Improvement in characteristics of a Nafion membrane by proton conductive nanofibers for fuel cell applications

Improvement in characteristics of a Nafion membrane by proton conductive nanofibers for fuel cell applications

Journal of Membrane Science 530 (2017) 65–72 Contents lists available at ScienceDirect Journal of Membrane Science journal homepage: www.elsevier.co...

1MB Sizes 0 Downloads 1 Views

Journal of Membrane Science 530 (2017) 65–72

Contents lists available at ScienceDirect

Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci

Improvement in characteristics of a Nafion membrane by proton conductive nanofibers for fuel cell applications

MARK



Takahito Makinouchi, Manabu Tanaka, Hiroyoshi Kawakami

Department of Applied Chemistry, Tokyo Metropolitan University Hachioji, 1-1 Minami-osawa, Hachioji, Tokyo 192-0397, Japan

A R T I C L E I N F O

A BS T RAC T

Keywords: Fuel cell Nanofiber Polymer electrolyte membrane Proton conductivity Gas permeability

A novel proton conductive nanofiber composite membrane, which was composed of sulfonated polyimide (SPI) nanofibers and a typical polymer electrolyte, Nafion, was developed to improve electrolyte characteristics, especially proton conductivity at low relative humidity, for polymer electrolyte fuel cells. In addition to the contribution of the SPI nanofibers that were revealed to possess ultra-high proton conductivities in our previous study, the Nafion matrix indicated unique morphology in the composite membrane. The TEM observation suggested that the sulfonic acid groups in the Nafion matrix were attracted to the SPI nanofibers during the membrane formation process to construct acid-condensed layer near the nanofibers. Such acid-condensed layer at the interface of SPI nanofibers and Nafion matrix should have a positive impact on the proton conduction and water diffusion, especially under the low relative humidity conditions. Improved gas barrier property and membrane stability were also achieved by using the proton conductive SPI nanofibers as a framework for the polymer electrolyte membrane.

1. Introduction Proton exchange membranes (PEMs) are key materials for polymer electrolyte fuel cells (PEFCs) and are desired various characteristics including proton conductivity, gas barrier property, membrane stability, and so on [1–4]. At the present moment, perfluorosulfonated polymers, such as Nafion® (Du Pont), has been most widely utilized for PEMs because of their outstanding proton conductive characteristics. However, many problems remain to be solved on the Nafion membranes for future PEFCs that will operate under more severe conditions to reduce system costs [5]. For example, insufficient proton conductivity at low humidity, relatively high gas permeability, and deficient membrane stabilities are major issues of the Nafion membranes for the long-term PEFC operation [6]. In order to overcome these problems, there are mainly two approaches, except for most recently reported special study [7–10]: One is the development of alternative PEMs based on non-fluorinated polymers. In past decades, sulfonated aromatic hydrocarbon polymers, such as sulfonated polyimides (SPIs) [11–13], sulfonated poly(arylene ether)s (SPAEs) [14–16], and sulfonated polyphenylenes (SPPs) [17,18], have been actively studied to exceed the conventional Nafion membranes. Nevertheless, to the best of our knowledge, there have been few commercial PEMs based on sulfonated aromatic hydrocarbon polymers due to their low proton conductivity at low relative humidity and insufficient membrane



durability for long-term PEFC operation. Another is a more realistic approach, in which characteristics of Nafion membranes are improved by functional additives. Various kinds of fillers, such as silica nanoparticles, titania nanotubes, and graphene oxide nanosheets, have been attempted for the improvement in membrane stability and gas barrier property of the Nafion membranes [19,20]. In the most cases, however, the proton conductivity of the Nafion composite membrane decreased compared to the pristine Nafion membrane without additives because the proton conductive pathway was inhibited and water uptake for better proton conduction was suppressed by the fillers. Although sulfonation of these additives is a possible alternative to maintaining proton conductivity of the composite membrane, these hydrophilic nanomaterials have risks to elute from the composite membranes by water diffusion during the PEFC operation. Recently, much attention has been attracted to the development of composite PEMs containing polymer nanofibers, which were fabricated by an electrospinning method [21,22]. The electrospun polymer nanofibers with nanometer-scale diameters possess unique characteristics including extremely large surface areas, distinguished mechanical properties, and fast materials transport properties [23,24]. We have fabricated a series of nanofibers from various functional polymers by the electrospinning method. The obtained three-dimensional nanofiber network nanostructures have been attempted to wide variety of applications, such as catalysts [25], filters [26], flexible electrodes

Corresponding author. E-mail address: [email protected] (H. Kawakami).

http://dx.doi.org/10.1016/j.memsci.2017.02.018 Received 4 October 2016; Received in revised form 28 January 2017; Accepted 13 February 2017 Available online 16 February 2017 0376-7388/ © 2017 Elsevier B.V. All rights reserved.

