Fabrication and proton conductivity of sulfonated silica composites prepared by postoxidization of mercaptomethoxysilane

Article abstract of DOI:10.1002/pola.26109

Sol-gel derived organic-inorganic hybrids containing phosphotungstic acid (PWA) have been prepared previously to obtain proton conductive membranes. However, leaking of PWA was a serious problem to achieve the higher proton conductivity. In this study, polyelectrolyte membranes functionalized with sulfonic acid groups were fabricated by the sol-gel method. Proton conductivity measurements were performed on an impedance analyzer at 80°C/95% RH. The functionalized polyelectrolyte membranes exhibited the proton conductivity σ ～ 0.9 (S/cm) which was much higher than the previously reported hybrids containing PWA. Although the hybrids exhibited fairly high proton conductivity irrespective to the catalysts used, that under the low relative humidity strongly depends on the catalysts. Among the hybrids prepared in this study, the membrane synthesized with HCl showed outstanding proton conductive properties even at the low humidity thanks to the proton transport channel formed by the swelling of ionic clusters. This fact was confirmed by measuring the ion exchange capacity, water uptake, swelling rate, Fourier transform infrared spectroscopy, atomic force microscopy, and thermogravimetric analysis.

Full text of DOI:10.1002/pola.26109

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
Fabrication and Proton Conductivity of Sulfonated Silica Composites
Prepared by Postoxidization of Mercaptomethoxysilane
1
1
1
1
Kazuhiro Yasumoto, Tomohisa Norisuye, Yusuke Teranishi, Qui Tran-Cong-Miyata,
2
Shigeki Nomura
1
Department of Macromolecular Science and Engineering, Graduate School of Science & Technology,
Kyoto Institute of Technology, Matsugasaki, Sakyo-ku, Kyoto 606-8585, Japan
2
R&D Center, Sekisui Chemical Co., Ltd., 32 Wadai, Tsukuba-shi, Ibaraki 300-4292, Japan
Correspondence to: T. Norisuye (E-mail: nori@kit.jp)
Received 22 September 2011; accepted 12 April 2012; published online
DOI: 10.1002/pola.26109
ABSTRACT: Sol–gel derived organic–inorganic hybrids contain-
ing phosphotungstic acid (PWA) have been prepared previ-
ously to obtain proton conductive membranes. However,
leaking of PWA was a serious problem to achieve the higher
proton conductivity. In this study, polyelectrolyte membranes
functionalized with sulfonic acid groups were fabricated by the
sol–gel method. Proton conductivity measurements were per-
strongly depends on the catalysts. Among the hybrids prepared
in this study, the membrane synthesized with HCl showed out-
standing proton conductive properties even at the low humid-
ity thanks to the proton transport channel formed by the
swelling of ionic clusters. This fact was confirmed by meas-
uring the ion exchange capacity, water uptake, swelling
rate, Fourier transform infrared spectroscopy, atomic force mi-
croscopy, and thermogravimetric analysis. VC 2012 Wiley Period-
ꢀ
formed on an impedance analyzer at 80 C/95% RH. The func-
tionalized polyelectrolyte membranes exhibited the proton
conductivity r ꢁ 0.9 (S/cm) which was much higher than the
previously reported hybrids containing PWA. Although the
hybrids exhibited fairly high proton conductivity irrespective to
the catalysts used, that under the low relative humidity
icals, Inc. J Polym Sci Part A: Polym Chem 000: 000–000, 2012
KEYWORDS: atomic force microscopy; heteropolyacid; proton
conductivity; sol–gel; inorganic polymers; membranes; nano-
composites; networks; structure-property relations; silicas
INTRODUCTION A nano-composite consisting of a flexible or-
ganic unit and a rigid inorganic junction provides effective
means to preserve functional molecules in network, provid-
ing functionalities such as superior thermal and mechanical
stabilities compared with conventional polymer films. As an
example of such composites, 1, 8-bis(triethoxysilyl)octane
for the proton transfer through the continuous phase is
strongly demanded. The most popular methods would be (1)
7
,12–18
manipulating the continuous conductive phase
forming the condensed ion-clusters by cluster growth.
or (2)
1
9,20
1
The former corresponds to ion-rich continuous structures
constructed by a delicate balance between the hydrophobic
and hydrophilic interactions as often seen in perfluorinated
(
TesOct) containing a moderate length of methylene bridge
2
1–25
between two alkoxide groups, enables rapid gelation without
membranes.
