Rapid Proton Conduction through Unfreezable and Bound Water in a

Mar 16, 2009 - Figure 3 shows the scanning electron microscopy (SEM) image of a cross-section view ..... Gierke , T. D.; Munn , G. E.; Wilson , F. C. ...
0 downloads 0 Views 626KB Size
Download PDF
4656

J. Phys. Chem. B 2009, 113, 4656–4663

Rapid Proton Conduction through Unfreezable and Bound Water in a Wholly Aromatic Pore-Filling Electrolyte Membrane Nobuo Hara,† Hidenori Ohashi,‡ Taichi Ito,‡ and Takeo Yamaguchi*,†,‡ Department of Chemical System Engineering, The UniVersity of Tokyo, Hongo 7-3-1, Bunkyo-ku, Tokyo 113-8656, Japan, and Chemical Resources Laboratory, Tokyo Institute of Technology, Nagatsuta 4259, Midori-ku, Yokohama, Kanagawa 226-8503, Japan ReceiVed: December 2, 2008; ReVised Manuscript ReceiVed: January 21, 2009

We found that protons rapidly conduct through unfreezable and bound water in a pore-filling electrolyte membrane (PF-membrane), although many ions usually conduct through free water contained in polymer electrolytes. PF-membrane is a unique membrane that can suppress the swelling of filled sulfonated poly(arylene ether sulfone) (SPES) because of its rigid polyimide substrate. Based on low-temperature DSC measurements, this strong suppression of swelling resulting from the special structure of the polymer electrolyte results in unfreezable and bound water only; it does not contain any free water. Protons rapidly conduct through this structure. In addition, the activation energy of the proton conduction decreased from 16.3 to 9.1 kJ/mol in proportion to the increase in the ion exchange capacity (IEC) of the filled SPES, unlike the almost constant values of the SPES-cast membranes. This tendency of PF-membrane occurred because of the structure of the membrane, where the concentration of the sulfonic acid groups increased with increase in IEC, which became possible by squeezing free water using the swelling suppression of filled SPES. Without being constrained by the PF-membrane, this unique proton conduction through the structured water and highly concentrated sulfonic acid groups will help to develop future polymer electrolytes, particularly in the fuel cell field where protons need to conduct at various conditions such as temperatures below 0 °C, combined high temperature and low humidity, and the presence of fuels. Introduction 1

Much attention has been paid to fuel cell technologies, i.e., polymer electrolyte membrane fuel cells (PEFCs)2-4 including direct methanol fuel cells (DMFCs)5-7 and more recently solidstate alkaline fuel cells (SAFCs)8,9 because of their potential in electrical devices, automobiles, and stationary applications. Among the ion exchange materials for these fuel cells, many types of proton-conducting materials, e.g., hydrated acidic polymers,10,11 oxoacids,12 heterocycles,13,14 organic-inorganic composites,15,16 etc., have been developed. These protonconducting materials must show fast proton conduction and, at the same time, enhancement of proton conduction above 80 °C at low humidity or below 0 °C, and reduce the crossover of fuels and gases. Thus, the investigation of the mechanism of proton conduction is especially important, and their transport properties and the mechanism of proton conduction have been widely reported.17 Based on previous studies, two main proton conduction mechanisms through aqueous media have been proposedsthe vehicle and the Grotthuss mechanisms.18 The vehicle mechanism is based on the diffusion of protonated water molecules through aqueous media, the same diffusion mechanism as shown by many ionic species such as sodium ion. The Grotthuss mechanism is referred to as “structure diffusion”, because the protonic charges are transported by a rearrangement of the hydrogen-bonding structure of water molecules, which is faster than the diffusion of protonated water molecules. * Corresponding author. Tel: +81-45-924-5254. Fax: +81-45-924-5253. E-mail: [email protected]. † The University of Tokyo. ‡ Tokyo Institute of Technology.

Among many proton-conducting materials, water-containing material is one of the most important for low-temperature fuel cells. Hydrated acidic polymers are typical proton-conducting materials; e.g., a perfluorosulfonic acid polymer electrolyte such as Nafion,19,20 sulfonated hydrocarbon polymer electrolyte based on styrene and other vinyl structures,21,22 and sulfonated hydrocarbon polymer electrolyte based on polycondensation polymers such as poly(ether ether ketone)s,23,24 poly(sulfone)s,25,26 poly(imide)s,27,28 etc. For these aqueous polymer electrolytes, proton conductivity is highly correlated with the presence of water molecules and the dissociation of sulfonic acid groups. The vehicle and Grotthuss mechanisms both contribute to the proton conduction, and they have been compared and discussed in various ways, such as the activation energy29-33 and the selfdiffusion properties obtained from pulsed-field gradient NMR (PFG-NMR).34-36 The activation energy for proton conduction is supposed to be lower in the Grotthuss mechanism than in the vehicle mechanism, and speculation on the dominant mechanism has been based on the comparison of the values of the activation energy.29-31,33 However, the properties and the precise factors that affect the proton conduction have not been sufficiently clarified. Besides, these previously reported membranes did not show sufficient proton conductivity at low humidity. Above all, fuels such as methanol pass through these previously reported membranes, including Nafion. Generally, proton conduction and fuel crossover correlate closely with each other. Various approaches have been proposed to overcome the limitation of previous polymer electrolyte membranes. From the viewpoint of molecular design, many polymer electrolytes have been designed and synthesized.37-43 They are designed to incorporate nonsulfonated hydrophobic segments and sulfonated

