Ab initio study of thermoelectric phenomena via the exact solution of the Boltzmann equation

Student thesis: Doctoral ThesisDoctor of Philosophy


A detailed understanding of electrical transport and energy dissipation phenomena is crucial for the discovery and engineering of new, high-performance materials for application ranging from nano-electronics to thermoelectric energy conversion. The experimental investigation of charge and heat conduction in controlled conditions is often challenging and expensive, and therefore accurate and efficient theoretical and simulation approaches are pivotal to foster new advancements.
In my PhD work I have developed an ab initio computational approach to predict electronic transport properties of bulk materials. This work goes beyond the state-of-the-art in the field by merging an accurate, first-principles description of the fundamental interaction between electrons and lattice vibrations, and the exact solution of the Boltzmann transport equation . This computationally challenging task is accomplished by the development of a novel, fully parallel computational infrastructure based on state-of-the-art high-performance computing techniques.
This approach allows the calculation of a range of electronic transport coefficients (electrical conductivity, mobility, electronic thermal conductivity, Seebeck coefficient and Lorenz number), also accounting for the effects of non-equilibrium phonon populations due to thermal gradients. In addition, the analysis tools developed in this work provide accurate ways to examine the microscopic details of the relevant scattering mechanisms and decay channels that determine each specific transport coefficient. These developments are of fundamental importance to assess and validate a number of approximations and phenomenological models that are popular within the electronic transport and thermoelectric communities. In addition, this work provides accurate tools to test design rules to enhance the thermoelectric performance and guide the experimental synthesis and characterization of more efficient compounds.
In this work, I have applied this computational framework to the investigation of many aspects of thermoelectricity in diverse classes of materials, such as metals and doped semiconductors, in a wide range of temperature and, where suitable, doping concentration.
In elemental metals, predictions for resistivity and Lorenz number have been compared against experiments and previous first-principles results, confirming the effectiveness of approximate approaches to Boltzmann transport for these systems. In n-doped silicon, I have focused on a wide range of electric and thermoelectric quantities. In addition to a detailed characterization of resistivity and mobility, this work provides a unique insight on the Lorenz number, a quantity that is not directly accessible with experiments and plays an important role in thermoelectric engineering. Moreover, the analysis of the Seebeck coefficient provides a better understanding of the coupled electron and phonon dynamics in this material, also suggesting strategies towards higher thermoelectric efficiency. Finally, I have analysed the transport properties of boron-doped diamond. Here I have used phenomenological models to simulate the complex doping mechanism in place in this system. I have focused on the prediction of the best transport properties achievable in Boron-doped diamond, in order to provide a useful reference for the experiments that typically display a large uncertainty (due to the difficult assessment of impurity content in synthetic samples). The characterization of the different phonon scattering channels (also in comparison with the case of silicon) offers an insight into the origin of the extraordinary hole mobilities of this system.
Date of Award2017
Original languageEnglish
Awarding Institution
  • King's College London
SupervisorMark Van Schilfgaarde (Supervisor) & Nicola Bonini (Supervisor)

Cite this