Abstract
Song, Benson, and Watkins have recently reported injection-annealing studies of electron-irradiated float-zone-grown Si. From these, they concluded that a long-accepted identification of a deep-level transient spectroscopy level, at 0.10 eV in n-type Si, with the carbon interstitial is incorrect. They conclude that instead this level must belong to a complex of a vacancy with some unidentified impurity or in some self-trapped configuration. They also concluded that another long-established identification of a level at 0.44 eV with the P-vacancy complex (E center) is partly wrong in that a major fraction of the defects contributing to that level are a complex of a single vacancy with another unidentified impurity and perhaps also P. They further conclude that this complex exhibits a remarkable fourfold configurational metastability. Here we offer an explanation of the injection-annealing data which preserves the previous identifications of these levels. The explanation centers on analysis of the expected interaction between C interstitials and Si vacancies with small dislocation loops, which are assumed to be formed during irradiation. The same argument offers an explanation of the effects of oxidation upon dopant diffusion that does not require the presence of Si self-interstitials.