Thermal conductivity engineering of bulk and one-dimensional Si-Ge nanoarchitectures

Various theoretical and experimental methods are utilized to investigate the thermal conductivity of nanostructured materials; this is a critical parameter to increase performance of thermoelectric devices. Among these methods, equilibrium molecular dynamics (EMD) is an accurate technique to predict lattice thermal conductivity. In this study, by means of systematic EMD simulations, thermal conductivity of bulk Si-Ge structures (pristine, alloy and superlattice) and their nanostructured one dimensional forms with square and circular cross-section geometries (asymmetric and symmetric) are calculated for different crystallographic directions. A comprehensive temperature analysis is evaluated for selected structures as well. The results show that one-dimensional structures are superior candidates in terms of their low lattice thermal conductivity and thermal conductivity tunability by nanostructuring, such as by diameter modulation, interface roughness, periodicity and number of interfaces. We find that thermal conductivity decreases with smaller diameters or cross section areas. Furthermore, interface roughness decreases thermal conductivity with a profound impact. Moreover, we predicted that there is a specific periodicity that gives minimum thermal conductivity in symmetric superlattice structures. The decreasing thermal conductivity is due to the reducing phonon movement in the system due to the effect of the number of interfaces that determine regimes of ballistic and wave transport phenomena. In some nanostructures, such as nanowire superlattices, thermal conductivity of the Si/Ge system can be reduced to nearly twice that of an amorphous silicon thermal conductivity. Additionally, it is found that one crystal orientation, <100>, is better than the <111> crystal orientation in one-dimensional and bulk SiGe systems. Our results clearly point out the importance of lattice thermal conductivity engineering in bulk and nanostructures to produce high-performance thermoelectric materials.