Polyolefin–carbon nanotube composites by in-situ polymerization
The in-situ polymerization of olefins such as ethylene and propylene (PP) in the presence of carbon nanotubes (CNTs) is one of the most efficient and versatile ways to synthesize polyolefine nanocomposites. Since a good interfacial adhesion between polyolefin matrix and carbon nanotubes is crucial for successful preparation, the polymerizations were performed with methylaluminoxane (MAO) anchored covalently to the surface of the nanotubes. Thus, a heterogeneous co-catalyst is formed and the polymerization starts with the addition of ethylene or propylene directly on the tube surface. The resulting nanocomposites contain 0.1–25.0 wt% of homogeneous distributed CNT and a good separation of the individual tubes. They show better crystallization temperatures, temperature stability and mechanical properties than unfilled polyolefins.
Polyolefins such as polyethylene (PE) and polypropylene (PP) are the biggest group of polymers with a high growth rate potential, due to their versatile properties, their easy tailored microstructure, and their low costs (Galli and Vecellio, 2004). In 2009, about 120 million tons of polyolefins were produced worldwide (Kaminsky, 2008). Temperature stability, the stiffness, and electric conductivity can be increased by the incorporation of nanoparticles and nanotubes (Moniruzzaman and Winey, 2006; Ajayan and Tour, 2007). The uniform dispersion of the carbon nanotubes (CNTs) in the polymer matrices is required to enhance the good physical and mechanical properties of the composite material. Various techniques to prepare nanocomposites have been studied, such as ultrasonication in combination with solution blending (Barrau et al., 2003; Okamoto et al., 2008), melt blending (Bhattacharyya et al., 2003; Wu et al., 2008), and in-situ polymerization (Bonduel et al., 2005; Wiemann et al., 2005; Jeong et al., 2008).
If nanocomposites are prepared via solution blending, the filler is mixed with a polymer in solution. In addition to mechanical mixing, ultrasound can be used to separate the filler particles. When the dispersion is satisfactory, the solvent is evaporated to yield the filled polymer. This method is suitable for systems that consist of a polymer which is soluble in common solvents (Andrews et al., 1999) and also where the nanofiller can be dispersed well. Polyethylene or highly tactic polypropylene, especially with high molecular masses, is not a good soluble in most low boiling organic solvents, which makes this method problematic for the preparation of PE-CNT- or PP-CNT-based nanocomposites.
In the case of a melt-compounding technique, molten polymer and the nanofillers are mixed intensively under the exposure of shear forces. Different kinds of mixing devices, such as a twin-screw extruder, are available for this task. It is important that the shear forces exerted by the mixer are sufficient to tear the individual particles apart. Especially in the case of carbon nanotubes (CNT), this can be a serious problem because they have a high tendency to agglomerate due to a very high surface energy. This method is not suitable for highly viscous melts of polymers because the mixing is not optimum and the high shear forces can lead to a partial degradation of the polymer itself. A good dispersion of CNT in polyolefin matrices can only be achieved by simple melt-compounding at low filler contents and low molecular masses and an accordingly low viscosity of its melt (Alexandre et al., 2001).
A wide variety of polymer carbon nanotube composites have been prepared by in-situ polymerization using epoxy resins, polyesters, polyurethanes, polyacrylates, polyamides, and polyolefins as the matrix (Harris et al., 2004; Coleman et al., 2006; Tasis et al., 2006).
For polyolefin nanocomposites, the in-situ polymerization of polyolefins in the presence of CNT is one of the most efficient and versatile ways to synthesize nanocomposites. Metallocene-methylaluminoxane (MAO) catalysts allow the tailoring of the polymer microstructure, tacticity, and stereoregularity, by the utilization of a metallocene with a suitable ligand structure (Brintzinger et al., 1995; Kaminsky, 2004). Polypropylene with an isotactic, syndiotactic, or atactic configuration can be obtained by using C2 −, C2v–, or Cs– symmetric zirconocenes (Ewen et al., 1988; Resconi et al., 2000). These catalysts are also excellent tools in the production of copolymers. Metallocene-MAO catalysts are highly active in the production of precisely designed polyolefins and engineering plastics (Coates, 2000; Scheirs and Kaminsky, 2000; Razavi and Thewalt, 2006) and also permit the control of the molecular masses.
The forces necessary to separate the CNT agglomerates into individual particles by melt-compounding or solution blending are applied during the compounding process, while during in-situ polymerization they can be applied beforehand. The in-situ polymerization provides the possibility of a previous CNT separation during the ‘pre-treatment’ in a low viscosity medium like toluene at a comparatively low energy input (e.g. by ultrasound). The matrix is synthesized afterwards in the presence of the (already separated) nanotubes. The resulting nanocomposites are generally indicated by a good separation, homogeneous distribution, and a good wetting with polymer (Andrews et al., 1999; Dong et al., 2006; Funck and Kaminsky, 2007). Theoretically, it should be possible to produce a nanocomposite with any molecular weight at any filler content desired. But there are some polymer fiber agglomerizations which occur when high molecular weight polyolefins are synthesized.
