Composite Membranes Using Hydrophilized Porous Substrates for Hydrogen Based Energy Conversion

: Poly(tetraﬂuoroethylene) (PTFE) porous substrate-reinforced composite membranes for energy conversion technologies are prepared and characterized. In particular, we develop a new hydrophilic treatment method by in-situ biomimetic siliciﬁcation for PTFE substrates having high porosity (60–80%) since it is di ﬃ cult to impregnate ionomer into strongly hydrophobic PTFE porous substrates for the preparation of composite membranes. The thinner substrate having ~5 µ m treated by the gallic acid / (3-trimethoxysilylpropyl)diethylenetriamine solution with the incubation time of 30 min shows the best hydrophilic treatment result in terms of contact angle. In addition, the composite membranes using the porous substrates show the highest proton conductivity and the lowest water uptake and swelling ratio. Membrane-electrode assemblies (MEAs) using the composite membranes (thinner and lower proton conductivity) and Naﬁon 212 (thicker and higher proton conductivity), which have similar areal resistance, are compared in I–V polarization curves. The I–V polarization curves of two MEAs in activation and Ohmic region are very identical. However, higher mass transport limitation is observed for Naﬁon 212 since the composite membrane with less thickness than Naﬁon 212 would result in higher back di ﬀ usion of water and mitigate cathode ﬂooding.


Introduction
The unstable crude oil prices and global warming caused by greenhouse gases drive us to use alternative energy sources. Companies and governments have made substantial investments in new and renewable energy over the past few years [1]. Among the new and renewable energy, hydrogen had begun to be translated into the alternative energy source area due to easy deployment into electricity infrastructure, diversified energy sources to produce hydrogen from fossil fuels to biomass, improvement of local air pollution, and recent matured technologies such as fuel cells and electrolyzers [2,3]. Using the electrochemistry-driven energy conversion technologies, hydrogen could be produced from water by using electricity and could be consumed to convert into the water along with generating electricity (H 2 + O 2 ↔ H 2 O) [4][5][6].
Fuel cells are the most promising technology utilizing hydrogen. Proton exchange membrane fuel cells (PEMFCs) using hydrogen as fuel show high efficiency during the direct conversion of chemical energy to electric energy. In addition, there is no greenhouse gas emission when hydrogen is produced by water electrolysis using electricity supplied from renewable energy sources such as wind, solar, biomass, etc. [1,7]. Nevertheless, the installation cost is still higher than conventional energy conversion technology such as internal combustion engine and the technical level of durability and reliability must be raised to a higher level to enter the full-fledged fuel cell market [8][9][10]. Membrane-electrodeassembly (MEA) is a key component in PEMFCs, which consists of a piece of proton exchange membrane (PEM) sandwiched between two catalyst layers as an electrode [11]. substrates and proton conductivity of composite membranes. Afterward, fuel cell performance of composite membranes is measured and discussed to investigate the effect of the biomimetic coating materials on the properties of porous PTFE-reinforced composite membranes.
Energies 2020, 13, x FOR PEER REVIEW 3 of 14 substrates and proton conductivity of composite membranes. Afterward, fuel cell performance of composite membranes is measured and discussed to investigate the effect of the biomimetic coating materials on the properties of porous PTFE-reinforced composite membranes.

Materials
Aeos TM ePTFE and PTU0214210 PTFE substrates were purchased from Zeus, the USA, and Sterlitech, USA, respectively. The main physical properties are summarized in Table 1. The 3,4,5trihydroxybenzoic acid (GA), APTES, TMPEDA, and TMPDETA were obtained from Sigma Aldrich. Their chemical structures are shown in Table 2. For pH 8.5 buffer, trizma  hydrochloride (Tris-HCl) was purchased from Dongin Biotech Co., Ltd., South Korea. All the chemicals were used as received without further purification. composite membranes is measured and discussed to investigate the effect of the biomimetic coating materials on the properties of porous PTFE-reinforced composite membranes.