Journal of Membrane Science 530 (2017) 65–72

T. Makinouchi et al.

2.3. Fabrication of SPI nanofibers

[27,28], and PEMs [29–32]. For instance, a nanofiber composite membrane consisted of proton conductive SPI nanofibers and SPI matrix showed better fuel cell performance compared with an SPI membrane without nanofibers [29]. In our recent study, it was revealed that the electrospun SPI nanofibers showed much higher proton conductivity along the nanofiber axis (5.1×10° S cm−1 at 90 °C, 95% RH) than the corresponding SPI membrane (8.3×10−2 S cm−1) under the same conditions [33]. It is assumed that the SPI nanofibers formed an effective proton conduction pathway by molecular orientation in the nanofibers. Y. A. Elabd and coworkers also reported similar high proton conductivity (1.5×10° S cm−1 at 30 °C, 90%RH) on the electrospun Nafion nanofibers [34]. In addition to their outstanding proton conductive characteristics, the nanofibers are expected to improve other characteristics of polymer electrolyte composite membranes, such as gas barrier property and membrane stabilities. In this study, in order to improve proton conductivity and other characteristics of a Nafion membrane, the proton conductive SPI nanofibers were utilized as a framework because the SPI nanofibers [33] possess better proton conductive characteristics than the Nafion nanofibers [34]. A series of PEM characteristics, such as proton conductivity, gas barrier property, and membrane stability, of the nanofiber composite membrane consisted of SPI nanofibers and Nafion matrix were investigated. In addition, morphological analyses including transmission electron microscopy (TEM) and small angle X-ray scattering (SAXS) were also attempted to the nanofiber composite membrane to discuss the effects of nanofibers on the Nafion membrane. Interestingly, the TEM observation first revealed the formation of acid-condensed layers at the interface of SPI nanofibers and the matrix Nafion. The acid-condensed layers have potential to construct efficient proton conductive pathways in the nanofiber composite membrane for future fuel cell applications.

The SPI nanofibers were fabricated by an electrospinning apparatus (ES-1000, Fuence, Co., Ltd., Tokyo, Japan) based on our previous study [33]. A typical electrospinning procedure is as follows: First, SPI (triethylamine form) was dissolved in anhydrous DMF with a concentration of 10 wt%, and the polymer solution was loaded into a 1 mL syringe as the spinneret. A syringe pump was used to squeeze out the polymer solution at a constant speed (20 μL/min) through a needle with an inner diameter of 0.21 mm. The distance between the spinneret and a collector (7.6 cm×7.6 cm aluminum foil) was 10 cm. The applied voltage between the spinneret and the collector was fixed to be 24 kV. Temperature and relative humidity inside the apparatus were maintained at 20–25 °C and 5–10%RH, respectively. The SPI nanofibrous membranes (nonwoven nanofiber mats) were collected on the aluminum foil, and their thickness was controlled by the electrospinning deposition time. The SPI nanofibrous membrane was dried under vacuum at 80 °C for 12 h to remove the residual solvent from the fabricated nanofibers. Finally, the SPI nanofibrous membrane was acidified with a 1 M HCl solution for 3 h to obtain SPI (proton form) nanofibers. After washing the SPI nanofibrous membrane by water several times, it was dried under vacuum at 80 °C for 12 h. The surface and cross-sectional images of the SPI nanofibrous membrane were obtained by scanning electron microscopy (SEM, JXP-6100P, JEOL, Tokyo, Japan). The average diameter of the SPI nanofibers was determined from the SEM images (at least five pictures and 25 nanofibers for each sample) by the Image-J software (Research Services Branch, NIH). The porosity of the SPI nanofibrous membrane was estimated by its apparent volume and weight.

2.4. Preparation of nanofiber composite membranes consisted of SPI nanofibers and Nafion matrix

2. Experimental 2.1. Materials

The nanofiber composite membranes were prepared using a solvent-cast method based on our previous study [29]. First, porosity (P) of the SPI nanofibrous membrane was calculated from Eq. (1):

1,4,5,8-Naphthalene tetracarboxylic dianhydride (NTDA) and Nafion dispersion (Nafion® perfluorinated resin solution, 5 wt% in lower aliphatic alcohols and water, contains 15–20% water) were purchased from Sigma-Aldrich Co. (St. Louis, MO, USA), and were used as received. 2,2-Bis[4-(4-aminophenoxy)phenyl]-hexafluoropropane (APPF) was purchased from the Wako Pure Chemical Industries Co. (Osaka, Japan), and was recrystallized twice from an ethanol solution prior to use. 4,4′-Diamino-biphenyl-2,2′-disulfonic acid (BDSA) was purchased from Tokyo Chemical Industry Co. (Tokyo, Japan). BDSA was purified by dissolution in an aqueous triethylamine solution, followed by precipitation in 1 M sulfuric acid. Finally, BDSA was dried in a vacuum oven at 80 °C for 12 h. All other chemicals were purchased from Kanto Chemical Co. (Tokyo, Japan) and were used as received.