Metallocene-MAO and other single site catalysts are soluble in hydrocarbons and therefore, they can perfectly cover the surface of carbon nanotubes, especially if the surface of multi-walled carbon nanotubes (MWCNTs) shows defects. In a first step, MAO as co-catalyst can be adsorbed or anchored on the surface of MWCNT to change the surface to a hydrophobic one (Fig. 1.1). The absorption of MAO is increased if the carbon nanotubes are oxidized with the formation of hydroxyl and carboxy groups. These groups can react with MAO by the formation of covalent oxygen aluminum bond, without deactivation effects on the catalyst. Excess MAO is washed out. As a second step, the metallocene is added to form catalytically active polymerization sites on the nanosurface (Funck and Kaminsky, 2007). The metallocene, together with MAO, forms an ion pair as active site with an anionic MAO and a cationic metallocene. The cations are sitting outside on the nano surface and through the positive load prevent the aggregation of the tubes. This is one reason why there is such a good dispersion of the fillers in polyolefins obtained by in-situ polymerization with metallocene catalysts. The thickness of the polymer coat, formed by the addition of ethylene or propylene, depends on the polymerization time and the pressure of the monomer. The in-situ polymerization leads to polymer-encapsulated carbon nanotubes.
At the beginning of an experiment, single-walled carbon nanotubes or oxidized multi-walled CNT (2–20 walls, average outer diameter 15 nm, length up to 50 μm supplied by Nanocyl S.A., Sambreville, Belgium) were sonificated in a toluene suspension, using a Sonopuls homogenizer HD 2200 equipped with a KE 76 sonotrode. The amplitude (10–50%) and the sonication time (5–120 min) were varied to achieve an optimum distribution of fillers in the polymer. The sonicated fillers were then either introduced directly into the toluene-charged reactor (without pre-reaction) or stirred with 2 ml of MAO solution (equivalent to 200 mg MAO) for 24 hours (with pre-reaction). MAO was purchased from Crompton as a 10 wt% solution in toluene. After it was filtered over a D4 fritted glass filter, toluene and trimethylaluminum were removed under reduced pressure. The MAO was used as a 100 mg/mL freshly prepared solution in dry toluene. Triisobutylaluminium (TIBA) was obtained from Aldrich and used as a 1 mol/L solution in toluene. The metallocene catalysts such as [rac-dimethylsilylbis(2-methyl-4-(1-naphyl)indenyl)zirconium dichloride] (see Fig. 1.4 (c)) were purchased from Boulder Scientific. The concentration of the solution was 5 mmol/L.
All polymerizations were carried out in a 1-litre glass reactor (from Büchi AG, Ulster, Switzerland) that was heated to 90 °C for 1 hour and then flushed with argon. The reactor was charged with 200 ml of toluene and heated to 30 °C polymerization temperature. In the case of the polymerizations without pre-reaction, MAO and the fillers were introduced into the reactor immediately after sonication, and the dispersion was then saturated with ethylene or propylene at the desired pressure using a mass-flow controller (from Brooks Instruments, 5850 series). The reaction was started by an injection of the metallocene about 10−6 mol zirconocene, Al:Zr = 3500 to 7000. In case of the polymerizations with pre-reaction, 2 ml of TIBA solution (1 mmol/ml toluene) were added to the toluene before saturation with propylene. The pre-activated catalyst (filler/MAO/metallocene) was injected into the reactor when saturation was completed using a pressure lock. Polymerizations were typically quenched after 45 min by adding 5 ml of ethanol. All polymers were stirred with a quenching solution (water, ethanol, hydrochloric acid) overnight, filtered, washed and dried under vacuum at 60 °C. In this way, polyolefin-carbon nanotube composites with 0.5–50 wt% filler content were obtained.
Melting temperatures, Tm, were determined by differential scanning calorimetry (DSC) with a DSC 821e (from Mettler-Toledo) from the second heating cycle at a heating rate of 20 K/min. Crystallization temperatures, Tc, were determined by DSC from the cooling curve (cooling rate 10 K/min) after complete melting at 200 °C for 5 min. All crystallinities were calculated from the melting peak of the second heat of the basis of a crystallization enthalpy of 207 J/g for 100% crystalline isotactic polypropylene (Wang et al., 2001). The half-time of crystallization (τ0.5) was determined by isothermal DSC experiments. The samples were quenched to the desired isothermal crystallization temperature (cooling rate 40 K/min) after they were entirely molten at 200 °C for 5 min. Electron microscopy was performed on a Leo 1530 FE-REM (SEM) and on a JEOL JEM-1011 (TEM). For studies of morphology, composite cryofractures were used. Gel permeation chromatography was carried out with a Waters GPC 2000 Alliance system equipped with a refractive index detector, viscosimetric detector, and a set of three columns, Styragel type. The particle size for each column was 10 μm, and the pore sizes were 103 Å (HT3), 104 Å (HT4), and 106 Å (Hto). 1,2,4-Trichlorobenzene was used as solvent. The analyses were performed at 140 °C and 1.0 mL/min.
