Over the past few decades, direct methanol fuel cells (DMFCs) have attracted significant attention as a potential candidate for environmentally friendly energy conversion devices. Compared with hydrogen proton exchange membrane fuel cells (PEMFCs), DMFCs are safer to operate because of the liquid characteristics of methanol fuel ^{[1, 2]}. A membrane electrode assembly (MEA) is a central part of DMFCs ^{[3, 4]}. A MEA usually consists of a gas diffusion layer (GDL) having a microporous layer (MPL), a catalyst layer, and a proton exchange membrane (PEM). The GDL controls water, gas, and electron transfer ^[5-7]. PEMs transfer protons and separate fuel from oxygen.

The cathode catalyst layer (CL) is the crucial zone of the electrochemical reaction. The heterogeneous reaction occurs at a three-phase boundary comprising reactants, catalysts and an electrolyte ^[8]. The state-of-the-art catalysts used for methanol oxidation and oxygen reduction are PtRu/C and Pt/C, respectively. The electrolyte is an perfluorosulfonic acid polymer, such as Nafion ^{[9, 10]}. There are numerous reports investigating the detailed structure of CLs in PEMFCs, which have led to improvements in cell performance through optimizing the CL structure ^{[9, 11-20]}. Xing et al. ^[20] presented a two-dimensional, two-phase flow model detailing how ionomer swelling results in both a decrease in cathode CL porosity and an increase in the ionomer coating thickness, thereby increasing the oxygen transport resistance. Optimal cell performance in the presence of fully humidified gases utilizes an initial dry ionomer volume fraction (initial ionomer content) close to 10%, corresponding to 0.3 mg/cm². Mashio et al. ^[21] employed a mathematical model to analyze the water sorption isotherm of PEMFC catalyst layers and observed water adsorption in the CLs initiates on the adsorption sites. Soboleva et al. ^[22] reported that ionomer co-deposition in the CL strongly influenced its porosity, covering a pore range ＜ 20 nm, which originated from the microporous nature of the carbon particles (pore sizes ＜ 2 nm) and the agglomeration of the carbon particles (pore sizes of 2-20 nm). Passalacqua et al. ^[23] investigated Nafion content in the PEMFC catalyst layer and observed the optimal ionomer content to be ~33 wt.%. Multiple investigations have reported hybrid and multi-graded catalyst layer combinations to optimize the function of the CL in PEMFCs ^[24-27]. Lee et al. ^[28] designed a hybrid anode CL to study methanol and water transport in DMFCs. However, there are few reports studying the optimi zation of cathode CLs in DMFCs.

Herein, cathode CLs were prepared as a function of ionomer content and the influence on the cell structure and catalytic behavior was investigated. Electrochemical surface areas (ESAs) were found to increase with increasing Nafion content until a constant value was reached. Additionally, further increasing the ionomer content beyond this threshold level results in the inability to control water and oxygen mass transfer effectively. N30 (30 wt.% Nafion content) is found to be the optimal level to effectively extend the three-phase boundaries, improving cell performance.

In our experiment, MEAs were prepared as follows: the MPL ink was prepared by mixing Vulcan XC-72 carbon with PTFE (25 wt.%), followed by spraying onto carbon paper (anode: TGPH060, 20 wt.% PTFE; cathode: TGPH090, 20 wt.% PTFE). The carbon loading was 1 mg/cm². The MPL-treated carbon paper was heat-treated at 320 °C for 0.5 h to prepare the GDL. The anode and cathode catalysts were Pt40%Ru20%/C and Pt60%/C (Johnson Matthey Corp.), respectively. The catalyst ink was prepared by dispersing the catalysts and 5 wt.% Nafion solution in a solution composed of water and isopropanol (v:v = 2:3). Subsequently, the ink was sprayed onto the prepared GDL at 60 °C to form the gas diffusion electrode (GDE). Precious metal loading with respect to the anode was controlled at 2 mg/cm², and the content of Nafion was fixed at 25 wt.%. Pt loading at the cathode was also 2 mg/cm²; however, the Nafion content varied at: 15, 25, 30, 35 and 45 wt.%, respectively. Thereafter, the MEA was prepared by hot-pressing the electrodes on two sides of Nafion 115 at 135 °C at 7.5 MPa for 2 min. The electrode area was 9 cm², and the MEAs were named as N15, N25, N30, N35 and N45 corresponding to the Nafion content at the cathode.

The cyclic voltammetry (CV) tests of MEAs were performed with an EG&G model 273 potentiostat/Galvanostat (Oak Ridge, USA). The cathode was employed as the working electrode and the dynamic hydrogen electrode (DHE) at the anode was used as the reference electrode and counter electrode. Deionized water at 35 °C was bubbled to the cathode at a flow rate of 8.5 mL/min, whereas hydrogen was bubbled to the anode at a flow rate of 150 mL/min. The cell remained at a constant temperature of 35 °C throughout the duration of the test. The potential window was 0-0.8 V at a scan rate of 50 mV/s.

The polar curve of the single cell having an active area of 9 cm² was evaluated using an Arbin FCT test station (Arbin Instrument Inc., USA). Methanol flow rate (2 mol/L) and O₂ were fixed at 6 mL/min and 0.5 L/min, respectively. All performance measurements were performed at 60 and 70 °C. Electrochemical testing measurements and operating conditions are shown in Table 1.

The CL hydrophobic character with varying Nafion content was investigated using a drop-shape analysis system, DSA30 (KRUSS, Hamburg, Germany). CL surface morphology and roughness was observed by a 3D non-contact optical surface profilometer (Nanovea, California, USA) and scanning electron microscope (SEM, Philips XL30, field emission gun environmental SEM operating at 20 kV, Amsterdam, Netherlands).

