July 1998
Contributors: X Fan, M Shaffer
Carbon nanotubes are a radically different type of materials [1]. In one sense they can be classified as ultra fine fibres, in another as rather fat and stiff examples of polymer molecules. Their commercial potential as exciting new types of filler for polymeric materials is thus also matched by a considerable level of scientific fascination. The project, being in a new area, has grown rapidly, and in some instances in unexpected directions. In particular it has lead to the entry of Materials Chemists into the field who are opening up a new route to the production of the carbon nanotubes themselves. The availability within the department of both high resolution scanning (FEGSEM) and transmission (TEM) electron microscopy was critical to this work. The TEM was able to image the graphitic planes, and reveal the thin amorphous layer of carbon on the surface of the tubes, whether treated or not (See upper frontispiece). The FEGSEM, with resolution down to one nanometre, provided quite unparalleled images dry aggregates of the nanotubes, revealing local orientational order where it was present, as well as the degree of entanglement and straightness (See lower frontispiece and below).
The carbon nanotube materials were received from Hyperion (Cambridge Mass) in the form of loosely entangled aggregates [2]. They were examined using very high resolution electron microscopy and high resolution TEM. The key step to processing as a potential filler was to learn how to form stable dispersions of them in aqueous media. The as received nanotubes were treated with a mixture of concentrated nitric and sulphuric acids (1:3) and refluxed at 135°C for 30 min. The treated tubes were washed until the residual acidity was at a minimum. The surface treatment was shown to involve the addition of both carbonyl and hydroxyl groups, the former associated with carboxylic acid and lactone groups, the latter with carboxylic acids and phenolic groups - as shown by FTIR spectroscopy.[3] High resolution TEM showed that the crystalline structure of the nanotubes was unaffected by the acid treatment, and that their mean length was around 1 mm. The distribution of lengths corresponded to a polydispersivity [Mw/Mn] of 1.6. Once oxidised the nanotubes were shown to spontaneously disperse in water and remain stable over many months. The stability of the dispersions appears to arise from the negatively charged surfaces resulting from the oxidation. Precipitation could be induced by the addition of ionic additions, particularly acidic ones which, in addition to reducing the Debye screening length, also reduce the degree of surface ionisation. The production of stable dispersions opens up routes to handling the tubes, to creating concentrated aggregates, gels and liquid crystalline type order. It is also shown to be the most agile route to the incorporation of nanotubes in polymeric matrices.
The near-molecule type dimensions of the nanotubes tempted us to try to reduce their length through the application of ultrasound, which is know to be a route to the reduction of molecular weight of a polymer molecule. However, the strength of the nanotubes seemed to be a dominant factor, and while damage at high power levels could be detected it did not lead to any neat scission of the tubes. Instead a bromine method was adopted in which the nanotubes are sealed in liquid bromine at room temperature for 24h. The nanotubes are then oxidised in 4% oxygen at 450°C for periods of 24h. The result was a reduction of tube length. It appears that the bromine intercalates into imperfect regions of the tube, which then preferentially decomposes under the oxidation [4]. Subsequent acid oxidation of the type used on the original material again yielded an aqueous dispersion. (iii) Graphitisation As with other carbon materials, graphitisation improved the crystallinity [5]. Treatment of the nanotubes at 2800°C in a vacuum carbon tube furnace, concentrated the bend defects distributed along the tubes into much more localised kinks. Experiments to bromine degrade such kinked tubes, into what might be short but very straight segments are still in progress. The process also removed the thin amorphous layer of carbon seen on the as received and surface treated fibres.
Suspensions of nanotubes in water are analogous to polymer molecules in solution, although their separation is ensured more by surface repulsion than by entropy derived from the chain flexibility. A series of experiments whereby the viscosity of the nanotube suspensions was measured as a function of concentration using a capillary, showed that the viscosity increased gradually until a critical point at about 0.7 vol% where there was a very rapid increase in viscosity typical of entanglement transitions in polymer solutions. The viscosity data could be fitted to the Schulz-Blaschke equation for polymer solutions, originally an empirical relationship but subsequently justified on theoretical grounds by Huggins [6]. It is possible to achieve continually higher concentrations by permitting the aqueous medium to evaporate. The increasing viscosity leads to a gel which forms around 5 vol% nanotubes. A full rheological study of this rather novel gel is projected for the developing programme. In the case of nanotubes whose length had been decreased as described above the mean length was 370 nm whilst the polydispersivity was now 1.7. In this case the rapid increase in concentration occurred at a predictability higher concentration of around 1 vol%.
The stiffness of the carbon nanotubes would suggest that their natural state of aggregation would be as parallel clusters. However, the same stiffness means that defects which cause deviations from straightness will compromise significant tube parallelism. Solid nanotube films were prepared by filtration onto a 0.2 mm membrane filter under 0.6 bar negative pressure. Examination of the films using the high resolution FEGSEM showed two particular regimes. For dispersions with concentrations greater than about 0.3 vol%, a random arrangement of nanotubes occurs in the plane of the film, although there is a high level of alignment into the plane. For concentrations below this critical value, local mutual alignment was clearly apparent (see SEMimage). The films showing such alignment were tougher than those with random in plane orientation. The films had high densities of around 1.2 gcm-3 corresponding to a packing fraction of 65%. SEM image of a film of nanotubes filtered from a low concentration dispersion The degree of alignment was characterised by obtaining the orientation function as a function of area examined from fast Fourier transformation of the digitised images. At 0.2 mm (approximately 10 tube diameters) the value of the orientation function was 0 for the unaligned sample, and 0.2 for the aligned one filtered from the dilute gels. These data are evidence for a highly localised liquid crystallinity in appropriately prepared films [3]. Further work will endeavour to increase the volumes over which orientational order is achieved through the use of graphitised and bromine cut nanotubes (i.e. shorter and straighter).
The continual evaporation of the aqueous medium from the suspensions leads to nanotube aggregates of ever increasing 'concentration' until a brittle grey-black solid is formed. The density of this solid was measured to be 0.9 ±0.1 gcm-3. This value corresponded to a packing fraction of nanotubes of the order of 50%, which was remarkably high for an aggregate of non-parallel tubes. It is considered that local tube bending occurs as the maximum density is approached, so that the available volume is filled as efficiently as possible. It was also interesting to note that re-addition of water to the solid resulted in a reversible swelling to a concentration of approximately 25 vol%. However, the tubes did not re disperse, and would not again form dilute solutions.
Material not yet released by industrial collaborators.
The work has also been presented at the following conferences:
![[icon]](images/Poly_tiny.gif)