10.1021/jp810575u CCC: $40.75  2009 American Chemical Society Published on Web 03/16/2009

Pore-Filling Electrolyte Membranes

J. Phys. Chem. B, Vol. 113, No. 14, 2009 4657

Figure 1. An illustration of the pore-filling electrolyte membrane and the chemical structure of the porous polyimide substrate and the filled polymer electrolyte of sulfonated poly(arylene ether sulfone).

hydrophilic segments to control phase separation and water hydration, and these materials succeeded in improving the mechanical stability and decreasing the crossover of fuels such as methanol. However, these materials still show the tradeoff relationship of proton conductivity and the fuel crossover which is the same order as Nafion and do not show enough performances for the application to PEFCs and DMFCs. As another method, we have recently proposed a novel wholly aromatic hydrocarbon pore-filling electrolyte membrane.44 This membrane is composed of polyimide as the substrate45 and sulfonated poly(arylene ether sulfone) (SPES) as the filling polymer,26 as shown in Figure 1. The porous polyimide substrate is completely inert to liquid fuels or to gas, and also rigid and mechanically very strong, which reduces any swelling of the filled polymer electrolyte. The filled polymer exhibits proton conductivity, and the porous substrate matrix mechanically prevents excess swelling of the filled polymer electrolyte. In addition, the substrate matrix restricts the change in membrane area from the dry to the swollen state. In this concept, the suppression of swelling is important and the substrate matrix needs to possess strong mechanical properties to restrict the highly swollen polymer electrolyte. This membrane showed an extremely low methanol crossover (300 times lower permeability than Nafion 117 membranes with a 30 wt % methanol feed solution at 25 °C). At the same time, it still showed high proton conductivity at 25 °C, in the order of 10-2 S/cm, even though the chemical structure of the filled polymer electrolyte is simple and is a previously reported structure. These compatible performances are much higher than the tradeoff relationship of Nafion and other previous polymer electrolyte membranes. It also showed a high oxidative and thermal durability. We believe that the key to compatibility of high proton conductivity and extremely low methanol crossover exists in the proton conduction mechanism of the pore-filling membrane. Focusing on the proton conduction, the concentration of sulfonic acid groups along the proton conduction passes and the state of hydration water of the polymer electrolyte are thought to be critical. The protons dissociate from sulfonic acid groups and then act as carriers of the proton conduction; hence, the proton conductivity is closely related to the acid concentration and the proton mobility.46,47 The hydration properties of the polymer, in particular the state of the water in the polymer electrolyte, is also essential and recently focused on, because it closely relates to the hydrogen-bonding network, which may relate to the proton conduction.39,48 These properties have usually been defined as follows: (i) free waterswater molecules that are not bound to the sulfonic acid groups and behave like bulk water; (ii) bound waterswater molecules that are weakly bound to the sulfonic acid groups; and (iii) unfreezable waterswater molecules that are strongly bound to the sulfonic acid groups.49,50 This method can be a useful way for elucidating the performance

and the proton conduction mechanism of a pore-filling electrolyte membrane. In this study, we prepared both cast and pore-filling electrolyte membranes with various ion exchange capacities (IECs). Porefilling membranes were prepared using the polymer-filling method, which was newly developed in this research. The synthesized SPES was impregnated into the pores of the porous polyimide substrate because different pore-filling membranes with varying IEC can be prepared, which made it possible to compare directly between cast membranes and pore-filling membranes. Then, we compared their properties, such as the state of hydration water from DSC measurements, and proton conductivity and its activation energy. The pore-filling membrane can control the swelling of the polymer electrolyte and the hydration degree independently without changing the basic chemical structure; thus, the relationship between proton conduction and hydration state was systematically studied. Experimental Methods Materials. 4,4′-Dichlorodiphenyl sulfone (DCDPS), purchased from Tokyo Chemical Industry Co., Ltd., was recrystallized from toluene and dried in vacuo. 4,4′-Biphenol (BP), purchased from Tokyo Chemical Industry Co., Ltd., and potassium carbonate, purchased from Wako Chemical Co., Ltd., were dried in vacuo before use. Anhydrous N-methyl-2pyrrolidone (NMP), anhydrous toluene, dimethylformamide (DMF), and sulfuric acid, purchased from Wako Chemical Co., Ltd., were used as received. 3,3′-Disulfonated-4,4′-dichlorodiphenyl sulfone monomer (SDCDPS) was synthesized from 4,4′-dichlorodiphenyl sulfone and fuming sulfuric acid (30% SO3, Wako Chemical Co., Ltd.) according to the previous method. Synthesis of SDCDPS monomer was confirmed by 1H NMR and FT-IR measurements. Synthesis of the SPES. Materials for synthesis of the SPES, i.e., DCDPS, SDCDPS, BP, and potassium carbonate, were dried at a temperature above 100 °C in vacuo overnight before synthesis to remove residual water. The synthesis reaction of SPES was nucleophilic aromatic substitution polycondensation, and was conducted according to the previous method.26 The degree of sulfonation, x (%) (Figure 1), was determined as the mole fraction (SDCDPS/(SDCDPS + DCDPS)), and was controlled by changing the ratio of the SDCDPS monomer to DCDPS monomer in the synthesis. Membrane Preparation. SPES bulk membranes were prepared by casting the SPES polymer solution in NMP (mass fraction of 10 wt %) on clean glass plates, followed by heating at 90 °C on a hot plate for 1 day. Pore-filling membranes (PISPES) were prepared by the polymer-filling method. Porous polyimide substrate was obtained from UBE Industries Ltd.; it was 25 µm thick and had a pore size of 100-300 nm and a porosity of 50 vol %.45 Porous polyimide substrate whose pores