Conversely, the latter can be obtained
2
intramolecular cyclization and thus can be suitably used as
by interconnection of the well-developed clusters without
macroscopic phase separation.
a host to fabricate organic–inorganic hybrid materials. In the
3
,4
previous studies, phosphotungstic acid (PWA), a superacid
as a catalyst, was effectively doped in the TesOct matrix to
obtain proton conductive membranes which exhibited the
So far, we have studied the microscopic structure and the
proton conductivity of sol–gel derived organic–inorganic
membranes by means of thermal and scattering investiga-
ꢂ2
practically acceptable conductivity r ꢁ 10 (S/cm) although
3,4
tions. Besides the structure analysis at equilibrium state,
the material was in the amorphous state. The materials has
a potential of being used as a crosslinking host for fuel cell
the reaction process was also characterized by systematic
sampling of the growing clusters in a pregel solution during
5
–11
applications,
preventing the chemical short due to a leak
the gelation by dynamic light scattering and atomic force mi-
of fuels throughout the membrane.
19,20
croscopy (AFM).
Although emergence of a specific struc-
Although the better proton conductivity or thermostability
was one of the most important strategies for the fuel cell
applications, designing structures at molecular level efficient
ture is indispensable for the proton pathway, isolation of the
conducting domains takes place if crosslinking do not have
sufficient role to suppress the concentration fluctuations. For
Additional Supporting Information may be found in the online version of this article.
2012 Wiley Periodicals, Inc.
V
C
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2012, 000, 000–000
1

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
example, macroscopic phase separation was observed if
methanol was used as a solvent for the synthesis of the
sary or only used as a superacid catalyst. Hereafter, the HCl
and PWA catalyzed hybrids are, respectively, abbreviated to
HCl-Sul-TesOct and PWA-Sul-TesOct.
3
PWA-TesOct composites in spite of the fact that it was a com-
monly used solvent in the field of the sol–gel synthesis. It is
well known that hydrogen bonding and the inductive effect
influence the branching structure via the change in the rela-
PWA (Aldrich), IPA (Wako Chemical, Japan), MPTS (Shin-etsu
Chemical), 1 N hydrochloric acid (HCl, Wako Chemical, Ja-
pan), and hydrogen peroxide (H O , Wako Chemical, Japan)
2
6
2 2
tive reactivity at each reaction front. In the case of tetrae-
thoxysilane, four functional groups can take part in the reac-
tion, where the states of the functional groups, for example,
hydrolyzed silanol groups, residual ethoxy groups or gener-
ated siloxane bonds can affect the subsequent reactivity and
the overall network structure.
were used as received. The reaction was initiated by mixing
a TesOct/MPTS/IPA solution into either PWA/IPA or HCl/IPA
ꢀ
solution at 20 C. After vigorous stirring, the solution was fil-
tered through a Teflon membrane with the pore size 0.25
lm and was poured into a plastic petri dish, followed by an
aging period of 24 h. After solidification, the resulting mem-
branes were further aged under saturated vapor pressure at
As for the PWA-TesOct composites studied so far, the PWA con-
4
27
ꢀ
centration and the preparation temperature dependence of
60 C for another 24 h. The dried membranes were then
3
the conductivity, the effect of solvent on the conductivity have
carefully washed with distilled water for three days. Subse-
quently, mercapto groups of MPTS were oxidized into sul-
fonic acid groups by immersing the membrane into a satu-
been investigated. Although the proton conductivity reached r
ꢂ2
ꢁ
10 (S/cm), which was practically acceptable value for the
2
8
fuel-cell applications, leaking of PWA from the hybrid mem-
branes limited further improvement of the proton conductivity,
motivating us to permanently introduce acid-groups, for exam-
ple, sulfonic-acid groups into polymer networks by a silane
coupling reaction. However, such acids could take part in the
acid-catalyzed reaction itself, resulting in the difficulty to
design the suitable structure of proton pathway. Therefore,
rated hydrogen peroxide solution. The concentration of
PWA and MPTS were respectively adjusted in the range
½
PWAꢃ
½MPTSꢃ
CPWA ¼
¼ 0.025–0.250 and CMPTS ¼
¼ 0–3.73
½
TesOctꢃ
½TesOctꢃ
where the bracket indicates the molar concentration of each
½H2Oꢃ
component. The water concentration defined as r ¼
6
½TesOctꢃ
is variable in range 0.25–4. The prescribed volume of 1 N
HCl solution was mixed into the reactor batch to adjust the
water concentration defined by r value. The thickness of the
sample was in the range 10–160 lm, typically 100 lm.
(3-mercaptopropyl)trimethoxysilane (MPTS) was incorporated
into the sol–gel network, followed by postoxidation of the mer-
28
capto groups into sulfonic acid after the structure was fixed.
