"Nanocomposites, Polymer–Clay". In ... - Wiley Online Library

ample of isotropic nanoparticles which either provide increased composite stiff- .... Small-angle x-ray scattering may provide some answers re- garding average ...
328KB taille 2 téléchargements 93 vues
NANOCOMPOSITES, POLYMER–CLAY Introduction Polymer matrix nanocomposites are a fairly new class of engineered materials which offer for a broad range of properties, an interesting and even radical alternative to more conventional filled polymers, yet at much lower filler loadings. They can be defined as polymer matrix systems in which the dispersed inorganic reinforcing phase has at least one of its dimensions in the nanometer range, which is quite close to the scale of elementary phenomena at the molecular level. The resulting unique combination of large interfacial area and small interparticle distance strongly influences nanocomposite behavior. Current status of research and industrial development of polymer nanocomposites clearly outlines the prominent position of clay nanocomposites and the present review is mainly devoted to the latter materials. From a general point of view, filler aspect ratio is a pertinent parameter to distinguish between various types of nanocomposites. Table 1 summarizes typical dimensions of particles under concern. Spherical silica particles are an example of isotropic nanoparticles which either provide increased composite stiffness while retaining matrix transparency, or exhibit novel optical properties by forming colloidal crystals. Although it is obviously critical to the optical behavior, it is generally observed that optimum mechanical properties are not achieved in conjunction with the best state of dispersion (1). When only two dimensions are in the nanometer range, fiber-like structures, such as whiskers or carbon nanotubes, with aspect ratios ranging between 50 and 1000 are dealt with. For instance, cellulose whiskers extracted from tunicate shells have been shown to dramatically improve composite stiffness in the case of poly(vinyl chloride) or 336 Encyclopedia of Polymer Science and Technology. Copyright John Wiley & Sons, Inc. All rights reserved.

Vol. 3



Table 1. Typical Nanofiller Dimensions Material


Typical dimensions

Silica particles Cellulose whiskers Carbon nanotubes Layered silicates

Spheres Rigid rods Flexible tubes (multiwall) Flexible discs

Diameter: 30–150 nm Diameter: 15 nm; length: 1 µm Diameter: 30 nm; length: 10–50 µm Diameter: 50–500 nm; thickness: 1 nm

poly(styrene-co-butyl acrylate) matrix in the rubbery state. In the latter material a 3 order of magnitude modulus improvement is achieved at only 6 mass% whisker content. Percolation of a rigid whisker network is evoked to account for such a property increase (2). Reaching a conductivity percolation threshold at very low loadings, owing to the nanoscale dispersion of carbon nanotubes with large aspect ratio, has also been of prime importance for designing conductive polymer blends with electrostatic paintability, while limiting resin embrittlement (3). Preliminary experiments with carbon nanotube–polymer composites also underline the potential of this nanofiller for entering the field of composite materials for structural applications as well (4). Opportunities for functional and/or structural applications of carbon nanotubes will mainly depend on the capacity to reach large-scale production at moderate cost and to monitor composite elaboration while retaining nanotube integrity. Among the variety of composites that display unique structure and behavior at the nanometer level, as compared to classical micrometer scale particulate filled materials, the use of layered silicates as a reinforcing phase is by far the most successful way of designing polymer nanocomposites with a broad range of markedly modified properties. A report from Toyota Central Research Laboratories of the development of a polyamide-6 (PA6)-based clay nanocomposite with remarkable thermomechanical behavior, at low clay loadings relative to conventional filler additives (below 5 mass% instead of 20–30 mass%), triggered extensive research efforts worldwide (5). In fact the benefits were shown not only for strength and stiffness, but also for thermal stability and barrier properties (6). Accordingly, the following presentation focuses on these materials, starting with a brief description of the layered silicates commonly used, followed by nanocomposite structural characterization and elaboration routes. Various properties of interest are reviewed together with the currently emerging structure–property relationship schemes.

Polymer-Layered Silicate Nanocomposites Layered silicates, the more widely used in polymer nanocomposites, belong to the same structural group, the 2:1 phyllosilicates, and more specifically to the smectite group. They comprise natural clay minerals such as montmorillonite, hectorite, and saponite and also synthetic layered minerals, fluorohectorite, laponite, or magadiite. An idealized structure for montmorillonite is presented in Figure 1. Elementary clay platelets consist of a 1-nm-thick layer made of two tetrahedral sheets of silica fused to an edge-shared octahedral sheet of alumina or magnesia. Isomorphic cation substitution results in an excess of negative charges within



Fig. 1. Structure of montmorillonite.