4658 J. Phys. Chem. B, Vol. 113, No. 14, 2009

Hara et al.

were filled with NMP in advance was dipped in the dilute SPES polymer solution of mixed NMP and DMF (mass fraction less than 10 wt %, mixed ratio of NMP/DMF ) 1:1 (v/v)). Porous polyimide substrate was then placed on a clean glass plate heated at 120-150 °C on a hot plate. The concentration of SPES inside the pores of the porous polyimide substrate was gradually increased by vaporization of the solvent. Finally, the pores were almost completely filled with SPES. Cast and pore-filling membranes were soaked in deionized water and then dried at 80 °C in a vacuum oven for more than 1 day to completely remove the residual solvent. Membranes were then protonated in 0.5 M sulfuric acid aqueous solution for 1 day and then refluxed in deionized water at 60 °C to fully remove residual sulfuric acid and to normalize the membrane pretreatment conditions that would affect the state of water and the proton conductivity of SPES. Characterization. 1H NMR analyses were conducted on a Lambda-300 (JEOL) spectrometer. Spectra were obtained from an SPES solution in DMSO-d6. The FT-IR spectra of the SPES cast membranes and PI-SPES pore-filling membranes were measured using a MAGNA 550 (Nicolet) spectrometer. Synthesis of SPES and the degree of sulfonation and the IECs normalized by the dry weight (IECdry) were characterized from the results of 1H NMR measurements. Gel permeation chromatography (GPC, L-7100 Hitachi Co. Ltd., Shodex Oven A030, Showa Denko K.K., Shodex RI-71, Showa Denko K.K., UV L-7400, Hitachi Co. Ltd., Shodex Asahipak GF-7MHQ columns, Showa Denko K.K.) measurements of SPES solution in DMF were conducted. A standard curve used to determine the molecular weight was created from a polystyrene standard solution (TSK, Tosoh Corp.). The pore-filling ratio, φf (%), was estimated using the following equation:

φf )

d wmem - wdsub

wdsub

× 100

(4)

where L (cm) is the distance between the electrodes, R (Ω) is the resistance of the membrane, and S (cm2) is the cross-sectional area of the measurement. The activation energy Ea (kJ/mol) of the proton conduction was estimated using the following equation:

( ) Ea RT

(5)

(1)

× 100

(2)

s d is the membrane area in the swollen state and Smem where Smem is the dry substrate area, respectively. The entire membrane area was measured, including that of the substrate. The water content (%) was estimated using the following equation:

water content )

L RS

where σ0 is a pre-exponential factor, R is the universal gas constant, and T is the temperature.

s d Smem - Smem d Smem

σ)

σ ) σ0 exp -

d d is the dry membrane weight and wsub is the dry where wmem substrate weight. The membrane area was measured in water, and the change in membrane area ratio between the swollen state and the dry substrate, φa (%), was evaluated using the following equation:

φa )

characterized by differential scanning calorimetry (DSC) measurements (DSC7, PerkinElmer Inc.). Membranes were fully soaked in deionized water overnight and then put into sealed aluminum pans after removing surface water. Sample pans were first cooled to -50 °C, and then DSC data were taken while sample pans were heated to 10 °C at a scanning rate of 5 °C/ min. Calibration was conducted using an acetone standard, and the area of the melt endothermic peak was calculated from the heat of fusion of pure water, 334 J/g. Proton Conductivity. The proton conductivity σ (S/cm) of the membranes was determined by an ac impedance method using a Solatron 1260 impedance analyzer, the same method as reported in our previous study. Two platinum foil sheet electrodes were contacted in parallel with a membrane sample. The distance between the two electrodes was varied to verify any interface resistance effect. Measurements were conducted in the fully hydrated condition with changing temperatures below 60 °C. The frequency range of the measurements was 1 Hz-10 MHz. The conductivity of the membranes in the longitudinal direction was calculated using the following equation:

s d wSPES - wSPES d wSPES

× 100

Results and Discussion Synthesis of SPES. The FT-IR measurement of SPES-g is shown in Figure 2A, in which characteristic peaks were observed at 1030 and 1098 cm-1, assigned to symmetric and asymmetric stretching of the sulfonic acid group, respectively. From the FT-IR measurements of all SPES, the intensity of these two characteristic peaks increased with an increase in the degree of sulfonation. Figure 2B shows the result of 1H NMR measurement of SPES-g. Successful polymerization of SPES with various sulfonation degrees was confirmed from the peak assignments of the aromatic region.51 The degree of sulfonation x (%) of SPES was estimated from the integral ratios of proton peaks. The following equation shows the ratio between the 1H NMR integral values of aromatic regions and the degree of sulfonation x (%) of the copolymers:

(3) A7 Aothers

s is the weight of SPES in the swollen state and where wSPES d wSPES is the weight of SPES in the dry state. The weight of filled SPES inside the pore-filling membrane was estimated from the weight of the membrane and the filling ratio, φf. State of the Water. The state of the water contained in the SPES cast membranes and PI-SPES pore-filling membranes was

) [xH7]/[(100 - x)(H1 + H2 + H3 + H4) + x(H5 + H6 + H8 + H9)] x 2x (6) ) ) 16(100 - x) + 12x 800 - 2x

Therefore

Pore-Filling Electrolyte Membranes

J. Phys. Chem. B, Vol. 113, No. 14, 2009 4659

Figure 2. (A) FT-IR spectrum of porous polyimide substrate (PI-substrate), sulfonated poly(arylene ether sulfone) (SPES-g), and pore-filling electrolyte membrane (PI-SPES-g). (B) 1H NMR spectrum of sulfonated poly(arylene ether sulfone) (SPES-g) in DMSO-d.

x ) 100 ×

8A7 2A7 + Aothers

(7)

where A7 is the integral value of H7 proton signal which is ortho to sulfonic acid group, Aothers is the sum of the integral value of the other protons of H1,2,3,4,5,6,8,9, as shown in Figure 2B. The value of the degree of sulfonation x (%) was directly obtained by the insertion of the experimental and values into the equation. The degree of sulfonation ranged from 15.5 to 65.6% (SPESa,b,c,d,e,f,g), and the IECs of the SPES normalized by dry weight (IECdry) were calculated from the degree of sulfonation x and the molecular weight of the components of the copolymers, as shown in Table 1, which are similar to the values expected from the synthesis. From these measurements and the results of GPC measurements, we confirmed successful synthesis of SPES with high molecular weight. Preparation of the Pore-Filling Membranes. We succeeded in preparing all of the aromatic pore-filling electrolyte membranes by the polymer-filling method. FT-IR spectra of the porous polyimide substrate (PI) and the pore-filling membrane (PI-SPES-g) are shown in Figure 2A. The spectrum of PISPES-g shows characteristic peaks at 1030 and 1098 cm-1 assigned to the sulfonic acid groups that are the same as SPES, in addition to the characteristic peaks of the porous polyimide substrate. Figure 3 shows the scanning electron microscopy (SEM) image of a cross-section view and the surface view of the porous polyimide substrate and the pore-filling membrane (PI-SPES-a). The surface electrolyte polymer thickness is less than 2 µm for each side, much less than the whole membrane thickness of around 25-28 µm. The pore-filling ratio φf is 80-90%, and the pores of the PI substrate are almost completely filled with SPES from the rough estimation of the pore volume and the SEM image of a cross-section view and the surface view as shown in Figure 3. The membrane area change ratio φa of the pore-filling membrane is too small to detect at 0-1%, different from the large area change ratio of Nafion 117 and SPES cast membranes. The successful preparation of pore-filling membrane is verified from these results. Pore-filling electrolyte membranes made by the polymer-filling method were confirmed to be sufficiently rigid compared with those made by the monomer impregnation polymerization method.

TABLE 1: Properties of Sulfonated Poly(Arylene Ether Sulfone) (SPES) SPES

sulfonation degree xa (%)

IECb (mequiv/gdry)

Mwc (-)

Mnc (-)

Mw/Mnc (-)

SPES-a SPES-b SPES-c SPES-d SPES-e SPES-f SPES-g

15.5 24.1 34.8 43.8 46.5 62.9 65.6

0.73 1.10 1.53 1.86 1.96 2.51 2.60

79 000 97 000 92 000 66 000 81 000 92 000 93 000

39 000 42 000 43 000 32 000 35 000 43 000 40 000

2.0 2.3 2.1 2.1 2.3 2.1 2.3

a x ) 100 × 8A7/(2A7 + Aothers), where A7 is the peak area of the proton next to the sulfonic acid group and Aothers is that of the other protons, both estimated from the 1H NMR measurements partially shown in Figure 2B. b IEC in dry condition was calculated from the sulfonation degree x. c Weight-average molecular weight (Mw) and number average molecular weight (Mn) were calculated from the result of GPC measurements using polystyrene standard samples.

Porous Polyimide Substrate Suppresses the Swelling of Filled SPES Even at High IEC. The water content of the polymer electrolyte membrane is an important factor affecting the proton conductivity and transport phenomena. Excess swelling of the electrolyte membrane leads to low mechanical durability and dimensional change, which could be a reason for low fuel cell performances when fabricated into a membrane electrode assembly (MEA). In this study, we focused on comparing the SPES cast membranes and the SPES polymer filled inside the pore of the pore-filling electrolyte membranes. Figure 4 shows the water content of the SPES cast membranes, SPES inside the pore-filling membranes, and Nafion 117 membrane in the fully hydrated condition at 25 °C. The water content of the SPES cast membranes increased from 26 to 250% in proportion to the increase in IECdry. The water content of the SPES inside the pore-filling membranes also increased with the increase in IECdry; however, the degree of water content was much less than in the SPES cast membranes, especially for the SPES with high IECdry. This is because of the swelling-suppression effect, by which the porous polyimide substrate suppresses any excess swelling of the filled SPES. The water content properties of the SPES cast membranes have been reported previously, in which a drastic increase of water content was observed between 35 and 40% degree of sulfonation because of the increasing connectivity of the

4660 J. Phys. Chem. B, Vol. 113, No. 14, 2009

Hara et al.

