A Nanoscale “Tune-Up” for Fuel Cells
So much of what scientists do is based on what we think is going on in the test tube, the flask, or the beaker. We make inferences of what actually occurred based on the results of an experiment and observations made before and after the action. But what if, just as the science fiction author Isaac Asimov detailed in his famous book Fantastic Voyage, we could shrink ourselves and walk among the atoms and molecules during the experiment, seeing first-hand what was happening?
Now we aren't shrinking people (yet), but researchers at the Energy Materials Center at Cornell (emc2), a DOE Office of Science-supported Energy Frontier Research Center (EFRC), are using a sophisticated microscope that allows them to sneak a peek at the very atoms and molecules that make up a hydrogen fuel cell and figure out how the cells work and how their performance degrades as they age. Their conclusions are already providing insights into how to build a better fuel cell, with the ultimate goal of using clean hydrogen, or small organic (and renewable) molecules to supply electricity for transportation and a range of other applications.
A fuel cell has some similarities to a battery. Both provide electricity that can, for example, run an electric motor. Unlike a battery, which needs to be recharged, usually over an extended period of time, the fuel cell generates its electricity on board, and can be repowered by a fast refilling of hydrogen (or another) fuel.
At their core, all fuel cells have three main parts: an electrolyte, an anode, and a cathode. The anode contains a catalyst that breaks the hydrogen fuel into protons and electrons. Beyond the anode is an electrolyte, usually a solid membrane, designed to allow protons to pass through while blocking electrons. The electrons are instead directed to an external circuit, providing the electric power to drive a motor or other applications. The circuit returns the electrons to the cathode side of the fuel cell, where they meet up with the protons that passed through the membrane. The cathode also contains a catalyst—one that is fed oxygen, usually from the surrounding air. Oxygen molecules, with the help of the catalyst, meet up with the free protons and electrons from the anode side and form water, the hydrogen fuel cell's only byproduct (other than some heat).
Splitting hydrogen molecules at the anode turns out to be relatively easy on the right catalyst surface. A much harder task is splitting the strong bond between oxygen molecules at the cathode. For this reaction, known as the oxygen reduction reaction, the catalysts we know of are notoriously inefficient (slow) and require a significant energy input (known as "overpotential" in fuel cell parlance) to split oxygen. This ultimately diminishes the amount of electricity (and the efficiency at which it is produced) that can be obtained from the fuel cell. Currently platinum-alloys are the best options for both of these catalysts. The high cost of platinum, however, is a major obstacle to the widespread adoption of fuel cells.
“Understanding exactly how the Pt3Co nanocatalyst ages is just as important as knowing that it does.”
Research into new platinum alloys, or combinations of platinum and less expensive metals, is yielding promising results. In particular, results from General Motors (GM) reveal that a platinum-cobalt nanocatalyst alloy, Pt3Co, shows improved power and efficiency when compared to platinum alone. Unfortunately Pt3Co is not very durable, as demonstrated in accelerated aging tests. The relatively rapid aging of the catalyst results in a short life span for the fuel cell. Thus, while the initial power output and efficiency of this new alloy are attractive, its rapid degradation remains a major barrier to making hydrogen fuel cells a viable option in cars and other applications.
Enter emc2, where researchers are developing novel materials for advanced fuel cells while also developing and applying new characterization tools to better understand these nanoscale materials. In collaboration with GM Electrochemical Energy Research Labs, emc2 began analyzing the Pt3Co material in hopes of providing new insights into how the cathode catalyst ages or degrades.
The aging or degradation occurs through a process called coarsening. The catalyst begins in the form of individual grains or nanoparticles with a diameter of around of 2-3 nanometers each (about 4 million times smaller than a centimeter) and, over time, ends up in the form of larger clumped particles that range in size to over 10 nanometers in diameter.
Now this change may seem subtle, but it turns out to have a big impact on how well the catalyst is able to do its job. Each of the Pt3Co grains has a certain amount of surface area to interact with oxygen to break it down so that it is available to react with the hydrogen's protons and electrons to form water. Think of it as using kids' building blocks to build a model house. Each block has 6 sides available to touch another brick. When you stack two together, one side for each block is now unavailable, so you end up with 10 sides open to touch another brick compared to the 12 total sides if the two bricks had remained separate. As you build towers into walls, some blocks may have 2 or fewer sides available. Now each Pt3Co nanocatalyst grain has a certain amount of surface area to react with oxygen. As the individual catalyst grains begin to clump together (coarsen), just like the building blocks, they are left with fewer available "faces," or less available surface area to react with oxygen. This reduces the speed and ease with which the catalyst can interact with the oxygen to produce water. This relatively rapid catalyst aging or degradation that results in catalyst clumping makes the fuel cell progressively less efficient and less powerful at producing electricity. A fuel cell that doesn't last very long is not a very helpful technology.
Understanding exactly how the Pt3Co nanocatalyst ages is just as important as knowing that it does.
"The two primary aging mechanisms that are well understood for platinum based electrocatalysts are Ostwald ripening and particle coalescence," says Héctor D. Abruña, the Director of emc2. "Ostwald ripening occurs when platinum atoms dissolve from the surfaces of smaller particles and redeposit onto others, creating larger particles. Particle coalescence describes how particles migrate and fuse into one larger particle." Both of these coarsening mechanisms are active in the fuel cells studied at emc2.
An everyday example of Ostwald ripening (named for the German chemist Wilhelm Ostwald, who first observed the phenomenon in the 1890s) is when water re-crystallizes within ice cream. As ice cream sits in your freezer, the smaller ice crystals within it melt and refreeze to form larger ones, giving the normally smooth, velvety texture of ice cream that unpleasant gritty, crunchy consistency.
