1. Introduction
It has been reported that polymer electrolyte fuel cells (PEFCs), including direct methanol fuel cells (DMFCs), show large deviations in current distribution
[1-
8]. The deviations are affected by device operating conditions such as the reactant flow rate, reactant concentrations, temperature, and configuration of flow field, as well as by the material properties of the membrane-electrode-assembly (MEA). Deviations from a homogenous current distribution can affect the overall cell performance. Furthermore, continuous deviations can reduce the lifetime of the cell
[5,
7]. Therefore, it is necessary to alleviate the inhomogeneity of current distribution.
In spite of the necessity, few studies have been reported on methods to reduce the inhomogeneity of current distribution. One such study by Santis
et al. investigated the homogeneity of current distribution in a hydrogen-fueled polymer electrolyte fuel cell
[9]. In this investigation, redistribution of catalyst loading along the channel was achieved to realize homogenous current densities in these devices. A mild or steep increase in catalyst loading along the channel was found to be effective. Studies on the variation in catalyst loading have also been conducted. Wilkinson and St-Pierre
[10] reported improved device performance by using an in-plane decreasing gradient of catalyst loading from the cathode inlet to the outlet with three separate sections of different catalyst loadings. However, the homogeneity of the current density was not considered. Prasanna
et al. [11] reported improvement in catalyst utilization to reduce the need for platinum by using a catalyst-gradient electrode with increased catalyst loading toward the gas outlet. Cell performance was shown to remain unaffected by this loading approach. Prasanna
et al. focused on the effects of the gradient method on single cell performance, and did not consider the homogeneity of current density.
In this study, based on our previous results
[7], the amount of catalyst loading along the surface of the cathode side was varied to reduce the inhomogeneity of current distributions and to obtain higher cell performance. The strategies used to achieve spatial variation in catalyst loading included increasing the catalyst loading at the cathode side where the current densities were found to be lower in our previous study, whilst maintaining a constant total catalyst loading between the fabricated fuel cells.
3. Results and Discussion
The current distributions at different air-flow rates of the cell with a current loading of 2 A and catalyst loadings as described in
Fig. 2 are presented in
Fig. 3,
4, and
5. The current density for the T4 cell was observed to be lower near the cathode outlet
[3,
5,
7]. This observation was found to be more apparent with low air-flow rates at a low applied current of 2 A. This inhomogeneity in current density was alleviated by the catalyst loading of U3D5, and it noticeably reduced for the catalyst loadings for L5M3R5 and U5D3. For U2D6, in contrast, the inhomogeneity of the current density worsened. The “alleviating” and “worsening” mechanisms are different for each catalyst loading and are detailed below.
Fig. 3.
Current distributions and the corresponding voltages for T4, U3D5, and U2D6 with a loaded current of 2 A at different cathode flow rates (λ) of (a) 2, (b) 3, and (c) 5.
Fig. 4.
Current distributions and the corresponding voltages for T4 and L5M3R5 with a loaded current of 2 A at different cathode flow rates (λ) of (a) 2, (b) 3, and (c) 5.
Fig. 5.
Current distributions and the corresponding voltages for T4 and U5D3 with a loaded current of 2 A at different cathode flow rates (λ) of (a) 2, (b) 3, and (c) 5.
The current distribution profiles of U3D5 were measured to be different from those measured for the T4 cell. This difference in the current distribution profile can be ascribed to the increase in the catalyst loading near the cathode outlet. In this case, even though water flooding remains important near the cathode outlet, the increased catalytic active area at the lower half with 5 mg·cm
-2 of Pt increases the reaction rate at this region
[10,
11]. For U3D5, therefore, the current densities near the cathode outlet are slightly higher than those for T4, as shown in
Fig. 3 (a),
(b), and
(c). It was also found that the cell voltage enhanced at all cathode flow rates for this catalyst loading (
Fig. 3)
[7,
11]. However, the catalyst loading of U2D6 had an opposing effect on the current distribution. The inhomogeneity of the current distribution with U2D6 was measured to be worse than that measured for T4, resulting in a decrease in the cell voltage (
Fig. 3). This inhomogeneity in the current distribution for U2D6 may be ascribed to the thick catalyst layer at the lower half of the cell with a catalyst loading of 6 mg Pt·cm
-2. Thick catalyst layers can increase the mass transport resistance; thus, oxygen transport through the catalyst layer is hindered and local water content cannot be easily removed. As a result, current densities at the thick region reduced. In addition, the periodic drops near the U-bend of the cathode inlet are more clearly observed, which suggests that water management with this catalyst loading is poor
[7].