Figure 3. SEM cross-section view and surface view of (A) the porous polyimide substrate, and (B) the pore-filling electrolyte membrane (PISPES-a).

Figure 5. Typical low-temperature DSC thermograms of SPES cast membranes and pore-filling membranes (SPES-b, PI-SPES-b, SPESg, PI-SPES-g).

Figure 4. Influence of the IEC on water content of Nafion 117 membrane, SPES cast membranes, and SPES filled in the PI-SPES porefilling membranes.

hydrophilic domain structure of the aggregation of sulfonic acid groups.26 In the same range of degree of sulfonation, the water content of the pore-filling membranes remained quite low, also because of the strong swelling-suppression effect. The precise characterizations of the hydration water of the filled SPES in the pore-filling membrane are discussed in the next section. Filled SPES Contains Only the Unfreezable and the Bound Water Molecules. The state of water is expected to be very important for those properties of the electrolyte membranes such as proton conductivity and methanol permeability. Particularly in the case of very low water content of the pore-filling membrane, the state of water may greatly affect the membrane properties. Generally, the state of water in a hydrated polymer can be defined as free water, bound water, and unfreezable water, and they are detected by lowtemperature DSC measurements.48,50 Figure 5A shows typical low-temperature DSC thermograms of the fully hydrated SPES cast membranes (SPES-b, SPES-g), and Figure 5B shows low-temperature DSC thermogram of all the porefilling membranes. SPES cast membranes showed a sharp melting endothermic peak at 0 °C derived from the free water shown by the dark region, and a broad melting endothermic peak below 0 °C derived from the bound water shown by the bright region. The pore-filling electrolyte membranes

showed only a single broad peak shown by the bright region, which was derived from the bound water. The content of the freezing water, both the free and the bound water content, were estimated from the total area of the overlapped broad peak of the DSC thermogram. The free water content was separated from the freezing water content by the fitted peak area of Gaussian distribution curve to the sharp endothermic peak shown by dark region, and the bound water content was calculated by subtraction of the free water content from the freezing water content. The content of unfreezable water was estimated by subtraction of freezing water content from the total water content. The number of water molecules per single sulfonic acid group, the hydration number λ (-), was calculated by dividing each water content by the molar quantity of sulfonic acid groups in the samples. Figure 6 shows the influence of the IECdry on the state of water, the hydration number λ for the SPES cast membranes, the pore-filling membranes, and Nafion 117 membrane. The hydration number of free and bound water of SPES cast membranes increased in proportion to the IECdry. It increased especially dramatically above the IECdry of 2.0, which is the same tendency as previously reported.48 In contrast, pore-filling membranes contained no free water molecules even for those filled with highly sulfonated SPES. Pore-filling membranes contained a very few bound water molecules, five at most, which slightly increased in proportion to the IECdry. Pore-filling membranes that were filled with SPES below the IECdry of 1.10 contained almost no bound water molecules and only unfreezable water molecules. The hydration numbers of unfreezable water molecules in pore-filling membranes are approximately the same as the SPES cast membranes, and slightly decreased in proportion to the increase in the IECdry.

Pore-Filling Electrolyte Membranes

J. Phys. Chem. B, Vol. 113, No. 14, 2009 4661

Figure 7. Influence of the IEC on the proton conductivity at 25 °C in the fully hydrated condition.

Figure 6. Influence of the IEC on the hydration number (λ): (A) free water, (B) bound water, (C) unfreezable water.

Therefore, only the unfreezable water molecules and a very small amount of bound water molecules and no free water molecules remained in the filled SPES inside the pore of the polyimide substrate, in which the hydration number is 10-15. This implies that all of the free water molecules and most of the bound water molecules are squeezed by the effect of swelling suppression of the porous polyimide substrate, and only the minimum number of structured water molecules are contained in the pore-filling membranes. Many types of polymer electrolyte membrane with various chemical structures have been developed to maintain high proton conductivity, mechanical stability, and low permeability to molecules such as methanol, etc. However, there is no membrane that has succeeded in squeezing out the free water molecules, especially for high IECdry. The pore-filling membrane succeeded in controlling the state of the water, containing only structured water, and this might be the reason for the extremely low methanol permeability that we reported previously.44 Proton Conduction through the Unfreezable and the Bound Water Molecules. The pore-filling membrane showed the same order of proton conductivity with Nafion 117 membrane in spite of very low water content, as we reported previously.44 In this study, we conducted the proton conductivity measurements precisely in order to evaluate the influence of the IECdry and the state of the water. Figure 7 shows the influence of IECdry on the values of proton conductivity of SPES cast membranes, the pore-filling membranes, and Nafion 117 membrane at 25 °C in the fully hydrated condition. The proton conductivity of SPES cast membranes increased in proportion to the IECdry below a value of 2.0; however, the value of the proton conductivity became constant when the IECdry was above the value of 2.0. This is because of the excess swelling of SPES cast membranes when the IECdry is above 2.0, as shown in the water content in Figure 4, which affected the concentration of the proton carriers. In contrast, the proton conductivity of pore-filling membranes increased in proportion to the IECdry. The highest proton conductivity of the pore-filling membrane is 0.052 S/cm in

