SSR

SSR - Surface Structure and Reactivity Group

research
  projects and collaborations

Research activities

Scanning Tunneling Microscopy Laboratory

In recent years, the formation and reactivity of surface oxides have received a growing interest within the surface science community, since they might play a fundamental role in catalysis under real operation conditions. Transition metal nanoparticles, which are commonly used in real catalysis, offer a wealth of defects that may easily undergo 1D oxidation under operating conditions. A thorough study of 1D oxide reactivity is therefore crucial for the understanding of the nanoparticle reactivity.

In a previous study we already investigated the reactivity at RT of a 1D oxide formed on Rh(110), i.e. the (10x2)-O structure. In that work we evidenced that upon exposure of the surface to molecular hydrogen, the water formation reaction proceeds in two steps. In a first step a reaction front propagates, starting from wide unreconstructed areas, removing half of the O atoms initially present, and leaving one O atom out of two adsorbed along the metal segments in a zigzag fashion. In a second step, the reaction is completed starting homogeneously on the surface from the segment ends towards the middle.

Here we show that the reactivity of the 1D oxide is driven by the local geometry of the atoms along the segments, which is in turn determined by the initial lattice expansion and its relaxation during the reaction.
We expose the 1D oxide to H₂ in the 238-263K temperature range. The reaction front is now clearly not uniform: it crosses the surface in a 'comblike' fashion, following the segmentation of the first metal layer. Zooming on the segments, we see that the comblike shape is related to a limited propagation of the reaction along the expanded oxide structure, i.e. starting from the ridge vacancy channels, it expands onto the adjacent segments but it stops in the middle of each segment.

Subsequent images extracted from an STM movie acquired at 253K with an acquisition time of 35s/image. The comblike front is indicated by the arrows.

Zoom on the segments immediately after the passage of the reaction front in the upper channel. It can be clearly seen that only the upper halves of the segments in the central part of the image have been involved in the reaction.

We repeat the experiment at RT by following the reaction at a frame rate of 30Hz with our new Fast-STM module. Indeed we see that the propagation of the reaction front follows the same 'comblike' pathway also at this temperature but at a much faster speed than in the low T measurements.

Snapshots from a Fast-STM movie acquired at room temperature with an acquisition time of 33ms/image. The reaction is propagating along the lowest RV channel.

This peculiar 'comblike' shape for the reaction front is therefore peculiar of the reactivity of the Rh 1D oxide.
DFT calculation, performed in collaboration with the group of Georg Kresse (University of Vienna), support this mechanism that can be traced back to a progressive increase of the activation barrier for OH formation as the reaction proceeds moving towards the middle of a segment. Such an increase is related to a longer O - H distance that stabilizes the initial state.

In the low temperature range, the second reaction step is never completed, with chains of zigzag O atoms remaining in the central part of the segments, leading to a regular low-coverage oxygen pattern. This pattern might result in an ordered array of active sites for partial oxidation reactions of large molecules.

Regular pattern of O atom chains after the second reaction step.

Ref:
C. Africh, L. Köhler, F. Esch, M. Corso, C. Dri, T. Bucko, G. Kresse and G. Comelli; Effects of Lattice Expansion on the Reactivity of a One-Dimensional Oxide; J. Am. Chem. Soc. 131, 3253-3259 (2009).

Participants: C. Africh, F. Esch, M. Corso, C. Dri, G. Comelli