Materials
Aeos TM ePTFE and PTU0214210 PTFE substrates were purchased from Zeus, the USA, and Sterlitech, USA, respectively. The main physical properties are summarized in Table 1. The 3,4,5trihydroxybenzoic acid (GA), APTES, TMPEDA, and TMPDETA were obtained from Sigma Aldrich.  Their chemical structures are shown in Table 2. For pH 8.5 buffer, trizma  hydrochloride (Tris-HCl) was purchased from Dongin Biotech Co., Ltd., South Korea. All the chemicals were used as received without further purification. materials on the properties of porous PTFE-reinforced composite membranes.

Hydrophilic Coating of GA/ATS on PTFE Substrates
PTFE substrates were immersed in acetone and ethanol for 6 h, respectively, and then dried in the air prior to hydrophilic treatment. GA of 0.2 g was dissolved in 10 mM Tris-HCl of 100 mL for a GA solution, and three different ATS solutions of 0.13 M using APTES (coded as #1), TMPEDA (coded as #2), and TMPDETA (coded as #3) were prepared in ethanol for ATS solutions. GA/ATS solutions were prepared by mixing 100 mL of the GA solution and 20 mL of the corresponding ATS solution. For the coating of hydrophilic materials, PTFE substrates were immersed in zipper bags filled by the corresponding solutions placed on a plate orbital shaker and were incubated at room temperature for a certain time period under shaking, followed by rinsing with distilled water and drying in the air. The abbreviated incubation conditions are summarized in Table 3.

Preparation of Porous PTFE-Reinforced Composite Membranes
Reinforced composite membranes were prepared by hydrophilized PTFE substrates and D1021 Nafion™ dispersion (EW 1100, 10 wt.%, Chemours). As shown in Figure 1, the hydrophilically treated porous substrate provides better wettability for the Nafion™ dispersion. However, the untreated pristine one completely repels the dispersion of the substrate surface. The procedure consists of six steps as follows: (1) the PTFE substrates were immersed into the dispersion for 5 min; (2) the membranes were placed on a glass plate doctor-bladed by a film applicator with a thickness of 10 µm; (3) the membranes were dried in a convective oven at 70 • C for 10 min; (4) another 10 µm-layer Nafion was coated on another side of the membranes; (5) the membranes were dried again in a vacuum oven at 70 • C for 6 h; (6) finally the membranes were annealed in a convective oven at 190 • C for 12 min. The scheme of the preparation is illustrated in Figure 2. All the composite membranes were boiled in a 0.5 M H 2 SO 4 solution for 6 h for protonation and boiled in distilled water for 6 h for removal of residual acid prior to use.
were prepared by mixing 100 mL of the GA solution and 20 mL of the corresponding ATS solution. For the coating of hydrophilic materials, PTFE substrates were immersed in zipper bags filled by the corresponding solutions placed on a plate orbital shaker and were incubated at room temperature for a certain time period under shaking, followed by rinsing with distilled water and drying in the air. The abbreviated incubation conditions are summarized in Table 3.

Preparation of Porous PTFE-Reinforced Composite Membranes
Reinforced composite membranes were prepared by hydrophilized PTFE substrates and D1021 Nafion TM dispersion (EW 1100, 10 wt.%, Chemours). As shown in Figure 1, the hydrophilically treated porous substrate provides better wettability for the Nafion TM dispersion. However, the untreated pristine one completely repels the dispersion of the substrate surface. The procedure consists of six steps as follows: (1) the PTFE substrates were immersed into the dispersion for 5 min; (2) the membranes were placed on a glass plate doctor-bladed by a film applicator with a thickness of 10 m; (3) the membranes were dried in a convective oven at 70 °C for 10 min; (4) another 10 m-layer Nafion was coated on another side of the membranes; (5) the membranes were dried again in a vacuum oven at 70 °C for 6 h; (6) finally the membranes were annealed in a convective oven at 190 °C for 12 min. The scheme of the preparation is illustrated in Figure 2. All the composite membranes were boiled in a 0.5 M H2SO4 solution for 6 h for protonation and boiled in distilled water for 6 h for removal of residual acid prior to use.