In comparison to Ziegler-Natta catalysts, metallocene-MAO catalysts show only one type of active site and their chemical structure can easily be changed. These properties allow one to accurately predict the properties of the resulting polyolefins by knowing the structure of the catalyst used during their manufacture and to control the resulting molecular weight and distribution, comonomer content and tacticity by careful selection of the appropriate reactor conditions. In addition, their catalytic activity is 10–100 times higher than that of the classic Ziegler-Natta systems.
A key factor in the activation of metallocenes is the co-catalyst. MAO is a mixture of different structures in which the unit Al4O3(CH3)6 is the main product (Fig. 1.2). The compound is electronically unsaturated (Fig. 1.2) as a Lewis acid and forms bigger cages or clusters such as from 3 or 4 units and separation of trimethylaluminum as condensation reactions (Sinn, 1995; Koide and Barron, 1996).
1.2 Structure of methylaluminoxane (MAO) (Sinn, 1995).
It is generally assumed that the function of MAO is first to undergo a fast ligand exchange reaction with the metallocene dichloride, thus rendering the metallocene methyl and dimethyl aluminium compounds. In a further step, either Cl− or CH3− is abstracted from the metallocene compound by an Al-centre in MAO, thus forming a metallocene cation and a MAO anion (Sishta et al., 1992). The alkylated metallocene cation represents the active centre. Meanwhile, other weakly coordinating cocatalysts, such as tetra(perfluorophenyl)borate anions [(C6F5)4B]−, have successfully been applied to the activation of metallocenes (Duchatean and Bochmann, 1997; Sishta et al., 1992).
Using metallocene catalysts it is possible to produce polyolefins with a narrow molecular weight distribution (Mw:Mn = 2) and high activities. The catalyst of biscyclopentadienyl zirconiumdichloride (Cp2ZrCl2)/MAO produces 95 °C polyethylene with an activity of 40 million g PE/g Zr · h. Polypropylenes, with different microstructures and characteristics, can be custom-made just by varying the ligands on metallocene (Fig. 1.3).
The different microstructures are shown in Fig. 1.4. Bridged fluoren-cyclopentadienyl catalysts (a) with a Cs-symmetry produces a syndiotactic polypropylene, C1-symmetric catalysts such as bisfluorenyl (b), or unbridged bisdyclopentadienyl zirconocenes produce atactic polypropylenes and C2-symmetric metallocenes such as bridged and substituted bisindenyl complexes (c), produce isotactic polypropylenes.
1.4 Structures of zirconocenes used as catalyst components for olefin polymerization. The Cs-symmetric catalyst [Ph2C(Cp)(Flu)]ZrCl2/MAO (a) produces syndiotactic poly(propylene) (sPP); the C1-symmetric catalyst [Me2Si(9-Flu)2]ZrCl2/MAO (b) atactic poly(propylene) (aPP); and the C2-symmetric catalyst rac-[Me2Si(2-Me-4-(1-Naph)-1-Ind)2ZrCl2/MAO (c) isotactic poly(propylene) (iPP).
Isoblock and stereoblock structures are obtained if insertions happen by higher temperatures or by different substitutions of the metallocene structure. By combining different olefins and cycloolefins with one another, the range of characteristics can be further broadened (Kaminsky, 1998).
Copolymers, block copolymers, and long-chained copolymers such as ethylene/propylene copolymers, ethylene/norbornene-, propylene norbornene copolymers are of interest because of their high melting and glass transition points up to 200 °C (Kaminsky et al., 2004; Arriola et al., 2006). Long-chain branched polyethylenes can be obtained with a crystallizing polyethylene backbone chain and oligoethylene or oligopropylene side chains. The oligopropylene side chain can have iso-, syndio-, or atactic microstructures. In the case of long-chain branched polypropylenes, the variety is even higher (Arikan et al., 2007). All these homo- and copolymers can be synthesized by in-situ polymerization to coat carbon single- and carbon multi-walled nanotubes.
Polyethylene multi-walled carbon nanotube–polyethylene composites (PE–MWCNT) were prepared by in-situ polymerization with different metallocene catalysts. Park et al. (2008) used biscyclopentadienyl zirconiumdichloride (Cp2ZrCl2) as the catalyst. The content of MWCNT in the polymer was varied between 1 and 5 wt%. The electrical, thermo- and dynamic mechanical properties are of great interest (Yang et al., 2007). Low amounts of MWCN can reduce the electrostatic load of polyethylene films drastically.
