocycles often leach from the polymer membrane when
ꢀ
theyareoperatedattemperatureslessthan100 C,whereliquidwater
is present, and some of them also poison the cathode catalyst [14]. To
alleviate this issue, various nitrogen-containing bases have been
covalently linked to polymer chains within the membrane [15];
however, such approaches often entail complicated syntheses and/or
require multiple post-polymerization modifications.
Another strategy to reduce methanol permeability while
maintaining high proton conductivities has been to crosslink the
polymer chains in the membrane [16]. It is believed that the
crosslinks formed limit membrane swelling and decrease the size
of the proton conducting channels in the hydrated state. Smaller
channels slow the diffusion of methanol and ultimately lower
methanol crossover. However, the post-polymerization cross-
linking reaction is practically challenging because it must be per-
formed during or after membrane casting. One successful strategy
for increasing the proton conductivity while decreasing the
methanol permeability is through the synthesis of membranes
comprised of block copolymers [17]. In particular, block copolymers
with hydrophilic and hydrophobic segments that are capable of
phase separation often generate organized highly proton conduc-
tive hydrophilic regions [17j].
ꢀ
operating conditions: 65 C, 100% relative humidity), but is expen-
ꢂ6
2
sive and suffers from high methanol permeability (1.2 ꢁ10 cm /s)
[4]. Sulfonation of commercially-available polyaromatic materials,
such as poly(sulfones) [5e7] and poly(ether ether ketones) [8e11],
has been demonstrated to afford protonconductive membranes that
exhibit reduced methanol permeability relative to Nafion
ꢂ7
2
(
1e9 ꢁ10 cm /s). One drawback, however, is that these sulfonated
membranes are hindered by relatively low proton conductivities
11e17 mS/cm), particularly under low humidity conditions [10]
where there is insufficient water present to act as a proton carrier.
(
*
Corresponding author.
E-mail address: bielawski@cm.utexas.edu (C.W. Bielawski).
0
032-3861/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.polymer.2010.09.041

B.C. Norris et al. / Polymer 51 (2010) 5352e5358
5353
Herein we describe a method that overcomes many of the
aforementioned challenges and enables simultaneous installation of
tethered N-heterocycles (to enhance proton transfer) and covalent
crosslinking (to reduce swelling and methanol crossover) [18,19].
The procedure utilizes the copper-catalyzed azide-alkyne 1,3-
dipolar cycloaddition (CuCAAC) reaction between poly(sulfone)s
containing pendant azides with alkyne additives. This “click” reac-
tion [20,21] was selected because it not only proceeds to high
conversions, but also is highly selective, tolerant to many functional
The membranes were thenweighed and measured. This process was
repeated until the weight of the membrane reached a constant
value. The percent uptake was calculated using the following
equation: % uptake ¼ 100 ꢁ (wet weight ꢂ dry weight)/dry weight.
The percent swelling was calculated using the following equation: %
swelling ¼ 100 ꢁ (wet length ꢂ dry length)/wet length.
2.2. Methanol permeability measurements
groups and generates Brønsted basic 1,2,3-triazoles (pK
which were envisioned to serve as the N-heterocycles [13,23] in the
proton shuttling processes described above.
B
: 0e1) [22],
Methanol permeability measurements were conducted in
a glass cell consisting of two chambers, each with a total volume of
100 mL [26]. Magnetic stir bars were added to each chamber. The
membrane was sandwiched between two rubber circular gaskets
(internal diameter of 3 cm) and then tightly clamped together. One
side of the cell was filled with 80 mL of 1 M methanol (aq.) and
ethanol ([ethanol]0 ¼ 0.01 M) as an internal standard. The other
side (analyte) was filled with 80 mL of 0.01 M ethanol (aq.) The
chambers were then sealed with septa and each side was stirred
with a magnetic stir bar. The concentration of methanol in the
analyte side was measured by gas chromatography and integrated
against the ethanol internal standard over time. The methanol
permeability was calculated according to equation (1), where Ca
and Cb refer to the methanol concentration in the feed and the
permeate, respectively, Vb refers to the solution volume of
permeate, and L, A, and t refer, respectively, to the membrane
thickness, membrane area, and time [27].
2
. Experimental
2.1. General considerations
N,N-Dimethylacetamide (DMAc) was purchased from Fisher
Scientific, 1,3-propanesultone was purchased from MP Biomedicals
LLC, and propargyl alcohol and 1,7-octadiyne were purchased from
Ò
Acros Organics. Udel P-1700 was purchased from Solvay Advanced
Ò
Polymers. Nafion 117 was purchased from DuPont. All reagents
were used as received. Membranes were cast from solution in
a 10 cm diameter custom-made flat-bottom sealed vessel under an
atmosphere of nitrogen. Infrared (IR) spectra were recorded on
a Perkin Elmer Spectrum BX FTIR spectrophotometer. High resolu-
tion mass spectra (HRMS) were obtained with a VG analytical ZAB2-
E instrument (CI). NMR spectra were recorded on Varian UNITYþ
P ¼ Cb
V L
b
(1)
3
00, Varian Mercury 400, and Varian INOVA 500 spectrometers.