which the pore-filling membrane is filled with SPES with IECdry equal to 2.60. This value is the same order with Nafion, and high enough for the application to PEFCs and DMFCs. The excess swelling of the filled SPES is suppressed by the swellingsuppression effect of the porous polyimide substrate, and the water content is kept lower than for the SPES cast membranes. This effect prevents the available protons from being diluted, and the proton conductivity keeps increasing in the pore-filling membranes filled with highly sulfonated SPES. We confirmed a high proton conductivity through the pore-filling membranes, although they contain no free water molecules and only unfreezable and a very small amount of bound water molecules from the state of the water as shown in Figure 6. Furthermore, pore-filling membranes that were filled with SPES when the IECdry was less than 1.10 showed proton conduction through the unfreezable water molecules only. Therefore, in the pore-filling membrane, we found the proton conduction through the structured water molecules, unfreezable, and very small amount of bound water molecules, without using any free water molecules, which is unique compared with other polymer electrolyte membranes previously reported that contain free water molecules. With proton conduction through the structured water molecules contained in the pore-filling membrane, the mechanism can be different from the membranes previously reported. Wide Range of Activation Energies of Proton Conduction through the Pore-Filling Membrane. The activation energy is very important for discussing the mechanism of proton conduction. The values of the activation energy for proton conduction of the pore-filling membranes and Nafion 117 membrane were calculated from the Arrhenius plots shown in Figure 8. The values of the SPES cast membranes were also calculated using the same method. The activation energies of proton conduction are plotted in Figure 9A against the IECdry. The observed activation energy for Nafion 117 is 11.3 kJ/mol, which is in good agreement with the value of 10.8 kJ/mol in the previous study.52 The activation energy for the SPES cast membranes ranges from 9.8 to 12.7 kJ/mol, and the values are almost in proportion to the increase of IECdry, which is also in good agreement with the previous study.32 On the other hand, the activation energy of the pore-filling membranes decreased from 16.3 to 9.1 kJ/mol in proportion to the increase of IECdry. In addition, it is very important to consider the effective concentration of the sulfonic acid groups by considering the water swelling effect of the polymer electrolyte. The IECs of the SPES in the swollen condition are lower than those in the dry condition because of the increase in the total weight from

4662 J. Phys. Chem. B, Vol. 113, No. 14, 2009

Hara et al.

Figure 8. Arrhenius plot of the proton conductivity of PI-SPES porefilling membranes and Nafion 117 membrane.

Figure 9. Influence of the IEC, (A) IECdry (mequiv/gdry) and (B) IECswollen (mequiv/gswollen), on the activation energy of Nafion 117 membrane, SPES cast membranes, and PI-SPES pore-filling membranes.

condition ranged from 0.6 to 1.7 mequiv/gswollen because of the low water swelling owing to the swelling-suppression effect of the porous polyimide substrate. As shown in Figure 9B, the activation energies of the SPES cast membranes and Nafion 117 membrane show only a small range; however, those of the pore-filling membranes have a wide range in proportion to the IECswollen. Therefore, the pore-filling membranes can change the IECswollen systematically even using the same chemical structure of SPES, which enables us to study the proton conduction mechanism by controlling the IECswollen. Relationship between the State of Water and the Activation Energy of the Proton Conduction. Focusing on the state of the water as a medium of proton conduction, previously reported polymer electrolyte membranes contain free water molecules in proportion to the IEC and also bound and unfreezable water molecules. With SPES cast membranes, the values of the activation energy ranged from 9.8 to 12.7 kJ/mol in this study. The water content of the pore-filling membrane is kept quite low because of the swelling-suppression effect of the porous polyimide substrate, and the concentration of the sulfonic acid groups is quite high, which is achieved only by the pore-filling membrane. The values of the activation energy of the proton conduction decreased from 16.3 to 9.1 kJ/mol in proportion to the increase of IECswollen, as shown in Figure 9B. We succeeded in controlling the IEC of filled SPES without increasing any free water by the pore-filling method and found a systematic change in the activation energy of proton conduction in proportion to the IEC. The water is highly structured by the effect of the high concentration of the sulfonic acid groups, and the dynamic structure of the hydrogen-bonding network is different from that of free water. The proton transition can be more favorable and frequent through the hydrogen bonding, which is affected by the high concentration of sulfonic acid groups, and the value of activation energy can be determined by this unique structure. We suggest that control of the state of water and the presence of highly concentrated sulfonic acid groups are necessary to develop new polymer electrolyte membranes for PEFCs in the future. Without being constrained by the pore-filling membrane, these basic physicochemical phenomena of proton conduction through the structured water will give a clue for developing future polymer electrolytes. Especially in the fuel cell field, new polymer electrolytes can be developed that show high proton conductivity in various operating conditions such as temperatures below 0 °C and high temperature and low humidity, and also low fuel permeability. Further research on the proton conduction mechanism can also be a strong aid in developing new proton conduction materials in the future. Conclusions

water swelling. The IEC of the SPES normalized by swollen weight IECswollen (mequiv/gswollen) was obtained by the following equation:

IECswollen ) IECdry ×

100 water content + 100

(8)

The activation energies of the proton conduction are plotted again in Figure 9B against the IECswollen. The IECswollen of the SPES all decreased greatly between 0.6 and 0.9 mequiv/gswollen because of the excess swelling of water. In contrast, the IECswollen of filled SPES in the pore-filling membrane in the swollen

We have prepared a wholly aromatic pore-filling electrolyte membrane from a porous polyimide substrate and SPES using a newly developed polymer-filling method. The SPES inside the pores contained no free water and only a little bound water and unfreezable water. The hydration number λ of the filled SPES in the pore was less than 15 and only structured water molecules were present. The pore-filling membranes with high IEC showed high proton conductivity, which is of the same order as Nafion. In the swollen condition, the activation energy of the pore-filling membranes decreased in proportion to the increase in the IECdry of the filled SPES, in contrast to the almost constant values of the SPES cast membranes that IECswollen decreased below 1.0 because of the excess swelling of water. We could control the water hydration of the polymer electrolyte

Pore-Filling Electrolyte Membranes and the IEC independently by the pore-filling method, and found that the activation energy of the proton conduction decreased in proportion to the increase of the IEC. Therefore, unique and rapid proton conduction can occur because of the structured water and highly concentrated sulfonic acid groups. Acknowledgment. Part of this research was supported by NEDO. Supporting Information Available: Supplementary data of the characterization of the water content and the state of water and the proton conductivity can be found in the online version. This material is available free of charge via the Internet at http:// pubs.acs.org. References and Notes (1) Steele, B. C. H.; Heinzel, A. Nature 2001, 414, 345–352. (2) Zhang, J. L.; Xie, Z.; Zhang, J. J.; Tanga, Y. H.; Song, C. J.; Navessin, T.; Shi, Z. Q.; Song, D. T.; Wang, H. J.; Wilkinson, D. P.; Liu, Z. S.; Holdcroft, S. J. Power Sources 2006, 160, 872–891. (3) Mehta, V.; Cooper, J. S. J. Power Sources 2003, 114, 32–53. (4) Prater, K. B. J. Power Sources 1994, 51, 129–144. (5) Deluca, N. W.; Elabd, Y. A. J. Polym. Sci., Part B: Polym. Phys. 2006, 44, 2201–2225. (6) McGrath, K. M.; Prakash, G. K. S.; Olah, G. A. J. Ind. Eng. Chem. 2004, 10, 1063–1080. (7) Heinzel, A.; Barragan, V. M. J. Power Sources 1999, 84, 70–74. (8) Varcoe, J. R.; Slade, R. C. T.; Yee, E. L. H.; Poynton, S. D.; Driscoll, D. J.; Apperley, D. C. Chem. Mater. 2007, 19, 2686–2693. (9) Varcoe, J. R.; Slade, R. C. T. Fuel Cells 2005, 5, 187–200. (10) Gierke, T. D.; Munn, G. E.; Wilson, F. C. J. Polym. Sci., Part B: Polym. Phys. 1981, 19, 1687–1704. (11) Zawodzinski, T. A.; Derouin, C.; Radzinski, S.; Sherman, R. J.; Smith, V. T.; Springer, T. E.; Gottesfeld, S. J. Electrochem. Soc. 1993, 140, 1041–1047. (12) Dippel, T.; Kreuer, K. D. Solid State Ionics 1991, 46, 3–9. (13) Decoursey, T. E. Phys. ReV. 2003, 83, 475–579. (14) Kawada, A.; Mcghie, A. R.; Labes, M. M. J. Chem. Phys. 1970, 52, 3121–3125. (15) Alberti, G.; Casciola, M. Annu. ReV. Mater. Res. 2003, 33, 129– 154. (16) Costamagna, P.; Yang, C.; Bocarsly, A. B.; Srinivasan, S. Electrochim. Acta 2002, 47, 1023–1033. (17) Kreuer, K. D.; Paddison, S. J.; Spohr, E.; Schuster, M. Chem. ReV. 2004, 104, 4637–4678. (18) Kreuer, K. D.; Rabenau, A.; Weppner, W. Angew. Chem., Int. Ed. 1982, 21, 208–209. (19) Costamagna, P.; Srinivasan, S. J. Power Sources 2001, 102, 242– 252. (20) Costamagna, P.; Srinivasan, S. J. Power Sources 2001, 102, 253– 269. (21) Ding, J. F.; Chuy, C.; Holdcroft, S. Chem. Mater. 2001, 13, 2231– 2233. (22) Serpico, J. M.; Ehrenberg, S. G.; Fontanella, J. J.; Jiao, X.; Perahia, D.; McGrady, K. A.; Sanders, E. H.; Kellogg, G. E.; Wnek, G. E. Macromolecules 2002, 35, 5916–5921. (23) Robertson, G. P.; Mikhailenko, S. D.; Wang, K. P.; Xing, P. X.; Guiver, M. D.; Kaliaguine, S. J. Membr. Sci. 2003, 219, 113–121.

