Structure-Property Relations of Unidirectional Carbon Nanotube Polymer Composites through Quantitative 3-D Nanoscale Characterization

B. Natarajan, N. Lachman, I.Y. Stein, D. Jacobs, B. Wardle, R. Sharma, J.A. Liddle
Center for Nanoscale Science and Technology, The National Institute of Standards and Technology,
United States

Keywords: carbon nanotubes, nanocomposites, electron tomography, structure-property relationships, nanostructure

Summary:

Carbon nanotube (CNT) filled polymer nanocomposites (PNC), which typically contain randomly aligned and dispersed CNTs, are observed to display sub-optimal, isotropic properties. A potentially useful method to enhance the functionality of CNT PNCs is to fabricate materials with uniaxially-aligned, continuous nanofibers in a matrix, which enable the transfer of the highly anisotropic CNT properties to the bulk system. The composite properties are, naturally, dependent on the volume fraction (Vf) of the aligned CNTs in the PNC, and thus increasing the CNT volume fraction without losing their alignment is expected to enhance the axial properties (be it conductivity or stiffness), as well as increase the PNC anisotropy. Indeed, the axial thermal and electrical conductivities, and the elastic modulus values reported from our own work are some of the highest found in the extant literature for CNT-based PNCs, with much smaller effects in the transverse direction. However, even these exciting properties fall short of theoretical predictions based on simple models, sometimes by over an order of magnitude. It is therefore clear that such simple models, which assume perfectly-aligned, uniformly-distributed CNTs in a polymer matrix fail to predict the real performance of such PNCs. Adjustments that reflect the real nature of the CNT morphology within the PNC have to be made in order to account for processing effects and artifacts such as bundling, phase separation, and waviness, as well as their evolution with varying Vf. While the need for such adjustments has been highlighted earlier, the ability to ascertain, quantitatively, the true nature of the CNT morphology in composites had so far been lacking, due to the limitations of 2D imaging techniques and scattering methods in resolving the complex 3D structure of the embedded CNTs. In this work we present a novel imaging methodology, using electron tomography, that aids in the extraction and quantification of the 3D arrangement of CNTs in these composites. High-quality, electron-transparent samples were prepared from these materials using a dual-beam focused ion beam microscope. The samples were then imaged at multiple tilt angles using zero-loss energy-filtering, which created greater contrast between nanotubes and their surrounding matrix. An automated phase-identification tool was then implemented that performed unbiased segmentation at approximately 50 times the speed of conventional segmentation methods to obtain 3D reconstructions of the nanotube morphology. Morphological metrics describing the network structure, alignment, proximity and waviness may be readily measured from these digitized reconstructions with nanometer resolution. We employ such metrics obtained from four different volume fractions of aligned multi-walled CNT composites (0.44 %, 2.6 %, 4 %, and 6.9 %) in combination with simulation studies to revisit and explain the properties measured from the very same PNCs reported in our highly cited work earlier. In a broader sense, we believe that the implications of the correlations made here will provide valuable insight into processing-structure-property relationships in similar advanced composite materials.