This article aims to provide an introductory overview of the topic of nanocomposites, with a particular emphasis on their use in electrical applicatio... Nanostructured materials - Nanocomposites - Mechanical factors - Gold - Electric breakdown - Blades - Polymers - Terminology - Partial discharges - Control systems - dielectric materials - electric breakdown - nanocomposites - nanoparticles - nanotechnology - partial discharges - nanocomposites - dielectric response - filler particles - nanodielectrics - layered silicates - partial discharge resistance - breakdown strength - engineering materials
Nanodielectrics – How Much Do We Really Understand? Key Words: nanocomposites, nanodielectrics, interfaces, layered silicates, mechanical properties, partial discharge resistance, breakdown strength Nanocomposites - A Historical Perspective
anotechnology represents one of the defining scientific concepts of the first years of the twenty first century. The reason for this stems from the potential that scientists, engineers, venture capitalists, and politicians see in the possibilities afforded by controlling material systems on a scale comparable with molecular dimensions. At present, this appears both vast and uncertain. In the broadest sense, a nanostructured material is one in which structural features exhibit one or more dimensions of the order of 10 nm in size. Nonetheless, despite the current surge in interest in nanostructured materials, in reality, mankind has been exploiting synthetic nanostructured materials for millennia. A remarkable example of nanotechnology in antiquity is the so-called Lycurgus cup that resides in the British Museum in London. This example of fourth century Roman craftsmanship depicts the myth of the eighth century BC Thracian king Lycurgus, entwined in the maenad Ambrosia, after she transformed herself into a vine. The extraordinary feature of this artifact is, however, its dichroic optical properties; it appears an intense ruby-red (see Figure 1) when it is viewed in transmission, but green when examined in reflected light. This effect is induced through the incorporation of nanoparticles of gold, or a gold/silver alloy, within the glass matrix ; the properties of colloidal gold were, rather more recently, studied by Faraday . Some seven hundred years after the fall of the Roman Empire, the Crusaders discovered the value of nanotube reinforcement, when they encountered the Damascus sword blades used by the Saracens. These exhibited remarkable mechanical properties, which are now believed to be a result of a combination of ore composition, the use of specific additives during smelting and the forging/annealing techniques used in their manufacture. When these factors are combined, using a long-forgotten process, the result is the development of cemenite nanowires and carbon nanotubes within the metallic matrix of the blade . A more recent polymer-based example of a nanocomposite is the layered-silicate elastomer prepared by cation exchange that was patented by the National Lead Company, New York, in 1947 . In this article, we aim to provide an introductory overview of the topic of nanocomposites, with a 6
Christopher Green and Alun Vaughan University of Southampton, School of Electronics and Computer Science, Southampton SO17 1BJ, UK
We have sought to introduce some of the salient features of nanocomposite research that relate to the development and application of useful dielectrics.
particular emphasis on their use in electrical applications in which their dielectric response is likely to be of importance. A comprehensive review of this complete topic would not be practical here; more detail can be found in review articles such as -. Composites are widely used engineering materials because of the enhancement in properties that can result from combining a number of distinct components. For example, adding aluminum hydroxide (ATH) filler to a plastic will reduce its flammability and modify its mechanical properties. However, this is often at the expense of some other characteristic; introducing micron-sized inorganic particles into a polymer rarely improves the electrical breakdown strength of the system. However, as this article will show, reducing the size of the filler particles into the submicron range can overcome such problems and influence combinations of properties in a complex and often non-intuitive manner. It is this that has resulted in the rapid expansion of nanocomposites research in recent years. The topic of nanocomposites is one that is intrinsically multidisciplinary in nature and, therefore, it employs terminology derived from many different areas of science. Consequently,
before embarking on any discussion of these systems, it is worth introducing some of the terminology and, in particular, defining terms in the way in which we intend them to be understood. Nanostructured material: any material in which structural features exhibit one or more dimensions that can be considered nanometric in size. The precise definition of the term nanometric is open to debate, but a value of the order of 10 nm is not uncommon  and will therefore be used here. Nanocomposite: a material consisting of a small proportion, typically of the order of 5% by mass, of nanofiller within a host material. Nanofiller/nanometric filler: a filler which has at least one dimension of the order of 10 nm in size. As such, a nanofiller may be nanometric in one, two or three dimensions. Nanodielectric/nanometric dielectric: a nanocomposite of specific interest in connection with its dielectric characteristics. Layered silicate: an inorganic material consisting of strong, covalently bound sheets separated by galleries, such that the sheets are held together only by weak Van der Waals forces.