Thus, the aim of this article is to fabricate sulfonated hybrid
membranes exhibiting the better proton conductivity by the
two-step synthesis and also to elucidate the role of catalyst on
the proton conductivity.
Conductivity Measurement
Proton conductivity measurements were performed on a
Hioki 3532-80 impedance analyzer coupling with two plati-
num electrodes placed in a thermostat chamber (Espec, LHL-
ꢀ
1
13) at 80 C/95%RH. The conductivity was determined
EXPERIMENTAL
from the Nyquist plot by an AC impedance method in the
range 10 Hz–1 MHz. Warburg impedance was extrapolated
to the real axis to obtain the resistance of membranes.
Samples
3
2,33
The sol–gel precursor, 1,8-bis(triethoxysilyl)octane (TesOct),
2
9
The measurements were iteratively performed until the im-
pedance reached the plateau value. The humidity and tem-
(
OC H ) ASiA(CH ) ASiA(OC H ) , was synthesized
by
2
5 3
2 8
2 5 3
hydrosilylation of octadiene with triethoxysilane in toluene
in the presence of the Karstedt catalyst
diene (Alfa Aesar), triethoxysilane (Tokyo Chemical Industry)
and toluene (Wako Chemical, Japan), and Karstedt catalyst
ꢀ
3
0,31
ꢀ
at 20 C. Octa-
perature dependences were also examined in range 40–80 C
and 45–95%RH, respectively. Because of a stability issue of
V
R
the membranes, the conductivity for Nafion112 was meas-
ꢀ
ured at 50 C when the humidity dependence was examined.
(
Umicore Japan) were used as-received. TesOct was purified
at least twice by reduced pressure distillation.
Ion Exchange Capacity Measurements
The mixture of 6-functional TesOct and tri-functional (3-mer-
captoproyl)trimethoxysilane (MPTS) in isopropanol (IPA) and
water lead to a branched hybrid network via the hydrolysis
and condensation reactions. The resulting structure strongly
depends on the type of catalysts. For example, when an acid
catalyst is used, a weakly branched network is obtained
because of the rapid hydrolysis and slow condensation. Con-
versely, highly branched clusters are obtained if a basic cata-
lyst promotes the nucleophilic reaction. In this study, two
types of acids are used, that is, one is regular acid, hydro-
chloric acid (HCl) and the other is superacid, PWA. Note that
PWA has a role not only as a catalyst but also as a proton
conductor as described in our previous works. In this study,
sulfonic-acid groups are introduced in the network by oxidi-
zation of mercapto groups of MPTS, thereby PWA is unneces-
The ion exchange capacity (IEC) was determined to find the
amount of acid groups carried in the hybrid membranes.
First, a membrane was soaked in a saturated sodium chlo-
ride solution to replace proton by sodium ion. Second, the
acid solution was titrated by a sodium hydroxide solution to
determine the amount of acid released from the membrane.
Subsequently, the dried membrane was weighed to calculate
the equivalent weight, EW, defined by
wðgÞ
EW ðg molꢂ1Þ ¼ n
(1)
NaOH
ðmolÞ
where w and nNaOH are weight of the dried membrane and
mole of NaOH required to neutralize proton, respectively.
Finally, IEC was calculated by
2
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2012, 000, 000–000

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
IECðmeq gꢂ1Þ ¼ 1
000
(2)
EW
Water Uptake
The water uptake was calculated as the ratio of the differ-
ence between the wet and dry masses to the dry mass of the
hybrid film. To obtain the mass of the wet hybrid, the film
was quickly weighted on a microbalance after wiping water.
Normalized Swelling Ratio
The swelling ratio was defined as the ratio of the diameters
of the hybrid membrane before and after the oxidization.
The obtained swelling ratio was normalized by the diameter
of the membrane before oxidization. Note that the variation
in the thickness was almost negligible within the error of ex-
perimental ranges, and noticeable expansion was only seen
for the diameter.
Atomic Force Microscopy
AFM was operated in tapping mode using a Digital Instru-
ments Multimode AFM, controlled by a Nanoscope IIIa scan-
ning probe microscope controller with an extender module.
