Proton exchange membrane fuel cells (PEMFC) are devices that convert chemical energy into electricity through electrochemical reaction. PEMFCs are efficient, with fast startup, and produce no emission. They have been applied in vehicles, power stations, uninterruptible power system, and aircraft.
To fully commercialize such fuel cells, their durability must be prolonged, their performance improved, and their cost reduced [1]. The membrane electrode assembly (MEA) performs the core functions of a PEMFC. The MEA consists of two catalytic electrodes and gas diffusion layers (GDLs) on each side of a proton exchange membrane, respectively, for the anode and cathode. Each component of the MEA is optimized to improve the performance, durability, and robustness of the overall PEMFC [1, 2, 3, 4, 5]. The GDL sandwiched between the catalyst layer and gas flow channel is an important component concerning the water management of the MEA.
Liquid water in MEAs arises from condensed humidified vapor and as a product of oxygen reduction, some of which is necessary for maintaining the ionic conductivity of the membrane and catalyst layer (CL). Excess water may occlude active sites for electrochemical reaction and affect the gas permeability [6, 7, 8]. Most water is transported out of the cathode CL and accumulates at the interface between the CL and GDL before draining from the MEA through the GDL by capillary action [9].
GDLs commonly consist of a substrate and microporous layer (MPL). To achieve water removal and gas diffusion, the substrate is typically waterproofed by introducing polytetrafluoroethylene (PTFE) within micropores. This prevents the GDL surface and pores from being clogged with liquid water and facilitates gas transport to the CL. The treatment process normally results in inhomogeneous PTFE distribution across the GDL substrate, with higher concentrations at the surface. Neutron radiography reportedly indicated that water concentrates in the center of the carbon paper when the fuel cell is operating [10, 11, 12], reflecting the fact that GDLs are easily flooded.
In the current study, hydrophobization treatment was used to achieve an even PTFE distribution in carbon paper. By vacuumizing the residual gas within carbon paper pores, PTFE solution can more effectively penetrate the pores. The resulted homogeneous hydrophobic pores exhibit favorable anti- flooding properties.
PTFE was introduced within the micropores of carbon paper (Toray, Japan, TGP-H-060) for hydrophobization. Carbon papers were immersed in PTFE solution, and excess solvent was removed by drying, leaving PTFE remaining in micropores. Carbon papers were treated by this process at pressures of 0, −0.08, and −0.05 MPa, with the resulting samples named CP-1, CP-2, and CP-3, respectively. The PTFE contents of the three carbon paper samples were controlled to ensure that they were constant. The resulting carbon papers were sintered at 240 °C for 30 min and then at 340 °C for 30 min in a N2 atmosphere.
The water contact angle of the carbon paper surface was measured using the KRÜSS DSA100 Drop shape analysis system. A total of 3 ml of distilled water was dropped onto the surface of the carbon paper at room temperature. The contact angle image was recorded 5 s after the drop was placed. The contact angle was analyzed using the tangent method of Sessile Dropfitting.
A 3 cm × 3 cm section of carbon paper was removed and weighed after hydrophobization. The paper was vacuumized and immersed in water at 20 °C for 2 h. The surface was wiped dry, and the total weight of the paper and water within hydrophilic pores was recorded. This process was repeated until a constant weight was achieved, from which the hydrophilic pore volume was calculated. Residual water was removed from the pores by drying at 90 °C under vacuum, and the paper was immersed in dodecane under vacuum at 20 °C for 2 h. The subsequent procedure was the same as for that above, and the total pore volume was calculated. The porosity of the carbon paper was measured by the mercury intrusion method with a Poremaster GT60 instrument (Quantachrome).
Pt/C (Johnson Matthey, 70 wt%), Nafion® solution (DuPont 5 wt%), and isopropyl alcohol were ultrasonically mixed to prepare the catalyst ink, which was sprayed onto a Nafion® 211 membrane to form the catalyst coated membrane (CCM). Carbon powder (XC-72), PTFE solution, and ethanol were mixed to form a homogeneous ink, which was painted onto the carbon paper. The GDL was prepared after sintering at 240 °C for 30 min and then at 340 °C for 30 min in a N2 atmosphere. GDLs were fabricated using CP-1 and CP-2.
A CCM was sandwiched between two GDLs, and two MEAs were hot-pressed, which were relevant to CP-1 and CP-2, respectively (Fig. 1). The active area of the MEAs was 270 cm2.
A short stack was assembled with a graphite flow field and two stainless steel bipolar plates. Evaluation of green light (G100) with a temperature, gas flow rate and pressure controlled system was used to test the cell performance. i-V curves were measured under an electronic loading (Kikusui, PLZ1004&PLZ2004). Electrochemical impedance spectra (EIS) were measured with a Kikusui KFM2150 instrument.
