New proton conducting membranes with high retention of protic ionic liquids†
Yun Sheng Ye,
aMing Yao Cheng,
bJi Yong Tseng,
aGao-Wei Liang,
aJohn Rick,
aYao Jheng Huang,
cFeng Chih Chang
cand Bing Joe Hwang*
aReceived 31st August 2010, Accepted 11th November 2010 DOI: 10.1039/c0jm02878a
New proton conducting membranes, exhibiting a high proton conductivity (102S cm1, without
humidification at 160C), based on the ionic liquid (IL) 1-butyl-3-methylimidazolium
tetrafluoroborate (BMIm-BF4) were investigated. The incorporation of a similar initial IL loading was
found to reduce the proton conductivity loss of a sulfonated polydiacetylene based membrane, i.e. SPDA70-IL70 by 19.2% and IL-based Nafion by 64.6%; additionally, the SPDA70-IL70 was able to
maintain a proton conductivity of more than 103S cm1after centrifugation. The introduction of both
a cross-linked structure and a high degree of sulfonation into the polymer matrix, resulted in significant improvements to the membranes’ thermal properties, the dispersion of ionic domains, and the retention of IL—together with a reduction of proton conductivity loss. The membranes, which have a high degree of sulfonation with thermally cross-linkable diacetylene in the polymer’s main chain, avoid the drawback of poor miscibility commonly associated with sulfonated polymers and cross-linking agents.
When the composite membrane was used in a H2/O2fuel cell without humidification at 100C, a power
density > 250 mA cm2was achieved together with a maximum current denisty of 122 mW cm2.
Introduction
Ionic liquids (ILs) are composed of organic cations (e.g., imi-dazolium, pyridinium, or tetraalkylammonium) and anions (halide, tetrafluoroborate, hexafluorophosphate, triflate, ami-dotriflate, etc); they have highly variable aqueous solubilities (depending upon the anion), ranging from highly soluble to immiscible. ILs are attracting significant attention, in many fields of chemistry and industry, because of their unique physico-chemical properties, such as their high ionic conductivity, high polarity, high density, high heat capacity, and high thermal and
chemical stability.1–5 These characteristic features of ILs offer
suitable properties for polymer electrolyte membranes (PEMs) and have been extensively studied to replace liquid electrolytes,
which are toxic, flammable and able to leach out.6–8
For polymer electrolyte membrane fuel cells (PEMFCs), the PEM is one of the key issues because it acts as an electrolyte to transport protons from the anode to the cathode. One of the main challenges in the development of PEMFCs is to develop membrane materials with high operating temperatures (above
100 C), to achieve higher reaction kinetics at both electrodes,
less poisoning at the anode, and easy heat and water
manage-ments of the stacks.9–13 For water-swelled polymeric systems,
however, the conductivity is strongly humidity dependent, so
they cannot be used above 100C due to the loss/evaporation of
water, which results in a significant loss of conductivity.14 In
contrast to which ILs are suitable for high temperature appli-cations, because they can act as proton transferring carriers in anhydrous conditions, while exhibiting: excellent chemical and thermal stabilities, non-flammability and non-volatility. There-fore, many studies have been carried out to develop IL-based membranes with high proton conductivities that can operate at
higher temperatures (100–200C) in anhydrous conditions.15–29
The development of IL-based PEMs mainly focuses on mixtures of well-known polymers with IL using various strategies, such
as: the in situ polymerization of vinyl monomers in ILs,18–20
doping polymers with ILs,21–23and solution casting IL-polymer
solutions.24–29
Although the IL-based membranes allow PEMFCs to operate at higher temperatures in anhydrous conditions while avoiding water dependent conductivity, they have a major drawback, i.e. the progressive release of ILs leading to a long-term decline in the
performance of the fuel cells.30,31To solve the leaching problem,
approaches incorporating inorganic fillers32,33into the polymer
membranes and the formation of cross-linkable structures appear to be efficient approaches towards retarding the degree of methanol diffusion and water uptake, while enhancing the
poly-mer’s mechanical properties and dimensional stability.34–38More
recently, Yan and Lu et al.20reported protic IL-based hybrid
proton conducting membranes, wherein the addition of silica
aDepartment of Chemical Engineering, National Taiwan University of
Science and Technology, Taipei, Taiwan. E-mail: bjh@mail.ntust.edu.tw
bGraduate Institute of Engineering, National Taiwan University of Science
and Technology, Taipei, Taiwan
cInstitute of Applied Chemistry, National Chiao-Tung University,
Hsin-Chu, Taiwan
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c0jm02878a
ª The Royal Society of Chemistry 2011
Materials Chemistry
Cite this: J. Mater. Chem., 2011, 21, 2723
www.rsc.org/materials
PAPER
fillers into the composite membranes was shown to both enhance the conductivity and prevent the release of the IL component from the composite membranes. To the best of our knowledge, the incorporation of cross-linking structures in the PEMs will enhance the retention of IL in the membrane; however, the presence of the cross-linking structure appears to lower proton
conductivity. Thus, the development of more efficient
membranes with both improved proton conductivities and enhanced retention of IL in the membrane remains an important challenge.