) are given in ppm and are referenced downfield
AC t
a
Chemical shifts (
from residual solvent ( H: DMSO-d
9.5 ppm). Molecular weights were determined by gel permeation
d
1
13
6
, 2.49 ppm; C: DMSO-d
6
,
2.3. Syntheses
3
chromatography (GPC) using a Waters HPLC system consisting of
three Viscotek I-series columns (2 ꢁ GMHHRH and 1 ꢁ G3000HHR)
arranged in series, a 1515 pump, and a 2414 RI detector and are
2.3.1. Sodium 3-(prop-2-ynyloxy)propane-1-sulfonate (1)
A 200 mL round bottom flask was charged with a magnetic stir
bar, sodium hydride (95%; 1.03 g, 40.9 mmol) and N,N-dime-
thylformamide (DMF) (30 mL). The flask was cooled in an ice bath
and a solution of propargyl alcohol (2.39 mL, 40.9 mmol) in DMF
(30 mL) was slowly added under continuous stirring over 10 min.
A solution of 1,3-propanesultone (5.00 g, 40.9 mmol) in DMF (30 mL)
was then added slowly. The resulting mixture was stirred on ice for
reported relative to polystyrene standards in DMF (0.1 M LiBr) at
ꢀ
4
0
C (column temperature). Melting points were obtained using
a Mel-Temp apparatus and are uncorrected. Thermogravimetric
analyses were performed on a Mettler Toledo TGA/SDTA851e and
differential scanning calorimetry analyses were performed on
a Mettler Toledo DSC 823e. The IEC (ion exchange capacity) was
ꢀ
10 min, warmed to 60 C and then stirred for an additional 2 h. The
ꢀ
measured by drying the polymer in a vacuum oven at 70 C and then
reaction was then concentrated under reduced pressure (with the
ꢀ
stirring in a 2 M NaCl solution for 1 h. The resulting polymers were
then titrated with an 8 mM solution of NaOH using phenolphthalein
as an endpoint indicator [24]. Proton conductivity values of the
membranes were obtained from the impedance data, which were
collected with a computer interfaced Solartron SI 1260 Impedance
Gain Phase Analyzer coupled to a SI 1287 Electrochemical Interface.
Samples were measured in the frequency range of 5 Hze13 MHz
with an applied voltage of 10 mV. The impedance measurement was
carried out using a home-made two-electrode setup and stainless
steel foil was used as the electrodes. The impedance was measured
in the plane of the membrane sample and the cell size used was
aid of a rotary evaporator) at 60 C and diethyl ether (500 mL) was
added which caused the product to precipitate. The desired product
(7.9 g, 97% yield) was collected via filtration as a white powder. Note:
in the solid state,1 was found to develop a reddish color and became
insoluble over time. However, the material was found to be stable
under ambient conditions when dissolved in N,N-dimethylaceta-
ꢀ
mide (DMAc) ([1]
0
¼ 0.503 M). m.p.120e150 C (the material turned
1
from white to dark red). H NMR (400 MHz, DMSO-d
J ¼ 2.4 Hz, 2H), 3.4 (t, 2H, overlaps with H O), 3.35 (t, J ¼ 2.4 Hz,1H),
2.47 (2H, m, overlaps with DMSO), 1.78 (2H, m). C NMR (100 MHz,
6
): d 4.05 (d,
2
13
D
2
O; referenced to an internal methanol standard, 49.5 ppm):
d
80.1,
ꢂ
of the following dimensions:
1
cm (width)
ꢁ
1
cm
76.5, 69.0, 58.2, 48.5, 24.8. HRMS [M ] calcd. for C
6
9
H O
4
S,177.02215;
(
length) ꢁ 50e200 m (thickness). The selectivities were calculated
m
found 177.02282. IR (KBr):
2112, 1638, 1199, 1066, 630 cm
n
¼ 3478 (broad), 3290, 2945, 2870,
ꢂ1
by determining the quotient of a membrane’s conductivity over
permeability, as described previously [25]. The relative selectivities
were calculated by determining the quotient of the selectivities of
the new membranes described below over that measured for Nafion
.
2.3.2. Polysulfone containing pendant azides (2)
The azide-functionalized poly(sulfone) 2 was synthesized using
a modified literature procedure [28]. A 1 L flask was charged with
poly(sulfone) (UdelÒ P-1700) (6.0 g, 13.5 mmol based on its repeat
117. Liquid uptake and swelling were measured by drying the
membranes in a vacuum oven for 12 h before measuring their dry
weight and thickness. After soaking the membranes in water or 1 M
methanol for various amounts of time, they were removed from the
aforementioned solutionsandtheir surfaces weregentlywiped with
a paper towel to remove the residual liquid present on the surface.
unit, M
w
¼ 45.6 kDa, PDI ¼ 2.01), a Teflon coated magnetic stir bar
and THF (400 mL) under nitrogen. After cooling the reaction
ꢀ
mixture to ꢂ78 C, n-butyl lithium (2.5 M; 2 equiv, 10.8 mL,
27.0 mmol) was added dropwise over a period of 2 h. In a separate

5354
B.C. Norris et al. / Polymer 51 (2010) 5352e5358
flask, a solution of p-toluensulfonyl azide (7.99 g, 40.5 mmol) in THF
2.3.5. Representative crosslinked poly(sulfone) copolymer
membrane casting procedure used to prepare 5b
A 50 mL flask was charged with a magnetic stir bar, polymer 2
(550 mg, 1.05 mmol based on the molecular weight of the repeat
unit), and DMAc (5 mL). The polymer was allowed to completely
ꢀ
(
20 mL) was cooled to ꢂ40 C and then added to the aforemen-
tioned polymer solution which caused solids to precipitate. The
reaction mixture was then slowly warmed to ꢂ30 C at which point
ꢀ
it became homogeneous. A 4:3 v/v mixture of water : ethanol
(
800 mL) was then added. The precipitated solids were collected by
dissolve before adding a solution of 1 in DMAc (([1]
0
¼ 0.503 M;
filtration to afford the desired polymer (6.7 g, 95% yield) as a white
powder. Spectroscopic data were consistent with literature values
2.49 mL, 1.25 mmol), and CuI (23 mg, 0.121 mmol). The flask was
then sealed with a septum, degassed under reduced pressure for
10 min and then backfilled with nitrogen. The resulting reaction
mixture was stirred at ambient temperature for 8 h to generate
polymer 4. In a separate vial, polymer 2 (200 mg, 0.452 mmol based
on the molecular weight of the repeat unit) was dissolved in DMAc
1
[28]. H NMR (400 MHz, CDCl
3
):
d
8.16 (d, J ¼ 8.8 Hz, 2H), 7.26 (d,
J ¼ 8.9 Hz, 4H), 6.98 (d, J ¼ 8.5 Hz, 4H), 6.76 (dd, J ¼ 8.9, 2.3 Hz, 2H),
6
4
.71 (d, J ¼ 2.3 Hz, 2H), 1.71 (s, 6H). GPC (DMF with 0.1 M LiBr,
ꢀ
ꢂ1
.
