Inﬂuence of Ionomer Content on Both Cell Performance and Load Cycle Durability for Polymer Electrolyte Fuel Cells Using Pt/Nb-SnO 2 Cathode Catalyst Layers

The steady-state current-voltage performance and load cycle durability of polymer electrolyte fuel cells using Pt catalysts supported on Sn 0.96 Nb 0.04 O 2- δ (Pt/Nb-SnO 2 ) cathode catalyst layers (CLs), without any carbon additive, were evaluated with various ionomer contents. The apparent mass activity at 0.80 V (MA app @0.80 V) of the cell using Pt/Nb-SnO 2 CL improved with decreasing volume ratio of the Naﬁon ionomer to the support (I/S). At I/S = 0.12, surprisingly, the MA app @0.80 V of the cell using Pt/Nb-SnO 2 CL approached 2 times higher than that using a commercial Pt catalyst supported on GCB (Pt/GCB) CL with the optimized I/S ratio. The current density at 0.60 V of the cell using Pt/Nb-SnO 2 (I/S = 0.12) CL also reached the same value of the cell using Pt/GCB CL. The electrochemically active surface area and MA app @0.80 V of Pt/Nb-SnO 2 CLs during the load cycle durability test maintained higher values than those of Pt/GCB throughout the test, indicating that the Pt/Nb-SnO 2 CL had higher durability than the Pt/GCB CL. Evaluation by means of low acceleration voltage transmission electron microscopy proved that the Naﬁon ionomer covered the hydrophilic surface of the Pt/Nb-SnO 2 uniformly, whereas the ionomer did not do so on the hydrophobic surface of the Pt/GCB. The thin, uniform coverage of the Naﬁon ionomer, which was thus easily obtained on the Pt/Nb-SnO 2 surface, was able to lower the overpotential originating from the oxygen diffusion resistance in the Naﬁon ionomer, thus improving the cell performance while maintaining the high load cycle durability.

The practical application of polymer electrolyte fuel cells (PE-FCs) to electric vehicles and residential co-generation systems can lead to the reduction of carbon dioxide emissions and improved distribution of electric power generation, respectively. The enhancement of the cathode catalyst activity for the oxygen reduction reaction (ORR) and the improvement of the durability are important issues for the widespread application of PEFCs. [1][2][3] The commonly used Pt and PtCo catalysts loaded on carbon black (Pt/CB, PtCo/CB) have relatively high catalytic activity, with their high electrical conductivity and well-developed pore structure. [4][5][6] These catalysts, however, have the well-known vulnerability to serious degradation during the startup and shutdown processes of the PEFC, due to the following reaction: [7][8][9][10][11][12] C + 2H 2 O → CO 2 + 4H + + 4e − (E o = 0.207 V vs. SHE) [1] A Pt catalyst loaded on graphitized carbon black (Pt/GCB) was confirmed to moderate the degradation during startup and shutdown compared with that of Pt/CB, but graphitized carbons have not been able to completely overcome the corrosion at high potential. [13][14][15] The development of other candidate supports, avoiding the use of carbon, is expected to further improve the durability. [16][17][18][19][20][21][22] As further examples, Pt catalysts loaded on Magneli phase titanium oxides or on tin dioxide have been reported to be highly stable and to exhibit high ORR activity. [23][24][25][26][27][28][29][30][31] Our group has also reported that Pt catalysts supported on Nb-doped SnO 2 (Pt/Nb-SnO 2 ) and Ta-doped SnO 2 (Pt/Ta-SnO 2 ), without any carbon additive, showed high durability during startup/shutdown, while maintaining high ORR activity. [32][33][34][35][36][37][38] The high durability during startup/shutdown of the Pt/Nb-SnO 2 and Pt/Ta-SnO 2 is dependent on the stability of Nb-SnO 2 and Ta-SnO 2 .