Al or Mg;

Vol. 3

OH;  O;

exchangeable cations.

the layer. Stacking of the layers leads to a regular van der Waals gap called gallery or interlayer. Cations located in the galleries (eg Ca2+ , Na+ ) counterbalance the excess layer charges. They are usually hydrated. This negative surface charge is quantified as the cation-exchange capacity (CEC), usually in the range from 80 to 150 meq/100 g for smectites. The interlayer space can be penetrated by organic cations or polar organic liquids as well. Exchange reactions with organic cations (aminoacids, alkylammonium ions, etc) enable to render the silicate surface organophilic. A key parameter of the stacking is the basal spacing or d-spacing, which ranges from 0.96 nm (ie, the layer thickness) in the fully collapsed state to about 2 nm, depending on the nature of the interlayer cation and the amount of adsorbed water (7). Individual lamellae have a high aspect ratio, with a diameter typically in the range 50–500 nm. Primary crystals (also called tactoids) consist of 8–10 lamellae, with usually disordered stacking. Their aggregation leads to a turbostratic structure. This organization is reflected in the x-ray diffraction (xrd) pattern where diffuse hk bands rather than sharp hkl reflections are observed. The basal reflection (001) is of more interest since it is used to derive the d-spacing. It will be dependent on the amount and nature of intercalated molecules in the galleries. Characterization of Nanocomposite Microstructure. Mixing clay with a polymer does not necessarily lead to a nanocomposite. Elaboration strategies are aimed at monitoring dispersion of the inorganic compound at the nanometer level, that is down to the elementary clay platelet. Figure 2 provides a schematic illustration of the various microstructures readily achievable, namely a conventional filled polymer with clay particles in the micrometer range, an

Vol. 3






Fig. 2. Schematic representation of clay platelet dispersion as (a) primary crystals, (b) intercalated, or (c) exfoliated structures.

intercalated nanocomposite in which extended polymer chains are intercalated in the gap between silicate layers while the stacking order is retained (note that in this case the host gallery height is much smaller than the radius of gyration of the polymer chain), and a delaminated (or exfoliated) nanocomposite where clay layers are individually dispersed in the host polymer matrix. The two basic tools used to elucidate nanocomposite morphology are x-ray diffraction (xrd) and transmission electron microscopy (tem). They provide complementary information on clay dispersion in the host matrix. The basal plane reflection (00l) diffraction peak yields a direct evaluation of the d-spacing between the clay lamellae, as long as layer registry is preserved to some extent. It allows therefore to follow the structural evolution from the pristine clay stacking to any intercalated state. With changes in d values in the range 1–4 nm, such data are accessible at low diffraction angle, ie, for 1 < 2 < 9◦ . Follow up of the increase in peak width at half maximum also reflects the increase in the degree of disorder in layer stacking during the intercalation process [see eg, (8)]. Figure 3 illustrates the change in d-spacing of a stearylammonium-modified montmorillonite (C18 Mt) blended with a maleic anhydride-modified polypropylene



Vol. 3


II 6.0 nm

6.3 nm

(d) 5.0 nm

Intensity, a.u.

(c) 3.4 nm

Intensity, a.u.


5.7 nm

(c) 3.5 nm

(b) (b) 2.1 nm 2.1 nm

(a) 0.5


3.5 5 6.5 2Θ -Cu, deg



(a) 0.5


3.5 5 6.5 2Θ -Cu, deg



Fig. 3. Evolution of the d-spacing showing montmorillonite intercalation in PP-MA oligomer (I), and no subsequent change upon further blending with PP (II). Reproduced from Ref. 9.

oligomer (PP-MA). As the amount of the latter component is increased, the interlayer distance increases drastically (spectra Ia–d), indicating the efficiency of the intercalation process. Spectrum IIa reveals the lack of intercalation of C18 Mt by pure PP. Spectra IIb–d do not show any change in the state of intercalation assessed in Spectra I when the PP-MA intercalated organoclay is further dipersed in a PP matrix. It is concluded that PP penetration in the galleries is not favored in this particular case (9). Whenever the xrd pattern becomes silent, meaning that the interlayer spacing is no longer accessible by conventional wide-angle x-ray diffraction, the need still remains to have access to a mesoscale description of the spatial distribution of the clay platelets. Small-angle x-ray scattering may provide some answers regarding average interparticle distance, and also about platelet orientation with respect to process geometry (10). Transmission electron microscopy has the great advantage to give a direct view of the microstructure. Typical micrographs reveal alternating dark and bright bands refering to the silicate layers and polymer matrix respectively. The illustrations of Figure 4 show exfoliated and intercalated-exfoliated structures in the case of 1 and 5 mass% dimethylditallow ammonium-modified montmorillonite

Vol. 3


500 nm


200 nm

Fig. 4. Transmission electron microscopy of water-aided melt-dispersed organoclay in PA6. Courtesy of Dr M. van Es, DSM Reseach, Geleen.

100 nm

Fig. 5. Transmission electron microscopy of a melt-intercalated organoclay tactoid in a PP matrix. Courtesy of Dr M. Bacia, UST Lille.

dispersed by melt extrusion in a PA6 matrix. Additionally tem enables imaging of any mesoscale long-range ordering of the clay. It also reveals platelet flexibility promoted by the high aspect ratio and nanometer thickness as illustrated in Figure 5 for an intercalated organoclay tactoid in a PP matrix. An accurate description of polymer clay nanocomposites is therefore available from a combination of electron microscopy and xrd techniques.

