Catalysis: A 30-year-old explanation overturned

In the Fischer–Tropsch synthesis, a gas mixture of carbon monoxide and hydrogen (“syngas”) reacts to give liquid hydrocarbons that serve as synthetic fuels. In the modern industrial process, the reaction runs at 20 to 40 bar and around 200 °C over a catalyst containing metallic cobalt. As syngas can also be produced from biomass or from CO2 and green hydrogen, the process is considered a possible building block for climate-neutral fuels. A team led by LMU chemist Professor Joost Wintterlin has now joined forces with theoreticians at Ulm University to determine what the surface of the operating catalyst actually looks like.

A common explanation put to the test

Under reaction conditions, the cobalt surface is heavily covered with carbon atoms released when carbon monoxide decomposes. Based on a study in the 1990s, the field has considered it likely that the carbon atoms cause a roughening of the smooth metal surface: small cobalt islands form, and their edges serve as the active sites of the reaction. The idea is appealing, as it can explain why fresh catalysts become more active during the first few days of operation. Until now, there has been no way of verifying this directly. There is no surface analysis method that works at 20 to 40 bar.

A carbon reservoir as a detour

The group uses a scanning tunneling microscope housed in a reactor cell, producing atomically resolved images at operating pressures of up to about 1 bar. The products formed are analyzed at the same time by gas chromatography. To reproduce the high carbon coverage of the industrial process, despite the lower pressure, first author Sebastian Kläger took a detour: using a cobalt single crystal as a model catalyst, he deposited a thin layer of cobalt carbide that acts as a carbon source. When the reaction was then started with roughly 1 bar of syngas at 200 °C, some of the carbide rapidly reacted away with hydrogen, but a portion of the carbon remained on the surface and kept the coverage at levels that would otherwise only be expected at far higher pressures.

Triangles instead of islands

However, the expected roughening did not occur at all. Instead, the images showed patterns of equilateral triangles, the smallest with edge lengths of 1.5 nanometers.

Dr. Sung Sakong and Professor Axel Groß at Ulm University, applying quantum chemical calculations, identified the cause of this pattern. Within the triangular regions, the cobalt atoms of the topmost layer form what are known as stacking faults. Along the lines between the stacking faults and the regular regions outside the triangles, the carbon atoms find particularly favorable sites. According to the calculations, this arrangement is considerably more stable than the roughened state assumed so far and remains so under the conditions of the technical process.

“Roughening has been the standard explanation for thirty years. At elevated carbon coverages, we don't see it, and the calculations also tell us why: the triangular structure is simply more favorable,” says Wintterlin.

The activity remains unchanged

Remarkably, the catalytic Fischer–Tropsch activity is not affected by this structure. “The surface looks different, but the catalyst works just the same. That fits into our picture that what matters are the atomic steps, and their number barely changes,” says Kläger. The group had already shown in earlier work that the step edges between the terraces are the active sites. What the data do not support is the widespread explanation of the activity in terms of a carbon-induced roughening of the industrial catalyst.

The work was funded by the German Research Foundation (DFG, project number 552794718).

S. Kläger et al.: A Carbon-Induced Surface Reconstruction on an Operating Fischer–Tropsch Catalyst. Angewandte Chemie International Edition 2026, e5104870.