Unusual thermal diffusion of carbon nanotubes in a thermoplastic polymer V. Tishkova, E. Pavlenko, M. Legros, G. Bonnet, P. Puech and W.S. Bacsa CEMES-CNRS, 29 Jeanne Marvig, Université de Toulouse, wolfgang.bacsa@cemes.fr ABSTRACT The dispersion of agglomerated carbon nanotubes in polymers is a major challenge. We find that when annealing, the agglomerated tubes disperse spontaneously in a highly viscous thermoplastic polymer. We attribute the thermal diffusion induced dispersion to the difference in the dielectric properties of the nanotubes and the polymer matrix. Strong infrared absorption of the nanotubes leads to selective heating of the nanotubes. The formation of a temperature gradient between the nanotubes and polymer matrix enhances diffusion at the tube surface and along the tube axis. As a result annealing disperses the carbon nanotubes efficiently in a viscous thermoplastic polymer. Transmission electron microscopy, Raman spectral maps and insitu optical microscopy studies show the formation of a uniform diffusion layer. Keywords: carbon nanotubes, thermoplastic polymer, TEM, Raman, dispersion 1 INTRODUCTION Using carbon nanotubes (CNT) in a thermoplastic polymer, the electrical conducting properties of CNTs are combined with the solvent resistant properties of the polymer to improve the electromagntic shielding properties of thermoplastics for their application in aeronautics industry. The high aspect ratio and small diameter of CNTs reduces the filling factor making it possible to observe percolation of the nanotube network in the range of 0.1-1.0 w% [1]. We use here as a thermoplastic polymer Poly(ether ether ketone) (PEEK) [2]. The major challenge of using CNTs in composites is in controlling the dispersion in the polymer matrix. Carbon nanotubes are highly polarizable and hence inherently agglomerate. Agglomerated tubes reduce percolaction and do not show the same record breaking properties of individual tubes. The use of surfactants to disperse the tubes is not useful because they increase the contact resistance and reduce the electrical conductivity of the percolating CNT network [6]. Yet, to improve the electrical conductivity, agglomeration of tubes at a smaller degree, is beneficial for the formation of conductive network of CNTs in bundles and complex secondary structures depending on their diameter and length distribution. The presence of other forms of disordered carbons can also influence agglomeration. CNTs have so far been dispersed in thermoplastic polymers by mixing the powder with carbon nanotubes solutions followed by hot pressing or using twin screw extruders [3]. First electron microscopy images and Raman maps show that a substantial fraction of the nanotubes remain agglomerated. To have a better insight of the wetting properties of the CNTs with PEEK we have dispersed nanotubes on the polymer substrate and heated the polymer above the melting temperature. It has been been reported that diffusion of polyethylene is 30% larger along the tube axis than perpendicular to the tube axis in a polymer SWNT composite above the glass transition temperature [4]. The higher diffusion rate along the tubes has been attributed to Brownian motion of the nanotubes in the polymer matrix leading to a larger excluded volume at higher temperature. We have observed that small molecules such as methanol can form uniform shells around CNTs [5]. We expect that larger diffusion of polymer molecules along the tube axis influences CNT agglomeration. 2 EXPERIMENTAL Multi wall carbon nanotubes (MWNT) were used from Arkema (density 50-150kg/m 3 , 5-15 layers, average diameter 15 nm, averagle length 2-10 μm, carbon purity of 90%). We use PEEK as supplied by Victrex, UK (Victrex® PEEK TM 90P high flow unfilled powder). To form a polymer film, PEEK powder was mixed with acetone and several droplets were deposited on a glass slide and backed after evaporation of the solvent at 380°C (PEEK melting temperature 343°C) for 10 minutes in argon atmosphere. Dispersions of carbon nanotubes was prepared by sonication using acetone. The CNTs were deposited on the PEEK film by placing droplets and annealed at 380°C for 10 min in argon atmosphere. The temperature in the oven was monitored seperately by using a thermo couple. To analyze the tube distribution the thin films were placed in an epoxy matrix and cut with Ultratome equipped with a diamond knife, in a direction perpendicular to the film interface. Multispectral Raman maps were obtained on a Xplora (Horiba Inc.) spectrometer using the 785 nm excitation wavelength and using a piezoelectric XY table. Transmission electron microscopy (TEM) has been performed on a Phillips CM30 microscope operating at 150 kV. We used a TEM thermal heating stage to heat up to 450°C on a JEOL 2010 at 160kV to observe in situ the melting, diffusion of the polymer and the formation of the CNT nano composite layer. We also used a cryostat with a heating element to study the in situ annealing in vacuum at 380°C in combination with a optical video microscope. NSTI-Nanotech 2012, www.nsti.org, ISBN 978-1-4665-6274-5 Vol. 1, 2012 298