Elaboration of Polymer–Organoclay Nanocomposites Various methods have been developed in order to prepare polymer-layered silicate nanocomposites. These include in situ polymerization, polymer intercalation in



Vol. 3

solution, emulsion polymerization in the presence of layered silicates, and melt intercalation. A few examples are given below to illustrate the different strategies. Review articles are available for more detailed information [see eg (11–13), and references therein]. In situ Polymerization. Early works in the 1960s (14) demonstrated the feasability of intercalation polymerization of methyl methacrylate (MMA) after insertion–adsorption of the polar monomer between the lamellae of a sodium montmorillonite. Polymerization was initiated by free-radical catalysts or with γ irradiation. Renewed interest for the method in the last decade came from the work on PA6–clay hybrids at Toyota Central Research Laboratories (5), following a scheme adopted earlier by Unitika Co. (15). Sodium montmorillonite was first cation exchanged with ω-amino acids [H3 N+ (CH2 )n − 1 COOH]. X-ray diffraction data show that the basal spacing is highly sensitive to the length of the alkyl chain. For n > 11, the ω-amino acid chain lies slanted to the layer and provides optimum swelling behavior by ε-caprolactam. 12-Amino-l-lauric acid (n = 12) modified montmorillonite (12-Mt) was used for performing the in situ ringopening polymerization of ε-caprolactam. The carbonyl end groups of 12-Mt initiate polymerization and an increasing amount of polyamide-6 is bonded to the clay as the 12-Mt clay content is increased as clearly established by nmr studies (16). The resulting hybrid is either exfoliated at low 12-Mt content or increasingly intercalated beyond 10 wt% clay. In an alternative approach, the same team successfully developed a so-called one-pot synthesis of PA6–clay hybrid without preliminary cation exchange of the montmorillonite. Ring-opening polymerization of ε-caprolactam with 6-amino-l-caproic acid as an accelerator was performed in a water dispersion of montmorillonite in the presence of an acid. Phosphoric acid seems to be the best candidate to achieve true exfoliation in this particular process (17). 12-Amino-l-lauric acid was also successfully used both as a fluorinated silicate modifier and as a monomer in order to prepare a polyamide-12-based nanocomposite with exfoliated–intercalated structure (18). Some attempts to produce polystyrene (PS)-based nanocomposites through the in situ polymerization route have been reported (19). Intercalated structures are obtained. Intercalative polymerization of ε-caprolactone has also been achieved in the presence of α-protonated amino acid-exchanged montmorillonite. Upon heating, the organic acid groups initiate ring-opening polymerization of the monomer and the resulting polymer is ionically bound to the silicate platelets with a good level of delamination as revealed by xrd (20). In situ polymerization has also been extended to polyolefins, polyesters, or polycarbonate in recent years. The production of thermoset-based nanocomposites by the same method has been investigated by many authors (21–23). In the case of epoxy–clay nanocomposites, the organophilic clay is first swollen in the mixture of epoxy prepolymer and curing agent. The gel state and final structure are strongly dependent on the nature of the onium ion, cross-linking amine, and curing conditions. In particular, larger chain length of the alkylammonium and ion-exchange with protonated primary amines should be preferred. Polyurethane networks are equally good candidates for clay nanolayer reinforcement (22). The approach consists in solvating polyol precursors in montmorillonite exchanged with long-chain onium ions and further adding the diisocyanate curing agent. Intercalated tactoids are obtained in the final cured nanocomposite.

Vol. 3



Intercalation in Solution. In the case of water-soluble polymers, it is possible to prepare solvent-cast nanocomposites by using water as a cosolvent. Pristine montmorillonite can be easily dispersed in water, owing to its hydrophilic character, and blending for instance with polymers such as poly(ethylene oxide) (PEO) or poly(vinyl alcohol) (PVOH) is thus achievable (see eg (24–26)). Intercalated– exfoliated structures are observed for PVOH-based nanocomposites. On the contrary, in the presence of PEO, intercalated silicate layers are organized as large clay tactoids; this indicates that reaggregation of the initial silicate water suspension occurred during the film casting process. A similar elaboration strategy has also been developed, starting from an organoclay. In the examples of poly(εcaprolactone) and poly(l-lactide), chloroform was used as a cosolvent. In both situations no evidence of intercalation could be found but the clay tactoids displayed a remarkable geometric arrangement, with their surfaces lying parallel to the cast film surface (27). These few illustrations are indicative of the high sensitivity of the final materials structure to the nature of the host matrix and to the interplay of polymer–polymer and polymer–clay interactions. The above technique can be advantageously adapted to a situation where the polymer of interest is not soluble in any solvent, as is the case for polyimides. The polyimide precursor, ie, a poly(amic acid) solution in dimethylacetamide, is prepared and blended to a dispersion of an organomodified montmorillonite in the same solvent. Dimethylacetamide is then gradually removed, and upon heating the poly(amic acid) film at 300◦ C under nitrogen atmosphere the polyimide–clay hybrid is obtained. When an ammonium salt of dodecylamine is used in the ionexchange process, true exfoliation is achieved and furthermore the clay platelets align parallel to the film surface (28). Emulsion Polymerization. Considering the high hydrophilic character of sodium montmorillonite, it was anticipated that polymerization in an aqueous medium might provide an alternative route for polymer–clay nanocomposite preparation. The first report dealing with such an approach concerned the emulsion polymerization of MMA dispersed in a water phase in the presence of Na+ montmorillonite. The structural analysis confirms intercalation of the PMMA in the clay galleries. A PS–clay nanocomposite has been elaborated according to the same procedure, with the same resulting intercalated morphology. Intercalation by the emulsion technique was also achieved for an epoxy system, again without requiring any ion-exchange treatment (29). Melt Intercalation. Although emulsion polymerization or in situ polymerization may be considered as viable routes for industrial-scale production, as exemplified for PA6 with the latter process, melt compounding remains the most obvious route for cost-effective development of nanoreinforced polymers. The first industrial applications refer to the PA6–clay hybrid in situ polymerization process patented by Toyota (30), but growing interest in achieving clay nanodispersion by melt compounding is observed worldwide (31). Patent literature is clearly indicative of the current trend with many references to polyamides and other polar polymers such as polyesters or polyimides [see eg (32)]. Emphasis is put on the nature of the surfactant and processing conditions. Nonpolar polymers such as polyolefins are also highly attractive candidates with research efforts primarily driven by the automotive market. Polypropylene–clay hybrids have been prepared by melt mixing an organoclay, maleic anhydride-modified PP oligomers (PP-MA),