The surface morphologies of the GDEs possessing varying Nafion contents were investigated using SEM and a 3D non-contact optical surface profilometer. 2D contour plots and the 3D surface morphology of the CLs are shown in Fig. 1. The CL roughness and thickness increased as a function of increasing Nafion content. As shown, the surfaces of N15 and N25 are relatively flat. Further increases in Nafion content result in a greater degree of surface roughness, as indicated by the higher peak densities homogeneously distributed across the sample. The area of sample studied, as indicated by the scale bars, is 5×5 mm²; therefore it can be observed that as surface roughness increases, the resulting micrometer-sized pores also increase; this may facilitate the mass transfer process. As shown in Fig. 2, increasing Nafion content resulted in the formation of a cracked surface with the gap between the islands extending farther apart as a function of increased Nafion content. The surface of N15 and N25 were relatively flat when compared with higher Nafion loadings. Significant gaps initially appear upon increasing Nafion loading to between 30 wt.% (N30) and 45 wt.% (N45). The results were consistent with 3D surface morphology observations of the CLs. Regarding the microstructure, observations reveal that the granular structure changed to a flocculent-type structure with increased Nafion loading. The granular structure mainly reflected the microstructure of the catalysts, whereas the flocculent-type structure originated from the microstructure of the Nafion ionomer. Additionally, we found that GDE performance declined dramatically with lower Nafion contents because Nafion acts not only as a proton conductive material but also as the binder in the CL. Therefore, CLs with lower Nafion contents exhibit poor bonding with the MPL and proton exchange membrane.

The sulfonic functionality on the Nafion increases its hydrophilic nature when compared with the catalysts; and as a result, the contact angle decreases as a function of increasing Nafion content in the CL. As seen in Fig. 3, the contact angles decrease from 166.8° to 143.1° as Nafion loading increases from 15 to 45 wt.%. The hydrophobicity of the CLs directly affects the mass transfer of liquid and gases, thereby affecting the DMFC cell performance. Cathode CLs are easily flooded because the hydrophilicity increases such that the water generated by the reaction is unable to be removed efficiently, thus blocking the pores for efficient mass transfer.

Cyclic voltammetry (CV) is an effective approach to evaluate the utilization of Pt nanoparticles (NPs) in the CL. As shown in Fig. 4, in-situ CV was tested for DMFCs with varying Nafion loading at the cathode CL. The cathodes adopting Pt/C catalyst showed larger current density for proton adsorption/desorption (around 0.3 V) and double layer charging (around 0.45 V). The EAS obtained from the integrated charge for electrochemical proton desorption without double layer charging. As shown, ESA increases as a function of Nafion content. At low ionomer content, such as N15, the majority of Pt NPs are not in direct contact with the continuous ionomer phase, thus becoming inactive for proton transfer. Adjusting the Nafion content from 15 to 25 wt.% results in a sharp increase in ESA. However, when the Nafion mass percentage is increased to more than 35 wt.%, no obvious increase in ESA is observed, which is limited by the Pt loading (2 mg/cm²). Therefore, a Nafion loading of 35 wt.% was found to be optimal for the Pt NP ESA. Total coverage of Pt NPs only allows O₂ dissolved in Nafion to contact the surface of the catalysts. Thus, cell mass transfer processes are also influenced by the ionomer content; therefore, it is preferable to directly measure the optimal CL Nafion loading from in-situ cell tests.

All performance measurements were conducted at 60 or 70 °C. Fig. 5(a) shows the polarization curves of DMFCs with varying Nafion content at 60 °C. Single cell performance increases as a function of increasing Nafion content when ＜ 30 wt.%, which can be ascribed to the increase in the catalyst utilization due to the facilitated proton mass transfer. However, further increasing the Nafion content beyond 30 wt.% results in a decrease in the cell performance, which may be ascribed to the higher hydrophilicity of the cathode that leads to mass transfer issues, i.e., water control becomes increasingly difficult as Nafion loading increases resulting in facile cathode flooding. According to the investigations made by Soboleva et al. ^[22], the ionomer co-deposition in the CL strongly influences the electrode porosity for pores sizes ＜ 20 nm. Therefore, mass transfer along the mesopores will be limited with excess ionomer, resulting in lower catalyst efficiency. When the operating temperature is raised to 70 °C, the cell containing 30 wt.% Nafion (N30) still exhibits optimal performance, as shown in Fig. 5(b), with growing performance differences between the cells observed. Fig. 6 shows the differences in peak power densities between 60 and 70 °C. The change in performance of N15 was insignificant between the two temperatures tested. In contrast, for N30, the performance increased by 15 mW/cm² as the proton conductivity of Nafion and catalytic activity increased with temperature.

In summary, DMFCs with varying ionomer loadings in the cathode CLs were fabricated and their the physical characteristics were studied according to their surface morphologies and contact angle measurements. CL roughness and thickness were shown to increase with the increasing of Nafion content. CLs with lower Nafion loading exhibited poor bonding with MPLs and proton exchange membranes. ESA increased as Nafion loading was increased, indicating higher catalyst utilization at elevated ionomer content. However, further increasing the ionomer content significantly affected hydrophilicity and the pore structure, which resulted in mass transfer issues. To improve the efficiency of CLs, proton conductivity and mass transfer should be considered simultaneously. This study found that a Nafion loading of 30 wt.% yielded the optimal composition in the DMFC.