Characterization of Material Structure and GA/ATS Solutions
Measurement of contact angle of PTFE substrates with and without hydrophilic material coating was carried out using a contact angle meter (Attension  Theta Lite, Bioline Scientific, Sweden) as soon as a droplet of the Nafion dispersion falls on substrates. Microcosmic morphology of the

Characterization of Material Structure and GA/ATS Solutions
Measurement of contact angle of PTFE substrates with and without hydrophilic material coating was carried out using a contact angle meter (Attension ® Theta Lite, Bioline Scientific, Sweden) as soon as a droplet of the Nafion dispersion falls on substrates. Microcosmic morphology of the substrates and the composite membranes were obtained using a field emission scanning electron microscopy (FE-SEM) system (ZEISS Sigma 500, Germany). Attenuated total reflectance-Fourier transformed infrared spectroscopy (ATR-FTIR) characterization of samples was obtained by a JASCO FT-IR 4700 spectrometer (USA) with an ATR accessory containing a Ge crystal with a wavenumber resolution of 4 cm −1 and range of 600-4000 cm −1 . The microscopic state of GA/ATS solutions was detected by dynamic light scattering (DLS) (ELSZ-1000, Otsuka Electronics Co., Ltd., Japan).

Characterization of Porous PTFE-Reinforced Composite Membranes
Water uptake of composite membranes was measured by immersing the membranes into distilled water at room temperature for 12 h after the dry membranes (W dry ) were weighed. Afterward, the membranes were taken out, surface water was wiped out, and the wet weight (W wet ) was then measured. Finally, the water uptake of composite membranes was calculated by using the following equation [37]: The swelling ratio is calculated by the following equation: where L dry and L wet are the dimension (length, width, and thickness) of composite membranes. The ion exchange capacity (IEC) of composite membranes is firstly soaked in 1.0 M H 2 SO 4 for 24 h to replace functional groups with a proton. The excess acid solution on the surface of the membrane was thoroughly washed off with distilled water, and the membranes were immersed in 1.0 M NaCl (precisely 20 mL) for 24 h. The amount of proton ion-exchanged with Na + was measured by a titration method with 0.01 N NaOH by an auto-titrator (848 Titrino plus, Metrohm, Switzerland). The IEC was calculated by the following equation: where n is the number of electrons gain or lost, M is the molar concentration of the titration solution (mmol/mL), V is the consumed volume of the titration solution (mL) and m is the weight of the dry membrane [38]. The transport number of composite membranes was measured by using the electromotive force (emf ) method which is a method of estimating the transport number of conducting species in ionic and mixed conductors in a two-compartment cell with a composite membrane as a diaphragm in two separate solutions having different concentrations. First of all, composite membranes with a size of 2 × 2 cm 2 were immersed in 0.001 M NaCl for 24 h, were taken out, and were installed between two compartment cells which were filled with 0.001 and 0.005 M NaCl on both sides of the cells. Air bubbles on the membrane surfaces were completely removed. Finally, two Luggin capillaries with Ag/AgCl wires filled in a saturated KCl solution were mounted to the nearest locations to the installed membrane. Two wires exposed out of each Luggin capillary reference electrode were then connected to Energies 2020, 13, 6101 6 of 14 a digital voltage meter (34401A, Agilent, USA) to record voltage between two electrodes. The transport number of composite membranes was calculated by the following equation: where, E m is the measured membrane potential, T is the temperature of the solution (K), F is the Faraday constant (96485 C eq −1 ), R is the ideal gas constant (8.3145 J mol −1 K −1 ), and C 1 is the lower electrolyte concentration, C 2 is the higher electrolyte concentration [38].
For the determination of proton conductivity of composite membranes, the electrical impedance of composite membranes was measured. The membranes were cut into a piece with 2 × 2 cm 2 and then were impregnated with 1 M H 2 SO 4 for 12 h to replace functional groups into a proton. A piece of a sample was washed with distilled water and was placed in a four-electrode in-plane conductivity cell. The in-plane cell was immersed in distilled water, and the impedance value was measured at a frequency range from 10 6 to 10 −3 Hz using a potentiostat/galvanostat with a frequency response analyzer (SP-150, BioLogic, France). Proton conductivity of composite membranes is calculated by the following equation: where, L is the thickness of a membrane, R is the impedance of a membrane at zero phase angle and A is the cross-sectional area of a membrane.