⎞ ⎛ W (g ) P (%) = ⎜1 − ⎟ × 100 ⎝ d × l × w × t⎠

(1)

where d, W, l, w, and t are density of SPI, weight, length, width, and thickness of a piece of SPI nanofibrous membrane. The electrospun SPI (proton form) nanofibrous membrane, whose porosity was estimated to be 90%, was put in a petri dish. Then, a Nafion dispersion was slowly poured into the petri dish to fill in the void of the SPI nanofibrous membrane. The amount of Nafion dispersion was calculated from the volume of the SPI nanofiber, density of Nafion (not the Nafion dispersion but the Nafion solid after evaporating solvents), and the concentration of the Nafion dispersion to be the volume ratio of SPI nanofiber: Nafion matrix=10: 90 by considering the porosity of the SPI nanofibrous membrane. After pouring the Nafion dispersion, the petri dish covered with a clock glass was allowed to remain at room temperature for 12 h, to soak the Nafion matrix dispersion through the whole nanofibrous membrane. Then, the solvent was slowly evaporated in a vacuum oven at 60 °C for 12 h. After drying, the nanofiber composite membrane containing the SPI nanofibers was peeled off from the petri dish and acidified three times with a 1 M HNO3 solution at 80 °C for 30 min each, and finally washed with deionized water. The resulting membrane was dried in a vacuum oven at 60 °C for 24 h. For comparison, a recast-Nafion membrane without nanofibers was also prepared according to the similar procedure. Finally, both membranes were pre-treated by a typical annealing and acid immersion process reported elsewhere [35].

2.2. Synthesis and characterizations of SPI Sulfonated polyimide (SPI: NTDA-BDSA-r-APPF) was synthesized according to our previous study [33]. The successful synthesis of the objective SPI with triethylammonium counter ions was confirmed by 1 H NMR spectroscopy (AVANCE III 500, Bruker BioSpin K.K.). The number-average molecular weight (Mn) and weight-average molecular weight (Mw) of the SPI were estimated to be 2.3×105 and 5.8×105, respectively, by gel-permeation chromatography (detector: RI-2031, JASCO Co., Tokyo, Japan) using two Shodex SB-806HQ and SB804HQ columns with a eluent of N,N-dimethylformamide containing 0.01 M lithium bromide at a flow rate of 1.0 mL min−1. 66

Journal of Membrane Science 530 (2017) 65–72

T. Makinouchi et al.

(oxygen) was introduced on one side of the membrane at a flow rate of 50 mL min−1 at the appropriate temperature and dew-point. The carrier gas (helium) was introduced on the other side of the membrane at a flow rate of 20 mL min−1 and was analyzed by the GC. The membrane was equilibrated until reliable permeation data were obtained. The oxygen gas permeability coefficients PO2 (cm3(STP) cm cm−2 s−1 cm Hg−1) and the water vapor permeance QH2O (g m−2 24 h−1) were calculated by the following equations:

2.5. Characterization of PEMs [12] The thickness of the nanofiber composite and recast-Nafion membranes were determined to be approximately 53 and 54 µm, respectively, by a film thickness meter (LE-300, Kett Electric Laboratory, Tokyo, Japan). The surface and cross-section of the membranes were observed by SEM. The ion exchange capacity (IEC) value was measured by a back titration using NaCl and NaOH solutions. The tensile test was performed by a precision universal tester (AGS-X5kN, Shimadzu Co., Tokyo, Japan). The membranes were cut into rectangle shapes (10 mm×40 mm) and set on the fixtures with 20 mm gap. The stressstrain tests were performed at a controlled velocity of 1 mm min−1 until breaking under ambient conditions (ca. 25 °C and 50%RH). In-plane proton conductivities of the nanofiber composite and recast-Nafion membranes were measured by an electrochemical impedance spectroscopy (Hioki 3532-50, Tokyo, Japan) over the frequency range from 50 Hz to 500 kHz at the appropriate temperature and relative humidity in a thermo-controlled humidity chamber. Proton conductivity (σ) of the membranes were determined from Eq. (2):

σ (Scm−1) =

d (cm) × R (Ω ) A (cm2 )

PO2 =

Ws−Wd × 100 Wd

qw⋅60⋅60⋅24 k a⋅t

(7)

where q, a, t, ΔP, k, T, L, and qw are the volume of the test gas permeated through the membrane (cm3), the permeation area (cm2), the sampling time (s), the difference in pressure by considering water vapor partial pressure (cmHg), a correction factor of the apparatus (-), the test temperature (°C), the thickness of the membrane (cm), and the weight of the water vapor permeated through the membrane (g), respectively. Inner morphology of the membranes and SPI nanofibers were examined by a transmission electron microscope (TEM). First, the samples were stained with lead ions by an ion exchange reaction of the sulfonic acid groups. The stained samples were embedded in epoxy resin and sectioned to give 100-nm thick membranes by a microtome (Leica EM UC6). The images were taken on a TEM (JEM-3200FS, JEOL, Tokyo, Japan) with an accelerating voltage of 300 kV. Inner morphology of the membranes were also evaluated by a SAXS (Rigaku NANO-Viewer) measurement. The fully hydrated membranes were enveloped in polyethylene bags and were irradiated by X-ray (CuKα, λi=1.54 Å) with 40 kV. The range of scattering vectors explored (q=4πsinθ/λi) was from 0.15 to 2.5 nm−1, where λi and 2θ are the incident wavelength and total scattering angle, respectively.

