Manipulating the percolation threshold of carbon nanotubes in polymeric composites

Abstract

The latex-based technique to introduce carbon nanotubes (CNTs) into polymers has shown to be highly versatile and to produce conductive composites with low loadings of CNTs (<1wt%). For certain applications, these loadings are still too high to be commercially viable. In this work we have examined various manipulation strategies by which the minimum loading of CNTs required for composites conductivity, also called the percolation threshold, is reduced. By using a latex-based composites production technique, segregated CNT networks are formed in the final composite structure. These networks may undergo conformation changes during the various processing steps unique to the latex-based route. In Chapter 2 the role of these conformation changes on the final percolation threshold is investigated. Plasticized matrixes are prepared using low molecular weight polymer and surfactants. By plasticizing the matrix, the probability of network conformational changes in the melt is increased. It was seen that by allowing network changes, or equilibration of the network structure, through lowering the matrix viscosity, reductions in the percolation threshold could be induced. The process by which the initial colloidal mixture is lyophilized was also found to influence the percolation threshold. From all these observations, it is clear that the CNT network structure undergoes substantial conformational changes when processed from the initial rod-sphere colloidal mixture, through the packing of spheres and rods ending in the liquid-like melt-state. While distinct changes in the percolation threshold of the CNT – polymer composites were observed, only small changes in the ultimate conductivity were observed between all the systems studied. A large reduction in the ultimate conductivity was only observed when filmforming polymers were used in ombination with fast lyophilisation. Not only does the network formation play a large role in the electrical conduction of a CNT network, the nature of the junctions within the network is of a high priority. The high conduction of CNTs is never transferred to CNT – polymer composites purely due to the high resistance of CNT junctions. Interfacial polymer layers act as electron tunnelling barriers that limit the conductivities of the final composites. By increasing the conductivity of these interfacial polymer layers, it was hoped that the composite conductivity would increase. Chapter 3 deals with systems where conductive polymeric surfactants were utilised as surfactants for the dispersion of CNTs in water. The close interaction between the conductive polymeric surfactants and the tubes results in film of polymer over the CNT surface. In this way, the conductivity of the polymer in the inter-tube junctions is assured to be high. Increases in the ultimate composite conductivity, along with reductions in the percolation threshold, were observed. Modelling of these three component systems was performed to understand the co-operative nature of the conductive polymer and carbon nanotubes. Due to the preparation technique required to prepare these composites, the concentration of conductive polymer was always much higher when compared to the CNTs. For this reason the contribution of the CNTs to the final composite conductivity was questioned. Replacing the highly conductive CNTs with a lower quality gave similar percolation thresholds and ultimate conductivities. This substantiates the fact that the use of a conductive surfactant negates the contribution of the CNTs, and that the role of the CNTs is predominantly that of a template-like structure that organizes the conductive polymer. The use of conductive material to bridge adjacent CNTs was investigated in Chapter 3. This concept was expanded in Chapter 4 where a second filler was introduced into a two component composite, in the first instance a semi-conducting filler and in the second a conductive filler. Different synthetic techniques were utilized for the two fillers. It was realised that the loading of the second filler would be severely restricted by the synthetic steps. Little differences were observed, with respect to the ultimate conductivity and percolation threshold, upon the introduction of the second filler. This is attributed to minimal interactions between the CNTs and the second filler, and the low loadings of the second filler. Polymer blends have been used to reduce the percolation threshold of conductive fillers in polymeric composites. This technique uses large volume fractions of polymer with low filler affinity as an excluded volume which constricts the localization of the filler. In Chapter 5, the use of air as such an excluded volume is investigated. Inverse emulsions, and more specifically high internal phase emulsions (HIPEs), were utilized to prepare foam-like polymer – CNT composites. The percolation threshold observed for the foam composites was lower when compared to nonporous composites when the foam morphology was kept independent of the CNT loading (conventional polyHIPEs). For systems where the foam morphology was dependent on the CNT loading, called Pickering-polyHIPEs, a distinct insulator-conductor transition was never observed. This was attributed to the CNT concentration-dependent morphology. Carbon nanotubes have for a long time been hailed as the conductive filler that would take over the electronics world. Many efforts have shown that this might have been too wild a dream. To gain exclusive use as the ultimate conductive material, CNT-mats and CNT – polymer composites must still gain much ground. In this light, the manipulation strategies discussed in this thesis are critically evaluated in Chapter 6. Whilst many of the strategies achieved their main aim, the final product characteristics will often fall short of the application required. For applications in field-effect transistors (FETs), the electrical conductivity of CNT – polymer composites prepared by latex technology proved to be too high. This system should further be optimized in such a manner that the concentration of metallic CNTs are reduced giving rise to a semi-conducting nature of the network.