The Oxide-Silicon Interface

Oxidized silicon surfaces are used on virtually all of today´s integrated circuits and silicon devices. These oxide layers encase the silicon surface in a protective sheath; they provide convenient insulating layers on which to deposit intraconnecting leads and bonding pads; and, perhaps most important, they terminate the silicon lattice in as ordered a fashion as has yet been found--interface state densities in the range of 1010 to 1011 cm¿2 ev¿1 or lower, commonly observed at the oxide-silicon interface, are by far the lowest surface state densities measured on any silicon surface. Surface related failures still exist, however, as has been noted at previous Physics of Failure Symposia. Device surface operation, whether it be any of the MOS family or simply a planar p-n junction, is predicated upon the action of the space charge region in the silicon that is adjacent to the oxide-silicon interface. The observed device properties are well described by simple models in which the surface conductance, capacitance, recombination velocity and other electrical properties are uniquely related to the surface potential. Extraneous charges in the oxidesilicon system that are not part of the silicon space chage region do exist, however, and constitute a source of degraded device performance, instability, and failure. Sources of extraneous charge include: 1. Ionic surface contamination (the Atalla model). anions and cations can inadvertently be left on the oxide surface during device manufacture or can be acquired at some time during operation.