- Top of page
- CONCLUDING REMARKS
- REFERENCES AND NOTES
Polymer nanocomposites are distinguished by the convergence of length scales corresponding to the radius of gyration of the polymer chains, a dimension of the nanoparticle and the mean distance between the nanoparticles. The consequences of this convergence on the physics of the polymer chains are considered, and some of the outstanding issues and their potential consequences on structure—property relations for polymer nanocomposites are highlighted. © 2007 Wiley Periodicals, Inc. J Polym Sci Part B: Polym Phys 45: 3252–3256, 2007
- Top of page
- CONCLUDING REMARKS
- REFERENCES AND NOTES
Significant scientific and technological interest has focused on polymer nanocomposites (PNCs) over the last two decades.1–3 Most of the early and some of the recent efforts focused on the understanding of the underlying synthetic and physical chemistry4, 5 along with aspects of colloidal particle physics6 and mechanical, thermal and physical properties7 of such nanocomposites. Only recently more attention has turned towards understanding and exploiting the unique physics of the polymers in such materials. This has been motivated by a growing recognition that to move beyond formulating polymers with nanoparticle fillers, and towards truly engineered, designed, and functional nanocomposites, and thus broader commercial utilization, the current incomplete understanding of the fundamental aspects of the structure-property-processing relationship of these materials must be addressed.
Investigation of PNCs was heralded by the development of nylon-6 based clay nanocomposites by the Toyota research group where an extraordinary set of mechanical, thermal, and physical properties were achieved by exfoliating individual clay layers with a significant fraction of the polymer chains ionically attached by the use of a clay-bound initiator for the ring-opening polymerization of ε-caprolactam.8 While the early focus remained on the exfoliation of the clay layers and its consequences on property enhancement, recently it has become evident that the nature of the polymer–clay interactions and the high density of surface grafting transforms the crystalline polymorph of nylon-6, and these changes in the polymorph can be reasonably correlated with the changes in the mechanical and thermal properties of those nanocomposites.9 With the exponentially growing availability of competing nanofillers (for example carbon-based (single wall carbon nanotubes, multiwall carbon nanotubes, carbon black, nanographite, nanodiamond), oxides (clay, ziroconium phosphate, halloysite, silica, titania, BaTiO3, Fe2O3, etc.), and metals and intermetallics (gold, silver, platinum, palladium, CoPt, etc.)), almost any property suite can now be imagined. Nevertheless, such successes on par with nylon-clay have remained largely rare and attributable, perhaps, to an incomplete understanding of the changes in polymer behavior when subjected to confinement, and when in the presence of the large specific area solid surfaces that allow for specific interactions.
A growing number of reports are emerging that are beginning to provide multifaceted views of this complex problem. Crucial to these studies is the wide array of synthetic chemistry tools for nanoparticle preparation along with polymerization methods such as grafting-to and grafting-from that have been developed and practiced over the last two decades. The synthetic issues,4, 5 colloidal particle physics and processing10, 11 as applied to PNCs, while quite mature, have many outstanding issues and these have been recently articulated in substantial detail.1, 3, 6, 10, 12–14 This article will concentrate on the polymer physics aspects of this complex subject.
PNCs are distinguished by the convergence of the dominant length scales associated with the polymer matrix (polymer radius of gyration, Rg), the nanoparticle (diameter of nanospheres and nanotubes or thickness of nanoplates, 2r) and the composite morphology (mean distance between constituents, 2ξ). For instance, for 10 nm diameter spherical nanoparticles dispersed in a polymer matrix, at a loading of 10 vol %, results in an average minimum distance between nanoparticles of ∼ 8 nm. When these dimensions are of a comparable magnitude (Rg ∼ r ∼ ξ), geometrical considerations have profound impact on the conventional underlying constructs of polymer behavior. Geometrically, this relation implies a predominance of nanoparticle–polymer interface area,3, 14 where every polymer chain is within 5Rg from an interface, and in many instances (sphere and rod), the interface exhibits curvature that is comparable to Rg.3 Using colloidal blend terminology, this is the rarely explored regime of r/Rg ≤ 1. Corollary to these geometrical considerations is that the ratio of the number density of particles and chains approaches unity. Also, the local density oscillations resulting from the presence of the impenetrable particle overlap with those from neighboring particles, and thus inhibit a complete decay to bulk-equilibrium characteristics. Taken together, the physics of the polymer constituent within a PNC will be far from conventional equilibrium bulk polymer behavior. The key question is how far from conventional constructs does the convergence of these dominant length scales take us, and to what extent does this deviation result in substantial changes to physical properties, especially with regard to other multiphase polymer systems (blends, semicrystalline polymers, filled resins), and not just in comparison to the bulk homogenous polymer.
