Twin-screw extrusion of polypropylene-clay nanocomposites: influence of masterbatch processing, screw rotation mode, and sequence.

INTRODUCTION

Polymer-clay nanocomposites (PCNs) have received intense research interest over the past decade, principally due to the significant macroscopic material property enhancements (e.g., increased tensile strength and modulus and improved permeation resistance, flame retardation, and heat distortion characteristics) achieved at a fraction of the loadings required for traditional fillers [1-4]. The most widely used nanoclay is derived from [Na.sup.+]-montmorillonite (MMT), whose 2:1 layered silicate structure consists of several stacked layers separated by a charged intergallery. Each silicate layer is 0.98-nm thick with a lateral dimension of 400-1000 nm [5]. The sum of the intergallery spacing and an individual silicate layer thickness is referred to as the basal (or [d.sub.001]) spacing, which is ~11.7 [Angstrom] for natural MMT [3]. Importantly, improvements in mechanical properties are often ascribed to the large surface area (ca. 750 [m.sup.2]/g) of MMT, which improves coupling between filler and polymer matrix and thereby stress transfer [6]. To promote compatibility between hydrophobic polymers (e.g., polypropylene (PP), polyethylene) and MMT, cation exchange reactions with alkylammonium surfactants are frequently performed to render the clay organophilic. Additionally, highly nonpolar polymers such as PP have an unfavorable enthalpic interaction with the organoclay that prohibits diffusion of polymer into the intergallery; consequently, a low molecular weight copolymer of polypropylene and maleic anhydride (PP-g-MA) is often added as a compatibilizer to facilitate intercalation and exfoliation of the organoclay and maximize its interfacial contact with the matrix polymer [7, 8].

The dispersion of organoclay in a polymer matrix is classified according to the degree to which individual silicate layers are separated from one another. In the order of decreasing layer-layer coherence, these states are referred to as agglomerated, intercalated, or exfoliated. In an agglomerated composite, the organoclay remains virtually unchanged from its native state. Alternatively, in an intercalated nanocomposite, the distance between silicate layers increases and the number of silicate layers per organoclay domain (or tactoid) decreases, leading eventually to the idealized exfoliated nanocomposite. Since an ideally exfoliated sample possesses the largest surface area to volume ratio for the filler, exfoliation leads to better phase homogeneity and the greatest improvements in thermal and mechanical properties [9, 10]. However, complete exfoliation is practically unrealizable, and the best dispersed systems typically consist of a mixture of intercalated and exfoliated structures. The ability to promote organoclay exfoliation is intimately connected to the preparation and blending strategy of PCNs. While in situ polymerization and solution blending methods of producing nanocomposites have been successfully employed for a number of polymers, these methods require large quantities of expensive, environmentally unfriendly solvents. As a result, melt-blending with high shear remains the most practical method of preparing PCNs [4].

Although melt-blending via extrusion is by far the most common and industrially practical blending strategy for PCNs, few research efforts have undertaken the considerable task of optimizing the extrusion process. Most scientific literature has focused on surface modification of MMT [11-16] and compatibilizer type and concentration [17-21] to maximize compatibility between the organoclay and polymer matrix. By contrast, investigations into the role of processing history and screw configuration are relatively scarce. In studying polyamide six nanocomposites, Dennis et al. concluded that increasing the mean residence time and shear intensity generally improved delamination and dispersion, although they noted that increasing the shear intensity beyond an optimal level led to poorer exfoliation and dispersion [22]. Wang et al. examined the influence of PP-g-MA compatibilizer and shear stress on dispersion of a PP-clay nanocomposite in a dynamic packing injection molding system, confirming the dual impact of chemistry and shear in determining dispersion [23]. Wang et al. also studied the role of the weight fraction and maleic anhydride (MA) content of the PP-g-MA compatibilizer [17, 18] and compatibilizer molecular weight [24]. They observed that compatibilizer to organoclay weight ratios beyond 3:1 gave rise to similar dispersions and that the balance between the stress and diffusivity must be considered in tandem in selecting compatibilizer molecular weight. Furthermore, the authors found that better dispersion was obtained by side-feeding pure PP into a blend of PP-g-MA and clay fed directly from the feed hopper rather than loading all three components simultaneously. They explained that such an approach afforded additional time for PP-g-MA to penetrate the clay galleries and exfoliate silicate layers prior to introducing the base resin. Similar conclusions have been drawn from studies of nylon 6 and PP, which employed a masterbatch approach to melt-blending organo clay with compatibilizer followed by a second dilution step with the matrix polymer [25, 26]. Other researchers have examined the effect of various processing variables, including feed rate, temperature, and screw speed (shear) on dispersion of organoclay in PP [27, 28]. Applying a range of extruder conditions, Lertwimolnun and Vergnes inferred from rheology that dispersion and exfoliation improved with increasing shear stress and mixing time but reduced barrel temperatures [27]. Similar conclusions were reached by Modesti et al., who claimed that the magnitude of the shear exerted on the polymer rather than its duration more strongly influenced exfoliation [29]. These results motivate a high shear, low temperature extrusion process.

One of the significant challenges associated with optimizing the extrusion conditions for preparing PCNs is the lack of common, accepted practice for characterizing exfoliation and dispersion. For example, X-ray diffraction (XRD), which is popular for its simplicity and availability, is widely used to quantify the mean basal spacing between silicate layers after processing. However, XRD is sensitive to the orientation distribution of organoclay domains relative to the incident X-ray beam, which affects the number of diffraction events during an XRD scan. As a result, XRD cannot be reliably used to quantify the number of silicate layers per tactoid, an essential parameter in characterizing the degree of exfoliation, and it provides little information on the state of the dispersion. Transmission electron microscopy (TEM) is often employed as a complementary method, since it offers a direct image of organoclay tactoids and their distribution in the polymer matrix. On the negative side, TEM requires rigorous, time consuming sample preparation, is limited to small sample volumes that may not represent the average state of the sample, and offers only a qualitative view of dispersion [30]. Rheology has proven to be a reliable means by which organoclay dispersion may be differentiated in PCNs, as it is well-known that both steady shear and small-amplitude oscillatory shear (SAOS) rheology are sensitive to the microstructure of filled polymer melts [30-43]. For example, divergent viscosity at low shear rates in steady shear rheology and flattening of the storage modulus in the terminal regime of SAOS are both associated with the increased solid-like response resulting from filler-filler interactions. In the case of PCNs, those interactions become stronger with superior exfoliation and dispersion of organoclay, and the resulting rheological signatures more pronounced. Consequently, rheology offers a powerful and convenient characterization tool to systematically evaluate extruder processing conditions and their relationship to organoclay dispersion.

Thus, we rely primarily upon rheology to appraise the dispersion of PP-clay nanocomposites melt-blended in a twin-screw extruder. Compatibilized (3:1 weight ratio of PP-g-MA to organoclay) and uncompatibilized blends of 1, 3, and 5 wt% PP-clay nanocomposites were prepared in a two-step extrusion process. For the compatibilized blends, a masterbatch of PP-g-MA and organoclay was melt-blended in the first step, followed by dilution with matrix PP in the second extrusion step. Uncompatibilized blends were simply processed twice with the appropriate ratio of organoclay and matrix PP. In both cases, the first and second extrusion steps were each conducted using corotating or counterrotating screws. The compatibilized masterbatch samples are contrasted with a compatibilized sample subjected to only a single pass. Steady shear and SAOS rheology were performed on all blends, and a novel holistic examination of key rheological features demonstrates the power of rheology in characterizing dispersion in PCNs. The importance of masterbatch processing, screw rotation mode, and sequencing is discussed within the context of these results.