The high ORR activity of Pt/Nb-SnO 2 and Pt/Ta-SnO 2 relies on a strong interaction between the Pt catalyst and the tin oxide. The cell resistivity of a membrane electrode assembly (MEA) using our Pt/Nb-SnO 2 cathode was also equal to that using a Pt/CB cathode, and the cell performance using the Pt/Nb-SnO 2 cathode matched that obtained J3084 Journal of The Electrochemical Society, 165 (15) J3083-J3089 (2018) volume ratios of Nafion ionomer (dry basis) to the support (I/S) were adjusted in steps from 0.12 to 0.67. A commercial Pt/GCB catalyst (TEC10EA30E, Pt loading amount 29 wt%, Tanaka Kikinzoku Kogyo K.K., Japan) was used as a reference for the cathode measurements. Commercial Pt/CB (TEC10E50E, Pt loading amount 46 wt%, Tanaka Kikinzoku Kogyo K.K., Japan) was utilized for all anodes of the test cells in a manner similar to that described above. The I/S values of both the reference cathode and the anode were adjusted to 0.67, previously found to be the optimal ratio. 39 These catalyst inks were directly sprayed onto both sides of the Nafion membrane (NRE 212, 50 μm thickness, Du Pont, U.S.A.) as the anode and cathode by use of the pulse-swirlspray technique (PSS, Nordson Co., U.S.A) to prepare the catalystcoated membranes (CCM), which were then dried at 60 • C in an electric oven. The CCMs were annealed by hot-pressing at 140 • C and 10 kgf cm −2 for 3 min, and then they were assembled with two gas diffusion layers (GDLs, 25BC, SGL Carbon Group Co., Ltd., Germany) and a single serpentine flow-channel cell (Japan Automobile Research Institute standard cell) [36][37][38] with an active geometric area of 29.2 cm 2 . The Pt loading amount on the cathode side was 0.05 ± 0.01 mg cm −2 for all of the cathodes. The Pt loading of the anodes with Pt/CB was controlled at 0.55 ± 0.05 mg cm −2 . The high Pt loading amount at the anode made the polarization negligibly small under all of the experimental conditions, and therefore the anodes were able to be used as a reversible hydrogen electrode (RHE) for the evaluation in single cell measurements. Pt/GCB and Pt/Nb-SnO 2 CLs formed on polypropylene film were observed by use of low acceleration voltage transmission electron microscopy (low acceleration voltage TEM, HT 7700S, acceleration voltage 80 kV, Hitachi High-Technologies. Co., Japan) with a high resolution lens (EXALENS, Hitachi High-Technologies. Co., Japan). The cumulative pore volume of Pt/Nb-SnO 2 CLs (I/S = 0.24 and 0.67) was estimated by the Kr gas adsorption method (BELSORP-max II, MicrotracBEL Co., Japan).
Fuel cell operation.-The cell voltage as a function of current density was measured with hydrogen and oxygen/air at 80 • C under ambient pressure (1 atm). Hydrogen gas was supplied to the anode and oxygen or air to the cathode. The flow rates of all gases were controlled by mass flow controllers. The utilizations of the reactant gases were 70% for H 2 and 40% for air. These gases were humidified at 80% relative humidity (RH) by purging through a hot water reservoir, controlled by a measurement system (FCE-1, Panasonic Production Technology Co., Ltd., Japan). The current-voltage (I-V) curves were galvanostatically measured under steady-state operation, after a stabilization time of 5 min. for each current by use of an electronic load (PLZ-664WA, Kikusui Electronics Co., Japan) controlled by the FCE-1. All of the potentials are referred to the hydrogen anode, which was assumed to approximate the RHE potential. The digital ac milliohmmeter (Model 356E, constant frequency 10 kHz, Tsuruga Electric. Co., Japan) was used to obtain the cell resistances of MEA using Pt/Nb-SnO 2 CL. In the case of the MEA using Pt/GCB CL, a digital ac milliohmmeter (Model 3566, constant frequency 1 kHz, Tsuruga Electric. Co., Japan) was used. 37 Cyclic voltammetry (CV) measurements were performed at 80 • C, 80% RH by use of a potentiostat (PGSTAT128N, Metrohm Autolab B.V., Netherlands) in order to evaluate the electrochemically active surface area (ECA) of the Pt catalyst in the cathode CLs. The cathode compartment was purged with N 2 (100 mL min −1 , 80% RH), while H 2 gas (100 mL min −1 , 80% RH) was supplied to the anode. Prior to the potential sweep, the potential was maintained at 0.075 V for 3 s to ensure that the Pt was in a reproducibly reduced state. Then, the potential was swept from 0.075 V to 1.000 V at 20 mV s −1 and reversed back to 0.075 V. The N 2 flow was stopped during the entire CV measurement to avoid perturbing the H 2 partial pressure. 36 The values of ECA were determined from the hydrogen adsorption charge referred to 0.21 mC cm −2 , the conventional value for a monolayer of adsorbed hydrogen on clean polycrystalline platinum. 45 The Pt utilization (U Pt ) value of the Pt catalyst is defined as follows: The ECA is divided by the total specific Pt surface area, S Pt , which is not calculated from the mean particle size in the conventional manner but is estimated from the area accumulated from the surface of a large number of Pt particles observed by TEM, in order to evaluate the catalyst properties more precisely. 46