Commercially available silicon tip with a spring constant of
30–40 N/m, single beam cantilevers of 125 lm long and a
typical radius of curvature in the range 5–10 nm was used
at the resonance frequencies ranging from 310 to 340 kHz
depending on cantilever. Both topography and phase images
with various scan sizes 500, 1000, 2000, 4000 nm were col-
lected using tapping mode with a scan rate of 1 Hz in ambi-
ent atmosphere at room temperature. These images were
transformed into a single master curve of the power-spectra
after a second-order background subtraction, fast Fourier
FIGURE 1 The proton conductivity, r, for a series of the Sul-
TesOct membranes (open markers) and PWA-TesOct mem-
branes (solid markers), respectively, as functions of CMPTS and
CPWA observed at 80ꢀC/95%RH.
hydrogen peroxide solution. The results for the conventional
PWA-TesOct membranes without MPTS were also plotted in
closed markers for comparison. The solid and dotted lines
are guided for the eyes. In the case of the conventional PWA-
TesOct membranes, it has already known that r exhibits a
maximum around CPWA ¼ 0.1 due to competition between
the moderate domain growth and aggregation of PWA. As
seen from the figure, the (HCl- or PWA-)Sul-TesOct mem-
branes systematically exhibited the proton conductivity
higher than these for the PWA-TesOct membranes regardless
of the range of the MPTS concentration. The significant
improvement of the proton conductivity was achieved over
the wide range of CMPTS and the maximum conductivity r ¼
4
transform, and radial averaging by homemade software.
Thermogravimetric Analysis
Thermogravimetric analysis (TGA) was conducted using a
TGA2950 (TA Instruments thermogravimetric analyzer) with
high-resolution mode. The weight of samples was in the
range 5–9 mg. Nitrogen was used as a purge gas. The heat-
ing rate was dynamically varied in response to the changes
in the decomposition rate of the sample so as to improve the
resolutions of the weight change. Maximum ramp heating
ꢀ
rate was set at 20 C/min.
0
.91(S/cm) was obtained for C
¼ 2.24. Surprisingly, the
Fourier Transform Infrared Spectroscopy
MPTS
behavior seemed to be invariant to the catalysis used (HCl
or PWA) for the network formation.
Fourier transform infrared spectroscopy (FTIR) measure-
ments were performed on a Perkin-Elmer Spectrum GX sys-
tem equipped with an attenuated total reflection (ATR)
attachment. The ATR spectra were collected in the wave
Figure 2(a) shows the IEC as a function of CMPTS. IEC
increased with CMPTS, suggesting the effective incorporation
of sulfonic acid groups into the network. If we simply
assume that fully oxidized MPTS is incorporated in the net-
work without loss and monomers are fully hydrolyzed, the
dashed line is expected from a simple stoichiometric calcula-
tion. Obviously, the line systematically overestimates the IEC
values, probably due to the simplified assumption. As all the
monomers can be treated as fully hydrolyzed condition, par-
ticularly for the acid catalyzed system, the discrepancy might
be attributed to incomplete oxidization of the SH groups or
unsuccessful incorporation of MPTS monomer. As described
later, the actual composition can be estimated by peak
ꢂ1
number ranging from 650 to 4000 cm with a resolution of
ꢂ1
4
cm . The background spectra including the absorption of
moisture and carbon dioxide were obtained by measuring
the absorption of an empty ATR crystal used as reference.
RESULTS AND DISCUSSION
Figure 1 shows the proton conductivity, r, for a series of the
Sul-TesOct membranes (open markers) as a function of the
ꢀ
MPTS concentration, CMPTS observed at 80 C/95%RH. The
maximum concentration of the MPTS studied here was CMPTS
¼
3.0 because the membrane became dissociated in the
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2012, 000, 000–000
3

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
Figure 3 shows the relative humidity dependence on r with
different CMPTS obtained, respectively, (a) for the HCl-Sul-
TesOct and (b) the PWA-Sul-TesOct membranes. The results
V
R
for Nafion112 were also plotted in the figures. The proton
conductivity decreased with decreasing the relative humidity
for all the data. However, the reduction of r was much
smaller for the HCl-Sul-TesOct membranes compared to those
for PWA-Sul-TesOct. In the case of the HCl-Sul-TesOct mem-
branes, particularly for CMPTS ¼ 2.98, r was much larger
than Nafion112 throughout the humidity range studied here.
Conversely, for the PWA-Sul-TesOct membrane, although r
(
CMPTS ¼ 2.24) at 95%RH exhibited a pronounced value of r
¼
0.3 (S/cm) as indicated in Figure 3(b), the low humidity
performance was fairly poor compared with those for the
HCl-Sul-TesOct or Nafion112. In other words, comparing the
HCl to the PWA catalyzed membranes, r at the water-satu-
rated condition are almost equivalent, whereas those at the
lower humidity condition are completely different.
It is well known that there are two types of mechanism for
the proton transfer, that is, the vehicle mechanism and the
3
5
Grotthuss mechanisms. The former is based on the direct
transportation of proton as oxonium ions, whereas the latter
originates from the exchange of proton between oxonium ion
and water. The Grotthuss mechanism requires less activation
energy thereby suitable for the better proton conduction.
