Morphology-deformation correlations in nanocomposites.

Traditionally, conductive fillers, such as carbon black, chopped carbon fiber or metallic flakes are used in elastomers to obtain high electrical conductivity. The resultant conductive elastomers are critical for applications ranging from the creation of seals between pipes used for transferring flammable gases, electrostatic automotive painting and electromagnetic shielding for mobile electronics. Conductivity is established by percolative network formation of the fillers and limited by carrier transport (hopping or tunneling) between filler particles. Thus, the extent of filler dispersion, aspect ratio of the filler and wettability of the filler by the elastomeric medium are key morphological characteristics in determining the conductivity of the system. The relatively large volume fractions (>20 vol. %) necessary, though, negatively impact deformability, processibility, surface finish and limit the ability to maintain desired conductivity at extreme deformations (>100%).

To address the limitations of classic conductive fillers, numerous investigators are examining the utility of nanoscale multiwall carbon nanotubes with large aspect ratios (>100) and high electrical conductivity ([sigma] ~ 18 kS/cm along tube axis) (ref. 1). In general, additions of small amounts (0.5-10 vol. %) of multiwall carbon nanotubes (CNT) produce elastomeric nanocomposites with high electrical conductivity ([sigma] ~1-10 S/cm), low electrical percolation ([PHI] ~ 0.005) and enhancement of mechanical properties including increased modulus and yield stress without the loss of deformability and elongation to break. The magnitude of the electrical conductivity above the percolation threshold is limited by charge transport between the nanotubes. If tube-polymer interactions are strong, which are considered necessary for complete tube dispersion, tube-tube contact will be mediated by a layer of insulative polymer, compromising the ultimate conductance. In contrast, poor polymer-tube interactions that are weaker than tubetube interactions will compromise dispersion, raising the percolation threshold and the concentration of tubes necessary to achieve satisfactory conductivity. Higher tube loadings will also compromise deformability. Similar qualitative discussions can be framed to speculate on the impact of deformation on conductivity, such as the collective response of the tube-network, the distribution in tube-tube separation or strain-induced changes in elastomer morphology (strain-induced crystallization). Further compounding a general understanding of conductivity-deformation-morphology correlation in these systems are factors such as the nonuniform distribution of strain (local strain, strain-enhancement, etc.) in a composite comprised of high aspect ratio inclusions with moduli orders of magnitude greater than the matrix (ref. 2), and the accumulative impact of strain cycling on morphology and matrix-tube interactions (Mullin effect).

Herein, in-situ x-ray diffraction-deformation studies of a carbon nanotube-polyurethane nanocomposite system are discussed, providing experimental insight on one facet of these complex systems--the correlation between uniaxial deformation, strain induced crystallization and nanotube orientation. The studies indicate that nanotube orientation and polymer deformation are coupled, altering the strain induced soft-segment crystallinity and the mechanical response of the polyurethane at increasing strain.

Results and discussion

The fabrication of the carbon nanotubes-polyurethane nanocomposite (CNT/PU) is discussed elsewhere (refs. 3 and 4). In brief, after short, light grinding of the carbon nanotubes (PRTHT-19, Applied Science.) with a mortar and pestle, they are combined with a small amount of polymer (Morthane PS455203, Huntsman Polyurethanes, aromatic polyester based thermoplastic polyurethane) in a polar medium, such as THE for several hours. This mixture is then added to a polymer solution in the same solvent, stirred for an additional two to four hours, and transferred into casting molds to allow slow solvent evaporation. The samples were then dried in a vacuum oven at 50 [degrees]C until a constant weight was achieved.

The chemical composition of Morthane (PU) was evaluated by quantitative [sup.13]C and [sup.1]H NMR. The peaks were assigned in the proton spectra using 2D NMR methods (COSY and TOCSY) and in the carbon spectrum using DEPT. Table 1 lists the different moieties present and their compositions. The quantitative analysis shows that the Morthane contains only 10% hard-segments and almost exclusively linear diols, with no more than 5% cyclic diols. Morthane exhibits low glass transition (-45 [degrees]C) and soft-segment melting (48 [degrees]C) temperatures, and displays extreme deformation (~700%) and strain-induced crystallization. Severe deformation in the rubbery state at room temperature ([T.sub.g] < RT < [T.sub.m]) induces crystallization of the flexible segments, creating physical crosslinks that restrict strain recovery upon removal of the applied stress.

In-situ x-ray diffraction-deformation experiments were performed at X27C, National Synchrotron Light Source in Brookhaven National Laboratory. Samples with concentrations of 0, 0.1, 0.5, 1.0, 5, 10 and 20 wt. % of carbon nanotubes (PRT-HT-19, Applied Science) and 1 wt. % carbon black (Sid Richardson Carbon) were stretched in an Instron at different strain rates (1, 3, 6, 12 mm/mm) and their x-ray diffraction patterns obtained at time intervals of 30 or 60 seconds, depending on strain rate. Samples were cut into 40 x 4 x 0.4 mm strips to allow for a gauge length of 2 cm for each sample in the tensile tests. WAXD images were collected using a Mar CCD detector. Two dimensional data were background corrected and azimuthally averaged to obtain 2-theta (2[theta]) scans for deconvolution of crystallinity for amorphous and crystalline phases of the polyurethane (figure 1A). The insets of figure 1A show X-ray patterns before and after deformation. Hermans orientation parameter, [S.sub.d] was determined from the azimuthal intensity distribution at 0.338 nm (2[theta] [lambda] = 0.1366nm = 23.3[degrees]), 0.446 nm (2[theta][lambda] = 0.1366nm- 17.6[degrees]) and 0.413 nm (2[lambda] = 0.1366nm- 19.0[degrees]) for CNTs, soft-segment crystallites and amorphous polymer chains, respectively. Orientation parameters evaluated from the small-angle streak (2[theta] = 1.5 [degrees]) caused by aligned carbon nanotubes reflects a similar trend as the CNT WAXS data. Note that since the crystalline and amorphous reflections overlap and the intensity of the crystalline reflection increases with deformation, the degree of orientation, especially at 2[theta][lambda] = 0.1366nm = 17.6 [degrees], is approximate, reflecting a complex superposition of amorphous and crystalline regions.

