Chapter 7: Elastomer–carbon nanotube composites – Polymer-Carbon Nanotube Composites

7

Elastomer–carbon nanotube composites

J. Fritzsche,     German Institute of Rubber Technology, Germany

H. Lorenz,     German Institute of Rubber Technology, Germany

M. Klüppel,     German Institute of Rubber Technology, Germany

A. Das,     Leibniz Institute of Polymer Research Dresden, Germany

R. Jurk,     Leibniz Institute of Polymer Research Dresden, Germany

K.W. Stöckelhuber,     Leibniz Institute of Polymer Research Dresden, Germany

G. Heinrich,     Leibniz Institute of Polymer Research Dresden, Germany

Abstract:

Different techniques to disperse multi-walled carbon nanotubes (MWCNT) in elastomers using an internal mixer are applied and the physical properties of the resulting composites are evaluated. It is demonstrated that the dispersion can be improved if the CNT are suspended in a liquid agent in a first step to break up the bonds. Here, ethanol proved to be a good dispersion agent without any additional surfactant and could be vaporized during the mixing process. These investigations are extended to technologically more relevant filler hybrid systems, where parts of the reinforcing filler are exchanged by CNT. Additionally, the use of ionic liquids for improved CNT–polymer interaction is discussed. The dielectric properties, electrical DC conductivity, thermal diffusivity, dynamic-mechanical as well as the stress–strain and fracture mechanical behavior of CNT-filled composites are investigated. The effect of nanoscopic gaps between adjacent CNT on the electrical and thermal conductivity of the composites and the missing percolation behavior of the thermal conductivity are discussed.

Key words

carbon nanotubes

elastomers

hybrid systems

ionic liquids

reinforcement

7.1 Introduction

Carbon nanotubes have been shown to be attractive fillers in various applications due to their unique combination of outstanding mechanical properties, high electrical conductivity and reinforcement quality. Various reports can be found in the literature describing the properties of individual single-walled carbon nanotubes (SWCNT) and multi-walled carbon nanotubes (MWCNT) such as high flexibility (Cooper et al., 2001), extremely high tensile strength (150–180 GPa) and high electrical and thermal conductivity (De Heer, 2004). Due to these excellent properties, a great enthusiasm to explore the potential as nanofillers exists around the world. Nevertheless, the number of commercial products based on CNT at the moment is very small (Moniruzzaman et al., 2006). As a matter of fact, the dispersion of CNT in a polymer matrix is still a great challenge due to the attractive interaction between neighboring CNT and the very large specific surface area of the tube, leading to heavy agglomeration. Considerable effort has been made to disperse CNT in a polymer matrix by solution mixing (Du et al., 2003; Islam et al., 2003; Sundararajan et al., 2004; Cho et al., 2005; Koerner et al., 2005; Haggenmueller et al., 2006) or in-situ polymerization (Ajayan et al., 2000; Gong et al., 2000; Zhu et al., 2003; Jung et al., 2006; Moniruzzaman et al., 2006; Xia et al., 2006) in the presence of nanotubes. In contrast, there are only a few examples of melt mixing, which is the most promising way from the viewpoint of the industrial compounding practice. Successful dispersion was reported for CNT–polycarbonate (Pötschke et al., 2003), CNT–nylon (Liu et al., 2004; Zhang et al., 2004), SWCNT–polypropylene (Battacharyya et al., 2003) and SWCNT–polyimide (Siochi et al., 2004). Pötschke et al. (Pötschke et al., 2007) reported on the successful dispersion of CNT in thermoplastic polyurethanes with promising properties. Similar good results were recently found for CNT–PDMS (polydimethylsiloxan) composites (Bokobza, 2008).

Nevertheless, the distribution of CNT in technical elastomers, which often is not comparable to thermoplastics, has not been well investigated. Considerable improvement of physical properties was reported when carbon nanotubes were incorporated in styrene–butadiene rubber (SBR) (Bokobza, 2007; Bokobza and Belin, 2007). In another work, the reinforcing effect of SWCNT in natural rubber was revealed by dynamic mechanical analysis and Raman spectroscopy (Lopez-Manchado et al., 2004). Here, a noticeable decrease of the loss tangent (tan δ) peak height, as well as a marked shift of glass transition temperature (Tg) towards higher temperature was observed. Fakhru’l-Razi et al. (Fakhru’l-Razi et al., 2006) showed that the initial modulus of a natural rubber (NR) composite was increased by a factor of up to 12 times in relation to pure NR, followed by Wang et al. (Wang et al., 2006) who dispersed CNT in NR via latex mixing.

Kim et al. (Kim et al., 2006) evaluated the mechanical, thermal and electromagnetic shielding properties of ethylene propylene diene monomer rubber (EPDM)–CNT composites. They concluded that the alignment of tubes in an EPDM rubber matrix, arising from the mill processing, resulted in significant improvements in mechanical, electrical and thermal properties. Frogley et al. (Frogley et al., 2003) reported the improvement of the mechanical properties of CNT-filled silicone rubber. In their work, it was reported that with the increase in the number of carbon nanotubes, a remarkable enhancement of the initial modulus (Young’s modulus) was observed, accompanied by a reduction of the ultimate tensile strength. However, at higher strains (≈ 20%), the modulus was found not to change upon further increase of CNT concentration. Hydrogenated nitrile rubber was also used to prepare nanocomposites with CNT (Yue et al., 2006). The authors reported a serious breakage of the CNT during the ultrasonic dispersion, which resulted in a poor electrical conductivity of such nanocomposites. Additionally, the dispersion behavior of CNT in epoxy materials has been investigated (Shaffer and Sandler, 2006).

In order to obtain thermoplastic elastomer–CNT composites for tribological applications, EPDM with a high ethylene content was spray-coated with an aqueous dispersion of the CNT, which was then dried and melt blended with the thermoplastic elastomer (Karger-Kocsis et al., 2008). Synthetic polyisoprene-CNT composites were also prepared by a solution method and the effect of stretching on the electrical properties has been discussed (Knite et al., 2007). Nevertheless, until now, the mechanism of reinforcement and the details of the polymer interaction of CNT with polymer matrices have not been well understood.

7.2 Processing

7.2.1 Variation of the polymer type

When mixed into rubber, normally a poor dispersion of the CNT results. This is due to attractive van der Waals bonds between the outer planes of neighboring nanotubes which result in agglomeration of the powder material. The quality of dispersion depends significantly on the polarity of the used rubber and, therefore, from its interaction with the CNT as well as the rheological properties and the mixing procedure.

To obtain a first impression of the rubber–CNT compatibility, a series of five rubbers varying in polarity and in rheological properties are mixed with 3 phr (parts per hundred rubber) of CNT (Nanocyl 7000, from Nanocyl S. A., Sambreville, Belgium, industrial grade, 90% purity). The macrodispersion was first evaluated analyzing the sample surface by light microscopic techniques in combination with the evaluation program DIAS (dispersion index analyzing system) resulting in a value for the amount of dispersed material. The results are shown in Fig. 7.1.

7.1 Dispersion characteristics of different elastomers as indicated with 3 phr CNT dry mixed, analyzed by light microscopic techniques.

The CNT consists of 90% carbon in graphite layers as well as a small amount of amorphous carbon and has, therefore, a more nonpolar behavior which is only slightly influenced by catalysator waste and undesired functional groups. As a matter of principle, therefore, the best dispersion is obtained by dispersing the CNT in nonpolar natural rubber (NR, SVR CV 50). Here, due to the high molecular weight, also high shear forces can be obtained during the mixing process which improves the dispersion additionally. Similar results can be found for styrene–butadiene/butadiene–rubber (SBR/BR) (SBR: S-SBR VSL 2525–0, Lanxess, BR: CB 24, Lanxess (50:50)). Here the polarity is similar but the shear forces obtained during mixing are lower due to the slightly lower viscosity of the used SBR. The polar rubbers hydrogenated nitrile rubber (HNBR, Therban A3406, Lanxess) and nitrile rubber (NBR, Perbunan NT 3470) show very bad dispersion characteristics as expected due to the weak interaction between the rubber and the CNT, although the shear forces for both types of rubber are high. Additionally, the nonpolar EPDM (Keltan 512, DSM) also shows large agglomerates; here probably the rheological properties are the crucial factor, since EPDM has the lowest viscosity and therefore only low shear forces can be obtained during the mixing process.