In the present study, we describe the synthesis and character-ization of IL-based cross-linked membranes. The thermal cross-linkable diacetylene-containing polymer, sulfonated poly-diacetylene (SPDA), was synthesized and used as the polymer matrix because the diacetylene groups have the ability to cross-link under irradiation or with heat treatment without the
evolution of volatile byproducts.39,40 In addition, such main
chain type cross-linked polymers avoid the drawback of poor miscibility often associated with common sulfonated polymer and cross-linking agents. Polymers with various degrees of sulfonation (DS) of SPDA were synthesized via oxidative coupling polymerization and were blended with
1-butyl-3-methylimidazolium tetrafluoroborate (BMIm-BF4) to form
IL-based membranes (SPDA-IL) after a thermal cross-linking process. The influences of the DS and the network structure on the properties of the resulting composite membrane were systematically investigated with respect to: the IL uptake, the leaching of IL by centrifugation, morphology, ionic conductivity and thermal stability. In addition, the influence of IL retention properties on proton conductivity, based on SPDA and recast Nafion containing IL, has also been compared in this study.
Experimental section
Materials
Nafion (5%) dissolved in a solvent comprising water and meth-anol, Bisphenol A ($99%), 2,5-dihydroxybenzenesulfonate, n-butylammonium bromide ($98.0%), sodium
tetra-fluoroborate ($98.0%), N,N,N0,N0-tetramethylethylenediamine
(TMEDA) ($99%), copper-iodine (CuI) ($99.5%), pyridine ($99%), hydrochloride acid (36.5–38.0%), dimethylformamide (DMF) ($99%), acetonitrile (HPLC grade), acetone ($99.9%), ethanol (EtOH) ($99.5%) and dimethyl sulfoxide (DMSO) ($99.9%) were purchased from Sigma-Aldrich and used as received. 1-Bromobutane ($99%), 1-methylimidazole ($99%), propargyl bromide (80 wt. % in toluene) and potassium carbonate ($99.0%) were purchased from Acros and used as
received. 4,40-(Propane-2,2-diyl)bis[(prop-2-ynyloxy)benzene]
(PBPB), potassium 2,5-bis(prop-2-ynyloxy)benzenesulfonate
(PBPBS) and the ionic liquid [1-butyl-3-methylimidazolium
tet-rafluoroborate (BMIm-BF4)] were synthesized according to
a modified procedure described elsewhere.41,42(See the ESI†).
Sulfonated polydiacetylene (SPDA)
To a dried three-neck, round-bottomed flask were added: CuI
(11.5 mg, 0.11 mmol), N,N,N0,N0-tetramethylethylenediamine
(TMEDA) (14 mg, 0.12 mmol), pyridine (0.185 g, 2.35 mmol), and 5 mL of anhydrous dimethyl sulfoxide (DMSO). To the flask
was attached a water condenser and an inlet and outlet for
oxygen. The reaction mixture was warmed to 50C while oxygen
was bubbled through. After 1 h, different ratios of the monomers PBPB and PBPBS (15 wt% solid content) in 10 mL of DMSO were added to the catalyst mixture to generate different polymers (Table 1). After 24 h, 8 mL of DMSO was added to the viscous solution, and the polymer was precipitated in 400 mL of meth-anol containing 2% HCl. The gray fibrous polymer was filtered and immersed in 50 mL of methanol containing 5% HCl with stirring over night. The solid was filtered and washed with
methanol–water several times and dried under vacuum at 60C
(yield 73%) (Scheme 1).
1H NMR (d
6-DMSO): d 1.58 (g), 4.91 (a), 6.86 (e), 6.95 (c),
7.00 (d), 7.13 (f), 7.32 ppm (b) (Fig. 1).;13C NMR: d 31.33 (19),
41.99 (18), 56.37–58.35 (3, 10, 13), 70.76 (1, 11), 77.04 (2, 12), 114.74 (15), 116.23–117.63 (5, 6, 8), 128.30 (16), 138.38 (7), 144.08 (17), 148.96 (9), 151.61 (4), 155.33 ppm (14) (Fig. 2).
Membrane preparation
SPDA and the BMIm-BF4(cf. Table 1) were dissolved in DMSO
at room temperature, to give a 15 wt% solution which was stirred overnight. The resulting solution was cast onto a glass plate and
heated at 60C for 24 h to slowly evaporate the solvent and then
heated at 180C for 4 h. The resulting membranes were blotted
with kimwipes to remove surplus IL. The Nafion-IL membrane
was prepared by a solution casting method described elsewhere.26
Monomer and polymer characterization
1H NMR and13C NMR spectra were recorded at 25C using an
INOVA 500 MHz NMR spectrometer. Solid state 13C NMR
spectra were recorded on a Buuker-400 MHz instrument using tetramethylsilane as an internal reference. FTIR spectra were obtained with a Nicolet Avatar 320 FTIR spectrometer; 32 scans
were collected at a spectral resolution of 1 cm1. The ion
exchange capacities (IECs) were determined by titration as in our
previous study.38,39 Wide-angle X-ray diffraction (WAXD)
spectra were recorded for powdered samples using a Rigaku D/ max-2500 type X-ray diffraction instrument. SAXS experiments were performed using the BL23A SWAXS instrument at the National Synchrotron Radiation Research Center (NSRRC),
Taiwan, using a 10 keV (wavelength l¼ 1.24 A) beam with a 0.5
nm diameter. The scattering wavevector transfer q¼ 4pl1sinq is
defined by l and the scattering angle 2q of X-rays. Samples for
SAXS (thickness¼ 150200 mm) were sealed between two thin
Kapton windows (80 mm thickness each) and measured at an
ambient temperature 26C. The proton conductivity of the
membrane was determined with an ac electrochemical impedance analyzer (PGSTAT 30), and the experiments involved scanning the ac frequency from 100 kHz to 10 Hz at a voltage amplitude of 10 mV. A DuPont Q100 thermo-gravimetric analyzer (TGA) was utilized to investigate the thermal stability of the membranes; the samples (10 mg) were heated from ambient temperature to 850
C under a nitrogen atmosphere at a heating rate of 20C min1.
