Energetics and kinetics of Cu atoms and clusters on the Si(111)-7 x 7 surface: first-principles calculations

Exploring the properties of noble metal atoms and nano- or subnano-clusters on the semiconductor surface is of great importance in many surface catalytic reactions, self-assembly processes, crystal growth, and thin film epitaxy. Here, the energetics and kinetic properties of a single Cu atom and previously reported Cu magic clusters on the Si(111)-(7 x 7) surface are re-examined by the state-of-the-art first-principles calculations based on density functional theory. First of all, the diffusion path and high diffusion rate of a Cu atom on the Si(111)-(7 x 7) surface are identified by mapping out the total potential energy surface of the Cu atom as a function of its positions on the surface, supporting previous experimental hypothesis that the apparent triangular light spots observed by scanning tunneling microscopy (STM) are resulted from a single Cu atom frequently hopping among adjacent adsorption sites. Furthermore, our findings confirm that in the low coverage of 0.15 monolayer (ML) the previously proposed hexagonal ring-like Cu-6 cluster configuration assigned to the STM pattern is considerably unstable. Importantly, the most stable Cu-6/Si(111) complex also possesses a distinct simulated STM pattern with the experimentally observed ones. Instead, an energetically preferred solid-centered Cu-7 structure exhibits a reasonable agreement between the simulated STM patterns and the experimental images. Therefore, the present findings convincingly rule out the tentative six-atom model and provide new insights into the understanding of the well-defined Cu nanocluster arrays on the Si(111)-(7 x 7) surface.This work was supported by the National Natural Science Foundation of China (Grant No. 11504332), the China Postdoctoral Science Foundation (2015M580633), the Outstanding Young Talent Research Fund of Zhengzhou University (1521317006) and the National Research Foundation of Korea (Grant No. 2015R1A2A2A01003248). We also thank Y. Jia for valuable discussions