A comprehensive picture of the initial stages of silicene growth on graphite is drawn. Evidence is shown that quasiparticle interferences play a crucial role in the formation of the observed silicene configurations. We propose, on one hand, that the charge modulations caused by those quantum interferences serve as templates and guide the incoming Si atoms to self-assemble to the unique ( $\sqrt{3}\times \sqrt{3}$)R30° honeycomb atomic arrangement. On the other hand, their limited extension limits the growth to about 150 Si atoms under our present deposition conditions. The here proposed electrostatic interaction finally explains the unexpected stability of the observed silicene islands over time and with temperature. Despite the robust guiding nature of those quantum interferences during the early growth phase, we demonstrate that the window of experimental conditions for silicene growth is quite narrow, making it an extremely challenging experimental task. Finally, it is shown that the experimentally observed three-dimensional silicon clusters might very well be the simple result of the end of the silicene growth resulting from the limited extent of the quasi-particle interferences.

The fascinating properties of two dimensional (2D) crystals have gained increasing interest for many applications. The synthesis of a 2D silicon structure, namely silicene, is attracting great interest for possible development of next generation electronic devices. The main difficulty in working with silicene remains its strong tendency to oxidation when exposed to air as a consequence of its relatively highly buckled structure. In this work, we univocally identify the Raman mode of air-stable low-buckled silicene nanosheets synthesized on highly oriented pyrolytic graphite (HOPG) located at 542.5 cm^-1. The main focus of this work is Raman spectroscopy and mapping analyses in combination with ab initio calculations. Scanning tunneling microscopy images reveal the presence of a patchwork of Si three-dimensional (3D) clusters and contiguous Si areas presenting a honeycomb atomic arrangement, rotated by 30° with respect to the HOPG substrate underneath, with a lattice parameter of 0.41 ± 0.02 nm and a buckling of the Si atoms of 0.05 nm. Raman analysis supports the co-existence of 3D silicon clusters and 2D silicene. The Raman shift of low-buckled silicene on an inert substrate has not been reported so far and it is completely different from the one calculated for free-standing silicene and the ones measured for silicene grown on Ag(111) surfaces. Our experimental results are perfectly reproduced by our ab initio calculations of deposited silicene nanosheets. This leads us to conclude that the precise value of the observed Raman shift crucially depends on the strain between the silicene and the HOPG substrate.