0 C): M
w
¼ 56.9 kDa, PDI ¼ 2.14. IR (KBr):
y
N3 ¼ 2119 cm
(3 mL) and added to the reaction vessel containing 4 followed by
2.3.3. Synthesis of sulfonated polymer 3 (not crosslinked)
a solution of DMAc (1 mL) containing 1,7-octadiyne (56 L,
m
A 25 mL flask was charged with polymer 2 (500 mg, 0.954 mmol
0.43 mmol). The resulting mixture was degassed for 10 min under
reduced pressure, backfilled with nitrogen, and then filtered
through a cotton plug into a custom-built, air-free chamber as it
was purged with nitrogen. After the purging was ceased, the
based on the molecular weight of its repeat unit), DMAc (5 mL), and
a magnetic stir bar. The solution was stirred until the polymer
was completely dissolved. After adding a DMAc solution of 1
ꢀ
(
[1]
0
¼ 0.756M; 2.52 mL,1.908 mmol), the reactionvessel was sealed
chamber was sealed and heated at 60 C in an oven for 12 h. The
with a septum. After degassing the vessel under reduced pressure,
chamber was then removed from the oven and allowed to cool to
ambient temperature. To allow the residual solvent to evaporate,
the top of the chamber was removed and the chamber was
CuI (38.3 mg, 200
m
mol) was added. The resulting mixture was
ꢀ
stirred for 30 min at ambient temperature and then heated at 60 C
in an oil bath for 12 h. The residual solvent was removed on a rotary
ꢀ
reheated at 60 C for 6 h with the aid of small circulation fan to
ꢀ
ꢀ
evaporator at 60 C. The resulting polymer was then acidified by
facilitate drying. The membrane was then dried at 80 C for 12 h
ꢀ
heating at 80 C for 2 h in the presence of H
2
SO
4
(2 M, 100 mL).
2 4
and released from the cell by adding 100 mL of H SO (2 M;
Finally, the polymer was heated in 100 mL of de-ionized water at
100 mL). Upon removal of the membrane, it was acidified by
ꢀ
ꢀ
ꢀ
8
0 C for 2 h and then dried in avacuum oven at 70 C for 12 h toyield
heating at 80 C for 2 h in the presence of H
2
SO
4
(2 M, 400 mL) in
ꢀ
ꢀ
716 mg (86% yield) of 3. T
g
¼ 67 C. T
d
(onset) ¼ 170 C. On account of
a 1 L Erlenmeyer flask. Finally, the membrane was heated in 400 mL
ꢀ
stoichiometry of the starting materials employed, the membranes
contained residual azides, as determined by IR spectroscopy
of de-ionized water at 80 C for 2 h and then dried in a vacuum oven
ꢀ
at 70 C for 12 h (812 mg, 70% yield).
ꢂ1
D
(y
N3 ¼ 2120 cm ; KBr). Both the crude polymer 3Na (sodium
sulfonate salt) as well as its acidified derivative 3 were analyzed by
GPC. GPC of 3Na (DMF with 0.1 M LiBr, 40 C): M
D
ꢀ
ꢀ
3. Results and discussion
w
¼ 59.1 kDa;
¼ 27.5 kDa;
PDI ¼ 2.99. GPC of 3 (DMF with 0.1 M LiBr, 40 C): M
w
1
3.1. Sulfonation of poly(sulfone)s containing pendant azides via
click chemistry
PDI ¼ 1.43. H NMR of 3 (300 MHz, DMSO-d
6
): 8.12 (s, 2H), 7.49 (d,
J ¼ 8.7 Hz, 2H), 7.37 (d, J ¼ 8.4 Hz, 4H), 7.1e7.2 (m, 8H), 4.48 (s, 4H),
13
3
.52 (t, J ¼ 6.0 Hz), 1.83 (s, 4H), 1.69 (s, 6H). C NMR of 3 (100 MHz,
As a potentially suitable reagent to modify poly(sulfone)s con-
taining pendant azides, alkyne sulfonate 1 features a flexible alkyl
chain that was envisioned to distance the sulfonic acid from the
main chain of the polymer to which it may be attached. Hence, the
hydrophobic regions of the resulting polymer backbone may be
effectively separated from the hydrophilic regions and result in
better swelling properties [29]. As shown in Scheme 1, 1 was
synthesized by the nucleophilic ring-opening of 1,3-propane-
sultone with sodium propargylate under mild conditions and iso-
lated in high yield (97%) following precipitation and collection via
filtration. Although 1 was found to decompose over a period of
days, forming an insoluble red powder, a solution of this compound
in N,N-dimethylacetamide (DMAc) was found to be stable under
ambient conditions for extended periods of time.