Particle coalescence is a somewhat different process. Think of blowing soap bubbles into the air. Now and again two separate bubbles will collide in mid-air and fuse into a larger bubble. That's coalescence.
“Our approach dramatically improved the speed of data collection, letting us collect years' worth of data in a single day
”
David Muller
Professor of Applied and Engineering Physics, Cornell University
So there are two different paths to coarsening. In the end you have larger particles, but the kind of particles and the way you get there are different. Understanding whether Ostwald ripening or particle coalescence predominates, can help improve the fuel cell electrode structure and hopefully help to inhibit the aging, and thus performance degradation.
The researchers at emc2 have developed a way to measure these two kinds of catalyst nanoparticle aging. Using a very specialized version of a sophisticated tool called a scanning transmission electron microscope equipped with electron energy loss spectroscopy (STEM/EELS), the researchers are able to zoom in on the individual atoms that make up the catalyst. They have developed a way of using the microscope that allows them to literally look directly at what is happening at the atomic scale in 3-dimensions before and after the fuel cell goes through a simulated 30,000 start-stop cycles (the equivalent of driving approximately 100,000 miles in a car).
"Our approach dramatically improved the speed of data collection, letting us collect years' worth of data in a single day," says David Muller, a professor of applied and engineering physics at Cornell and a member of the emc2 team.
Muller has trained Cornell graduate students to use these advanced microscopy techniques for a variety of different applications. As part of emc2 and his collaboration with Abruña, he has adapted the techniques to study materials issues in fuel cells and batteries.
Yingchao Yu and Huolin Xin, graduate students working with both Abruña and Muller, designed the experiment. They took the small piece of metal with a grid pattern typically used in these microscopes as a sample holder and coated it with gold and then the Pt3Co nanoparticles. They then used the STEM/EELS to evaluate 300 of the nanoparticles for size, shape, and chemical composition. Each square in the grid is uniquely marked, enabling examination of the exact same catalyst material before and after aging. The researchers next used the grid in a unique way: they placed it in an electrochemical cell that replicates the fuel cell environment. After going through a simulated 100,000 miles in a car, the grid is returned to the microscope, and the researchers can then zero in on the exact same set of now aged catalyst nanoparticles. The researchers also produced 3-D images and videos of these catalyst nanoparticles before and after aging, providing clear evidence of how the nanoparticles changed during aging. (The videos are available at http://pubs.acs.org/doi/suppl/10.1021/nl203920s)
Yu and Xin used this experiment to look at degradation or aging of a fuel cell catalyst aged in the lab and another batch of catalyst aged in fuel cell stacks provided by the GM laboratory to see whether the stacks would age the same as particles in the lab. (Individual fuel cells produce small amounts of electricity, about 0.7 volts, so to increase the voltage and electricity output to meet an application's power needs, cells are "stacked," or placed in series or parallel circuits.)
Lab conditions, especially the upper voltage limit of the aging cycles, were well controlled to 1.0V, whereas the particles from the GM fuel cells may have seen voltages in excess of 1.4V during their aging cycles.
So what did they learn? On the lab grid, cycling caused the Pt3Co nanoparticles to increase in size as they expected, and based on the makeup of the nanoparticles the larger size was due to particle coalescence. However, with particles from the GM fuel stacks, they saw both coalesced particles and evidence of Ostwald ripening. Why would they be different?
The main reason lies in how the emc2 researchers controlled the experimental environment. With the electrochemical cell grid experiment, the upper voltage limit of the cycling did not exceed 1.0V through all 30,000 cycles. The fuel cell stacks, by contrast, saw greater variations in fuel and oxygen within the stack, especially during start up and shut down of the system, resulting in higher voltages at the cathode, possibly up to 1.4V. The differences in catalyst aging under controlled laboratory conditions versus real-life aging cycles was clearly distinguished and also quantitatively measured by this study. Note that these differences are virtually impossible to measure with traditional imaging techniques.
Based on this initial emc2 study, the team is now expanding the research to determine how to use this understanding to build a better fuel cell. For example, analysis of the results from the fuel cell stack suggested that one effect—particle coalescence—tends to reinforce the other—Ostwald ripening. Coalesced particles not only become larger through their coalescence; they also become favored targets for Ostwald ripening, which further enlarges them. This suggests that by decreasing catalyst particle coalescence, it may be possible to decrease Ostwald ripening, too. One way to do that is to improve the nanoparticle supports to prevent migration, then particle coalescence should be reduced and Ostwald ripening can be better controlled—potentially improving fuel cell durability.
Through new visualization techniques and improved cathode catalyst development, emc2 researchers are on course to develop a knowledgebase that will improve the durability of automotive fuel cell stacks, early steps toward the ultimate goal of an automobile operating on clean, efficiently converted hydrogen fuel.
—Dawn Adin, DOE Office of Science, Dawn.Adin@science.doe.gov
Funding
Research (Energy Materials Center at Cornell): DOE Office of Science, Office of Basic Energy Sciences
Microscopy Facility (Cornell Center for Materials Research): National Science Foundation
Publications
Y.C. Yu, H.L. Xin, R. Hovden, D. Wang, E.D. Rus, J.A. Mundy, D.A. Muller, H.D. Abruña. "Three-Dimensional Tracking and Visualization of Hundreds of Pt-Co Fuel Cell Nanocatalysts During Electrochemical Aging," NanoLetters Articles ASAP (2011).
H. L. Xin, J. A. Mundy, Z. Liu, R. Cabezas, R. Hovden, L. F. Kourkoutis, J. Zhang, N. P. Subramanian, R. Makharia, F. T. Wagner, and D. A. Muller, "Atomic-Resolution Spectroscopic Imaging of Ensembles of Nanocatalyst Particles Across the Life of a Fuel Cell," Nano Letters 12, 490 (2012).