Recently, nanocomposite technology has been explored as a means for tailoring the mechanical, barrier, and thermal properties of foams. By nature of their smaller size, nanoparticles effectively alter the nucleation of gas during depressurization of thermoplastic melts. As a consequence, foams of varying density and cell structure can be obtained. Increased thermal stability and barrier properties demonstrated in solid nanocomposites should also transfer to foam. Furthermore, nanoparticles as reinforcements to the foam structure become more effective than larger particles because their smaller size has less of an effect on the very thin walled cells of the foam. For these reasons, nano materials offer a potential method for tailoring performance properties of foams in conventional processing methods. Interfacial Solutions has explored in-line compounding and extrusion processes for foaming polyolefin nanocomposites using supercritical carbon dioxide (CO
2
) as a physical blowing agent. In this paper, results from experiments that investigate the type of nanofiller, coupling agents, and the balance between process variables will be discussed. Results demonstrate that the combination of these factors produce foams of a widely varying quality. irradiation or curing processes from the production of extruded polyolefin foams.
To explore this idea, Interfacial Solutions set up a small, twin-screw extrusion line to run supercritical CO
2 as a physical foaming agent. The performance benefits of supercritical CO
2
as a blowing agent has been described by several authors.[1,2] In particular, supercritical CO
2 has been shown to assist in the exfoliation and dispersion of nanoclays into polymer composites.[3,4] This paper describes our early efforts in developing flexible nanocomposite foams based on ethylene-octene copolymer (EOC). Nanocomposites produced from EOC resins by solvent blending methods have shown significant improvement in modulus, strength, and elongation at break, all potentially interesting improvements for foam applications.[6] Various nanofiller materials were explored in EOC foams for their nucleation effects and for their nanocomposite structure.
In the flexible foams market, polyolefins such as polyethylene (PE) and ethylene vinyl acetate (EVA) are typically cross-linked to impart improved properties of strength, resilience to compression, and chemical resistance. For compression-molded foams, batch processing allows for simultaneous foam formation and chemical cross-linking of the polymer within the cavity of the mold. However for extrusion foams, a secondary process is typically required to complete cross-linking of the polymer. After extrusion of the foam into its final shape, foams can be cross-linked by electron irradiation, or chemically cross-linked in an oven or sauna. Although used extensively in industry, these processes add cost to manufacturing extruded foams. The recent findings in nanocomposite technology suggest that nanocomposites could be used to impart substantially enhanced properties to foam. The very small size of nanofillers suggests opportunities for mechanical enhancement of the thin cell walls of foams without disrupting cell structure. If a nanocomposite could impart sufficient strength and resilience, it may be possible to completely remove
An ethylene-octene copolymer (EOC) (Engage 8003,
Dow Chemical) was used in this study to produce flexible polyolefin foams. Two polymeric coupling agents were additionally investigated, a maleated polyethylene
(MAPE) produced by Interfacial Solutions (XP433) and an ethylene acrylate terpolymer of ethylene, n-butyl acrylate, and glycidyl methacrylate monomers (Elvaloy
PTW, DuPont).
Several types of nucleating agents were explored, including both nano and micoparticles. Nucleating agents included talc (Silverline 303 Talc, Luzenac), montmorillonite nanoclays (Nanofil 116 and Cloisite 15A,
20A, and 30B, Southern Clay Products), a hydrotalcite nanoclay (Perkalite FR100, AkzoNobel), a halloysite nanoclay (Dragonite XR, Applied Minerals Inc.), a micronized volcanic ash, and thermally expanded mirospheres produced from micronized volcanic ash. Of the montmorillonite clays, Nanofil 116 contains no surfactant, but the three grades of Cloisite clays contain various types of quaternary ammonium salts as surfactants to aid in exfoliation and intercalation. The montmorillonite and hydrotalcite clays exhibit platelet like morphology, whereas halloysite clay was a tubular structure. The micronized volcanic ash is produced by grinding a natural volcanic ash (Kansas Minerals), and is