Fuel Cell Performance Using Porous PTFE-Reinforced Composite Membranes
The membrane-electrode assemblies (MEAs) were made by a spraying technique using a commercially available Pt-based electrocatalyst (Alfa Aesar HiSPEC™ 4000, USA) with the catalyst loading of 0.4 mg cm −2 and Nafion ® 212 (Chemours) or a lab-made composite membrane. MEAs prepared in this study were mounted in a unit cell with an electrode active area of 9 cm 2 and were evaluated in a test station (CNL Energy Co. Ltd., Republic of Korea). Hydrogen crossover was measured by using a test station (CNL Energy Co. Ltd., Republic of Korea) at 70 • C, with H 2 /N 2 flows fixed at 0.2/0.2 L min −1 and 2.5 bar absolute pressure in the anode and cathode compartment. The cell operation conditions were in 100% relative humidity for both anode and cathode and the cell temperature of 70 • C. The fuel and oxidant were hydrogen and air to the anode and cathode with a stoichiometry ratio of 1.2 and 2.0, respectively.

Characterization of GA/ATS Solutions
Surface coating of adhesive hydrophilic materials is based on the synthesis of polysiloxane from GA/ATS, which is caused by adhesive deposition of the mussel-inspired coating [35]. To confirm in-situ biomimetic silicification, the reactions between GA with the catechol hydroxyl moiety and ATS with the amino moieties have been investigated. Firstly, the visual observation of GA/ATS solutions with incubation time was carried out. Figure 3 shows the co-deposition of GA and ATS. Color and transparency of the GA/#1 change and the co-deposit particles in the GA/#1 solution are not clearly formed. It is, however, observed that the GA/#2 and GA/#3 solutions successfully formed the co-deposition. Initially, both solutions are transparent and yellow. As the incubation time increases, the color of the solutions becomes opaque and dark brown by visual observation. This is also confirmed by the laser transmission of the synthetic solutions in Figure 3. In all cases, laser light penetrates clear solutions immediately after mixing GA and #2 but does not penetrate any more after 120, 30, and 30 min for the GA/#1, GA/#2, and GA/#3, respectively. In addition, the GA/#3 shows more scattered laser light penetration at the beginning compared to the GA/#1 and GA/#2. It infers that the homogeneous reaction is faster than the others. In the case of the GA/#3, after the incubation time of Energies 2020, 13, 6101 7 of 14 120 min, laser light penetrates again since the synthetic material begins to grow massively and to sink, finally permitting laser light to transmit slightly again through the supernatant.
deposition. Initially, both solutions are transparent and yellow. As the incubation time increases, the color of the solutions becomes opaque and dark brown by visual observation. This is also confirmed by the laser transmission of the synthetic solutions in Figure 3. In all cases, laser light penetrates clear solutions immediately after mixing GA and #2 but does not penetrate any more after 120, 30, and 30 min for the GA/#1, GA/#2, and GA/#3, respectively. In addition, the GA/#3 shows more scattered laser light penetration at the beginning compared to the GA/#1 and GA/#2. It infers that the homogeneous reaction is faster than the others. In the case of the GA/#3, after the incubation time of 120 min, laser light penetrates again since the synthetic material begins to grow massively and to sink, finally permitting laser light to transmit slightly again through the supernatant. The hydrophilic treatment of the porous PTFE-substrates could be attained by the self-coating of the adhesive polysiloxane on the surface of the two kinds of substrates when the substrates are immersed in the GA/ATS/alkaline buffer solutions. The self-coating results in a decrease in the contact angle of the surface of the substrates in a solid-water-air system. Normally, the surface is hydrophobic when the contact angle larger than 90° and hydrophilic when the contact angle is less than 90°. As shown in Figure 4, the treatment of GA and #1 results in no decrease in contact angle as the incubation time increases, which means that the surface of both substrates was not treated to be hydrophilic. It was reported