Theory of Defects in the MOS System

Without doubt, the single most important technology for electronic device fabrication relies on the metal-oxide-silicon (MOS) system. In recent years III-V semiconductors have promised great increases in device speed and, because they are direct-gap materials, have been very important in electro-optical applications (i. e. light emitting diodes). However, these semiconductors have very complicated surfaces with high densities of surface-states and, hence, very low surface electron densities. While bare silicon surfaces also have low electron densities due to surface-states, most of these states can be eliminated by growing a thermal oxide. For other semiconductors, no such passivating process exists, so that, for instance, inversion devices such as enhancement-mode MOSFET’s (field effect transistors) are practical only in silicon. It is not surprising, then, that the MOS system has been studied intensely over the past thirty years. One of the most important areas of investigation has been the study of point defects. Defects in the oxide lead to shifts in threshold voltage in MOSFET’s [1]. Also, the residual unpassivated defects at the Si/SiO2 interface are responsible for decreased surface mobility, and for “soft” threshold characteristics. We should note that while some defects are inherent, i. e. arise from differences in the thermal expansion coefficients of silicon and silicon dioxide, and from lattice-network mismatch, most exist due to ionizing radiation or hot-electron injection. They can be created during fabrication (during X-ray lithography, or during the various plasma-assisted etching steps), or they can be a result of exposure to a space environment where even a modest flux of gamma rays can cause dramatic changes in device characteristics [2].