|
Defect types include atom vacancies, adatoms, steps, and kinks which occur most frequently at surfaces due to finite material size causing crystal discontinuity. What all types of defects have in common, whether they be surface or bulk, is that they produce dangling bonds which have specific electron energy levels not similar to those of the bulk. This is because these states cannot be described with periodic Bloch waves due to the change in electron potential energy caused by the missing ion cores just outside of the surface. Hence, these are localized states which one must solve the Schrödinger equation for separately such that electron energies can be properly described. The break in periodicity results in a decrease in conductivity due to defect scattering. ==Electronic Energy Levels of Semiconductor Dangling Bonds== A simpler and more qualitative way of determining dangling bond energy levels is with Harrison diagrams.〔Harrison, Walter A., Electronic Structure and the Properties of Solids: The Physics of the Chemical Bond. San Francisco: Freeman, 1980.〕〔Rockett, Angus, The Materials Science of Semiconductors. New York: Springer, 2007〕 Metals have non-directional bonding and a small Debye length which, due to their charged nature, makes dangling bonds inconsequential if they can even be considered to exist. Semiconductors are dielectrics so electrons can feel and become trapped at defect energy states. The energy levels of these states are determined by the atoms that make up the solid. Figure 1 shows the Harisson diagram for the elemental semiconductor Si. From left to right, s-orbital and p-orbital hybridization promotes sp3 bonding which, when multiple sp3 Si-Si dimers are combined to form a solid, defines the conduction and valence bands. If a vacancy were to exist, such as those on each atom at the solid/vacuum interface, it would result in at least one broken sp3 bond which has an energy equal to that of single self hybridized Si atoms as shown in Figure 1. This energy corresponds to roughly the middle of the bandgap of Si, ~0.55eV above the valence band. Certainly this is the most ideal case whereas the situation would be different if bond passivation (see below) and surface reconstruction, for example, were to occur. Experimentally, the energies of these states can be determined using absorption spectroscopy or X-ray photoelectron spectroscopy, for example, if instrument sensitivity and/or defect density are high enough. Compound semiconductors, such as GaAs, have dangling bond states that are nearer to the band edges (see Figure 2). As bonding becomes increasingly more ionic, these states can even act as dopants. This is the cause of the well known difficulty of GaN p-type doping where N vacancies are abundant due to its high vapor pressure resulting in high Ga dangling bond density. These states are close to the conduction band edge and therefore act as donors. When p-type acceptor dopants are introduced, they are immediately compensated for by the N vacancies. With these shallow states, their treatment is often considered as an analogue to the hydrogen atom as follows for the case of either anion or cation vacancies (hole effective mass, m *, for cation and electron m * for anion vacancies). The binding energy, Ec-Edb, is where U=-q2/(4πεεrr) is the electrostatic potential between an electron occupying the dangling bond and its ion core with ε, the free space permittivity constant, εr, the relative permittivity, and r the electron-ion core separation. The simplification that the electron translational energy, KE=-U/2, is due to the virial theorem for centrosymmetric potentials. As described by the Bohr model, r is subject to quantization . The electron momentum is p=mv=h/λ such that resulting in and . This treatment loses accuracy as the defects tend away from either band edge. 抄文引用元・出典: フリー百科事典『 ウィキペディア（Wikipedia）』 ■ウィキペディアで「Carrier scattering」の詳細全文を読む スポンサード リンク
|