The technologically useful properties of a crystalline solid depend upon the concentration of defects it contains. Here we show that defect concentrations as deep as 0.5 microm within a semiconductor can be profoundly influenced by gas adsorption. Self-diffusion rates within silicon show that nitrogen atoms adsorbed at less than 1% of a monolayer lead to defect concentrations that vary controllably over several orders of magnitude. The results show that previous measurements of diffusion and defect thermodynamics in semiconductors may have suffered from neglect of adsorption effects.

No abstract

In the same way that gases react with surfaces from above, solid-state point defects such as interstitial atoms can react from below. Little attention has been paid to this form of surface chemistry. Recent bulk selfdiffusion measurements near the Si͑100͒ surface have quantified Si interstitial annihilation rates, and shown that these rates can be described by an annihilation probability that varies by two orders of magnitude in response to saturation of surface dangling bonds by submonolayer gas adsorption. The present work shows by modeling that the interstitial annihilation kinetics are well described by a precursor mechanism in which interstitials move substantial distances parallel to the surface before incorporation.

High-temperature real-time observation of surface defects induced by single ion irradiation using scanningtunneling-microscope/ion-gun combined system Point defects such as vacancies and interstitial atoms serve as primary mediators of solid-state diffusion in many materials. In some cases, the defects encounter surfaces where annihilation can occur. Quantification of annihilation rates presents formidable challenges, since point defect concentrations are typically low and therefore difficult to monitor directly. The present work develops a method for such quantification based upon measurements of diffusional profile spreading of a foreign species, using as an example isotopically labeled silicon implanted into a silicon matrix. Optimal experimental design techniques together with maximum-likelihood estimation indicate that the loss probability for Si interstitials on nitrogen-covered Si͑100͒ lies at 7.1ϫ 10 −4 .

No abstract