characterized by small, irregular, jagged pieces of a silicate glass less than 2
 m in diameter. By processing the micronized ash through a thermal expansion process identical to the production of perlite, the irregular glass particles expand to form hollow microspheres with an average diameter of about 5
 m.
For all foams produced in this work, a Micro 27 lab extruder (American Leistritz Extruder Company) was used in combination with supercritical CO
2
as the physical blowing agent. The parallel twin-screw extruder was operated in co-rotation mode, and the length to diameter ratio (L/D) was 32. CO
2
was supplied to the extruder from a positive displacement pump designed for supercritical CO
2
(Thar SFC, Waters Corporation). The pressure of CO
2
was controlled by a needle valve at the site of injection, keeping the pressure above 1100 psi to maintain CO
2
in its supercritical state. A special screw configuration was designed to produce a melt seal upstream of the supercritical fluid injection site. Proper design of the melt seal is critical for maintaining pressurization inside of the extruder barrel, otherwise gas will escape from the throat. In addition, the barrel configuration was changed so that each section was solid, without a vent port or insert. A hand-held CO
2
detector was used to verify that there were no leaks during operation. A simple, three-strand die with 3mm die openings was used for all experiments.
Polymer pellets and nucleating agent were blended in a bag in small batches with varying concentrations of nucleating agent. A single screw, volumetric feeder was used to feed the blended material into the throat of the extruder. Feed rate was set manually to the minimum required mass rate that would fill the screws in the melt seal section of the extruder (approximately 5 lbs/hr). By filling the melt seal section with melted polymer, blowing agent was contained inside of the pressurized extruder. A minimum mass rate was used to facilitate cooling of the melt during the process and to minimize torque.
The process was started initially at high temperature,
180 o C, across all barrel sections. Blowing agent is initially introduced into the barrel at startup at bottle pressure (approx. 750psi) to inhibit blockage of the injection port. The screw speed was set between 35 and
45 rpm. As the extruder fills with polymer and mixes with the bowing agent, the temperatures in various barrel sections were reduced, and the pressure of CO
2
is increased above its critical point (>1100psi). The initial extrudate is foamed, but at the high temperature, lacks sufficient melt strength to produce a low-density foam.
Gas exits the extrudate quickly. Temperatures are reduced until the extrudate develops sufficient melt strength to stabilize a low-density foam. It was noticed that the elastic nature of the Engage resin lead to minor shrinkage of the foam after cooling, causing a wrinkled surface texture. After cell stabilization several days later, the surface did not recover a smooth, taunt surface.
A combination of EOC resin and talc was used to determine the best processing conditions producing a foam with less than 0.1 specific gravity, and these conditions were attempted to be replicated for all nucleating agents and concentrations explored. Process conditions were varied as required to best match a specific gravity of less than 0.1.
The specific gravity of each foam was determined by a water displacement method. Since the foams were always less than a specific gravity of 1, a weighted pouch was used to submerge the samples in water. To evaluate the nanocomposite structure, x-ray diffraction (XRD) was used (Miniflex II with Cu-Ka source 1.5404 Å, Rigaku).
Scanning electron microscopy (1450VP, Carl Zeiss) was used to image the cross-section of foam samples. Image-J analysis software (National Institutes of Health) was used to measure the average diameter of cells from SEM micrographs by measuring more than 30 sites on each image. Dynamic mechanical analysis (DMA, Mettler
Toledo DMA/SDTA861e) was used in compression mode to evaluate modulus of foam samples as a function of temperature.
Several nano and microparticle fillers were compared for their relative nucleation effectiveness by using a concentration of 0.5wt% of the filler in Engage 8003 resin. The objective of this work was to compare the nucleating agents for their ability to produce low-density foams at a fixed process condition. Following the experimental procedure outlined in the Experimental section, the foams were stabilized to the best extent possible by cooling. The strand die used for these experiments was not optimized for foaming, and resulted in non-uniform quality of the foams. Samples were taken from representative samples of average quality. Table 1 displays the measured specific gravity, the average cell diameter, the calculated average number of cells per volume, calculated mean wall thickness, and displays a
SEM micrograph of a cross-section of the foam for each nucleating agent explored. The calculated average number of cells per volume ( N c
) and mean wall thickness
(