The dispersion characteristics are reflected in the mechanical as well as electrical properties which are displayed in Table 7.1. The mechanical properties were analyzed by measuring the tensile properties using a Zwick 1456 (model 1456, Z010) with a cross-head speed of 200 mm/min (ISO 527). In most tested materials, an increase of the tensile strength is obtained with 3 phr CNT. The highest improvements of 176% and 150% are obtained for the SBR/BR and the NR samples, respectively. In the elongation at break values for the SBR/BR blend a reduction of the elongation at break is always observed. Probably, the less than optimal dispersion of CNT results in agglomerates which act as stress concentrators and, therefore, reduce the failure strain. To analyze the electrical properties, dielectric investigations were carried out at room temperature at frequencies from 0.1 Hz to 10 MHz using a broadband dielectric spectrometer BDS 40 (from Novocontrol GmbH, Germany). With the limit of 0.1 Hz, the measured real part of the conductivity σ′ reaches a plateau value which corresponds directly to the DC conductivity σ0. In these measurements, for all materials, no high electrical conductivity is observed, the highest value here is obtained for NR with an AC conductivity of 5.2 × 10–6 S/cm in the limit of 0.1 Hz.

Table 7.1

Influence of 3 phr CNT (dry mixed) on the mechanical and electrical properties of different rubber types

7.2.2 Influence of predispersing solvents

The literature shows that for epoxy and thermoplastic materials the application of additives has proved to be successful in dispersing CNT. Therefore, one approach is to transfer and adapt this method to rubber materials. The used solvents have to fulfill various requirements such as low vapor pressure, an evaporation temperature in the range of the mixing temperature, incompatibility with the rubber used, environmental friendliness and no interaction with the curing system. Therefore, the CNT (Nanocyl 7000, industrial grade) is first mixed with the solvents and dispersed in an ultrasonic bath. Afterwards the CNT-solvent mix was incorporated into the rubber by classical melt mixing in an internal mixer. The rubber used was an SBR/BR-blend (S-SBR 2525–0/CB 24 70:30), and the solvents ethanol and 2-propanol were applied in two weight ratios of solvent:CNT as 5:1 and 10:1. Further details about the sample preparation can be found in (Lorenz et al., 2009).

The effect of the predispersing agent on the dispersion of the CNT can first be evaluated by microscopic techniques. As can be seen in the light microscopic images in Fig. 7.2, significant improved macro dispersion can be stated by use of ethanol with only a few and small agglomerates of CNT. The best results are obtained with an ethanol to CNT ratio of 10:1.

7.2 Dispersion characteristics of 3 phr CNT in SBR/BR (70:30) illustrated by light microscopic images: (a) dry mixed; (b) with CNT:ethanol in a weight ratio of 1:10.

For this ethanol-containing sample, additionally TEM measurements were carried out to investigate the distribution on a microscale. Therefore thin sections of the investigated sample were cut with a microtome. With magnification from Fig. 7.3, it is obvious that the distribution of the CNT is not completely homogeneous, regions of high CNT concentration can be seen that are surrounded by polymer containing only a few CNT. However, it is also revealed from the measure of electrical conductivity that the carbon nanotubes are forming a percolating network at 3 phr loading.

7.3 Dispersion characteristics of 3 phr CNT in SBR/BR (70:30) (CNT:ethanol = 1:10) illustrated by transmission electron microscopy (TEM) images.

Due to the best dispersion, the SBR/BR-composite prepared using ethanol shows the largest gain in the AC electrical conductivity (Fig. 7.4 (a)) with up to σ0 = σ0.1Hz = 10–1 S/m. This value is similar to the values of elastomer composites filled with a highly active carbon black at 10 times higher loading (Klüppel, 2003; Meier et al., 2007). The composite produced with the aid of 2-propanol delivers slightly lower conductivity values, probably due to the slightly lower dispersion. Fig. 7.4 (b) shows the mechanical strength of the different systems. With ethanol, a threefold amplification of stress can be achieved in relation to the unfilled rubber. It is observed that the stress–strain curves of the composites prepared using 2-propanol exhibit somewhat smaller stress values and mechanical strength than those mixed with ethanol.

7.4 Influence of the dispersion on (a) quasi-static mechanical properties and (b) frequency dependent conductivity.

7.2.3 Variation of the carbon nanotube (CNT) concentration

To investigate the influence of the CNT concentration on the investigated properties, a variation of the CNT content between 1 and 4 phr of CNT was carried out for SBR/BR composites (S-SBR 2525–0/CB 24 = 70:30). In the preparation of these composites, ethanol was used as an additive in a weight ratio of 10:1 as described before, since in this case the best properties have been obtained.

As seen in Fig. 7.5, the mechanical reinforcement of these samples rises quite linearly with the CNT content. There is already a considerable reinforcement of 1.6 at 1 phr below the electrical percolation threshold without a continuous filler network. With only 4 phr, nearly five times higher tensile strength values can be achieved. The elongation at break stays nearly constant, with 4 phr, only a slight decrease is observed. Here the fracture surfaces show that cracks are initiated at small dots with diameters of some 100 μm. Therefore, it is supposed that for highly CNT-filled materials, the dispersion is more critical and cracks originate at CNT agglomerates which behave as stress concentrators.

7.5 Quasi-static stress–strain curves of SBR/BR composites (SBR:BR = 70:30) mixed with ethanol (10:1) for different CNT concentrations as indicated.

Figure 7.6 demonstrates a drastic change of the dielectric properties of the described composites. Above a critical volume fraction of carbon nanotubes φc, the percolation threshold, an interconnecting filler network is formed, which results in a sharp increase of the electrical conductivity of the nanocomposites. It is evident from Fig. 7.6 that the percolation threshold is reached when the filler concentration increases from 2 to 3 phr. The very high aspect ratio of the carbon nanotubes is mainly responsible for this percolation behavior at very low loadings of CNT. Supposing a mass density of the CNT of 1.7 g/cm3, which corresponds to the density of carbon black and 0.9 g/cm3 for the matrix, which is typical for elastomers, this results in a volume-related percolation threshold of φc = 0.01, which is quite low but about one order of magnitude higher than theoretical estimations φc = 0.001. This indicates that there is still the potential for an improvement of the dispersion.

7.6 Influence of the CNT concentration in SBR/BR composites (SBR:BR = 70:30) mixed with (10:1) ethanol on (a) frequency dependent conductivity and (b) frequency dependent permittivity.

A closer look at the frequency dependent behavior of the dielectric properties results in a deeper understanding of the CNT networking. In the real part of the frequency dependent AC conductivity σ′ (Fig. 7.6 (a)), a cross-over from a plateau at low frequencies to a power-law behavior at higher frequencies is observed for all samples, which is also well known for carbon black-filled samples (Jonscher, 1977). The characteristic cross-over frequency moves to higher values with increasing concentration of CNT. For the unfilled and the low filled sample (1 phr), the slope m of the high frequency part above the cross-over frequency is nearly m = 1, indicating an insulator-like behavior with an almost linear increase of the conductivity with frequency. For the filled samples with 2 phr CNT this slope decreases to a value around m = 0.6. This transition of the scaling behavior results from the formation of a conducting CNT network on mesoscopic length scales. In the case of carbon black composites a similar scaling behavior with m = 0.6 has been observed and related to anomalous diffusion on a fractal percolation network (Klüppel, 2003).

The conducting nature of the composites becomes very pronounced above the percolation threshold of filler particles. The composites containing 3–4 phr of CNT are definitely above the percolation threshold as the power law regime is particularly shifted to higher frequencies and thus the material behaves like an Ohmic conductor with the conductivity almost independent of frequency.

