504

DEVELOPMENT OF A 10 kW HYDROGEN/AIR PEM FUEL CELL STACK

♦Energy Partners, Inc. 1501 Northpoint Pkwy. #102 West Palm Beach, FL 33407

fW.L. Gore & Associates, Inc. 101 Lewisville Road P.O. Box 1100 EIkton,MD 21922-1100

PEM fuel cells have potential for meeting automotive industry's power density and cost requirements, such as 0.8 kW/kg, 0.8 kW/1 and $30/kW. For automotive applications, the fuel cell power requirements are in the 10-100 kW range. As the first phase in reaching this power output, a 10 kW PEM fuel cell stack has been developed at Energy Partners [1].

The stack consists of 50 cells with relatively large active area of780 cm2. The main feature of the stack is the advanced membrane electrode assembly (MEA) developed by W.L. Gore & Associates, Inc. These novel MEAs consist of a thin composite perfluorinated polymer membrane with a catalyst layer with platinum loading of 0.3 mg/cm2 on each side. The combination of reinforcement and thinness provides high membrane conductance and improved water distribution in the operating cell. In addition, the membrane has excellent mechanical properties (particularly when it is hydrated) and dimensional stability [2],

Two different gas diffusion layers were employed. A thin, soft, microporous, and hydrophobic gas diffusion material (GoreCarbel™) was applied directly on the back of the catalyst layer to protect the thin membrane as suggested by the MEA manufacturer. It provides much better electrical contacts at the surface of the catalyst layer, and may also contribute to water removal, as suggested by Wilson, el al., [3], A thicker, macroporous, Teflon-treated carbon fiber paper provides membrane support and makes contacts with the bi-polar collector plate.

The collector/separator plates are made of graphite in a proprietary two step compression molding process, thus avoiding expensive and time-consuming machining. The plates have the embedded hydrogen and air flow fields on opposite sides. Internal manifolds are placed around the perimeter of the plate. Cooling of the stack is provided by de-ionized water which circulates through cooling cells strategically distributed between the active cells. The current is drawn from the stack at two gold-plated copper bus plates placed at opposite ends of the stack. The stack is held together by two aluminum plates bolted together with tie-rods. All the connections for hydrogen, air and cooling water are provided on one of the end-plates. The stack weighs 69 kg (or 17.7 kg/m2) and has a volume of 38 liters.

The stack normally operates on hydrogen/air at 300 kPa (although it has been successfully operated at 170 kPa). The prescribed operating temperature (defined as the temperature of the cooling water outlet) is 65°C. Nominal power output of the stack is 10 kW, but a peak of 11.4 kW has been achieved during testing. Figure 1 shows the stack polarization curve at standard operating conditions. Operation is very stable, even at high power levels. The stack was operated at 9.5 kW for three 8-hour periods. More than 100 operating hours have been accumulated during testing.

0 100 200 300 400

current (Amps)

Figure 1 Performance of a 50-cell stack (H^Air, 300 kPa, 65°C)

0 100 200 300 400

current (Amps)

Figure 1 Performance of a 50-cell stack (H^Air, 300 kPa, 65°C)

The stack is capable of delivering >7 kW at cold start in less than two seconds, a feature of interest for automotive applications. The stack was also very responsive to load changes varying between 20% and 95% of the load (i.e., 2.0 and 9.5 kW). The responsivity of a fuel cell system in practical application will be limited by a compressor, not the fuel cell.

Due to a very thin membrane, water back-diffusion from the cathode to the anode may be higher than the electro-osmotic drag at certain current densities. During 2 hours of operation at 9.5 kW 11 kg of water accumulated at the cathode outlet, and 2.5 kg at the anode outlet, although hydrogen gas was not humidified prior entering the stack.

Nominal power (10 kW) is achieved at 410 mA/cm2 and 31.3 V (corresponding to 0.626 V/cell). The performance of the entire stack appears to be limited by a few cells that seem to suffer from severe mass transport problems (Fig. 2). Figure 3 shows the individual cell potential distribution at different current densities. A majority of the cells are well above 0.65 V at 410 mA/cm2, thus having potential of achieving much higher power densities (>0.36 W/cm2).

current density (mA/cm2) Figure 2 Performance of individual cells in 50-cell stack
Figure 3 Cell potential distribution in 50-cell stack at different current densities

There is a big discrepancy in published data on the performance of PEM fuel cells in laboratories and in practical applications. Significant development efforts are needed to cross the gap from a small single cell in a controlled environment to a stack of large active area cells operating in a practical application, such as a fuel cell powered vehicle. A variety of technical problems must be addressed in developing a large fuel cell stack, such as:

• sealing of the cells (particularly around the manifolds),

• uniform reactants supply to each cell,

• uniform reactants supply over large active areas inside each cell,

• product water removal,

• electrical contacts over large areas.

Development of the 10-kW stack was accompanied by the experiments with a 150 cm2 single cell, using the same MEA (Gore PRIMEA™). As opposed to a cell in a multi-cell stack, the single cell experiment allows operation at controlled conditions, i.e., flow rate, pressure temperature and humidity of both reactant gases. Figure 4 shows the performance of a single cell (150 cm2) with hydrogen/air at 60 °C and 300 kPa.

Power density of -0.5 W/cm2 was achieved at 0.6 V. When the hydrogen flow rate was increased the power density at 0.6 V increased up to >0.6 W/cm2 (max. power density was 0.75 W/cm2 but at lower voltage). It was also noted that the cell potential responded to the changes in hydrogen pressure much more than what should be expected just based on the Nernst equation. Such sensitivity to hydrogen flow rate and pressure indicates that the cell performance is affected by net water transport across the membrane from the cathode to thé anode. Higher pressure on the hydrogen side apparently reduced the net water flux, and higher flow rate helped removing water from the anode. This was confirmed in operation of the 50-cell stack which needed constant purging of the hydrogen side.

0 200 400 600 800 1000 1200 current density (mA/cm2)

Figure 4 Performance of a single 150 cm2 cell (H^Air, 300 kPa, 60°C)

0 200 400 600 800 1000 1200 current density (mA/cm2)

Figure 4 Performance of a single 150 cm2 cell (H^Air, 300 kPa, 60°C)

With the performance currently achievable in single cell (i.e., >6 kW/m2) the next generation Energy Partners' fuel cell stacks (with projected weight of <12 kg/m2) will reach power density of >0.5 kW/kg and 0.8 kW/1, thus coming another step closer to the automotive industry requirements.

0 0

Post a comment