The direct synthesis of nanostructures by epitaxial growth is an important ingredient in the bottom up approach of the Si/Ge device fabrication. However, the creation of nanostructures with well defined properties still remains a challenge. In the present work we study the formation of atomically straight and well ordered Ge nanowires on a Si(111) surface and show how to optimize their size uniformity. We succeeded in growing Ge nanowires attaching to the step edges of a highly ordered kink-free stepped Si template. The Ge nanowires are single-crystalline and feature minimal kink densities allowing them to span length larger than 1 µm at a width of ~4 nm, as confirmed by STM. To achieve the desired growth behavior we introduce a new concept in surfactant mediated epitaxy: controlling the surfactant coverage at the surface and/or at the step edge to modify the growth properties of the surface.
In the usual growth of Ge on Si the SiGe intermixing is a severe problem for the formation of Si-Ge nanostructures. To effectively suppress GeSi intermixing one monolayer of Bi at the surface was used as surfactant. However, the presence of Bi at the surface modifies the free step energies leading to a drastic loss of initial step ordering. The key ingredients to maintain the original step structure are: very slow Bi termination to allow a gradual Si mass transport toward the step edges during the lifting of the Si(111)-7x7 reconstruction, and a careful control of the Bi concentration at the surface during Ge deposition influencing the growth properties of the surface steps in a desired way. In particular, we show that a slight decrease of the Bi content at the steps during the Ge deposition is sufficient to keep the steps straight and preserve their original <-1-12> orientation.
We show that this modified surfactant-mediated epitaxy is relevant to nanotechnology since it allows the fabrication of long equidistant Ge nanowires with a width in the single-digit nanometer range. These nanowire arrays could serve as templates for the selective attachment of molecules, decoration with metals, or attachment of clusters and other nanoscale building blocks such as fullerenes. The study of charge transport through such very small nanowires will be another challenge.
|