Silicon Dry Oxidation Kinetics: Effects of the Choice of Integration Constant on Fits to the Linear Parabolic Model

On the Mechanism of Interaction Between 18 O 2 Gas and Noncrystalline SiO2

“…Thus, 18On concentration profiles can be described by an effective diffusion coefficient of the 'sOn; this is labeled as D in this paper. Also note that the observation, to be discussed below, that no change has occurred in the 'sOn concentration profile during subsequent heat-treatment in N2, indicates that reactions [1] and [2] are essentially not reversible. The reason for this is that, except for the area in the vicinity of the outer oxide surface, the concentration of 'sOn is SO small compared to that of 16On that reactions [1] and [2] going from right to left can be neglected.…”

“…There have been several recent studies which have used SIMS or nuclear reaction microanalysis techniques to examine the interaction between '80~ molecules and films of Si'sO2 thermally grown on silicon (1)(2)(3)(4). These studies have shown that an oxide film consisting predominantly of '80 atoms forms at the Si/SiO2 interface and that ~sO atoms are incorporated into the outer surface region of the Si'602 film.…”

“…These studies have shown that an oxide film consisting predominantly of '80 atoms forms at the Si/SiO2 interface and that ~sO atoms are incorporated into the outer surface region of the Si'602 film. On the basis of these results, and analyses of the growth rate of the oxides (1), some investigators (1,3) have suggested that there are two independent, noninteracting mechanisms responsible for the thermal growth of oxide on silicon. One of these is related to molecular oxygen, which diffuses through the oxide to react with silicon at the oxide/ silicon interface, and is generally accepted as a necessary mechanism of oxide growth, at least for thicker oxides.…”

“…It is suggested in Ref. (1) that growth resulting from this second mechanism is especially important for thin oxides, and is responsible for the well-known enhancement of oxide growth rates for thin oxides grown in dry oxygen. This departure from the standard linearparabolic growth rate law is the so-called "fast growth" regime.…”

“…In this paper we analyze the data of these studies [especially the SIMS work (1), because of its higher spatial resolution] in terms of a model based on the wellknown isotopic exchange process between '802 gas molecules and the 160 network atoms of silica glass (5). This analysis is similar in principle to that applied earlier by Revesz and Hughes (6) to the data of Rochet et al (3).…”