The frequency dependency of the real part of the permittivity ε′ is shown in Fig. 7.6 (b) and characterizes the polarization of the sample in an alternating field. In the unfilled sample and the sample filled with 1 phr of CNT, the permittivity has low values and behaves almost independent of frequency. With the increasing number of nanotubes, this behavior changes. At low frequencies the permittivity is significantly increasing with the amount of filler. Additionally, at a certain frequency, a relaxation process takes place, resulting in a drastic drop of the permittivity with increasing frequency. With increasing loading of CNT, the relaxation transition is shifted towards higher frequencies. We point out that the behavior of the dielectric response functions shown in Figs 7.6 (a) and 7.6 (b) is very similar to that of carbon black-filled elastomers above the percolation threshold (Jonscher, 1977; Klüppel, 2003; Meier and Klüppel, 2007).

7.3 Structure–property relationships

7.3.1 Mechanical properties

The observed amplification of stress is undoubtedly connected to the high specific surface area of CNT, but it is not typical of ‘hydrodynamic reinforcement’. This mechanism is caused by a ‘shielding’ and immobilization of polymer chains, e.g. by stiff clusters of carbon black. In CNT-filled elastomers, it can only play a role when larger network-like filler clusters are present. This seems to be the case for the ‘dry’ mixed compounds in Fig. 7.4(b). Here, also an upturn of the stress strain curves above 100% is found. The missing upturn for the samples containing ethanol may be caused by the geometry of dispersed nanotubes which hardly allows for a ‘shielding’ effect, but also by a lower CNT/polymer interface strength resulting in de-bonding. It is pointed out that hydrodynamic reinforcement is marked by a ‘shift’ of the upturn from high strains in unfilled rubber to lower strains in the filled rubber, since the local strain in the rubber matrix is essentially a product of the global strain multiplied by a strain amplification factor.

Considering a CNT content of 3 phr in SBR/BR (S-SBR 2525–0/CB 24 70:30) composites mixed with ethanol in a weight ratio of 1:10 CNT:EtOH, the so-called ‘discontinuous damage’ test depicted in Fig. 7.7 (a) reveals a strong stress softening and hysteresis behavior that increase with the applied maximum strain. This effect is well known from carbon black-reinforced elastomers and is typically not observed for the unfilled reference, also shown in Fig. 7.7 (a). It may be related to a breakdown of the CNT network or alternatively to a sliding between the CNT and the polymer, corresponding to an insufficient CNT/matrix coupling. For the dry mixed composite shown in Fig. 7.7 (b), a much less pronounced stress softening and hysteresis is observed. This indicates that no CNT network is formed, which is in agreement with the low electrical DC conductivity of σ < 10–11 S/m. Furthermore, the marked stress relaxation and permanent set increasing with strain are observed for the samples mixed with ethanol, which may be caused by low filler/matrix interaction but may also be due to a softener effect of the remaining ethanol within the rubber matrix. Nevertheless, the insufficient filler/matrix interaction seems probable since CNTs have no structured, active surface like e.g. carbon black, consequently, there will be no tightly bound rubber.

7.7 Quasi-static stress–strain cycles of unfilled and CNT-filled SBR/BR samples mixed (a) with (1:10) ethanol and (b) dry mixing; five repeated cycles for various maximum strain values from 25% up to 100% are shown.

From stress–strain experiments, Young’s modulus can be evaluated and utilized to roughly estimate an apparent effective aspect ratio (length/width) of the tubes following the Guth–Gold–Smallwood equation (Guth and Gold, 1938; Smallwood, 1944; Guth, 1945):

[7.1]

where Ef is Young’s modulus of the filled elastomer, Eu is the value for the unfilled rubber and φ is the volume fraction of the filler. The shape factor f is defined as the aspect ratio of a non-spherical particle. The quadratic term of Equation 7.1 can be ignored at low filler concentrations and for low f-values. The aspect ratios were calculated according to Equation 7.1 and it was found that the f-values vary between 15–20 for dry mixed compounds and 50–60 for ethanolic dispersed compounds. The estimated values are far below the expected values for CNT with an aspect ratio for a typical single stretched nanotube of nearly 1000. The main factors hindering the utilization of the high aspect ratio for the reinforcement process probably are agglomerates of bundled nanotubes due to low dispersion and breakage of CNT during ultrasonication or processing.

Recently, it has been reported that the Guth model and the Halpin–Tsai model fit very well after considering the aspect ratio of 40–45 for multi-walled carbon nanotubes in styrene-butadiene rubber (Bokobza, 2007). The Halpin–Tsai model also describes the relation between the aspect ratio of a fiber and the modulus of a reinforced polymer. It must be noted that originally Equation 7.1 was derived and validated for anisotropic fillers with relatively low aspect ratio. These fillers were assumed to behave totally rigid. For the CNT, the rigid condition is probably not fulfilled if the aspect ratio is large. Accordingly, the longer CNT visible in the TEM micrographs of Fig. 7.3 (b) will bend in the stress field of the rubber matrix and the strain amplification factor estimated in Equation 7.1 can be considered only as an effective one. For that reason, the fitted values for the aspect ratio are expected to be smaller than the values obtained from morphological measurements. Nevertheless, the procedure presented for a rough estimation of an apparent f-value is considered to be useful.

7.3.2 Dynamic-mechanical properties

The dynamic-mechanical properties are also influenced by the incorporation of CNT. Therefore, the properties were analyzed with a Rheomix ARES at a constant frequency of 1 Hz and dynamical strain amplitude of 0.5%, varying the temperature between − 80 °C and + 80 °C. Figures 7.8 (a) and 7.8 (b) show the temperature-dependent behavior of the shear storage moduli G′, the shear loss moduli G″, and the loss factor tan δ of the unfilled sample and the nanocomposite filled with 3 phr CNT. Figure 7.8 (a) illustrates that with an increase in temperature the storage modulus of both samples decreases, which is associated with the glass transition phenomenon of the elastomer chains. Above room temperature, the values of G′ and G″ increase significantly with CNT. As seen in Fig. 7.8 (b), the glass transition temperature at the maxima of the tan δ plot does not change with the incorporation of CNT in the rubber matrix. However, the peak height is considerably reduced. This behavior also indicates the strong reinforcement efficiency with only a low content of CNT.

7.8 Dynamic-mechanical analysis of the unfilled and CNT-filled SBR/BR composite mixed with 10:1 ethanol: (a) temperature sweep of G' and G″; (b) loss tangent (tan δ).

The storage modulus of unfilled rubbers, E′, is independent of the deformation amplitude. In contrast, E′ for the filled rubber shows a significant dependency on the dynamic deformation (Payne, 1965; Heinrich and Klüppel, 2002), i.e. E′ considerably decreases with increasing strain amplitude. This non-linear behavior of filled rubbers is known as the ‘Payne effect’ (Payne, 1965) and has been explained by the existence of a filler network in the rubber matrix above the percolation threshold. With increasing strain amplitude, the filler network breaks down, which results in the lower E′ values. In a SBR/BR-blend with various concentrations of CNT (dispersed with ethanol), no Payne effect is observed at small CNT loading (up to 2 phr) as seen in Fig. 7.9 (Das et al., 2008). With the increase of the CNT content, an increase of E′ at low deformation amplitudes is observed. So, even with 3 phr of CNT, the tubes obviously form a continuous filler network in the rubber matrix. The filler–filler network can be reformed again after a certain time interval. Payne revealed that the value of E′ is largely recoverable upon return to smaller amplitudes in the linear regime. So, flexible rubber chains allow the filler particles to rearrange again to form a three-dimensional filler network in the rubber matrix (Payne, 1965). In order to investigate the ability to recover, the strain-sweep experiments were also carried out in the reverse direction from higher to lower strain amplitudes for the samples with unmodified CNT dispersed with an ethanol suspension. It is observed that the E′ values do not reach their original value within the relaxation time of the experiment, but a recovery of the E′ values in the limit of low strain has been attained (Fig. 7.9). This behavior of a rubber can be explained by the stress–softening effect during the dynamic strain. So at least it can be said that rather than damage or permanently break the tubes, the amplitude sweep disrupted the filler–filler network in the rubber matrix.

