This dopant diffused extensively after implantation and long-term annealing. The results could be explained by assuming that the diffusivity depended upon the square of the electron concentration. The dopant diffusion was affected by the presence of implantation damage; the higher the concentration of extended defects, the slower being the diffusivity as compared with the values for conventional diffusion from a solid source. If the sample was amorphized during implantation, extended defects did not form and the diffusivity of the ion was very close to that in material which had been diffused from a solid source. When amorphization did not occur, extended defects formed after implantation, and diffusion was inhibited; especially after low doses, in the short term, or at low temperatures. The higher the density of extended defects, the greater was the suppression of diffusion. The diffusion was time-dependent. It was concluded that the results (table 22) were consistent with a diffusion mechanism in which the mobile species was the donor that was coupled with a charged Ga vacancy. The equilibrium vacancy concentration was thought to be suppressed by the presence of extended defects and/or excess Ga interstitials which resulted from implantation.

E.L.Allen, J.J.Murray, M.D.Deal, J.D.Plummer, K.S.Jones, W.S.Rubart: Journal of the Electrochemical Society, 1991, 138[11], 3440-9

Table 22

Diffusivity of Implanted Si in GaAs