Reduced thermal conductivity of nanotwinned random layer structures: a promising nanostructuring towards efficient Si and Si/Ge thermoelectric materials

Abstract

Si and Si/Ge based nanostructures of reduced lattice thermal conductivity are widely attractive for developing efficient thermoelectric materials. In this study, we demonstrate the reduced thermal conductivity of Si nanotwinned random layer (NTRL) structures over corresponding superlattice and twin-free counterparts. The participation ratio analysis of vibrational modes shows that a possible cause of thermal conductivity reduction is phonon localization due to the random arrangement of twin boundaries. Via non-equilibrium molecular dynamic simulations, it is shown that ~23 and ~27% reductions over superlattice counterparts and ~55 and 53% over twin-free counterparts can be attained for the structures of total lengths of 90 and 170 nm, respectively. Furthermore, a random twin boundary distribution is applied for Si/Ge random layer structures seeking further reduction of thermal conductivity. A significant reduction in thermal conductivity of Si/Ge structures exceeding the thermal insulating performance of the corresponding amorphous Si structure by ~31% for a total length of 90 nm can be achieved. This reduction is as high as ~98% compared to the twin-free Si counterpart. It is demonstrated that application of randomly organised nanoscale twin boundaries is a promising nanostructuring strategy towards developing efficient Si and Si/Ge based thermoelectric materials in the future.