Durability evaluation simulating load changes of the PEFC.-
The durability evaluation simulating the load change operation of the PEFC was performed according to the procedure of the standard potential cycle protocol recommended by the Fuel Cell Commercialization Conference of Japan (FCCJ). The load change cycling was operated with H 2 (anode, 100 mL min −1 ) and N 2 (cathode, 100 mL min −1 ) atmospheres at 80 • C, 80% RH. In the initial stage, I-V curves and CVs were measured. Then, the durability was evaluated by means of the square wave cycling, in which the voltage was stepped between 0.6 and 1.0 V, with a holding time of 3 s at each voltage (FCCJ protocol). The variations in ECA at the cathode were examined by CV. After the durability evaluations, the cross-sections of the CLs were observed from small samples cut from the centers of the CCM with a focused ion beam apparatus (FIB, FB 2200, Hitachi High-Technologies Co., Japan) by use of scanning transmission electron microscopy (STEM, HD 2700, Hitachi High-Technologies Co., Japan).

Characterization of the catalysts and cathode catalyst layers.-
A TEM image of the Pt/Nb-SnO 2 catalyst is shown in Fig. 1. The Pt nanoparticles, several nanometers in diameter with darker contrast, are dispersed on the Nb-SnO 2 support particles, several tens of nanometers in diameter with lighter contrast. The Pt catalyst nanoparticles are hemispherical in shape and have been shown to be highly oriented on the Nb-SnO 2 support, indicating a strong interaction between Pt and the support. 36 The average Pt particle size was estimated to be 2.9 ± 0.6 nm from the measurement of 500 particles in the TEM images, demonstrating good control of size within a narrow range. The Nb-SnO 2 nanoparticles were single crystallites (particle size: 20-30 nm), fused partially with nearest neighbors, constructing a network structure, a so-called "fused aggregate network structure." This network structure has the ability to construct both electronically conductive pathways, originating from the nanoparticle network, and gas mass transport pathways for reactants (oxygen, hydrogen), originating from open pores (diameter: 10-50 nm) surrounding the nanoparticle network. [32][33][34][35][36][37][38] These characteristics are desirable for improving the cell performance, as described in the next section.
Typical low acceleration voltage TEM images of Pt/Nb-SnO 2 and Pt/GCB sampled from the CLs are shown in Figs. 2a-2c. Thin layers of weak contrast (thicknesses of several nm) were observed on the surface of the Pt/Nb-SnO 2 and Pt/GCB agglomerates and were stable, without any discernable changes occurring during the TEM observation. These thin layers were determined to consist of Nafion ionomer and were found to cover the Pt/Nb-SnO 2 surface uniformly (Figs. 2a, 2b). In the case of Pt/GCB, the ionomer coverage was partial and nonuniform (Fig. 2c). In the case of Nb-SnO 2 , the surface can include oxygen vacancies and dangling bonds, which would bind to hydroxyl groups originating from water, due to its intrinsic hydrophilicity. The hydrophilic surface of the Nb-SnO 2 support has adsorbed hydroxide groups with negative charge, 47 which would interact with the positive charge of the sulfonic acid groups in the ionomer. The electrostatic interaction between the two groups would assist in constructing the uniform coverage of the ionomer on Pt/Nb-SnO 2 . The proton-conducting pathways are easily constructed on the Pt/Nb-SnO 2 and enhance the Pt utilization compared with that on Pt/GCB. The modification of the hydrophobic surface of a carbon nanotube support to a hydrophilic state was shown to enhance the Pt catalytic activity by improving the coverage of the Nafion binder. 48,49 The uniform ionomer coverage also has a beneficial effect in maintaining wide pores in the CLs. The cumulative pore volume in the primary pore region of Pt/Nb-SnO 2 CLs with I/S = 0.24, estimated by the Kr gas adsorption method (Fig. 2d), was higher than that with I/S = 0.67. The cumulative pore volume per gram of Pt/GCB with Nafion ionomer (I/S = 0.67) was found in previous work to be of the same order as that of Pt/Nb-SnO 2 (I/S = 0.67). 50 We conclude that the gas diffusion pathways were better developed in the CLs with I/S = 0.24 in comparison with I/S = 0.67.