7.9 Strain dependency of the elastic modulus E' for CNT filled S-SBR/BR blends (SBR/BR 50:50).

7.3.3 Electrical transport processes

The data in Fig. 7.6 (b) showed that the permittivity of the samples above the percolation threshold φ > φc is very high and a relaxation transition appears, which shifts to higher frequencies with increasing CNT loading. This behavior is also observed for carbon black-filled elastomers (Klüppel, 2003; Meier et al., 2007) and can be explained qualitatively by a combined effect of nanoscopic gaps between the adjacent CNT and the predictions of the percolation models, e.g. the resistor-capacitor (RC) model (Kirkpatrick, 1973; Havlin and Bunde, 1991; Meier et al., 2007). In such systems there is no perfect electrical contact between the carbon clusters, due to a layer of ‘bound rubber’ that fills the gap between contacting clusters (Klüppel, 2003; Meier et al., 2007). Adopting an RC model of the percolation network, this small layer corresponds to a large contact resistance and a capacitance parallel to it (Fig. 7.10). This unit forms a micro capacitor with a capacity CG and the bound rubber a resistor RG. A detailed derivation can be found elsewhere (Sheng et al., 1978; Sichel et al., 1978; Klüppel, 2003, Meier et al., 2007; Meier and Klüppel, 2008). In the case of nanometer-sized gaps between the conducting particles, free electrons can tunnel from one particle to another.

7.10 Schematic representation of a percolation network based on CNT, description of the CNT–CNT connection as an RC-unit.

In this frame, the relaxation process can be referred to the movement of electrons by tunneling over these gaps. The characteristic frequency of the relaxation transition observed at high frequencies (~ 1 MHz) can be related to the distance between adjacent filler particles (Meier et al., 2007).

In detail, the resistivity can be calculated as a quantum mechanical tunneling current and does follows an exponential function of the gap distance δ:

[7.2]

where A is the cross-section of the gap, k0 describes the potential barrier with the potential height V, the Planck constant ħ and the electron mass me. The characteristic frequency of a tunneling process over such a CNT–CNT connection can then be described by Equation 7.3:

[7.3]

Therefore, with the knowledge of the characteristic frequency ωG and the assumption of a typical potential barrier of V = 0.3 eV and a dielectric constant of the polymer in the gap of ε = 3 the gap distance can be calculated. We point out, that in (Sheng et al., 1978) and (Sichel et al., 1978), the value of V = 0.2 eV was assumed to be characteristic for carbon black polymer composites. The present somewhat larger choice of V = 0.3 eV is based on recent TEM investigations of the gap distance in carbon black–SBR composites (Fritzsche and Klüppel, 2008). It delivers slightly smaller values for δ.

The characteristic frequency ωG has been determined by a quantitative evaluation of the dielectric spectra by fitting Cole-Cole functions (Equation 7.4) simultaneously on the quantities ε′ and ε″:

[7.4]

As fit parameters beside the relaxation frequency mG, the DC conductivity odc, the relaxation strength Δε = εs–ε as well as the broadening parameter α are obtained. For the fitting procedure, one fit function was sufficient, additionally, the term for the DC conductivity has been added. It was clearly shown that the relaxation time τ =1/ω decreased with the rising CNT concentration, which is in correlation to the shift of the relaxation process to higher frequencies.

The resulting gap distances δ calculated in Equation 7.3 are plotted against the filler volume fraction in Fig. 7.11. Additionally, the dependence of the DC resistivity plateau measured at 0.1 Hz on the filler content is included. An important result in this context is that the gap distance δ is exponentially decreasing with increasing CNT concentration and does reach a plateau value. This behavior is identical to carbon black composites (Kohjiya et al., 2006; Fritzsche and Klüppel, 2008). Additionally, it can be stated that the decreasing gap distance is connected to a decreasing resistivity.

7.11 Percolation behavior of the plateau value of the AC resistivity for SBR/BR (70:30) and gap distances between nanotubes calculated by the RC model of the CNT network based on permittivity data.

The amount of polymer bound to the filler particles is dependent not only on the specific surface area, but also on the surface activity. Due to the low activity of graphitic carbon, it is not clear if such a rubber layer also exists in CNT–polymer networks, but it is indicated by the dielectric data. However, it is clear that the charge transport through the tube network is strongly hindered by hopping or tunneling of charge carriers over small gaps, indicating that adjacent tubes do not touch but are separated by thin polymer layers. This is also confirmed by the obtained maximum conductivity values of 10–1 S/m for the percolated systems which lie several orders of magnitude below that found for bundles or single CNT. The percolation threshold indicated by the steep increase of the conductivity lies around 1 vol.% CNT. This indicates that the dispersion of the CNT is quite good though it may be further improved.

7.3.4 Thermal transport processes

The thermal diffusivity of the CNT composites was evaluated by measuring the temperature development in SBR/BR samples (S-SBR 2525–0/CB24 70:30) filled with 3 phr CNT (Nanocyl 7000). Therefore two plates of one sample were placed in a heating press with a temperature of Tu = 100 °C. A temperature sensor was placed in between the two sample plates and the temperature increase T(t) is measured in dependence on time t compared to the start temperature T0. To calculate the thermal diffusivity a, the resulting values can be described by Equation 7.5 with δ = 10 cm the thickness of the sample plates:

[7.5]

To evaluate, the measured values were plotted as ln{(T(t)-Tu)/(T0-Tu)} against time t. The slope in that plot is then directly connected to the thermal diffusivity a by using only that part of the temperature curve showing a linear dependence.

The regression statistical error of the calculated values was about 1%. For the unfilled rubber, one obtains a thermal diffusivity of a = 0.83 × 10–7 m2/s, which increased to 1 × 10–7 m2/s with 3 phr CNT mixed 1:10 with ethanol. A similar value of a = 1.01 × 10–7 m2/s is obtained with 3 phr CNT mixed 1:10 with 2-propanol. This results in a thermal conductivity of λ = 0.19 W(m K)–1 for the unfilled polymer and λ = 0.21 W(m K)–1 for the composite with 3 phr CNT, respectively. This means that λ increases by only 20%, while the electrical conductivity increased by about 10 orders of magnitude.

Since the charge transport through the tube network is controlled by hopping or tunneling over gaps, it is clear that the DC conductivity of the composite would be dominated by the gap, rather than the much higher conductivity of the CNT. This also affects the thermal conductivity induced by electron transport. In addition, the transport of heat in isolated CNTs is dominated by phonons showing a characteristic quadratic dependence on temperature in the range between 50 and 300 K. The thermal conductivity exhibits a maximum at about 320 K of more than 3000 W(m K)–1, and for higher temperatures the thermal conductivity decreases due to phonon back-scattering effects. For CNT dispersed in a polymer matrix, further scattering effects, e.g. interfacial boundary and defect scattering, will appear, leading to a drastic reduction of thermal transport properties (Gojny et al., 2006). In addition, the thermal transport through the CNT network by phonons will be strongly hindered by the gaps between adjacent tubes. From these arguments we conclude that the thermal conductivity of the CNT/rubber composites should lie several orders of magnitude below that of the isolated CNT (up to 3000 W(m K)–1 (Berber et al., 2000; Kim et al., 2001). Accordingly, the thermal conductivity of the CNT network in rubber composites should be not much larger or even smaller than that of the pure polymer (about 0.1 W(m K)–1).

7.4 Systems with ionic liquids for increased coupling activity

Ionic liquids with imidazolium ions can be transformed into a ‘gel’ in the presence of a small amount of SWCNT (Fukushima et al., 2003) due to the strong physical interaction of the π electrons of the nanotubes with the imidazolium ions. By assuming that some ionic liquids may behave like a coupling agent between rubber polymer and CNT, a series of ionic liquids were considered in order to find good rubber/nanotubes compatibility. Further details can be found in (Das et al., 2009). The ionic liquids used are summarized in Table 7.2. As a basis polymer an SBR/BR blend (S-SBR 2525–0/CB 24 50:50) was used with a constant CNT concentration of 3 phr CNT. The composites were prepared by mixing the CNT with ethanol in a weight ratio of 1:10 and adding 3 m mol ionic liquid to this mixture as well in a weight ratio of 1:10 CNT to ionic liquid. This pre-batch was then remixed with the polymer in an internal mixer.

