Anomalous Diffusion of B and P in Si Directly Masked with Si₃N₄

Abstract

Diffusion of B and P in Si directly masked with Si3N4 films is found to have anomalous characteristics depending on temperature, time, and Si crystal growing method. The discrepancy in diffusion between CZ and FZ substrates under Si3N4 films is explained well assuming the Si–SiO2 interface acts as a sink for super-saturated interstitials. The Si–Si3N4 interface does not have such an effect. It is also found that super-saturated interstitials enhance the diffusion of B and P in CZ Si crystals directly masked with Si3N4 films.

Point defects and dopant diffusion in silicon

“…In addition, it is possible that oxygen precipitation in the Czochralski (CZ) wafers floods the silicon with interstitials for times on the order of hours, thereby causing changes in the inert diffusivity, as observed. Such "anomalous" effects in CZ material have been reported by Hu (1980), Mizuo andHiguchi (1982b), . For this reason, the extraction of a time dependence of OED from this early work is questionable.…”

“…The values of D", & and Q"&& from these experiments do not show good agreement between different investigations, especially when compared to expressions for D"&f in germanium, which show rather good agreement [see the review article by Frank, Gosele, Mehrer, and Seeger (1984)]. In part, this may be due to the fact that oxygen in silicon, which is now known to affect point-defect behavior in the bulk (Mizuo and Higuchi, 1982b), was an uncontrolled factor in early investigations. Carbon content in silicon may also prove to be a factor.…”

“…Lin et al showed that these data were consistent with previous observations based on stacking fault kinetics during oxidation, but the data cannot provide an independent measure of the oxidation-rate dependence of OED. Ishikawa et al (1982) made measurements on (100) and (111) A range of times corresponding to an oxidation-rate difference of =3 at 1100'C was investigated by Mizuo and Higuchi (1982b), but no quantitative analysis of the results was given. Instead, junction depths and ratios of junction depths were plotted as a function of time.…”

“…When the thermal nitride film is removed, the antimony diffusivity returns to the inert value (Ahn, Kennel, Plummer, and Tiller, 1988a). It is also known (Mizuo and Higuchi, 1982b;Ahn et al , 1988b) that deposited nitride films cause enhanced Sb diffusivity, the degree of which is closely related to the stress level in the silicon nitride films. The stress level can be varied by varying the stoichiometric composition of the films.…”

Generation and annihilation of point defects by doping impurities during FZ silicon crystal growth

“…Even the enhanced diffusions [32] of the antimony by the vacancy and the boron by the silicon interstitials, which proceed in a wafer without a thermal gradient, are not contradicted by the above explanation.…”