(2)

3. Results and discussion (3)

3.1. Preparation of a nanofiber composite membrane

where Ws and Wd are the weights of the wet and dry nanofibrous membranes, respectively. Water uptake of the membranes by immersion in water at 25 °C for 24 h were also measured by the similar procedure, except that water on the membrane surface was quickly wiped off after the immersion in water. Size (in-plane length) and thickness changes of the membranes by water uptake were similarly measured by comparing the initial dry state and wet state at different relative humidity. The gas barrier property of the nanofiber composite and recastNafion membranes were measured by two apparatuses. A differential pressure type high vacuum apparatus (Rika Seiki, Inc., K-315-H, Tokyo, Japan) was attempted to discuss the difference of diffusivity and solubility of oxygen in the dry membranes. The gas permeation measurements of the membranes were carried out at 35 °C and 76 cmHg. The gas permeability coefficient (P), diffusion coefficient (D), and solubility coefficient (S) of the membranes were calculated from

P = DS

(4)

L2 6θ

(5)

D=

(6)

Q H2 O =

where R, A, and d are the impedance value, the cross-sectional area of the membrane, and the distance between the electrodes, respectively. Proton conductivity measurement in the through-plane direction was also performed by sandwiching the membranes in the platinum electrodes under ambient temperature and relative humidity (27 °C and 24%RH). Water uptakes of the membranes were gravimetrically measured from the dried and humidified samples. The nanofibrous membranes were dried in a vacuum oven at 80 °C for 10 h. After weighing the membrane at dry state, the samples were put in a thermocontrolled humidity chamber at 80 °C and 90, 60, or 30%RH for 24 h each. After 24 h at each relative humidity, the nanofibrous membranes were quickly weighed. The water uptake was calculated using Eq. (3):

W (%) =

q 1 273 L k a⋅t ΔP T + 273

SPI with high molecular weight (Mw > 500 kDa) and narrow molecular weight distributions (Mw/Mn=2.5) was successfully synthesized according to our previous study [33]. The SPI were easily soluble in polar aprotic solvents, such as DMF, N,N-dimethylacetamide, dimethylsulfoxide, and N-methyl pyrrolidone, and were insoluble in water, alcohol, and other common organic solvents. Such solubility of the SPI was suitable for the fabrication of the nanofiber composite membrane with Nafion dispersion containing water and alcohol. The SPI was attempted to the electrospinning to fabricate nanofibers. After optimizing the electrospinning parameters, uniform nanofibers with the average diameter of 192 ± 13 nm were obtained. The porosity of the SPI nanofibrous membrane with its thickness of ca. 50 µm was estimated to be 90% based on its apparent volume, density, and weight of the SPI nanofibers. Typical surface and cross-sectional SEM images of the SPI nanofibrous membrane and the nanofiber composite membrane were shown in Fig. 1. No visible voids were observed in the cross-sectional SEM image of the nanofiber composite membrane (Fig. 1(d)). The thickness of the nanofiber composite membrane (54 µm) was close to that of the original SPI nanofibrous membrane (ca. 50 µm), meaning the appropriate amount of Nafion matrix was completely filled in the void space of the SPI nanofibrous membrane. The IEC value of the nanofiber composite membrane was determined to be 0.96 meq g−1, which was higher than that of the recast-Nafion membrane (0.86 meq g−1) because the nanofiber composite membrane contains SPI nanofibers with a higher IEC value (1.65 meq g−1) (Table 1). The mechanical strength of the nanofiber composite membrane was evaluated by the tensile test. As shown in Fig. 2, the

using the time lag, θ. Oxygen and helium (as an alternative to hydrogen) were used for the measurements. The gas barrier property and water vapor permeance of wet membranes were measured by another equal pressure type gas permeation measurement apparatus (Round Science Inc., RGP-3000Z), equipped with a gas chromatograph (J-science Lab Co. Ltd. GC7100). Each membrane was placed in the center of the temperature-controlled permeation cell, and the test gas 67

Journal of Membrane Science 530 (2017) 65–72

T. Makinouchi et al.

(a)

(b)

(c)

(d)

Fig. 1. SEM images of the SPI nanofibrous membrane ((a) surface and (b) cross-section) and the nanofiber composite membrane ((c) surface and (d) cross-section). Arrows indicate the thickness of the membranes.