We investigated PE–MWCNT composites synthesized by the highly active zirconocene b (see Fig. 1.4). For the in-situ polymerization, oxidized MWCNTs were used (Table 1.1). PE–MWCNT composites can be produced with very high activities; 20,000 kg of polymer is obtained in one hour with one mol of the zirconium complex. As suspected, the activity decreases with an increasing amount of tubes in the starting suspension from 29,000 kg PE/mol Zr·h (pure PE) down to 4700 kg PE/mol Zr·h (18.4 wt% of CNT in PE). There is nearly no effect of increasing CNT content on the molecular weight of the polyethylene. The molecular weight is high, about 700 000–900 000. Compared to pure PE, the nanocomposites show a decrease of 1 or 2 °C in the melting point. Only the crystallization temperature shows a significant effect. Small amounts (less than 1 wt% of CNT) are sufficient to increase the crystallization temperature by 10 °C. The tubes act as nucleation agents.
Synthesis of PE–MWCNT composites by 30 °C in toluene using [Ph2C(Cp)(Flu)]ZrCl2/MAO (structure (a) in Fig. 1.4)
All PE–MWCNTcomposites are characterized by the strong increase in the tensile strength; the content of 1 wt% of MWCNT results in a tensile strength of 20 MPa while this value is only 5 MPa for pure PE, 10 wt% of MWCNT gives 47 MPa. Even higher is the effect for the electric conductivity (Table 1.2). For these measurements, plates of 1 mm thickness were pressed from the PE–CNT composite materials. The voltage used was 85 V.
For the synthesis of polypropylene (PP)–CNT composites, double-walled (DWCNTs) and multi-walled carbon nanotubes (MWCNTs) were used. Composites were obtained with a matrix of isotactic polypropylene (iPP), ultra high molecular weight isotactic PP and syndiotactic PP (sPP) depending on the metallocene structure (for the metallocene structures, see Fig. 1.3).
Table 1.3 shows the most important polymerization conditions and properties of iPP/DWCNT composites. Zirconocene was used, dimethylsilyl(bis-2-methylindenyl)zirconiumdichloride [Me2Si(2-Me-Ind)2]ZrCl2.
In-situ polymerization of iPP/DWCNT composites depending on the filler content; catalyst [Me2Si(2-Me-Ind)2]ZrCl2/MAO (zirconocene structure in Fig. 1.1)
The polymerization activity as well as the molecular mass of the obtained iPP matrix does not depend a lot on the filler content. Activities up to 27 000 kg PP/mol cat·h·[P] and molecular weights of the iPP of 380 000 were achieved. Surprisingly, the melting point of the iPP increased with increasing amounts of DWCNT. The double-walled carbon nanotubes not only increase the crystallization rate as found for PE–CNT composites, but also the tacticity of the polypropylene is increased. Less mis-insertion occurs. The crystallization temperature of highly filled PP is about 15 °C higher than that of pure iPP.
Similar results were obtained with MWCNT as fillers (Table 1.4) and the synthesis of high molecular weight iPP. The polymerization activities and the molecular weights of the iPP measured by viscosity do not depend greatly on the MWCNT content in the polymer.
In-situ polymerization of iPP/MWCNT composites depending on the filler content, catalyst Me2Si(2-Me4-(1-naph)-1-Ind)2ZrCl2/MAO (structure (c) in Fig. 1.4)
The molecular weight (viscosity average) of the pure high molecular weight iPP is Mv = 1 400 000 g/mol for a metallocene catalyst with a typical polydispersity of two. The polymer matrix of the composites discussed here have a molecular weight in the range of Mv = 1 300 000–1 900 000 g/mol.
The wide scattering is probably a result of CNT remaining during the viscometric and GPC analyses. The polypropene also shows for this catalyst a characteristic percentage of isotactic pentads of 97 ± 2%. The supported MAO is sterically hindered and, due to the removal of solved MAO, the Al:Zr ratio is not optimal. However, the polymerization activity is determined to an average quantity of 5000 kgPol/(molZr/h × molMon/l).
To reinforce high and ultra-high molecular mass polypropylene with nanotubes through in-situ polymerization, a CNT pre-treatment that avoids the uncontrolled gelation of the polymer inside the reactor is necessary. This task is accomplished by linking the co-catalyst covalently to the nanotube surfaces.
To prevent the PP from forming an unwanted gel, the growing polymer strings must entwine themselves around the nanotubes. Therefore, the polymerization must take place directly and only on the CNT surfaces. To force the catalyst to form the active complex only with the inhibited MAO, immobilized on the CNT, no dissolved MAO must be left in the reactor. Consequently, any unreacted and dissolved co-catalyst is removed by a filtration and washing procedure (Kaminsky et al., 2008).
The resulting and extremely strong nanocomposites were obtained in powder form while the high molecular weight iPP becomes insoluble during the polymerization process and tends to accumulate as a block around the stirrer. The obtained pure iPP normally must be cut off the stirrer in one piece and has a fibrous surface structure after drying. It is a great advantage that CNT-high molecular weight iPP can be obtained in powder form and this makes the processing easier.
The morphology of the iPP–CNT nanocomposites was investigated by using scanning electron microscopy (SEM), while transmission electron microscopy (TEM) was used to prove the coating ability of the pre-treatment discussed here and the encapsulation of individual tubes (Fig. 1.5).
