The role of stabilized back-surface damage in controlling internal SiOx nucleation and denudation zones in Si

Abstract

Using two-step annealing of wafers containing mechanically induced back-surface damage, we have shown that the presence of stabilized damage regions controls the development of front-surface defect denudation zones and internal SiOx nucleation. Using data from transmission electron microscopy, secondary ion mass specrometry profiling, and secondary ion microscopy measurements we have demonstrated that little or no direct correlation exists between measured (front-surface) oxygen depletion widths and defect denudation zone widths. Primary anneals at temperatures <600 °C have also been shown to produce no significant or measurable depletion of oxygen at the front surface of wafers.

Denuded Zones in Czochralski Silicon Wafers

“…Work has been done to investigate the formation of the denuded zone on silicon wafers under different heat-treatment conditions. Denuded zones ranging from a few microns up to 170 microns have been observed by different authors (Matyja et al, 1975;Magee et al, 1981;etc. ).…”

“…Magee et al (1981) describe how lightly damaging the back surface of silicon wafers can generate a stress field at the back surface. The stress field will introduce defects such as dislocations, microtwins, polycrystals, etc.…”

“…Precipitates at the surface region introduced during the first annealing step were then redissolved back into the solid solution, due to the undersaturation, and left a clear denuded zone. Magee, Leung and Kawayoshi (1981) reported that after a low-high temperature, two-step annealing (600°C for 12 hr. in Ar + 1050°C for 3 hr.…”

“…Some authors have reported (Tan, Gardner and Tice, 1976;Yamamoto et al, 1979;Magee et al, 1981;Series et al, 1981) that the denuded zone could form at the silicon wafer surface during preannealings in an inert ambient, such as Ar and N 2 , since a great oxygen concentration gradient exists between the silicon wafer and the ambient, and this allows the process of oxygen out-diffusion to occur. Tan, Gardner and Tice (1977) showed that after annealing for 6 hr.…”

“…After being subjected to back-surface damage, the wafers are annealed for some time to allow impurities to diffuse to the back-surface-damaged sites of the wafers (Magee, 1981). Magee et al (1981) examined denuded zones on back-surfacedamaged wafers. The samples used were n-type, 75-mm-diameter, 500 ym 18 thick, CZ-<100> wafers.…”