PNNL

Abstract

A repeatable, in situ process was developed for preparing clean, low-defect-density Ge(100) surfaces based on scanning tunneling microscopy (STM), Auger Electron Spectroscopy (AES), and x-ray photoelectron (XPS) measurements. Surfaces prepared by ioin sputtering followed by annealing, wet chemical etching followed by oxidation and thermal desorption of the oxide layer, growth of a Ge buffer layer, and in situ plasma cleaning were compared. Both sputter/annealing and etching followed by oxidation and thermal desorption of the oxide layer could produce surfaces with no impurities detectable by AES, however, STM images of surfaces prepared using these methods revealed high densities of protrusions, vacancies and other defects. These surfaces also exhibited a wide distribution of terrace widths due to step pinning by the protrusions. Auger electron spectra and STM images recorded before and after annealing C contaminated surfaces suggested that the protrusions were due to aggregation of surface C to form three-dimensional clusters st elevated temperatures. The low surface to volume ratio of the clusters makes them difficult to detect by either XPS or AES, and leads to a low cross-section for removal by ion bombardment. Although XPS indicated that oxygen plasma treatment effectively removes C from Ge surfaces, STM images of the plasma treated surfaces still showed the protrusions attributed to C. In contrast, surfaces prepared by the growth of Ge buffer layers followed by annealing exhibited no protrusions, low-defect-densities on the terraces (less than 2% of a monolayer), and evenly spaced 60 nm wide terraces that reflect the misorientation of the crystal. Thus the surfaces prepared by Ge buffer layer growth were determined to be the most suitable for the atomic level nucleation and growth studies.