July/August 2008 — Vol. 24, No.4
Tactoid: a micrometric group of undispersed silicate layers. Compatibilizer: a chemical compound introduced into a composite to modify the interactions between the filler and the surrounding matrix. This may be to alter the surface properties of the filler particles or to encourage the intercalation and exfoliation of layered silicates. Intercalated: a layered silicate in which matrix molecules have been introduced into the galleries. Exfoliated: a layered silicate which has been dispersed into aggregates of a few layers or less. In a fully exfoliated system, the silicate would exist as isolated layers. The above concepts are represented schematically in Figure 2, where Figure 2(a) seeks to show how the chemistry at a surface can be modified through reactions with a chosen compound and Figure 2(b) represents the introduction of a compatibilizer, subsequent intercalation and complete exfoliation in a layered silicate. This explosion of commercial and academic interest in nanocomposites is, primarily, concerned with materials consisting of relatively small quantities of nanoscopic filler within a host matrix; Figure 3 contains data extracted from the ISI Web of Science showing the exponential increase in publications referring to nanocomposites that has taken place during the present decade. For such materials, a number of dimensionally distinct systems can be envisaged. First, one could introduce spherical particles with a diameter of the order 10 nm into a host matrix, in which case the nanofiller could be thought of as exhibiting nanoscopic character in three dimensions. Alternatively, in nanowires or nanotubes, only two dimensions are nanoscopic, while systems containing thin platelets are nanoscopic in just one. Such materials can exhibit interesting properties, as discussed below, and are relatively inexpensive to produce, in sharp contrast to the high cost density of, for example, band-gap engineered materials for the optoelectronic industry and nanoporous catalysts for chemical engineering. Indeed, there are already hundreds of examples of nanoparticle-containing substances being commercially exploited in, for example, the medical, cosmetic and hygiene industries.
Nanodielectrics Although pre-dated by some research into the dielectric response of nanocomposites , , most contemporary authors ascribe the origin of the term nanodielectrics, or more correctly “nanometric dielectrics”, to the seminal paper of Lewis in 1994 , in which he considers the consequences of reducing the phase size in a two-phase dielectric system. To illustrate this, consider two composite systems, each containing 10% by volume of spherical inclusions of material I within a matrix M. If, as might typify a conventional micrometric composite, the inclusions have a diameter of 10 µm, then the interfacial area will be ~6 x 104 m2 per cubic meter of composite. If the diameter of the spheres is reduced to 10 nm, then the interfacial area increases by 3 orders of magnitude. An alternative means of envisaging this is to think, not of the interfacial area, but rather, to consider the composite in terms of the typical separation between the surface of one inclusion and its nearest neighbor. In the micrometric case, this equates to about 10 µm whereas, in the nanometric equivalent, it is just 10 nm.
Which of the above calculations is the more enlightening depends upon the perspective one takes. For example, in his 2005 paper , Lewis divides interfaces into two types, passive and active, where active interfaces well illustrate how behavior of a composite dielectric can be strongly influenced by the magnitude of the interfacial area through polarization, hopping conduction from one interface to its neighbor, etc. However, as pointed out
by Lewis , other authors have sought to define an interface without recourse to any particular characteristic. In this approach, an interfacial region is defined as that portion of the system where the characteristic of interest, which may relate to the material’s structure or property, changes from being characteristic of the inclusion, I, to being characteristic of the unfilled matrix, M. This therefore leads to the idea of a fuzzy interphase region, which is not well represented by either bulk material I or bulk material M. In the literature, this interphase is often though to extend for about 10 nm. In this case, the separation between neighboring particles becomes critical; in the extreme case of interparticle separations never exceeding 10 nm or so, the complete bulk matrix is perturbed by the included nanofiller.
Preparation and Characterization A. Preparation
Figure 2. Schematic representations of (a) the modification of the chemistry of a surface through chemical reaction with a molecule that is compatible with the host matrix and (b) the various stages by which layered silicates are chemically modified and dispersed within a host.
The dielectrics community is interested principally with insulating fillers, such as layered silicates (e.g. montmorillonite clay) and ceramic particles (e.g. TiO2 and SiO2), rather than conducting fillers such as graphite platelets and carbon nanotubes (CNT). However, the latter have received much interest for use in mechanical applications and may be of use in connection with electromagnetic shielding  or applications in which a degree of controlled electrical conductivity is required. Nanoparticles can be prepared by mechanical attrition, although the generation of crystal defects and impurities from the milling media is problematic. Furthermore, it is very difficult to produce narrow particle size distributions and, therefore, wet chemical methods are generally preferred. Ceramic nanoparticles are, for example, best produced by the sol-gel process . Before compounding a nano-filler into a host matrix, it is important to ensure good thermodynamic and chemical compatibility between the filler and the matrix. For example, vinyltriethoxysilane can be used to form covalent bonds between silica and polyethylene . Maleic anhydride will hydrogen bond with oxygen groups on silicate layers and will therefore enhance dispersion in polyethylene and polypropylene based nanocomposites if it is first grafted onto the nonpolar polyolefin chains (<1% by mass) . Each layer of montmorillonite (MMT), consisting of two octahedral sheets sandwiched between two tetrahedral sheets , is negatively charged. The cations (e.g. Na+) that exist in the gap between adjacent clay layers in order to balance these negative charges can be exchanged with a range of surfactant-like ions, which render the inorganic clay more compatible with the host matrix. Replacing small ions such as Na+ with large structures such as alkylammonium ions also serves to increase the gallery spacing, such that polymer chains can more easily intercalate between the silicate layers. Generally, compounding of nanocomposites is economically feasible only by in-situ polymerization or mechanical blending methods. The latter are inherently simple, although a considerable amount of process optimization is necessary if reagglomeration of particles is to be prevented. In the case of layered silicates, it is considerably more difficult to achieve fully exfoliated structures than intercalated ones. Dennis et al.  has conducted a detailed study of extrusion conditions during the compounding of
IEEE Electrical Insulation Magazine
Figure 3. Graph showing the fraction of the total number of papers published between 1970 and 2006 on the subject of nanocomposites for each year between 2000 and 2006. Data relating to all matrix types are represented in blue, while papers concerning polymer-based nanocomposites are represented in red. Some 27% of all papers concerned with polymeric nanocomposites were published in 2006. nylon-6/MMT nanocomposites, finding that too much backmixing or intensive shearing reduces the likelihood of exfoliation. In addition, mechanical mixing is not universally suitable; in the case of epoxy/clay systems, it is difficult to generate high enough shear forces to delaminate the platelets . Another disadvantage is that at typical melt extrusion temperatures, the alkylammonium compatibilizers included to give compatibility with the host polymer are thermally unstable . Some promising studies exist concerning the use of polar solvents to expand the intergallery gap instead of organic compatibilizers. Sun et al.  have studied the preparation of epoxy-clay nanocomposites in this manner, using acetone to carry epoxy monomers into the galleries prior to polymerization.