EXPERIMENTAL

Materials

PP (Dow Chemical, grade H700-12, MFI = 12 g/10 min at 230[degrees]C and 2.16-kg load, [M.sub.w] = 229 kg/mol, [M.sub.w]/[M.sub.n] = 3.98) was used as the matrix polymer. Southern Clay Products provided Cloisite[R] 15A (C15A, 125 mequiv./100 g, [d.sub.001] = 3.15 nm), a natural MMT clay modified with a quaternary ammonium salt. Polybond[R] 3200 (PP-g-MA, Crompton, MFI = 110 g/10 min at 190[degrees]C and 2.16-kg load), a 1 wt% MA functionalized PP, was used as the compatibilzer.

Processing

Melt-blending was conducted with a Leistritz Micro 27 twin-screw extruder (L/D = 52, D = 27 mm) to prepare 1, 3, and 5 wt% clay nanocomposites. For all compatibilized samples, the ratio of PP-g-MA to C15A was fixed at a weight ratio of 3:1. C15A was introduced via a side-stuffer positioned at the halfway down the length of the screw (26D) and set at a metering rate appropriate to the desired loading.

Most samples were prepared using a two-step extrusion process. For the compatibilized samples, the first step blended PP-g-MA with C15A, creating a masterbatch (MB). Compatibilized MBs were prepared in both co- and counterrotation screw modes. In both cases, PP-g-MA was metered at 5 lb/hr to the screws while C15A was simultaneously added via the side-stuffer to obtain the desired ratio (75 wt% PP-g-MA and 25 wt% C15A). A third uncompatibilized MB sample was prepared by melt-blending in corotating mode with PP in place of PP-g-MA. Since compatibilizer is widely known to be critical to exfoliation and dispersion, a comprehensive examination of screw modes on uncompatibilized samples was judged unnecessary. The barrel temperatures and screw speeds were set at 200-210[degrees]C and 200 rpm, respectively for the MB extrusion step.

In the second dilution step, each of the three MB samples were dry mixed with appropriate amounts of PP (yielding 1, 3, and 5 wt% PCNs) and subsequently metered to the screws at 20 lb/hr. The dilution step was also conducted using both co- and counterrotating screws, but the barrel temperatures and screw speeds were set to 190[degrees]C and 300 rpm, respectively. The rationale behind the two sets of conditions was that preparing the MBs at a higher temperature (210[degrees]C) and lower screw rotation speed (200 rpm) would enhance diffusion of compatibilizer into clay intergalleries, while diluting with matrix PP at a lower temperature (190[degrees]C) but higher screw rotation speed (300 rpm) would increase stress and promote exfoliation and dispersion. In the subsequent discussion, samples prepared with PP-g-MA in the two-step process will be notated according to the screw rotation mode employed in each extrusion step. For instance, a compatibilized sample prepared with a counterrotating MB step and a corotating dilution step will be referred to as "Counter-Co". Uncompatibilized samples will be similarly described but will have the moniker "(None)" added to specify the absence of PP-g-MA.

Finally, single-throughput 1, 3, and 5 wt% compatibilized samples, subsequently denoted as "Co (No MB)", were also prepared in corotating mode by dry mixing PP and PP-g-MA and metering it to the screws at a rate of 5 lb/hr. C15A was again introduced via the side stuffer. Matrix PP and a blend of PP and 9 wt% PP-g-MA were also extruded once in corotating mode and serve as control materials. These single-throughput samples were processed at the same conditions as the MB samples described previously. All 23 samples (as well as the three intermediate MBs) were extruded into a water bath, pelletized, and oven dried at 80[degrees]C for 12 hr prior to the second melt-blending step or characterization.

The screw element stacks for the counter- and corotating twin screw extrusions employed in this work are of paramount importance in affecting exfoliation and dispersion of organoclay. The mixing and dispersion activity of the corotating screw configuration mainly results from four banks of kneading block elements with different stagger angles (30[degrees], 60[degrees], and 90[degrees]) between successive forward-conveying paddles and one dispersive mixing element trio. The kneading block elements in the corotating mode are forward conveying elements, leading to a mean residence time of 60 s. In the counterrotating configuration, dispersive mixing arises mainly from the four banks of helical lobe dispersive mixing element pairs and one combing distributive mixer element trio. These groupings induce a partial reverse conveying character that increases the mean residence time to 105 s. Detailed schematics of the screw element stacks are available from the manufacturer [44, 45]. For the single-throughput compounding process and MB step, only the 28D to 52D mixing elements (i.e., those downstream of the side-stuffer where organoclay is first introduced) contribute to the dispersion and exfoliation of nanoclay particles. For the second step of the two-step processes, all mixing elements contribute to clay exfoliation and dispersion.

X-Ray Diffraction

Samples used in XRD experiments were first compression molded at 210[degrees]C to a thickness of ~1.5 mm. XRD experiments were conducted using a Scintag XDS 2000 diffractometer with a Cu K[alpha] radiation source ([lambda] = 1.540562 [Angstrom]). Data were collected with a step size of 0.02[degrees] and a scan rate of 0.5[degrees]/min. The results presented are the average of two scans. The angle at which the primary peak in diffraction intensity is observed corresponds to the mean basal spacing of the organoclay, which is computed from Bragg's Law.

Rheology

All rheological tests were conducted with a TA Instruments AR 2000 rheometer using a 25-mm parallel plate geometry and environmental test chamber (ETC). A nitrogen purge of 10 l/min was continuously supplied to the ETC to inhibit oxidative degradation of the PP during experiments. Compression molded samples were melted at 210[degrees]C for 10 min in the rheometer, and the upper plate was then lowered to a gap distance of 1 mm for testing. All experiments were performed at 180[degrees]C.

Each sample was subjected to three types of rheological experiments: (1) oscillatory strain sweeps, (2) SAOS frequency sweeps, and (3) steady shear rate sweeps. These tests were selected because of their sensitivity to filler loading and dispersion and the unique experimental signatures associated with the development of a solid-like mechanical response. First, oscillatory strain sweeps were performed at a fixed frequency of 1 rad/s by varying the amplitude of the applied strain ([[gamma].sub.0]) from 0.001 to 10, although it should be noted that reliable data were only obtained up through a strain of ~8, beyond which melt fracture at the edges of the parallel plate geometry occurs. Typically, oscillatory strain sweeps serve to identify the critical strain corresponding to the boundary between linear and nonlinear viscoelastic behavior. For PCNs, several authors have reported that the critical strain ([[gamma].sub.c]) is inversely proportional to the volume fraction of clay, though the steepness of the dependence is influenced by the degree of exfoliation and dispersion [39, 41, 46]. Here, we define the critical strain as the strain at which the complex viscosity has decreased to 90% of its linear viscoelastic plateau value. Additionally, the magnitude of the complex viscosity (|[eta]*|) at low strain is sensitive to organoclay loading and dispersion. Therefore, we will focus upon these two quantities in examining oscillatory strain sweep data.