As expected, the polymer grows directly on the tube surface, which led to very good coverage and wetting of individual CNTs with a thin PP layer. The encapsulation is almost complete and nearly every tube is affected. It is also obvious that almost every nanotube is coated with about 10 nm in-situ grown polypropylene film and that there is a good attraction on the tubes. Even the small agglomerate on the lower left side is permeated with polymer and seems to be widening by the growth of the polymer chains. Figure 1.6 shows a detailed TEM micrograph of a funnel-shape opened MWCNT. The nanotube consists of 20 walls and is layered by an 8 nm thick in-situ grown PP film. The opened cap is covered, too.
The morphology of nanocomposites prepared by in-situ polymerization is, in comparison to those produced by melt-compounding, generally indicated by a good CNT separation, homogeneous distribution in the matrix, and a good polymer wetting, which in turn indicates a tight adhesion. Especially the pre-treatment discussed here led to an encapsulation of the tubes with a homogeneous iPP layer. SEM micrography taken from cryofractures of different high molecular weight iPP–MWCNT nanocomposites shows no pull-out effect.
If the polymerization time is reduced to some seconds by adding only traces of propylene, it can be seen where there are active sites on the surface of MWCNT formed by defects and anchoring the MAO–metallocene catalyst. Only at some parts of the tube does the polymerization form a ball of polypropylene (Fig. 1.7). The same can be found at the end of the tube (Kaminsky and Funck, 2008).
As previously discussed, a homogeneous distribution and a good interfacial adhesion are crucial for the successful preparation of nanocomposites. The main challenge in this context is to avoid the slipping of the polymer off the nanotubes (pull-out) under mechanical stress.
The differently loaded nanocomposites were examined according to their crystallization behavior which plays an important role during the processing of polymeric materials. It is known that CNTs act as nucleating agents (Valentini et al., 2003; Wang et al., 2006), therefore improvements in the crystallization behavior were expected.
Important parameters for the characterization of the crystallization behavior are the crystallization temperature (Tc), the melting temperature (Tm), and the halftime of crystallization (τ0.5). The melting temperature (see Table 1.4) is not influenced by the presence of nanotubes or their amount. The average melting temperature is 160 ± 1 °C. Also, the crystallinity is not affected by the CNT, i.e. in the range of 45 ± 4%. In contrast to Tm, the effect on the crystallization temperature (Tc) is emphasized more. Pure high molecular mass iPP has a Tc of about 118 °C. The addition of only 0.9 wt% MWCNT led to an increase in crystallization temperature by 5 °C. At higher filler contents the crystallization temperature rises rapidly by up to 7 or 8 °C above the one of pure iPP, leveling at a MWCNT content of about 3 wt%.
Different pre-treatments before polymerization were necessary to achieve a homogeneous distribution of the MWCNT. Preliminary experiments had shown that MWCNTs stayed aggregated if not treated with ultrasound prior to the polymerization. The effect of different ultrasonic amplitudes (which are a measure of the energy input) and various sonification times was investigated.
For experiments without pre-reaction, the MWCNTs were treated with ultrasound in a toluene suspension for 15 minutes prior to the polymerization. The amplitude was varied from 10 to 30%, and no pre-reaction with MAO was carried out. A higher ultrasonic amplitude led to a slightly better dispersion of the nanotubes. It was still not very good after treatment with an amplitude of 30%, however (Kaminsky and Wiemann, 2006).
The activity (Table 1.5) does not show any significant dependence on the amplitude used for the sonication of the fillers. It is 3500 to 4000 kgPol/(molZr·h·[P]) and, therefore, in the range of the activities for the polymerizations without MWCNTs (4300 kgPol/(molZr·h·[P])).
Activities of propylene polymerization after different pre-treatments of the MWCNTs, catalyst [(/p-MePh)2C(Cp)(2,7-bis-tBuFluo)]ZrCl2/MAO (compare structure (a) in Fig. 1.4)
To improve the homogeneity of the dispersion, polymerizations were carried out after the same ultrasonic treatments, but this time a pre-reaction of the MWCNTs with MAO for approximately 24 hours was performed. The improvement of the dispersion with rising ultrasonic amplitude was somewhat more pronounced when the sPP–MWCNT nanocomposites were prepared with pre-reaction than when they were prepared without pre-reaction.
What is clearly more important than the effect of the amplitude is the effect of the pre-reaction with MAO on the dispersion. Obviously, the pre-reaction leads to a better distribution of the nanotubes in the sPP-matrix. This effect is probably due to the hydroxyl groups and other defects present on the surface of the MWCNTs. As described, they can react with the MAO, thus anchoring the co-catalyst to the surface of the nanotubes. This would, on one hand, prevent the nanotubes from reagglomerating after the ultrasonic treatment because of repulsive forces between the MAO ions. On the other hand, it would lead to the polymer formation directly on the surface of the fillers. The growing polymer chains would then separate the MWCNTs from each other during the polymerization.
