The polymer industry is continually researching new materials that offer increased performance at lower costs, often necessitating the introduction of fillers. Nanoparticles offer significant advantages over macro-or micro-sized fillers, including a greater surface area to mass ratio, low percolation threshold and often very high aspect ratios, and consequently there has been significant research and investment into the production of polymer-based nanocomposites. Much of this research has focussed on carbon nanotubes (CNT) in this role, although few studies have demonstrated their significant potential. In this thesis a range of techniques, including ball milling and covalent functionalisation, were employed to create several varieties of carbon nanotubes. The effect of these well characterised CNT samples on the mechanical properties of thermoplastic polyurethanes (TPUs) was investigated.
The effectiveness of shortening CNT by ultrasonication, chemical treatment and high-energy ball milling was investigated in a preliminary study. High-energy ball milling was chosen as the preferred option due to its high throughput and effectiveness at shortening CNT samples. An increase in ball milling time corresponded to a reduction in CNT median length; however, fine control of sample length distribution was not possible. Ball milling was also found to qualitatively increase amorphous carbon content, whilst a slight increase in side wall oxidation with increased ball milling time was also observed.
The impact of CNT morphology and length distribution on the properties of TPU-based nanocomposites was investigated through the production of a series of solution processed nanocomposites. CNT length distribution and morphology was found to have minimal impact on nanoscale dispersion, microphase morphology and mechanical behaviour of the resultant nanocomposites. Tensile testing did reveal a small increase in tensile strength with decreasing CNT length, which was accomplished without the traditional loss of the host TPU’s ‘rubbery’ behaviour. However, the improvements were minimal and ultimately not statistically significant. In contrast, the TPU / purified multi-walled carbon nanotube (pMWNT) nanocomposite, demonstrated acceptable dispersion, a much higher interfacial shear strength compared to the other nanocomposites and consequently demonstrated enhanced mechanical properties, even under cyclic loading.
In conjunction with the investigation into the effects of CNT morphology, this study also explored a variety of processing methodologies in an attempt to optimise CNT dispersion and mechanical
CNT dispersion. However, the mechanical properties of all materials, including the host TPU, were decreased with increased melt processing time, and this behaviour was significantly worse in the CNT-based nanocomposites. It was concluded that the poor performance of the melt processed samples was primarily caused by TPU degradation, exacerbated by the addition of CNT. behaviour of the nanocomposites. Ultimately, processing methodology had minimal impact on
Based on these findings, the functionalisation of the carbon nanotube materials, to improve both dispersion and CNT interactions with the host TPU, was investigated. Specifically, the effect of functional group length and conformation was studied through several iterations to select an optimal functionality. Tensile testing of rudimentary nanocomposites formed from these functionalised CNT materials, showed that an increase in hydrophobic tail length resulted in an increase in tensile strength and a reduction in Young’s modulus. As a result, octadecylamine with a toluene 2,4-diisocyanate linker was chosen as the optimal functionality for use in TPU-based nanocomposites. Several varieties of CNT materials were later modified with the optimal functionality, although conversion rates were lower than expected due to poor CNT dispersion during the reaction process. The impact of this functionalisation on the properties of the resultant nanocomposites was investigated. Contrary to expectation, dispersion was slightly reduced for all nanocomposites, whilst the mechanical properties remained largely unchanged. It was thought that reduced dispersion and apparent preferential functionalisation of CNT end caps over sidewalls were major contributing factors in the failure of the functionalised CNT nanofillers to improve tensile behaviour.
Finally, to investigate the effects of CNT loading on the properties of TPU-based nanocomposites, the pMWNT sample was chosen as the optimal nanofiller. Scanning electron microscopy demonstrated that dispersion, specifically the ratio of aggregates to individual nanotubes, was unaffected. As expected, Young’s modulus increased with pMWNT loading; however, the trend observed did not follow any established predictive models. In contrast, tensile strength peaked at 3wt% implying that the sheer number of aggregates present overwhelmed any additional reinforcement provided by the increased loading. It was concluded that the optimal nanocomposite was the TPU / pMWNT (1%) sample based on the significant increase in tensile strength obtained without compromising its ‘rubbery’ behaviour.
This work has contributed to the understanding of TPU / CNT nanocomposites and to the greater field of CNT-based polymeric nanocomposites. Furthermore, this thesis continued the development of a novel microwave-assisted functionalisation procedure. The failure of the functionalisation to greatly enhance CNT dispersion and mechanical behaviour in the nanocomposite, despite
acceptable functional group loadings is significant for the field. It highlights not only the need for improved dispersion during functionalisation but also the need for preferential functionalisation of the sidewalls, over the CNT end caps.