The homogeneity of the cell current distribution is improved by L5M3R5 loading (
Fig. 4). This improvement is more evident at the lower air-flow rates. To increase the active catalyst area in this region, catalyst loading near the U-bend where the current densities periodically drop is enhanced in L5M3R5
[7]. As expected, the periodic drops are not observed with this catalyst loading. Moreover, the current distribution of the L5M3R5 cell has a homogenous current distribution profile, which may be attributed to the periodic enhancement of the catalyst loading. Even though the oxygen mass fraction decreases as the air goes through the channel toward the outlet, the catalyst loading regularly increased at the region near the U-bend along the air path, thus increasing the catalytic active area at this region. The increase in the catalytic active area of this cell subsequently prevents any decrease in the current density toward the cathode outlet. In addition to the homogenization of current distribution, the cell voltage was also enhanced by the L5M3R5 loading at a low air-flow rate (
λ = 2).
The homogeneity of the current distribution was also shown to be increased by the U5D3 catalyst loading approach (
Fig. 5). This approach involves the opposite catalyst loading to that of U3D5. However, the “alleviating” effect on the current distribution profile is more evident for this cell. Wilkinson
et al. [10] reported a similar approach with regard to the catalyst loading variation that led to the previously reported improvement in the cell performance of hydrogen-fueled PEFCs. In their work, Wilkinson
et al. explained that this approach reduces both kinetic and mass transfer losses near the cathode outlet, where liquid water mass transfer is most prevalent. The catalyst loading is reduced at the cathode outlet so that the oxygen mass transfer rates increase. Similar current distribution mechanisms were observed in this study.
Current distributions at higher catalyst loading with an applied current of 5 A are presented in
Fig. 6. Compared to those at 2 A, the current distributions were more consistent at 5 A. At the high applied current of 5 A, the T4 cell is shown to have a homogenous current distribution regardless of the air-flow rate
[7]. As a general summary, there did not appear to be any improvement in the homogeneity of the current distribution for all catalyst loadings at the higher loading currents (5 A). The current densities with the U2D6 catalyst loading remained inhomogeneous in distribution. For L5M3R5, we observed a slight increase in current density toward the cathode outlet (
Fig. 6 (b)), which became more obvious as the cathode flow rate increased. Therefore, we can say that the strategy of varying the catalyst loading is more effective in maintaining a homogenous current distribution at lower air flow rates and lower applied currents.
Fig. 6.
Comparisons of the current distributions and the corresponding voltages of T4 with those of (a) U3D5 and U2D6, (b) L5M3R5, and (c) U5D3 with a loaded current of 5 A at the lowest cathode flow rate (λ) of 2.
To compare the overall effects of the spatial variations in catalyst loading to one another, the half-sum ratio is defined as the ratio of the half-sum current densities near the cathode outlet to those near the cathode inlet:
Cn is the current measured at the n
th segment position. The normalized standard deviation is defined as the standard deviation of the current distribution from the average divided by each loaded current (2 A and 5 A)
[8].
Fig. 7 summarizes the standard deviations of the current density and the half-sum ratios for the five spatial variations of catalyst loading as a function of the cathode flow rate. These values were calculated from the data shown in
Fig. 3 to
Fig. 6 (current distributions at 2 A for
λ = 10 and those at 5 A for
λ = 3, 5, and 10 have not been shown in this article). U3D5, L5M3R5, and U5D3 have lower standard deviations than T4 at a current load of 2 A for all the examined cathode flow rates. Among these three loading variations, the deviation for L5M3R5 was lowest at the low air flow rates of
λ = 2 and 3. The standard deviations of T4, U3D5, and U5D3 were similar at the loaded current of 5 A for all flow rates, as seen in
Fig. 7 (b). This result suggests that the variations in the catalyst loading are not effective at the high loaded current of 5 A, which may be attributed to the relatively high air flow rate needed to satisfy the required stoichiometric factors. The half-sum ratios at 2 A were nearest to unity (marked with a dotted line) when L5M3R5 was used at
λ = 2 and 3 and when U3D5 and U5D3 were used at
λ = 5 and 10. The half-sum ratios at 5 A were nearest to unity for T4, U3D5, and U5D3 at all cathode flow rates examined. The ratios for L5M3D5 were higher than unity at all flow rates, and those for U2D6 were lower than unity at all flow rates.
Fig. 7.
Standard deviations at (a) 2 A and (b) 5 A and half-sum ratios at (c) 2 A and (d) 5 A for T4, U3D5, U2D6, L5M3R5, and U5D3 as a function of the cathode flow rate from the data of Fig. 3 to Fig. 6.