The CVD process for the spinning of carbon nanotube (CNT) fibres combines the nucleation, growth and aggregation of CNTs in the form of an aerogel with fibre spinning into a single process step. The optimisation of the process requires agility in multi-dimensional parameter space, so one tends to find parameter 'islands' where spinning is possible, while exploration tends to follow 'routes' through this space. Here, we follow two such routes, one of which drastically improves fibre purity, the other changes the nature of the nanotubes comprising the fibres from multiwall to single wall. In the first case there is only a modest enhancement of the mechanical properties, but in the second a very considerable improvement is seen. In terms of the conditions required to make fibres comprising predominately single wall CNTs, the key factor appears to be the rigorous control of the sulphur addition, in trace quantities, coupled with the availability of carbon atoms at the earliest stage after injection, typically in the range 400-500 °C. A model is presented for the role of sulphur in floating catalysts CNT synthesis.

Thermal conductivity of carbon nanotube (CNT) films and micro-fibres synthesised by floating catalyst chemical vapour deposition was measured by the parallel thermal conductance method. CNT films showed in-plane thermal conductivities of 110 W m 1 K 1. Online condensation into a micro-fibre morphology-a two-dimensional reduction in the transverse plane, including some axial stretching during solvent evaporation-resulted in room-temperature thermal conductivity values as high as 770 ± 10 W m 1 K 1 , which is the highest thermal conductivity reported for CNT bulk materials to date. In specific terms, this matches the maximum thermal conductivity of heattreated carbon fibre, but with a higher onset temperature for Umklapp scattering processes (300 K rather than 150 K). We selected four sample types to investigate effects of alignment, purity, and single-or multi-wall character on their thermal conductivity. For both the electrical and thermal conductivity of as-spun material, i.e. without any post-synthesis treatment, we show that the density and quality of CNT bundle alignment are still the predominant factors affecting these properties, outweighing the influence of single-or multi-walled character of the nanotubes. This raises the promise that, with optimal alignment and junction points, even higher values of thermal conductivity are achievable for macroscopic CNT fibres.

No abstract

A protocol has been developed for the production of epoxy-based composites containing high-volume fractions of aligned carbon nanotubes. The nanotubes were fabricated as continuous fibres or aligned mats directly from the CVD reactor, in which they were synthesized. The block composites with highly aligned and tightly packed nanotube assemblies were prepared via epoxy resin infiltration, and their volume fraction, distribution, and internal porosity being analysed prior to mechanical testing. The samples were tested in both axial tension and three-point bending. The results show that the strength and stiffness enhancements were close to pro rata with the volume fraction of the carbon nanotubes added. The failure modes were distinctly different from those characteristic of the conventional aligned carbon fibre composites. The fracture surface showed considerable evidence of pull-out of bundles of (*50) nanotubes, but the pull-out appeared to involve the resin matrix which drew out along with the bundles. Subsidiary cracks were bridged by nanotube bundles giving structures reminiscent of crazes in glassy polymers, what constitutes the distinct toughness mechanism and higher resistance to the transverse cracks propagation.

The uniaxial orientational order in a macromolecular system is usually specified using the Hermans factor which is equivalent to the second moment of the system's orientation distribution function (ODF) expanded in terms of Legendre polynomials. In this work, we show that for aligned materials that are two‐dimensional (2D) or have a measurable 2D intensity distribution, such as carbon nanotube (CNT) textiles, the Hermans factor is not appropriate. The ODF must be expanded in terms of Chebyshev polynomials and therefore, its second moment is a better measure of orientation in 2D. We also demonstrate that both orientation parameters (Hermans in three dimensional (3D) and Chebyshev in 2D) depend not only on the respective full‐width‐at‐half‐maximum of the peaks in the ODF but also on the shape of the fitted functions. Most importantly, we demonstrate a method to rapidly estimate the Chebyshev orientation parameter from a sample's 2D Fourier power spectrum, using an analysis program written in Python which is available for open access. As validation examples, we use digital photographs of dry spaghetti as well as scanning electron microscopy images of direct‐spun carbon nanotube fibers, proving the technique's applicability to a wide variety of fibers and images.