Thermal conductivity of silicon nanostructures containing impurities

Abstract

In recent years the thermal conductivity of materials has become an important area of research. High thermal conductivity is useful for heat removal in electronic devices, and low thermal conductivity results in a larger thermoelectric figure of merit. The question is how to control the thermal conductivity. Many physical properties of semiconductors are successfully controlled by adding impurity atoms. The electrical conductivity can be substantially increased using dopants, the mechanical strength of Si can be improved by introducing O and N, optical properties are tailored using rare-earth elements such as Er. But is it possible to control the thermal conductivity with impurities as well? The impact of 'impurity scattering' (whatever this really means) on the thermal conductivity of materials is well known, but only qualitatively. However, it is never described at the atomic level. A better understanding of the physics behind the impact of impurities on the thermal conductivity could lead to devices which are tuned to the the desired thermal conductivity through the use of impurities. This dissertation introduces a first-principles method for calculating the thermal conductivity of semiconductor nanostructures containing impurities. The method is used to study the effects of various impurities on the thermal conductivity of silicon nanowires. The effects of impurity type, isotope, and concentration are studied. Surprisingly, the isotopic mass of the impurity appears to have a substantial impact on the thermal conductivity. Attempts to correlate this effect with the vibrational lifetime of impurity-related (localized) vibrational modes are discussed.