B. Characterization The properties of any material will depend upon both its composition and its structure. Therefore, having prepared a nanocomposite, it is necessary to determine how the structure that has evolved will subsequently influence the properties of the
July/August 2008 — Vol. 24, No.4
material, potentially, over decades. While it is convenient to use the term structure in this way, in the case of a nanocomposite, what exactly do we mean by “structure” and which dimensional levels are of greatest importance (see Figure 4)? It is currently thought that many of the characteristics of nanodielectrics are determined by the interactions that occur at nanoparticle-matrix interfaces, which are represented by the regions immediately adjacent to the nanoparticles in Figure 4. Since these will be related to molecular factors of various sorts, many examples exist - in which the standard techniques of electromagnetic spectroscopy (Raman, infra-red, and UV-Vis techniques) have been used to probe the local interactions that occur in these regions. Whilst many effects have been observed, these are generally rather subtle in nature and their utility is limited to specific problems. At intermediate scales, our use of the term interphase refers to a region where material properties are not well represented by either bulk component; how far do interphase regions extend into the matrix and what is their nature (matrix effects in Figure 4)? When it comes to electrical breakdown, for example, aggregation
is a well-appreciated problem and, here, real difficulties exist in determining the level of nanoparticle dispersion and the extent to which agglomeration exists (aggregation in Figure 4). Provided an appropriate contrast mechanism exists, current generation scanning electron (SEM) and transmission electron microscopes (TEM) can both be used to examine specimens at high magnification, but a large number of images are then needed to build up a statistically representative morphological picture. Even then, much care needs to be exercised to minimize bias and to discern genuine features from artifacts. Scattering techniques provide an alternative means of producing a measure of structure that is averaged over a significant sample volume. Light scattering is an inexpensive way of analyzing the structure of transparent nanoparticle-containing materials , though sophisticated models, which may not be reliable, are needed to extract quantitative data. Wide-angle X-ray diffraction is used extensively in the literature to characterize the inter-gallery spacing in layered silicates . As intercalation proceeds and the gap widens, the diffraction peak that characterizes the inter-layer periodicity shifts to smaller angles before finally disappearing. However, it goes without saying that the disappearance of this peak does not necessarily imply exfoliation; it could equally indicate that the clay layers are still aggregated, but in a non-periodic arrangement. Nevertheless, at least one attempt has been made to analyze X-ray scattering data quantitatively in order to obtain a size distribution that reflects the state of dispersion of the clay in the polymer . Unfortunately, this process is complex and
its utility appears limited; comparison with TEM images reveals the presence of a few large tactoids, which are not represented in the X-ray-derived distribution function. It would seem therefore that much work remains to be done to develop reliable methods for the characterization of nanocomposites.
Properties of Nanocomposites Insulation is used in a wide variety of different electrical components, ranging from field effect transistors to high voltage transformers. Within this diverse range of plant, the characteristics required by each insulating element may include many different properties, such as mechanical strength and high thermal conductivity, not just low dielectric loss, high breakdown strength, etc. Nanocomposites can exhibit many potentially interesting properties and, therefore, in the discussion that follows, we have not limited ourselves to purely electrical issues.