Note that apparent activation energy could be larger at high
temperature because of the swelling (less local density of
ion). Thus, the obtained slopes may not represent the correct
activation energy. Although the results are not shown here,
the activation energy estimated by the Arrhenius plot (irre-
ꢂ1
spective to the fitting range) was around 20–50 kJ mol
which is close to the value reported in the literature.
,
3
6
Therefore, Grotthuss mechanism could mainly contribute to
the proton transportation. To understand the roles of low-
humidity in the proton conductivity, the swelling properties
of the membranes were examined. Figure 4(a) shows the
water uptake of the membranes prepared with HCl (solid
symbols) and PWA (open symbols) catalysts. As seen from
the figure, the PWA-catalyzed membranes contained more
water molecules than the HCl catalyzed membranes. There-
fore, it is expected that the former membranes swell more
than the latter one. However, as evidenced from the swelling
ratio measurement shown in Figure 4(b), the normalized
swelling ratio, which is defined as the ratio of the diameters
of the membrane before and after the oxidization in the
equilibrium swelling state, is systematically higher for the
HCl-Sul-TesOct membranes compared to the PWA-Sul-TesOct
membranes. From this experimental evidence, it is expected
that the PWA-Sul-TesOct membranes have porous and rigid
structures compared to the HCl-Sul-TesOct membranes as
schematically illustrated in Figure 4(c). For the HCl-Sul-
TesOct membranes, the gel-like structure may be formed as
expected from pronounced swelling and moderate water
uptake. This is easily understood from the fact the acid-cata-
FIGURE 2 (a) The IEC as a function of CMPTS. The dashed and
solid lines are the theoretical prediction (See context). (b) The
ꢀ
double logarithmic plot of r versus IEC observed at 80 C/
95%RH. The dotted line is the best fit with a power-law
function.
separation of the TGA data. By taking into account the
amount of MPTS successfully incorporated into the network,
the predicted curve nicely agreed with the experimental data
as indicated by the solid line. It is also noted that IEC for the
PWA-Sul-TesOct membranes at CMPTS ¼ 0 has a finite value
because of the conductivity of PWA itself. The IECs were
almost independent on the catalysts similar to the result of
the conductivity obtained above. Then, the double logarith-
mic plot of r versus IEC is shown in Figure 2(b). As indi-
cated by the dotted line, the data well falls on a single mas-
ter curve with a power-law with an exponent of 1.52 6 0.16
which is smaller than 2.0 predicted by the three-dimensional
3
4
(
3D) percolation theory. The figure indicates that the con-
lyzed reaction leads to a weakly branched network
3
7
ductivity can be determined almost all by the IEC measure-
ments at least 95%RH.
immersed in a solvent. When PWA, a superacid is used, the
chain extension might occur drastically in preference to
4
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2012, 000, 000–000

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
1200 min, respectively, for HCl- and PWA- Sul-TesOct mem-
branes. Therefore, all the conductivity measurements shown
above were performed after immersion of the membranes
more than 1 day.
FIGURE 3 The relative humidity dependence on r with differ-
ent CMPTS obtained, respectively, (a) for the HCl-Sul-TesOct and
ꢀ
(
b) the PWA-Sul-TesOct membranes at 80 C.
crosslinking of tetrafunctional monomers, leading to instabil-
ity of the system and the domain structure would noticeably
grow.
Figure 5 shows the oxidation time dependence on the proton
conductivity. The HCl- and PWA-Sul-TesOct membranes were
divided into several pieces to prepare the samples with the
different oxidation time. As can be seen from the figure, the
data for the PWA-Sul-TesOct membranes reached a plateau
value at the shorter soaking time. From all above experimen-
tal findings, one can speculate that the porosity of the mem-
brane was much larger for the PWA-Sul-TesOct membranes
than those for HCl-Sul-TesOct. Although the SEM micrographs
showed the structure (not shown here), further investigation
is required to ensure the hypothesis. It is noted that the
proton conductivity reached plateau values at 800 and
FIGURE 4 (a) The water uptake of the membranes (b) normal-
ized swelling ratio, and (c) the schematic model of the mem-
brane structures.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2012, 000, 000–000
5

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
groups. However, the reaction process accompanying three
functional MPTS under superacid environment becomes
tricky. If the concentration fluctuations are not well sup-
pressed by crosslinks due to the presence of trifunctional
species, the system becomes unstable as large molecular-
weight species are formed, leading to phase separation or
formation of the porous structures.
Subsequently, the ratio of the peak areas attributed to MPTS
and TesOct was calculated to evaluate the MPTS composition.