that GA and #1 in alkaline buffer successfully deposited hydrophilic polysiloxane on MF and UF filters [33]. However, no significantly hydrophilic coating was attained for the substrates with higher porosity than MF and UF. In contrast, #2 and #3 with GA in alkaline buffer successfully achieved the hydrophilic coating on the substrates. Thus, the GA/#1 was excluded for further hydrophilic treatment. Figure 5 shows the change in the contact angle of the surface of two substrates treated by GA/#2 and GA/#3 with incubation time. Hydrophilicity in all the cases increased with the incubation time. The minimum incubation time for the increase was 30 min for both GA/ATS solutions. It was observed that there was no change in hydrophilicity prior to 30 min. A decrease rate in the contact angle for the thinner substrate (Zeus) was greater than the thicker substrate (Sterlitech). There is no distinct difference in hydrophilic coating for the thicker substrate with the type of the solutions, but better hydrophilic treatment by the GA/#3 for the thinner substrate is observed than that by the GA/#2. In addition, the contact angle of the GA/#3 treated thinner substrate only becomes less than 90° with an incubation time of 30 min. As a result, the combination of the thinner substrate having ~5 m and the GA/#3 solution is effective for hydrophilic coating by in-situ biomimetic silicification. In particular, the ATS with the longer amino moieties results in better The hydrophilic treatment of the porous PTFE-substrates could be attained by the self-coating of the adhesive polysiloxane on the surface of the two kinds of substrates when the substrates are immersed in the GA/ATS/alkaline buffer solutions. The self-coating results in a decrease in the contact angle of the surface of the substrates in a solid-water-air system. Normally, the surface is hydrophobic when the contact angle larger than 90 • and hydrophilic when the contact angle is less than 90 • . As shown in Figure 4, the treatment of GA and #1 results in no decrease in contact angle as the incubation time increases, which means that the surface of both substrates was not treated to be hydrophilic. It was reported that GA and #1 in alkaline buffer successfully deposited hydrophilic polysiloxane on MF and UF filters [33]. However, no significantly hydrophilic coating was attained for the substrates with higher porosity than MF and UF. In contrast, #2 and #3 with GA in alkaline buffer successfully achieved the hydrophilic coating on the substrates. Thus, the GA/#1 was excluded for further hydrophilic treatment. Figure 5 shows the change in the contact angle of the surface of two substrates treated by GA/#2 and GA/#3 with incubation time. Hydrophilicity in all the cases increased with the incubation time. The minimum incubation time for the increase was 30 min for both GA/ATS solutions. It was observed that there was no change in hydrophilicity prior to 30 min. A decrease rate in the contact angle for the thinner substrate (Zeus) was greater than the thicker substrate (Sterlitech). There is no distinct difference in hydrophilic coating for the thicker substrate with the type of the solutions, but better hydrophilic treatment by the GA/#3 for the thinner substrate is observed than that by the GA/#2. In addition, the contact angle of the GA/#3 treated thinner substrate only becomes less than 90 • with an incubation time of 30 min. As a result, the combination of the thinner substrate having 5 µm and the GA/#3 solution is effective for hydrophilic coating by in-situ biomimetic silicification. In particular, the ATS with the longer amino moieties results in better Michael addition and Schiff base reactions with the catechol hydroxyl moiety of GA. In addition, according to the results of the size distribution of co-deposits formed in the GA/#2 and GA/#3 solutions (see Figure 6 and Table 4), the GA/#3 forms narrower and smaller particles than the GA/#2 as the incubation time increases. Hence, the porous substrate immersed in the GA/#3 results in better hydrophilic treatment. It could be concluded that the hydrophilic treatment of the Zeus porous PTFE substrates in the GA/#3 solution for 30 min incubation time is optimal.

