6 Grain-boundary and dislocation diffusion in semiconductors and silicides

This chapter presents a collection of experimental data on the triple or double products of grain-boundary and dislocation diffusion in semiconductors and silicides. The most important method - serial sectioning in type-B kinetic regime - is described. From these measurements the product of the diffusion coefficient (D' or D" for grain boundaries or dislocations, respectively), the segregation factor (K', or K", respectively) and the grain boundary width δ (or the square of the pipe radius, a, for dislocation diffusion) can be determined. From profiling measurements carried out in type-A or C kinetic regimes, the diffusion coefficient of the given short circuit can be estimated directly. The methods of measurements include low angle grain boundary method, defect annealing method, Pavlov-Panteleev method, indirect methods in type-A kinetic regime, determination of double products from creep and sintering experiments, isoconcentration contour method, Hwang-Balluffi method, first appearance method, and Gilmer-Farrell method. Diffusion processes in semiconductors proved to be more complex phenomena than the transport in metallic systems. Dislocation diffusion in semiconductors and grain boundary diffusion in semiconductors and silicides are tabulated. Semilogarithmic plot of the dislocation diffusivity K"a ² D" of different impurities in silicon versus inverse temperature 1/T and semilogarithmic plot of the grain boundary diffusivity KδD' of arsenic in silicon, aluminum and boron in silicon, phosphorus in silicon, antimony in silicon, self- and impurity diffusivity KδD' in germanium, self- and impurity diffusivity KδD' in silicides and In and Sb in InSb compound versus inverse temperature 1/T are figured out.