In addition to the effect of the amplitude, the effect of the sonication time of the nanotubes on their distribution in the polymer matrix was investigated. The amplitude was set to 10% and the sonication times were 15, 30, 60 and 120 minutes. From Table 1.5, it can be seen that a longer sonication time generally improves the dispersion. The most homogeneous distribution is achieved for the sPP–MWCNT composite after sonication of the nanotubes for 120 minutes. This sPP–MWCNT nanocomposite contains 0.5 wt% of nanotubes as compared to 0.8 wt%, 0.4 wt%, and 0.8 wt% for the samples that had been sonicated for 60, 20, and 15 minutes, respectively. The activities were still around 2500 kgP0l/(molzr·h·[P]).
For the subsequent polymerizations, an ultrasonic amplitude of 10% and a sonication time of 60 minutes were chosen as a compromise between a dispersion as homogeneous as possible and the possible damage of the nanotubes due to high amplitudes or long sonication times (Lu et al., 1996). A series of sPP–MWCNT nanocomposites with filler contents between 0.1 and 0.9 wt% was prepared.
The quality of the adhesion of the matrix sPP to MWCNTs was estimated by SEM. There is an excellent dispersion of all individual tubes in the polymer matrix, no bundles are present. Moreover, no pull-out of nanotubes out of the polymer matrix is observed. This is proof of a very good adhesion of the polymer matrix to the nanotube surface because otherwise MWCNTs pointing straight out of the polymer surface and holes where they could have been pulled out of the polymer should have been observed. The carbon nanotubes were wet well, which also is a sign of a good adhesion between the components of the nanocomposite.
The molecular weights of the sPP–MWCNT nanocomposites were elevated slightly with regard to the pure sPP. There is almost no dependence on the molecular weights of sPP and the preparation technique (see Table 1.5). The molecular weight varies between 270 000 and 315 000 and is slightly higher for polymerizations with pre-reaction time. For neat sPP, a value of 310 000 g/mol synthesized by the same metallocene catalyst (Kaminsky et al., 2006) was found. These materials were also characterized with respect to their crystallization behavior, their thermal degradation, their electrical properties, and their tensile properties (see next section).
The differently loaded syndio- and isotactic polypropylene–CNT nanocomposites were examined with regard to their degradation, tensile strength, crystallization properties, and selected mechanical properties. The results for the different filler contents of MWCNT in a syndiotactic PP matrix are summarized in Table 1.6.
The melting temperatures of the sPP–MWCNT composites are compared with pure sPP and increased slightly with the filler content. In the composites the melting temperatures lay at least two or four degrees above those of the neat sPP.
The crystallinity increased slightly with respect to the pure polymer as well. In this case, no trend regarding the dependence on the filler content is observed. All crystallinities were calculated from the melting peak of the second heat on the basis of a crystallization enthalpy of 164 J/g for the 100% crystalline material (Wang et al., 2001).
A more pronounced effect of the filler content was observed for the crystallization temperatures. All crystallization temperatures of sPP–MWCNT were located above the crystallization temperature of pure sPP (96 °C). Already at a filler content as low as 0.1 weight-%, the crystallization temperature was raised by 5 °C in comparison to that of neat sPP. This effect is even more pronounced at higher filler contents. The nanocomposites containing 0.9% MWCNTs exhibit a crystallization temperature of 111 °C which is 15 °C higher than that of pure sPP. The rise in the crystallization temperature upon incorporation of carbon nanotubes is an indication of the nucleating ability of the MWCNTs.
Agreement exists about the fact that the crystallization temperature of PP is raised by the addition of carbon nanotubes. The extent of the increase shows great differences, however. The crystallization temperature of iPP–MWCNT films obtained by solution casting is raised to higher values by about 8 °C upon addition of 0.5 wt% of carbon nanotubes (Sandler et al., 2003). In contrast to that, the addition of 0.5 wt% single-walled carbon nanotubes (SWCNTs) to PP–EPDM blends by melt-compounding led to a rise in crystallization temperature of 4 °C only. At higher filler loadings, even a lowering of the Tc is found (Valentini et al., 2003). In iPP–SWCNT nanocomposites that had been prepared by solution blending of PP and SWCNT modified with octadecylamine, a filler content of 1.8 wt% SWCNTs was necessary to reach a 5 °C increase in crystallization temperature. The peak sharpened when SWCNTs were present.
A melt-compounded sample of PP–SWCNT was subjected to isothermal and non-isothermal crystallization. The crystallization temperature of the composite containing 0.8 wt% SWCNT is raised by 11 °C as compared to the neat polymer (Bhattacharyya et al., 2003). The differences in the effect of comparable amounts of carbon nanotubes are probably due to differences in the homogeneity of the dispersion in the polymer matrix and to the different types of nanotubes used.
In our work, an enhancement of the crystallization temperature of 5 to 15 °C of the sPP–MWCNT nanocomposites in comparison to the pure polymer could be achieved. At a filler content of 0.5 wt% of MWCNTs, the crystallization temperature was raised by 10 °C which is more of an increase than in the publications cited above, even though the cooling rate was the same in all experiments. One possible reason for this is the different preparation method used in this work which could have led to a better dispersion of the nanotubes.