of 14
Michael addition and Schiff base reactions with the catechol hydroxyl moiety of GA. In addition, according to the results of the size distribution of co-deposits formed in the GA/#2 and GA/#3 solutions (see Figure 6 and Table 4), the GA/#3 forms narrower and smaller particles than the GA/#2 as the incubation time increases. Hence, the porous substrate immersed in the GA/#3 results in better hydrophilic treatment. It could be concluded that the hydrophilic treatment of the Zeus porous PTFE substrates in the GA/#3 solution for 30 min incubation time is optimal.   solutions (see Figure 6 and Table 4), the GA/#3 forms narrower and smaller particles than the GA/#2 as the incubation time increases. Hence, the porous substrate immersed in the GA/#3 results in better hydrophilic treatment. It could be concluded that the hydrophilic treatment of the Zeus porous PTFE substrates in the GA/#3 solution for 30 min incubation time is optimal.    As shown in Figure 7, the microscopic surface morphology of porous substrates shows an obvious coating layer after simple immersion in the GA/#2 and GA/#3 solutions in the corresponding SEM images. However, after the incubation time of 120 min, the coating material covers most of the surface of porous substrates. It could cause improper impregnation of PFSA ionomers into porous substrates due to the blocking of the surface pores of substrates by the covered hydrophilic material. As a result, Zeus porous substrates immersed in the GA/#3 show a significant decrease in contact angle less than 90 • at the incubation time of 30 min due to the formation of smaller co-deposits.  Table 4. Average sizes of the particles formed in the GA/#2 and GA/#3 with incubation time.

Time (min)
Average Particle Sizes (nm) GA/#2 GA/ #3  0  185  205  30  1142  1050  60  1264  1377  90  1938  1781  120 3023 1810 As shown in Figure 7, the microscopic surface morphology of porous substrates shows an obvious coating layer after simple immersion in the GA/#2 and GA/#3 solutions in the corresponding SEM images. However, after the incubation time of 120 min, the coating material covers most of the surface of porous substrates. It could cause improper impregnation of PFSA ionomers into porous substr

Characterization of GA/ATS-Treated Porous PTFE Substrate Based Composite Membranes
As determined earlier, the GA/#3 and Zeus porous substrates are only used for the preparation of composite membranes. Table 5 summarizes the results of the physical characterization of the composite membranes using GA/#3 treated Zeus porous substrates. All data in Table 5 are the average values of five samples with less than 10% standard deviation. The thickness of the composite membranes increased as the incubation time increased in all composite membranes. When the hydrophilically treated porous substrates were immersed in PFSA dispersion in the first step of the preparation of composite membranes, the porous substrates with higher incubation time absorbed