Figure 6(c) shows the composition of MPTS obtained by
TGA, rTGA, versus the feed composition, rfeed. As the result of
nonleast squares fitting, slope was about 0.55 irrespective of
the type of catalyst, indicating that 55% of MPTS monomer
was actually incorporated into the TesOct network. By taking
FIGURE 5 The oxidation time dependence on the proton con-
ꢀ
ductively r observed at 80 C.
To clarify the structural difference between the HCl- and
PWA- catalyzed Sul-TesOct membranes, TGA were carried
out. The weight less curve (weight W vs. temperature T) was
differentiated with respect to the temperature. dW/dT exhib-
ꢀ
ꢀ
ited two peaks around at 350 C (region I) and at 500 C
region II) as shown in Figure 6(a,b). The former corre-
sponds to the dissociation of MPTS, whereas the latter corre-
(
3
8
sponds to that of the highly condensed block made of
TesOct, which is confirmed by independent FTIR measure-
ments. It should be noted that there is another peak in the
region I probably attributed to the different number (i.e., sin-
glet, doublet, or triplet) of siloxane links in MPTS. Subse-
quently, peak separation was carried out to obtain the ampli-
tude and the locations of the three peaks. The curve was
well reproduced by the sum of three Lorentzian functions.
As indicated by the arrow, the peak heights in region I
increased with CMPTS because of incorporation of MPTS into
the network. Conversely, those in region II decreased with
CMPTS, indicating that the highly condensed cage-like struc-
tures made of 6-functional TesOct became less branched due
to incorporation of trifunctional MPTS. For both systems, all
the peaks became broadened with CMPTS. However, noticea-
ble peak shift was only observed for the PWA-Sul-TesOct
samples as indicated in the region II in Figure 6(b).
Without MPTS, it has been already known that the PWA cat-
alyst leads to a rigid structure of TesOct network having
4
high thermal stability. The dissociation temperature of the
previously reported PWA-TesOct membranes appeared
ꢀ
around 500–520 C, which was fairly larger than that of the
FIGURE 6 Temperature derivative of the weight loss dW/dT
evaluated by TGA obtained for the (a) HCl and (b) PWA-Sul-
TesOct membranes. The solid lines indicate the best fit by
ꢀ
conventional HCl-TesOct membranes (490 C). This is prob-
ably due to the high charge density of superacid, PWA, which
sufficiently promotes the condensation reaction of the silanol
three multiple Lorentzian functions. (c) rTGA versus rfeed
.
6
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2012, 000, 000–000

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
Sul-TesOct membranes, it was quite difficult to find the dif-
ference of the structures between PWA- and HCl-Sul-TesOct
membranes for the larger CMPTS only from the AFM analysis.
Conversely, the TGA results suggested that the low thermo-
stable species were dominant (region I in Fig. 6) in the
PWA-Sul-TesOct membranes. Thus, it provides an insight for
the structure of PWA-Sul-TesOct membranes where the
chemical bonding is weaker than those for the HCl-Sul-
TesOct. In general, multifunctional precursor such as TesOct
with a super acid would lead to a well-developed network.
However, when trifunctional coupling agent is incorporated
into the network, the concentration fluctuations grow so rap-
idly because of the loose structure involving less junction
points, resulting in macrophase separation. This suggests
that superacid PWA is not always the best catalyst for this
particular purpose, instead a normal acid like HCl could pro-
duce a gel like structure suitable for the better proton
conduction.
FIGURE 7 FTIR spectra for the HCl-Sul-TesOct membranes
before the oxidization procedure indicating the presence of (a)
organic bridge of TesOct and (b) mercapto group of MPTS.
The spectra were acquired after TGA under different destina-
tion temperatures. The dotted line shows the data after the oxi-
dization procedure.
CONCLUSIONS
Sol–gel derived hybrids containing sulfonic acid groups were
fabricated to obtain thermostable proton conductive mem-
branes. The conductivity reached r ¼ 0.9 (S/cm), which was
the real composition instead of the feed value into account,
we obtained the good agreement between the experimental
and predicted lines of IEC as shown in Figure 2(a).
Figure 7(a,b) show FTIR spectra for the HCl-Sul-TesOct mem-
branes before the oxidization procedure. The spectra indicate
the presence of (a) organic bridge of TesOct and (b) mer-
capto group of MPTS, which are respectively evidenced by
ꢂ1
ꢂ1
the peaks appeared at 2900 cm ; (CH2)8 and 2550 cm
;
(
CH SH. The spectra were acquired after TGA under differ-
2 3
)
ent destination temperatures where disappearance of the
corresponding peaks on heating was clearly confirmed.