Table 7.2

Chemical structures of the ionic liquids

As previously described, the addition of 3 phr CNT without any ionic liquid increases the elongation at break value to a certain extent. However, the presence of ionic liquid enhances this property even further (Table 7.3). Here the ionic liquid AMIC (1-Allyl-3-methyl imidazolium chloride) along with 3 phr CNT shows an elongation at break value of 457% which is nearly three times higher than the value of the unfilled sample. For these particular composites, the tensile strength as well as the 100% modulus has the highest values among all other 3 phr filled composites. From these physical data, it is clear that AMIC with a double bond in the tail shows the best reinforcing activity. It is believed that the double bond is chemically linked to the double bond of the diene rubber molecules by sulphur bridges and simultaneously has strong interactions with the π electron cloud of the CNT due to delocalization of the π electrons in the imidazolium carbocation.

Table 7.3

Mechanical and electrical properties of SBR/BR (50:50) filled with different concentrations of CNT with and without additional use of the ionic liquid AMIC

In order to check the dispersion and the CNT network in the rubber matrix, TEM studies were done with the AMIC containing CNT composites shown in Fig. 7.12. In the evaluated pictures, no agglomeration of the CNT has been observed and a continuous percolating network within 3 phr of CNT loading is assumed. However, at less magnification, some spheres with dark color were observed (Fig. 7.12 (b)). One of the dark phases was further focused upon and it was found that CNTs were forming a cluster of a spherical phase along with the wrapping network of CNTs on the surface of the spherical phase (Fig. 7.12 (c)). These three-dimensional interconnected globular-like structures are defined as ‘cellular structures’ of the composites comprised of CNTs (Endo et al., 2008). At 11 wt% loading, the formation of cellular structures was also reported in fluoro-rubber filled with CNTs. In the present case, at relatively low loading (3 wt% of CNT), cellular structures are formed, promoted by AMIC ionic liquid.

7.12 TEM images of a SBR/BR (50:50) matrix loaded with 3 phr CNTs in the presence of 1-allyl-3-methylimidazolium chloride: (a) exhibiting the evenly distributed CNTs with an interconnecting network; (b) overall dispersion of the CNTs; (c) clusters of ‘cellular structure’ formed by CNTs; (d) a magnified image of the cellulation structure.

Since the use of AMIC as ionic liquid gave rise to the highest mechanical performance compared to other composites at same CNT loading, it is also interesting to determine the electrical properties of the composite containing AMIC. Figure 7.13 shows the dependence of the electrical conductivity on the volume fraction of CNT. Again, at least 102 times higher conductivities have been found with AMIC compared to other composites at 3 phr CNT. It was clear from the conductivity data of the unfilled rubber compounds with only 3 mmol phr ionic liquids that the significant increase of the conductivity is not related to the ionic liquid itself. Therefore, all effects can be traced back to the CNT percolation network.

7.13 Dependence of the conductivity on the volume fraction of CNT in the presence and absence of 1-allyl-3-methylimidazolium chloride ionic chloride. The lines are a guide only.

Dynamic mechanical properties of the pure S-SBR/BR blend and its composites with 3 phr CNT and different ionic liquids were studied at 1 Hz and 0.5% strain amplitude over a wide temperature range (− 80 to + 80 °C) and in some cases from − 80 to + 140 °C. The dependency of the tan δ on the temperature is illustrated in Fig. 7.14 (a). Here the addition of AMIC in the composite filled with 3 phr CNT clearly reduces the peak height. In the storage modulus (Fig. 7.14 (b)), it can be seen that the storage modulus sensibly increases in the presence of AMIC which is attributed to the higher reinforcement effect. A distinguishable increment of the modulus at the same filler loading directly indicates a strong rubber–filler interaction. Additionally, from Figs 7.14 (a) and 7.14 (c), it can be seen that an extra relaxation process is taking place at high temperatures for AMIC containing composites with a peak position around 80 °C. This observation indicates the presence of a relatively rigid rubber polymer in the whole rubber matrix. A schematic presentation of the dynamic mechanical behavior is illustrated in Fig. 7.14 (c).

7.14 Dynamic-mechanical properties of S-SBR/BR with 3 phr CNT in the presence of different ionic liquids: (a) dependence of tan δ over temperature; (b) storage modulus over temperature; (c) dependence of tan δ over temperature with and without 1-allyl-3-methylimidazolium chloride and schematic representation of the chemical interaction between polymer and carbon nanotubes.

7.5 Hybrid systems based on silica filler

Conventional fillers like silica are widely used in the rubber industry especially for high performance rubber goods like tires (Wypych, 1999). The use of silica results in improved resistance to wear and tear, decreasing heat build-up (Berriott et al., 2003), increased stiffness, modulus, rupture energy, tear strength, tensile strength, cracking resistance, fatigue resistance and abrasion resistance (Dannenberg, 1975). However, the dispersion of silica nano-particles in the polymer matrix is also problematic in the case of strong polar bonds between the filler particles, especially for non-polar rubbers. In particular, the application of precipitated silica in high performance tires could become a successful technology only after the dispersion problem was solved by coating the silica with a bifunctional silane in a complicated multi-step mixing procedure. However, in contrast to carbon black-filled samples, silica-filled materials always result in non-conducting samples. Here in several applications the tailor-made silica-based properties in combination with sufficient electrical conductivity, would be advantageous, for example, in antistatic materials.

Usually the low electrical conductivity of polymers is improved by the incorporation of conductive fillers like carbon black. For sufficient high conductivity depending on the structure and the size of the primary aggregates and the specific interaction with the used polymer, percolation thresholds in the range of 8 to 20 vol.% of carbon black are found (O’Farrell et al., 2000). However, with such high amounts of carbon black necessary for a high electrical conductivity, the typical silica properties, such as low rolling resistance and high wet traction, are lost and therefore other techniques to obtain a sufficient conductivity are necessary. Due to the low percolation threshold of CNT, which can potentially be reduced to 0.1 vol.% in the case of optimal dispersion, the addition of a small amount of CNT is one possibility to make silica-filled materials conductive. Additionally, the use of silica is expected to improve the dispersion of the CNT due to the higher shear forces experienced during the mixing process.

7.5.1 Styrene/butadiene (SBR/BR) with silica and addition of CNT

Figure 7.15 shows the quasi-static stress–strain behavior of SBR/BR composites (S-SBR 2525/CB 24 70:30) filled with CNT and silica (Ultrasil 7000 GR, from Evonik Industries). With 3 phr CNT and ethanol as the dispersion agent, the mechanical stress increases by a factor of about 3 related to the unfilled SBR/BR. A similar reinforcing effect can also be observed when 3 phr CNT are added to silica-filled systems. This is indicated in Fig. 7.15 by the two arrows, demonstrating a further increase of the stress for both systems with 40 phr and 60 phr silica. Obviously, CNTs cause an increase in tensile strength of the rubber when added to silica-filled SBR/BR. The ultimate properties of the samples are summarized in Table 7.4. The last column of Table 7.4 shows that the conductivity plateau σdc = 0.10 S/m of the system with CNT is quite high compared to σdc = 10–11 S/m for unfilled S B R/BR. Adding 40 phr silica, σdc increases further to 0.15 S/m. It appears that silica helps to disperse the CNT and, accordingly, lowers the percolation threshold. This can be related to the higher shear forces in the highly filled rubber, causing the breakage of CNT agglomerates. In all cases, only some small agglomerates could be seen in the light microscopic image (inset of Fig. 7.15) confirming that the dispersion is good. Accordingly, the mechanical and electrical properties of the SBR/BR composites filled with CNT and silica appear promising.