The stability of the membranes was tested by measuring their
proton conductivities at 160C, after thermal treatment at the
same temperature for a designated time. The centrifugal test was carried out using a centrifugal process as detailed in our previous
ª The Royal Society of Chemistry 2011
study.43 The IL-based membrane was poured into centrifugal filters (MICORSEP 30 K OMEGA, PALL). The centrifugal process was performed at 20000 rcf (relative centrifugal force)
and 25C for several cycles until no weight loss was observed in
the membrane. The proton conductivity of these membranes was
measured at 160 C. The electrode system, electrochemical
procedure, fabrication of membrane electrode assembly (MEA), and the single-cell test procedure are described in the ESI†.
Results and discussion
Monomer and polymer synthesisMonomers PBPB and PBPBS were prepared with high yields
from propargyl bromide. The 1H-NMR (d
6-DMSO, 25 C)
spectra of the monomers are shown in Fig. 1a, wherein the presence of propargyl protons, at 2.54 and 3.53 ppm respectively, for the monomers PBPB and PBPBS is apparent. The
corre-sponding 13C NMR (d
6-DMSO, 25 C) shows resonances at
78.28 and 79.76 ppm, indicating the propargyl structure. The
FTIR spectrum is in agreement with the structure of the mono-mer PBPBS as indicated by the presence of bands at 3286 and
1181 cm1, corresponding to propargyl and sulfonic acid
struc-tures, respectively.
The synthesis of the three sulfonated polydiacetylenes (SPDA) SPDA50, 60 and 70, was carried out using the oxidative coupling
method developed by Hay (Scheme 1).44 SPDA50-70 were
obtained using three different PBPB to PBPBS ratios. The degree
of sulfonation (DS) in the final copolymers was determined by1H
NMR (d6-DMSO, 25 C) from the integration of aromatic
Table 1 Characterization of sulfonated polydiacetylene membrane and composite membrane
Polymer Theoretical content of PBPBS (mol%)a Measured content of PBPBS (mol%)b Theoretical IEC
(meq g1)b Measured IEC
(meq g1)c Reaction temp.
(C)d SPDA50 50 49.5 1.68 1.57 200.9 SPDA60 60 58.8 2.04 1.99 187.9 SPDA70 70 68.0 2.41 2.32 175.9 PDA 0 0 0 0 209.4 a
Determined from the initial feed ratio of PBPBS.bCalculated from1H NMR spectra.c
IEC measured with titration.dReaction exotherm temperature obtained from DSC at a heating rate of 20C.
Scheme 1 Schematic representation of the preparation of the cross-linked SPDA-IL membrane.
Fig. 1 (a)1H-NMR (d
6-DMSO, 25C) spectrum of PBPB, PBPBS and
SPDA50. (b)1H-NMR (d
6-DMSO, 25C) spectrum of SPDA50-70.
ª The Royal Society of Chemistry 2011
groups of the monomers PBPB [marked as ‘f’ and ‘e’ in the molecular formula in Fig. 1b] and PBPBS, [marked as ‘b’, ‘c’, and ‘d’ in the molecular formula in Fig. 1b]. Table 1 shows that the PBPB to PBPBS ratios in the final polymers were found to be very close to those in the feed ratios, allowing them to be easily adjusted and controlled. Moreover, a similar result was also observed for the measured ion exchange capacity (IEC) values for SPDA membranes, which were close to the theoretical values.
The 1H NMR and 13C NMR spectra of the polymers were
consistent with their assigned structures. No residual ethynyl protons (–C^CH), due to unreacted acetylene groups (–C^C–),
were observed for the polymers by1H NMR (Fig. 1a and 1b).
The 13C NMR spectra also showed that the terminal ethynyl
carbons were transformed into diacetylenic carbons after poly-merization. For example, the two diacetylenic carbons at 70.59 and 77.04 ppm in the spectrum of SPDA50 (Fig. 2) replace the two ethynyl carbons at 78.28 and 79.76 ppm in the spectra of the monomers PBPB and PBPBS respectively. These results indicate that the different DS of SPDA were successfully synthesized by oxidative coupling.
Cross-linking behavior of polymer with ionic liquid
The cross-linking of SPDA50–70 and PDA (polydiacetylene) was examined by DSC as shown in Fig. 3a. Pure PDA exhibits
a single sharp peak at 209.9C, which is attributed to the
cross-polymerization of diacetylene. It has been reported that the cross-polymerization of diacetylene often takes place in this
temperature range for aliphatic diacetylenes.45–47Fig. 3a shows
that the SPDA membranes (SPDA50–70) display an exotherm
that began around 130.8 C, peaked above 175.9 C, and
extended to 318.4C. The nature of such a dramatic difference
may be due to structural rearrangements facilitated by the higher DS of aromatic diacetylenes during cross-linked polymerization. In addition, for SPDA membranes, the exotherm peak temper-ature decreased with an increase in the DS of the membrane. These observations simply reflected the increase in the local concentration of sulfonated segments with reactive diacetylene
groups facilitating the ease of thermal activation.48,49 The
SPDA50 was taken as an example to study the curing behavior
by DSC after thermal treatment (at 180C) with designed times
(30–240 min). As shown in Fig. 3b, the exotherm decreased
gradually and disappeared completely after 240 min. These results indicate the complete cross-polymerization of the
diac-etylene in the SPDA membrane after 240 min at 180C.
