Wafers with B-doped buried marker layers were implanted with 100keV Si to a fluence of 2 x 1015/cm2. This caused amorphization. Further B implantation was carried out before post-implantation annealing, so that the implanted B was located entirely within the pre-amorphized region. After implantation, the samples were annealed at various temperatures for various times, and secondary ion mass spectroscopy was used to obtain the dopant profiles. It was found that the buried marker layer exhibited normal transient-enhanced diffusion behavior. However, B in the pre-amorphized regions did not exhibit any significant degree of migration. It was suggested that solid-phase re-growth of the amorphous layer did not cause redistribution of the dopant atoms within that layer. Also, a plane of dislocation loops which formed at the amorphous/crystalline interface was an effective barrier to the diffusion of interstitial damage from the non-amorphized tail of the amorphizing Si implant. The same behavior was observed when As or P were used instead of B. This type of behavior was simulated by using a model that considered the growth of stacking faults that were bounded by dislocation loops.

H.S.Chao, P.B.Griffin, J.D.Plummer: Applied Physics Letters, 1996, 68[25], 3570-2