Research into carbon nanotubes (CNT):polymer composites for thermoelectric applications have greatly expanded in recent years, aiming to combine the high power factor of CNT networks and the low thermal conductivity of polymer matrix. Many attempts have been made on achieving the electron-crystal-phonon-glass concept and the simultaneous increase of electrical conductivity and the Seebeck coefficient. However, the thermoelectric transport within composites has not been fully understood due to the complexity in materials properties, morphologies and experimental methods. This thesis reports the thermoelectric properties and transport in composite materials by studying a model system that embeds highly-purified semiconducting single-walled carbon nanotube network (s-SWCNTs) with poly(2,5-bis(3-alkylthiophen-2-yl)thieno[3,2-b]thiophene) (PBTTT). s-SWCNTs and PBTTT are selected due to their fairly well-studied thermoelectric properties and to achieve a good understanding of their charge transport mechanisms. Through investigating the s-SWCNT:PBTTT model system, some general design guidelines for optimising the power factor and figure of merit of CNT:polymer based composites are established. Charge accumulation spectroscopy (CAS) can reveal the charge density of a semiconducting layer. The charge distribution within the composite is investigated by the CAS. The charge carrier distribution is used to model macroscopic electronic properties such as electrical conductivity and linked to the Seebeck coefficient. Insights into the transport energetics and the density of states (DoS) are studied by the charge density and temperature dependence of the gated Seebeck coefficient. The electronic and thermoelectric properties are modelled by a heterogeneous transport that involves three conduction pathways, CNT-CNT, PBTTT-PBTTT and CNT-PBTTT. The CNT-PBTTT pathway plays a significant role in the charge transport of composites. Firstly, it is demonstrated that power factors of the composites can be optimised by tuning the density of CNTs in composites, doping species and doping levels. The important role of thermal conductivity in designing thermoelectric composites is highlighted as one of our composites with a lower power factor exhibiting a higher value of ZT. It has also been found that the choice of dopants can tune thermal properties. Insights into the thermal transport are studied by measuring the temperature-dependent thermal conductivity and the Lorenz number.

Subsequently, the charge densities of charge carriers in CNTs and PBTTT within the composites are characterised by the CAS. It is shown that the distribution of charge carriers depends on the network density of CNTs in composites and the total charge density. More than 70% of charge carriers accumulate at low charge densities in the 3wt% CNTs, highlighting that the volume/mass ratio of components in composites should not be used to model charge transport. The charge density-dependent distribution and transition of the dominant phase for charge carriers affect the macroscopic electronic properties such as the mobilities. The charge transport in composites can be described by overcoming an energetic barrier by thermal activation. It is found that the activation energy increases with charge density which is the opposite trend of the neat CNTs and PBTTT, suggesting that the CNTPBTTT interface cannot be neglected. The mobilities of composites are well modelled by the three-conduction pathway analytically model. Contribution percentages of CNT-CNT, PBTTT-PBTTT and CNT-PBTTT are extracted. These findings suggest that the CNT-PBTTT pathway contribute significantly to the charge transport, and the competition among the three pathways determines the electronic properties of composites.

The charge density and temperature-dependent Seebeck coefficient of composites are measured to gain insights into transport energetics and the density of states (DoS). A simultaneous increase of the Seebeck coefficient and electrical conductivity with increasing charge densities is observed in the 3wt% s-SWCNT:PBTTT composite. This positive slope of S(n) cannot be explained using a single transport model, and therefore we propose an analytic model that considers the Boltzmann transport formalism and charge transport in heterogeneous media. The unusual behaviour is explained by the transition of the dominant phase of charge carriers and the separation of Fermi and transport levels into different regimes of the DoS. These results demonstrate that blending two materials into composites might create unique transport energetics to achieve unusual thermoelectric properties. Observations and analysis of the current model system provide design guidelines towards CNT:polymer or other composites for thermoelectric applications.