I. Tonozuka et al. / Polymer 52 (2011) 6020e6028
6023
with a JMS-SX102A (JEOL) in electron impact mode. Elemental
analysis was obtained with a PE2400-II (PerkinElmer Inc.) at 975 ꢂC.
The molecular weights of the polymers were determined by gel
permeation chromatography (GPC). The chromatography system
was equipped with two Shodex columns (KF-804L and KF-805 for
THF, KD-805 and KD-805 for DMF), and measurements were made
at 40 ꢂC with THF eluent or 50 ꢂC with DMF eluent at a flow rate of
1.0 mL minꢁ1. Molecular weights were estimated with a calibration
curve constructed using polystyrene (THF) or poly(ethylene oxide)
standards (DMF). GPC measurements were performed more than
three times for more than three samples synthesized under each
condition. Thermogravimetric analyses (TGA) were performed wiꢁth1
a TG-DTA 200 (Seiko Instruments) at a heating rate of 10 ꢂC min
with nitrogen flow. Glass transition temperatures and melting
points were measured by ꢁa1 DSC6200 (Seiko Instruments) at
a heating rate of 10 ꢂC min with nitrogen flow. Dried S-PPBP
membranes were placed for 6 days at room temperature in glass
desiccators humidified at 30e100% RH using saturated solutions to
prepare hydrated samples. Water uptakes (W) were estimated from
the weights of dried (Wdry) and hydrated (Wwet) membranes as
chlorinated monomer, unreacted starting materials, and protic
compounds is essential for the preparation of high molecular
weight polymers via the following Ni(0)-catalyzed coupling poly-
merization [4]. The purification of DPBP was carried out using two
or more recrystallizations from an ethyl acetate and hexane mixed
solvent (1/5, v/v). The identity of DPBP was confirmed by EI-MS, FT-
IR, 1H and 13C NMR, and elemental analyses. The purity of DPBP was
estimated to be 99.94% by liquid chromatography. Polymerization
of DPBP was first carried out according to our previous Ni(0)-
catalyzed coupling polymerization procedure [7,27]. PPBPs with
1e2 ꢀ105 g molꢁ1 of weight-average molecular weights (Mw) were
isolated in more than 80% yield. The identification of PPBP was
carried out by FT-IR, GPC, 1H and 13C NMR, and elemental analyses.
The 13C NMR spectrum of PPBP exhibited two chemical shifts in the
carbonyl region; these shifts resulted from different regioisomers in
the polymer backbone. This result was also observed for all of the
PPBP preparations irrespective of the reaction conditions and
resultant molecular weights, suggesting that the PPBPs obtained in
this study have a less regioregular structure compared with the
structure of poly(2,5-benzophenone) reported elsewhere [4,5].
Sheares et al. observed a reduction of nearly 15% of the carbonyl
groups in poly(2,5-benzophenone) prepared by Ni(0)-catalyzed
coupling polymerization [6,28,29]. We did not observe any
similar phenomenon for our PPBP preparations. This difference
might be dependent on the substituents on the carbonyl group and
on the polymerization conditions.
W ¼ (Wwet ꢁ Wdry)/Wdry ꢀ 100%, and hydration numbers (
l) were
calculated as
l
¼ (number of sorbed water molecules)/(number of
sulfo groups). The ion exchange capacity (IEC) was determined by
back titration and also by elemental analysis. The swelling ratios in
both the in-plane (PS) and through-plane (TS) directions were
expressed as follows:
All of the PPBPs were amorphous and all were soluble in chlo-
roform, dichloromethane, THF, NMP, DMF, and DMAc. All of the
PPBP membranes prepared from 3 to 5 wt% chloroform solutions
via the casting method exhibited an amorphous feature and dis-
played no melting point until the decomposition temperature at
500 ꢂC. The glass transition temperatures of PPBPs measured by
DSC were 150e160 ꢂC.