The cross-linking of the membrane after the incorporation of IL into the polymer matrix was examined by FTIR. Fig. 4 shows the FTIR spectra of the non-crosslinked SPDA50, crosslinked SPDA50, SPDA50-IL composite membrane, and pure
BMIm-BF4. SPDA50 was selected as a typical example to analyze the
cross-linking behavior of IL-based membranes. The spectrum of
pure BMIm-BF4shows various characteristic peaks associated
with the IL structure. The first two bands at 3166 and 2943 cm1
Fig. 2 13C-NMR spectrum of non-crosslinked and crosslinked SPDA50.
Fig. 3 (a) DSC curves of non-crosslinked membranes (PDA and SPDA50–70). (b) DSC curves of uncross-linked SPDA50 after thermal treatments (at 180C) with designated time (30–240 min).
Fig. 4 FTIR spectrum of non-crosslinked SPDA50, crosslinked SPDA50, SPDA50-IL composite membrane, and pure Bmin-BF4.
ª The Royal Society of Chemistry 2011
represent the –CH groups of the imidazole ring of the Bmin-BF4,
where the signal for the C]C double bonds of the imidazole ring
is around 1572 cm1. The bands at 1121 and 754 cm1correspond
to strength vibrations of the B–F bonds in BF4, relating to the
structure of the anion.50,51 The spectrum of SPDA50 before
curing was characterized by the symmetric stretching vibration of
the C–O–C groups at 1223 cm1, the _C–C stretching vibration
band at 912–970 cm1, and the symmetric stretching vibration of
the sulfonic acid groups around 1186 cm1. After cross-linking,
the intensity of the _C–C stretching vibration band at 912–970
cm1decreased, while a new peak appeared at 1733 cm1(due to
carbonyl groups). These results suggest that the –CH2O– linkage
is involved in the thermal-curing process to a certain extent and is responsible for the formation of the new carbonyl in the cross-linking process. These results are similar to the previous results
for thermal cross-linked polydiacetylene,46 indicating that
a Claisen rearrangement reaction has taken place during the
thermal cross-linking process (Scheme 1).52,53 The 13C NMR
spectrum of SPDA50 (Fig. 2) also confirmed extensive reactions involving the diacetylene and methyleneoxy groups, as the peaks at 70.59 and 77.04 ppm (alkyne carbons) and at 56.37–58.27 ppm
(–CH2O–) almost completely disappeared. Similar results also
showed in other SPDA-IL membranes which indicate that the diacetylene groups were completely reacted, incorporating IL into the polymer matrix. Fig. 5 shows a photograph of the resulting SPDA-IL membrane (SPDA70-IL70 with the thickness of 160 mm). These membranes are flexible and can be cut into any desired sizes and shapes, whereas the pure form polymeric matrix (SPDA) (without ILs) is stiff and brittle.
Ionic liquid uptake and centrifugal test
The ionic liquid used in the present study was
1-butyl-3-meth-ylimidazolium tetrafluoroborate (BMIm-BF4). This IL has been
reported to a have dynamic viscosity of 2.33 P at 30 C and
a conductivity of 0.86 mS cm1 at 60 C.54 The loss after
centrifugation and thermal decomposition temperatures of SPDA membranes with different ratios of IL loading are shown in Table 2. The amount of absorbed IL into the membranes impregnated after the cross-linking procedure is less than the theoretical content of IL. In comparison with the Nafion-IL membrane, the difference between the theoretical and the measured content of IL in SPDA is similar. The lower IL uptake for the measured value can be mainly attributed to the presence of cross-linking in the polymer chains that restricts IL swelling in
Fig. 5 A photograph of the SPDA-IL membrane (SPDA70-IL70). Table
2 Quantit y o f ionic liquid uptake, IL loss after centrifu gal test and ther mal de compo sition tem peratu re of memb ranes Membr ane Theore tical IL uptake (wt% ) a M easured IL upt ake (wt%) b Loss after centrifu gal test (wt% ) R emain ing IL in memb rane (wt%) Td5 c Td10 c SPDA50 -IL5 0 5 0 28.6 1.6 27.4 311.1 340.1 SPDA50 -IL6 0 6 0 34.0 4.5 30.9 312.6 349.4 SPDA50 -IL7 0 7 0 40.4 11.5 32.7 316.6 348.4 SPDA60 -IL7 0 7 0 43.8 11.3 36.6 319.4 348.9 SPDA70 -IL7 0 7 0 47.5 10.5 41.3 317.2 349.1 Nafio n-IL50 50 47.8 22.0 33.1 359.6 382.8 a De termine d from the initial IL loadin g in the polym er. bC alculate d from TG A, using pris tine polym er co mpare d w ith compo site me mbrane. cDecomposition tempe rature as calculat ed at 5% and 10% weight loss in N2 , a s assessed b y TGA at a h eating rate of 20 C min 1.
ª The Royal Society of Chemistry 2011
the membrane. Generally, the absorption of IL in the membrane is dependent on the availability and accessibility of ionic groups. As the membrane content of sulfonic acid groups increases, more become accessible and available for ion exchange with the cation of the IL. Therefore, the content of IL in the SPDA membrane is expected to increase with increasing DS. A decrease in the membrane weight was observed after the leaching test.
Choud-hury et al.22demonstrated that the IL molecules in the membrane
can be classified into two parts: (a) where the IL is physically residing within the membrane and is not bound through inter-action, and (b) where the IL interacts with the membrane through electrostatic interactions. In this study, the centrifuge was used to probe the IL holding levels by physical separation of the solid (polymer) and the liquid phase (IL). The centrifugal test was carried out to determine the amount of IL that may be lost from the membrane. This test does not reflect real operating conditions for the membranes, but it helps to evaluate the long-term stability of these composite membranes. The percentage losses (%) and residual weights (wt %) of IL are shown in Table 2. All samples were subjected to a rotation speed of 20 000 rpm in the centrifugal test. Fig. 6 shows the amount of IL lost from the sample SPDA70-IL70 as a function of rpm. It was evident that a centrifuge rotation speed of 20 000 rpm could easily differen-tiate between strongly interacting IL and weakly interacting IL.
