Elsevier

Polymer

Volume 40, Issue 21, October 1999, Pages 5967-5971
Polymer

Polymer Communication
Development of a dispersion process for carbon nanotubes in an epoxy matrix and the resulting electrical properties

https://doi.org/10.1016/S0032-3861(99)00166-4Get rights and content

Abstract

To avoid electrostatic charging of an insulating matrix an electrical conductivity above σ=10−6 Sm−1 is needed. At present, the most common practice to achieve this conductivity is to use a conductive filler such as carbon black. In this work, untreated catalytically-grown carbon nanotubes were dispersed in an epoxy matrix. After curing the epoxy, the electrical properties of the composite were measured in order to relate the filler volume fraction to the electrical conductivity. The intense stirring process used to disperse the carbon nanotubes has made it possible to achieve a matrix conductivity around σ=10−2 Sm−1 with filler volume fractions as low as 0.1 vol.%. These figures represent an advance on best conductivity values previously obtained with carbon black in the same epoxy matrix. These low filler fractions ensure that the mechanical properties of the matrix are not compromised.

Introduction

An increasing number of components are being made from fibre reinforced composites. For example, on aircraft, radomes and the leading edge of the vertical stabilisers are generally made from glass fibre reinforced composites using an insulating epoxy matrix. For these applications, some electrical conductivity is required to provide electrostatic discharge (ESD), and as electromagnetic-radio frequency interference (EMI/RFI) protection. Currently, a highly conductive filler such as carbon black (CB) is added to the matrix in order to ensure electrical conductivity above the required level of σ=10−6 Sm−1[1]. This approach reduces the manufacturing and maintenance costs of components as compared with those previously coated with an anti-static paint. The technology is also relevant to other applications where static electrical dissipation is needed, such as computer housings or exterior automotive parts, and where different polymer matrix systems may be involved.

The matrix system investigated in this work is a resin commonly used for fibre reinforced composites for aircraft applications. When dispersing a particulate conductive filler such as carbon black it is important to keep the filler volume fraction as low as possible in order to maintain the fracture toughness and the tensile properties of the matrix. The epoxy matrix can reach a stiffness of about 3–5 GPa, but it is relatively brittle, with an ultimate elongation of about 4%; this strain to failure is only slightly above the critical value required for aircraft applications. The addition of a filler, which is usually harder than the matrix, generally leads to an increase in the Young's modulus, and a reduction in the ultimate elongation of the matrix. High filler contents also lead to an increase in the viscosity which, in turn, reduces processing ease. For these reasons it is important to minimise the conductive filler loading.

To provide a conductive path throughout a component, a three-dimensional network of conductive filler particles is needed, a situation is known as percolation; the percolation threshold is characterised by a sharp drop in the electrical resistance. Many theoretical percolation models have been developed to define the conditions, especially the critical filler volume content, at which a network is formed in conductive polymer compounds. In the most prominent geometrical models, created by Kirkpatrick [2] and Zallen [3], a regular array on which spherical particles are distributed statistically is examined. In their models, the required minimum network of touching particles is 16 vol.%. This value is in approximate agreement with most experimental determinations since for most polymers filled with powdery materials the critical volume fraction for percolation is between 5 and 20 vol.%. However, this model cannot describe the experimentally observed percolation thresholds for carbon black-filled resins that are conductive at filler concentrations as low as 0.5 vol.%.

Colloid theory describes the interactions amongst particles dispersed in liquids as well as the structure and dynamics of agglomeration. Assuming that the carbon black particles are charged during the preparation, and taking the subsequent interactions into account, the agglomeration process of carbon black dispersed in an epoxy resin was described by Schueler [4]. It was found that the percolation threshold not only depends on the particle size and fractal dimension but also on the shearing rate used to disperse the carbon black particles in the matrix. Using this approach it was possible to both achieve and explain a percolation threshold of about 0.3 vol.%. Further optimisation of this dispersion process by adding copper-chloride has led to a matrix conductivity of σ=10−2 Sm−1 at around 0.06 vol.% [4]. A new preparation method now allows the adjustment of the final resistivity within a range of about seven decades by applying a voltage during the curing of the epoxy [5].

The experimentally observed percolation threshold values strongly depend on the particle shape. Extremely low percolation thresholds have been reported for particles with a high aspect ratio (length to diameter). Dispersing short carbon fibres with an aspect ratio of 280 in an epoxy matrix, Carmona found a percolation threshold of 0.25 vol.% [6]. The recent discovery of carbon nanotubes [7] offers a new opportunity to modify the electrical conductivity of polymer matrix systems. Multi-walled carbon nanotubes are generally conducting [8] and typically have aspect ratios of around 1000. It was expected that a low percolation threshold could be obtained, by dispersing nanotubes in an epoxy matrix using the process developed for carbon black [4].

Section snippets

Experimental

The matrix used in this study is an epoxy polymer based on bisphenol-A resin (ARALDITE LY 556, CIBA GEIGY) and an aromatic hardener (ARALDITE HY 932, CIBA GEIGY). The carbon nanotubes used are supplied under contract by Hyperion Catalysis International, Cambridge USA. They are generated by decomposition of hydrocarbon gases [9]. The supplied powder consists, almost exclusively, of balls of loosely aggregated nanotubes that are non-coiled but generally curved. Their outer diameter is about 10 nm

Results and discussion

It is not possible to break up all the entanglements of the supplied catalytically-grown carbon nanotube material by the dispersion process used, although the exposure to ultrasound in ethanol and the subsequent intense stirring process of the resin leads to a dramatic improvement in the dispersion of the nanotubes in the epoxy. These aggregated phases then form a conductive three-dimensional network throughout the whole sample. The viscosity of the resin is not affected so much as to

Conclusions

Catalytically-grown carbon nanotubes were dispersed as conductive fillers in an epoxy matrix, and the resulting electrical properties compared to those obtained using an optimised process for carbon black. Sufficient matrix conductivity for anti-static applications can be achieved at lower filler concentrations using carbon nanotubes in place of carbon black. The use of carbon nanotubes both reduces the percolation threshold to below 0.04 wt.%, and increases the overall conductivity achieved. At

Acknowledgements

Thanks to S. Friend of Hyperion Catalysis for making possible this work on their catalytically-grown material. J. Sandler and Prof. Dr.-Ing. K. Shulte especially thank the Volkswagen Foundation, Germany, for financial support.

References (13)

  • F Carmona et al.

    Solid State Communications

    (1984)
  • M.S.P Shaffer et al.

    Carbon

    (1998)
  • Schüeler R. Entwicklung polymerer Verbundwerkstoffe mit elektrischer Leitfähigkeit, PhD Thesis, Technischen Universität...
  • S Kirkpatrick

    Review of Modern Physics

    (1973)
  • R Zallen

    The physics of amorphous solids

    (1983)
  • R Schüeler et al.

    Journal of Applied Polymer Science

    (1997)
There are more references available in the full text version of this article.

Cited by (1430)

View all citing articles on Scopus
1

On leave from: Polymers and Composites Section, Technical University Hamburg-Harburg, 21071 Hamburg, Germany.

View full text