The low acceleration voltage TEM images of the Pt/Nb-SnO 2 also indicate that, in the case of the image with I/S = 0.24, most of the tops of the Pt particles were covered with extremely thin ionomer layers on top. In contrast, in the case of I/S = 0.67, all of the Pt particles were buried in the ionomer. In the case of uniform ionomer coverage on the support, I/S is estimated from the Eq. 3: where S s is the surface area of the Nb-SnO 2 support (4πr s 2 , r s : radius of Nb-SnO 2 support) and d i is the estimated thickness of the ionomer on the Nb-SnO 2 support. From Eq. 3, the estimated ionomer thickness (d i ) is shown below: Nb-SnO 2 approached half the Pt particle size (ca. 1.5 nm), so that the tops of the Pt catalyst particles would be exposed to the gas phase, without a covering of ionomer. The low acceleration voltage TEM images of the Pt/Nb-SnO 2 of I/S = 0.24 (Fig. 2a) substantiated this absence of ionomer covering on the tops of the Pt nanoparticles and the construction of three-phase boundaries of gas phase/Pt/ ionomer, which would lead to the reduction of the overpotential due to gas diffusion in the ionomer. 51 We considered that the thin, uniform coverage of the ionomer on the catalyst would be desirable to reduce or even eliminate the oxygen diffusion resistance in the cathode CLs while maintaining the proton conducting pathways.

Single cell performance of MEAs using Pt/Nb-SnO 2 with varying ionomer content.-The ohmic loss (IR) included I-V curves
and cell resistances of the cells using Pt/Nb-SnO 2 cathode CLs (Pt amount; 0.05 mg cm −2 ) at 80 • C with ambient pressure air humidified at 80%RH and hydrogen are shown in Fig. 4a. In previous literature, Pt loadings have been in the 0.2-0.5 mg/cm 2 range, which is the present standard. However, the Nb-SnO 2 support discussed here is compatible with much lower Pt loadings (1/10 reduction of present Pt loadings,  i.e., 0.05 mg/cm 2 ), which are the target for the future fuel cell stack development over the next decade in Japan. Thus, we have unified the Pt loading at the value of 0.05 mg cm −1 and compared the performance of both Pt/Nb-SnO 2 and Pt/GCB CLs. The data were obtained at the beginning of testing (BOT), prior to the load cycle durability test. We found that the current density of the single cell using Pt/Nb-SnO 2 CL without any carbon additive as the conductive adjuvant was superior to that using commercial Pt/GCB CLs at each cell voltage. Each cell resistance using Pt/Nb-SnO 2 CLs was less than 0.2 cm −2 over 0.05 A cm −2 , which was lower than that using Pt/GCB CL. The low cell resistance using our Pt/Nb-SnO 2 CL would rely on the characteristics of microstructure of the Nb-SnO 2 support. The microstructure of the fused-aggregated network structures can supply electrically conductive pathways while diminishing the contact resistivity between nanoparticles of the support. [32][33][34][35][36][37][38] In the low current density region, less than 0.05 A cm −2 (cathode voltage >0.80 V), the electron donation from the Pt decreases due to the Pt itself becoming oxidized, 34,37 which leads to the slight increase of the cell resistivity.
The apparent mass activity (MA app , at 0.80 V) of the cell using each of the Pt/Nb-SnO 2 cathode CLs at BOT (Fig. 4b) increased with decreasing I/S and reached a maximum value of 305 A g Pt −1 at I/S = 0.12, which was 2 times higher than that using the Pt/GCB cathode CL with the optimized I/S ratio 0.67. The current density at 0.60 V of the cells using Pt/Nb-SnO 2 cathode CLs at BOT also increased with decreasing I/S and reached the value of the cell using a Pt/GCB CL with the optimized I/S ratio 0.67. The Tafel plots for each cell using Pt/Nb-SnO 2 CLs at BOT are shown in the inset of Fig. 4a. The plot of the cell using Pt/GCB CL at BOT (open circles) is also shown as a reference. Each of the Tafel plots exhibited a linear portion at potentials greater than 0.70 V; the Tafel slopes at BOT (inset of Fig. 4b) for the Pt/Nb-SnO 2 CLs with greater than I/S = 0.24 maintained values of the same order as that of Pt/GCB CL (89.8 mV decade −1 ). In the case of I/S = 0.12, the Tafel slope for the Pt/Nb-SnO 2 CL at BOT decreased to 80 mV decade −1 . From the observation of low acceleration voltage TEM images and the estimation of ionomer thickness (Figs. 2a-2b and Fig. 3), the tops of the Pt nanoparticles would be left uncovered, and the three-phase boundary of gas phase/Pt/ionomer would be better constructed at I/S = 0.12. Part of the cathode overpotential is known to originate from the oxygen diffusion in the ionomer. 51 An optimized three-phase boundary would reduce the gas diffusion overpotential in the ionomer and thus enhance the cell performance.