[FIGURE 1 OMITTED]

Figure 2 summarizes representative data for 0, 1, 10 and 20 wt% CNT/PU samples deformed uniaxially at 1 mm/sec. Overall, the addition of carbon nanotubes to the thermoplastic polyurethane decreases the strain at which strain-induced crystallization occurs. The extent of crystallite orientation is also greater at lower strains as the concentration of CNT is increased. Moreover, the crystallinity index increases with higher volume fractions of CNTs. To a first order, these observations are consistent with strain-enhancement concepts underlying hydrodynamic models of rubber reinforcement (ref. 5).

[FIGURE 2 OMITTED]

For the CNTs, alignment occurs most readily at the low deformations and around [lambda] ~ 3-4 becomes six times less sensitive. The extent of alignment and trend with increasing [lambda] is the same for both 1 and 10 wt. %. This behavior qualitatively mirrors the plateauing of crystalline index with increased elongation. The CNT alignment deviates from a kinematic description of alignment of isolated, rigid rods (ref. 6). This deviation probably arises from the finite flexibility of the bamboo-like CNTs, a CNT concentration that is in excess of [[PHI].sub.C], and the presence of the relatively rigid crystalline phase and physically crosslinked medium.

With regard to the stress-strain response, the similarity between the effective modulus (d[sigma]/d[epsilon]) of the pure Morthane and the nanocomposites at deformations greater than the yield point imply that the presence of the CNTs does not alter the mechanistics of viscoelastic/viscoplastic deformation. This is very surprising, given the substantial morphological change, both for the content of soft-segment crystallites and the continual increase of CNT orientation with deformation. A possible explanation for this behavior is the poor interfacial bonding between the thermoplastic elastomer and CNT. Even though SEMs indicate good wetting of the matrix on the tube, residual void formation is observed, for example within a 10 wt. % (5.9 vol. %) CNT nanocomposite elongated to [lambda] = 5 and then allowed to fully recover to a zero-stress state.

Conclusion

In summary, a complex interplay between nucleation and strain induced crystallization, polymer crystallite orientation and tube alignment underlies the reinforcing effect of multi-wall CNT on Morthane. Incorporation of as little as 2.9 vol. % of CNTs into the thermoplastic polyurethane increases yield stress, stress at break and modulus, without losing the ability to stretch the elastomer above 1,000%. These properties are influenced by a strain induced crystallization of the soft-segments of the PU, which leads to a complex synergism in the mechanical reinforcement of PU nanocomposites. This synergism also carries over to the alignment of CNTs during deformation.

References

(1.) N. Sheng, M.C. Boyce, D.M. Parks, G.C. Rutledge, J.L. Abes and R.E. Cohen, "Multiscale micromechanical modeling of polymer/clay nanocomposites and the effective clay particle, "Polymer, 45, 487-506, 2004.

(2.) H. Koerner, G Price, N.A. Pearce, M.D. Alexander and R.A. Vaia, "Remotely actuated polymer nanocomposites--stress-recovery of carbon-nanotube-filled thermoplastic elastomers," Nature Materials; 3, 115-120, 2004.

(3.) U.S. Patent #6,680,016.

(4.) H.M. Smallwood, J. Appl. Phys., 15, 758-766, 1944; E. Guth, J. Appl. Phys., 16, 20, 1945; J.S. Bergstrom and M.C. Boyce, Rubber Chem. Tech., 72, 633-656, 1999; J.S. Bergstrom and M.C. Boyce, Mechanics of Materials, 32, 627-644, 2000.

(5.) M. Taya, W.J. Kim, K. Ono, Mechanics of Materials, 28, 53, 1998.

Richard Vaia, Peter Mirau and Max Alexander, Air Force Research Laboratory, WPAFB, OH; Hilmar Koerner, University of Dayton Research Institute., WPAFB, OH; and Benjamin S. Hsiao and Igors Sics, State University of New York at Stony Brook

Table 1--Morthane composition

Monomer               Carbon       Percentage

a: Urethane          -CO-NH-          9.9%      [??]
                                                 -O-C-N-[??]-C[H.sub.2]
                                                [??]-N-C-O [??]

b: Dicarboxylic       -COO-          31.8%      [??]
acid                                            -O-C-C[H.sub.2]-
                                                C[H.sub.2]-C[H.sub.2]-
                                                C-O- [??]

c: Diol           -C[H.sub.2]-O-     58.2%      -O-C[H.sub.2]-
                                                C[H.sub.2]-C[H.sub.2]-
                                                C[H.sub.2]-O-

                                                -O-C[H.sub.2]-[??]
                                                -O-C[H.sub.2]-O-

The fraction of urethane (a: hard-segments), dicarboxylic acid and
diols (b, c: soft segments) are calculated from the intensities of the
peaks at 153, 172 and 63 ppm from the [C.sup.13] spectrum.