Date of Award

Spring 1-1-2012

Document Type


Degree Name

Doctor of Philosophy (PhD)


Mechanical Engineering

First Advisor

Ronggui Yang

Second Advisor

Yung-Cheng Lee

Third Advisor

Scott Bunch


Nano-scale thermal conduction has been an interesting research topic over the past two decades due to the intriguing electron and phonon physics in the low-dimensional materials, the ever-increasing challenges in the thermal management, and the potentials in using nanostructures for enhanced energy conversion, storage and thermal management. Atomistic simulation methods, such as molecular dynamics (MD) and atomistic Green's function (AGF), have been powerful theoretical tools for understanding and predicting the phonon transport and thermal conduction in nanostructured materials and across dissimilar interfaces. The research conducted in this thesis aims at further developing these atomistic simulation methods and broadening the understanding of phonon transport in nanostructured materials and across the interfaces of dissimilar materials. There is strong artificial simulation-domain-size effect on the prediction of thermal conductivity using equilibrium MD (EMD) for high thermal conductivity materials. In this thesis, spectral method is further developed to avoid the simulation domain size effect and obtain the intrinsic phonon thermal conductivity of materials from EMD simulations. This method is then applied for the study of strain effects on thermal conductivity of low-dimensional nanostructures. We show that thermal conductivity of nanostructures can be greatly tuned by applied mechanical strains. The underlying mechanism on strain/stress tuning of the one-dimensional and two-dimensional nanostructures of carbon and silicon are explained. Phonon transport across dissimilar material interfaces plays a critical role in determining the effective thermal properties of nanostructures. Although AGF has been widely used for the prediction of frequency-dependent phonon transmission across material interfaces, most of the past works have been limited to lattice-matched interfacial structures. In this thesis, an integrated MD and recursive AGF approach is developed to study the frequency-dependent phonon transmission across lattice-mismatched dissimilar interfaces. Due to the induced atomic disorder, lattice mismatch greatly affects the phonon transmission across dissimilar material interface. The reduction of thermal conductance is correlated with adhesion energy at the interface. Further studies are performed for phonon transport across interfaces with defects and specie diffusion. The numerical tools developed in this thesis can be useful for the understandings of phonon transport in nanostructured materials.