The amount of IL lost from the SPDA membranes (all with DS¼
50), was compared to the amount of IL initially loaded. The results indicated that a fraction of the IL was physically residing within the membrane and not bound through interactions, and was hence lost during centrifugation. The retaining capability within the polymer for the IL was improved by increasing the DS of SPDA membrane, indicating that to a significant degree ILs interact with the membrane through electrostatic interactions (Scheme 2). To clarify the effects of cross-linking on the retention of the IL, SPDA membranes (SPDA50-IL70) with various degrees of cross-linking (49.5–100%) were investigated (ESI†). It is not surprising that the IL loss after the centrifugation test decreased as the degree of membrane cross-linking increased, due to the presence of more cross-linked structure able to prevent the release of IL. The retention of IL, from SPDA50-IL70, was compared to that of SPDA70-IL70 and Nafion-IL50, since
similar IL loadings are found in these membranes. Higher amounts of IL were lost during centrifugation from the Nafion-IL50 membrane (22.0%) than the SPDA70-IL70 membrane (10.5%), indicating a higher number of sulfonic acid groups
(IEC ¼ 2.32 meq g1) to be present in the SPDA70-IL70
membrane. In addition, the network structure throughout the membrane allows for the entrapment of more IL in the membrane and prevents the release of IL component from the composite membranes.
Morphological changes in the presence of ionic liquid
In order to understand and explain the ionic conductivity results obtained from the cross-linked SPDA containing hydrophilic IL
(BMIm-BF4), the morphology of the membranes was studied by
X-ray diffraction (XRD) and small-angle X-ray scattering
(SAXS). In general, the cluster network model55,56and the core
shell model57,58 are the two main models most widely used to
explain the morphologies and ionic conductivities of water swollen Nafion membranes. These models feature a spherical ionic cluster phase within a continuous matrix phase. On the basis of quantitative analysis of published SAXS data, it has been shown that the angular position of the ionomer peak in the SAXS profile represents the average spacing between centers of ionic clusters. Recently, the morphology of proton conductivity membranes containing ILs has been investigated by small-angle
X-ray scattering (SAXS).17,26,59Swelling of the membrane ionic
domains was expected to cause changes in the ionomer’s morphology, especially in the ionomer peak of the SAXS profile. SAXS scattering profiles, plotted as an intensity vs. scattering vector q on a log–log plot, at various IL swelling levels and DS of the cross-linked SPDA are shown in Fig. 7a. The ‘‘ionomer peak’’
is presented in SPDA50-IL50 membrane at q¼ 0.077 A1(d¼
81.6 A) and q¼ 0.107 A1(d¼ 58.7 A). The presence of a weak
ionomer peak at q¼ 0.107 A1(d¼ 58.7 A) may result from the
restriction of the ionic domains by the cross-linked structure, leading to the small size of ionic domains containing IL. The membrane used in the present study was a cross-linked polymer
Fig. 6 Amount of IL lost from sample SPDA70-IL70 as a function of rpm.
Scheme 2 Graphical representation of a crosslinked SPDA-IL membrane.
ª The Royal Society of Chemistry 2011
and such cross-linked structures generally form small ionic
domains.34–38The ionomer peak systematically shifted to a lower
q as the IL content increased, i.e., 0.077 A1and 0.107 A1at 50
wt % to 0.072 A1 and 0.103 A1 at 70 wt % BMIm-BF
4,
respectively. This increase in inter-cluster spacing may corre-spond to an increase in cluster size through plasticization with IL. An increase in d spacing is generally observed for many
ionomers in the literature.17,26,59In case of the membrane
con-taining the same IL loading, the ionomer peak slightly decreases with an increase in the DS of the SPDA matrix (Table 2), which is related to the presence of the higher content of sulfonic acid
groups in the membrane. Moreover, the ionomer peak at q ¼
0.103 A1(d¼ 61.0 A) disappeared with higher DS (DS¼ 70).
Although the clustering of IL occurs due to aggregation in the hydrophobic cross-linked structure, as more sulfonic acid groups in the membrane can favor the accessibility of hydrophilic IL to ion exchange with sulfonic acid groups of the membrane. Therefore, a higher degree of DS in the polymer matrix allows more IL to ion exchange with sulfonic acid groups and results in homogeneous distribution of IL in the membrane matrix.
The morphological change in the presence of IL was also checked by XRD, as shown in Fig. 7b. For a pure SPDA membrane, the boarder amorphous peak appeared at a value of
2q¼ 30.06(d¼ 2.97 A). Upon incorporating 10 wt% of
BMIm-BF4, a new boarder shoulder appeared at 2q¼ 21.04(d¼ 4.22
A), an increased trend was observed with an increase in the IL loading in the membrane, suggesting that the polymer matrix was
compressed in the membrane due to the distribution of IL along the polymer chains, or that the polymer chains were restricted by IL domains to produce the different formation of cross-linked structure.
