Fuel-Cell Power-Up

In her lab at MIT, chemical-engineering professor Paula Hammond pinches a sliver of what looks like thick Saran wrap between tweezers. Though it appears un­remarkable, this polymer membrane can significantly increase the power output of a methanol fuel cell, which could make that technology suitable as a lighter, longer-­lasting, and more environmentally friendly alternative to batteries in consumer electronics such as cell phones and laptops.

Methanol is a promising energy source for fuel cells because it is a liquid at room temperature, so it’s easier to manage than hydrogen. But so far, its commercial applications have been limited. One reason has to do with the properties of the proton-conducting membranes at the heart of fuel-cell technology.

On one side of a methanol fuel cell, a cata­lyst causes methanol and water to react, yielding carbon dioxide, protons, and free electrons. The protons pass through a membrane to a separate compartment, where they combine with oxygen from air to form water. The electrons, which can’t cross the membrane, are forced into wires, generating a current that can be used to power electronic devices.

The more protons cross the membrane, the more power is generated. But the polymers that conduct protons well also tend to let the methanol solution into the other compartment. The resulting loss of fuel lower­s the cells’ power output. To limit such “metha­nol crossover,” researchers have to either use polymers that don’t conduct protons as well or make thicker membranes. But both of those options decrease efficiency, too.

In work published last spring in Advanced Materials, Hammond used an elegant, inexpensive process to reduce methanol crossover in a commercial fuel-cell membrane, increasing the efficiency of a methanol fuel cell by more than 50 percent. “What we’ve done is generate a very thin film that actually prevents the permeation of methanol but at the same time allows a rapid rate of proton transport,” says Hammond. Encouraged by this success, her team is now working to build such membranes from scratch, which could make them less expensive.

A Modified Process
A layer-by-layer assembly technique is the key to Hammond’s membranes. In earlier work, her team altered a membrane made of Nafion, a polymer manufactured by DuPont that is commonly used in fuel cells. It conducts protons well but also permits some methanol leakage, and it’s rela­tively expensive to make.

To begin the new process, Avni Argun, a postdoc in the lab and lead author on the Advanced Materials paper, mounts a specially treated silicon disc in a lab hood and starts the disc slowly rotating. Facing the membrane are four sprayer ­nozzles. Each nozzle is connected to a separate container. One contains a positively charged polymer solution and one a negatively charged polymer solution; two hold water.

Argun starts the sprayer system, which mists the disc with the positive solution for a few seconds, then with a water rinse, then with the negatively charged polymer, and finally with water again. A two-layer film forms within about 50 seconds. The thickness of this “bilayer” depends on the polymers and can range from 3 to 50 nanometers. In about six hours, the sprayer can apply between 400 and 600 bilayers, creating a membrane about 20 micro­meters thick. The membrane described in Advanced Materials was made up of three bilayers on top of a Nafion membrane, adding only 260 nanometers to its thickness. By using a combination of positive and negative polymers, the researchers maintained Nafion’s high conduc­tivity while reducing its methanol crossover.

Other researchers have tried to reduce membrane permeability by using new polymers or blending two different polymers. Blending often doesn’t work well, though, because polymers with different structures tend to separate, making the membrane less stable. With the layer-by-layer assembly process–common in other areas of materials science–“we combine two different materials, but on a nanometer-length scale so they’re really intermingled,” Hammond says.

Testing Grounds
After the membrane dries, Argun carefully peels it off the disc and tests its permeability and electrical resistance, which allows him to calculate its conductivity. With a large clip, he fastens the membrane between a plastic chip and a base that holds platinum wires that will measure resistance. After putting the assembly in a sealed plastic box that allows him to control temperature and humidity, he manipulates the membrane using a pair of gloves that reach through the box and into the chamber. Most membranes perform bette­r under high temperature and humidity, so both conditions must be noted. Argun connects the assembly to an external analyzer to test the membrane’s resistance. Measuring its permea­bility is more straightforward; he simply notes the amount of methanol that diffuses through it over a specific amount of time.

If a membrane fares well in these initial tests, Argun couples it to a positive and a negative electrode (where the electricity-­producing reactions take place) to see how it would perform in an actual fuel cell. He places the electrodes–two black, circular carbon cloths studded with particles of plati­num and a metal alloy–on either side of the membrane. Then he sandwiches the whole apparatus inside an insulating gasket that looks like thin cardboard. Finally, he seals the unit using a hot press.

Graduate student Nathan Ashcraft takes over from here. Ashcraft puts the membrane-electrode assembly into an active fuel cell, into which air and methanol are carefully pumped. Two square slabs of steel, about the size of slices of bread, make up the outside of the cell; they contain heaters that allow Ashcraft to precisely control the temperature of the reaction. Between the steel slabs, two gold-plated electrodes sandwich graphite blocks with small channels etched into them. Ashcraft places the membrane-electrode assembly between the blocks and secures it with screws. He then pumps methanol and air through the channels to either side of the assembly. He measures and records the resulting current, along with the system’s temperature.

Hammond’s team has not yet devised a completely new membrane that conducts as well as Nafion. However, “we feel like we’re very close,” she says. The team is also experimenting with membrane thickness; if a membrane is too thin, it will tear in the fuel cell, but thicker membranes don’t conduct protons as well. The membrane that the lab ends up with will probably be about 50 micrometers thick, Ashcraft says. ­Hammond also plans to try building membranes that incorporate additional polymers.