Atomistic Characterization of Thermal Conductivity Modulation by Nanostructuring


Waste heat recovery and heat dissipation are two of the most important research topics right now due to the world energy crisis and thermal destruction. It has been proved that nanomaterials have the potential in effective utilization and sustainable development of energy. Thermoelectric materials, which can convert thermal energy into electric power, provide a potential pathway to solve the world’s energy crisis. The most popular method to obtain the high ZT coefficient, which depends on the power factor positively and thermal conductivity negatively, is to decrease the thermal conductivity tremendously while keep the power factor as a constant. Nanostructuring the materials has been proved to be a quite efficient way to reduce the thermal conductivity then reach a very high ZT value1. Therefore, understanding the underlying mechanisms how nanostructuring can affect the thermal and electrical properties in nano structures is quite important for designing competitive thermoelectric candidates.

In my PhD research, I focus on the methods to characterize the phonon level thermal conductivity of the nanostructures, such as nanowire (NW), nanomebranne (NM) and bulk nanostructures (BNS). Currently, the two most popular approaches to predict the thermal conductivity in MD simulations are the equilibrium molecular dynamics (EMD) simulation based on the linear response theory and the non-equilibrium molecular dynamics (NEMD) simulation which is just applying the Fourier’s law directly. However, both EMD and NEMD can not provide the phonon level information. Recently, the increase of computational efficiency makes the rigorous calculation of individual phonons’ contribution possible. In the framework of EMD simulations, the spectral energy density (SED) or equivalently the time domain normal mode analysis (TDNMA) method, which only considers the depopulation of phonons, is the most popular way to predict the phonon behavior. On the other hand, from the aspect of NEMD simulations, th

e time domain direct decomposed method (TDDDM) and the frequency domain direct decomposed method (FDDDM), where the key equations are deduced by Kimmo et. al., were generated to obtain the detailed information of phonons by Y. G. Zhou et. al. Most recently, the FDDDM is also extended to deal with the interface problem. Meanwhile, the electrical transport properties of the nanostructures are also considered by the semi-classical Boltzmann transport equation (BTE).

The goal of my PhD thesis is to propose methods, to calculate the contributions of individual phonon mode contribution and electrical properties of nanostructures. Then, using such methods, we can predict the thermoelectric performance of the nanostructures.