The Atomistic Mechanism of Si (111)-7×7 Surface Reconstruction Revealed Using a Machine-Learning Force Field

While the formation of the Si (111)-7×7 surface upon annealing is well known, the mechanism underlying this surface reconstruction remains elusive. We employ molecular dynamics simulations based on a machine-learning force field to unravel the atomistic mechanism. Our simulations reveal two possible pathways. The first pathway involves the growth of a stacking fault from the metastable 5×5 structure to the stable 7×7 reconstruction. The second pathway shows the direct formation of the 7×7 structure. Both mechanisms involve key steps, including the creation of dimers and five-membered rings, stabilization of the triangular halves, and the formation of a corner hole by joining several five-membered rings. Adatoms and dumbbell configurations are formed through the diffusion or rearrangement of extra atoms during the evolution of triangular subunits and dimer formation. Our findings offer insights into surface reconstruction in semiconductor materials.