Numerous studies of filled polymeric liquids have cited a flattening, or plateau, of the storage modulus (G') in the terminal regime during SAOS frequency sweeps. Such behavior has been attributed to the formation of a mesoscale clay network, which is sensitive to organoclay loading, extent of exfoliation, and dispersion [37, 42, 43, 47, 48]. Moreover, while (G') tends to increase at low frequencies as organoclay loading increases and dispersion improves, the loss modulus (G") remains largely unaffected. As a result, the loss tangent (tan [delta] [equivalent to] G"/G'), which is another measure of solid-like character in a fluid, decreases in the terminal regime. By contrast, the loss tangent of polymeric liquids grows indefinitely with decreasing frequency. As in the case of oscillatory strain sweeps, the complex viscosity in the terminal regime is a third measurable by which the solid-like nature of the samples may be probed, as it also increases with loading and dispersion. For the SAOS experiments reported here, the strain value was chosen to be 0.01 to ensure a linear response, and a single frequency sweep was conducted at 180[degrees]C from high to low frequency (100-0.01 Hz). While frequency sweeps were conducted at other temperatures and time-temperature superposition (TTS) validated, the TTS curve is not germane to this study, and we report only the results of the sweep performed at 180[degrees]C.

Finally, steady shear rate sweeps were performed from low to high shear rate (0.01-10.0 [s.sup.-1]) at 180[degrees]C. Again, we note that some high shear rate data were discarded as unphysical, probably resulting from edge fracture in the sample. Prior work has shown that organoclay loading and dispersion can cause PCNs to behave as constant shear-thinning materials [37, 39, 43, 49]. That is, they exhibit a monotonically decreasing viscosity even at low shear rates where the matrix polymer would otherwise plateau to the zero-shear viscosity. This divergence of viscosity at low shear rates (defined so because the viscosity asymptotes to infinity in the limit of zero shear rate) is consistent with the presence of a finite yield stress and interpreted as a strong indicator of the hypothesized volume-spanning, mesoscale clay network [37, 38, 47, 49]. The yield stress may be inferred by fitting the low shear rate data to Casson's equation, given by:

[[sigma].sub.xy.sup.1/2] = [[sigma].sub.0.sup.1/2] + a[dot.[gamma].sup.1/2] (1)

where [[sigma].sub.xy] is the shear stress, [[sigma].sub.0] is the yield stress, and the parameter a is an arbitrary constant [39, 43, 50]. Upon imposition of shear, the highly anisotropic organoclay domains tend to orient with the flow direction even at low shear rates, because the characteristic time scale for rotational diffusivity is typically much longer than that associated with the deformation (i.e., the reciprocal of the shear rate). Flow-induced orientation of organoclay domains leads to concurrent disruption of the network and yield in the sample. Moreover, when organoclay domains increasingly orient in the flow direction, their projection of orientation in the velocity gradient direction diminishes, resulting in a monotonically decreasing viscosity. Thus, for highly filled, well-dispersed samples, flow-induced orientation of organoclay domains and divergent shear thinning behavior are expected. On the other hand, samples with low clay loadings and poor exfoliation are expected to exhibit the liquid-like rheology characteristic of the polymer melt with only a modest increase in the zero-shear viscosity resulting from the inclusion of a dilute concentration of solid particles. In summary, the yield stress and steady shear viscosity at low shear rates offer two more experimental markers by which organoclay dispersion may be probed rheologically.

RESULTS

X-Ray Diffraction

The results of XRD scans for all 1, 3, and 5 wt% PCNs are shown in Fig. 1. It is immediately apparent that the non-MB and uncompatibilized samples are characterized by stronger diffraction peaks at all organoclay loadings. While this result does not certify that these materials are less well exfoliated than the others, the increase in the number of diffraction events at a particular angle indicates a greater coherence between silicate layers with the requisite orientation relative to the X-ray beam. Hence, the XRD scans are in accord with expectations that MB processing with compatibilizer should enhance exfoliation in processing.

Table 1 summarizes the organoclay basal spacing values calculated from the angle associated with the peak intensities in Fig. 1. We first note that the uncompatibilized samples possess basal spacings markedly smaller than all compatibilized samples, as expected, as well as the as-received C15A organoclay (3.222 nm). The reduction relative to C15A may result from degradation of surfactant molecules, which is known to occur at temperatures above 200[degrees]C [51, 52]. Since PP is thermodynamically prohibited from diffusing into the intergallery, thermal degradation of surfactant causes a decrease in basal spacing. Additionally, within the uncompatibilized samples, the [d.sub.001] spacing of the sample processed in counterrotation mode for the dilution step is consistently smaller than that processed with corotating screws during dilution. Since the dilution step is the only difference in preparation between these two samples, the shift to smaller basal spacing for the Co-Counter (None) sample highlights the importance of processing history even in the absence of compatibilizer. Whether the difference in basal spacing translates to changes in dispersion and solid-like rheology will be discussed in the subsequent sections.

The [d.sub.001] spacings of the compatibilized samples are similar to that of the as-received C15A clay. While some are slightly larger, others are somewhat smaller with no definitive dependence upon processing history and organoclay loading. Furthermore, the basal spacings of the single-throughput samples are indistinguishable from those of the compatibilized MB blends. Hence, we do not believe that the results for the compatibilized samples are indicative of any meaningful differences in organoclay structure, at least on the length scale probed by XRD. Because the organically modified C15A already has a large [d.sub.001] spacing (relative to the 1.17 nm of natural MMT), additional swelling of the intergallery without exfoliating layers may be difficult to achieve. It should be noted that surfactant degradation will also occur in the compatibilized samples, and differences in basal spacing between compatibilized and uncompatibilized blends highlight the ability of the PP-g-MA compatibilizer to expand the clay intergallery. However, because XRD cannot generally provide information about the number of platelets per tactoid, it is a poor standalone technique for assessing the degree of exfoliation [43, 53]. Consequently, changes in organoclay structure at the mesoscale, which may be inferred from rheology, are more likely to reveal the effect of processing history on exfoliation and dispersion. Therefore, we simply note that the intergallery distances of the compatibilized samples remain larger than those of the uncompatibilized counterparts, suggesting that PP-g-MA successfully intercalated between silicate layers.

[FIGURE 1 OMITTED]

Rheology

As stated previously, rheology has proven to be an effective means of indirectly probing clay exfoliation and microstructure [9, 35-40, 49, 54]. In this section, we will present data and identify qualitative features and trends upon which we will elaborate in the Discussion section to follow. Frequently, we will refer to samples as having more or less solid-like rheology. This designation implies greater microscopic interaction between organoclay domains, which leads to increases in viscosity and storage modulus and decreases in loss tangent. For a given organoclay loading, the increased interactions between organoclay domains that contribute to these rheological signatures can only derive from improved exfoliation and dispersion. Although a detailed analysis of the forthcoming results in the context of exfoliation and dispersion will be deferred to the Discussion section, it is helpful to retain a vision of the underlying source of the rheological features identified below.

We begin with oscillatory strain data for all three organoclay loadings presented in Fig. 2, where we draw attention to the magnitude of the complex viscosity in the linear viscoelastic (low strain amplitude) regime and the critical strain at which the transition to nonlinear rheology occurs. First, in accordance with expectations, we observe that the dependence of the low strain complex viscosity on processing history becomes stronger with organoclay loading. However, more pertinent to the goals of this study are the trends that develop with regard to processing history and sample formulation. At 1 wt%, most PCN samples are not noticeably distinct from the matrix PP, and most observed rheological differences are likely due to changes in polymer composition rather than organoclay exfoliation and dispersion. For instance, the uncompatibilized materials exhibit a higher low strain complex viscosity than all but the compatibilized Counter-Co sample. Since the low strain complex viscosity of the PP/9 wt% PP-g-MA control sample is also less than that of the matrix PP, it is clear that the low molecular weight compatibilizer is more strongly affecting the viscoelastic response than the organoclay. As a result, the low strain complex viscosity of the compatibilized samples tends to be less than that of the uncompatibilized blends. The lone exception is the Counter-Co MB sample, which has the highest low strain complex viscosity and smallest critical strain for shear thinning, pointing to a stronger organoclay influence.