3.2. Proton conductivity and water uptake of the nanofiber composite membrane

Table 1 Compositions of the polymer electrolyte membranes. Membrane

Blend ratio Nafion: SPI nanofiber

Thickness (μm)

IECa (meq g−1)

Elastic modulusb (GPa)

Recast-Nafion membrane Nanofiber composite membrane

100: 0

52.5 ± 4.9

0.86 ± 0.04

0.23 ± 0.02

90: 10

54.0 ± 4.1

0.96 ± 0.03

0.40 ± 0.05

a b

Fig. 3(a) shows the temperature dependence of the in-plane proton conductivity on the nanofiber composite and recast-Nafion membranes at a constant relative humidity (95%RH). The nanofiber composite membrane showed higher proton conductivity than the recast-Nafion membrane at all the range of temperatures between 30 and 90 °C. The activation energy of the proton conductivity on the nanofiber composite membrane (18.0 kJ mol−1) was almost equal to that of the recastNafion membrane (17.1 kJ mol−1). Relative humidity dependence of the proton conductivity on these membranes at 80 °C is depicted in Fig. 3(b). It was demonstrated that the nanofiber composite membrane had superior proton conductivity to the recast-Nafion membrane, especially at low relative humidity. Furthermore, through-plane proton conductivity of the membranes were measured under ambient temperature and relative humidity (27 °C and 24%RH). The nanofiber composite membrane showed 1.02×10−2 S/cm, which was higher than that of the recast-Nafion membrane (8.4×10−3 S/cm) under the same condition. These results indicated that the SPI nanofibers had better proton conductive pathway than the Nafion matrix. In our previous study, we first revealed that the intrinsic proton conductivity of the SPI nanofiber was much higher than that of the corresponding SPI membrane. Although the chemical structure and IEC of the SPI were same in the nanofiber and membrane, the nanofiber formation process by the electrospinning enhanced polymer chain orientation and hydrophilic/hydrophobic phase separation inside the SPI nanofibers, yielding their distinguished proton conductivity [33]. In addition to the contribution of SPI nanofibers, it is possible that the effective proton conductive pathway was constructed in the nanofiber composite membrane. The detail will be discussed later. Water uptake, which is an important property related to the proton conductivity, was measured on the nanofiber composite and recastNafion membranes at various relative humidity (Fig. 4(a)). The nanofiber composite membrane showed similar water uptake compared to the recast-Nafion membrane at 30–90%RH. On the other hand, water uptake of the nanofiber composite membrane under water immersion was lower than that of the recast-Nafion membrane. It is

Ion exchange capacity determined by a titration method. under ambient condition.

Fig. 2. Stress-strain curves of the nanofiber composite and recast-Nafion membranes.

nanofiber composite membrane showed higher maximum stress (19.8 MPa) and lower degree of elongation (15.4%) than the recast Nafion membrane (9.2 MPa and 17.9%, respectively). The elastic modulus of the nanofiber composite membrane was estimated to be 0.40 GPa, which was almost twice higher than that of the recast-Nafion membrane (0.23 GPa). Such distinguished mechanical strength is definitely derived from the mechanically tough three-dimensional network structure composed of SPI nanofibers.

68

Journal of Membrane Science 530 (2017) 65–72

T. Makinouchi et al.

Fig. 3. (a) Temperature and (b) relative humidity dependence of proton conductivities on the nanofiber composite and recast-Nafion membranes.

nated polymer membranes have been widely used for PEMs, their relatively low gas barrier property is a critical issue for long-term fuel cell operation. Table 2 summarizes the oxygen gas permeability measured by the high vacuum apparatus at a dry state. The oxygen gas permeability coefficient (PO2) of the nanofiber composite membrane was lower than that of the recast-Nafion membrane that was mainly derived from the lower diffusivity coefficient (DO2) of the nanofiber composite membrane. Furthermore, as an alternative to hydrogen permeability measurement, helium (He) gas permeability of the membranes were also measured. The He gas permeability coefficient (PHe) of the nanofiber composite membrane was 1.9×10−9 cm3(STP) cm/(cm2 s cmHg), which was lower than that of the recast-Nafion membrane (2.3×10−9 cm3(STP) cm/(cm2 s cmHg)). This result promises that the nanofiber composite membrane has lower hydrogen gas crossover than the recast-Nafion membrane during fuel cell operations. Fig. 5(a) shows relative humidity dependence of PO2 on

considered that the three-dimensional nanofiber network structure inhibited excessive swelling of Nafion matrix to reduce water uptake. A similar tendency was observed on the length and thickness change of the membranes (Fig. 4(b)). Lower water uptake of the nanofiber composite membrane than the recast-Nafion at high relative humidity or in water will promise higher membrane stability, even though its water retention property at low relative humidity was comparable to the recast-Nafion membrane. That is to say, the nanofiber composite membrane possesses desired property on the water uptake. 3.3. Gas and water vapor permeabilities of the nanofiber composite membrane PEMs are required to have low gas permeabilities of the fuel and oxidant to inhibit unexpected gas crossover that will lead degradation of fuel cell components. Although the Nafion and other perfluorosulfo-

Fig. 4. (a) Water uptakes and (b) length and thickness changes of the nanofiber composite and recast-Nafion membranes at various relative humidity at 80 °C.

