Optimized design of nanofibers for low thermal conductivity in nanocomposites
Date
2015
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
The design of efficient thermoelectric materials involves the reduction of thermal conductivity attributable to heat conduction by thermally generated elastic waves called phonons, without degradation of electrical conductivity. Improved scattering of phonons can be attained by introducing a small volume fraction of embedded nanoparticles inside a semiconductor alloy matrix material. Calculating the thermal conductivity tensor from microscopic principles requires knowledge of the scattering cross-sections spanning all possible incident elastic wave orientations, polarizations and wavelengths. In this thesis, analytical continuum mechanics is used to construct the scattering cross-section of incident elastic waves from embedded nanofibers where the waves have arbitrary orientation, polarization, and wavelength. The model is then incorporated using Boltzmann transport theory to predict the thermal conductivity tensor. The thermal model is used to study optimal methods of producing low thermal conductivity nanocomposites. We consider the specific case of Si 0.5 Ge0.5 alloy matrix materials as the active thermoelectric component and nanoparticle scatterers including Ni, Co, and Pt silicide compounds. The thermal conductivity tensor is studied as a function of fiber size, elastodynamic contrast, and the degree of orientation in an effort to minimize thermal conductivity for a given volume fraction of nanoparticles. The optimal fiber size is found to be quite small, corresponding to fiber diameters of only ∼2nm. For the range of practical materials studied, the anisotropy of the thermal conductivity tensor is found to be <2 for aligned fibers. PtSi was found to be the most effective silicide material due to its high density and elastic contrast. Compared to an unnanostructured Si-Ge alloy, the addition of 3.4% volume fraction of PtSi is found to reduce the thermal conductivity by a factor of 9 for fully aligned and optimally sized fibers with heat flow perpendicular to the fiber axis. This is found to be more effective than an equivalent amount of PtSi utilized in the form of optimally sized spherical nanoparticles. In some cases like NiSi and CoSi 2 , we find two locally optimal nanofiber sizes due to the separate optima associated with each phonon mode. So, a multimodal distribution of nanofiber sizes can be observed to be optimal.