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

For the 3 wt% samples shown in Fig. 2b, the state of the organoclay dispersion plays a prominent role. The four compatibilized MB samples possess low strain complex viscosities above those of the two uncompatibilized samples and appear to shear thin at smaller strains as well. Of the four compatibilized MB samples, Counter-Co again possesses the highest low strain complex viscosity. It is also interesting to note a mixed influence of organoclay on the results for the single-throughput Co (No MB) sample. Despite containing PP-g-MA, its low strain complex viscosity is only slightly less than that of Co-Co (None), suggesting fewer clay-clay interactions and a poorer dispersion, yet its critical strain appears to be smaller indicating the converse. The trends of the 3 wt% samples continue at the 5 wt% organoclay loading. All five compatibilized materials have substantially higher low strain complex viscosities and lower critical strains than the uncompatibilized samples. In this case, Co-Co supplants Counter-Co as the sample with the highest low strain complex viscosity, although it is important to note that all four MB materials exceed the single-throughput compatibilized sample by a wide margin in this regard, emphasizing the efficacy of the MB processing approach.

[FIGURE 4 OMITTED]

In SAOS, a transition from liquid-like (i.e., G' [proportional] [[omega].sup.2] and G" [proportional] [omega]) to solid-like (i.e., G', G" [proportional] [[omega].sup.0]) behavior in the terminal regime has been related to the formation of a solid-like organoclay network [36, 37, 40-43, 47]. In Figs. 3-5, we present data for the storage modulus, loss tangent, and complex viscosity for the 1, 3, and 5 wt% PCNs, respectively. The data at 1 wt% organoclay loading are similar to those for oscillatory strain in the sense that the rheological response of most samples is unaffected by organoclay, displaying a liquid-like, polymeric response that includes the expected slope of 2 for the storage modulus, diverging loss tangent, and plateau in complex viscosity in the terminal regime. Again, the Counter-Co sample is the exception to this description. Its shallower G' slope, flattening of the loss tangent, and modestly diverging complex viscosity are all signatures of solid-like rheology.

[FIGURE 5 OMITTED]

[FIGURE 6 OMITTED]

Shifting to the 3 wt% samples in Fig. 4, the qualitative differences between the compatibilized MB, single-throughput compatibilized, and uncompatibilized samples are manifest. All compatibilized MB materials exhibit significant flattening of G', decrease in the loss tangent, and increased and diverging complex viscosity in the terminal regime relative to uncompatibilized samples. Moreover, the effects are accentuated for the MB samples relative to the single-throughput sample. Differentiating between the MB samples is more challenging, although the results for storage modulus and complex viscosity suggest that the Co-Counter sample is less solid-like than the others. Interestingly, Co-Counter (None) is also less solid-like than Co-Co (None). Finally, similar trends prevail for the 5 wt% blends in Fig. 5. The four compatibilized MB samples, which are almost indistinguishable from one another graphically, are unequivocally more solid-like than the uncompatibilized samples, whereas the single-throughput sample falls in between. Also, the differences in rheology between Co-Counter (None) and Co-Co (None) are even more pronounced, where the latter exhibits much more solid-like behavior in the terminal regime. We will return to this point when discussing the impact of processing sequence.

In Fig. 6, we present the final set of rheological data for steady shear rate sweep experiments of 1, 3, and 5 wt% PCNs. Similar to the SAOS complex viscosity data, we focus upon increases in viscosity at low shear rates as well as the appearance of divergent behavior, both of which are believed to derive from a mesoscale organoclay network [2, 37, 39, 49]. The results for 1 wt% organoclay mirror the earlier observations for complex viscosity in oscillatory strain and SAOS at the same loading. Compatibilizer seems to have a stronger effect than organoclay on the viscosity, except for the Counter-Co material, which again exhibits a higher viscosity. At 3 wt%, only the uncompatibilized samples show the low shear plateau characteristic of unfilled polymeric fluids, although the fact that the zero-shear viscosity of the Co-Co (None) sample exceeds that of Co-Counter (None) again points to a strong processing history influence in the uncompatibilized samples. Alternatively, all compatibilized samples show the divergence in viscosity characteristic of a finite yield stress. Since we are focusing on qualitative features at this time, we will defer discussion of the yield stresses inferred from Casson's equation to the Discussion section. The viscosity of all four MB samples at any given low shear rate exceeds that of the single-throughput compatibilized blend.

DISCUSSION

In the preceding section, a large amount of rheological data was presented from which we may draw the following qualitative conclusions concerning the processing of PP-clay nanocomposites:

* The addition of PP-g-MA as a compatibilizer leads to more solid-like rheology for 3 and 5 wt% organoclay loadings, presumably due to its ability to intercalate between silicate layers and facilitate exfoliation and dispersion. At a 1 wt% loading, the effect of the low molecular weight PP-g-MA on exfoliation and dispersion has a minimal influence on rheology, but it does tend to decrease the small deformation viscosity except in the case of the compatibilized Counter-Co sample.

* Two-stage MB processing, in which organoclay is first melt-blended with PP-g-MA compatibilizer before subsequent dilution with matrix polymer, improves exfoliation and dispersion (as inferred from rheology) relative to single-throughput processing at 3 and 5 wt% loadings. At 1 wt%, the effect is again indeterminate, except for the Counter-Co blend.

While important, these conclusions are not entirely unexpected. To comment specifically upon the more subtle influences of screw rotation mode and sequencing, a more detailed, quantitative analysis of these data is required. However, analyzing data for three types of rheological experiments performed on a total of 23 samples is a potentially daunting task. The challenge lies in identifying key quantities that capture the behavior of greatest interest. We will focus on seven quantities extracted from the three types of experiment that are believed to best reflect the rheological changes engendered by improved exfoliation and dispersion. For oscillatory strain sweeps, we will examine the plateau value of the complex viscosity at small strain amplitude (0.01) and the critical strain marking the boundary between linear and nonlinear viscoelasticity. For SAOS, the storage modulus, loss tangent, and complex viscosity at 0.01 Hz will be included. Lastly, the shear viscosity at low shear rate and the yield stress inferred from Casson's equation will be used. The shear rate at which the viscosity is taken will depend upon the organoclay loading, because the rheometer is able to access lower shear rates as the sample viscosity increases. Hence, the reported values of shear viscosity for the 1, 3, and 5 wt% samples will correspond to those recorded at shear rates of 0.0251, 0.0126, and 0.01 [s.sup.-1], respectively.

In Tables 2-4, we present the quantitative values of the seven rheological parameters for the three organoclay loadings. Table 2 also contains similar data for the processed PP and PP/9 wt% PP-g-MA control samples. However, while quantitative measures of organoclay exfoliation and dispersion are now summarized in tabular form, a strategy for averaging the results for a given processing history and sample composition is necessary to use rheology to holistically assess the influence of those variables. Therefore, we have taken the results in Tables 2-4 and normalized them to a unit scale, where zero represents the most liquid- or polymer-like response and unity the most solid-like, well exfoliated, and dispersed result. For example, consider oscillatory strain sweep experiments at fixed organoclay loading, where an increase in complex viscosity is indicative of increased clay-clay interactions and improved exfoliation and dispersion. Hence, its normalized value is calculated as (|[eta]*| - |[eta]*|[.sub.min])/(|[eta]*|[.sub.max] - |[eta]*|[.sub.min]), where |[eta]*|[.sub.max] and |[eta]*|[.sub.min] represent the maximum and minimum values of all the samples at that organoclay loading. All quantities for which a large value of the parameter represents improved exfoliation and dispersion are computed in this fashion. For the critical strain and loss tangent, smaller values reflect better exfoliation and dispersion; consequently, the normalized critical strain, for instance, is calculated as ([[gamma].sub.c,max] - [[gamma].sub.c])/([[gamma].sub.c,max] - [[gamma].sub.c,min]). Each organoclay loading is normalized and evaluated independently of the others; however, because the processed PP and PP/9 wt% PP-g-MA represent control materials, they are included in the normalizations at each organoclay loading.

