Dispersion and Organization with Carbon Nanotubes

The introduction of a small concentration of nanoparticles can substantially improve the thermomechanical, optical, and electrical properties of polymers, but the specific particle-level mechanism responsible for the improvement is generally unclear.  In several cases, there has been significant investment in systems based on flawed assumptions regarding the scalability and generality of a specific observation.  The resulting performance of the “nano”-composites is generally inconsistent and well below expectations, which pose significant barriers to commercial applications.  These early problems reflect inherent problems in the field related to the lack of available techniques to process and characterize materials on the nanoscale.  Most processing techniques used today are top-down, mechanical methods developed for micron-scale fiber composites and particulate filled systems.  Common characterization techniques are similarly limited by compromises between time, resolution, and sampling size, and may not be sensitive to the relevant processes for a desired property of interest. 

Single-walled carbon nanotubes very exciting materials that are very challenging to use because they really like to group together rather than spreading out.  This is because interactions between particles also increase with surface area, which poses a big challenge (this is also part of the reason why electrospun nanofibers can stay together without a binder when used as a dry adhesive!).  The challenge is how to separate the tubes without changing their properties too much.  The first approach is to basically destroy the tubes through an oxidation process.  This is akin to polishing a stained-glass mirror with a hammer, but it does kind of work.  It reduces big bundles of tubes down to smaller bundles.  But it introduces lots of defects (which are the points where the tubes are disconnected and non-aggregating, so critical to the whole concept), which hurts the electrical and mechanical properties.  To improve on this, you can introduce bigger functional groups that push the tubes farther apart without having to damage them too extensively.  Again, this is OK, but we are still hurting the very objects we are trying to use, and that seems immoral.

Illustration showing nanoplatelet-assisted dispersion of carbon nanotubes. Carbon nanotubes (CNTs) are initially received in aggregated or bundled state. Zirconium phosphate (ZrP), a synthetic clay-like material with plate-like structure, is dispersed with CNTs and is electrostatically tethered to the nanotube surface. With mild ultrasonication, the nanotubes are reduced in size until individual particles are obtained. The ZrP nanoplatelets remain in the system and form a super-structure with the nanotubes [see  "Rheology of electrostatically tethered nanoplatelets and multi-walled carbon nanotubes in epoxy,"  Polymer ,  84 , p. 223-233 (2015) ]. Depending on the application, the nanoplatelets can be retained in the final material for unique improvements in mechanical properties [see  “Influence of trace amount of well-dispersed carbon nanotubes on crystal structure and mechanical properties of injection-molded polypropylene,”  Macromolecules ,  46 (2), p. 463-473 (2013)  and  “Electrical conductivity and thermal stability of polypropylene containing well-dispersed multi-walled carbon nanotubes disentangled with exfoliated nanoplatelets,”  Carbon ,  50 (12), p. 4711-21 (2012) ], or removed to obtain individual nanotubes [for example, see  “Electrical conductivity of well-exfoliated single-walled carbon nanotubes,”  Carbon ,  49 (15), p. 5124-5131 (2011) ].

Illustration showing nanoplatelet-assisted dispersion of carbon nanotubes. Carbon nanotubes (CNTs) are initially received in aggregated or bundled state. Zirconium phosphate (ZrP), a synthetic clay-like material with plate-like structure, is dispersed with CNTs and is electrostatically tethered to the nanotube surface. With mild ultrasonication, the nanotubes are reduced in size until individual particles are obtained. The ZrP nanoplatelets remain in the system and form a super-structure with the nanotubes [see "Rheology of electrostatically tethered nanoplatelets and multi-walled carbon nanotubes in epoxy," Polymer, 84, p. 223-233 (2015)]. Depending on the application, the nanoplatelets can be retained in the final material for unique improvements in mechanical properties [see “Influence of trace amount of well-dispersed carbon nanotubes on crystal structure and mechanical properties of injection-molded polypropylene,” Macromolecules, 46(2), p. 463-473 (2013) and “Electrical conductivity and thermal stability of polypropylene containing well-dispersed multi-walled carbon nanotubes disentangled with exfoliated nanoplatelets,” Carbon, 50(12), p. 4711-21 (2012)], or removed to obtain individual nanotubes [for example, see “Electrical conductivity of well-exfoliated single-walled carbon nanotubes,” Carbon, 49(15), p. 5124-5131 (2011)].

The approach developed in Professor Hung-Jue Sue’s research group at Texas A&M University was to add a second dispersant particle that assists in separating the tubes from the bundles, and may  remain in the system to keep them separated.  In this case, you can then introduce the same functional molecules as before to space out the tubes with a much lower degree of damage and then remove the dispersant.  We did this using a plate-like nanoparticle called zirconium phosphate (ZrP; more on this one in the future!).  This nanoplatelet-assisted dispersion technique allowed the aggregates to be broken down to the individual tube level without introducing significant damage or relying on additional molecules to keep them separated.

In this case, you can a lot of very interesting things.  On a fundamental side, you can see what the effect of an individual nanotube really is!  In most cases, people just tossed the bundled tubes into a plastic and assumed that what they were seeing was the effect of the individual tubes.  They would then make ridiculous extrapolations after what the behavior of the material “might have been” based on these observations, and that did help fund further research based on their ridiculous claims, but screwed over future researchers that weren’t able to achieve their non-sensical boasts and empty promises (I’m not passionate about this, I swear).  What we were able to do was several studies that compared individual tubes with aggregated tubes and we found a number of things that made a lot of sense in hindsight.  If you have individual tubes, you require less of them to see a change in behavior.  This is an effect called “percolation” that is very neat and worthy of a write-up all its own (if you have institutional access, see “Electrical conductivity of well-exfoliated single-walled carbon nanotubes,” Carbon, 49(15), p. 5124-5131 (2011) and “Electrical conductivity and thermal stability of polypropylene containing well-dispersed multi-walled carbon nanotubes disentangled with exfoliated nanoplatelets,” Carbon, 50(12), p. 4711-21 (2012) for brief and semi-coherent introductions).

For non-spherical nanoparticles, most research to date has focused solely on achieving random dispersions on local length scales.  Any resulting properties that do not meet expectation are generally explained based on experiences with micron-scale fillers, in particular, inadequate dispersion and interfacial adhesion.  This level of discussion ignores the most exciting features of nanoparticles, which is that they interact with their environment on a much smaller length scale than their micron-scale counterparts.  This distinction offers tremendous opportunities to design functional materials with properties tailored to specific applications by controlling material structure from the bottom-up.

For simultaneous improvement of multiple properties, more complex morphologies of the dispersed phase may be desirable (for interesting examples, see Torquato, Journal of Applied Physics 2003; 94(9): 5748-5].  If the shape and interaction of nanoparticles are well controlled, the mesoscale organization of nanoparticles may be potentially guided to favor formation of higher-order structures (some good examples will be listed under “nanocrystals” at the end).  These approaches have potential advantages because properties may be modified across multiple length scales.  This provides unique opportunities for simultaneous improvements in properties that are typically in opposition, such as stiffness and ductility.  Higher order structures may be able to interact with their environment on multiple length scales, which is desirable for photonic crystals and other optical systems (see Joannopoulous, Nature 1997; 386(6621): 143-9).  The parameters governing dispersion and organization, and the resulting macroscale properties of the composite material, remain poorly understood from theoretical and empirical viewpoints.