Although the peak associated to the sulfonic acid group over-
lapped with other peaks, it was found that disappearance of
the MPTS peak after oxidization was clearly observed as
indicated by the dotted curve in Figure 7(b), suggesting
most of the SH groups are transformed into the sulfonic acid
groups at least in the detectable level of the FTIR
measurements.
Figure 8(a) shows the AFM height morphology obtained for
the (a) HCl and (b) PWA-Sul-TesOct membranes at CMPTS¼0.
As can be seen from the images, the latter has the larger do-
main structure than the former one. To quantitatively evalu-
ate the structural difference, the 1D power spectra were
evaluated from the height morphology by taking fast Fourier
transformation, followed by radial average. As shown in Fig-
ure 7(b), the correlation length n evaluated from the power
spectra (inset) decreased rapidly with CMPTS for the PWA-
Sul-TesOct membranes, whereas those for the HCl-Sul-TesOct
was rather invariant. The behavior was quite similar to that
obtained for the characteristic temperature of the mem-
branes evaluated by TGA (region II in Fig. 6). Although the
AFM results indicated that addition of MPTS led to dissocia-
tion of the rigid structure of TesOct precursor in the PWA-
FIGURE 8 The AFM height morphology obtained for the (a)
HCl and (b) PWA-Sul-TesOct membranes at CMPTS ¼ 0. (c) The
correlation length n versus CMPTS. The power spectra were
shown in the inset.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2012, 000, 000–000
7

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
respectively two orders and one order higher than the previ-
ously reported values for the similar hybrids and Nafion
12 Won, J.; Park, H. H.; Kim, Y. J.; Choi, S. W.; Ha, H. Y.; Oh, I.-
H.; Kim, H. S.; Kang, Y. S.; Ihn, K. J. Macromolecules 2003, 36,
3
228–3234.
112. Beside the high conductivity at the high humidity condi-
1
3 Serpico, J. M.; Ehrenberg, S. G.; Fontanella, J. J.; Jiano, X.;
tion, the conductivity depends on the type of catalysts partic-
ularly at the lower humidity. When PWA was used as a cata-
lyst, the conductivity became poor with decreasing the
humidity, which was attributed to the porous structure
where water molecules are easily released on drying. The
swelling ratio, water uptake, thermo-gravimetry analysis, and
AFM results supported the hypothesis. Although multifunc-
tional monomers such as bis(triehoxysilyloctane) have poten-
tial to produce versatile structures, the catalyst used here
plays an important role to control the mesostructures suita-
ble for the better proton conduction. Such differences are
not realized from the IEC measurements, suggesting the im-
portance to carry out the structural or thermal analysis to
understand the mechanism of the proton conduction.
Perahia, D.; McGrady, K. A.; Sanders, E. H.; Kellogg, G. E.;
Wnek, G. E. Macromolecules 2002, 35, 5916–5921.
1
4 Sel, O.; Soules, A.; Ameduri, B.; Boutevin, B.; Laberty-Rob-
ert, C.; Gebel, G.; Sanchez, C. Adv. Funct. Mater. 2010, 20,
1
090–1098.
1
4
5 Bae, B.; Miyatake, K.; Watanabe, M. Macromolecules 2010,
3, 2684–2691.
1
6 Li, N. W.; Liu, J.; Cui, Z. M.; Zhang, S. B.; Xing, W. Polymer
2009, 50, 4505–4511.
17 Peckham, T. J.; Schmeisser, J.; Rodgers, M.; Holdcroft, S. J.
Mater. Sci. 2007, 17, 3255–3268.
18 Miyatake, K.; Chikashige, Y.; Higuchi, E.; Watanabe, M. J.
Am. Chem. Soc. 2007, 129, 3879–3887.
19 Takata, Y.; Norisuye, T.; Hirayama, S.; Takemori, T.; Tran-Cong-
Miyata, Q.; Nomura, S. Macromolecules 2007, 40, 3773–3778.
ACKNOWLEDGMENTS
20 Aoki, Y.; Norisuye, T.; Tran-Cong-Miyata, Q.; Nomura, S.;
Sugimoto, T. Macromolecules 2003, 36, 9935–9942.
This work is supported by Grant-in-Aid, No. 18750190 (for T.
N.) and Grant-in-Aid for Scientific Research on Priority Area,
2
1 Rollet, A. L.; Diat, O.; Gebel, G. J. Phys. Chem. B 2002, 106,
3
033–3036.