These normalized parameters are reported in Tables 5-7 along with the ordinal ranking for that parameter. Additionally, an average score is reported for each blend, where equal weight is given to each type of experiment (not each rheological parameter) to address the imbalance in the number of parameters extracted from each experiment. The expectation is that the average score will provide a simple and effective quantitative assessment of the effect of both compatibilization and processing history on organoclay exfoliation and dispersion. Beginning with the results for the 1 wt% organoclay loading, we note quantitative support for the observation that the compatibilized Counter-Co sample was more solid-like than the other 1 wt% materials. At first glance, it is surprising that the uncompatibilized samples ranked second and fourth; however, we recall that the rheology at this organoclay loading is dominated by polymer rather than clay-clay interactions. Hence, the low molecular weight PP-g-MA significantly reduces the viscosity of compatibilized blends, invalidating a holistic analysis predicated on clay-clay interactions having greater influence on rheology. The lone rheological signature that is not affected by PP-g-MA is the yield stress. Interestingly, if the results were sorted based solely on the ordinal values of yield stress, the four compatibilized MB materials would occupy rankings one through four, followed by the single-throughput compatibilized blend and finally the two uncompatibilized samples. Within the compatibilized samples, Counter-Co has the highest yield stress, followed distantly by Counter-Counter, Co-Co, Co-Counter, and Co (No MB). Although the yield stresses at this low loading are relatively small and the fitting to Casson's equation somewhat subjective, we will recall this ranking based on yield stress when considering holistic rankings at higher loadings. In summary, while the holistic rheological analysis of the 1 wt% materials has been shown to be inadequate when clay-clay interactions do not dominate rheological behavior, rankings based upon yield stress produce results consistent with expectations.

Turning next to the 3 wt% samples whose normalized values are shown in Table 6, we see that the holistic rheological analysis is more robust at differentiating between not only compatibilized and uncompatibilized samples but also screw rotation mode and sequence. First, the uncompatibilized blends are less solid-like than the compatibilized blends based on average score and nearly all individual parameters, reinforcing the premise that PP-g-MA facilitates exfoliation and dispersion of clay silicates. The four MB samples similarly distinguish themselves relative to the Co (No MB) sample. As a result, it appears that MB processing generates better exfoliated and dispersed materials, likely due to the prolonged direct mixing between PP-g-MA and organoclay before PP is introduced. Examining the individual and average scores of the four compatibilized MB blends, we observe stronger dependence on processing history than at the 1 wt% loading. Intriguingly, however, the rankings of their average scores match those reported above for 1 wt% based solely on yield stress. That is, Counter-Co proves to be the most solid-like, followed by Counter-Counter, Co-Co, and Co-Counter at discrete intervals. Put another way, the counterrotation processing of PP-g-MA and organoclay in the MB step leads to better exfoliation and/or dispersion in the final product. Counterrotating screws are known to impart significantly higher stresses over longer extruder residence times relative to corotating screws (~75% longer in this case). It is possible that the longer residence time in counterrotation mode allows the compatibilizer more time to diffuse into clay intergalleries, while the higher stresses facilitate the exfoliation of clay silicates. The manner in which material is processed in the subsequent dilution step also appears to be an important consideration. Here, we compare the results for fixed MB processing conditions but varying dilution conditions (e.g., Counter-Co vs. Counter-Counter, and Co-Co vs. Co-Counter). In both cases, corotating screws in the dilution step yield more solid-like materials than counterrotating screws, a result that is also reflected in the uncompatibilized samples. The explanation for this behavior may lie in the thermomechanical degradation of PP during extrusion, which is known to be accelerated at higher stresses, longer residence times, and in the presence of filler [55-58]. The result is [beta]-scission, a reduction in the average molecular weight of the matrix PP, and a corresponding decrease in viscosity. Since the zero-shear viscosity of the matrix PP used in this study decreased by nearly a factor of two after extrusion with corotating screws (data not reported), thermomechanical degradation is important during the processing of these blends. The stresses experienced by compatibilized MB samples in the dilution step should be larger due to the presence of organoclay, which increases the stress imparted by the screws by virtue of higher sample viscosity. For identical MB processing conditions, both corotating and counterrotating screws will induce thermomechanical degradation, but the higher stresses of the counterrotating mode undoubtedly exacerbate the effect. A more pronounced decrease in matrix viscosity during the counterrotating dilution step would reduce the shear stress transfer between the polymer melt and clay tactoids, leading to less effective exfoliation and dispersion. As a result, materials processed in corotating mode for the dilution step tend to have more solid-like rheological properties. Finally, we noted a similar pattern in the uncompatibilized Co-Co (None) and Co-Counter (None) blends, where the former exhibited more solid-like rheology than the latter at all loadings in all experiments. The most significant reduction in basal spacing was also observed for the Co-Counter (None) blend. We suspect that the combination of smaller separation between silicate layers and greater thermomechanical degradation both contribute to the less solid-like rheology of the Co-Counter (None) sample, although their relative importance is likely dependent upon organoclay loading.

Finally, we examine the normalized rheological data for the 5 wt% blends summarized in Table 7. While the average scores of the compatibilized MB, uncompatibilized, and single-throughput compatibilized materials remain quite distinct from one another, we immediately note much less differentiation between the four MB samples than at 3 wt%. Both the average and individual parameter scores for the MB blends are compressed into a narrow range of values, and the ordinal rankings are inconsistent across parameters. Hence, although the Co-Counter sample (heretofore the least solid-like of the four) unexpectedly scores the highest and Counter-Counter the lowest of the four, the individual parameters (see Table 4) are often not meaningfully different from one another. For example, it is difficult to argue that the reported values of critical strain and loss tangent are appreciably different from one another (a fact that applies to a lesser degree for the 3 wt% samples as well), which renders parameters with greater variation more influential in computing the average. However, the inability of the holistic rheological analysis to resolve microstructural differences in the four compatibilized samples at 5 wt% organoclay loading may itself represent an important result. At high organoclay loadings, screw rotation mode and sequence (and thereby stress and residence time in the extruder) may not significantly impact the degree of exfoliation and dispersion of organoclay. In this more highly percolated state, clay-clay interactions are the dominant rheological influence, as evidenced by the over 20-fold increase in shear viscosity relative to the processed PP. However, subtle changes in the degree of exfoliation may have only a secondary effect on rheological properties, resulting in the capricious parameter rankings in Table 7. Nonetheless, it is apparent that the compatibilized MB processing yields a more solid-like product than the compatibilized single-throughput and uncompatibilized blends, serving notice that introduction of compatibilizer in the MB processing step has a profound impact on organoclay exfoliation and dispersion.