Single cell performance of MEA using Pt/Nb-SnO 2 with varying ionomer content at the end of testing.-The IR-included I-V curves
and cell resistances of the cells using Pt/Nb-SnO 2 cathode CLs at the end of testing (EOT: 50000 cycles) are shown in Fig. 5a. The I-V curve of the cell using a Pt/GCB cathode CL with the optimized I/S ratio 0.67 at EOT is also shown as a reference in the same figure.
The current density at each voltage improved with decreasing I/S and exceeded that of commercial Pt/GCB at I/S = 0.24 and or less. The significant increase in cell resistance for I/S values higher than 0.24 after the load cycle test is an important issue for the Pt/Nb-SnO 2 CLs. The Tafel plots at EOT showed that, for I/S greater than 0.36, the slope increased up to 200 mV decade −1 , indicating that either the oxygen diffusion or the proton conductance in the Nafion binder might have degraded during the load cycle test. Pt catalyst nanoparticles can easily dissolve at high Nafion binder content, even though a Pt band did not appear in the membrane. We need further investigation to elucidate the reasons for the resistivity increase, but we consider that soluble Pt cations might interrupt the proton conductivity in the Nafion ionomer and thus increase the cell resistivity. The MA app values at 0.80 V of the cells using each of the Pt/Nb-SnO 2 cathode CLs at EOT (Fig. 5b) were also 2 times higher than that using Pt/GCB with I/S = 0.67. The current density at 0.60 V of the cell using Pt/Nb-SnO 2 at EOT (Fig. 5b) also increased with decreasing I/S and approached the value obtained using Pt/GCB with I/S = 0.67.
The CV curves of the Pt/Nb-SnO 2 cathode CLs obtained at BOT are presented in Fig. 6. Peaks of both the adsorption/desorption of hydrogen and the oxidation of the platinum surface are clearly observed in these CVs. The initial values of ECA and U Pt of Pt/Nb-SnO 2 (I/S = 0.12, 0.24, 0.36, 0.45, 0.67) were more than 50 m 2 g −1 and 50%, respectively, significantly greater than those of Pt/GCB (22 m 2 g −1 , 27%). All Pt catalyst nanoparticles were located on the outer surface of the Nb-SnO 2 support, due to the lack of nanopores, and the ionomer covered uniformly the surface of the Pt/Nb-SnO 2 (Figs. 2a, 2b). However, in the case of Pt/GCB, 30% of the Pt nanoparticles on GCB were located in the interiors of nanopores of the GCB, which would lead to difficulty in making contact to the ionomer. 46,52 Also, the Nafion ionomer itself did not cover the Pt/GCB surface uniformly, as shown in the low acceleration voltage TEM image (Fig. 2c). The higher ECA and U Pt values obtained for Pt/Nb-SnO 2 in comparison with those of Pt/GCB originates from the uniform coverage of the ionomer and the high dispersion of Pt on the outer surface of the Nb-SnO 2 support particles. The CV curves of the Pt/Nb-SnO 2 cathode CLs at EOT are also shown in Fig. 6. The electric double layer current at EOT was maintained at nearly the same values as that of BOT, indicating that the Nb-SnO 2 support did not degrade during the load cycles. These adsorption/desorption currents for hydrogen and those for the oxidation/reduction of the platinum were smaller than those at BOT. The changes of ECA for Pt/Nb-SnO 2 and Pt/GCB during the load cycles are summarized in Fig. 7. For example, the ECA of Pt/Nb-SnO 2 (I/S = 0.12) and that of Pt/Nb-SnO 2 (I/S = 0.67) at EOT were 30.8 m 2 g −1 and 28.0 m 2 g −1 , which were about 51.0% and 55.2% of the values at BOT, respectively. The ECA of Pt/GCB (I/S = 0.67) at EOT was 11.2 m 2 g −1 , which was about 46.6% of the value at BOT. Figs. 8a-8c show the TEM images of Pt/Nb-SnO 2 (I/S = 0.12), Pt/Nb-SnO 2 (I/S = 0.67) and Pt/GCB (I/S = 0.67) at EOT. The