Ionic conductivity and single-cell test of composite membranes The ionic conductivities (s) of SPDA-IL and Nafion-IL membranes in anhydrous conditions were plotted as a function of temperature, see Fig. 8a. The corresponding conductivity
increased exponentially with temperature from the order of 104
S cm1 up to 102 S cm1 at 160 C. In general, the proton
conductivity is strongly dependent on its nanostructure and water content. The decrease in the water content of the membrane results in low conductivity, due to not all of the acid sites being dissociated and a low level of interaction among water molecules, via hydrogen bonding. The conductivity of the IL-based membranes shows a similar trend to the water-IL-based membrane. It may be due to the presence of IL resulting in an enhancement in the dissociation of protons from the sulfonated polymer causing an increase in proton transport mobility. The SPDA with a higher DS shows better ionic conductivity in the
same range of temperatures (80–160C), which may be due to the
larger number of sulfonic acid groups. In addition, the SPDA
Fig. 7 (a) SAXS spectra of IL50, IL60, SPDA50-IL70, SPDA60-IL70 and SPDA70-IL70. (b) XRD spectra of SPDA50, SPDA50-IL10, SPDA50-IL50, SPDA50-IL70 and SPDA70-IL70.
Fig. 8 Ionic conductivities as a function of temperature of SPDA-IL and Nafion-IL membranes in anhydrous conditions. (b) Ionic conductivity of the IL-uptake membranes before and after centrifugation.
ª The Royal Society of Chemistry 2011
with higher DS may provide better ion-conducting paths for proton conduction in the composite membrane. These results are consistent with previous results for membranes with different DS.60–62
Fig. 8b shows the ionic conductivities of the IL-uptake membranes before and after the centrifugal process. The
Nafion-IL50 membrane shows a higher conductivity (1.7 mS cm1)
compared to the SPDA-IL membranes at 160C, which is related
to a longer flexible spacer (free volume) that facilitates the
transportation of protons.63,64After centrifugation, the
conduc-tivities of all membranes decreased by more than the as-prepared ones (made without any centrifugation test), indicating a strong dependence on IL uptake in the membrane. In the case of a water-swelled electrolyte membrane, there are two main mechanisms for proton transport, namely proton hopping
(Grotthuss mechanism) and diffusion (vehicular mechanism).64
The two mechanisms may be adopted in this work as well. Before centrifugation, protons transport may occur by both hopping and diffusion in the membrane. However, after centrifugation, proton transport via diffusion is limited, since the amount of IL in the bulk phase that can facilitate proton diffusion is reduced. It is similar to proton transport under low humidity conditions for water-based electrolyte membranes. It is simply the case that after centrifugation, the SPDA membrane with a higher DS exhibits better proton transport, implying the presence of more sulfonic acid groups as well as a higher amount of IL in the membrane. In addition, the trend of variation in the proton conductivity is consistent with our supposition, i.e. that the conductivity reduction is less for membranes with higher DS.
For Nafion-IL composite membranes, a 64.7% reduction in its conductivity was observed after centrifugation (from 1.70 mS
cm1 to 0.63 mS cm1). In contrast, the conductivity of the
SPDA70-IL70 membrane decreased from 1.04 mS cm1to 0.84
mS cm1, i.e. only a 19.2% in reduction. Although the
Nafion-based composite membrane possesses a relatively higher conductivity than the sulfonated polymer, its proton conduc-tivity significantly decreases after centrifugation, indicating the benefits of both a higher DS and the presence of a network structure in the membrane.
The H2/O2fuel cell was operated at 100C using a
SPDA50-IL70 and SPDA70-SPDA50-IL70 composite membrane without
humidi-fication (ESI†). The maximum power density was 98 mW cm2
for the SPDA50-IL70 membrane and 122 for the SPDA70-IL70, which is approximately a 24% increase. This may be due to the higher DS of the SPDA70-IL70 increasing both the membrane IL retention and the proton conductivity. However, the perfor-mance is still much lower than that of conventional PEMFCs with hydrated ion exchange membranes. It is worth noting that the SPDA membrane was operated without humidification of the
supplied gases (H2and O2). Hence, this approach allows this type
of IL-based cross-linked membrane to have potential applica-tions in high-temperature polymer electrolyte membrane fuel cells.
Thermal stability of composite membranes
The thermal stabilities of pure IL (BMIm-BF4), SPDA50,
Nafion-IL50 and SPDA50-IL membranes were investigated by TGA as shown in Fig. 9. The corresponding onset temperatures
for weight losses of 5% and 10% are listed in Table 2. The TGA curve of pure IL exhibits excellent thermal stability up to 400.5
C, followed immediately by one-step decomposition. For
cross-linked SPDA50 membranes, it shows a continuous weight loss
between 225C and 520C, which is attributed to the
decom-position of sulfonic acid groups and aliphatic segments of propargyl. However, the onset temperatures for 5 and 10 wt%
weight losses are above 310 C and 340 respectively, for all
SPDA-IL membranes. It is evident that the decomposition by desulfonation of sulfonic acid groups has been suppressed and shifted to higher temperatures for all of the SPDA-IL membranes. It has been reported that the thermal stability of sulfonated polymers in the sodium form is much higher in
comparison with that in acid form.55,56This observation confirms
that the decomposition of sulfonic acid groups and aliphatic segments of propargyl is inhibited by the presence of IL since the protons originating from the sulfonic acid groups are partially replaced by the bulky cation of the IL, resulting in the increase of the thermal stability of the membrane.
Membranes employed in PEFCs operating in the range
100200 C, require a high degree of stability for long-term
Fig. 9 TGA curves of pure IL (BMIm-BF4), SPDA50, Nafion-IL50 and
SPDA50-IL membranes.
Fig. 10 Time-dependant proton conductivities of the SPDA-IL membranes under high temperature treatment (160C).
ª The Royal Society of Chemistry 2011
operation. The thermal stability of SPDA-IL membranes was
also evaluated by thermal treatment for various times at 160C.