The carbon papers were characterized after hydrophobization treatment. Cross-sectional SEM images of fresh and treated carbon papers are shown in Fig. 2. The straight line indicates the analyzed cross-section, and the flexural curve indicates the elemental F distribution measured by an electron probe. Fresh carbon paper contained carbon fibers and organic materials, with F distributed across the entire section. Introducing PTFE within micropores did not change the morphology of the cross section, but the F distribution was influenced by vacuum PTFE treatment. CP-1 treated at ambient pressure had a higher PTFE distribution near the paper surface, which may subsequently result in water flooding in the central area and therefore limited gas penetration. CP-2 and CP-3 exhibited more even PTFE distributions.
The proportion of hydrophilic and hydrophobic pores was investigated by immersing a carbon paper in water and dodecane. After immersing in water, the hydrophilic pore volume was calculated from the increased weight and water density. After immersing in dodecane, the total pore volume was calculated from the increased weight and dodecane density. The total pore volume consisted of the hydrophilic and hydrophobic pore volume contributions. The proportion of hydrophilic and hydrophobic pores was obtained, as shown in Table 1. The percentage of hydrophobic pores for CP-1 and CP-2 were 56.8% and 60.3%, respectively, which indicated that PTFE distribution within the carbon paper was improved by the treatment.
The total porosities of carbon papers before and after hydrophobization were also measured by mercury infusion, and pore size distribution curves are shown in Fig. 3. The porosity of carbon paper largely depended on the density of the carbon fiber and binder. Thus, porosity did not significantly change upon hydrophobization. The predominant pore size of fresh carbon paper was concentrated near 40 μm, while those of CP-1 and CP-2 were 35 and 30 μm, respectively. This suggested that the PTFE within the carbon paper may have influenced the pore size distribution.
The total amount of immersed PTFE was constant for CP-1 and CP-2. The PTFE distribution within carbon paper was improved by vacuumizing the hydrophobization treatment, so the PTFE content on the CP-2 surface was lower than that on CP-1. Water contact angles on carbon paper were measured to determine surface hydrophobicity, which may have changed with surface PTFE content. A base line was drawn across the two droplet-paper intersections, as shown in Fig. 4. The left and right contact angles between the baseline and tangents across the two intersections were obtained. The contact angles of CP-1, CP-2, and fresh carbon paper were 150°, 145°, and 134.7°, respectively. This indicated that the surface hydrophobicity of carbon paper was improved by introducing PTFE and was related to the PTFE content of the surface.
Two GDLs were formed by adding a microporous layer to the carbon paper. By sandwiching a self-fabricated CCM with the prepared GDLs, two MEAs were prepared. MEA-1 and MEA-2 incorporated CP-1 and CP-2, respectively. A short stack was assembled and activated to measure cell performance, as shown in Fig. 5. MEA-2 containing carbon paper treated by immersing under vacuum conditions exhibited better performance at high current density, especially at current densities > 0.8 A/cm2. For MEA-1, water tended to accumulate in the center of the carbon paper because of the lower PTFE content at this point, and this may have influenced gas diffusion. Hydrophobization treatment under vacuum resulted in PTFE more efficiently penetrating into the center of the carbon paper, which ensured effective gas diffusion. Such characteristics were manifested in the better cell performance for MEA-2.
Nyquist plots of the two MEAs at 1 A/cm2 are shown in Fig. 6 as dispersed point. Hydrogen oxidation was sufficiently fast that the impedance of the anode could be neglected. Thus, the EIS polarization resistance reflected that of the cathode. The high frequency intercept with the real axis typically represents the ohmic resistance (Rohm) of a single cell, indicating the resistance of materials and the contact resistance between them. The first EIS arc relates to the charge transfer resistance (Rct) during oxygen reduction and double layer capacitance within the catalyst layer. The second EIS arc relates to the gas diffusion resistance (Rm). Using the equivalent circuit shown in Fig. 7, EIS spectra were fitted using commercial ZSimpWin software. Comparison of the fitting results (Fig. 8) showed that Rohm decreased slightly, while Rct and Rm decreased more significantly. All components within the two MEAs were constant except for the hydrophobization treatment, so the decreased resistances were attributed to the more even PTFE distribution. Such PTFE distribution discouraged water accumulation in the center of the carbon paper and enhanced gas diffusion. Oxygen reduction was subsequently accelerated and Rct decreased. The greater resistance to water flooding may have further favored Rohm.
PTFE could not thoroughly penetrate into the center of carbon paper upon conventional hydrophobization treatment, and water flooding occurred at high current density. Vacuumizing the carbon paper during immersion in PTFE solution improved PTFE distribution in the GDL. The surface hydrophobicity of the carbon paper decreased because of the lower PTFE content resulting from the vacuum treatment. However, the cell performance increased significantly. We investigated the proportion of hydrophobic and hydrophilic pores by immersing the carbon paper in different solvents and measuring its increased weight. Hydrophobization treatment under vacuum increased the proportion of hydrophobic pores, resulting in a more even PTFE distribution. Cell performance and fitted EIS spectra indicated that polarization resistance was improved for the MEA containing vacuum treated carbon paper. The mass transport controlling region may have shifted to higher current density.
2014, Vol. 35 Issue (4): 468-473
PDF (984KB)
2014, Vol. 35 Issue (4): 468-473
PDF (984KB)