This thesis presents theoretical treatments of vacancies and core excitons on semiconductor surfaces. After an introduction to the subject, the general Koster-Slater approach in the tight-binding approximation is reviewed. Realistic tight-binding parameters for a number of III-V and II-VI semiconductors are obtained. Chapter 3 then deals with semiconductor surface core excitons. Vacancies near a semiconductor surface are treated in Chapter 4.
We present in the third chapter calculations of the binding energies and oscillator strengths of surface core excitons. Results for the (110) surface of GaAs, GaP, GaSb, and InP are included. We find that transitions from core levels to the surface dangling bonds allow the confinement of the electron to the region near the hole, producing a Frenkel exciton, We find that transitions out of the cation d core to the large p component of the dangling bond lead to observable excitons. Transitions from the cation p core couple to the small s component of the surface state and have similar binding but damped oscillator strength. It is found that transitions from the anion core should produce no exciton.
In the fourth chapter, we show calculations of the bound state levels of vacancies near semiconductor surfaces. We present results for ideal vacancy levels in bulk and near the (110) surface of GaSb, GaAs, GaP, InAs, AlAs, InP, ZnTe, CdTe, ZnSe, GaxIn1-xAs,GaxAl1-xAs, and GaxIn1-xP. We find that ideal vacancy levels are changed by less than 0.1 eV as the vacancy is moved near the surface until it reaches the second layer from the surface . On the surface, the highest occupied level in the neutral anion vacancy goes to higher energy and that in the cation vacancy goes to lower energy.
General trends are noted for comparisons between covalent and more ionic materials. Cation vacancy levels tend to move toward the valence band and anion levels toward the conduction band as one considers increasingly ionic compounds. We consider also the Jahn-Teller effect (lattice distortion) and Coulomb repulsion among the bound electrons. These effects cause the various charge states of vacancies to have different energies. Vacancies in III-V's and II-VI's show qualitatively different behavior. A neutral vacancy in a III-V compound has an odd number of electrons and so can have both donor and acceptor roles. A neutral vacancy in a II-VI semiconductor has an even number and takes on either double acceptor or double donor character, depending on the specifics.
The Defect Model of Schottky barrier formation is also investigated in Chapter 4, and it is concluded that anion vacancies in III-V semiconductors can account for the observed properties of Fermi level pinning. The role of vacancies in pinning the Fermi level of II-VI materials is uncertain.