Table 7.4

Influence of CNT on mechanical and electrical properties in silica-filled SBR/BR (70:30) blends

7.15 Stress–strain behavior of SBR/BR(70:30) composites with 3 phr CNT mixed with ethanol (10:1) and varied silica content; inset: light-microscopy image characterizing the filler dispersion.

7.5.2 Natural rubber with silica partially exchanges against CNT

Compounds with high silica content (up to 90 phr) are technologically widely used in passenger car tire treads, particularly in SBR/BR/NR blends. Therefore the incorporation of small amounts of CNT in such systems is of high technological interest. Since the dispersion and distribution of CNT by dry melt mixing have proved to be most successful in NR (Lorenz et al., 2009), the first investigation was made with this material combination. This method offers the possibility of using such highly filled NR systems as the masterbatch as well as adapting this concept to other polymer types. It might also serve as a guideline for the development of highly conductive silica-filled truck tire treads with reduced rolling resistance, which meets the growing requirements of improved energy efficiency by reducing the high fuel consumption of trucks. Therefore composites based on NR (SVR-CV 50) with 90 phr silica (Ultrasil 7000 GR, from Evonik) were prepared. To allow adequate processing, 30 phr paraffinic processing oil (Sunpar 2280) was added to all samples. To distinguish between properties based on silica and based on CNT, up to 10 phr of silica was successively replaced by the same amount of carbon nanotubes (Nanocyl 7000). Further details can be found in (Fritzsche et al., 2009).

The dispersion of both types of filler determined by light microscopy is displayed in Fig. 7.16 (a) and 7.16 (b) for the only silica-filled sample and the sample filled with additionally 6 phr of CNT. In both cases, the dispersion is very good with 92–98% indicating that only 2–8% of the filler is not macroscopically dispersed. In TEM measurements (Fig. 7.17), in a huge number of spherical silica particles, small and long nanotubes are visible in the background. These nanotubes are well dispersed in the rubber matrix and almost no agglomeration takes place. The aspect ratio for the tubes lies around 100, based on an average length of 1 μm and an average diameter of 10 nm. The orientation of the CNTs can be ignored.

7.16 Dispersion characteristics of natural rubber samples filled with: (a) 90 phr of silica; (b) 84 phr of silica and 6 phr of CNT illustrated by light-microscopic images.

7.17 Dispersion characteristics of the sample filled with 84 phr of silica and 6 phr of CNT illustrated by TEM image.

The mechanical properties are shown in Fig. 7.18. With increasing exchange of silica against CNT, a significant increase in the tensile strength from 14 to 17 MPa is observable. The elongation at break seems to be nearly constant up to an amount of 6 phr CNT. With an exchange of 10 phr CNT against 10 phr of silica, a complete change of the stress–strain behavior sets in, resulting in a significant increase of the stress at low strains (σ100) from 3 MPa to 8 MPa at 100% elongation and a considerable drop of the elongation at break from over 400 to 300%. The characteristic shape of the stress–strain curve changes significantly to a more linear behavior. In the literature, the influence of nanoparticles on the strain-induced crystallization of natural rubber is discussed. For example, this has been investigated by Carretero-González et al. (Carretero-González et al., 2008) concerning NR/nanoclay-composites. Here depending on the aspect ratio and the orientation of the single platelets, crystallization at smaller strains that does not appear for the unfilled samples has been observed. Similar effects could occur with CNT as the reinforcing filler, if during the stress–strain measurements an orientation of the tubes takes place. Such behavior could explain the increase of the stress, especially at low strain, compared to the silica–filled system, since with most filler systems, like silica or carbon black, the strain–induced crystallization is rather reduced.

7.18 Quasi-static stress–strain behavior of the investigated NR samples with varying concentrations of silica and CNT.

The dynamic property measurements depending on the strain amplitude can deliver information about the filler–filler network as well as the rubber–filler interaction. In the present investigations at room temperature and at a frequency of 1 Hz, the plots of G′ versus double strain amplitude are shown in Fig. 7.19. It can clearly be seen that up to 0.1% strain amplitude, the values of G′ remain constant in all cases and thereafter a significant drop of the values is observed. The G′ value at low strains does significantly increase with the increase of CNT loading. Especially with the sample with 10 phr of CNT, a significant increase of G′ is observed. This behavior is attributed to hydrodynamic effects and an increase of the formation of bound or trapped rubber with the rising amount of CNT.

7.19 Strain-dependent storage modulus of the NR samples in varying concentrations of silica and CNT showing the characteristic non-linear behavior of filled rubbers (Payne effect).

In the dielectric properties (Fig. 7.20), it is found that the plateau conductivity measured by dielectric spectroscopy in the limit of 0.1 Hz significantly increases with the amount of CNTs. With an amount of 6 phr of CNT, the percolation threshold is already passed. This indicates that for a non-conducting material with a high amount of silica, less than 6 phr CNT are necessary to obtain a conducting network. With 10 phr CNT, the conductivity even slightly increases, resulting in a maximum conductivity of about 10–2 S/cm. The relaxation process described in section 7.3.3 appears at relatively low frequencies for the sample containing 3 phr of CNT. With increasing amounts of CNTs, this relaxation process increases in strength and moves to higher frequencies as seen below. This behavior can be related to the decrease in the gap distance between adjacent tubes with rising CNT content.

7.20 Dielectric properties of NR samples filled with silica and CNT: (a) conductivity; (b) permittivity.

Additionally, measurements depending on the temperature between − 100 °C and + 100 °C were performed. In Fig. 7.21 the temperature-dependent behavior at 0.1 Hz was investigated in detail. Here the conductivity is plotted against the reciprocal temperature for the sample with 3 phr and the sample with 6 phr of CNT. At low frequencies at around − 60 °C (1/T = 0.0047 K–1), a relaxation process is visible for both samples. It is related to the thermal expansion of the polymer with increasing temperature. Since the glass transition temperature lies in that temperature region, a connection between the glass transition and the thermal expansion of the polymer is plausible. The thermal expansion of the polymer results in increasing gaps between adjacent filler particles and therefore in a slight decrease of the conductivity. Usually these effects are more pronounced in systems with unstable polymer networks directly above the percolation point. In both samples at low temperatures, beside the mentioned relaxation process, the conductivity is more or less independent of temperature. At higher temperatures, a temperature-dependent behavior sets in, resulting in a pronounced increase of the conductivity with temperature. For the sample with 6 phr CNT, the transition starts around 60 °C. This transition indicates a change in the conductivity mechanism from tunneling to a thermally activated hopping mechanism. In the higher filled sample, the transition to the hopping mechanism starts very sharply; this means for all transport processes the hopping mechanism is allowed at nearly the same temperature. Hence, the network is more homogeneous; the distance between the filler particles is similar in the whole sample. The activation energy can be estimated by describing the increase of the conductivity by a simple Arrhenius equation. The slope is then related to the activation energy of the hopping mechanism resulting in EA ≈ 20 kJ/mol for the higher filled sample. In the lower filled sample, the transition is in two steps. Therefore, different temperatures are necessary for the hopping transition, the network is more heterogeneous. The first hopping transport processes start at room temperature. Since the gaps between the clusters are small, the activation energy is quite low with EA 15 kJ/mol. At higher temperatures, the hopping processes for the clusters with larger gaps set in where higher activation energies of around EA 57 kJ/mol are necessary. The results are in agreement compared to those found for carbon black-filled samples. Hence the interpretation based on the percolation theory can also be applied to describe the qualitative properties of the conduction mechanism in CNT–polymer composites.

7.21 Arrhenius plot of the temperature-dependent conductivity at 0.1 Hz for the samples filled with 87 phr silica/3 phr CNT and 84 phr silica/6 phr CNT and evaluation of the hopping activation energy.