‘‘Soft Matter Physics’’ (No. 463/19031018) from the Ministry of
2
2 Gierke, T. D.; Munn, G. E.; Wilson, F. C. J. Polym. Sci.
Education, Science, Sports, Culture, and Technology. The
authors thank Shinji Sugihara, Department of Applied Chemis-
try and Biotechnology, University of Fukui, for the valuable
discussion.
Polym. Phys. Ed. 1981, 19, 1687–1704.
2
3 Gebel, G.; Lambard, J. Macromolecules 1997, 30, 7914–7920.
4 Ma, X. H.; Zhang, C. J.; Xiao, G. Y.; Yan, D. Y.; Sun, G. M. J.
2
Polym. Sci. Part A: Polym. Chem. 2008, 46, 1758–1769.
2
5 Mikami, T.; Miyatake, K.; Watanabe, M. J. Polym. Sci. Part
A: Polym. Chem. 2011, 49, 452–464.
REFERENCES AND NOTES
26 Brinker, C. J.; Scherer, G. W. Sol-Gel Science; Academic
Press: London, 1990.
1
2
Shea, K. J.; Loy, D. A. Chem. Mater. 2001, 13, 3306–3319.
2
7 Norisuye, T.; Ikezawa, Y.; Takemori, T.; Tran-Cong-Miyata,
Q.; Nomura, S. Kobunshi Ronbunshu 2008, 65, 716–729.
8 Daiko, Y.; Katayama, H.; Araki, T.; Mineshige, A.; Kobune, M.;
Loy, D. A.; Carpenter, J. P.; Alam, T. M.; Shaltout, R.; Dorh-
out, P. K.; Greaves, J.; Small, J. H.; Shea, K. J. J. Am. Chem.
Soc. 1999, 121, 5413–5425.
2
Yusa, S.; Yazawa, T.; Nagasaka, K.; Murata, S.; Matsumoto, S.;
Koiwai, A.; Konomi, I. J. Phys. Chem. C 2009, 113, 1891–1895.
3
Sato, H.; Norisuye, T.; Takemori, T.; Tran-Cong-Miyata, Q.;
Nomura, S. Polymer 2007, 48, 5681–5687.
Nakanishi, T.; Norisuye, T.; Sato, H.; Takemori, T.; Tran-
Cong-Miyata, Q.; Nomura, S.; Sugimoto, T. Macromolecules
2
9 Loy, D. A.; Jamison, G. M.; Baugher, B. M.; Myers, S. A.;
4
Assink, R. A.; Shea, K. J. Chem. Mater. 1996, 8, 656–663.
3
0 Karstedt, B. D. U.S. Patent No. 3,775,492, 1973.
2007, 40, 4165–4172.
3
1
1 Oviatt, H. W., Jr.; Shea, K. J.; Small, J. H. Chem. Mater.
993, 5, 943–950.
5
Miyatake, K.; Tombe, T.; Chikashige, Y.; Uchida, H.; Wata-
nabe, M. Angew. Chem. Int. Ed. 2007, 46, 6646–6649.
3
2 Kim, J. D.; Honma, I. Electrochim. Acta 2004, 49, 3429–3433.
6
Miyatake, K.; Watanabe, M. J. Mater. Chem. 2006, 16,
4465–4467.
33 Lefebvre, M. Chem. Mater 1999, 11, 262–268.
7
Takimoto, N.; Wu, L.; Ohira, A.; Takeoka, Y.; Rikukawa, M.
34 Stauffer, D. Introduction to Percolation Theory; Taylor &
Francis: London, 1985.
Polymer 2009, 50, 534–540.
8
Tung, S. P.; Hwang, B. J. J. Mater. Chem. 2005, 15,
35 Nogami, M.; Daiko, Y.; Akai, T.; Kasuga, T. J. Phys. Chem. B
2001, 105, 4653–4656.
3532–3538.
9
Uma, T.; Nogami, M. Anal. Chem. 2008, 80, 506–508.
36 Onizuka, H.; Kato, M.; Shimura, T.; Sakamoto, W.; Yogo, T.
J. Sol-Gel Sci. Technol. 2008, 46, 107–115.
10 He, S. L.; Yu, J. J.; Li, C. H.; Fang, J. H.; Xu, J.; Guan, R.
Polym. Eng. Sci. 2011, 51, 264–271.
37 Norisuye, T.; Inoue, M.; Shibayama, M.; Tamaki, R.; Chujo,
Y. Macromolecules 2000, 33, 900–905.
11 Lee, C. H.; Wang, Y. Z. J. Polym. Sci. Part A: Polym. Chem.
008, 46, 2262–2276.
2
38 Ni, Y.; Zheng, S. X. Chem. Mater. 2004, 16, 5141–5148.
8
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2012, 000, 000–000