Raman and Conduction Studies on Carbon Nanotube Networks and Cupric Oxide Nanostructures

Abstract

This thesis presents detailed temperature-dependent Raman and conduction studies on two materials: cupric oxide (CuO) nanostructures and single-walled carbon nanotube (SWNT) networks. SWNT networks are a promising alternative to indium tin oxide as the transparent conducting material in electronic displays. A key factor that complicates fundamental studies on SWNT networks is surfactant residue. Our study utilises SWNTs (HiPCO) dispersed in volatile solvents, which can be removed by annealing to only 100°C. Using this unique solvent system, we have systematically studied charge transport mechanisms in surfactant-free networks using a percolation approach; where sample resistance can be controlled by the amount of deposited material. The chemical environment of these networks was found to be unchanged using Raman spectroscopy; i.e. film fabrication did not cause any significant doping of the network. Around 15 surfactant-free networks, with resistances between 300 kΩ and 8 kΩ, were found to follow a 'universal' charge transport model. Most of the networks could be described by two dimensional variable range hopping (VRH) and thermal activation. The barrier energy or, To parameter, for the VRH mechanism was independent of resistance with a value of 20x 10ᵌ ± 9.4x 10ᵌ K. The activation energy also had a resistance-independent value of 160 ± 20 meV. Four terminal measurements confirm that the activation mechanism is due to processes within the network not, Schottky barriers at the nanotube/metal interface. The effects of extrinsic adsorbants on the network resistance provides evidence for dominant non-metallic conduction pathways within the studied range of resistances. These results strongly suggest a characteristic barrier size in our SWNT networks where non-metallic tubes dominate the resistance. The surfactant-free networks were also used to study the temperature-dependent behaviour of the radial breathing modes (RBM) in bundled nanotubes. Deconvolution of complex RBM spectra was made possible using an interactive routine: based on the higher resolution of second derivatives for fitting spectra with washed-out features. Using this routine, the temperature-dependent characteristics of the RBM lineshape could be identified. We find that RBM modes in our bundled networks soften at the same rate as individual tubes; the linewidth follows a three phonon decay process with a temperature-independent component and the intensity can be modelled from the change in Eii with temperature. The second part of this thesis addresses the unique asymmetric lineshape of the A1g mode in CuO nanowire forests. A symmetric lineshape is recovered at low powers indicating that the underlying mechanism is thermal. To study this effect, the high temperature behaviour of the A1g mode is first analysed in a `bulk' form of CuO. The analytical temperature dependence of the A1g mode frequency, linewidth and intensity were used as the basis of a physical model that connects lineshape asymmetry to laser-induced, spatial temperature gradients in the sample.The peak temperature (under the laser hotspot) was found to be proportional to laser power until it reaches a critical value. We believe that regions with temperature above the critical value cool by radiation rather than convection.