A. Mechanical Properties A key point in the current explosion of interest in nanocomposites research came in the early 1990s, when researchers at Toyota prepared nylon-6/clay nanocomposites by in-situ polymerization. They reported that these materials exhibit a 50% increase in strength and a doubling in elastic modulus over the base resin, together with an increase of 80 °C in heat distortion temperature . A decade on, polymer/layered-silicate (PLS) nanocomposites can now be found in diverse automotive applications, ranging from timing-belt covers and fuel-system components to
Figure 4. Schematic diagram indicating the range of different dimensional levels that need to be considered when attempting to characterize a nanocomposite. At the smallest dimensions, local interactions, molecular conformations and the properties of the matrix material need to be considered within the interphase regions. At intermediate scales, how far do interphase regions extend into the matrix and what is their nature? Finally, how well dispersed is the nanofiller within the matrix? Do the interactions that occur between the matrix and the nanofiller tend to promote dispersion, or do they lead to demixing and aggregation of the nanofiller? 10
IEEE Electrical Insulation Magazine
doors and seats . The extensive article by Tjong , summarized briefly here, reviews the enormous amount of research into the mechanical behavior of nanocomposites that has been undertaken in the last decade. In general, silicate platelets have a strengthening and stiffening effect on the base polymer, but they tend to adversely affect the toughness and tensile ductility of semicrystalline polymers. On the other hand, the toughness of elastomers and epoxy resins is generally improved, due to microcrack deflection, whereas the toughness of glassy polymer nanocomposites can be somewhat compromised because of the difficulty of obtaining adequate clay dispersion in these systems. The inclusion of well-dispersed carbon nanotubes can result in even greater modulus enhancements without compromising toughness in both semicrystalline and glassy thermoplastics. Greater toughness can also be obtained with ceramic nanoparticles, due to cavitation and shear yielding effects. In an attempt to understand these phenomena, micromechanical models have been extrapolated to the nanometric level, even taking into account effects such as incomplete exfoliation in clay/ polymer nanocomposites . Silicates and CNTs have aspect ratios of ~200 and ~1000 respectively, and so their intrinsic capability for matrix reinforcement is enormous. It must not be forgotten, however, that the filler-matrix interaction zone is capable of behaving in a completely different manner than that of the bulk polymer. For example, Shah et al.  observed that the addition of an organically modified clay to poly(vinylidene fluoride) (PVF2) resulted in a change in the dominant crystal form from the usual α to β and the production of material with enhanced stiffness and toughness. The question of how nanofillers affect the glass transition behavior (Tg) of glassy polymers is a particularly interesting one, since this parameter is intimately linked to molecular conformations, molecular mobility, and free volume, and therefore could provide direct information concerning the nature of nanofiller/matrix interfaces. For example, if the matrix adheres well to the filler, as reported by Sarwar et al. , then Tg can be increased as a result of local molecular confinement. If adhesion is poor, then the resultant lack of filler/matrix interactions may, in the limiting case, result in the polymer molecules behaving as if they are in ultra-thin films, with a consequently reduced Tg, as discussed by Ash et al. .
ing the kind of temperatures likely to be experienced by material surfaces during dry-band arcing, confirmed the impressive erosion resistance of the nanocomposites. The reason for this was shown to be the formation of a silica-rich protective barrier on the surface, whereupon the material behaves in a manner that is comparable to nanocomposite systems that exhibit superior flame resistance . Kozako et al.  drew attention to the crystallization of polyamide spherulites on layered synthetic mica, resulting in partial discharge-resistant “stair-like” structures forming out of less resistant amorphous regions. Sarathi et al.  remark that well-exfoliated clay structures can act as oxygen barriers to inhibit bulk degradation, a phenomenon that is being developed for exploitation in the food packaging industry. It is gas permeation through conventional PET bottles that makes this type of packaging unsuitable for use with premium ranges of beer. A study of complex permittivity as a function of frequency is fundamental to the characterization of dielectrics. For example, a promising development for high energy density storage materials comes in the form of metal nanoparticle-based composites, in which the confinement of electron wavefunctions are reported to yield massive permittivities, as high as 1010 . So-called high κ dielectrics are attracting great interest at present for use in applications such as flexible, organic electronics, where it is necessary to optimize both the breakdown strength of the gate dielectric and the electric field distribution in the device. In the field of more conventional electrical insulation, it seems possible to generate nanocomposites exhibiting a permittivity that is lower than that of either of the constituent components. It has been suggested that this counterintuitive result is due to polymer chain immobilization at the interfaces. To illustrate the value of dielectric spectroscopy as a probe for nanocomposite dielectric behavior, consider the work of Roy et al. , who compared the dielectric properties of micro- and nano-filled silica/XLPE composites, in particular studying the effect of compatibilizer chemistry. Figures 6(a) and 6(b) illustrate the real permittivity and tan-delta behavior of the materials
B. Electrical Properties Following the pioneering lead of Lewis in 1994 , the idea of nanocomposites finding useful application in electrical insulation applications has become increasingly appealing. Not only do nanocomposites seem to have enhanced mechanical and thermal properties over their microfilled cousins, but an increasing body of research indicates that improvements in electrical properties can also be achieved with only a few weight percent of filler. Particularly promising are materials that offer greater resistance to electrical erosion. El-Hag et al.  compared the electrical erosion resistance of silicone rubber filled with 12 nm fumed silica particles with the same matrix filled with 5 µm silica. Using the ASTM D2303 liquid contaminant inclined-plane test, they found that a 5% by mass loading of nanofiller resulted in less eroded mass than did 30% by mass of microfiller, as illustrated in Figure 5. Laser ablation, which is a controllable way of locally reproduc-
July/August 2008 — Vol. 24, No.4
Figure 5. Data showing the effect of filler loading level and particle size on the electrical erosion resistance of various micro- and nanocomposites using the ATSM D2303 liquid contaminant inclined-plane test. Data derived from . All stated percentages are by mass.
