Computational modeling and spectral analysis of nanoscale energy transfer

Swift progress in the synthesis and processing of materials with nanoscale feature sizes has spawned new possibilities to control the flow of thermal energy. New materials and devices with engineered thermal properties are expected to enable, e.g., clean and more efficient production of energy from waste heat by thermoelectric converters, reducing the energy consumption of digital electronics, and generating novel technologies such as heat-assisted magnetic recording and phase-change memories. As the classical laws of energy transfer do not generally apply in nanoscale, practical realization of such applications calls for powerful computational methods delivering scientific understanding of nanoscale heat transfer. The goal of this thesis is to develop new computational models and methods for describing energy transfer in atomic-scale structures and to apply the methods to generate useful insight into various thermal phenomena. The work is founded on classical molecular dynamics simulations and quantum-mechanical Green's function approaches, both using the fluctuation-dissipation theorem to couple the studied systems to external heat baths. To enable detailed analysis of energy transfer mechanisms in thermal conduction, new methods to spectrally decompose the lattice heat current into frequency components are also developed. Spectral analysis is applied in the thesis to identify non-linear energy transfer mechanisms at material interfaces and to determine the mean free paths of heat carriers in carbon nanotubes. The results also suggest that the thermoelectric efficiency of silicon nanowires can be increased by a specific superlattice structure and that the electromagnetic energy transfer rate between dielectric nanoparticles can be tuned by a mirror cavity. In addition, the thesis initiates the development of a unified fluctuational model for describing energy transfer by lattice vibrations, electromagnetic fields, and electrons in a single mathematical framework that can generate extensive understanding of the energy conversion phenomena present in small structures. As a whole, the methods and results of the thesis provide new analytical and numerical tools for describing nanoscale energy transfer within a framework that may, with further development, become instrumental also in modeling energy conversion and transfer processes in multiscale systems involving heat, light and electricity.