Fuel-cell APUs: A STEP TOWARD ALL-ELECTRIC CARS

Delphi's 5-kW solid-oxide fuel-cell auxiliary power unit


SOFCs make electricity from fuel and oxidant gases in an electrochemical process that takes place across an ion-conducting, ceramic membrane. Reformed fuel (reformate) feeds to an anode, and air, to a cathode. A solid electrolyte separates the two electrodes.


The first test of Delphi's APU based on Generation 3 2 30-cell stacks. The APU went through a fully automated cold start, warm-up, and produced 1,617 W on gasoline.


Generation 3 30-cell stack and its electrical output at 750°C with simulated reformate (20% H2, 23% CO, 3% H2O, rest N2).


Automakers are increasingly looking at electric motors as a replacement for the venerable internal-combustion engine (ICE). Candidate power sources for electric prime movers include batteries and fuel cells. Of the two, fuel cells have emerged as the more viable-option. Batteries are bulky and heavy, provide severely limited range per charge, and take a long time to recharge.

Fuel cells, in contrast, convert hydrogen and air to electricity and exhaust mostly water vapor. Only the onboard supply of fuel limits vehicle range. However, fuel-cell-powered, all-electric cars aren't a panacea. Batteries must still provide some amount of energy storage. And no infrastructure yet exists for hydrogen refueling. This is especially problematic for proton-exchange membrane fuel cells (PEMFCs) that need relatively pure hydrogen to work.

PEMFCs are widely regarded as the technology of choice for automotive propulsion. PEMFCs run at about the temperature of boiling water and need a separate reformer to break down liquid fuels such as methanol and gasoline into hydrogen gas of suitable purity. The reformers tend to be large and complex, which is not necessarily a showstopper for the high-power PEMFC systems envisioned for all-electric cars. In fact, simple economics is a much bigger roadblock to their commercialization. Allelectric drivetrains are a tough sell because internal-combustion engines are highly refined and relatively inexpensive to build per unit of power delivered.

A parallel trend in automotive design — and one that could pave the way for fuel cells — is electrification. Many cars already come with intermittent electrical accessories and systems such as heated seats, heated windshields, power steering, and brakes. These may soon be joined by higher power, "continuous" systems including electric air conditioning and electromagnetic valve trains. Today, enginedriven alternators provide onboard electrical power. But these 14-V systems are being pushed to practical limits. Evolving dualvoltage 42/14-V systems based on the 42-V PowerNet spec will help meet the added power demands. This is where fuel cells come in, say Delphi Automotive Systems, Troy, Mich., and partner, German automaker BMW.

Delphi's prototype solid-oxide fuel cell (SOFC) auxiliary power unit (APU) is sized to generate 5 kW of electrical power at 42 Vdc. The APU would replace alternators and work with the engine on or off, a plus for military, worksite, and recreational vehicles that need autonomous electrical power. In practice, vehicles initially would retain their conventional ICE drivetrains. Delphi says the approach represents an easier transition to fuel cells than the "all-or-nothing" proposition of allelectric cars. The fuel cells could as well provide electrical power for hybrid vehicles that join electric motors with small ICEs.

WHY SOFCs?
SOFCs are highly efficient (>50% fuel-to-electric conversion efficiency in some cases). They are compatible with conventional hydrocarbon fuels such as gasoline, methanol, natural gas, and diesel fuel, when combined with a relatively simple partial-oxidation reformer.

SOFCs make electricity from fuel and oxidant gases in an electrochemical process that takes place across an ion-conducting, ceramic membrane. Reformed fuel (reformate) feeds to an anode, and air, to a cathode. A solid electrolyte separates the two electrodes.

There are two principal SOFC configurations, tubular and planar. Tubular designs bundle tube-shaped cells in parallel. One end of the tubes is closed, and the other hooks to a common airsupply manifold. Oxygen ions pass through the electrolyte and react with fuel flowing over the outside of the tubes, creating an electric current.

Planar types, in contrast, place individual cells in a multilayered "stack." The Delphi APU uses this arrangement. Each rectangular-shaped cell contains an anode, electrolyte layer, and a cathode, and produces about 0.7 Vdc. Cells are arranged in series or parallel to give the required voltage and current. Separator plates electrically connect the cells and direct fuel and air to them from gasdistribution manifolds. The stack mounts in a bolted frame whose thermal expansion closely matches that of the stack itself over the operating temperature range, eliminating the need for an active loading system. Compressible seals in the stack separate the reactant gases and accommodate thermal cycling and mechanical vibration. The cells are of an anode-supported design, as opposed to a cathode or electrolytesupported configuration. Of the three, anode-supported cells have been shown to give the highest power density at low operating temperatures.