DMSO-d ) 161.8, 152.0, 147.2, 143.8, 135.9, 131.56, 129.6, 128.7, 127.3,
6
119.7, 118.7, 118.1, 69.1, 62.9, 48.3, 42.11, 30.5, 25.7.
2.3.4. Representative crosslinked poly(sulfone) membrane cast
procedure used to prepare 3f
A 25 mL flask was charged with polymer 2 (700 mg, 1.29 mmol
based on the molecular weight of its repeat unit), DMAc (5 mL), and
a magnetic stir bar. The solution was stirred until the polymer was
completely dissolved. After adding
[1]
¼ 0.503 M; 2.44 mL, 1.23 mmol), the reaction vessel was
sealed with a septum. The vessel was then degassed under reduced
pressure, and CuI (38.3 mg, 200 mol) and 1,7-octadiyne (51.0 L,
.387 mmol) were added. The resulting mixture was stirred for
0 min at ambient temperature and then filtered through a cotton
a DMAc solution of 1
(
0
m
m
0
3
Upon the synthesis of 1, efforts shifted toward exploring the
utility of this compound to modify azide-functionalized poly
plug into a custom-built, air-free Petri dish as it was continuously
purged with nitrogen. The purging was stopped upon completion
ꢀ
(sulfone)s. As shown in Scheme 2, polysulfone 2 (M
PDI ¼ 2.01; prepared as described by Guiver [28]) was treated with
(2.0 equiv per repeat unit of 2) under Cu-catalyzed cycloaddition
w
¼ 45.6 kDa;
of the transfer. The chamber was then sealed and heated at 60 C in
an oven for 12 h. The chamber was then removed from the oven
and allowed to cool to ambient temperature. To allow the residual
solvent to evaporate, the top of the chamber was removed and the
1
conditions in DMAc. The progress of the reaction was monitored by
ꢀ
chamber was heated at 60 C in the oven for an additional 6 h
O
ꢀ
followed by heating at 80 C for 12 h. The membrane was released
O
NaH
S
+
O
3
SO Na
from the cell by adding 100 mL of H
removal of the membrane, it was acidified by heating at 80 C for
2
SO
4
(2 M; 100 mL). Upon
OH
O
ꢀ
DMF, 60 °C
2
h
1
2
h in the presence of H
2 4
SO (2 M, 400 mL) in a 1 L Erlenmeyer flask.
9
7% yield
Finally, the membrane was heated in 400 mL of de-ionized water at
ꢀ
ꢀ
8
0
C for 2 h and then dried in a vacuum oven at 70 C for 12 h
Scheme 1. Synthesis of alkyne sulfonate 1 via the ring-opening of 1,3-propanesultone
(
815 mg, 84% yield).
with propargyl alcohol under basic conditions.

B.C. Norris et al. / Polymer 51 (2010) 5352e5358
5355
O
S
O
S
O
O
1
O
O
O
O
N
3
N
3
n
5-10 mol% CuI
DMAC
75% yield
N
N
n
3
N
N
2
N
N
O
O
SO Na
3
SO Na
3
Scheme 2. Synthesis of poly(sulfone) 3 via Cu-catalyzed 1,3-dipolar cycloaddition of alkyne sulfonate 1 with an azide modified polysulfone (2).
the disappearance of the distinct IR absorption of the aryl azide
a similar approach. Poly(sulfone) 2 was dissolved in DMAc followed
by the addition of various quantities of alkyne sulfonate 1, 1,7-
octadiyne as a crosslinker, and copper catalyst (10 mol%); see Table 1.
Each of these solutions were independently degassed and then
ꢂ1
(
y
N3 ¼ 2119 cm
in a KBr matrix; see Fig. 1) as well as the
), diagnostic of
appearance of a singlet at
a triazole CeH proton, in the H NMR spectrum (see Fig. 2). Char-
d
¼ 8.1 ppm (DMSO-d
6
1
ꢀ
acterization of this crude reaction mixture (which contained the
heated at 60 C for 12 h in a Petri dish under nitrogen. The solvent
D
ꢀ
sodium sulfonate salt 3Na ) by GPC showed a small increase in
was then evaporated at 60 C to produce polymeric membranes that
molecular weight (M
compared to its starting material (2) (M
see ESI). The acidified poly(sulfone) 3 was isolated in 75% by
removal of the solvent under reduced pressure at 60 C followed by
washing with excess water and methanol and finally treatment
w
¼ 59.1 kDa) and polydisperisty (PDI ¼ 2.99)
2 4
were released from the dish by addition of 2 M H SO (aq.) The
w
¼ 56.9 kDa, PDI ¼ 2.14)
crosslinked polymeric materials obtained from these reactions were
insoluble; hence, they could not be analyzed via NMR spectroscopy.
However, as summarized in Table 1, a close correlation between the
theoretical and the measured IECs was observed. In addition, the
methanol permeabilities and proton conductivities of these
(
ꢀ
with 2 M H
molecular weight polymer (M
2
SO
4
(aq.) Analysis of 3 by GPC revealed a relatively low
¼ 27.5 kDa) of lower polydispersity
ꢀ
ꢀ
w
membranes were measured at 25 C and 65 C, with all membranes
showing significant improvements in conductivity at the later
temperature. Low levels of sulfonation resulted in membranes that
exhibited low methanol permeabilities and modest conductivities.
