The concentration profiles of B which had been diffused, from polycrystalline material, into underlying monocrystalline material were analyzed by means of secondary ion mass spectrometry. The co-diffusion of As and B was studied in an emitter and extrinsic base configuration. The process was investigated by carrying out simulations. These indicated that diffusion of the dopant at the lowest fluence was slowed much more by the in-depth inhomogeneous grain growth that was induced by amorphization and annealing, than by a built-in electric field. It was assumed that the dopant, at the highest doses, saturated the grain boundary traps. This was true of B. In a first poly-Si layer, the diffusivity (table 22) could be described by:
D (cm2/s) = 1.9 x 10-2exp[-2.5(eV)/kT]
In a second poly-Si layer, the diffusivity could be described by:
D (cm2/s) = 3.2 x 10-5exp[-1.86(eV)/kT]
Dopant Redistribution during Rapid Thermal Annealing in a Self-Aligned Polysilicon Emitter Bipolar Structure Compatible with a Complementary Metal-Oxide-Semiconductor Technology. A.Merabet, C.Gontrand: Physica Status Solidi A, 1994, 145[1], 77-88
Table 22
Diffusivity of B in Poly-Si Layers
Layer | Temperature (C) | D (cm2/s) |
first | 1150 | 2.7 x 10-11 |
first | 1100 | 1.3 x 10-11 |
first | 1050 | 6.3 x 10-12 |
first | 1000 | 2.5 x 10-12 |
second | 1150 | 7.2 x 10-12 |
second | 1100 | 4.8 x 10-12 |
second | 1050 | 2.6 x 10-12 |
second | 1000 | 1.2 x 10-12 |