CONCLUSIONS

PP-clay nanocomposites with 1, 3, and 5 wt% organoclay loadings were studied with the goal of elucidating the importance of MB processing, screw rotation mode, and sequence. XRD data demonstrate that uncompatibilized samples have smaller basal spacings than compatibilized blends, but the technique cannot differentiate the latter on the basis of processing history. On the other hand, a holistic rheological analysis of oscillatory strain, frequency, and steady shear rate sweep experiments has been shown to be sensitive to differences in organoclay exfoliation and dispersion caused by variation in processing history. Results not only support the well-known premise that PP-g-MA compatibilizer dramatically improves exfoliation and dispersion of clay silicates but also demonstrate that MB processing of compatibilizer and organoclay followed by dilution with matrix PP has a similarly efficacious effect. Moreover, at the 1 and 3 wt% organoclay loadings, the higher stress and longer residence time characteristic of the predominantly dispersive mixing counterrotation screw mode appear to improve exfoliation in the MB step, whereas the greater distributive mixing of a corotating screw configuration in the dilution step may reduce thermomechanical degradation relative to counterrotating screws. At the higher 5 wt% loading, the sequence of processing is less important, suggesting that the sheer volume of organoclay renders subtle changes in the degree of exfoliation and dispersion less consequential to the rheology.

ACKNOWLEDGMENTS

We thank Dr. Robert Sammler of The Dow Chemical Company for the generous donation of the H700-12 polypropylene resin, Dr. Leigh Allen of Crompton Corporation for the Polybond[R] 3200 PP-g-MA compatibilizer, and Dr. Douglas Hunter of Southern Clay Products for the Cloisite[R] 15A montmorillonite clay.

REFERENCES

1. R.A. Vaia, K.D. Jandt, E.J. Kramer, and E.P. Giannelis, Macromolecules, 28, 8080 (1995).

2. R. Krishnamoorti, R.A. Vaia, and E.P. Giannelis, Chem. Mater., 8, 1728 (1996).

3. E.P. Giannelis, R. Krishnamoorti, and E. Manias, Adv. Polym. Sci., 138, 107 (1999).

4. S.S. Ray and M. Okamoto, Prog. Polym. Sci., 28, 1539 (2003).

5. D.L. Ho, R.M. Briber, and C.J. Glinka, Chem. Mater., 13, 1923 (2001).

6. M. Alexandre and P. Dubois, Mater. Sci. Eng., 28, 1 (2000).

7. N. Hasegawa, M. Kawasumi, M. Kato, A. Usuki, and A. Okada, J. Appl. Polym. Sci., 67, 87 (1998).

8. D. Marchant and K. Jayaraman, Ind. Eng. Chem. Res., 41, 6402 (2002).

9. Y.T. Lim and O.O. Park, Rheol. Acta, 40, 220 (2001).

10. D.F. Eckel, M.P. Balogh, P.D. Fasulo, and W.R. Rodgers, J. Appl. Polym. Sci., 93, 1110 (2004).

11. F.L. Beyer, N.C.B. Tan, A. Dasgupta, and M.E. Galvin, Chem. Mater., 14, 2983 (2002).

12. T.D. Fornes, P.J. Yoon, D.L. Hunter, H. Keskkula, and D.R. Paul, Polymer, 43, 5915 (2002).

13. M.Y. Gelfer, H.H. Song, L.Z. Liu, B.S. Hsiao, B. Chu, M. Rafailovich, M.Y. Si, and V. Zaitsev, J. Polym. Sci. Part B: Polym. Phys., 41, 44 (2003).

14. K.M. Lee and C.D. Han, Macromolecules, 36, 7165 (2003).

15. M. Kurian, A. Dasgupta, F.L. Beyer, and M.E. Galvin, J. Polym. Sci. Part B: Polym. Phys., 42, 4075 (2004).

16. B.N. Jang, D. Wang, and C.A. Wilkie, Macromolecules, 38, 6533 (2005).

17. Y. Wang, F.B. Chen, Y.C. Li, and K.C. Wu, Compos. B, 35, 111 (2004).

18. Y. Wang, F.B. Chen, and K.C. Wu, J. Appl. Polym. Sci., 93, 100 (2004).

19. C.H. Hong, Y.B. Lee, J.W. Bae, J.Y. Jho, B.U. Nam, and T.W. Hwang, J. Appl. Polym. Sci., 98, 427 (2005).

20. J. Li, M.T. Ton-That, and S.J. Tsai, Polym. Eng. Sci., 46, 1060 (2006).

21. U.N. Ratnayake and B. Haworth, Polym. Eng. Sci., 46, 1008 (2006).

22. H.R. Dennis, D.L. Hunter, D. Chang, S. Kim, J.L. White, J.W. Cho, and D.R. Paul, Polymer, 42, 9513 (2001).

23. K. Wang, S. Liang, R.N. Du, Q. Zhang, and Q. Fu, Polymer, 45, 7953 (2004).

24. Y. Wang, F.B. Chen, K.C. Wu, and J.C. Wang, Polym. Eng. Sci., 46, 289 (2006).

25. R.K. Shah and D.R. Paul, Polymer, 45, 2991 (2004).

26. H.S. Lee, P.D. Fasulo, W.R. Rodgers, and D.R. Paul, Polymer, 46, 11673 (2005).

27. W. Lertwilmolnun and B. Vergnes, Polymer, 46, 3462 (2005).

28. P. Peltola, E. Valipakka, J. Vuorinen, S. Syrjala, and K. Hanhi, Polym. Eng. Sci., 46, 995 (2006).

29. M. Modesti, A. Lorenzetti, D. Bon, and S. Besco, Polymer, 46, 10237 (2005).

30. R. Wagener and T.J.G. Reisinger, Polymer, 44, 7513 (2003).

31. V.M. Lobe and J.L. White, Polym. Eng. Sci., 19, 617 (1979).

32. Y. Suetsugu and J.L. White, J. Rheol., 26, 581 (1982).

33. Y. Suetsugu and J.L. White, J. Appl. Polym. Sci., 28, 1481 (1983).

34. C.H. Suh and J.L. White, J. Non-Newtonian Fluid Mech., 62, 175 (1996).

35. R. Krishnamoorti and E.P. Giannelis, Macromolecules, 30, 4097 (1997).

36. G. Galgali, C. Ramesh, and A. Lele, Macromolecules, 34, 852 (2001).

37. M.J. Solomon, A.S. Almusallam, K.F. Seefeldt, A. Somwangthanaroj, and P. Varadan, Macromolecules, 34, 1864 (2001).

38. A. Lele, M. Mackley, G. Galgali, and C. Ramesh, J. Rheol., 46, 1091 (2002).

39. J.X. Ren and R. Krishnamoorti, Macromolecules, 36, 4443 (2003).

40. J.X. Ren, B.F. Casanueva, C.A. Mitchell, and R. Krishnamoorti, Macromolecules, 36, 4188 (2003).

41. J. Li, C.X. Zhou, G. Wang, and D.L. Zhao, J. Appl. Polym. Sci., 89, 3609 (2003).

42. S.Y. Gu, J. Ren, and Q.F. Wang, J. Appl. Polym. Sci., 91, 2427 (2004).

43. M.A. Treece and J.P. Oberhauser, J. Appl. Polym. Sci., 103, 884 (2007).

44. "Guidelines for installation, operation and maintenance of Leistritz twin-screw extruders," American Leistritz Extruder Corporation, Somerville, NJ.

45. "Operation manual for twin-screw extruder: Micro 27," American Leistritz Extruder Corporation, Somerville, NJ.