For example membrane 3f (which contained 30% crosslinker and an
IEC of 1.3) displayed a 10-fold reduction in methanol crossover
(
PDI ¼ 1.43). This change in molecular weight as well as the
decreased PDI was attributed to selective fractionation of the
polymer during the aforementioned purification and isolation
procedures. Regardless, the IEC (ion exchange capacity) of 3
(
2.1 meq/g) corresponded well with the theoretical value (2.2 meq/
ꢂ7
2
g). The thermal properties of 3 were also analyzed: the polymer
(1.03 ꢁ 10 cm /s) but exhibited a reduction in proton conductivity
from 120 to 40 mS/cm versus Nafion 117 (120 mS/cm). Utilization of
larger quantities of 1 as a means to increase the IEC of the material
even further resulted in membranes that were too mechanically
unstable for further testing, unfortunately. As summarizedinTable 1,
the relative selectivities of the membranes, which were determined
by comparing the conductivity/permeability ratio of the membrane
versus the same ratio measured for Nafion 117, were also calculated
and ranged from 0.02 to 5.21. In general, the membranes that
exhibited low methanol permeabilities exhibited the highest rela-
tive selectivities.
ꢀ
exhibited a glass transition temperature (T
g
) at 67 C which was not
present in any of the crosslinked systems (see below), as deter-
mined by differential scanning calorimetry, and was stable until
ꢀ
1
70 C, as determined by thermogravimetric analysis (TGA) (see
ESI).
3.2. Synthesis and study of crosslinked poly(sulfone) membranes
Once it was demonstrated that polymer 2 was successfully
modified with 1 using the aforementioned cycloaddition chemistry,
efforts shifted toward the synthesis of crosslinked polymers using
Fig. 1. FT-IR spectrum (KBr) of poly(sulfone)s 2 (bottom) and 3 (top). The asterisk
Fig. 2. 1H NMR spectrum (DMSO-d
) of poly(sulfone) 3. The diagnostic chemical shifts
6
denotes an azide stretching frequency.
are labeled.

5356
B.C. Norris et al. / Polymer 51 (2010) 5352e5358
Table 1
Selected methanol permeability and proton conductivity (
s
) data.
mol % 1,7-octadiyne
Polymer
IEC (theory),
meq/g a
IEC (exp)
MeOH perm
(cm /s)
relative
s
(mS/cm)
s (mS/cm)
(65 C)
b
2
c
selectivity (25 C) d
ꢀ
ꢀ
e
ꢀ
e
meq/g
(25 C)
ꢂ8
ꢂ8
ꢂ9
ꢂ7
ꢂ7
ꢂ7
ꢂ5
3
3
3
3
3
3
3
3
3
3
1
a
b
c
d
e
f
0.82
0.82
0.82
1.3
1.3
1.3
1.6
1.8
1.8
1.8
0.74
0.81
0.79
1.2
1.3
1.3
1.5
1.8
1.7
1.7
0
10
20
0
20
30
30
0
1.44 ꢁ 10
2.45 ꢁ 10
8.44 ꢁ 10
6.85 ꢁ 10
3.10 ꢁ 10
1.03 ꢁ 10
5.21
1.56
3.74
0.44
1.70
1.02
5.7
2.9
2.4
23
40
8.0
9.1
4.3
2.9
60
74
40
f
g
4.11 ꢁ 10
0.02
72
111
f
g
g
g
g
h
f
ꢂ6
ꢂ6
ꢂ6
i
j
20
25
0
3.90 ꢁ 10
4.83 ꢁ 10
1.25 ꢁ 10
0.17
0.13
e
50
49
95
53
50
120
f
h
17
0.91
0.91
a
b
c
The theoretical IEC was determined by the quotient of the total amount of sulfonate added in mmol over the total theoretical mass of the membrane in g.
The experimental IEC was determined by the quotient of the acid content of a representative sample, measured by titration, over the mass of the sample.
2
ꢀ
The methanol (MeOH) permeability data was acquired by measuring the diffusion of a 1 M solution of methanol through a membrane (area ¼ 7.1 cm ) at 25 C.
The relative selectivity was determined by the quotient of membrane selectivity over the selectivity of Nafion 117.
The impedance measurements were performed at 100% relative humidity.
d
e
f
These membranes formed mechanically weak gels upon acidification and were too fragile to incorporate into viable MEAs.
These membranes proved to be too weak for further testing.
Nafion 117 was tested for comparison.
g
h
3
.3. Synthesis and study of crosslinked poly(sulfone) copolymer
In general, we have observed that increasing either the degree
of sulfonation or the crosslink concentration resulted in more
brittle membranes. To balance these two factors, a new copol-
ymer membrane 5a, which has a more moderate IEC and only
10% crosslinker, was synthesized and tested. Qualitatively, this
membrane exhibited improved mechanical properties in the dry
state when compared to the other membranes reported herein.
The thermal stability of these membranes was probed by ther-
mogravimetric analysis, and found to exhibit a decomposition
membranes
To enhance the mechanical properties of the aforementioned
materials and to further separate the hydrophilic regions from the
hydrophobic regions as described in the Introduction, a series of
crosslinked copolymers comprised of polymer 2, a partially-
sulfonated polysulfone with varying degrees of sulfonation (4), and
a crosslinker (1,7-octadiyne) were synthesized as shown in Scheme
ꢀ
3
. As summarized in Table 3, this protocol afforded copolymers 5
temperature of 170 C, which is similar to that of the poly
that formed mechanically robust membranes, and exhibited lower
methanol permeabilities and similar conductivities as Nafion 117.