) were determined from Equations 1 and 2, respectively, using the density of the foam (

) as the measured specific

gravity, the density of the polymer (
 p
) as 0.885 g/cm 3 for
Engage 8003, and the average cell diameter ( d ) determined from image analysis.[5]
N
C

1

10


4
 p d
(1)
   d





1
1
   p

1





(2)

Table 1 . Measured and Calculated Properties of Ethylene-Octene
Copolymer (EOC) Foams With Various Nucleating Agents
The quality and density of the foams were found to vary significantly with the type of nucleating agent. The lowest density foams generally were produced from the largest nucleating agents, for example talc and expanded ash. Both materials exhibited closed-cell structure, with cells reasonably uniform in size. The average cell size for the composite containing expanded ash was significantly smaller than that for talc, requiring that there be a far greater number of cells per volume in the foam containing the expanded ash to achieve their resulting densities.
Interestingly, the micronized ash, also a relatively large particle compared to the nanoclays, did not achieve a lowdensity foam.
Nanoclay nucleating agents, such as Cloisite 15A,
Cloisite 20A, Cloisite 30B, and Perkalite FR 100 generally produced higher density foams, mostly with cell sizes smaller than that for talc. This is consistent with the conclusion from the literature that nano-sized nucleating agents produce smaller, more dispersed cells with less bubble growth during foaming.[4] This also suggests that these foams exhibit a nanocomposite structure due to exfoliation of the nanoclays. The notable exceptions to this finding are the foams containing Nanofil 116 and
Dragonite XR. For the foam containing Nanofil 116, the density is very low, created by very large cells with apparently thick cell walls. The foam containing
Dragonite XR is moderately denser, with a smaller cell diameter and thick cell walls. The unifying theme behind the difference in the behavior of these clays is that the
Nanofil and Dragonite clays are not synthetically modified by surfactants to assist in exfoliation or polymer intercalation, whereas the Cloiste and Perkalite clays contain surfactant. Without surfactant, the Nanofil and
Dragonite clays should act as larger particles within the composite, leading to larger cell sizes and greater bubble growth.
From the initial study on nucleation effects, it was evident that the nanoclays containing surfactant were more effective at producing nanocomposites than the nucleating agents without surfactant. For this reason, the Cloisite
20A nanoclay was used in subsequent formulation experiments to produce nanocomposite foams. The concentration of nanoclay was varied an order of magnitude, from 0.5wt% to 5wt% in Engage 8003, to determine if mechanical reinforcement concurrent with foam generation would be possible. For comparison, formulations containing talc were produced at 0.5, 1, and
2wt%. Foam strands were produced for each formulation following the same general procedure outlined in the experimental section. Some minor improvements were made during this round of processing that improved the quality of the foam and enabled lower density to be achieved. This included increasing the screw speed to
40rpm and increasing feed rate slightly to better fill the melt seal.

