4. Oxidation of Doped Silicon

THE DOPANT DISTRIBUTION in silicon is stronly influenced by thermal oxidation, because the dopants are redistributed by diffusion and segregation, especially near the silicon wafer surface [79]. However, this dopant redistribution is not the only effect of an oxidation step. Because of the oxide growth, the upper silicon zones are converted into SiO$_2$ and the Si/SiO$_2$ interface is moving into deeper silicon zones. Before oxidation, the dopant distribution exhibits generally a Gaussian-like profile, which means that the dopant concentration decreases stronly with the distance from the surface. Therefore, oxidation leads to a general decrease of the dopant concentration at the silicon surface. Furthermore, the formed oxide absorbs the dopants from the converted silicon material. This oxide doping influences the segregation of the dopant concentration at the Si/SiO$_2$ interface.

An influence of the dopants on the oxide growth rate was only found at very high dopant concentrations near the repective solubility limits of the used doping material, which are in the order of 10$^{20}$ atoms/cm$^3$ [80]. A high dopant concentration at the silicon surface beneath the SiO$_2$ (see Fig. 4.1b) causes crystal defects and so the silicon is easier to oxidize. A high number of dopants in the SiO$_2$ (see Fig. 4.1a) loosens the material and reduces its density, which enables a better oxidant diffusion through the SiO$_2$ to the interface. In both cases the oxide growth rate is increased.

Since very high dopant concentrations increase the oxide growth rate, theoretically the accelerated oxide growth at heavily doped zones could be used for selective oxidation. But unfortunately in practice this effect is too small to obtain noticeable differences in the oxide thickness. However, the different oxide growth velocities must be taken into account for an etching process. A faster oxide growth leads to a faster material removal by etching.