Vol. 3

and PP. The polar PP-MA intercalates in the clay galleries and the quality of clay dispersion in the hybrids are clearly affected by the degree of miscibility of the polar oligomers in PP (9,33). Direct melt intercalation of PEO in pristine montmorillonite by static annealing of cold-pressed powder mixture slightly above the melting temperature of PEO has been reported (34). The resulting d-spacing of the order of 1.8 nm indicates that the PEO chains are constrained in a 0.8-nm interlayer space. Differential scanning calorimetry studies reveal that the polymer is deprived of any thermodynamic transition. Neither the heat capacity jump characteristic of the glass transition nor the melting endotherm is observed. Chain dynamics appear quite peculiar in these systems. Thermally stimulated current results point at the essentially noncooperative nature of the motions of the confined chains. Intercalation kinetics have also been followed by xrd monitoring of the basal spacing reflection in model systems consisting of monodisperse PS and organically modified fluorohectorite. The most striking result is that the mobility of the polymer chains in this confined environment seems larger than that in the bulk melt (34,35). Spin-echo nmr experiments using deuterated PS suggest the coexistence of multiple environments, from solid-like to liquid-like, for intercalated chains. Molecular dynamics simulations relate this complex dynamic behavior to strong density inhomogeneity normal to the surface [(36), and Refs. 45 to 49 therein]. An extensive investigation of nanocomposite preparation with the aid of a swelling agent that is known to intercalate the clay provides some additional interesting experimental ground on the phase behavior. The results underline the influence of the polymer/swelling agent miscibility (as assessed by a solubility parameter approach) on nanocomposite formation. The example of an epoxy monomer as the swelling agent shows that either complete miscibility or strong immiscibility are preferable (37).

Structure Development in Polymer–Clay Nanocomposites Whatever the elaboration route is, understanding phase behavior of the resulting nanocomposites is of prime importance to achieve reliable material development. A lattice-based mean field model has been developed in order to address the problem of nanocomposite formation (38). By deriving the evolution of the free energy of the system with interlayer spacing, the model provides some basic predictions regarding the equilibrium states (ie, exfoliated, intercalated, or immiscible), in relation to the enthalpic and entropic factors of the interacting constituents, namely the silicate, the tethered surfactant chains, and the polymer. Polymer confinement results in entropy loss but the latter may be compensated in part by the entropy gain induced through the increase in conformational freedom of the tethered surfactant chain upon layer separation. As a consequence, melt intercalation is predicted to depend primarily on energetic (enthalpic) factors. Polar polymers or polymers containing groups interacting with the silicate surface will favor polymer–clay hybrid formation (35,38). A theoretical investigation of the phase behavior of model analogues of polymer–clay nanocomposites has been conducted (39). Combining a selfconsistent field model with density functional theory, the investigation underlines

Vol. 3



some trends regarding phase morphology and stability for these systems. In particular, the calculations point at the key influence of the surfactant chain length. Polymer-like values lead to enhanced miscibility of the clay platelets and polymer matrix (exfoliated structure), even at moderate level of interactions between the grafted chains and polymer melt (39). The equilibrium behavior of a mixture of functionalized and nonfunctionalized chains (taken chemically identical) in the presence of two infinite planar surfaces has alternatively been considered using scaling theory. Functionalized chains have a telechelic architecture; ie, the active groups are located at each end of the chain. Qualitative phase diagrams are derived with prime consideration of the respective influences of interaction energy between the surface and the end group, and volume fraction φ of functionalized chains. Provided the interaction energy ε is high enough, exfoliation occurs more easily as φ is increased, whereas low ε values result in an immiscible mixture (40).

Nanocomposites Behavior Thermal Stability. Thermal stability improvement was already recognized in the pioneering work by Blumstein on PMMA intercalated in montmorillonite. Intercalated PMMA degraded at a temperature 50◦ C higher than that of conventional unfilled PMMA. In recent years, thermogravimetric analysis of various polymer–clay systems have confirmed this observation even for low nanofiller loadings [see eg (13,41), and references therein]. A particularly striking example is that of cross-linked poly(dimethylsiloxane) incorporating 10 mass% exfoliated organomontmorillonite (42) for which thermal stability under nitrogen flow is enhanced by 140◦ C. Overall, restricted thermal motion in silicate interlayers and hindering of the diffusion of decomposition products are certainly key factors, but polymer structure and nature of the degradation mechanisms and degradation conditions are equally important to account for the disparities observed in literature. Owing to what has been said previously on the role of the organic modifier of the layered silicate in elaboration and phase behavior of the polymer–clay nanocomposites, understanding of the key structural factors that influence its thermal degradation is of prime importance. Major concern is of course toward processing conditions and fire-retardant behavior as well. Recent work (43) on alkyl quaternary ammonium montmorillonite by a combination of thermogravimetric analysis, Fourier transform infrared spectroscopy, and mass spectrometry points at the complex degradation behavior of the organic surfactant. The initial degradation temperature is insensitive to chain architecture and length, or exchange ratio. Compared to the parent alkyl ammonium, the thermal stability of a fraction of the surfactant is substantially lowered because of catalytic sites on the layered silicates. Polymer processing generally implying temperatures higher than 180◦ C, chemical degradation of the surfactant is expected to occur. Many questions remain regarding the role of the decomposition products on the melt intercalation mechanism and subsequent nanocomposite phase stability. In this respect the potential of nmr techniques to follow the fate of the organic modifier seems quite promising (44).