Unfortunately, further sulfonation of the hydrophilic block led to
films that were visually hazy or opaque. The phase separation in
this material is most likely due to the incompatibility of the
sulfonated polymer with the unsulfonated polymer [30]. In addi-
tion, these membranes showed poor mechanical and conductive
properties, and therefore were not investigated further [31].
(sulfone) 3 (not crosslinked) described above. The proton
conductivities of membranes 5 were measured at 25 and 65 C.
ꢀ
Consistent with literature reports, higher conductivities were
observed at the elevated temperature. As summarized in Table 3,
the relative selectivities of these membranes were also calcu-
lated and ranged from 0.55 to 1.85. In general, higher selectivities
were observed in membranes that exhibited relatively lower
proton conductivities.
SO
3
Na
3
SO Na
O
O
N
N
N
N
N
N
O
O
O
O
S
O
O
S
O
O
N
3
n
N
n
4
N
N
+
1
0 mol% CuI
DMAC
N
3
O
S
O
O
N
O
N
n
N
N
3
O
S
2
O
O
O
n
+
N
3
5
Scheme 3. Copolymer formation via crosslinking of two different polymer chains.

B.C. Norris et al. / Polymer 51 (2010) 5352e5358
5357
Table 2
a
Selected liquid uptake and swelling data.
%
Liquid Uptake
% Swelling
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
Polymer
2
H O @ 25
C
MeOH (1 M) @ 25
C
MeOH (1 M) @ 65
C
H
2
O @ 25
C
MeOH (1 M) @ 25
C
MeOH (1 M) @ 65
C
3
3
3
3
3
3
3
3
3
5
5
5
a
b
c
d
e
f
g
i
j
14.7
14.4
12.8
40.1
31.7
19.8
18.7
15.9
15.3
42.0
38.9
21.1
146
572
535
18
34.3
138.0
19.7
17.5
18.3
77.1
50.7
30.7
276
1589
581
24
5.4
3.9
2.2
11.7
9.2
4.5
28.6
124
62.5
5.7
6.8
3.9
3.3
13.3
11.7
7.1
38.1
129
62.5
5.7
8.1
5.9
5.5
20.0
15.8
10.7
124
47.6
b
486
528
16
31.2
129.3
75.0
5.7
14.0
43.5
a
b
c
38.3
174.4
7.0
28.3
9.3
30.4
a
The percent liquid uptake was calculated using the following equation: 100 ꢁ (wet mass ꢂ dry mass)/dry mass. The percent swelling was calculated using the following
equation: 100 ꢁ (wet length ꢂ dry length)/dry length.
b
This membrane formed a gel that proved to be too fragile to evaluate further.
3
.4. Liquid uptake and swelling
electrochemical performances of the MEAs were evaluated using
a computer controlled fuel cell testing setup (Scribner 840) at
65 C with 1 M methanol cycled through the anode at rate of
ꢀ
Membranes with high liquid uptake often exhibit high methanol
crossover. As summarized in Table 2, the membranes described
2.5 mL/min and oxygen fed to the cathode at rate of 200 mL/min.
As summarized in Fig. 3, the performance of the fuel cell con-
taining 5b was found to be highly dependent on MEA fabrication
ꢀ
herein were measured for their abilities to uptake water at 25 C as
ꢀ
well as methanol at 25 and 65 C. The uptake for these membranes
ꢀ
ranged from 14.4% to 1600% with larger uptakes observed at higher
methanol concentrations and elevated temperatures. Moreover,
liquid uptake appeared to be proportional with the acid content of
the membrane tested and inversely proportional with the
concentration of the crosslinker employed for membrane synthesis.
Membrane swelling is another critical parameter that should be
considered for proper function of proton exchange membranes in
fuel cells since swelling can have detrimental effects on fuel cell
performance. The swelling properties of the membranes described
herein were measured and the trends observed were similar to that
of the uptake (range: 2.2% to greater than 129%); see Table 2.
conditions. Assembly at 150 C (designated 5b 150C) afforded
a membrane that became brittle when drying during the hot
pressing procedure and was difficult to manipulate. In addition,
delamination of the membranes pressed under these conditions
was also observed, which may be due to the swelling properties of
the membrane. We believe that the hot pressed membrane
bonded to the electrode in the dry state and then wrinkled upon
wetting which resulted in higher interface resistance between
electrode and membrane [32], and ultimately resulted in cell
voltage loss. However, assembly of the MEA containing 5b at
ambient temperature (designated 5b 25C) afforded a membrane
that remained hydrated and flexible, which resulted in a better
contact between the electrode and membrane interface. More-
over, as shown in Fig. 3, improved performance compared to the
MEA fabricated at higher temperature was observed. The MEA
3.5. Membrane-Electrode Assembly (MEA): fabrication and testing
Upon synthesis, MEAs containing copolymers 5a and 5b as the
ꢀ
electrolyte were independently fabricated and tested. The MEAs
were fabricated by placing the membrane between two carbon
cloth electrodes coated in Pt (cathode) and Pt/Ru (anode) and
containing 5b (pressed at 25 C) also showed a relatively high
2
maximum power density (130 mW/cm ), comparable to that
measured for the MEA containing Nafion 117 (150 mW/cm ) under
2
2
a Nafion ionomer. Both catalyst loadings were 5 mg/cm and the
the same conditions.