J. Phys. Chem. B, Vol. 113, No. 14, 2009 4663 (24) Roziere, J.; Jones, D. J. Annu. ReV. Mater. Res. 2003, 33, 503– 555. (25) Ueda, M.; Toyota, H.; Ouchi, T.; Sugiyama, J. I.; Yonetake, K.; Masuko, T.; Teramoto, T. J. Polym. Sci., Part A: Polym. Chem. 1993, 31, 853–858. (26) Wang, F.; Hickner, M.; Kim, Y. S.; Zawodzinski, T. A.; McGrath, J. E. J. Membr. Sci. 2002, 197, 231–242. (27) Fang, J. H.; Guo, X. X.; Harada, S.; Watari, T.; Tanaka, K.; Kita, H.; Okamoto, K. Macromolecules 2002, 35, 9022–9028. (28) Genies, C.; Mercier, R.; Sillion, B.; Petiaud, R.; Cornet, N.; Gebel, G.; Pineri, M. Polymer 2001, 42, 5097–5105. (29) Dai, C. A.; Liu, C. P.; Lee, Y. H.; Chang, C. J.; Chao, C. Y.; Cheng, Y. Y. J. Power Sources 2008, 177, 262–272. (30) Gosalawit, R.; Chirachanchai, S.; Shishatskiy, S.; Nunes, S. P. Solid State Ionics 2007, 178, 1627–1635. (31) Jiang, Z. Y.; Zheng, X. H.; Wu, H.; Wang, J. T.; Wang, Y. B. J. Power Sources 2008, 180, 143–153. (32) Kim, Y. S.; Hickner, M. A.; Dong, L. M.; Pivovar, B. S.; McGrath, J. E. J. Membr. Sci. 2004, 243, 317–326. (33) Shahi, V. K. Solid State Ionics 2007, 177, 3395–3404. (34) Saito, M.; Ikesaka, S.; Kuwano, J.; Qiao, J.; Tsuzuki, S.; Hayamizu, K.; Okada, T. Solid State Ionics 2007, 178, 539–545. (35) Saito, M.; Hayamizu, K.; Okada, T. J. Phys. Chem. B 2005, 109, 3112–3119. (36) Saito, M.; Arimura, N.; Hayamizu, K.; Okada, T. J. Phys. Chem. B 2004, 108, 16064–16070. (37) Norsten, T. B.; Guiver, M. D.; Murphy, J.; Astill, T.; Navessin, T.; Holdcroft, S.; Frankamp, B. L.; Rotello, V. M.; Ding, J. F. AdV. Funct. Mater. 2006, 16, 1814–1822. (38) Liu, B. J.; Kim, D. S.; Murphy, J.; Robertson, G. P.; Guiver, M. D.; Mikhailenko, S.; Kaliaguine, S.; Sun, Y. M.; Liu, Y. L.; Lai, J. Y. J. Membr. Sci. 2006, 280, 54–64. (39) Kim, D. S.; Robertson, G. P.; Guiver, M. D.; Lee, Y. M. J. Membr. Sci. 2006, 281, 111–120. (40) Gao, Y.; Robertson, G. P.; Guiver, M. D.; Wang, G.; Jian, X.; Mikhailenko, S. D.; Li, X.; Kaliaguine, S. J. Membr. Sci. 2006, 278, 26– 34. (41) Sankir, M.; Bhanu, V. A.; Harrison, W. L.; Ghassemi, H.; Wiles, K. B.; Glass, T. E.; Brink, A. E.; Brink, M. H.; McGrath, J. E. J. Appl. Polym. Sci. 2006, 100, 4595–4602. (42) Li, Y. X.; Roy, A.; Badami, A. S.; Hill, M.; Yang, J.; Dunn, S.; McGrath, J. E. J. Power Sources 2007, 172, 30–38. (43) Ghassemi, H.; McGrath, J. E.; Zawodzinski, T. A. Polymer 2006, 47, 4132–4139. (44) Yamaguchi, T.; Zhou, H.; Nakazawa, S.; Hara, N. AdV. Mater. 2007, 19, 592–596. (45) Matsuo, M.; Fujii, Y.; Takagi, J.; Ohya, S. Ube Industries, Ltd. JP Patent 359860, 2004. (46) Peckham, T. J.; Schmeissert, J.; Holdcroft, S. J. Phys. Chem. B 2008, 112, 2848–2858. (47) Peckham, T. J.; Schmeisser, J.; Rodgers, M.; Holdcroft, S. J. Mater. Chem. 2007, 17, 3255–3268. (48) Kim, Y. S.; Dong, L. M.; Hickner, M. A.; Glass, T. E.; Webb, V.; McGrath, J. E. Macromolecules 2003, 36, 6281–6285. (49) Karlsson, L. E.; Wesslen, B.; Jannasch, P. Electrochim. Acta 2002, 47, 3269–3275. (50) Kim, S. J.; Park, S. J.; Kim, S. I. React. Funct. Polym. 2003, 55, 61–67. (51) Li, Y. X.; Wang, F.; Yang, J.; Liu, D.; Roy, A.; Case, S.; Lesko, J.; McGrath, J. E. Polymer 2006, 47, 4210–4217. (52) Park, Y. S.; Yamazaki, Y. Eur. Polym. J. 2006, 42, 375–387.

JP810575U