Methods for Multiscale Metrology and Process Optimization in Industrial Carbon Nanotube Composites

B. Natarajan, N.D. Orloff, R. Ashkar, S. Doshi, A.M. Forster, E. Thostenson, R. Sharma, J.A. Liddle
Center for Nanoscale Science and Technology, The National Institute of Standards and Technology,
United States

Keywords: nanocomposites, composites, advanced manufacturing, nanomanufacturing, carbon nanotubes, characterization


CNT nanocomposites are widely employed as light-weight, corrosion-resistant, readily-processable materials for mechanical reinforcement, electro-magnetic interference shielding, structural health monitoring, and flame retardancy. Most of these applications demand that the CNTs be individually and homogenously dispersed in the polymer matrix, to maximize polymer-filler interfacial area and to form connected networks at low CNT loadings. However, industrially-produced CVD-grown tubes often assume a “disorganized” entangled structure, which requires extensive mechanical processing to break down aggregates, and to disentangle and uniformly distribute CNTs into the polymer matrix. In light of this fact, several candidate methods for increasing CNT dispersion have been investigated. Unfortunately, most of these processing methods are not scalable to volume production. Even with scalable processing, another oft-neglected hindrance to the widespread incorporation of CNT composites is the absence of a high-throughput measurement method for metrology and continuous process optimization, through microstructural evaluation. Such a technique should ideally probe the sample in a non-contact, non-destructive fashion and must measure a bulk property that is sensitive to the matrix and filler properties, filler spatial organization and loading. However, prior to applying this technique in volume manufacturing, a quantitative understanding of the effects of morphological parameters on the bulk property of interest must be developed. To support this, a suite of techniques probing the microstructure at various length scales is required, since industrially produced CNT composites characteristically possess a hierarchical structure. To this end, we demonstrate the use of a novel combination of imaging and scattering techniques for the multiscale evaluation of CNT-epoxy samples with varying CNT loadings and dispersion states, prepared by an industrially-relevant calendering process. The various dispersions are realized by varying the gap size in the calendering mill. We use scanning gallium ion microscopy for the characterization of the CNT dispersion at large length scales (> 100 microns), with resolutions down to 100 nm. We employ transmission electron microscopy and small-angle and ultra-small-angle neutron scattering to study the morphological features in the 10 microns to 0.1 nm range. This microstructural information is analyzed to obtain metrics that enable quantitative evaluation of the differences in CNT spatial arrangement at the relevant measurement length scales. We then illustrate the use of the resonant cavity perturbation (RCP) technique for metrology and process optimization in carbon nanotube composites. RCP is a well-established, experimentally simple, non-contact, non-destructive technique to characterize the AC conductivity of materials at microwave frequencies. The sensitivity of the electrical conductivity to CNT dispersion and loading, and the ability of RCP to measure these properties accurately and in a non-contact fashion, make this technique well-suited for quality control of CNT-polymer composites produced at industrial scale. By correlating dispersion metrics, obtained from the characterization methods, to the bulk properties measurements (as determined by RCP) we develop a quantitative understanding of processing-structure-property relationships and estimate the sensitivity of RCP to dispersion variations at various length scales. We then discuss the utility of the combination of these techniques in identifying optimal processing conditions and how this methodology may be applied in an industrial setting.