The surface treatment and functionalization were carried out as follows. The CNFs (herringbone type) were placed in a flask, and H₂SO₄/HNO₃ (v/v = 4:1) was added; the solution was ultrasonicated and stirred for 4 h. The CNFs were separated from the acids and washed with deionized water. A Pt salt precursor, H₂PtCl₆·6H₂O (Sigma-Aldrich Corporation, St. Louis, MO, USA) was dissolved in ethylene glycol, and the CNF support was dispersed in the solution. The suspension was filtered and dried at 60°C for 4 h in a vacuum oven. The heat treatment was performed in argon atmosphere at 350°C for 2 h.

In order to study the effect of the catalyst, a thin film electrode was manufactured. A mixed slurry composed of the catalyst mixed with isopropyl alcohol and a 5 wt.% Nafion ionomer solution was sprayed onto the polymer electrolyte membrane (Nafion-212 membrane). The amount of Pt was 0.4 mg/cm² on both electrodes. Table 1 lists the manufactured MEAs using a commercial catalyst and a synthesized Pt/CNF.

The polarization curves of the unit cell were used to gauge the cell temperature at 70°C under atmospheric pressure using H₂ and air at the anode and cathode, respectively. After obtaining the polarization curves, cyclic voltammetry [CV] was performed in the range of 0.05 to 1.2 V at a sweep rate of 50 mV/s with 20 and 100 cm³/min flow rates of H₂ and N₂ to the anode and cathode, respectively.

The durability of the assembled MEA was determined by acceleration tests using reverse potential operation under fuel starvation conditions. The acceleration experiment was operated at a current density of 400 mA/cm². The hydrogen stoichiometry of the anode was maintained at 0.5. If the cell potential reached at -0.5 V, then the recovery system would be driven for 30 s under an open circuit voltage [OCV] state. In the OCV condition, the stoichiometric ratios of hydrogen and air were maintained at 1.5 and 2.0. This process was considered as 1 cycle, and experiments were repeated 200 times. After the acceleration tests, the performance curve and CV were obtained using the same method.

Table 2 summarizes the physical properties of the Pt/CNF. The weight percentages of the Pt catalyst to the weight of the synthesized Pt/CNF and the commercial Pt/C were approximately 47.5 wt.% and 50.7 wt.%, respectively, based on the TGA analysis. The nitrogen adsorption and the pore size distribution were measured in the Pt/C and Pt/CNF catalysts via BET analysis. The measured BET surface areas of the Pt/CNF and Pt/C catalysts were 58.4 m²/g and 347.3 m²/g, respectively. Figure 1 illustrates the difference in pore diameter and pore volume between the Pt/CNF and Pt/C. The pore sizes of the Pt/CNF were between 1.7 and 35.0 nm, whereas those of the Pt/C catalyst were between 1.6 and 101.5 nm. In general, the pores were confirmed to be micropores (< 2 nm), mesopores (2 to 50 nm), and macropores (> 50 nm). The carbon support relieved the degradation with graphitization and low specific surface area. Pores in the catalyst layer are known to affect cell performance due to gas diffusion, mass transfer resistance, and water treatment 33 34 .

The morphology and crystallography of the Pt/CNF and Pt/C catalysts were studied with XRD. Figure 2 shows that face-centered cubic Pt crystal planes were observed. The average platinum particle size was calculated from the Pt (111), (200), (220), and (311) peaks using Scherrer's formula. The particle sizes in the synthesized Pt/CNF and commercial Pt/C were 2.8 and 3.4 nm, respectively. A carbon peak of (002) at 26.2° indicated that the CNF support was a graphitic structure, whereas the peak of the CB support was smaller.

The Pt particle size distribution and carbon type of the Pt/CNF and Pt/C catalysts were observed using TEM. Figure 3 shows that the Pt particles on the CNF supports were more distributed than those on the Pt/C catalyst. The mean particle sizes were approximately 2.5 and 3.4 nm in the Pt/CNF and Pt/C, respectively. An increase in Pt particle size can degrade cell performance with a decrease in the electrochemical surface area [ESA]. The morphology and the surface treatment with the CNF support are important due to the effect on the dispersion, particle size, and activity of Pt catalysts 31 . Figure 4 shows the hydrophilicity of the synthesized Pt/CNF and commercial Pt/C catalysts. Each catalyst was mixed with deionized water in an ultrasonic bath at room temperature. At first, the catalysts were mixed perfectly, with a change observed after 2 h. The Pt/C catalyst remained in a mixed state, while the Pt/CNF catalyst exhibited phase separation. The hydrophilicity increased with the increasing acid treatment on the carbon surface. The carbon surface was oxidized as the number of functional groups, such as carboxyls and hydroxyls, increased. However, hydrophilicity can decrease with increasing graphitization and elimination of functional groups with heat treatment. In addition, the hydrophilicity and durability of the catalyst are closely related, where the hydrophilic nature is known to degrade the durability 32 .

The cell performance was measured as shown in Figure 5. The change in cell voltage was determined during the fuel starvation cycles shown in Figure 5a. As shown in Figure 5b, the performance of the MEA-1 with the Pt/C catalyst rapidly decreased from 917 to 378 mA/cm² at 0.6 V after 200 cycles of fuel starvation. Activation loss and mass transfer loss were observed at a high voltage and a low voltage, respectively, as carbon corrosion was caused by applying a reverse voltage. Fuel starvation is known to cause carbon corrosion at the anode. Therefore, Figure 5c shows the performance of the MEA with Pt/CNF at only the anode. The performance of the MEA-2 was reduced by approximately 58 mA/cm². Activation and mass transfer losses were observed, but they were smaller than those of MEA-1. The performance of MEA-3 with the Pt/CNF catalyst with both an anode and a cathode was only reduced by approximately 10 mA/cm². Furthermore, activation loss and mass transfer loss did not occur after 200 cycles. These results indicate that MEA-3 is more durable than MEA-1 with the Pt/C catalyst because the CNF catalyst support has a strong resistance to corrosion due to its highly graphitic structure.

CV measurement was performed in order to examine changes in the ESA before and after fuel starvation. Based on the CV analysis shown in Figure 6, the ESA of the manufactured MEA-1 was initially calculated to be 59.4 and 48.5 m²/g at the anode (Figure 6a) and cathode (Figure 6b), respectively, but these values decreased to 33.2 and 33.3 m²/g after 200 cycles. Carbon corrosion caused the reduction of the ESA with the agglomeration and migration of Pt particles. Figure 6c, d shows the results of the CV on MEA-2. The ESA decreased by approximately 2.3 and 4.4 m²/g at the anode and cathode, respectively. These results mean that carbon corrosion resistance has occurred at both the anode and cathode when the Pt/CNF catalyst was used on the anode. As shown in Figure 6e, f, the ESA of MEA-3 was nearly unchanged in both electrodes. This tendency is similar to the cell performance results. The results of the electrochemical analysis are summarized in Table 3. Electrochemical impedance analysis measured the change in resistance at 200 mA/cm² for the manufactured MEAs. The cell resistance of MEA-1 increased from 0.103 to 0.163 Ω after 200 cycles of fuel starvation. However, MEA-2 and MEA-3 showed little change in cell resistance. Therefore, it is clear that the Pt/CNF catalyst is more durable against carbon corrosion than the Pt/C catalyst.