Fig. 10 illustrates the time-dependant proton conductivities of the SPDA-IL membranes subjected to high temperature
treat-ment (160 C). Although the onset temperatures for 5 wt%
weight loss for all the SPDA-IL membranes are similar (Table 2), the corresponding proton conductivity decays with a reduction of IL loading in the membrane. The result is basically inconsis-tent with the TGA study, where a higher IL coninconsis-tent is able to improve the thermal stability of the membrane. Among these, the SPDA70-IL70 membrane with a higher IL loading maintains
adequate proton conductivity (>1 104S cm1) for one week
(at 160 C). This result, indicates that a higher DS results in
a better distribution of IL domains in the membrane matrix and leads to an effective improvement in membrane stability.
Conclusions
A new type of sulfonated polymer, containing diacetylene groups in the main chain, was successfully synthesized via an oxidative coupling approach. FTIR, NMR and DSC were used to confirm the structures and the crosslinking behavior of the polymer and the IL-based membrane. After the introduction of a network structure and a high DS into the polymer matrix, the IL-based membranes exhibited a significant improvement in their thermal properties, dispersion of ionic domains, retention of IL and a reduction of proton conductivity loss. These IL-based membranes with higher amounts of IL and DS in the polymer matrix show higher ionic conductivities without humidification
(ionic conductivity: SPDA50-IL50 ¼ 0.21 mS cm1;
SPDA70-IL70¼ 1.02 mS cm1). As a comparison, the incorporation of
a similar initial IL loading was found to reduce the proton conductivity loss of SPDA70-IL70 (19.2%) to about triple that of a IL-based Nafion membrane (64.6%), while showing a proton
conductivity of more than 103S cm1after centrifugation. In
H2/O2fuel cell operation using a composite membrane without
humidification, a current density higher than 250 mA cm2was
achieved with a maximum power denisty of 122 mW cm2at 100
C. We believe this approach gives this type of IL-based
cross-linked membrane potential applications in high-temperature polymer electrolyte membrane fuel cells.
Acknowledgements
We thank financial support from the National Science Council (under contract number NSC-99-2120-M-011-001), facilities from the National Taiwan University of Science and Tech-nology, and the National Synchrotron Radiation Research Center (NSRRC) is gratefully acknowledged.
References
1 T. Welton, Chem. Rev., 1999, 99, 2071.
2 Ionic Liquid: Industrial Applications for Green Chemistry; R. Rogers, K. Seddon, ed.; ACS Symposium Series 818; American Chemical Society: Washington, DC, 2002.
3 P. Wasserscheid and W. Keim, Angew. Chem., Int. Ed., 2000, 39, 3772. 4 R. D. Rogers and K. R. Seddon, Science, 2003, 302, 792–793. 5 J. Dupont, R. F. de Souza and P. A. Z. Suarez, Chem. Rev., 2002, 102,
3667.
6 M. A. Ratner and D. F. Shriver, Chem. Rev., 1988, 88, 109. 7 K. Xu, Chem. Rev., 2004, 104(10), 4303.
8 F. M. Gray, Polymer Electrolytes; RSC Materials Monographs; The Royal Society of Chemistry, Information Services: Letch-worth, U.K., 1997.
9 R. K. Ahluwalia and X. Wang, J. Power Sources, 2008, 177(1), 167. 10 G. Alberti, M. Casciola, L. Massinelli and B. Bauer, J. Membr. Sci.,
2001, 185(1), 73.
11 Q. Li, R. He, J. O. Jensen and N. Bjerrum, Chem. Mater., 2003, 15, 4896.
12 K. -D. Kreuer, S. J. Paddison, E. Spohr and M. Schuster, Chem. Rev., 2004, 104, 4637.
13 M. S. Dresselhaus and I. L. Thomas, Nature, 2001, 414, 332. 14 J. Roziere and D. Jones, Annu. Rev. Mater. Res., 2003, 33, 503. 15 P. Wasserscheid, T. Welton, ed. Ionic Liquids in Synthesis;
Wiley-VCH Verlag: Weinheim, Germany, 2003.
16 T. Ueki and M. Watanabe, Macromolecules, 2008, 41, 3739. 17 H. Chen, J. H. Choi, D. S. Cruz, K. I. Winey and Y. A. Elabd,
Macromolecules, 2009, 42, 4809.
18 M. Susan, T. Kaneko, A. Noda and M. Watanabe, J. Am. Chem. Soc., 2005, 127(13), 4976.
19 S. Yu, F. Yan, X. Zhang, J. You, P. Wu, J. Lu, Q. Xu, X. Xia and G. Ma, Macromolecules, 2008, 41, 3389.
20 Feng Yan, S. Yu, X. Zhang, L. Qiu, F. Chu, J. You and J. Lu, Chem. Mater., 2009, 21, 1480–1484.
21 M. Doyle, S. K. Choi and G. Oroulx, J. Electrochem. Soc., 2000, 147, 34.
22 M. K. Mistry, S. Subianto, N. R. Choudhury and N. K. Dutta, Langmuir, 2009, 25(16), 9240.
23 R. H. Brown, A. J. Duncan, J. H. Choi, J. K. Park, T. Wu, D. J. Leo, K. I. Winey, R. B. Moore and T. E. Long, Macromolecules, 2010, 43(2), 790.
24 Y. Y. He and T. P. Lodge, Chem. Commun., 2007, 26, 2732. 25 Y. Y. He, P. G. Boswell, P. Buhlmann and T. P. Lodge, J. Phys.
Chem. B, 2007, 111, 4645.