Table 2 displays data and SEM micrographs for the composites containing talc and Cloisite 20A in Engage
8003. Since the loading levels of filler were no longer negligible, the theoretical density was calculated for talc and clay filled polymer using the rule of mixtures. Some variability exists between the samples, but the general trend is that increasing the concentration of filler decreases density and the average cell diameter. This is consistent with the concept that increasing the number of sites for gas nucleation leads to a greater number of smaller cells. Comparing the nanocomposite to the talc composite further reinforces this finding, whereas the average cell size for the talc composite is larger than for the nanoclay.
Table 2 . Measured and Calculated Properties of Ethylene-Octene
Copolymer (EOC) Foams With Talc and Cloisite 20A
It is interesting from these results that filler loadings as high as 5% can still be supported in a relatively lowdensity foam. At these filler loadings, it is expected that the foam could be significantly mechanically reinforced.
To test this, dynamic mechanical analysis (DMA) was performed on cross-sections of the foamed strands in compression mode at temperatures from 25 to 75 o C.
Figure 1 plots the DMA results for storage modulus as a function of temperature for foams containing 0.5wt% and
2wt% filler. As seen in the plot, increasing the loading of filler in the composites increases modulus as expected.
Most significant is the marked increase in modulus for the nanocomposite foams compared to the foams containing talc. The moduli of the nanocomposite foams are more that three times greater than foams containing talc at corresponding loading levels. In fact, a foam containing only 0.5wt% Cloisite 20A exhibits a greater modulus than a foam reinforced with 2wt% talc. It is important to point out that the modulus could be affected by both the density and structure of the foam (number and size of cells) as well as mechanical reinforcement of the polymer by the filler. The samples evaluated in Figure 1 have similar densities, and in the case of the 0.5% loadings, the cell size and number of cells are similar. For this reason, it is likely that the observed effects on modulus are based on differences in the mechanical reinforcement of the polymer, with the nanoclay providing substantially more reinforcement on a weight basis. This demonstrates that the principles of reinforcement from nano-sized fillers in conventional, solid thermoplastics are likewise transferrable to foam. In fact, the exfoliation of nanoclay is likely enhanced in the foam process using supercritical
CO
2
as the blowing agent, due to the very high diffusivity of CO
2
in polymer melt, possibly facilitating penetration and intercalation of polymer chains into the clay.[3]
Figure 1 . Storage modulus as a function of temperature for
Engage foams containing 0.5% and 2wt% talc and Cloisite 20A nanoclay.

To evaluate the structure of the nanocomposites and to verify intercalation of the nanoclay with polymer, XRD was performed on the foam samples. Figure 2 displays the spectra recorded from these measurements on samples containing Cloisite 20A clay, a control sample containing talc, and pure Cloisite 20A clay as a powder. As expected, the talc filled system shows no significant structure of interest at nanometer length scales. The pure
Cloisite 20A clay shows the expected structure of the nanoclay with a primary 2

peak of 3.6
o , corresponding to a d -spacing of 2.45nm. Since the samples are low density foams, and the concentration of filler is low, it is difficult to generate a substantial signal during the XRD measurement. Compounded with the resolution limits of these instruments at low values of 2

, it is difficult to identify a peak for the clay layer spacing in the foam nanocomposites. If the shoulders in the spectra at a 2
 value of 2.5 are taken as the peak reflectance for these nanocomposites, the d -layer spacing of the nanoclay increases to 3.53nm, indicating a very substantial increase in polymer intercalation into the clay. The lack of a clearly distinguished peak is also a strong indication that substantial intercalation and possibly exfoliation has occurred in these systems. We conclude from these measurements that the foams produced with Cloisite 20A are indeed nanocomposites with exceptional intercalated or exfoliated structures.
Figure 2 . XRD spectra for foams containing talc and Cloisite 20A nanoclay. The XRD spectrum for pure Cloisite 20A powder is included for reference.
Coupling agents are typically employed to enhance compatibility of fillers within polymers, and nanocomposites are no exception. The effects of MAPE and an acrylate terpolymer (Elvaloy PTW) were investigated in the Engage/Cloisite 20A nanocomposite system, with the results shown in Table 3. Compared to the nanocomposites without coupling agents, the properties of the foams containing MAPE were similar in density, cell diameter, and the average number of cells per volume. The MAPE coupling agent apparently had little to no effect on the nucleation or bubble growth of the foam. However, for acrylate terpolymer there is a significant decrease in density, increase in cell diameter and decrease in the average number of cells. In effect, fewer, larger cells were formed. This suggests that the mechanism of bubble growth appeared to overcome any increase in the nucleation density that may have occurred by the formation of a nanocomposite. This further suggests that either the acrylate terpolymer is not a very good coupling agent for the clay or that compatibility of the acrylate terpolymer in Engage is poor, thus allowing the coupling agent itself to act as a nucleation agent.
Table 3 . Measured and Calculated Properties of Ethylene-Octene
Copolymer (EOC) Foams Cloisite 20A and Coupling Agents

