Electrical transport analysis of ultra-pure silicon

Abstract

Silicon with ever increasing purity and perfection is demanded for devices with smaller feature sizes and for specialized applications such as infrared detectors. Achievement of high purity Si has necessitated development of measurement procedures and data analysis techniques that are quantitatively sensitive to very low concentrations (<10 ppta) of electrically active impurities. Described are results from a research program that has improved all aspects of the temperature dependent Hall effect experiment while emphasizing p-type Si with very low concentrations of residual impurities. The improvements cover sample preparation and contacting, precise voltage and current measurements over a broad temperature range (4.2–400 K), sensitive temperature control, and calculations of the effective mass, m *( T), and the Hall scattering factor, r( T), that incorporate the valence bands' non-parabolicities and anisotropies. This is a quantitative technique that can be used as a standard for other analytical methods such as photoluminescence and photoabsorption. Two classes of high purity Si have been studied, namely undoped intrinsic material with boron as the dominant residual impurity at concentrations below 10 12 cm -3 (20 ppta) and material doped (generally with Group III elements) to concentrations in excess of 10 16 cm -3 in the presence of residual impurity concentrations near or below 10 12 cm -3. Results from the measurement of samples with a variety of doping and impurity concentrations are shown. Comparisons between the mobility calculations and the measured mobilities are discussed as well as direct measurements of r( T) made on intrinsic Si at low temperatures. Significant improvements in the ability to measure high purity samples have been realized and in some cases the accuracy has been increased by over 50% in the measured dopant and impurity concentration values.