26 S. S. Sekhon, J. S. Park, E. Cho, Y. G. Yoon and C. S. Kim, Macromolecules, 2009, 42, 2054.
27 Q. Che, B. Sun and R. He, Electrochim. Acta, 2008, 53, 4428. 28 E. K. Cho, J. S. Park, S. S. Sekhon, G. G. Park, T. H. Yang,
W. Y. Lee, C. S. Kim and S. B. Park, J. Electrochem. Soc., 2009, 156(2), B197.
29 S. Y. Lee, T. Yasuda and M. Watanabe, J. Power Sources, 2010, 195, 5909.
30 A. Fernicola, S. Panero and B. Scrosati, J. Power Sources, 2008, 178, 591.
31 J. Kim, S. Mulmi, C. Lee, H. Park, Y. Chung and Y. Lee, J. Membr. Sci., 2006, 283, 172.
32 F. Pereira, K. Valle, P. Belleville, A. Morin, S. Lambert and C. Sanchez, Chem. Mater., 2008, 20, 1710.
33 Y. Jin, S. Qiao, L. Zhang, Z. Xu, S. Smart, J. Costa and G. Lu, J. Power Sources, 2008, 185, 664.
34 D. S. Phu, C. H. Lee, C. H. Park, S. Y. Lee and Y. M. Lee, Macromol. Rapid Commun., 2009, 30(1), 64.
35 J. Chen, M. Asano, Y. Maekawa and M. Yoshida, J. Polym. Sci., Part A: Polym. Chem., 2008, 46(16), 5559.
36 Y. S. Ye, Y. C. Yen, C. C. Cheng, W. Y. Chen, L. T. Tsai and F. C. Chang, Polymer, 2009, 50, 3196.
37 Y. S. Ye, W. Y. Cheng, Y. J. Huang, M. Y. Cheng, Y. C. Yen, C. C. Cheng and F. C. Chang, J. Membr. Sci., 2010, 362, 29. 38 Y. S. Ye, Y. J. Huang, C. C. Cheng and F. C. Chang, Chem.
Commun., 2010, 46, 7554.
39 T. M. Miller, E. W. Kwock, T. Baird and A. Hale, Chem. Mater., 1994, 6(11), 1569.
40 Y. Qi and Z. Y. Wang, Macromolecules, 2003, 36, 3146.
41 Y. S. Ye, Y. C. Yen, C. C. Cheng, Y. S. Syu, Y. J. Huang and F. C. Chang, Polymer, 2010, 51, 430.
42 E. Cho, J. S. Park, S. S. Sekhon, G. G. Park, T. H. Yang, W. Y. Lee, C. S. Kim and S. B. Park, J. Electrochem. Soc., 2009, 156(2), B197. 43 M. Y. Cheng, C. J. Pan and B. J. Hwang, J. Mater. Chem., 2009, 19,
5193.
44 A. S. Hay, D. A. Bolon and K. R. Leimer, J. Polym. Sci., Part A-1, 1970, 8, 1022.
45 M.-F. Grenier-Loustalot, V. Denizot and D. Beziers, High Perform. Polym., 1995, 7, 157.
46 C. Badarau and Z. Y. Wang, Macromolecules, 2004, 37, 147.
ª The Royal Society of Chemistry 2011
47 A. Chernykh, T. Agag and H. Ishida, Macromolecules, 2009, 42, 5121. 48 H. W. Beckham and T. M. Keller, J. Mater. Chem., 2002, 12, 3363. 49 Y. Xiaowu, Y. Luo, Y. Deng, Q. Yan, G. Zou and Q. Zhang, Eur.
Polym. J., 2008, 44, 881.
50 K. M. Dieter, C. J. Dymek Jr, N. E. Heimer, J. W. Rovang and J. S. J. Wilkes, J. Am. Chem. Soc., 1988, 110, 2722.
51 N. Nanbu, Y. Sasaki and F. Kitamura, Electrochem. Commun., 2003, 5, 383.
52 S. K. Dirlikov and Y. Feng, Poly. Prepr. (Am. Chem. Soc., Div. Polym. Chem.), 1991, 32(1), 363.
53 L. G. Picklesimer, U.S. Patent 4, 2266, 800 (Oct 7, 1980). 54 H. Olivier-Bourbigou, J. Mol. Catal. A: Chem., 2002, 182, 419. 55 T. D. Gierke, G. E. Munn and F. C. Wilson, J. Polym. Sci., Polym.
Phys. Ed., 1981, 19(11), 1687.
56 C. L. Marx, D. F. Caulfield and S. L. Cooper, Macromolecules, 1973, 6, 344.
57 J. Kao, R. S. Stein, W. J. MacKnight, W. P. Taggart and G. S. Cargill, Macromolecules, 1974, 7, 95.
58 W. J. MacKnight, W. P. Taggart and R. S. Stein, J. Polym. Sci., Polym. Phys., 1974, 45(1), 113.
59 R. H. Brown, A. J. Duncan, J. H. Choi, J. K. Oark, T. Wu, D. J. Leo, K. I. Winey, R. B. Moore and T. E. Long, Macromolecules, 2010, 43(2), 790.
60 M. Doyle, S. K. Choi and G. Proulx, J. Electrochem. Soc., 2000, 147, 34.
61 Q. Che, B. Sun and R. He, Electrochim. Acta, 2008, 53, 4428. 62 M. Yoshizawa and H. Ohno, Electrochim. Acta, 2001, 46, 1723. 63 S. Washiro, M. Yoshizawa, H. Nakajima and H. Ohno, Polymer,
2004, 45, 1577.
64 E. K. Cho, J. S. Park, S. S. Sekhon, G. G. Park, T. H. Yang, W. Y. Lee, C. S. Kim and S. B. Park, J. Electrochem. Soc., 2009, 156, B197.
ª The Royal Society of Chemistry 2011