2
active cell area was measured at 5 cm . The electrodes were then
Although the membrane comprised of 5b was measured to be
pressed onto the membrane using a Carver hot press at varying
temperatures (25e150 C) under 1500 psi of pressure for 2.5 min.
thinner than the Nafion 117 membrane (110 mm vs. 175 mm,
ꢀ
respectively), the former showed slightly lower performance due to
its relatively low proton conductivity as well as the incompatibility
between the aromatic polymer membrane and Nafion ionomer
coated on the electrodes. However, the open circuit voltage (OCV)
For comparison, a MEA containing Nafion 117 was also fabricated
ꢀ
as previously described, and assembled by hot pressing at 120 C
under otherwise identical conditions as described above [15]. The
Table 3
Selected methanol permeability and conductivity data.
g
(mS/cm) g
C
Polymer
Degree of
sulfonated/
unsulfonated wt/wt
IEC (theory),
meq/g c
IEC (exp)
meq/g d
mol % 1,7-octadiyne
MeOH
perm (cm /s)
Relative
Selectivity
s
(mS/cm)
C
s
sulfonation a
b
2
e
f
25
ꢀ
65
ꢀ
ꢂ8
ꢂ7
ꢂ6
5a
5b
5c
1.0
1.2
1.6
2.0
4.0
3.8
0.93
1.3
1.5
0.86
1.2
1.4
10
30
30
5.55 ꢁ 10
1.85
1.33
0.55
7.8
36
78
14.2
86
111
3.57 ꢁ 10
1.87 ꢁ 10
a
b
c
Equivalents of acid per repeat unit of the sulfonated polymer.
This ratio refers to the relative amounts of sulfonated and unsulfonated polymers added to the reaction mixture.
The theoretical IEC was determined by the quotient of the total amount of sulfonate added in mmol over the total theoretical mass of the membrane in g.
The experimental IEC was determined by the quotient of the acid content of a representative sample, measured by titration, over the mass of the sample.
d
e
f
ꢀ
The methanol permeability data was acquired at 25 C.
Relative selectivity was determined by the quotient of membrane selectivity over the selectivity of Nafion 117.
The impedance measurements were performed at 100% relative humidity.
g

5358
B.C. Norris et al. / Polymer 51 (2010) 5352e5358
0
.8
.7
.6
.5
.4
.3
.2
180
150
References
Nafion 117
5b 150C
b 25C
a 25C
0
[1] Dillon R, Srinivasan S, Aricò AS, Antonucci V. J Power Sources 2004;127:112.
5
5
[2] Neburchilov V, Martin J, Wang H, Zhang J. J Power Sources 2007;169:221.
[3] Li Q, He R, Jensen JO, Bjerrum NJ. Chem Mater 2003;15:4896.
0
1
9
6
3
0
20
0
[
4] Tricoli V, Carretta N, Bartolozzi M. J Electrochem Soc 2000;147:1286.
[5] Lafitte B, Karlsson LE, Jannasch P. Macromol Rapid Commun 2002;23:896.
6] Karlsson LE, Jannasch P. Electrochim Acta 2005;50:1939.
[7] Park HB, Shin HS, Lee YM, Rhim JW. J Membr Sci 2005;247:103.
0
[
0
0
[8] Kopitzke RW, Linkous CA, Anderson HR, Nelson GL.
J Electrochem Soc
0
0
2000;147:1677.
[9] Kreuer KD. J Membr Sci 2001;185:29.
[10] Li L, Zhang J, Wang Y. J Membr Sci 2003;226:159.
0
0
100 200 300 400 500 600
[11] Yang B, Manthiram A. Electrochem Solid-State Lett 2003;6:A229.
2
Current density (mA/cm )
[12] (a) Li W, Norris B, Snodgrass P, Prasad K, Stockett AS, Pryamitsyn V, et al.
J Phys Chem B 2009;113:10063;
(
2
b) Münch W, Kreuer KD, Silvestri W, Maier J, Seifert G. Solid State Ionics
001;145:437.
Fig. 3. MEA tests of 5a, 5b and Nafion 117. Filled shapes represent the polarization
curve; open shapes represent the power density curve. A Pt-black catalyst was used at
the cathode and a Pt:Ru black catalyst was used at the anode (the loading of each was
[
[
13] Li W, Manthiram A, Guiver MD, Liu B. Electrochem Commun 2010;12:607.
14] Yang C, Costamagna P, Srinivasan S, Benziger J, Bocarsly AB. J Power Sources
2
2
5
5
mg/cm ). The active cell area was 5 cm . Sample 5b 150C refers to a MEA containing
b that was assembled via hot pressing at 150 C. Sample 5b 25C refers to a MEA
2
001;103:1.
15] (a) Fu Y, Manthiram A, Guiver M. Electrochem Commun 2006;8:1386;
b) Kreuer KD, Paddison SJ, Spohr E, Schuster M. Chem Rev 2004;104:4637;
ꢀ
[
containing 5b that was assembled via pressing at ambient temperature. 5a 25C was
(
ꢀ
also assembled at 25 C.
(c) Herz HG, Kreuer KD, Maier J, Scharfenberger G, Schuster MFH, Meyer WH.
Electrochim Acta 2003;48:2165.
[16] (a) Saxena A, Tripathi BP, Shahi VK. J Phys Chem B 2007;111:12454;
(
(
(
b) Jeong MH, Lee KS, Lee JS. Macromolecules 2009;42:1652;
c) Kerres JA. Fuel Cells 2005;5:230;
d) Kerres JA. J Membr Sci 2001;185:3.
of 5b was measured to be higher than that of Nafion 117, which was
attributed to the lower methanol permeability of the former and
the consequent smaller voltage loss at the cathode side.