containing only unmodified silica, for simplicity. The permittivity of the microcomposite is unusually high, apparently due to interfacial polarization; in the nanocomposite, there is not only a reduction in the permittivity compared with the virgin XLPE, but the low frequency loss peak near 1 Hz also appears to be absent. These authors suggest that this is also a consequence of chain immobilization. A “quasi-DC” feature is indicated by the large negative slopes in both the real permittivity and tan delta for the nanocomposite at low frequencies. This is explained in terms of charge carriers that are localized to individual particles, yet free to move on their surfaces; the equivalent circuit would be a network of large capacitors whose dynamics are limited by their internal resistances. Though this could be advantageous for space charge mitigation, the potential for percolative conduction due to particle agglomeration cannot be ignored. The workers found that certain polar compatibilizers resulted in the suppression of this feature; further experiments verified that these samples exhibited deep trapping and low mobility. However, these compatibilizers also resulted in a relaxation in the power frequency range, where the permittivity increased to around the level observed in the microcomposite. Indeed, some researchers have observed increased dielectric loss over a wide range of frequencies in their materials, which may be related to the compatibilizer chemistry . Another characteristic often ascribed to nanodielectrics is improved electrical breakdown strength relative to the base resin. Breakdown by electrical treeing was studied in an epoxy composite filled with a mixture of micro- and nano-silica  and improved endurance times were observed compared with a simple microcomposite. This feature was related to the denser, more tightly packed tree structures that evolved in the presence of the nanofiller. Returning to the work of Roy et al.  described above, these workers reported increases in the DC ramp breakdown strengths and voltage endurance times of their nanocomposites. Depending on the compatibilizer chemistry, increases in endurance time of over two orders of magnitude could be obtained compared with the unfilled resin. They attribute this to a combination of electron scattering and, in the case of the polar compatibilizers, deep trapping. In general, the physics of electrical breakdown depends critically on sample preparation routes, geometry, and test method. The critical instability may result from relationships between many, sometimes very subtle, factors. In the above work, for example, it is suggested that a scattering mechanism may be provided through enhanced nucleation in the case of the nonpolar compatibilizer. A recent paper  has discussed the improved breakdown strengths of magnesium oxide/LDPE nanocomposites in terms of the suppression of packet charge. Conversely, another paper reports that internal fields due to space charge build up in nano-alumina/LLDPE materials regularly initiated sample breakdown during pulsed electroacoustic testing . We cannot have a complete picture of the electrical breakdown performance of nanocomposites without a thorough understanding of their charge transport behavior. In summary, therefore, we should conclude that the dielectric properties of nanocomposites show considerable promise, but that much remains to be understood.
Nanodielectrics Research at the University of Southampton The account given thus far presents a picture of a blend of technological promise and scientific uncertainty. While it might be argued that “if it works, why worry too much about why?,” our view is that this is unacceptable for two reasons. First, the factors that determine the behavior of nanocomposites are manifold and this therefore opens up great potential for creating optimized “designer” materials. If we do not understand why a material behaves in the way that it does, how can we optimize its properties without extensive empirical testing? Second, if we do not understand why a material behaves in the way that it does at this moment in time, how can we gauge the way that it will behave in the future? Clearly, this is a particularly important consideration when designing plant for a 40+ year service lifetime. Consequently, much of
Figure 6. Dielectric data obtained from various crosslinked polyethylene (XLPE)-based systems, demonstrating that the scale of the filler (micro- or nano-) can have a marked effect on (a) the real permittivity of the material and (b) the dielectric loss (tan δ). Data derived from . All stated percentages are by mass.
IEEE Electrical Insulation Magazine
Figure 7. Scanning electron micrographs showing the way in which a nanofiller can perturb the development of structure in a semicrystalline polymer. Both images show the same blend of high density (10% by mass) and low density (90% by mass) polyethylene and both samples were crystallized at 117 oC: (a) no additives (scale bar 20 µm); (b) polymer plus 5% by mass of montmorillonite (scale bar 10 µm).