The half-time of crystallization (see Table 1.6) is the time needed for 50% of the crystallizable material to crystallize from the melt. Generally, the half-times of crystallization depend on the (isothermal) crystallization temperature and on the amount of filler incorporated and increase with increasing temperatures which could be confirmed for all the materials investigated. This was significantly lower for the sPP–MWCNT nanocomposites than for the pristine polymer which is due to the nucleation of crystallite growth from the MWCNT surface.
The comparison of the nanocomposites with different filler contents shows that the crystallization at a certain isothermal crystallization temperature proceeded faster when more MWCNTs were incorporated in the polymer. If one looks at the half-time of crystallization at 122 °C, for example, it can easily be seen that it was significantly reduced when more carbon nanotubes were present. The half-time of crystallization at this temperature was 15 minutes for the pure polymer. When only 0.1 wt% of carbon nanotubes were incorporated, it was lowered to one third of that value, accordingly, to only 5 minutes. When the percentage of MWCNTs was raised to 0.5 wt%, the half-time of crystallization is reduced by roughly three minutes to 2.3 minutes. The nanocomposite with a filler loading of 0.9 wt% crystallizes too fast to allow for the determination of the half-time of crystallization possible at this crystallization temperature.
In Fig. 1.8, the rate of the half-time crystallization is compared for sPP with different fillers and at different isothermal crystallization temperatures. The filler content is 0.1 wt%. It can be seen that the rate constant of crystallization of pure sPP is always the lowest compared to the nanocomposites.
With respect to the filler type, MWCNTs show the strongest nucleating ability again. The constant rate of crystallization for the pure sPP (7.5 × 10−4 min−1) is multiplied to more than the tenfold value (8.8 × 10−3 min−1) in the presence of the carbon nanotubes. Carbon nanofibers and carbon black show a much smaller effect but still act as nucleating agents for sPP. Lozano and Barrera (2001) and Sandler et al. (2003) have reported an increase in crystallization temperature and an acceleration of the crystallization upon addition of CNFs or CNTs to iPP. The extent to which the properties were enhanced and the influence of the filler content were discussed controversially, though. This could be the result of differences in the homogeneities of the nanocomposites as well as the different types of nanotubes and preparation methods used.
The degradation of the neat polymer and the nanocomposites containing MWCNTs was investigated using TGA. Experiments were performed at a heating rate of 5 °C/min. The results are shown in Table 1.6. The mass loss of the nanocomposites with different filler loadings and the pure polymer are presented depending on the temperature.
The degradation temperature of a composite filled by 0.1 wt% of MWCNT is 9 °C higher than that of neat sPP. Higher amounts of the tubes have no significant influence. The presence of the nanofillers generally enhanced the yield strength of the pure sPP (17.5 Mpa) slightly in all investigated cases. The tensile strength increased for the syndiotactic PP by low filler contents up to 10%. The effects of high molecular weight isotactic PP–MWCNT composites on the crystallization behavior and mechanical properties are similar to that of the composites with a syndiotactic PP matrix.
As previously mentioned, the melting temperature is not influenced by the presence of nanotubes or their amount. The average melting temperature is 160 ± 1 °C. Also, the crystallinity is not affected by the CNT and lies in the range of 45 ± 4%. In contrast to the melting temperature, the effect on the crystallization temperature is stronger. Pure high molecular weight iPP has a crystallization temperature of about 118 °C. The addition of only 0.9 wt% MWCNT led to an increase in crystallization temperature to 123 °C.
Pristine high molecular weight iPP does not crystallize or it takes too long to provide reliable results at temperatures above 130 °C. Compared with sPP (96 °C), this temperature is much higher. Generally, iPP–MWCNT composites crystallize much faster than sPP–CNT composites.
The half-time of crystallization for high molecular weight iPP–MWCNT composites is significantly reduced by low amounts of nanotubes. At 135 °C, it is 4.5 min for a composite with 0.9 wt% filler content. The same composite requires 6.3 min at 137 °C and does not crystallize at 139 °C. When 2.2 wt% of MWCNT are incorporated, the time is reduced to only 2.4 min at 135 °C and to 4.1 min at 137 °C. This nanocomposite crystallizes even at higher temperatures (7.1 min at 139 °C and 8.1 min at 140 °C). When the percentage of MWCNT rises to more than 3%, no significant reduction of the half-time is obtained.
Selected high molecular weight iPP–MWCNT composites were analyzed by dynamic mechanical analysis (DMA) (Kaminsky et al., 2009). Samples of the same form were pressed down by a constant force, the deflection was measured and also the temperature where the deflection was not reversible (Table 1.7).
n.d. = not detected.
Neat iPP shows a high deflection of − 487.8 pm. The form stability can be increased by 100% if 3.7 wt% of MWCNT is incorporated into the polymer matrix by in-situ polymerization. The nanocomposite material becomes much stiffer. Also the temperature stability increases for the nanocomposites. While this temperature is 48.7 °C for neat IPP, it can be increased by 22 °C to a value of 71.5 °C for a composite with 2.3 wt% of CNT.