[17] (a) Lee M, Park JK, Lee H-S, Lane O, Moore RB, McGrath JE, et al. Polymer
009;50:6129;
2
(
(
For comparison membrane 5a was also tested in a fuel cell. This
membrane had a smaller IEC and crosslink density which made it
much less brittle in the dry state but also reduced its conductivity
relative to 5b. The fuel cell showed a slightly higher OCV but a steeper
polarization curve which is typically associated with less conductive
electrolytes. However, the lower methanol permeability of these
membranes may not only help to lower the Pt catalyst loading at the
cathode but also lead to an improved long-term stability and
performance in the respective DMFC; such studies are underway.
b) Lee H-S, Roy A, Lane O, Dunn S, McGrath JE. Polymer 2008;49:715;
c) Ghassemi H, Ndip G, McGrath JE. Polymer 2004;45:5855;
(d) Kim DS, Robertson GP, Kim YS, Guiver MD. Macromolecules 2009;42:957;
e) Einsla ML, Kim YS, Hawley M, Lee H-S, McGrath JE, Liu B, et al. Chem Mater
008;20:5636;
f) Badami AS, Lane O, Lee H-S, Roy A, McGrath JE. J Membr Sci 2009;333:1;
(g) Lee H-S, Roy A, Lane O, McGrath JE. Polymer 2008;49:5387;
h) Wang H, Badami A, Roy A, McGrath JE. J Poly Sci Part A Polym Chem
006;45:284;
i) Lee H-S, Badami A, Roy A, McGrath JE. J Polym Sci Part B Polym Phys
2007;45:4879;
(
2
(
(
2
(
(
(
j) Roy A, Lee H-S, McGrath JE. Polymer 2008;49:5037;
k) Lafitte B, Jannasch P. Adv Funct Mater 2007;17:2823.
4
. Conclusions
[
[
18] Xu N, Guo X, Fang J, Xu H, Yin J. J Polym Sci Part A Polym Chem 2009;47:6992.
19] Zhang G, Guo X, Fang J, Chen K, Okamoto KI. J Membr Sci 2009;326:708.
A modular method for modifying poly(sulfone)s containing
[20] Meldal M, Tornøe CW. Chem Rev 2008;108:2952.
[
[
21] Kolb HC, Finn MG, Sharpless KB. Angew Chem Int Ed 2001;40:2004.
22] Abboud JLM, Roces-Foces C, Notario R, Trifonov RE, Volovodenko AP,
Ostrovskii VA, et al. Eur J Org Chem 2001;2001:3013.
pendant azides is reported. Using copper-catalyzed, 1,3-dipolar
cycloaddition chemistry, Brønsted basic 1,2,3-triazoles and cross-
links were successfully formed in a single step which was found to
significantly decrease the methanol permeability while maintain-
ing relatively high proton conductivity in the resulting membranes.
In addition, the fuel cell performances of MEAs containing these
materials were comparable to those containing Nafion-based
membranes, and exhibited a maximum power output of 130 mW/
[23] Zhou Z, Li S, Zhang Y, Liu M, Li W. J Am Chem Soc 2005;127:10824.
[
[
24] Chen T-Y, Leddy J. Langmuir 2000;16:2866.
25] (a) Kim YS, Sumner MJ, Harrison WL, Riffle JS, McGrath JE, Pivovar BS. J
Electrochem Soc 2004;151:A2150;
(b) Pivovar BS, Wang Y, Cussler EL. J Membr Sci 1999;154:155.
26] Okamoto K, Yin Y, Yamada O, Islam MN, Honda T, Mishima T, et al. J Membr
Sci 2005;258:115.
[
[27] Zhai F, Guo X, Fang J, Xu H. J Membr Sci 2007;296:102.
2
cm . On a broader level, the strategy described herein effectively
[28] Guiver MD, Robertson GP. Macromolecules 1995;28:294.
[29] (a) Karlsson LE, Jannash P. J Membr Sci 2004;230:61;
establishes a new and versatile route to functionalization of
aromatic polymers and extension of this methodology to other
proton conductive materials is currently underway.
(
(
b) Kim DS, Robertson GP, Guiver MD. Macromolecules 2008;41:2126;
c) Norsten TB, Guiver MD, Murphy J, Astill T, Navession T, Holdcroft S, et al.
Adv Funct Mater 2006;16:1814;
(
(
d) Li Z, Ding J, Robertson GP, Guiver MD. Macromolecules 2006;39:6990;
e) Liu B, Robertson GP, Kim DS, Guiver MD, Hu W, Jiang Z. Macromolecules
Acknowledgement
2
007;40:1934.
ꢀ
[
[
30] Karam HJ. In: Solc K, editor. Polymer Compatability and Incompatability
Principles and Practice. MMI Press; 1982. p.93.
This work was supported by the Office of the Naval Research
Multidisciplinary University Research Initiative (ONR-MURI; Grant
No. N00014-07-1-0758).
31] For examples highly sulfonated membranes with good mechanical properties,
see:
(
a) Fujimoto CH, Hicker MA, Cornelius CJ, Loy DA. Macromolecules
005;38:5010;
b) Kerres J, Zhang W, Cui W. J. Polym. Sci.: Part A: Polym. Chem.
998;36:1441.
32] (a) Pivovar BS, Kim YS. J Electrochem Soc 2007;154:B739;
(b) Kim YS, Kim DS, Liu B, Guiver MD, Pivovar BS.
008;155:B21.
2
(
1
Appendix. Supplementary data
[
Supplementary data associated with this article can be found in
the online version at doi:10.1016/j.polymer.2010.09.041.
J Electrochem Soc
2