AVS 49th International Symposium
    Thin Films Thursday Sessions
       Session TF-ThA

Paper TF-ThA6
Epitaxial Growth of Au and Ag Films on Si(111) using Cu Buffer Layers

Thursday, November 7, 2002, 3:40 pm, Room C-101

Session: Ultra Thin Films
Presenter: K. Pedersen, Aalborg University, Denmark
Authors: K. Pedersen, Aalborg University, Denmark
P. Morgen, University of Southern Denmark
T.G. Pedersen, Aalborg University, Denmark
Z.S. Li, Aarhus University, Denmark
S.V. Hoffmann, Aarhus University, Denmark
Correspondent: Click to Email

Deposition of metals on semiconductor surfaces often lead to films of poor quality due to lattice mismatch or chemical reaction between metal and the substrate. For both Au and Ag growing on Si(111)-7x7 the films are imperfect (111) metal layers. In the case of Au the deposition also creates a disordered, Si-rich layer that floats on top of the Au layer. Furthermore, the surface states usually found on (111) faces of noble metals are absent for these Au and Ag films. Though Cu reacts with Si, growth of a well ordered Cu(111) film starts after a thin reacted phase around the contact layer has been created. The reacted layer thus serves as buffer layer between crystalline Cu and the Si substrate. In synchrotron radiation photoemission experiments we demonstrate how the reacted Cu layer also serves as a buffer layer improving the properties of Au and Ag films on Si(111). Core level spectra show that the buffer layer prevents the reaction between Au and Si. This leads to an ordered surface of the film with a surface state. Thin Ag films grown directly on Si(111)-7x7 form quantum well states that are seen as peaks in valence band spectra. These peaks appear broader and weaker than for Ag grown epitaxially on for instance Fe(001). Furthermore, strains caused by mismatch between Ag and Si lattices shift the surface state above the Fermi level. The thin Cu buffer layer improves on both problems. The sharpness of the quantum well levels increases considerably and the surface state appears just below the Fermi level. From the sharpness of quantum well levels it is concluded that the optimum buffer layer thickness is 6 - 8 atomic layers.