the current research at Southampton, as elsewhere, is targeted at understanding why nanocomposites in general and nanodielectrics in particular behave in the way that they do. One topic has been stressed more than any other in the preceding discussion is that of interfaces. So, what is an interface? Is it abrupt? For what distance away from the interface is the matrix in some way perturbed by the presence of the nanofiller? What is this interaction zone? To illustrate the complex nature of such questions, the following account describes the influence of the layered silicate MMT on structural evolution in polyethylene. Figure 7 compares the structure of polyethylene [Figure 7(a)] with that of an MMT filled nanocomposite [Figure 7(b)]. Figure
July/August 2008 — Vol. 24, No.4
7(a) reveals an array of spherulites, ~40 µm in diameter, that are composed of radiating lamellar crystals which, in the SEM, appear predominantly as bright lines. The orientation of these crystals periodically changes along the spherulite radius, giving the characteristic concentric banded texture that is arrowed in Figure 7(a). This structure is highly organized and is very typical of polyethylene. Conversely, Figure 7(b) reveals a structure that is extremely difficult to interpret. The arrowed objects are MMT tactoids and, throughout the image, lamellar crystals can be seen which, as in Figure 7(a), generally appear as bright linear features. However, there is now little evidence of these crystals being arranged into the lamellar aggregates that typify polyethylene. In short, the addition of the MMT has completely disrupted the standard polyethylene crystallization process. The polymer molecules are interacting strongly with the nanoclay surface, such that the presence of the MMT results in massive crystal nucleation. In other words, there is a strong local interaction zone. However, interactions are not limited to molecular dimensions since, while the clay platelets promote nucleation, they also appear to inhibit subsequent crystal growth. Indeed, there is evidence in the literature of the presence of MMT greatly impeding the dynamics by which molecules rearrange themselves from a disordered melt into an ordered crystal. Effectively, the polymer molecules become “stuck” to the nanofiller. Unfortunately, this description is not universal, since, even when adding MMT to the same polymer, changing the surface chemistry of the MMT appears to be able to change the processes entirely. Is this a consequence of highly subtle chemical interactions between the polymer matrix and the nanofiller? Alternatively, is it related to variations in some form of epitaxial interaction that is related to the dispersion state of the MMT? At present, we have no answers to such questions but what is entirely clear is that these systems are not at all easy , . Continuing with this theme, dispersion is one of the most problematic areas for nanodielectrics, because systems containing large clumps behave at best like conventional micrometric composites. But how can dispersion best be ascertained? While small angle neutron scattering and detailed synchrotron studies may appeal to the academic researcher, they are of little technological use. Epoxies are widely used as the basis of many practical insulation systems and such materials are often processed in the liquid state. If the addition of nanofillers were to affect the flow behavior of the host polymer then we may potentially have a practical quality control tool with which to monitor dispersion. Figure 8(a) compares the rheological properties of a simple model epoxy with the same resin containing 5% by mass of silica. From this, the addition of 5% by mass of conventional micro-silica results in a significant increase in viscosity across the range of shear rates shown, but this is much less than the increase seen in the system containing the same proportion of nano-silica. Figure 8(b) compares the flow behavior of three systems: neat epoxy; epoxy plus 5% by mass of the mineral nanofiller boehmite; epoxy plus 5% by mass of a boehmite that has been chemically modified to render it more compatible with the host polymer. Although both boehmite nanocomposites are more viscous then the base resin, the increase in viscosity is much more marked in the chemically compatibilized system. Also, all three of the nanocomposite data
Figure 8. Rheological data comparing a number of composite systems with the behavior of the matrix resin alone: (a) data showing the effect of the filler size (systems containing 5% by mass silica); (b) data showing the effect of surface chemistry (nanocomposites containing 5% by mass boehmite). In all cases, data indicated ▲ were obtained with an increasing ramp rate whereas those indicated ▼ were obtained subsequently with a decreasing ramp rate. Hysteresis effects appear in the better-dispersed systems.
sets shown in Figure 8 exhibit some degree of hysteresis. Are these effects related to dispersion? Certainly, the material containing the chemically modified nanofiller exhibits significantly higher breakdown strength than the unmodified analogue. These are promising results worthy of further study.
ing sections we sought to illustrate this with reference to some examples. Also, how do we reliably manufacture well-dispersed systems, and how do we unequivocably characterize the state of dispersion? Can we develop robust online process monitoring systems for use during commercial manufacture? We have seen that certain aspects of dielectric behavior, namely electrical erosion and dielectric spectroscopic characteristics, are comparatively well understood, but that we still have very little understanding of the physics underpinning the charge dynamics and electrical breakdown behavior of these systems. One can well imagine that in the long run, an understanding of the aging mechanisms and behavior of nanocomposites will also be of critical importance to their use in practical high voltage plant. Finally, we understand little about the long-term toxicity of nanocomposites. It should never be assumed that nanoparticles are as harmless to health as their macroscopic counterparts. First, their massively increased surface area means that they have a greater chemical activity, perhaps as catalysts for the production of damaging radicals . Second, once they enter the body they can penetrate deeper into biological systems, potentially accumulating in individual cells. We only consider here the question of inhalation, though wider considerations such as marine toxicity are equally important. Soto et al.  have studied the toxicity of a range of nanoparticles, including Ag, TiO2, Fe2O3, Al2O3, ZrO2, Si3N4 and crysotile asbestos, remarking “it would seem unrealistic to repeat the failures of the asbestos industry, which largely ignored the product dangers for more than 2000 years.” Since Toyota reported their preparation of nylon-6/clay hybrids in the early 1990s, academic and industrial interest in nanocomposites has grown exponentially, using this term in its literal mathematical sense (see Figure 3). Many novel materials are being developed that are of great interest to the electrical insulation engineer, in that they appear to bring benefits including increased mechanical strength, reduced mechanical and electrical erosion, improved electrical breakdown/endurance behavior and space charge mitigation. Nevertheless, the field is still in its infancy; for example, the review by Tanaka et al.  describes reports of contradictory effects, indicating clearly that much remains to be understood. So, to return to our starting point; nanodielectrics—how much do we really understand? While the potential appears great, we would suggest that it is important to temper our enthusiasm by remembering that there is still much to understand in this fascinating class of materials.
Acknowledgements The authors wish to acknowledge the contributions of Dr. David Saunders of the British Museum, London, for supplying Figure 1 and Mr. Toby Matheson for supplying the data presented in Figure 8.