LOWER OPERATING TEMPERATURES
SOFCs typically operate at temperatures of about 900 to 1,000°C, high enough to internally reform hydrocarbon fuels. But advances in porous ceramics and manufacturing techniques have helped to push temperatures to a much lower 600 to 800°C. Operation at these lower temperatures requires a separate catalytic partial-oxidation reformer, but permits the use of less-expensive metal alloys for fuel-cell separators, interconnects, and balance of plant components. Lower operating temperatures also extend cell life.

Delphi's latest Generation 3 30-cell stacks, for example, are designed to operate at about 750°C. They use separators and interconnects made of a special alloy called Cro Fer 22 APU. In prototype units, cell components are brazed to "picture frames" made of the alloy. Cell/picture-frame assemblies are then laser welded to cell separators. The techniques could likewise scale to volume production.

Cro Fer 22 APU is characterized by a small thermalexpansion coefficient and the ability to remain electrically conductive at temperatures to 900°C. High chrome content helps the alloy match its thermal-expansion rate to that of the surrounding ceramic components. Elevated temperatures normally would cause some of the chrome to evaporate and penetrate the cathode up to the electrolyte interface, lowering cell efficiency. But the addition of such ingredients as lanthan, manganese, and titanium, let the alloy "self-seal" at high temperatures. Here, a protective layer of electrically conductive chrome manganous oxide grows on the surface of the interconnectors and helps prevent chrome evaporation. Furnace-based thermal cycle tests to 750°C of Generation 3 15-cell stacks show a minimal drop in power density. And a 250-hr durability test of interconnects on single cells at the same temperature show little power degradation after the initial 50 hr.

The APU's catalytic partial-oxidation reformer has been the subject of ongoing research as well. The reformer turns liquid fuel into small droplets, heats them, then catalytically converts the stream primarily into gaseous CO and H2, and less than a percent each of methane, ethylene, and ethane. Like the SOFC stack, the reformer must be capable of rapid cycling without a significant drop in performance over time. Generation 2 tubular reformers in cyclic tests reach a steadystate output of about 20% H2 gas in under 3 min from startup.

PUTTING IT ALL TOGETHER
Managing these systems is the job of an electronic control unit. In operation, an increase in electrical load signals the controller to boost flow rates of reformed fuel and air. Feedback comes from exhaust-gas analyzers and temperature and pressure sensors. The control system resides in the same package as the reformer and SOFC stacks. This makes it necessary to separate the package into two basic zones: a hot-zone module and a plant-support module. An insulated "hot box" houses the fuel-cell stack, the reformer, and a waste-energy recovery (WER) system. Output from the fuel reformer and an integrated heat exchanger feeds to the stack anode.

The electronic-control unit, sensors, and actuators sit in the relatively cooler (125°C design temperature) plant-support module along with an electric blower fan that supplies air to the stack cathode, purging, and cooling systems. Both the anode and cathode exhaust to the WER. A 40-V lithium-ion battery operates the automobile electrical accessories when the SOFC is coming up to temperature, and supplies power to the cathode air blower, SOFC sensors and actuators during startup and cool down.

NOT QUITE READY FOR PRIMETIME
Delphi has made considerable progress towards a commercially viable automotive APU, though much work remains to reach DOE performance objectives. For example, a recent test of a Generation 3 30-cell stack produced a stack power density of 308 mW/cm2, considerably less than the DOE target of >1 W/cm2. On the upside, Generation 3 stacks consume a volume of just 3.5 liters and weigh 28.6 lb (13 kg), a significant improvement over earlier Generation 2 stacks (6 liters and >20 kg).

Another key metric is start-up time. DOE projections say the SOFC portion of the APU should reach its operating temperature range of 600 to 800°C in under 2 min. Generation 2 stacks, for reference, take about 45 min.

Shrinking APU size and weight are equally important objectives. First-generation units (circa 2000) consumed 155 liters and weighed a hefty 450 lb (204 kg). For comparison, a Generation 2 APU consumes just 60 liters and tips the scales at 154 lb (70 kg). Ongoing work on Generation 3 APUs aims to further lower these numbers.

Meanwhile, a BMW 7 Series fitted with a prototype APU is undergoing testing. Delphi says the APU's primary target is luxury cars with a high electrical demand, which represents about 1 to 2% of the vehicle market. The technology could also work in conventional superultralow emissions vehicles (SULEVs), ships, portable, and stationary power systems.

MAKE CONTACT
Delphi Automotive Systems,
www.delphi.com