mean Pt particle size of Pt/Nb-SnO 2 at EOT was 8.9 nm (I/S = 0.12) and 4.7 nm (I/S = 0.67). The Pt catalyst particles undergo enlargement due to Ostwald ripening during the load cycling, and diminution due to the dissolution of Pt. The particle size of Pt/GCB at EOT was 8.5 nm, which was rather larger than that at Pt/Nb-SnO 2 (I/S = 0.67). The decrease rates of the ECA of Pt/Nb-SnO 2 and Pt/GCB during the load cycling were similar, but the ECA values of Pt/Nb-SnO 2 were always higher than those of Pt/GCB. The MA app at 0.80 V of Pt/Nb-SnO 2 at BOT with I/S = 0.24 was higher than that of Pt/GCB (Fig. 4c). We conclude that the Pt/Nb-SnO 2 (I/S <0.24) had higher load cycle durability in comparison with that of Pt/GCB. Some of the Pt particles on GCB aggregated partially with nearest neighbor Pt particles, and the number of the Pt particles on Pt/GCB was less than that on Pt/Nb-SnO 2 . The cross-sectional Z contrast images of the MEAs using Pt/Nb-SnO 2 and Pt/GCB at EOT are also shown in Figs. 9a-9c. Very few Pt particles existed in the membranes of the MEAs using Pt/Nb-SnO 2 CLs in low I/S (Fig. 9a), whereas large numbers (white small particles, Pt bands) were detected in the membrane of the MEA using Pt/GCB (Fig. 9c). We concluded that the Pt/Nb-SnO 2 CLs has an ability to diminish the deposition of Pt band in the membrane. The dissolution of Pt was clearly diminished on the Nb-SnO 2 support in comparison with that on GCB from the observation of TEM images of the catalysts of EOT (Figs. 8a-8c). The hydrophilic surface of the Nb-SnO 2 support has a negative charge, 47 which would be expected to enhance the electrostatic attraction of dissolved Pt cations. Although further investigation of the Pt/Nb-SnO 2 surface at the EOT is necessary for a deeper understanding, the present results clearly suggest that the low I/S can help to diminish the dissolution of Pt, and the electrostatic attraction of the support surface can help to prevent the diffusion of Pt cations into the membrane.

Conclusions
The steady-state I-V performance and load cycle durability of PEFC using Pt/Nb-SnO 2 cathode CLs with varying I/S values were evaluated at 80 • C, 80%RH. The MA app values at 0.80 V (I/S >0. 20) were at the same level as that using Pt/GCB CLs with an optimized I/S ratio, and increased up to 2 times higher at I/S = 0.12 in comparison with that using the Pt/GCB CL. The current density at 0.60 V of the Pt/Nb-SnO 2 CL (I/S = 0.12) also reached the same value as that of the Pt/GCB CL. From the observation of low acceleration voltage TEM, it was confirmed that the Nafion ionomer covered the hydrophilic surface of Nb-SnO 2 uniformly in the CLs, and that the tops of the Pt nanoparticles were left uncovered, so that an effective three-phase boundary of gas phase/Pt/ionomer was constructed. The hydrophobic surface of the GCB was not covered uniformly by the ionomer, and the Pt particles on GCB were either not covered or were deeply buried under the ionomer. The three-phase boundary of the cathode using Pt/Nb-SnO 2 (I/S <0.20) clearly has desirable characteristics that lead to decreased oxygen diffusion overpotentials, and thus to improved cell performance.
The ECA values of Pt/Nb-SnO 2 for all I/S were higher than that of Pt/GCB at each cycle number during the load cycle durability evaluation, which indicates that the Pt/Nb-SnO 2 catalyst had high durability in comparison with that of Pt/GCB. Essentially, no Pt band was observed in the MEAs using Pt/Nb-SnO 2 CLs after the load cycle durability evaluation. The thin, uniform coverage of the Nafion ionomer on the Pt/Nb-SnO 2 surface appears to increase the Pt utilization and mitigate the degradation of the Pt catalyst.