DMA was performed on foam samples containing
2wt% Cloisite 20A and 5wt% coupling agent. Results are shown in Figure 3 for the storage modulus as a function of temperature. For comparison, the results for 2wt% talc and Cloisite 20A without coupling agent from Figure 1 are plotted again. As seen from the modulus curves, the addition of MAPE coupling agent increases the modulus of the nanocomposite by more than about 10% across the entire temperature range explored. The fact that the
MAPE coupling agent did increase density could in part explain the increased modulus, however the MAPE did not appear at alter cell diameter or the number of cells.
This strongly suggests a very favorable coupling of the nanoclay into the EOC in the presence of MAPE.
Figure 3 . Storage modulus as a function of temperature for
Engage foams containing 2wt% filler and MAPE and acrylate terpolymer coupling agents.
Figure 4 . XRD spectra for foams containing Cloisite 20A nanoclay and coupling agents. The XRD spectrum for pure Cloisite
20A powder is included for reference.
The results for the acrylate terpolymer as a coupling agent are far less favorable. As shown in Table 3, the large cell diameter of the foam indicates that nucleation in this system has fundamentally changed by the addition of the acrylate terpolymer. The DMA data in Figure 3 accordingly demonstrates that the modulus of this foam decreased substantially and was less reinforced than the talc composite.
In order to understand if the poor mechanical performance of the foam containing the acrylate terpolymer was caused by a lack of intercalation or exfoliation, XRD analysis was used to determine the structure of the composites containing coupling agents.
The XRD spectra are shown in Figure 4. The spectra for the foams containing acrylate terpolymer are very difficult to interpret, and no strong indication of a peak 2

value is seen. As shown in Figure 2, pure Cloisite 20A clay exhibits a primary 2

peak at 3.6
o . At this value, no clear structure is found in the traces for samples containing the acrylate terpolymer. This suggests that intercalation and possibly exfoliation of the nanoclay in this system is still occurring. The poor mechanical performance of this clay is therefore believed to be caused by the poor coupling performance of the acrylate terpolymer. The acrylate terpolymer is very likely incompatible with the EOC, causing weakened mechanical properties and overwhelming nucleation from the polymer incompatibility.
The XRD spectra of the Cloisite 20A nanocomposites containing MAPE indicate peaks at 2

values between 2.6 and 2.7
o , representing layer spacing between 3.39 and
3.27nm, respectively. These values are in agreement with those found in Figure 2 for Cloisite 20A nanocomposites without coupling agent. Therefore, a well intercalated or partially exfoliated structure is apparently produced in these foams.
This work has shown successful production of nanocomposite foams in an ethylene-octene copolymer resin using supercritical CO
2
as a physical blowing agent.

The nanocomposite structure appears to be highly intercalated or even possibly partially exfoliated. Superior mechanical properties are clearly observed in these systems, suggesting that nanocomposites may be able to replace some favorable properties of cross-linked polyolefin foams without the additional cross-linking processes.
1.
S.G. Kazarian, Polymer Science , Series C , 42, 1, (2000), and references therein.
2.
V. Goodship, E.O. Ogur, Rapra Review Reports , 15, 8,
(2004).
3.
S. M. Lee, D. C. Shim, J. W. Lee, Macromol. Res.
, 16, 1,
(2008).
4.
L. J. Lee, C. Zeng, X. Cao, X. Han, J. Shen, G. Xu,
Composites Science and Technology , 65, (2005).
5.
D. Klempner, V. Sendijarevic, Eds., Handbook of
Polymeric Foams and Foam Technology , 2 nd Ed., (2004).
6.
M. Maiti, S. Sadhu, A. K. Bhowmick, J Appl Polym Sci ,
101, (2006).
Key Words: nanocomposite, extrusion, foam, supercritical
CO
2
.