This shows the catalyst sample in the ultrahigh-vacuum chamber of the photoemission electron microscope. Photo: TU Wien.
This shows the catalyst sample in the ultrahigh-vacuum chamber of the photoemission electron microscope. Photo: TU Wien.

The taste of the icing on a chocolate cake should not depend on whether it is served on a porcelain plate or a silver plate. Similarly, for chemical reactions on the surface of large precious metal grains, the substrate (the so-called support) should not play a crucial role. The catalytic grains often have a diameter spanning many thousands of atoms and so the support on which they rest should not affect chemical reactions on the other side, far away from the interface between them – at least that was always thought to be the case.

Now, experimental studies performed at the Vienna University of Technology (TU Wien) in Austria have shown that assumption to be incorrect. These studies found that chemical processes on palladium grains, which are used in exhaust gas catalysts, changed significantly when they were placed on specific support materials – despite the fact the support material is nearly inactive in the chemical reaction itself. This novel insight is reported in a paper in Nature Materials.

For vehicles using an internal combustion engine, toxic carbon monoxide (CO) must be converted into carbon dioxide (CO2), which is achieved using catalysts that contain palladium or platinum powder. "We have investigated chemical reactions on powder grains, which are often used in industrial catalysis," says Günther Rupprechter from the Institute of Materials Chemistry at TU Wien. "The precious metal grains have a diameter on the order of 100µm – this is very large by nanotechnology standards, one can almost see them with the naked eye."

When the surface of the powder particles is covered by oxygen atoms, CO molecules react with them to form CO2, leaving empty sites (holes) in the oxygen layer. These sites must be quickly filled by other oxygen atoms to sustain catalysis, as they can also be filled by CO molecules. If this happens on a large scale, the catalyst surface becomes covered by a CO layer, rather than an oxygen layer, preventing the formation of any more CO2. This phenomenon is called ‘carbon monoxide poisoning’ and deactivates the catalyst.

Whether this happens or not depends on the CO concentration in the exhaust gas supplied to the catalyst. But the current study now reveals that the support material on which the palladium grains are placed is also crucial. "If the palladium grains are placed on a surface of zirconium oxide or magnesium oxide, then poisoning of the catalyst occurs at much higher carbon monoxide concentrations," says Yuri Suchorski, also from the Institute of Materials Chemistry at TU Wien and first author of the paper.

At first glance, this is surprising behaviour for such large palladium grains. Why should the nature of the support have an effect on chemical reactions that take place on the surface of the entire metal grain? Why should the contact line between the palladium grain and substrate, which is only a few tenths of a nanometer wide, influence the behaviour of palladium grains that are a hundred thousand times larger?

This puzzle was finally solved with the help of a special photoemission electron microscope at the Institute of Materials Chemistry at TU Wien. With this device, the spatial propagation of a catalytic reaction can be monitored in real time. "We can clearly observe that carbon monoxide poisoning always starts at the edge of a grain – exactly where it contacts the support," explains Suchorski. "From there, the ‘carbon monoxide poisoning’ spreads like a tsunami wave over the whole grain."

It is mainly for geometrical reasons that the poisoning wave starts exactly there: the oxygen atoms at the border of the grain have fewer neighbouring oxygen atoms than those on the interior surface. When free sites open up there, it is easier for a CO molecule to populate these sites than sites somewhere in the middle of the free surface, where CO would easily react with surrounding oxygen atoms. In addition, it is not easy for other oxygen atoms to fill vacant areas at the border, since oxygen atoms always come in pairs, as O2 molecules. To fill an empty site, O2 needs two free sites next to each other, and there is not much room for this at the border.

The border where the palladium grain is in direct contact with the support is therefore of great strategic importance – and it is exactly at this interface that the support can influence the properties of the metal grain. "Calculations by our cooperation partners from the University of Barcelona show that the bond between the metal atoms of the grain and the protective oxygen layer is strengthened precisely at the borderline to the support," says Rupprechter. The palladium atoms in intimate contact with the oxidic support can thus form stronger bonds with oxygen.

One may assume that this does not matter for metal sites far away from the border of the grain, because the support can only energetically influence atoms at the border – and there are only very few of these compared to the total number of atoms in the palladium grain. However, because carbon monoxide poisoning starts at the border, this effect is of great strategic importance. The metal-oxide border is in fact the ‘weak point’ of the grain, and if this weak point is reinforced by the support, the entire micrometer-size catalyst grain becomes protected from carbon monoxide poisoning.

"Various oxide supports are already used in catalysts, but their exact role during catalysis in terms of CO poisoning has not yet been directly observed," says Rupprechter. "With our methods, the ongoing process and its wave-like long-range effect were directly visualized for the first time, and this opens up promising new routes towards improved catalysts of the future".

This story is adapted from material from TU Wien, with editorial changes made by Materials Today. The views expressed in this article do not necessarily represent those of Elsevier. Link to original source.