69

Journal of Membrane Science 530 (2017) 65–72

T. Makinouchi et al.

Table 2 Oxygen permeability coefficient (PO2), diffusivity coefficient (DO2), and solubility coefficient (SO2) of the nanofiber composite and recast-Nafion membranes. Membrane

Blend ratio Nafion: SPI nanofiber

PO2a

DO2b

SO2c

Recast-Nafion membrane Nanofiber composite membrane

100: 0 90: 10

0.91 0.60

2.06 0.52

0.44 1.16

Nafion matrix

SPI nanofiber 20 nm

The gas permeation measurements of the dry membranes were carried out using the differential pressure type high vacuum apparatus at 35 °C and 76 cmHg. a Oxygen permeability coefficient PO2: 10−10(cm3(STP) cm/(cm2 s cmHg)). b Oxygen diffusivity coefficient DO2: 10−8(cm2/s). c Oxygen solubility coefficient SO2: 10−2(cm3(STP)/(cm3 cmHg)).

100 nm

Fig. 6. Cross-sectional TEM images of the nanofiber composite membrane. The sulfonic acid groups in the SPI nanofiber and Nafion matrix were previously stained with lead ions. An arrow in the left enlarged image indicates the acid-condensed layer of Nafion matrix near the surface of the SPI nanofiber.

the nanofiber composite and recast-Nafion membranes. It is obvious that the nanofiber composite membrane possessed much lower gas permeability than the recast-Nafion membrane. These tendency are in good agreement with the previous study on PEMs containing nanofibers [29,30]. There are three possible reasons why the nanofiber composite membrane represented better gas barrier property than the recast-Nafion membrane: First, SPI had lower gas permeability than Nafion because the aromatic hydrocarbon polymers generally have low gas diffusion capability due to their π-stacked structures. Second, the electrospun nanofibers improved gas barrier property based on their stacked structures due to polymer chain orientation along the nanofiber axis. Third, the Nafion matrix around the nanofibers may change its morphology to improve its gas barrier property. In Fig. 5(b), water vapor permeances through the nanofiber composite or recast-Nafion membrane at various relative humidity are plotted. Unlike in the case of gas permeability, water permeability through the nanofiber composite membrane was almost equivalent to that of the recast-Nafion membrane. It is considered that the water diffusion effectively occurred inside SPI nanofibers and/or through some effective water transport channels in the nanofiber composite membrane. Such distinguished water permeability is preferable for PEMs, especially under low relative humidity conditions, because the membrane with high water permeability can effectively utilize backdiffusion of water from the cathode to the anode using electrochemically generated water at the cathode electrode during the fuel cell

operation. The nanofiber composite membrane is a promising PEM that can disturb gas diffusion without inhibiting water diffusion. 3.4. Inner morphological analyses of the nanofiber composite membrane As described above, the nanofiber composite membrane demonstrated higher proton conductivity and lower gas permeability than the recast-Nafion membrane. In the widely-reported related work, composite membranes that decrease gas permeability often sacrifices their proton conductivity at the same time. Therefore, high proton conductive, relatively high water diffusive, and low gas permeable characteristics of the nanofiber composite membrane are unique and valuable for PEFC application. In order to discuss the reasons for the advantageous effect of the SPI nanofibers in the composite membrane, the internal morphology of the membrane was analyzed by the TEM observation and the SAXS measurement along with the recast-Nafion membrane. Fig. 6 represents the cross-sectional TEM image of the nanofiber composite membrane. The sulfonic acid groups in the SPI nanofibers and Nafion matrix were stained with lead ions to be observed black dots. An SPI nanofiber was observed as a circle structure with its diameter of ca. 200 nm, and a contrast of the nanofiber domain was darker than that of the Nafion matrix due to higher IEC of the SPI nanofibers than the Nafion matrix. In the Nafion matrix domain, small black dots with from several to ten nanometers-

Fig. 5. (a) Oxygen gas permeability and (b) water vapor permeance of the nanofiber composite and recast-Nafion membranes at various relative humidity at 80 °C. The gas barrier property and water vapor permeance were measured by the equal pressure type gas permeation measurement apparatus equipped with the gas chromatograph.

70

Journal of Membrane Science 530 (2017) 65–72

T. Makinouchi et al.

31.4

d / nm 12.6 6.28

near the nanofibers. Such acid-condensed layer at the interface of SPI nanofibers and Nafion matrix should have a positive impact on the proton conduction and water diffusion, especially under the low relative humidity conditions. Although the SPI nanofibers and Nafion matrix were selected as typical polymer electrolytes in this study, other polymer electrolyte combination will be applicable to the prospective ion conductive nanofiber system. Now we are studying on the optimal combination of the polymer electrolyte nanofibers and matrices as well as the evaluation of their fuel cell performances.

3.14

Intensity (a.u.)