Characterization of GA/ATS-Treated Porous PTFE Substrate Based Composite Membranes
As determined earlier, the GA/#3 and Zeus porous substrates are only used for the preparation of composite membranes. Table 5 summarizes the results of the physical characterization of the composite membranes using GA/#3 treated Zeus porous substrates. All data in Table 5 are the average values of five samples with less than 10% standard deviation. The thickness of the composite membranes increased as the incubation time increased in all composite membranes. When the hydrophilically treated porous substrates were immersed in PFSA dispersion in the first step of the preparation of composite membranes, the porous substrates with higher incubation time absorbed more PFSA dispersion. It results in an increase in thickness for composite membranes with higher incubation time in the GA/#3. Thick composite membranes have thick skin layers on both sides of substrates since the thickness of substrates is similar. Accordingly, the water uptake of the composite membranes also increases as the incubation time increases. The main reason for using composite membranes in fuel cell applications is the suppression of swelling ratio in the area since the electrodes coated on polymeric electrolyte membranes could be cracked due to continuous areal expansion and contraction of membranes by hydration and dehydration during on and off cyclic operation of fuel cells [39]. Due to the skeleton effect of porous substrates, the variation of swelling ratio in the area is lower than that in thickness. Similarly, both swelling ratios also increase with the incubation time.  Table 6 summarizes the results of the characterization of the composite membranes using GA/#3 treated Zeus porous substrates. All data in Table 6 are the average values of five samples with less than 6% standard deviation. As discussed earlier, the higher amount of hydrophilic co-deposits is coated on the substrate as the incubation time increases. Consequently, the inactive volume for proton transport increases, leading to a decrease in the ion-conducting fraction of the composite membranes which was comprised of pore-impregnated PFSA ionomer. Hence, it is observed in Table 6 that the ion conductivity of the composite membranes using the porous substrates with higher incubation time decreases. Similarly, the ion exchange capacity of the composite membranes also decreases with incubation time since the equivalent weight of the sulfonic acid group of PFSA ionomer in dry weight of composite membranes decreases. In addition, the transport number representing the ability of permselectivity of proton for the composite membranes using porous substrates treated by the GA/#3 for the incubation time greater than 120 min decreases substantially. It means that the composite membranes still have pores to allow to penetrate co-ions. It might be due to the coverage of hydrophilic co-deposits on the surface of the substrate or the filling of those into the substrate. The co-deposits could not provide permselectivity and result in a decrease in transport number. Among the composite membranes, the GA/#3_30-treated composite membrane is chosen to compare with Nafion 212 as reference. As summarized in Table 7, the composite membrane has a very similar areal resistance to Nafion 212 even though the proton conductivity of the composite membrane is less than that of Nafion 212 due to thickness difference. It is presumed that they could lead to very similar fuel cell performance. Figure 8 shows I-V polarization curves of the MEAs using Nafion 212 and the composite membrane. First of all, it was found that the crossover of hydrogen in the MEA using the GA/#3_30-treated composite membrane exhibited approximately 0.5 mA cm −2 which confirms no defect composite membrane. As expected, the I-V polarization curves of two MEAs in activation and Ohmic region are very identical, but higher mass transport limitation is observed for Nafion 212. The main reason to show higher mass transport voltage loss at high current densities is due to water flooding at cathode since a water forming oxygen reduction reaction occurs as well as proton transport causes electro-osmotic drag from the anode to the cathode [40]. Under the same conditions except for the type of electrolyte membranes in the I-V polarization curves, electrolyte membranes could substantially affect the cathode flooding. The composite membrane facilitates back diffusion of water to prevail on electro-osmotic drag leading to the net water transport toward the anode [41]. Diffusion is driven by a gradient in concentration over a moving distance. Hence, the gradient becomes higher if the moving distance, i.e., the thickness of membranes, is shorter, leading to an increase in diffusion flux. In the same way, the composite membrane with less thickness than Nafion 212 would result in higher back diffusion of water and mitigate cathode flooding. Finally, it leads to lessening the mass transport voltage loss. Table 7. Physical properties of Nafion 212 and the composite membrane using the GA/#3_30 and the summary of the I-V polarization curves of the membrane-electrode assemblies using Nafion 212 and the composite membrane. lead to very similar fuel cell performance. Figure 8 shows I-V polarization curves of the MEAs using Nafion 212 and the composite membrane. First of all, it was found that the crossover of hydrogen in the MEA using the GA/#3_30-treated composite membrane exhibited approximately 0.5 mA cm -2 which confirms no defect composite membrane. As expected, the I-V polarization curves of two MEAs in activation and Ohmic region are very identical, but higher mass transport limitation is observed for Nafion 212. The main reason to show higher mass transport voltage loss at high current densities is due to water flooding at cathode since a water forming oxygen reduction reaction occurs as well as proton transport causes electro-osmotic drag from the anode to the cathode [40]. Under the same conditions except for the type of electrolyte membranes in the I-V polarization curves, electrolyte membranes could substantially affect the cathode flooding. The composite membrane facilitates back diffusion of water to prevail on electro-osmotic drag leading to the net water transport toward the anode [41]. Diffusion is driven by a gradient in concentration over a moving distance. Hence, the gradient becomes higher if the moving distance, i.e., the thickness of membranes, is shorter, leading to an increase in diffusion flux. In the same way, the composite membrane with less thickness than Nafion 212 would result in higher back diffusion of water and mitigate cathode flooding. Finally, it leads to lessening the mass transport voltage loss. Table 7. Physical properties of Nafion 212 and the composite membrane using the GA/#3_30 and the summary of the I-V polarization curves of the membrane-electrode assemblies using Nafion 212 and the composite membrane.

