Quantum Thermal Transport in Semiconductor Nanostructures

Kai Miao, Purdue University

Abstract

Modern semiconductor devices scale down to the nanometer range. Heat dissipation becomes a critical issue in the chip design. From a different perspective, energy conservation has attracted much of attention from researchers. The essence of heat dissipation and energy conservation is the heat transport. Thermal properties of semiconductors have been under intense investigation in recent decades. Classical models fail to consider the quantum effects in devices on the scale of nanometers. First-principle methods only can deal with small devices and is computationally intensive. Instead, a modified valence force field (VFF) model is applied to reproduce the phonon properties of different materials and devices. Phonon transport is explored using the Green’s functions. The concept of a Büttiker probe model is first used to mimic the scattering mechanisms in phonon transport. This energy conservation model is straightforward and efficient in describing scattering. In the quasiparticle approximation, phonon scattering will cause a phonon energy shift. This energy shift is represented by the scattering self-energy in a retarded Green’s function. Phonon lifetime is extracted from the scattering self-energy expression. Different relaxation time approximation (RTA) models are studied and coupled with the phonon Green’s function method for the first time. We prove that the widely used and proven RTA models in the Boltzmann transport equation (BTE) survive in the atomistic Green’s function method. This method can give accurate thermal properties agreeing closely with the experimental results for bulk devices. This atomistic method can also consider quantum confinement effects at the nanoscale. The heat transport across a Si/Ge interface is introduced in this work as an example for this application. The heat transfer across metal/semiconductor (MS) interfaces is investigated as well. Relaxation at the interface can be done in two different ways. Using VFF model to relax the interface and the second, using DFPT tool to relax the structure. Both methods show that thermal conductance increases after relaxation. The local temperature can be extracted for studying heat transfer. Our model shows that interfacial resistance is largely independent of the scattering rate and device length. Comparisons between our model and other models are introduced. The results show that the deviation is due to the interatomic force constant and scaling effects. Next, we introduce a new self-consistent electron-phonon coupling model. This model is still based on the concept of Büttiker probe and can capture the energy exchange from electrons to phonons. Proper scattering models are chosen for the Si and Ge. The simulation results show that in the normal case, the effect of electron-phonon scattering in the semiconductor is limited. For study purposes, we increase the scattering strength to understand the effect of large scattering rates on the overall transport simulation. Generally, temperature changes more rapidly as long as we include electron-phonon coupling. Finally, a summary is presented as well as an outlook for future work. The main improvement will be to implement the recursive Green’s function with the Büttiker probe model.

Degree

Ph.D.

Advisors

Klimeck, Purdue University.

Subject Area

Electrical engineering|Nanotechnology

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS