Arsenic overlayer on Si(111): Removal of surface reconstruction

Abstract

The complex reconstruction characteristic of annealed Si(111) surfaces is removed by the interaction of As with the clean Si(111) 7x7 surface. Surface-sensitive core-level spectroscopy measurements of the Si(111):As 1&1 surface reveal a well-ordered surface with threefoldcoordinated As atoms replacing the Si atoms in the outermost layer. The absence of dangling bonds in this structure leads to an unreconstructed surface which is highly resistant to contamination.Si(~&~) 4s:1x1Side view in the (110) plane of the Si(111):As 1 x 1 surface. The As atoms substitute for the top Si layer, resulting in a stable, unreconstructed surface.While many factors such as strain' and correlation energies combine with the electronic energy to determine the microscopic structure of a reconstructed surface such as Si(111)7 x 7, 2 the high energy of the ideal Si(111)surface arises from the existence of metallic dangling-bond orbitals. If these orbitals could be either removed or filled, the driving force for reconstruction would be removed, and a 1&&1 surface should result. We find that this situation is realized in practice by arsenic termination of the Si (111) surface. The atoms in bulk Si are tetrahedrally coordinated with sp3 hybridized bonding orbitals. At the (111)surface, Si atoms in the top layer have only three bonds, leading to instability and reconstruction. Arsenic, on the other hand, prefers a threefold coordination, with three orbitals taking part in bonding and the two remaining valence electrons in a nonbonding "lone-pair" state. If As atoms substitute for the surface Si atoms at the (Ill) surface, all atoms can be optimally bonded (see Fig. 1). Each Si atom in the bulk (b) is bonded to four Si atoms; each interface Si atom (a) is bonded to one Si and three As atoms; each As atom at the surface (shaded in Fig. 1) is bonded to three Si atoms. The result is an unreconstructed, stable Si(111):As 1 x 1 surface.This configuration is similar to the situation of As on Ge(111) which has been studied with angle-resolved photoemission spectroscopy. s However, in the germanium case the As atom is nearly the same size as the Ge atom it is replacing, so that terminating the surface with As does not lead to a large strain at the surface. For the silicon surface, the possibility exists that the larger As atom will lead to a strain-induced reconstruction. This does not happen, indicating that dangling-bond removal dominates over strain effects. It should be noted that As termination of the Si(111)surface is quite different from the 1 x 1 surface structure which results from chlorine adsorptions on Si(111)2x 1 or hydrogen adsorptions on Si(111)2&1 and Si(111) 7&7. In the case of Cl or H termination, the monovalent adatom passivates the surface by capping a single silicon dangling bond; in the case of As termination, we will show that the arsenic replaces the top silicon layer, bonding to three silicon atoms and terminating the surface with a nonreactive lone-pair orbital. We expect that the As atoms relax outwards from the bulk Si lat...

Effect of As preadsorption on InAs nanowire heteroepitaxy on Si(111): A first-principles study

“…As long as As chemical potential is below the bulk level, the segregated As will evaporate, limiting As adsorption to one monolayer. This is consistent with the experimental observation 5 that excess As can be removed by annealing at a low temperature ͑200°C͒.…”

“…The detailed structure is however unclear from the vapor-phase experiments while 1 ϫ 1-As is the only phase found in ultrahigh-vacuum ͑UHV͒ studies. [5][6][7][8][9][10][11] Although previous first-principles calculations have shown that the ͱ 3 ϫ ͱ 3 T 4 adatom, 12-14 ͱ 3 ϫ ͱ 3 trimer, 15 and 2 ϫ 1 zigzag chain 15 FIG. 1.…”

“…On a Si substrate, however, both vertical and tilted nanowires grow because inversion symmetry makes ͑111͒ and ͑111͒ equivalent. To increase the percentage of vertical InAs nanowires on Si, the experimentalists 2 tried exposing the substrate to arsine, thus turning the exposed Si͑111͒ surfaces into the unreconstructed arsenide structure ͑1 ϫ 1-As͒, 5 which resembles the ideal InAs͑111͒ surface ͑Fig. 1͒.…”

The physics and chemistry of the Schottky barrier height

“…, Si(111)6.3 Â 6.3-Ga, [424][425][426][427] Si(111) 1 Â 1-As, [428][429][430] Si(100) 1 Â 1-S, 431,432 and Si(111) 1 Â 1-Cl, 433,434 fabricated in situ, different metals were deposited at cryogenic temperatures and occasionally in an inert gas environment. Since the ATS surfaces were known to be stable to various degrees and the metals were deposited "softly," the expectation was that the surface dipole on the ATS might survive the metal deposition and be available for SBH adjustment.…”

Surfactant Mediated Heteroepitaxy versus Homoepitaxy: Kinetics for Group-IV Adatoms on As-Passivated Si(111) and Ge(111)

“…A full microscopic explanation of the surfactant effect is still missing, but several aspects have been clarified: Because of the additional electron, surfactant layers of group-V elements on Si or Ge modify the reconstruction of the surface. For example, the As-covered Si(111) surface shows a ͑1 3 1͒ structure [8] instead of the ͑7 3 7͒ reconstruction of pure Si(111). Group-V atoms have a reduced-surface free energy which makes them float on the surface without being incorporated into the growing crystal [1,9].…”