Conclusions
In this study, we have developed a new hydrophilic treatment method for PTFE porous substrates to impregnate PFSA ionomers into strongly hydrophobic porous substrates. Two different PFTE substrates were used: the one has the thickness of ~5 m and the porosity of ~80% and another has the thickness of ~25 m and the porosity of ~60%. For hydrophilic treatment, we have used insitu biomimetic silicification which the pyrogallol moiety in gallic acid (GA) with amino-terminated Figure 8. I-V polarization curves of the membrane-electrode assemblies using Nafion 212 (black) and the composite membrane (red) using the GA/#3_30-treated porous substrate.

Conclusions
In this study, we have developed a new hydrophilic treatment method for PTFE porous substrates to impregnate PFSA ionomers into strongly hydrophobic porous substrates. Two different PFTE substrates were used: the one has the thickness of~5 µm and the porosity of~80% and another has the thickness of~25 µm and the porosity of~60%. For hydrophilic treatment, we have used in-situ biomimetic silicification which the pyrogallol moiety in gallic acid (GA) with amino-terminated substances (ATS) such as siloxane generated a similar mussel-inspired adhesive coating via Michael addition/Schiff base reactions in alkaline conditions. We investigated three different ATS materials, i.e., 3-aminopropyltriethoxysilane (APTES), N- [3-(trimethoxysilyl)propyl]ethylenediamine (TMPEDA), and (3-trimethoxysilylpropyl)diethylenetriamine (TMPDETA) with GA. It was found that GA/APTES showed no hydrophilic treatment on both substrates with higher porosity than microfiltration or ultrafiltration membranes. On the other hand, GA/TMPEDA and GA/TMPDETA using ATS with the longer amino moieties showed effective hydrophilic treatment on both substrates. However, GA/TMPDETA has obtained the best contact angle result at less incubation time for the thinner substrates. The thinner substrate having~5 µm treated by the GA/TMPDETA solution with the incubation time of 30 min shows the best hydrophilic treatment result. In addition, the composite membranes using the porous substrates show the highest proton conductivity and the lowest water uptake and swelling ratio. MEAs using the composite membranes (thinner and lower proton conductivity) and Nafion 212 (thicker and higher proton conductivity), which have similar areal resistance, are compared in I-V polarization curves. The I-V polarization curves of two MEAs in activation and Ohmic region are very identical. However, higher mass transport limitation is observed for Nafion 212 since the composite membrane with less thickness than Nafion 212 would result in higher back diffusion of water and mitigate cathode flooding. It could be concluded that the composite membrane would be advantageous in proton exchange membrane fuel cell application since it has similar areal resistance to Nafion 212 to obtain similar Ohmic loss and less thickness than Nafion 212 to allow higher water flux from cathode to anode to obtain lower mass transport loss in I-V polarization curves.