In-situ polymerization offers a way to prepare CNT containing nanocomposites even with a high molecular weight polyolefin matrix. This makes in-situ polymerization one of the most efficient and versatile methods to synthesize nanocomposites.
Generally, a very good separation, homogeneous distribution and an excellent wetting of the nanotubes with polyethylene or polypropylene can be achieved. The polymerization takes place directly on the CNT surface which in turn led to an encapsulation of the nanotubes with about 10 nm thick polymer film. The thickness of the polymer cover can be varied by the polymerization time, the ethylene or propylene pressure, and the catalyst concentration.
The crystallization temperature is raised up to 7 or 8 °C above that of pure iPP at a MWCNT content of 3 wt% or higher or by 15 °C of pure sPP at a MWCNT content of 0.9 wt%. This reduces the beat times for an injection molding machine to form articles from sPP–MWCNT composites up to a factor of 20 and greatly increases the economy of the process. The CNT nanocomposites show exciting mechanical and electric properties. Some 2 to 3 wt% of MWCNT are sufficient to increase the conductivity 108 times and the stiffness and form stability by 100%. A big market for these composite materials could be the automotive and the electronic industries. Cars should weigh less and then they need less gasoline in the same way.
This can be achieved by using polyolefin–CNT composites. Important for the stiffness and the temperature stability are the aspect ratio and the filler content. The aspect ratio is the length of the tube divided by the diameter of the tube. This must be at least 10,000 to have the wanted properties. For an injection molding machine, the only way to produce cheap autobodies, the length of a fiber cannot be much longer than 1 mm. Otherwise, the nozzle will be blocked. These properties are given for multi-walled carbon nanotube reinforced polymers. For North America, Sinclaire (2001) forecast a large demand of polyolefin nanocomposites in the near future. Through in-situ polymerization, excellent dispersed CNT-composite materials can be produced and should stimulate an interesting next decade for the polyolefin industry.
Alexandre, M., Pluta, M., Dubois, P., Jerôme, R. Metallocene catalyzed polymerization of ethylene in the presence of graphite, synthesis and characterization of the composites. Macromol. Chem. Phys. 2001; 202:2239–2246.
Barrau, S., Demont, P., Perez, E., Peigney, A., Laurent, C., Lacabanne, C. Effect of palmitic acid on the electrical conductivity of carbon nanotubes–epoxy resin composites. Macromolecules. 2003; 36:9678–9680.
Bhattacharyya, 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.
Bonduel, D., Mainil, M., Alexandre, M., Monteverdeb, F., Dubois, P. Supported coordination polymerization: a unique way to potent polyolefin carbon nanotube nanocomposites. Chem. Commun. 2005; 2005:781–783.
Duchateau, R., Lancaster, S.J., Thornton-Pett, M., Bochmann, M. Synthesis of cyclopentadienyl, indenyl, and fluorenylbis(pentafluorophenyl)boranes as ligands in titanium and zirconium half sandwich complexes. Organometallics. 1997; 16:4995–5005.
Okamoto, M., Fujigaya, T., Nakashima, N. Individual dissolution of single-walled carbon nanotubes by using polybenzimidrole, and high effective reinforcement of their composite films. Adv. Funct. Mater. 2008; 18:1776–1782.
Park, S., Yoon, S.W., Choi, H., Lee, J.S., Cho, W.K., Kim, J., Park, H.J., Yun, W.S., Choi, C.H., Do, Y., Choi, J.S. Pristine multiwalled carbon nanotube/polyethylene nanocomposites by immobilized catalysts. Chem. of Materials. 2008; 20:4588–4594.
Sishta, C., Hathorn, R.M., Marks, T.J. Group 4 metallocene-alumoxane olefin polymerization catalysts. CPMAS-NMR spectroscopic observation of cation-like zirconocene alkyls. J. Am. Chem. Soc. 1992; 114:1112–1114.
Valentini, L., Biagiotti, I., Kenny, J.M., Manchado, M.A.L. Physical and mechanical behavior of single walled carbon nanotube/polypropylene/ethylene-propylene-diene rubber nano composites. J. Appl. Polym. Sci. 2003; 89:2657–2663.
Wang, Z.G., Wang, X.H., Hsiao, B.S., Philips, R.A., Medellin-Rodriguez, F.J., Srinivas, S., Wang, H., Han, C.C. Structure and morphology development in syndiotactic polypropylene during isothermal crystallization and subsequent melting. J. Polym. Sci, B. 2001; 39:2982–2995.
Wiemann, K., Kaminsky, W., Gojny, F.H., Schulte, K. Synthesis and properties of syndiotactic polypropylene/carbon nanofiber and nanotube composites prepared by in-situ polymerization with metallocene/MAO catalysts. Macromol. Chem. Phys. 2005; 206:1472–1478.