Nanofludic Analysis on Methanol Crossover of Direct Methanol Fuel Cells

Abstract

Direct methanol fuel cells (DMFCs) are considered as a competitive power source candidate for portable electronic devices. Nafion® has been widely used for the electrolyte of DMFCs because of its good proton conductivity and high chemical and mechanical stability. However, the major problem that must be solved before commercialization is the high methanol crossover through the membrane. There are a number of studies on experiments about the methanol crossover rate through the membrane but only few theoretical investigations have been presented [1–3]. In this paper, an atomistic model [4] is presented to analyze the molecular structure of the electrolyte and dynamic properties of nanofluids at different methanol concentration. In the same time, the nano-scopic phenomenon of methanol crossover through the membrane is observed. The simulation system consists of the Nafion fragments, hydronium ions, water clusters and methanol molecules. Fig. 1 shows the simplified Nafion fragment in our simulation. Both intra- and inter-molecular interactions were involved in this study. Intermolecular interactions include the van der Waals and the electrostatic potentials. Intramolecular interactions consist of bond, angle and dihedral potentials. The force constants used above were determined from the DREIDING force field. The SPC/E model was employed for water molecules. The three-site OPLS potential model was utilized for the intermolecular potential in methanol. Each proton which migrates inside the electrolyte is assumed to combine with one water molecule to form the hydronium (H3O+). The force parameters for the hydronium were taken from Burykin et al [5]. The atomistic simulation was carried out on the software DLPOLY. First, a 500 ps NPT ensemble was performed to make the system reach a proper configuration. This step was followed by another 500 ps NVT simulation. All molecular simulations were performed at a temperature of 323K with three-dimensional periodic boundary conditions. The intermolecular interactions were truncated at 10 Å and the equations of motion were solved using the Verlet scheme with a time step of 1 fs. Fig. 2 shows the calculated density of the simulation system for different methanol concentrations at 323K. It can be seen that the density decreases with the methanol uptakes. The volume of the system increases as the methanol concentration increases, which means that the membrane swelling with methanol uptakes. The radial distribution functions (RDFs) of the ether-like oxygen (O2) toward water and methanol molecules for different methanol concentrations at 323K are shown in Fig. 3. From this figure, we find that methanol molecules can reside in the vicinity of the hydrophobic part of the side chain while water can not. Fig. 4 shows the RDFs between the oxygen atom of the sulfonic acid groups (O3) and solvents for different methanol concentrations at 323K. As shown in Fig. 4, both water and methanol have a tendency to cluster near the sulfonic acid groups, but water molecules prefer to associate with the sulfonic acid groups in comparison with methanol molecules. The mean square displacements (MSDs) of water and methanol molecules for different methanol concentrations at 323K are displayed in Fig. 5. It is shown that MSD curves have a linear tendency, which means both water and methanol molecules are diffusing in the system during the simulation. As the methanol concentration increases, the slope of MSD curve increases for methanol and decreases for water. This indicates higher methanol content constrains the mobility of water molecules but enhances the mobility of methanol molecules that cross the electrolyte. In summary, molecular simulations of the Nafion membrane swollen in different methanol concentrations (0, 11.23, 21.40, 46.92 wt%) at 323K have been carried out. Both methanol migration mechanism and hydronium diffusion phenomenon have been visualized by monitoring the trajectories of the specific species in the system. MSDs are used to evaluate the mobility and shows that the higher the methanol concentration, the greater the tendency of methanol crossover.