46. T. Aubry, T. Razafinimaro, and P. Mederic, J. Rheol., 49, 425 (2005).

47. R. Krishnamoorti and K. Yurekli, Curr. Opin. Colloid Interface Sci., 6, 464 (2001).

48. J. Zhao, A.B. Morgan, and J.D. Harris, Polymer, 46, 8641 (2005).

49. R. Krishnamoorti, J. Ren, and A.S. Silva, J. Chem. Phys., 114, 4968 (2001).

50. N. Casson, Rheol Disperse Systems, University College of Swansea, Swansea (1959).

51. W. Xie, Z. Gao, W.-P. Pan, D. Hunter, A. Singh, and R. Vaia, Chem. Mater., 13, 2979 (2001).

52. M. Gelfer, C. Burger, A. Fadeev, I. Sics, B. Chu, B.S. Hsiao, A. Heintz, K. Kojo, S.L. Hsu, M. Si, and A. Rafailovich, Langmuir, 20, 3746 (2004).

53. M.A. Treece and J.P. Oberhauser, Polymer, 48, 1083 (2007).

54. J.X. Ren, A.S. Silva, and R. Krishnamoorti, Macromolecules, 33, 3739 (2000).

55. F.P. La Mantia and N.T. Dintcheva, Macromol. Symp., 194, 277 (2003).

56. A.V. Machado, J.M. Maia, S.V. Canevarolo, and J.A. Covas, J. Appl. Polym. Sci., 91, 2711 (2004).

57. F. Berzin, B. Vergnes, S.V. Canevarolo, A.V. Machado, and J.A. Covas. J. Appl. Polym. Sci., 99, 2082 (2006).

58. J.-C. Huang and J. Xu, Int. J. Polym. Mater., 52, 203 (2003).

Mark A. Treece, (1) Wei Zhang, (2) Ronald D. Moffitt, (2) James P. Oberhauser (1)

(1) Department of Chemical Engineering, University of Virginia, Charlottesville, Virginia 22904

(2) Advanced and Applied Polymer Processing Institute, 300 Ringgold Industrial Parkway, Danville, Virginia 24540

Correspondence to: James P. Oberhauser; e-mail: [e-mail]

Contract grant sponsor: National Science Foundation; contract grant number: NSF CAREER Award CTS-0134275.

TABLE 1. Organoclay basal spacings of all PCN samples inferred from the
XRD scans shown in Figure 1.

                   [d.sub.001] spacing (nm)
Process            1 wt%  3 wt%  5 wt%

Co-Co              3.222  3.475  3.044
Co-Counter         3.198  3.225  3.269
Counter-Co         3.245  3.319  3.044
Counter-Counter    3.153  3.153  3.175
Co-Co (None)       3.023  3.003  2.982
Co-Counter (None)  2.848  2.829  2.904
Co (No MB)         3.175  3.245  3.294

The [d.sub.001] spacing of as-received C15A is measured at 3.222 nm.

TABLE 2. Results for the seven characteristic rheological parameters for
the 1 wt% PCNs and control materials.

                   Oscillatory strain sweep   Frequency sweep
                   |[eta]*|  [[gamma].sub.c]  G'     tan      |[eta]*|
Process            (Pa-s)                     (Pa)   [delta]  (Pa-s)

Counter-Co         2024      0.662            39.0    4.86    3080
Co-Co (None)       1939      1.794            20.3    8.39    2729
Co-Co              1814      1.923            17.0    9.22    2505
Co-Counter (None)  1818      1.929            14.6   10.9     2534
Co (No MB)         1721      2.055            13.1   11.5     2401
Counter-Counter    1599      1.776            14.1    8.81    1995
Co-Counter         1503      2.196             9.68  11.3     1751
PP                 1729      2.687            10.2   13.4     2181
PP/9 wt% PP-g-MA   1394      2.745             5.17  18.4     1517

                   Steady shear rate sweep
Process            [[sigma].sub.0] (Pa)  [eta] (Pa-s)

Counter-Co         10.8                  3433
Co-Co (None)        0.012                2866
Co-Co               0.417                2751
Co-Counter (None)   0.017                2648
Co (No MB)          0.060                2526
Counter-Counter     0.730                2266
Co-Counter          0.307                1949
PP                  0.005                2332
PP/9 wt% PP-g-MA    0.0008               1644

Results for rheological parameters extracted from oscillatory strain
sweep, frequency sweep, and steady shear rate sweep experiments
performed at 180[degrees]C on the 1 wt% PCNs, processed PP, and PP/9 wt%
PP-g-MA samples. All frequency sweep parameter values were taken at 0.01
Hz. The steady shear viscosity was that recorded at 0.0251 [s.sup.-1].

TABLE 3. Results for the seven characteristic rheological parameters for
the 3 wt% PCNs.

                     Oscillatory strain sweep
Process            |[eta]*| (Pa-s)  [[gamma].sub.c]

Counter-Co         3296             0.140
Counter-Counter    2941             0.158
Co-Co              3067             0.194
Co-Counter         2689             0.193
Co (No MB)         2380             0.301
Co-Co (None)       2452             0.567
Co-Counter (None)  1921             1.194

                           Frequency sweep
Process            G' (Pa)  tan [delta]  |[eta]*| (Pa-s)

Counter-Co         942      0.55         17,130
Counter-Counter    853      0.53         15,360
Co-Co              719      0.70         13,940
Co-Counter         455      0.71          8,882
Co (No MB)         201      1.38          5,466
Co-Co (None)        98.7    2.92          4,854
Co-Counter (None)   70.6    3.01          3,567

                   Steady shear rate sweep
Process            [[sigma].sub.0] (Pa)  [eta] (Pa-s)

Counter-Co         230                   23,680
Counter-Counter    208                   21,730
Co-Co              159                   19,160
Co-Counter         151                   17,070
Co (No MB)          78.0                 10,870
Co-Co (None)        17.7                  5,412
Co-Counter (None)    1.82                 3,122

Results for rheological parameters extracted from oscillatory strain
sweep, frequency sweep, and steady shear rate sweep experiments
performed at 180[degrees]C on the 3 wt% PCNs. All frequency sweep
parameter values were taken at 0.01 Hz. The steady shear viscosity was
that recorded at 0.0126 [s.sup.-1].

TABLE 4. Results for the seven characteristic rheological parameters for
the 5 wt% PCNs.

                      Oscillatory strain sweep
Process            |[eta]*| (Pa-s)  [[gamma].sub.c]

Co-Counter         5860             0.064
Counter-Co         5753             0.058
Co-Co              6542             0.047
Counter-Counter    5338             0.087
Co (No MB)         4143             0.091
Co-Co (None)       3501             0.108
Co-Counter (None)  2263             0.603

                   Frequency sweep
Process            G' (Pa)  tan [delta]  |[eta]*| (Pa-s)

Co-Counter         3936     0.25         64,540
Counter-Co         3821     0.27         63,010
Co-Co              3677     0.28         60,750
Counter-Counter    3853     0.24         63,070
Co (No MB)         1417     0.44         24,600
Co-Co (None)        591     0.90         12,670
Co-Counter (None)    87.3   2.83          4,172

                   Steady shear rate sweep
Process            [[sigma].sub.0] (Pa)  [eta] (Pa-s)

Co-Counter         584                   60,170
Counter-Co         562                   59,220
Co-Co              482                   51,420
Counter-Counter    519                   53,040
Co (No MB)         319                   35,760
Co-Co (None)        64.5                 10,940
Co-Counter (None)   11.7                  4,464

Results for rheological parameters extracted from oscillatory strain
sweep, frequency sweep, and steady shear rate sweep experiments
performed at 180[degrees]C on the 5 wt% PCNs. All frequency sweep
parameter values were taken at 0.01 Hz. The steady shear viscosity was
that recorded at 0.01 [s.sup.-1].

TABLE 5. Results for the 1 wt% PCNs, processed PP, and PP/9 wt% PP-g-MA
samples.

