Evolution and Future Trends of SIMOX Material
- PDF / 5,615,989 Bytes
- 5 Pages / 576 x 777.6 pts Page_size
- 8 Downloads / 206 Views
MRS BULLETIN/DECEMBER 1998
lar this article will focus on development of the two types of SIMOX of greatest interest, "high-dose" (1.8 X 10w cm"2) and "low-dose" (3 to 5 X 1017 cm"2) material. Early developments of SIMOX are reviewed in this issue by Izumi, including the key developments leading to "contemporary" high-dose material.2-5 Whereas
"early" high-dose SIMOX had similar implant energy and dose to "contemporary" high-dose material, the structure and electrical properties differ dramatically because of a decade of progress in equipment, processing conditions, and understanding of the material. Structural differences between "early" and "contemporary" are clearly apparent in transmission-electron-microscopy (TEM) images in Figures la and lb. The reasons for these differences will be discussed next, whereas issues on the more recently developed "low-dose" material will be presented later. The BOX Layer in High-Dose SIMOX The choice of dose for high-dose material was set at a value between the minimum oxide thickness required for low electrical-leakage currents through the BOX and the minimum top silicon film thickness required to build devices. High-dose SIMOX is typically implanted to a dose of 1.8 X 1018 cirT2 at 150-200 keV
Figure 1. Evolution of high-dose SIMOX over the past decade as shown by cross sectional transmission-electron-microscopy (TEM) images of (a) "early" material with a high density of precipitates and threading defects in the top silicon layer, and (b) "contemporary" material with a precipitate-free top silicon layer with a much lower density of threading defects.
25
Evolution and Future Trends of SIMOX Material
and at 550-600°C, and is annealed at 1300-1350°C, for several hours, either oxide-capped or in an ambient of 99.5% Ar + 0.5% O2. Top silicon and BOX thicknesses are typically 200 nm and 400 nm, respectively. Early SIMOX was annealed at 11501250°C—the upper limit of quartz furnace tubes—and left a high density of residual precipitates in the top silicon, as shown in Figure la. When Celler et al.6 in 1986 used quartz lamps to heat wafers to ~1400°C for 30 min, they found that oxide precipitates were completely removed from the top silicon—similar to the structure seen in Figure lb. The thermodynamic driving force for this structural change is the reduction in surface free energy of the precipitates, called "Ostwald ripening." The processes that eliminate the precipitates appear in Figure 2 and include oxygen out-diffusion, precipitate growth, precipitate coalescence, and precipitate incorporation into the BOX. In practice, furnaces with silicon carbide tubes and refractory elements were developed for annealing above 1300°C because high-temperature annealing for long times and slow heating and cooling rates are required to form high-quality oxides and to avoid formation of slip planes in the silicon. A serious technology problem with early SIMOX was metallic impurity contamination due to sputtering of the implanter walls by the intense oxygen beam. This was solved by coating the
Data Loading...