Vol. 3

Crystal Organization. Semicrystalline polymer nanocomposites present a unique interplay between nanoscale morphology of crystal lamellae on one hand and clay platelet organization on the other. Considering the importance of interfacial interactions and the confined chain environment, one may expect drastic changes in crystal organization. The case of PA6–clay hybrids is rather well documented. The preferred orientation of the silicate layers under melt-flow conditions, together with polymer confinement, will affect crystallization behavior. For instance, in injection-molded bars, close to the surface both the silicate layers and polymer chain axes (and hence lamella thickness) are parallel to the surface, whereas the chain axes rotate by 90◦ inside the bar and remain perpendicular to the silicate layers (45). Clay also favors nucleation of the γ phase (16,44), contrary to bulk PA6 which predominantly crystallizes in the more stable α form. The elevated temperature (205◦ C) crystal morphology of PA6–clay hybrids has been examined by performing simultaneous small- and wide-angle x-ray scattering (46). These results clearly establish that the nanoscale correlations of the silicate layer organization (40–60 nm) affect polymer crystallization, resulting in a less-ordered crystal γ phase. Evidence is also provided of the impact of polymer-layer interactions (tethered vs nontethered chains), the more defective lamellae pertaining to the in situ polymerized (tethered) nanocomposites. Spherulitic structure is unable to develop as in bulk polymers, which ought to influence the nonelastic mechanical response. Similar crystallization behavior is also observed in other semicrystalline polymers such as poly(ε-caprolactone) or poly(ethylene terephtalate) nanocomposites (27,47). The observation of high melting temperature phases, though defective, might come from a reduced entropy of fusion Sf due to the confined crystallization environment. Fire-Retardant Behavior. Controlling polymer flammability remains a key issue in numerous applications of engineering plastics and commodity polymers as well. The fire-retardant additive approach provides cost-effective solutions, but generally at the expenses of some physical and mechanical properties. There is also growing pressure for environmentally safe products and processes, including recyclability and use of halogen-free compounds. For these reasons, recognition of improved flammability properties in the case of polymer–clay nanocomposites has triggered the development of extensive research programs on a large variety of materials (41,48,49). Cone calorimetry is used to evaluate the flammability under fire-like conditions. Relevant parameters such as the rate of heat release (HRR) and its peak value, heat of combustion (Hc), smoke yield (specific extension area, SEA), and carbon monoxide yield are obtained. Table 2 shows some typical data for layered silicate nanocomposites based on organically treated montmorillonite, with polyamide 6, poly(propylene-graft-maleic anhydride), and polystyrene as the host matrix. Nanocomposites under investigation have either delaminated (PA6) or intercalated–delaminated structures. In all cases there is a substantial reduction in peak HRR value (50–75%), whereas Hc and CO formation show little variation. Table 2 also compares the PS–clay nanocomposite with a PS–clay mix for which intercalation does not occur. The kinetics of heat release are displayed in Figure 6. In the case of the mix, particle dispersion is only achieved at the primary particle micrometer level. The peak HRR value remains identical to

Vol. 3



Table 2. Cone Calorimeter Dataa

Nylon-6 Nylon-6–clay 2% delaminated Nylon-6–clay 5% delaminated PS PS–clay mix 3% immiscible PS–clay 3% intercalated PP-g-MA PP-g-MA–clay 5% intercalated a After

Residue yield, %

Peak HRR, kW/m2

Mean HRR, kW/m2

Mean Hc, MJ/kg

Total heat released, MJ/m2

Mean SEA, m2 /kg

Mean CO yield, kg/kg

1.0 3.0

1011 686

603 390

27 27

413 406

197 271

0.01 0.01








0 3.2

1118 1080

703 715

29 29

102 96

1464 1836

0.09 0.09








0 8.0

2028 922

861 651

38 37

219 179

756 994

0.04 0.05

Ref. 48. 1400

Heat release rate, kW/m2

1200 1000 800 600 400 200 0 0



150 Time, s




Fig. 6. Kinetics of heat release rate for PS-based compounds. Reproduced from Ref. 48. PS pure; PS immiscible (3% silicate); PS intercalated (3% silicate).

that of pure PS while the intercalated PS–clay nanocomposite shows a 50% reduction. The observed reduced flammability in the nanocomposites may not be attributed to an additional retention of carbon alone since the residue yields are not markedly increased. Some key insights are provided by radiative gasification experiments which enable to follow pyrolysis either in a nitrogen or in a nitrogen–oxygen atmosphere. The example of PA6 nanocomposites is revealing



