Thermoelectric Transport in Disordered Materials

Material disorder comes in two forms, dynamic energetic disorder and static energetic disorder, both of which can greatly influence charge and thermoelectric (TE) transport properties by altering the energy dependences of the density-of-states (DOS) and carrier mobility. Yet, despite known implications for the optimization of charge and thermoelectric transport properties, the computational and experimental literature has traditionally ignored the crucial role of disorder in dictating the energy dependences of the density-of-states (DOS) and carrier mobility. In this dissertation, a systematic framework for characterizing and assessing the optimization of thermoelectric transport in disordered materials that exhibit these large disorder-induced modifications in the DOS and the energy-dependent mobility is presented, along with tail engineering strategies for improving thermoelectric transport at a lower (e.g., intrinsic) carrier concentration by moving carriers from valence states to tail states. While static energetic disorder in highly-disordered (amorphous) materials is shown to impact thermoelectric transport by uniformly decreasing power factor (PF) over the entire range of carrier concentrations (p) accessible by dopants, dynamic energetic disorder in disordered soft material systems is shown to greatly undermine TE performance by causing the Seebeck coefficient, electrical conductivity, and thermoelectric PF to attain their minimal metallic values in the limit of strong dopant-induced energetic disorder. As a consequence, getting to the optimal carrier concentration (and thus the optimal value of PF) in this material class requires more than 10 times the carrier concentration needed to optimize PF in a highly-disordered (amorphous) material or a material in which no disorder is present. These inherent limitations on the doping of disordered soft materials, in particular, are shown in this dissertation to be circumvented when mode-selective vibrations are used to create a specific kind of resonant distortion within the DOS tail that causes states on one side of the Fermi window to converge (thereby creating a region of carrier energies with high mobility) and states on the other side of the Fermi window to diverge (thereby creating a region of carrier energies with relatively low mobility). As a result, a very large room temperature thermoelectric PF is predicted to occur within the gap or tail region of the DOS (cf. state-of-the-art bismuth telluride), lending way to a novel approach for optimizing PF that doesn't rely on the introduction of dopants to maximize PF near the band edge. The potentially important implications of these findings for bulk thermoelectric transport as well as the measurement of thermoelectric properties in field-effect transistor geometry are discussed.PHDMechanical EngineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/174284/1/swbirch_1.pd