Defects, Diffusion, Ion Implantation, Recrystallization, and Dielectrics
Ideally pure single-crystal silicon shows an intrinsic electrical conductivity, which is low at room temperature and rises with increasing temperature. To produce devices, well-defined parts of the silicon specimen, usually in the form of a wafer, must ex
- PDF / 2,863,972 Bytes
- 24 Pages / 439.37 x 666.142 pts Page_size
- 62 Downloads / 197 Views
11.1 Introduction Ideally pure single-crystal silicon shows an intrinsic electrical conductivity, which is low at room temperature and rises with increasing temperature. To produce devices, well-defined parts of the silicon specimen, usually in the form of a wafer, must exhibit a well-defined surplus of either negative (electrons) or positive (holes) carriers, leading to n-type conductivity or ptype conductivity, respectively. n-type conductivity is obtained by doping with donors such as P, As, and Sb, and p-type conductivity is obtained with acceptors such as B, AI, Ga, and In. Donor or acceptor atoms must be on substitutional sites in the single-crystal silicon lattice to be electrically active, i.e. to form levels close to the conduction band or close to the valence band, respectively. The particular cases of N in Si and C are discussed in the context of ion implantation. The first, now conventional techniques used to dope a wafer or maskdefined areas of a wafer were based on the diffusion of dopants from the specimen surface, used as the source, into the bulk in long-lasting hightemperature furnace processes. The dopants are already electrically active at the end of the diffusion process. The now conventional implantation of high-energy dopant ions, introduced at the end of the 1960s is performed using high-voltage accelerators. It is a comparatively short process. However, the dopant ions do not necessarily come to rest at substitutional sites. Thus a short high-temperature process must follow to electrically activate the dopants and also to anneal the radiation damage caused by the ion implantation, so we do not really escape processes which involve high-temperature induced changes of sites, i.e. diffusion. Annealing can be done in a furnace, or by laser or electron beam irradiation, for example. Other methods of doping include doping during oxidation, CVD, epitaxy, or MBE, and doping by neutron transmutation. Thus the primary energies of the dopants range from approximately 0.1 eV to the high-MeV level, a range of 10 orders of magnitude. The continuing interest in silicon is demonstrated by the proceedings of large-scale conferences on the occasion of the 50th anniversary of silicon (see e.g. [1]). We had to learn that extraordinary precautions have to be applied concerning cleanliness in high-temperature processes to avoid deleterious conP. Siffert et al. (eds.), Silicon © Springer-Verlag Berlin Heidelberg 2004
208
E.F. Krimmel
tamination of the specimens due to the carrier lifetime killers gold and iron. Why were the Si wafers suddenly contaminated with traces of Au? It was finally found out that the supplier of the solvent used to clean the wafers had improved the production process of the solvent by introducing a gold catalyst. We and our distant partners in diffusion had to learn that annealing low-dose B-implanted Si specimens in standard furnaces may give odd results owing to cross-contamination.
11.2 High-Temperature Doping by Diffusion Sources for doping Si specimens from the gase phase include
Data Loading...