Vol. 3

enough of enhanced char formation and reduced mass loss rate in comparison to pure PA6. The current interpretation of these results is that the nanocomposite flame-retardant mechanism occurs through the build up of a reinforced char layer, which acts as an insulator, and a mass transport barrier so as to slow down the escape of the decomposition products. Developments according to this concept include the use of a PA6–clay nanocomposite in intumescent formulations as in the case of ethylene–vinyl acetate copolymers with ammonium polyphosphate (50). Enhanced flame-retardant performance is related to the formation of a thermally stable phosphocarbonaceous structure in the char, and the blend even shows a slight improvement in mechanical behavior. Processing conditions also strongly influence the flame-retardant behavior. For example, in the case of PS-based nanocomposites, extrusion above 180◦ C under partially oxidative conditions yields an intercalated nanocomposite but with no flammability improvement, whereas the melt-extruded system at 170◦ C under nitrogen or vacuum exhibits flame-retardant efficiency (41). The way thermal degradation of the organic modifier alters the flammability reduction mechanism has yet to be understood. Barrier Properties. Early work in the Toyota Research group acknowledged the great potential of polymer–clay nanocomposites to reduce moisture absorption and decrease water and gas permeability, even at low clay loadings (6). A further advantage for packaging applications lies in the fact that lowering of permeability is achieved while preserving transparency, owing to the suitable dispersion of platelets smaller than the wavelength of visible light. The example of polyimide clay films illustrates the dramatic decrease of permeability coefficients. Only 2 mass% montmorillonite loading reduces the permeability by more than 50% of the pure polymer value for water vapor, oxygen, or helium. Notwithstanding possible changes in diffusion and/or solubility, it has been postulated that the major role of the clay platelets consists in substantially increasing the path length of the permeant, that is by creating a highly tortuous path, due to the high aspect ratio of the clay. A simple theory derived by Nielsen expresses the relative permeability as follows: Pc /Po = 1/[1+(L/2W)Vf ] in which V f is the volume fraction of plates, L is the plate length, and W is the thickness. Pc and Po stand for the nanocomposite and polymer permeability respectively. Using equivalent loadings of clay but varying the aspect ratio yields results in fairly good agreement with the theoretical prediction (28). In the same way, a significant reduction in water vapor permeability was observed in the case of a poly(εcaprolactone)–organomontmorillonite nanocomposite, showing a fivefold reduction at only 4.8 vol% clay whereas it is only halved at best with a 20 vol% conventionally filled silicate composite (20). The tortuosity model shows discrepancies between actual aspect ratios measured by tem and values deduced from curve fitting. Among the reasons for that are the deviations from the ideal dispersion of platelets parallel to the film surface, and possible aggregation of individual platelets. Although tortuosity plays a role in barrier enhancement, some other factors ought to be taken into account (see BARRIER POLYMERS). Recent work on MXD6

Vol. 3



Table 3. Comparison of PA6 and PA6–Clay Hybrids Mechanical Propertiesa Property

PA6 clay hybrid



Tensile strength, MPa At 23◦ C At 120◦ C Tensile modulus, GPac At 23◦ C At 120◦ C Flexural strength, MPab At 23◦ C At 120◦ C Flexural modulus, GPac At 23◦ C At 120◦ C Izod impact, J/md Charpy impact, KJ/m2e HDT (1.82 MPa), ◦ C

97 32 1.9 0.6

69 27 1.1 0.2

143 33

89 12

4.3 1.2 18 6.1 152

2.0 0.3 21 6.2 65

a After

Ref. 5. convert MPa to psi, multiply by 145. c To convert GPa to psi, multiply by 145,000. d To convert J/m to ft·lbf/in., divide by 53.38. e To convert kJ/m2 to ft·lbf/in.2 , divide by 2.4. b To

nanocomposites is indicative of an oxygen permeation reduction in humid environment far beyond what is expected from the increase in path length alone (51). In the light of the findings regarding chain packing and dynamics in such confined environments (36), models for the prediction of barrier properties ought to take into account the changes induced in terms of solubility and diffusion. A conceptual model has been proposed (52). So far no general predictive approach is available. Mechanical Behavior. Being able to improve strength and stiffness with limited alteration of toughness is a goal readily achievable with polymer–clay nanocomposites (see MECHANICAL PROPERTIES; REINFORCEMENT). Table 3 gathers some key data of the original work by the Toyota group (5), which show the dramatic influence of organomontmorillonite on mechanical properties of PA6–clay hybrids at low mineral loading (4.7 mass%). The improvement in strength is claimed to have little or no influence on impact properties as evaluated from Izod or Charpy tests. Increase in modulus is paralleled by a substantial rise in heat distortion temperature. A concept of constrained polymer region related to the ion-bonding strength of clay and PA6 is introduced to account for the observed behavior. Linear Elastic and Rubber Elastic Behavior. Although stiffening is quite noticeable in the glassy regime of the amorphous phase, the most spectacular effect is seen in the rubber elastic regime phase, as already evoked in the case of reinforcement by cellulose whiskers (2). The PA6–clay hybrids example presented in Table 3 is quite representative of the situation encountered with semicrystalline thermoplastics, but elastomeric networks benefit as well of clay layer dispersion with a two- to threefold increase in modulus for polyurethane or epoxy networks