In this article, we have sought to introduce some of the salient References features of nanocomposite research that relate to the development  G. L. Hornyaka, C. J. Patrissia, E. B. Oberhausera, et al., “Effective medium theory characterization of Au/Ag nanoalloy-porous alumina and application of useful dielectrics. We have seen that although composites,” Nanostruct. Mater., vol. 9, pp. 571–574, 1997. nanocomposites appear to offer a variety of advantages over  M. Faraday, “The Bakerian lecture: experimental relations of gold their micron-filled counterparts, nonetheless, there are still many (and other metals) to light,” Phil. Trans. Roy. Soc. Lon., vol. 147, unanswered questions and obstacles to overcome. In the precedpp. 145–181, 1857.
IEEE Electrical Insulation Magazine
 M. Reibold, P. Paufler, A. A. Levin, et al., “Materials - Carbon nanotubes in an ancient Damascus saber,” Nature, vol. 444, pp. 286–286, 2006.  U. S. Patent no. 2, 531, 396.  T. Tanaka, G. C. Montanari, and R. Mülhaupt, “Polymer nanocomposites as dielectrics and electrical insulation - perspectives for processing technologies, material characterization and future applications,” IEEE Trans. Dielectr. Electr. Insul., vol. 11, pp. 763–784, 2004.  S. S. Ray and M. Okamoto, “Polymer/layered-silicate nanocomposites: A review from preparation to processing,” Prog. Polym. Sci., vol. 28, pp. 1539–1641, 2003.  S. C. Tjong, “Structural and mechanical properties of polymer nanocomposites,” Mater. Sci. & Eng. R: Reports, vol. 53, pp. 73–197, 2006.  C. J. T. Landry, B. K. Coltrain, M. R. Landry, J. J. Fitzgerald, and V. K. Long, “Poly(vinyl acetate) / Silica filled materials: Material properties of in situ vs fumed silica particles,” Macromolecules, vol. 26, pp. 3702–3712, 1993.  K. U. Kirst, F. Kremer, and V. M. Litvinov, “Broad-band dielectric spectroscopy on the molecular dynamics of bulk and adsorbed poly(dimethylsiloxane)”, Macromolecules, vol. 26, pp. 975–980, 1993.  T. J. Lewis, “Nanometric dielectrics,” IEEE Trans. Dielectr. Electr. Insul., vol. 1, pp. 812–825, 1994.  T. J. Lewis, “Interfaces: Nanometric dielectrics,” J. Phys. D: Appl. Phys., vol. 38, pp. 202–212, 2005.  W. S. Jou, H. Z. Cheng, and C. F. Hsu, “The electromagnetic shielding effectiveness of carbon nanotubes polymer composites,” J. Alloys & Comp., vol. 434, pp. 641–645, 2007.  G. Cao, Nanostructures & Nanomaterials: Synthesis, Properties & Applications, Imperial College Press, London, 2004, p. 51.  M. Roy, J. K. Nelson, R. K. MacCrone, et al., “Polymer nanocomposite dielectrics – The role of the interface,” IEEE Trans. Dielectr. Electr. Insul., vol. 12, pp. 629–643, 2005.  K. H. Wang, M. H. Choi, C. M. Koo, et al., “Synthesis and characterization of maleated polyethylene/clay nanocomposites,” Polymer, vol. 42, pp. 9819–9826, 2001.  H. R. Dennis, D. L. Hunter, D. Chang, et al., “Effect of melt processing conditions on the extent of exfoliation in organoclay-based nanocomposites,” Polymer, vol. 42, pp. 9513–9522, 2001.  L. F. Drummy, H. Koerner, K. Farmer, et al., “High-resolution electron microscopy of montmorillonite and montmorillonite/epoxy nanocomposites,” J. Phys. Chem. B., vol. 109, pp. 17868–17878, 2005.  K. Wang, L. Wang, J. S. Wu, et al., “Preparation of highly exfoliated epoxy/clay nanocomposites by “slurry compounding”: Process and mechanisms,” Langmuir, vol. 21, pp. 3613–3618, 2005.  L. Sun, W. J. Boo, J. Liu, et al., “Preparation of intercalating agentfree epoxy/clay nanocomposites,” Polym. Eng. & Sci., vol. 47, pp. 1708–1714, 2007.  A. W. Musumeci, G. G. Silva, J-W. Liu, et al., “Structure and conductivity of multi-walled carbon nanotube/poly(3-hexylthiophene) composite films,” Polymer, vol. 48, pp. 1667–1678, 2007.  M. Casciola, A. Donnadio, M. Pica, et al., “Characterization of Zr phosphate/PVDF nanocomposites by vibrational spectroscopy,” Macromol. Symp., vol. 230, pp. 95–104, 2005.  S. S. Kumar, C. S. Kumar, J. Mathiyarasu, and K. L. Phani, “Stabilized gold nanoparticles by reduction using 3,4-ethylenedioxythiophene-polystyrenesulfonate in aqueous solutions: Nanocomposite formation, stability, and application in catalysis,” Langmuir, vol. 23, pp. 3401–3408, 2007.  M. Hernandez, B. Sixou, J. Duchet, and H. Sautereau, “The effect of
July/August 2008 — Vol. 24, No.4