Nanofiber composite membrane Recast-Nafion membrane

Acknowledgment This work was partially supported by a grant (No. P10001) from NEDO, a grant (Platform for Technology and Industry) from Tokyo Metropolitan Government, and JSPS KAKENHI Grant Numbers 24750225 and 26410225. References

10

-1 -1

q/Å

10

0 [1] W. Vielstich, Handbook of Fuel Cells, Wiley, Chichester, England, 2009. [2] J. Rozière, G.J. Jones, Non-fluorinated polymer materials for proton exchange membrane fuel cells, Annu. Rev. Mater. Res. 33 (2003) 503–555. [3] M.A. Hickner, B.S. Pivovar, The chemical and structural nature of proton exchange membrane fuel cell properties, Fuel Cells 5 (2005) 213–229. [4] H. Zhang, P.K. Shen, Recent development of polymer electrolyte membranes for fuel cells, Chem. Rev. 112 (2012) 2780–2832. [5] J. Marcinkoski, B.D. James, J.A. Kalinoski, W. Podolski, T. Benjamin, J. Kopasz, Manufacturing process assumptions used in fuel cell system cost analyses, J. Power Sources 196 (2011) 5282–5292. [6] M.P. Rodgers, L.J. Bonville, H.R. Kunz, D.K. Slattery, J.M. Fenton, Fuel cell perfluorinated sulfonic acid membrane degradation correlating accelerated stress testing and lifetime, Chem. Rev. 112 (2012) 6075–6103. [7] J. Kamcev, B.D. Freeman, Materials science: cracks help membranes to stay hydrated, Nature 532 (2016) 445–446. [8] Y.M. Lee, Fuel cells: operating flexibly, Nat. Energy 1 (2016) 16136. [9] K.-S. Lee, J.S. Spendelow, Y.-K. Choe, C. Fujimoto, Y.S. Kim, An operationally flexible fuel cell based on quaternary ammonium-biphosphate ion pairs, Nat. Energy 1 (2016) 16120. [10] C.H. Park, S.Y. Lee, D.S. Hwang, D.W. Shin, D.H. Cho, K.H. Lee, T.-W. Kim, T.W. Kim, M. Lee, D.-S. Kim, C.M. Doherty, A.W. Thornton, A.J. Hill, M.D. Guiver, Y.M. Lee, Nature 532 (2016) 480–483. [11] J. Saito, M. Tanaka, K. Miyatake, M. Watanabe, Proton conductive polyimide ionomer membranes: effect of NH, OH, and COOH groups, J. Polym. Sci. A: Polym. Chem. 48 (2010) 2846–2854. [12] K. Yamazaki, G. Wang, M. Tanaka, H. Kawakami, Sulfonated block-graft copolyimide for high proton conductive and low gas permeable polymer electrolyte membrane, J. Power Sources 216 (2012) 387–394. [13] C.F. Kins, E. Sengupta, A. Kaltbeitzel, M. Wagner, I. Lieberwirth, H.W. Spiess, M.R. Hansen, Morphological anisotropy and proton conduction in multiblock copolyimide electrolyte membranes, Macromolecules 47 (2014) 2645–2658. [14] L. Sheng, T. Higashihara, M. Ueda, T. Hayakawa, Poly(arylene ether ether nitrile)s containing flexible alkylsulfonated side chains for polymer electrolyte membranes, J. Polym. Sci. A: Polym. Chem. 52 (2014) 21–29. [15] J. Miyake, T. Mochizuki, K. Miyatake, Effect of the hydrophilic component in aromatic ionomers: simple structure provides improved properties as fuel cell membranes, ACS Macro Lett. 4 (2015) 750–754. [16] S. Lee, J. Ann, H. Lee, J.-H. Kim, C.-S. Kim, T.-H. Yang, B. Bae, Synthesis and characterization of crosslink-free highly sulfonated multi-block poly(arylene ether sulfone) multi-block membranes for fuel cells, J. Mater. Chem. A 3 (2015) 1833–1836. [17] S. Seesukphronrarak, K. Ohira, K. Kidena, N. Takimoto, C.S. Kuroda, A. Ohira, Synthesis and properties of sulfonated copoly(p-phenylene)s containing aliphatic alkyl pendant for fuel cell applications, Polymer 51 (2010) 623–631. [18] Y. Takeoka, K. Umezawa, T. Oshima, M. Yoshida, M. Yoshizawa-Fujita, M. Rikukawa, Synthesis and properties of hydrophilic-hydrophobic diblock copolymer ionomers based on poly(p-phenylene)s, Polym. Chem. 5 (2014) 4132–4140. [19] C. Laberty-Robert, K. Vallé, F. Pereira, C. Sanchez, Design and properties of functional hybrid organic-inorganic membranes for fuel cells, Chem. Soc. Rev. 40 (2011) 961–1005. [20] N.N. Krishnan, D. Henkensmeier, J.H. Jang, H.-J. Kim, Nanocomposite membranes for polymer electrolyte fuel cells, Macromol. Mater. Eng. 299 (2014) 1031–1041. [21] S. Cavaliere, S. Subianto, I. Savych, D.J. Jones, J. Rozière, Electrospinning: designed architectures for energy conversion and storage devices, Energy Environ. Sci. 4 (2011) 4761–4785. [22] M. Tanaka, Development of ion conductive nanofibers for polymer electrolyte fuel cells, Polym. J. 48 (2016) 51–58. [23] D.H. Reneker, A.L. Yarin, Electrospinning jets and polymer nanofibers, Polymer 49 (2008) 2387–2425. [24] H. Matsumoto, A. Tanioka, Functionality in electrospun nanofibrous membranes