Vol. 3

(22). In the meantime, improved elongation at break is observed, contrary to what is seen in classical filled systems, presumably due in part to dangling chain formation in the network (see ELASTICITY, RUBBER-LIKE). The dynamic mechanical loss peak related to the glass-transition mechanism is equally informative on the extent of polymer–clay interaction. This shows mainly in the reduction of the loss peak area and/or in the evolution of the peak temperature with clay content and elaboration conditions (5,53). Predictive modeling of both storage and loss modulus faces a complex challenge in order to account for the mechanical coupling between the phases, the potential existence of an interphase, and/or a certain degree of connectivity between the fillers. In the latter situation, a percolation approach should be useful (2,54). Otherwise models derived from Halpin–Tsai equations seem quite promising for modulus prediction in relation to clay platelet arrangement (55,56). Plasticity and Rupture. The main drawback identified regarding the solidstate drawing behavior is certainly the limited elongation at break encountered for most thermoplastic–clay nanocomposites (9,33) in the vicinity or below the glass-transition temperature. Nanovoiding and subsequent extensive fibrillation of the polymer matrix is clearly evidenced from volume strain measurements during drawing (53) and from in situ tem observations (57). Such enhancement of nanoscale plasticity offers an opportunity for optimizing the stiffness/toughness balance. However critical microvoids may develop from areas where load transfer is no more achievable because of splitting of clay platelet aggregates. This points at the most critical issue in nanocomposite development, ie, monitoring of elaboration and processing conditions. Key Role of Processing. Scarce work has been devoted to the influence of processing on microstructure and properties of polymer–clay nanocomposites (58, 59). It is shown that twin-screw extrusion enables achieving a significant degree of dispersion of the clay platelets, provided residence time and degree of shearing are optimized in conjunction with the nature of the organoclay. Thermal stability of the organic modifier is again at the heart of the problem. In the same way as demonstration was made in the last decade of the importance of processing to design polymer blends, taking the full benefits of the interesting combination of properties displayed by polymer nanocomposites will mainly rely on key developments in the field of processing. Automotive and packaging markets are undoubtedly the driving force for it.

BIBLIOGRAPHY 1. J. M. Jethmalani and W. T. Ford, Chem. Mater. 8, 2138 (1996); Z. Pu and co-workers, Chem. Mater. 9, 2442 (1997). 2. L. Chazeau and co-workers, J. Appl. Polym. Sci. 71, 1797 (1999); V. Favier and co-workers, Polym. Eng. Sci. 37, 1732 (1997). 3. J. J. Scobbo, in SPE Antec 98 Proceedings, 1998, p. 2468. 4. D. Quian and co-workers, Appl. Phys. Lett. 76, 2868 (2000); L. Jin, C. Bower, and O. Zhou, Appl. Phys. Lett. 73, 1197 (2000). 5. A. Usuki and co-workers, J. Mater. Res. 8, 1174 (1993); Y. Kojima and co-workers, J. Mater. Res. 8, 1185 (1993). 6. A. Okada and co-workers, Polym. Prep. 28, 447 (1987).