7.5.3 Fracture mechanical investigation of silica-based hybrid-CNT composites

The fracture mechanical properties of the CNT composites were investigated by analyzing the dynamic crack propagation rates under cyclical loading by a tear fatigue analyzer (TFA, from Coesfeld) at a frequency of 10 Hz. The determination of the tearing energy, Tel, is based on the following semi-empirical equation (Gent, 2001):

[7.6]

Here, c is the crack length, λ is the strain ratio and Wel is the elastically stored energy density far away from the crack tip. The latter has been estimated approximately by numerical integration of the stress cycles measured on-line for every 1000th cycle during the test at the notched samples. Thereby the remaining cross-section of the sample after subtraction of the crack area was taken as the reference cross-section. Data points were only evaluated in the range where the crack length is small compared to the sample width. Figure 7.22 shows a double logarithmic plot of the fatigue crack propagation rate dc/dn against the tearing energy Tel at high severity under pulsed loading with n being the number of cycles. This type of power law behavior is known as Paris plot with α and β being polymer-specific constants.

7.22 Crack growth rate vs. tearing energy for NR–silica composites, with and without CNT.

[7.7]

This equation has been fitted to the experimental data which is shown as solid lines in Fig. 7.22. The tear fatigue measurements demonstrate that below Tel = 3 N/mm the crack growth rate of the NR-silica composites with 3 phr CNT is smaller compared to the composite without CNT. This indicates that, at small strains, the fatigue performance of the CNT samples is even better than for the silica-filled sample without CNT. Nevertheless, the power exponent β, corresponding to the steepness of the log–log plot, shows a slight increase for the systems with CNT, implying a cross-over of crack propagation rates at lower tearing energies. This behavior correlates with the already observed decrease in elongation at break (high tearing energy) by the addition of CNTs.

7.6 Conclusion

Huge progress in carbon nanotube applications based on elastomer composites can only be realized when a proper dispersion of the entangled agglomerates of as-prepared CNT products is achieved, without damaging their unique properties. Advancements in the use of carbon nanotubes as reinforcing filler in elastomers have been achieved by applying melt mixing techniques for the incorporation of the CNT in elastomers. The use of predispersing solvents like ethanol results in a good dispersion of the CNT as evinced by light microscopy (Fig. 7.2) and transmission electron microscopy (Fig. 7.3). It was observed that the mechanical properties of CNT-filled S-SBR/BR blends are improved considerably compared to the unfilled rubber.

High electrical conductivities of about 10–3 S/cm with the quite low amount of 2–3 phr of CNT are achieved, indicating a percolation threshold of 1 vol.% of CNT (Fig. 7.6). The conduction mechanism was proved to take place over nanoscopic gaps between adjacent tubes; hereby, the gap size decreased with increasing CNT concentration which is analogous to carbon black composites. An increase in thermal conductivity by 20% can be achieved in an SBR–BR blend by incorporation of 1.6 vol.% CNT with ethanol as dispersion agent. This relatively small effect has been explained by the presence of these gaps and the inability of phonons to overcome these distances.

In hybrid filler systems, based on conventional fillers, the best dispersion can be achieved related to the high shearing forces applied during the mixing process (Fig. 7.16). The resulting samples show an increased mechanical stiffness and tensile strength as well as a high electrical conductivity. Dynamic-mechanical measurements show a significant increase of the storage modulus at low strains with rising CNT content and a pronounced Payne effect (Fig 7.19). The fracture mechanical measurements show lower crack propagation rates under cyclic loading at low severity conditions (Fig. 7.22). This study has shown that there is a high potential of using CNT in technical rubber goods like tire treads, where a sufficient high electrical conductivity is necessary but the ideal properties of the silica-filled compounds, like low rolling resistance and high wet traction, should be retained.

7.7 References

Ajayan, P.M., Schadler, L.S., Giannaris, C., Rubio, A. Single-walled carbon nanotube-polymer composites: strength and weakness. Adv Mater. 2000; 12:750–753.

Battacharyya, A.R., Sreekumar, T.V., Liu, T., Kumar, S., Ericson, L.M., Hauge, R.H., Smalley, R.E. Crystallization and orientation studies in polypropylene/single wall carbon nanotube composite. Polymer. 2003; 44:2373–2377.

Berber, S., Kwon, Y.K., Tomanek, D. Unusually high thermal conductivity of carbon nanotubes. Phys. Rev. Lett. 2000; 84(20):4613–4616.

Berriot, J., Montes, H., Lequeux, F., Long, D., Sotta, P. Gradient of glass transition temperature in filled elastomers. Europhys Lett. 2003; 64(1):50–56.

Bokobza, L. Multiwall carbon nanotube elastomeric composites: a review. Polymer. 2007; 48:4907–4920.

Bokobza, L., Multiwall carbon nanotube elastomeric composites. Conference Proceedings, 8th Rubber Fall Colloquium. Hannover, Germany, 2008.

Bokobza, L., Belin, C. Effect of strain on the properties of a styrene-butadiene rubber filled with multiwall carbon nanotubes. J. Appl. Polym. Sci. 2007; 105:2054.

Carretero-González, J., Retsos, H., Verdejo, R., Toki, S., Hsiao, B.S., Giannelis, E.P., Lopez-Manchado, M.A. Effect of nanoclay on natural rubber microstructure. Macromolecules. 2008; 41(18):6763–6772.

Cho, J.W., Kim, J.W., Jung, Y.C., Goo, N.S. Electroactive shape memory polyurethane composites incorporating carbon nanotubes. Macrom. Rapid. Commun. 2005; 26:412.

Cooper, C.A., Young, R.J., Halsall, M. Investigation into the deformation of carbon nanotubes and their composites through the use of Raman spectroscopy. Composites, Part A. 2001; 32A:401–411.

Dannenberg, E.M. Rubber effects of surface chemical interactions on the properties of filler-reinforced rubbers. Rubber Chem. Technol. 1975; 48:410.

Das, A., Stöckelhuber, K.W., Jurk, R., Fritzsche, J., Klüppel, M., Heinrich, G. Coupling activity of ionic liquids between diene elastomers and multi-walled carbon nanotubes. Carbon. 2009; 47:3313–3321.

Das, A., Stöckelhuber, K.W., Jurk, R., Saphiannikova, M., Fritzsche, J., Lorenz, H., Klüppel, M., Heinrich, G. Modified and unmodified multiwalled carbon nanotubes in high performance solution-styrene–butadiene and butadiene rubber blends. Polymer. 2008; 49:5276–5283.

De Heer, W.A. Nanotubes and the pursuit of application. MRS Bull. 2004; 29:281–285.

Du, F., Fischer, J.E., Winey, K.I. Coagulation method for preparing single-walled carbon nanotube/poly(methyl methacrylate) composites and their modulus, electrical conductivity, and thermal stability. J. Polym. Sci., Part B, Polym. Phys. 2003; 41:3333–3338.

Endo, M., Noguchi, T., Ito, M., Takeuchi, K., Hayashi, T., Kim, Y.A. Extremeperformance rubber nanocomposites for probing and excavating deep oil resources using multi-walled carbon nanotubes. Adv. Funct. Mater. 2008; 18:3403–3409.

Fakhru’l-Razi, A., Atieh, M.A., Girun, N., Chuah, T.G., El-Sadig, M., Biak, D.R.A. Effect of multi-walled carbon nanotubes on the mechanical properties of natural rubber. Compos. Struct. 2006; 75:496.

Fritzsche, J., Klüppel, M., Filler networking and reinforcement of carbon black filled styrene-butadiene-rubber. Conference Proceedings, 8th Rubber Fall Colloquium. Hannover, Germany, 2008.

Fritzsche, J., Lorenz, H., Klüppel, M. CNT based elastomer-hybrid-nanocomposites with promising mechanical and electrical properties. Macromol. Mater. Eng. 2009; 294:551–560.

Frogley, M.D., Ravich, D., Wagner, H.D. Mechanical properties of carbon nanoparticle-reinforced elastomers. Compos. Sci. Technol. 2003; 63:1647.

Fukushima, T., Kosaka, A., Ishimura, Y., Yamamoto, T., Takigawa, T., Ishii, N. Molecular ordering of organic molten salts triggered by single-walled carbon nanotubes. Science. 2003; 300:2072–2074.

Gent, A.N. Engineering with Rubber, 2nd edn. München: Hanser Publishers; 2001.

