Silicon surface alloys and silicide nanolayers are highly important in heteroepitaxial growth, in heterogeneous catalysis, and as contact materials in integrated circuit devices. Recently, surface alloys have been reported to act as a precursor for growth of 2D materials such as silicene. In particular, although Si is immiscible in bulk Ag, the Si/Ag surface exhibits distinct Si/Ag exchange reactions and leads to the formation of unique nanostructured surface geometries with interesting physical properties. In this context, the Si/Ag(001)-(3x3) surface reconstruction has recently been proposed as a physical manifestation of the 2D Su-Schreiffer-Heeger (SSH) model of topological properties. However, the geometry of the surface remains unclear, despite a wide range of experimental and theoretical studies.
In this joint experimental-theoretical work, published in Nature Communications, the researchers carried out high resolution STM measurements allowing well defined criteria (symmetry, coverage, etc) for possible surface geometries to be established. Then, using DFT calculations based on particle swarm optimization techniques to explore a huge range of geometries and stoichiometries, a number of possible models were identified. Of these, only one (thermodynamically stable) model succeeded in explaining the bias-dependent STM measurements and previously published XRD and ARPES data. The model, dubbed "surface-alloy double-pentagon chain" by the authors, features unbroken chains of silicon in a double-pentagon shape, reflecting similar findings on the Si/Ag(110) surface. In this case however, the chains are shown to possess a quasi-one dimensional character thanks to the presence of Ag atoms in the surface layer, that define the structure as a non-pseudomorphic surface alloy. The findings have important consequences for realizing a true 2D SSH model on surfaces and highlight the pervasiveness of pentagonal geometries in 2D materials.
The work is a collaboration between researchers from CNR-ISM and Univ. Rome "Tor Vergata" in Italy and from the CNRS, Sorbonne and Aix Marseille Universities in France, and financed in part through the TeraExc MSCA.