Vol. 3



7. W. A. Deer, R. A. Howie, and J. Zussman, An Introduction to Rock-Forming Minerals, 2nd ed., Longman Scientific, Essex, U.K., 1992. 8. R. A. Vaia and co-workers, Chem. Mater. 8, 2628 (1996). 9. N. Hasegawa and co-workers, J. Appl. Polym. Sci. 67, 87 (1998). 10. K. Varlot and co-workers, J. Polym. Sci., Part B 39, 1360 (2001). 11. E. P. Giannelis, Adv. Mater. 8, 29 (1996). 12. P. C. Lebaron, Z. Wang, and T. J. Pinnavaia, Appl. Clay Sci. 15, 11 (1999). 13. M. Alexandre and P. Dubois, Mater. Sci. Eng. 28, 1 (2000). 14. A. Blumstein, J. Polym. Sci., Part A 3, 2653 (1965). 15. Jpn. Pat. JP-A- 51-109998 (1976), S. Fujiwara and T. Sakamoto (to Unitika). 16. R. D. Davis, W. L. Jarrett, and L. J. Mathias, Polym. Mater. Sci. Eng. 82, 272 (2000). 17. Y. Kojima and co-workers, J. Polym. Sci., Part A 31, 1755 (1993). 18. P. Reichert and co-workers, Acta Polym. 49, 116 (1998). 19. A. Akelah and A. Moet, J. Mater. Sci. 31, 3589 (1996); J. G. Doh and I. Cho, Polym. Bull. 41, 511 (1998). 20. P. B. Messersmith and E. P. Giannelis, J. Polym. Sci., Part A 33, 1047 (1995). 21. P. B. Messersmith and E. P. Giannelis, Chem. Mater. 6, 1719 (1994). 22. T. Lan, P. D. Kaviratna, and T. J. Pinnavaia, J. Phys. Chem. Solids 57, 1005 (1996); Z. Wang, T. Lan, and T. J. Pinnavaia, Chem. Mater. 8, 2200 (1996); Z. Wang and T. J. Pinnavaia, Chem. Mater. 10, 1820 (1998); Z. Wang and T. J. Pinnavaia, Chem. Mater. 10, 3769 (1998). ¨ 23. C. Zilg, R. Mulhaupt, and J. Finter, Macromol. Chem. Phys. 200, 661 (1999). 24. P. Aranda and E. Ruiz-Hitzky, Chem. Mater. 4, 1395 (1992); J. Wu and M. M. Lerner, Chem. Mater. 5, 835 (1993). 25. R. A. Vaia and co-workers, J. Polym. Sci., Part B 35, 59 (1997). 26. N. Ogata, S. Kawakage, and T. Ogihara, J. Appl. Polym. Sci. 66, 573 (1997); K. E. Strawhecker and E. Manias, Chem. Mater. 12, 2943 (2000). 27. N. Ogata and co-workers, J. Polym. Sci., Part B 35, 389 (1997); G. Jimenez and co-workers, J. Appl. Polym. Sci. 64, 2211 (1997). 28. K. Yano and co-workers, J. Polym. Sci., Part A 31, 2493 (1993); K. Yano, A. Usuki, and A. Okada, J. Polym. Sci., Part A 35, 2289 (1997). 29. D. C. Lee and L. W. Jang, J. Appl. Polym. Sci. 61, 1117 (1996); M. W. Noh and D. C. Lee, Polym. Bull. 42, 619 (1999); D. C. Lee and L. W. Jang, J. Appl. Polym. Sci. 68, 1997 (1998). 30. U.S. Pat. 4,739,007 (Apr. 19, 1988), A. Okada and co-workers (to Toyota Co.). 31. L. Liu, Z. Qi, and X. Zhu, J. Appl. Polym. Sci. 71, 1133 (1999). 32. WO Pat. 93-04117 (Mar. 4, 1993), M. R. Maxfield and co-workers (to Allied-Signal Inc.); WO Pat. 99-29767 (June 17, 1999), R. A. Korbee and co-workers (to DSM NV); WO Pat. 99-32403 (July 1, 1999), R. B. Barbee and co-workers (to Eastman Chemical Co.); WO Pat. 99-50340 (Oct. 7, 1999), H. Okamoto and co-workers (to Toyota Co.); WO Pat. 2000034180 (June 15, 2000), J. N. Gilmer and co-workers (to Eastman Chemical Co.). 33. M. Kawasumi and co-workers, Macromolecules 30, 6333 (1997). 34. R. A. Vaia and co-workers, J. Polym. Sci., Part B 35, 59 (1997); R. A. Vaia and co-workers, Macromolecules 28, 8080 (1995). 35. E. P. Giannelis, R. Krishnamoorti, and E. Manias, Adv. Polym. Sci. 138, 107 (1999). 36. R. A. Vaia and E. P. Giannelis, MRS Bull. 26, 394 (2001). 37. H. Ishida, S. Campbell, and J. Blackwell, Chem. Mater. 12, 1260 (2000). 38. R. A. Vaia and E. P. Giannelis, Macromolecules 30, 7990 (1997); 8000 (1997). 39. V. V. Ginzburg, C. Singh, and A. C. Balazs, Macromolecules 33, 1089 (2000). 40. D. V. Kuznetsov and A. C. Balazs, J. Chem. Phys. 112, 4365 (2000). 41. J. W. Gilman and co-workers, Chem. Mater. 12, 1866 (2000). 42. S. D. Burnside and E. P. Giannelis, Chem. Mater. 7, 1597 (1995).

352 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59.


Vol. 3

W. Xie and co-workers, Chem. Mater. 13, 2979 (2001). D. L. Vanderhart, A. Asano, and J. W. Gilman, Macromolecules 34, 3819 (2001). Y. Kojima and co-workers, J. Polym. Sci., Part B 33, 1039 (1995). D. M. Lincoln and co-workers, Polymer 42, 1621 (2001). Y. Ke, C. Long, and Z. Qi, J. Appl. Polym. Sci. 71, 1139 (1999). J. W. Gilman, T. Kashiwagi, and J. Lichtenham, SAMPE J. 33, 40 (1997); J. W. Gilman, Appl. Clay Sci. 15, 31 (1999). D. Porter, E. Metcalfe, and M.-J. K. Thomas, Fire Mater. 24, 45 (2000). S. Bourbigot and co-workers, Fire Mater. 24, 45 (2000). T. Lan, Y. Liang, G. W. Beall, and K. Kamena, in “Additives 1999” Conference Proceedings, San Francisco, Calif., 1999. G. W. Beall, in T. J. Pinnavaia and G. W. Beall, eds., Polymer-Clay Nanocomposites, (Wiley Series in Polymer Science), John Wiley & Sons, Inc., New York, 2001, p. 267. J. M. Gloaguen and J. M. Lefebvre, Polymer 42, 5841 (2001). E. Reynaud, C. Gauthier, and J. Perez, Rev. Metall.-CIT 169 (Feb. 1999). M. van Es, F. Xiqiao, J. van Turnhout, and E. van der Giessen, in “Additives 2000” Conference Proceedings, San Francisco, Calif., 2000. D. A. Brune and J. Bicerano, Polymer 43, 369 (2002). G. M. Kim and co-workers, Polymer 42, 1095 (2001). H. R. Dennis and co-workers, Plast. Eng. 57, 1, 56, (2001). J. W. Cho and D. R. Paul, Polymer 42, 1083 (2001); T. D. Fornes and co-workers, Polymer 42, 9929 (2001).

JEAN-MARC LEFEBVRE Universit´e des Sciences et Technologies de Lille