Activity of fuel cell catalysts doubled

In fuel cells, hydrogen reacts with oxygen to produce water, generating electricity in the process. Sophisticated catalysts at the electrodes are required in order to optimize this conversion. Platinum plays a central role in the oxygen-reduction reaction.

Searching for an ideal solution, the team created a computer model of the complete system. The central question: How small can a cluster of platinum atoms be and still have a highly active catalytic effect? "It turns out that there are certain optimum sizes for platinum stacks," explains Fischer.

Particles measuring about one nanometer and containing approximately 40 platinum atoms are ideal. "Platinum catalysts of this order of size have a small volume but a large number of highly active spots, resulting in high mass activity," says Bandarenka.

Interdisciplinary collaboration at the Catalysis Research Center (CRC) was an important factor in the research team's results. Combining theoretical capabilities in modelling, joint discussions and physical and chemical knowledge gained from experiments ultimately resulted in a model showing how catalysts can be designed with the ideal form, size and size distribution of the components involved.

In addition, the CRC also has the expertise needed to create and experimentally test the calculated platinum nano-catalysts. "This takes a lot in terms of the art of inorganic synthesis," says Kathrin Kratzl, together with Batyr Garlyyev and Marlon Rück, one of the three lead authors of the study.

The experiment exactly confirmed the theoretical predictions. "Our catalyst is twice as effective as the best conventional catalyst on the market," says Garlyyev, adding that this is still not adequate for commercial applications, since the current 50 percent reduction of the amount of platinum would have to increase to 80 percent.

In addition to spherical nanoparticles, the researchers hope for even higher catalytic activity from significantly more complex shapes. And the computer models established in the partnership are ideal for this kind of modelling. "Nevertheless, more complex shapes require more complex synthesis methods," says Bandarenka. This will make computational and experimental studies more and more important in the future.