                   Oscillatory strain sweep
Process            |[eta]*|   [[gamma].sub.c]

Counter-Co         1.000 (1)  1.000 (1)
Co-Co (None)       0.865 (2)  0.457 (3)
Co-Co              0.667 (4)  0.395 (4)
Co-Counter (None)  0.673 (3)  0.392 (5)
Co (No MB)         0.519 (5)  0.331 (6)
Counter-Counter    0.325 (6)  0.465 (2)
Co-Counter         0.173 (7)  0.264 (7)
PP                 0.532      0.028
PP/9 wt% PP-g-MA   0.000      0.000

                   Frequency sweep
Process            G'         tan [delta]  |[eta]*|

Counter-Co         1.000 (1)  1.000 (1)    1.000 (1)
Co-Co (None)       0.448 (2)  0.740 (2)    0.775 (2)
Co-Co              0.349 (3)  0.679 (4)    0.632 (4)
Co-Counter (None)  0.278 (4)  0.556 (5)    0.651 (3)
Co (No MB)         0.235 (6)  0.514 (7)    0.566 (5)
Counter-Counter    0.265 (5)  0.709 (3)    0.306 (6)
Co-Counter         0.133 (7)  0.524 (6)    0.150 (7)
PP                 0.148      0.368        0.425
PP/9 wt% PP-g-MA   0.000      0.000        0.000

                   Steady shear rate sweep
Process            [[sigma].sub.0]  [eta]      Avg.

Counter-Co         1.000 (1)        1.000 (1)  1.000
Co-Co (None)       0.001 (7)        0.683 (2)  0.552
Co-Co              0.039 (3)        0.619 (3)  0.471
Co-Counter (None)  0.002 (6)        0.561 (4)  0.436
Co (No MB)         0.005 (5)        0.493 (5)  0.371
Counter-Counter    0.068 (2)        0.348 (6)  0.343
Co-Counter         0.028 (4)        0.170 (7)  0.196
PP                 0.000            0.385      0.262
PP/9 wt% PP-g-MA   0.000            0.000      0.000

Results from Table 2 for the 1 wt% PCNs, processed PP, and PP/9 wt%
PP-g-MA samples, where each parameter has been normalized to a unit
scale, where zero corresponds to the least solid-like response and unity
to the most solid-like. The ordinal rank relative to all 1 wt% results
is shown in parentheses. All frequency sweep parameter values were taken
at 0.01 Hz. The steady shear viscosity was that recorded at 0.0251
[s.sup.-1].

TABLE 6. Results for the 3 wt% PCNs, processed PP, and PP/9 wt% PP-g-MA
samples.

                   Oscillatory strain sweep
Process            |[eta]*|   [[gamma].sub.c]

Counter-Co         1.000 (1)  1.000 (1)
Counter-Counter    0.813 (3)  0.993 (2)
Co-Co              0.880 (2)  0.979 (4)
Co-Counter         0.681 (4)  0.980 (3)
Co (No MB)         0.518 (6)  0.938 (5)
Co-Co (None)       0.556 (5)  0.836 (6)
Co-Counter (None)  0.277 (7)  0.595 (7)
PP                 0.176      0.022
PP/9 wt% PP-g-MA   0.000      0.000

                           Frequency sweep
Process            G'         tan [delta]  |[eta]*|

Counter-Co         1.000 (1)  0.999 (2)    1.000 (1)
Counter-Counter    0.904 (2)  1.000 (1)    0.887 (2)
Co-Co              0.762 (3)  0.991 (3)    0.796 (3)
Co-Counter         0.480 (4)  0.990 (4)    0.472 (4)
Co (No MB)         0.209 (5)  0.952 (5)    0.253 (5)
Co-Co (None)       0.100 (6)  0.866 (6)    0.214 (6)
Co-Counter (None)  0.070 (7)  0.861 (7)    0.131 (7)
PP                 0.005      0.279        0.043
PP/9 wt% PP-g-MA   0.000      0.000        0.000

                       Steady shear rate sweep
Process            [[sigma].sub.0]  [eta]      Avg.

Counter-Co         1.000 (1)        1.000 (1)  1.000
Counter-Counter    0.904 (2)        0.912 (2)  0.914
Co-Co              0.693 (3)        0.795 (3)  0.841
Co-Counter         0.658 (4)        0.700 (4)  0.719
Co (No MB)         0.340 (5)        0.419 (5)  0.526
Co-Co (None)       0.077 (6)        0.171 (6)  0.404
Co-Counter (None)  0.008 (7)        0.067 (7)  0.276
PP                 0.000            0.031      0.075
PP/9 wt% PP-g-MA   0.000            0.000      0.000

Results from Table 3 for the 3 wt% PCNs, processed PP, and PP/9 wt% PP-
g-MA samples, where each parameter has been normalized to a unit scale,
where zero corresponds to the least solid-like response and unity to the
most solid-like. The ordinal rank relative to all 3 wt% results is shown
in parentheses. All frequency sweep parameter values were taken at 0.01
Hz. The steady shear viscosity was that recorded at 0.0126 [s.sup.-1].

TABLE 7. Results for the 5 wt% PCNs, processed PP, and PP/9 wt% PP-g-MA
samples.

                   Oscillatory strain sweep
Process            |[eta]*|   [[gamma].sub.c]

Co-Counter         0.868 (3)  0.994 (3)
Counter-Co         0.847 (2)  0.996 (2)
Co-Co              1.000 (1)  1.000 (1)
Counter-Counter    0.766 (4)  0.985 (4)
Co (No MB)         0.534 (5)  0.984 (5)
Co-Co (None)       0.409 (6)  0.977 (6)
Co-Counter (None)  0.169 (7)  0.794 (7)
PP                 0.065      0.021
PP/9 wt% PP-g-MA   0.000      0.000

                           Frequency sweep
Process            G'         tan [delta]  |[eta]*|

Co-Counter         1.000 (1)  1.000 (2)    1.000 (1)
Counter-Co         0.971 (3)  0.998 (3)    0.976 (3)
Co-Co              0.934 (4)  0.998 (4)    0.940 (4)
Counter-Counter    0.979 (2)  1.000 (1)    0.977 (2)
Co (No MB)         0.359 (5)  0.989 (5)    0.366 (5)
Co-Co (None)       0.149 (6)  0.964 (6)    0.177 (6)
Co-Counter (None)  0.021 (7)  0.858 (7)    0.042 (7)
PP                 0.001      0.274        0.011
PP/9 wt% PP-g-MA   0.000      0.000        0.000

                   Steady shear rate sweep
Process            [[sigma].sub.0]  [eta]      Avg.

Co-Counter         1.000 (1)        1.000 (1)  0.977
Counter-Co         0.962 (2)        0.984 (2)  0.959
Co-Co              0.826 (4)        0.850 (4)  0.932
Counter-Counter    0.890 (3)        0.878 (3)  0.915
Co (No MB)         0.547 (5)        0.583 (5)  0.632
Co-Co (None)       0.111 (6)        0.159 (6)  0.419
Co-Counter (None)  0.020 (7)        0.048 (7)  0.274
PP                 0.000            0.012      0.048
PP/9 wt% PP-g-MA   0.000            0.000      0.000

Results from Table 4 for the 5 wt% PCNs, processed PP, and PP/9 wt% PP-
g-MA samples, where each parameter has been normalized to a unit scale,
where zero corresponds to the least solid-like response and unity to the
most solid-like. The ordinal rank relative to all 5 wt% results is shown
in parentheses. All frequency sweep parameter values were taken at 0.01
Hz. The steady shear viscosity was that recorded at 0.01 [s.sup.-1].