dispersion state on PMMA–epoxy–clay ternary blends: In situ study and final morphologies,” Polymer, vol. 48, pp. 4075–4086, 2007.  F. Hussain, M. Hojjati, M. Okamoto, and R. E. Gorga, “Review article: Polymer-matrix nanocomposites, processing, manufacturing and application: An overview,” J. Compos. Mater., vol. 40, pp. 1511–1575, 2006.  D. A. Brune and J. Bicerano, “Micromechanics of nanocomposites: Comparison of tensile and compressive elastic moduli, and prediction of effects of incomplete exfoliation and imperfect alignment on modulus,” Polymer, vol. 43, pp. 369–387, 2002.  D. Shah , P. Maiti , E. Gunn, et al., “Dramatic enhancements in toughness of polyvinylidene fluoride nanocomposites via nanoclaydirected crystal structure and morphology,” Adv. Mater., vol. 16, pp. 1173–1177, 2004.  M. Sarwar, S. Zulfiqar, Z. Ahmad, “Preparation and properties of polyamide-titania nanocomposites,” J. Sol-Gel Sci. & Tech., vol. 44, pp. 41–46, 2007.  B. J. Ash, L. S. Schadler, R. W. Siegel, et al., “Mechanical properties of Al2O3 / polymethylmethacrylate nanocomposites,” Polym. Compos., vol. 23, pp. 1014–1025, 2002.  A. H. El-Hag, L. C. Simon, S. H. Jayaram, and E. A. Cherney, “Erosion resistance of nano-filled silicone rubber,” IEEE Trans. Dielectr. Electr. Insul., vol. 13, pp. 122–128, 2006.  A. B. Morgan, “Flame retarded polymer layered silicate nanocomposites: A review of commercial and open literature systems,” Polym. for Adv. Tech., vol. 17, pp. 206–217, 2006.  M. Kozako, N. Fuse, Y. Ohki, et al., “Surface degradation of polyamide nanocomposites caused by partial discharges using IEC(b) electrodes,” IEEE Trans. Dielectr. Electr. Insul., vol. 11, pp. 833–839, 2004.  R. Sarathi, P. R. Kumar, R. K. Sahu, “Analysis of surface degradation of epoxy nanocomposite due to tracking under AC and DC voltages,” Polym. Degrad. & Stabil., vol. 92, pp. 560–568, 2007.  S. K. Saha, “Observation of giant dielectric constant in an assembly of ultrafine Ag particles,” Phys. Rev. B., vol. 69, art. no. 1254161, 2004.  M. Roy, J. K. Nelson, R. K. MacCrone, and L. S. Schadler, “Candidate mechanisms controlling the electrical characteristics of silica/ XLPE nanodielectrics,” J. Mater. Sci., vol. 42, pp. 3789–3799, 2007.  C. D. Green, A. S. Vaughan, G. R. Mitchell, and T. Liu, “Structure property relationships in polyethylene / montmorillonite nanodielectrics,” IEEE Trans. Dielectr. Electr. Insul., vol. 15, pp. 134–143, 2008.  T. Imai, F. Sawa, T. Nakano, et al., “Effects of nano- and microfiller mixture on electrical insulation properties of epoxy based composites,” IEEE Trans. Dielectr. Electr. Insul., vol. 13, pp. 319–326, 2006.  S. Masuda, S. Okuzumi, R. Kurniant, et al., “DC conduction and electrical breakdown of MgO / LDPE nanocomposite,” IEEE Conf. Electr. Insul. Dielectr. Phenomena (CEIDP), 2007, pp. 290–293.  G. Chen, C. Zhang, and G. Stevens, “Space charge in LLDPE loaded with nanoparticles,” IEEE Conf. Electr. Insul. Dielectr. Phenomena (CEIDP), 2007, pp. 275–278.  A. S. Vaughan, S. G. Swingler, and Y. Zhang, “Polyethylene nanodielectrics: The influence of nanoclays on structure formation and dielectric breakdown,” IEEJ Trans. Fund. & Mater., vol. 126, pp. 1057–1063, 2006.  R. D. Handy and B. J. Shaw, “Toxic effects of nanoparticles and nanomaterials: Implications for public health, risk assessment and the public perception of nanotechnology,” Health, Risk & Society, vol. 9, pp. 125–144, 2007.
 K. Soto, K. M. Garzaa, and L. E. Murra, “Cytotoxic effects of aggregated nanomaterials,” Acta Biomaterialia., vol. 3, pp. 351–358, 2007.
Christopher Green graduated from Southampton University in the UK with a first class B.Eng. degree in electrical engineering in 2001. Born in Worcester, England in 1980, he is currently a Senior Research Student in the Tony Davies High Voltage Laboratory and will submit his Ph.D. thesis in 2008.
Alun Vaughan has a B.Sc. degree in chemical physics and a Ph.D. degree in polymer physics. After working at the UK’s Central Electricity Research Laboratories and spending a period as an academic at The University of Reading, he is now Professor in Electronics and Computer Science at the University of Southampton. He is an EPSRC College member, Honorary Treasurer of The Dielectrics Group of the Institute of Physics, a Fellow of the Institute of Physics, Senior Member of the IEEE and a member of the IET.