Gojny, F.H., Wichmann, M.H.G., Fiedler, B., Kinloch, I.A., Bauhofer, W., Windle, A.H., Schulte, K. Evaluation and identification of electrical and thermal conduction mechanisms in carbon nanotube/epoxy composites. Polymer. 2006; 47(6):2036–2045.

Gong, X., Liu, J., Baskaran, S., Voise, R.D., Young, J.S. Surfactant-assisted processing of carbon nanotube/polymer composites. Chem. Mater. 2000; 12:1049–1052.

Guth, E. Theory of filler reinforcement. Appl. Phys. 1945; 16:20.

Guth, E., Gold, O. On the hydrodynamic theory of the viscosity of suspensions. Phys. Rev. 1938; 53:322.

Haggenmueller, R., Fischer, J.E., Winey, K.I. Single wall carbon nanotubes/polyethylene nanocomposites: nucleating and templating polyethylene crystallites. Macromolecules. 2006; 39:2964–2971.

Havlin, S., Bunde, A. Percolation II. In: Bunde A., Havlin S., eds. Fractals and Disordered Systems. Berlin: Springer; 1991:137–143.

Heinrich, G., Klüppel, M. Recent advances in the theory of filler networking in elastomers. Adv. Polym. Sci. 2002; 160:1.

Islam, M.F., Rojas, E., Bergey, D.M., Johnson, A.T., Yodh, A.G. High weight fraction surfactant solubilization of single-wall carbon nanotubes in water. Nano. Lett. 2003; 3:269–273.

Jonscher, A.K. The “universal” dielectric response. Nature. 1977; 267:673.

Jung, Y.C., Sahoo, N.G., Cho, J.W. Polymeric nanocomposites of polyurethane block copolymers and functionalized multi-walled carbon nanotubes as crosslinkers. Macrom. Rapid. Commun. 2006; 27:126–131.

Karger-Kocsis, J., Felhoes, D., Thomann, R. Tribological behavior of a carbon-nanofiber-modified santoprene thermoplastic elastomer under dry sliding and fretting conditions against steel. J. Appl. Polym. Sci. 2008; 108(2):724–730.

Kim, P., Shi, L., Majumdar, A., McEuen, P.L. Thermal transport measurements of individual multiwalled nanotubes. Phys. Rev. Lett. 2001; 87:215502.

Kim, Y.A., Hayashi, T., Endo, M., Gotoh, Y., Wada, N., Seiyama. Fabrication of aligned carbon nanotube-filled rubber composite. J. Scr. Mater. 2006; 54:31.

Kirkpatrick, S. Percolation and conduction. Rev. Mod. Phys. 1973; 45:574–588.

Klüppel, M. The role of disorder in filler reinforcement of elastomers on various length scales. Adv. Polym. Sci. 2003; 164:1–86.

Knite, M., Tupureina, Fuith A., Zavickis, J., Teteris, V. Polyisoprene-multi-wall carbon nanotube composites for sensing strain. Mater. Sci. Eng. 2007; 27:1125–1128.

Koerner, H., Liu, W., Alexander, M., Mirau, P., Dowty, H., Vaia, R.A. Deformation-morphology correlations in electrically conductive carbon nanotube-thermoplastic polyurethane nanocomposites. Polymer. 2005; 46:4405.

Kohjiya, S., Katoh, A., Suda, T., Shimanuki, J., Ikeda, Y. Visualization of carbon black networks in rubbery matrix by skeletonization of 3D TEM image. Polymer. 2006; 47:3298–3301.

Liu, T., Phang, L.Y., Shen, L., Chow, S.Y., Zhang, W.D. Morphology and mechanical properties of multiwalled carbon nanotubes reinforced nylon-6 composites. Macromolecules. 2004; 37:7214–7222.

Lopez-Manchado, M.A., Biagiotti, J., Valentini, L., Kenny, J.M. Dynamic mechanical and Raman spectroscopy studies on interaction between single-walled carbon nanotubes and natural rubber. J. Appl. Polym. Sci. 2004; 92(5):3394–3400.

Lorenz, H., Fritzsche, J., Das, A., Stöckelhuber, K.W., Jurk, R., Heinrich, G., Klüppel, M. Advanced elastomer nano-composites based on CNT-hybrid filler systems. Comp. Sci. Technol. 2009; 69:2135–2143.

Meier, J.G., Klüppel, M. Carbon black networking in elastomers monitored by dynamic mechanical and dielectric spectroscopy. Macromol. Mater. Eng. 2008; 293:12.

Meier, J.G., Mani, J.G., Klüppel, M. Analysis of carbon black networking in elastomers by dielectric spectroscopy. Phys. Rev. B. 2007; 75(054202):1–10.

Moniruzzaman, M., Du, F., Romero, N., Winey, K.I. Increased flexural modulus and strength in SWNT/epoxy composites by a new fabrication method. Polymer. 2006; 47:293–298.

Moniruzzaman, M., Winey, K.I. Polymer nanocomposites containing carbon nanotubes. Macromolecules. 2006; 39(16):5194–5205.

O’Farrell, C.P., Gerspacher, M., Nikiel, L. Carbon black dispersion by electrical measurements. Kautschuk Gummi Kunstst. 2000; 53(12):701–710.

Payne, A.R.Kraus G., ed. Reinforcement of Elastomers. New York: Interscience Publisher, 1965. [Chapter 3].

Pötschke, P., Battacharyya, A.R., Janke, A., Goering, H. ‘Melt mixing of polycarbonate/multi-wall carbon nanotube composites’, Compos. Interfaces. 2003; 10:389–404.

Pöetschke, P., Häeußler, L., Pegel, S., Steinberger, R., Scholz, G. Thermoplastic polyurethane filled with carbon nanotubes for electrical dissipative and conductive applications. Kautschuk Gummi Kunstst. 2007; 60:432–437.

Shaffer, M., Sandler, J. Carbon nanotube/nanofibre polymer composites. In: Advani S.G., ed. Processing and Properties of Nanocomposites. New York: World Scientific Publishing Company, 2006.

Sheng, P., Sichel, E.K., Gittleman, J.I. Fluctuation-induced tunneling conduction in carbon-polyvinylchloride composites. Phys. Rev. Lett. 1978; 40(18):1197–1200.

Sichel, E.K., Gittleman, J.I., Sheng, P. Transport properties of the composite material carbon-poly(vinyl chloride). Phys. Rev. B. 1978; 18:5712.

Siochi, E.J., Working, D.C., Park, C., Lillehei, P.T., Rouse, J.H., Topping, C.C., Battacharyya, A.R., Kumar, S. Melt processing of SWCNT-polyimide nanocomposite fibers. Composites, Part B. 2004; 35B:439–446.

Smallwood, H.M. Limiting law of the reinforcement of rubber. J. Appl. Phys. 1944; 15:758.

Sundararajan, P.R., Singh, S., Moniruzzaman, M. Surfactant induced crystallization of polycarbonate. Macromolecules. 2004; 37:10208–10211.

Wang, J.D., Zhu, Y.F., Zhou, X.W., Sui, G., Liang, J. Preparation and mechanical properties of natural rubber powder modified by carbon nanotubes. J. Appl. Polym. Sci. 2006; 100:4697–4702.

Wypych, G. Handbook of Fillers. Toronto: ChemTec Publishing; 1999.

Xia, H., Song, M. Preparation and characterisation of polyurethane grafted singlewalled carbon nanotubes and derived polyurethane nanocomposites. J. Mater. Chem. 2006; 16:1843.

Yue, D., Liu, Y., Shen, Z., Zhang, L. Study on preparation and properties of carbon nanotubes/rubber composites. J. Mater. Sci. 2006; 41:2541.

Zhang, W.D., Shen, L., Phang, I.Y., Liu, T. Carbon nanotubes reinforced nylon-6 composite. Macromolecules. 2004; 37:256–259.

Zhu, J., Kim, J., Peng, H., Margrave, J.L., Khabashesku, V.N., Barrera, E.V. Improving the dispersion and integration of single-walled carbon nanotubes in epoxy composites through functionalization. Nano Lett. 2003; 3:1107–1113.