PS ¼ (Lwet ꢁ Ldry)/Ldry ꢀ 100%
TS ¼ (Twet ꢁ Tdry)/Tdry ꢀ 100%
where Ldry and Lwet are the diagonal lengths in-plane of the
dried and hydrated membranes, and Tdry and Twet are the thick-
nesses of the dried and hydrated membranes, respectively.
Mechanical strengths were measured by tensile tests with an
3.2. Optimization of polymerization conditions
autograph AGAS-1kNA (Shimadzu Co.) at
a test speed of
2 mm minꢁ1. Optical micrographs were obtained by using an
optical polarized microscope (Opticalhot-Pol, Nikon) with a trans-
mitted white light source.
We explored the use of several catalysts for achieving high
molecular weight products. These results are summarized in
Table 1. Since Co catalysts tend to form five-coordinate complexes,
CoCl2(PPh3)2 was inactive for this polymerization. The features of
Pd catalysts are usually similar to Ni catalysts, except for the smaller
ion diameter, but PdCl2(PPh3)2 did not catalyze the polymerization
with the aryl chloride monomer. In contrast, the Ni catalysts
(NiCl2(PPh3)2 and Ni(PPh3)4) provided polymeric materials. NiCl2
with PPh3 was inactive, whereas NiCl2 with PPh3 and Bipy
(4.76 ꢀ 10ꢁ4 mol, equimolar Bipy relative to NiCl2) was slightly
effective (Entry 1-3). As reported by Colon and Kelsey and by other
groups [1,30], the transfer of aryl groups from triphenylphosphine
to nickel occurs. This side reaction is also known to be influenced by
the temperature and the ligand concentration. Our results, which
indicated that addition of Bipy suppressed the side reaction and
slightly improved the polymerization, were consistent with these
previous reports. Although fine tuning of the PPh3 and Bipy ratio
relative to nickel chloride and of other conditions might give higher
yields and molecular weights, NiCl2(PPh3)2 provided better results
without additional tuning.
We also explored the effect of other aspects of the polymeri-
zation reaction. Our studies on different ligands can be seen in
Entries 1-4e1-8 of Table 1. Neither of the bidentate ligands (dppf
and dppp) nor the larger ligands (P(PhMe3)3, P(Ph4Cl)3, and
P(Ph3Cl)3) could effectively catalyze the polymerization as well as
PPh3. The reaction temperature, time, and amount of solvent also
affected the molecular weights and PDIs of the obtained PPBPs.
As shown in Table 2, the molecular weight increased as the
reaction temperature was raised from 50 and 65 ꢂC, but both yield
Proton conductivity values of 1 ꢀ6 cm S-PPBP membranes were
measured with an electrochemical impedance analyzer SI-1260
(Solartron). Samples were clamped between Pt electrodes and
placed on a home-made Teflon four-probe cell for the in-plane
measurements and on a home-made Teflon pseudo-four-probe
cell for through-plane measurements. Cells were set in a thermo-
hygrostat that was electrically shielded and grounded to ensure
accurate measurements. Membrane electrode assemblies (MEA) of
S-PPBP membranes were prepared by the following protocol: Pt on
Vulcan XC-72 (46.9%), purified water, and Nafion solution (5 wt%)
were mixed, and the mixed catalyst inks were coated on gas
diffusion electrodes (GDL35BC, SGL Japan). The electrodes were
then pressed onto S-PPBP membranes at 50 kg cmꢁ2 and 130 ꢂC for
10 min. The active area was 5 cm2, and the Pt catalyst loading was
1 mg cmꢁ2. The MEA was assembled in a single cell between bipolar
plates made of graphite with serpentine flow channels. A single fuel
cell test fed by H2/air was conducted at a cell temperature of 80 ꢂC
under humidified condition by a FC test system (Toyo Co.).
3. Results and discussion
3.1. Polymerization of DPBP
We successfully synthesized DPBP by Friedel-Crafts acylation
with 2,5-dichlorobenzoic chloride, diphenyl ether, and AlCl3 in
59e72% yield. The removal of all impurities such as the mono-