Abstract for 1999 American Vacuum Society National Symposium
Seattle, WA  October 1999

Title:  Si Surface Passivation: Si(111):GaSe versus Si(111):As

Authors: Aaron Bostwick, Physics Dept., Univ. of Washington 
	 Shuang Meng, Physics Dept., Univ. of Washington
         Brett R. Schroeder, Physics Dept., Univ. of Washington
         Eli Rotenberg, Advanced Light Source, Lawrence Berkeley Nat'l Lab
         Fumio S. Ohuchi, Dept. of Mat. Sci. & Eng., Univ. of Washington
	 Marjorie A. Olmstead, Physics Dept., Univ. of Washington 

Abstract:

Heteroepitaxy on Si(111)7x7 requires removal of the deep reconstruction
and accompanying dangling bonds. One way to passivate the Si(111) surface
and remove the reconstruction is exposure to As, forming Si(111):As 1x1.1
The initial chemisorption of GeSe on Si(111)7x7 surfaces also results in a
nearly ideally terminated 1x1 surface, quite similar to Si(111):As. Ga and
Se occupy bulk Si sites, with Ga directly above Si (T1 site) and Se
forming three back-bonds to Ga (H3 site). This (1x1) structure is the same
as half a bulk GaSe layer, and initiates GaSe heteroepitaxy on Si(111).
The very small (less than 0.1 eV) Si 2p core level shift shows the
interface silicon to be in a bulk-like environment with minimal charge
transfer, in contrast to As terminated silicon (0.75 eV shift). Electron
counting arguments predict a lone-pair state on the Si(111):GaSe 1x1
surface, very similar to the As case. We observe such a state with
angle-resolved ultraviolet photoemission spectroscopy. It has a similar
E(k) dispersion to Si(111):As, though a somewhat larger bandwidth. We find
a second surface state between this lone-pair state and the first bulk
state, which we attribute to Ga-Se bonds. In addition, the zone-center
bulk state, degenerate in bulk Si, is split by about 0.5 eV. This
surprising result is not found for Si(111):As. We tentatively attribute
the splitting to the Si-Ga interaction.

M. A. Olmstead, R. D. Bringans, R. I. G. Urhberg and R. Z. Bachrach, Phys.
Rev. B 34, 6401 (1986).