For example, consider what the relation Rg ∼ r ∼ ξ implies to polymer behavior in the athermal situation. First, the end-to-end distance of a chain, due to the confinement introduced by the nanoparticles, will not necessarily reflect the ensemble average in the bulk. Although the mean end-to-end distance may only exhibit a small change, the constraints imply that the distribution will no longer be isotropic and Gaussian in all directions. Following this, the distribution of topological entanglements will also be modified, and will not be spatially uniform, but exhibit a dependence on the distance from the nanoparticle surface as well as on the nanoparticle shape and curvature. Associated with this, there will be a substantial gradient in dynamics resulting in a dynamic heterogeneity which will be spatially correlated.15 This will have implications on the underlying construct of a glass transition.16 With regard to the nanoparticle, the spatial distribution of the particle center of mass will depend on volume fraction, particle size distribution, and particle shape (sphere, rod, plate). For extended spheroids (rods and plates), excluded volume interactions between particles will lead to local orientation correlations, as well as the possibility of fractal association leading to percolation behavior and interpenetrating polymer–nanoparticle rich regions. Finally, all of these factors will have substantial impact on phase transitions and related morphologies, including polymer crystallization, polymer blend phase separation, and block copolymer mesophase organization. In reality, experimental systems are not athermal, and ascertaining the relative contributions of enthalpic versus entropic factors is a critical challenge to future investigations beyond the aforementioned suppositions. Probably even more of a challenge is the determination of techniques necessary to establish a trusted experimental database that reveals the details of the distributions of static and dynamic characteristics of both polymer and nanoparticle, critical for the valid discussion of the physics concepts.
Outstanding issues that might significantly impact the properties and processing of PNCs include changes in glass transition temperatures, as well as crystal polymorph and habit because of the extensive surface-polymer interactions.17, 18 It can be argued that in many thermodynamically stable cases due to the requirement of favorable interactions between polymer and nanoparticle, the Tg of the polymer in such nanocomposites is likely to be shifted upwards. However, because of the possibility of creating dispersed hybrids due to processing conditions even in the absence of favorable interactions, there exists the possibility of observing significant reduction in the Tg of the polymer.17, 19 For semicrystalline polymers, there have been many reports of changes in the dynamics and structure of polymer crystals.9, 20, 21 Detailed understanding of the basis and ways to control heterogeneous nucleation induced by the nanoparticles though remains largely untackled.21, 22 Changes in fundamental mechanisms of vitrification and formation of crystalline order are likely to have an enormous impact on the solid-state mechanical properties as well as the capability of creating textured materials with anisotropic mechanical properties. Further, these issues along with details of the interfacial interactions between the polymers and the nanoparticles also dictate phonon, electron, and photon transport in such materials and also remain relatively unexplored.
The ability to disperse the nanoparticles and to process and form final product from the PNC is affected by local and chain dynamics. The confinement experienced by the polymer chains between neighboring nanoparticles (Rg ∼ ξ) and the changes in the topological constraints imposed on the polymers due to the solid surfaces,23 must result in fundamental changes in the transport of polymer chains.24 Unique and highly unexpected processing advantages such as a possible reduction in the viscosity of the polymer by the inclusion of nanoparticles may be possible.19 Additionally, using weakly interacting rod-like nanoparticles in an elastomeric or thermosetting matrix leads to the possibility of forming interpenetrating polymeric network with highly disparate mechanical properties that are hitherto unanticipated based on traditional polymeric interpenetrating networks.25
While significant challenges still remain in predicting the dispersion of nanoparticles in single component polymeric matrices,26 the ability to disperse nanoparticles in multicomponent polymeric matrices, especially to direct them to specific phases or to the interface between phase-separated structures, promises even greater opportunities. Experimental and theoretical efforts are only just beginning. The pioneering work of Kramer has clearly demonstrated the ability to sequester spherical nanoparticles in particular phases/interface of a block copolymer.27 Extending this to systems with nonspherical nanoparticles and to complex architectures such as those present in polymer blends and polymeric microemulsions would certainly impact the utilization of PNCs significantly.12
Because of the comparable length scales of the nanoparticles and attached polymer chains, some of these nanoparticles can form exquisitely tailored crystalline and liquid-crystalline mesophases with unique topologies analogous to ordered block copolymers and small molecule surfactant systems.4, 12, 28 Such materials, if prepared with monodisperse nanoparticles and polymers, would suffer much less thermally induced fluctuations and exhibit long-range defect-free structures that are required for many technological applications especially as thin-films. Additionally, attaching block copolymers and other structured polymers onto the surfaces of nanoparticles can alter significantly the natural topologies adopted by those materials and therefore lead to interesting characteristics.