Fig. 7. SAXS profile of the nanofiber composite and recast-Nafion membranes.

diameter represent the hydrophilic clusters of the Nafion. By taking a closer look at the interface of the SPI nanofibers and Nafion matrix (Fig. 6, left), slightly darker area can be seen around the nanofibers with its width of around ten nanometers. This image suggests that the sulfonic acid groups in the Nafion matrix may be attracted to the SPI nanofibers during the membrane formation process to construct acidcondensed layer near the nanofibers. Such acid-condensed layer at the interface of SPI nanofibers and Nafion matrix can be a pathway for effective proton conduction and water diffusion even in the low relative humidity conditions. The SAXS measurement was also attempted to evaluate the internal morphology of the fully-hydrated nanofiber composite membrane. Fig. 7 shows the SAXS profiles of the nanofiber composite and recast-Nafion membranes. In the case of the nanofiber composite membrane, the peak top at q=1.10 Å−1 corresponds to the scattering between the hydrophilic domains. The distance between the hydrophilic clusters was calculated to be 5.71 nm, which was slightly longer than that of the recast-Nafion membrane (5.61 nm, q=1.12 Å−1). It is generally known that the longer distance between hydrophilic clusters will provoke lower proton conductivity due to the poor connectivity of the proton channels [36]. This result supports that the SPI nanofibers and acid-condensed layer around the nanofibers contributed higher proton conductivity than the Nafion matrix domains in the nanofiber composite membrane.

4. Conclusion A novel electrospun SPI nanofiber-based Nafion composite membrane was developed and characterized for polymer electrolyte fuel cells. The nanofiber composite membrane showed higher proton conductivity, better mechanical stability, and lower gas permeability than the recast-Nafion membrane, in which the improvement of these characteristics has been assumed difficult to accomplish at the same time for long years. In addition, in spite of such the distinguished gas barrier property, the nanofiber composite membrane indicated good water permeability comparable to the recast-Nafion membrane. Besides the contribution of the SPI nanofibers that had been revealed to possess ultra-high proton conductivity in our previous study, the Nafion matrix indicated unique morphology in the nanofiber composite membrane. The TEM observation suggested that the sulfonic acid groups in the Nafion matrix were attracted to the SPI nanofibers during the membrane formation process to construct acid-condensed layer 71

Journal of Membrane Science 530 (2017) 65–72

T. Makinouchi et al.

[25]

[26]

[27]

[28]

[29]

based on fiber's size, surface area, and molecular orientation, Membranes 1 (2011) 249–264. T. Arai, M. Tanaka, H. Kawakami, Porphyrin-containing electrospun nanofibers: positional control of porphyrin molecules in nanofibers and their catalytic application, ACS Appl. Mater. Interfaces 4 (2012) 5453–5457. Y. Karube, H. Kawakami, Fabrication of well-aligned electrospun nanofibrous membrane based on fluorinated polyimide, Polym. Adv. Technol. 21 (2010) 861–866. K. Sode, M. Tanaka, Y. Suzuki, H. Kawakami, Nickel nanoparticle chains inside carbonized polymer nanofibers: preparation by electrospinning and ion beam irradiation, Nanoscale 5 (2013) 8235–8241. K. Sode, T. Sato, M. Tanaka, Y. Suzuki, H. Kawakami, Carbon nanofibers prepared from electrospun polyimide, polysulfone and polyacrylonitrile nanofibers by ionbeam irradiation, Polym. J. 45 (2013) 1210–1215. T. Tamura, H. Kawakami, Aligned electrospun nanofiber composite membranes for fuel cell electrolytes, Nano Lett. 10 (2010) 1324–1328.

[30] R. Takemori, H. Kawakami, Electrospun nanofibrous blend membranes for fuel cell electrolytes, J. Power Sources 195 (2010) 5957–5961. [31] T. Tamura, R. Takemori, H. Kawakami, Proton conductive properties of composite membranes containing uniaxially aligned ultrafine electrospun polyimide nanofiber, J. Power Sources 217 (2012) 135–141. [32] M. Tanaka, Ion conductive polymer nanofibers, Sen.'i Gakkaishi 69 (2013) p57–p62. [33] R. Takemori, G. Ito, M. Tanaka, H. Kawakami, Ultra-high proton conduction in electrospun sulfonated polyimide nanofibers, RSC Adv. 4 (2014) 20005–20009. [34] B. Dong, L. Gwee, D. Salas-dela Cruz, K.I. Winey, Y.A. Elabd, Super proton conductive high-purity Nafion nanofibers, Nano Lett. 10 (2010) 3785–3790. [35] T.A. Zawodzinski Jr, C. Derouin, S. Radzinski, R.J. Sherman, V.T. Smith, T.E. Springer, S. Gottesfeld, Water uptake by and transport through Nafion® 117 membranes, J. Electrochem. Soc. 140 (1993) 1041–1047. [36] K. Schmidt-Rohr, Q. Chen, Parallel cylindrical water nanochannels in Nafion fuelcell membranes, Nat. Mater. 7 (2008) 75–83.

72