Since the discovery of carbon nanotubes (CNTs) more than a decade ago, polymer/CNT nanocomposites (NCs) have been extensively studied, because low levels of well-dispersed CNTs can yield NCs with enhanced mechanical properties, electrical conductivity, thermal conductivity, and flame retardancy. However, the structural nature of CNTs, which are commonly strongly aggregated due to their high specific surface area and heavy entanglements (the latter especially so with commercially available CNTs), has made it very difficult to employ conventional melt processing to achieve excellent dispersion of CNTs in a polymer matrix. Other processes that have been employed in selected polymer systems to achieve CNT dispersion include solvent blending and in-situ polymerization usually with the aid of surface functionalization and/or sonication, but these processes are limited for use in only a subset of polymer systems and are not commercially attractive routes for production of polymer/CNT NCs.

Here, we incorporate originally heavily entangled, multiwall CNTs of different dimensions (diameter and/or length) into polypropylene (PP) by a continuous process called solid-state shear pulverization (SSSP). With SSSP, deformation energy can be highly stored in the system and released by creating a new surface, which allows for effective pulverization of the system into fine powder. The dispersion of CNTs and NC properties are studied by scanning electron microscopy (SEM) as a function of the energy consumed during pulverization. After SSSP processing, SEM reveals an unusual structure in which loose CNT agglomerates are interpenetrated. We have found that these interpenetrated agglomerates can be further dispersed by additional melt mixing. A NC with 1 wt% MWCNT (30-50 nm outer diameter and aspect ratio of ~125) and fabricated by the two-step process of SSSP followed by melt mixing is shown via SEM to have CNTs that are almost fully debundled into individual tubes even under relatively low energy input during SSSP. This well-dispersed NC exhibits a greater than 50% increase in Young's modulus relative to neat PP. This enhancement in Young's modulus agrees well with predictions from the Halpin-Tsai model for a perfectly dispersed 99/1 wt% PP/MWCNT NC with a MWCNT aspect ratio of 125. Additionally, relative to neat PP, the NC exhibits a PP crystallization half-time that is reduced by more than 75 % and a thermal degradation temperature that is elevated by more than 25 ˚C.

In contrast, when we use the same conditions for our two-step SSSP plus melt mixing process on PP and 1 wt% of significantly thinner (< 8 nm outer diameter) and originally much more heavily entangled MWCNTs, we find that our hybrid product contains MWCNT agglomerates with diameters that are less than 500 nm. This nano-agglomerated 99/1 wt% PP/MWCNT NC exhibits properties that are slightly better than those of our other NC. However, because the aspect ratio of the thinner MWCNTs is greater than that of the thicker MWCNTs, the observed Young's modulus in the agglomerated NC is well below the value predicted from the Halpin-Tsai model. These results demonstrate that the ability to disperse CNTs in polymers is a function of their dimensions and the degree of CNT entanglements and that our novel, two-step process method must be accordingly tailored to the energy required to disperse each MWCNT system.

Studies are underway involving other CNT systems, optimization of our two-step SSSP plus melt mixing process approach, assessment of any damage of the PP chains and MWCNTs during our two-step process, and characterization of the electrical conductivity and rheological properties of our resultant NCs.