Preparation, Characterization, and Physical Properties of Multiwall Carbon Nanotube/Elastomer Composites
Leon D. Perez,1,2 Manuel A. Zuluaga,1 Thein Kyu,3 James E. Mark,2 Betty L. Lopez1 1 Instituto de quı´mica Universidad de Antioquia, Apartado 12-26, Medellı´n, Colombia 2
Polymer Research Center, Department of Chemistry, The University of Cincinnati, Cincinnati, Ohio 45221-0172 3
Department of Polymer Engineering, University of Akron, Akron, Ohio 44325-0301
The present article describes preparation, characterization, and physical properties of nanoparticles-filled composites consisting of multiwall carbon nanotubes (MWCNT) and styrene-butadiene rubber and nitrilebutadiene rubber. The reinforcing MWCNT fillers were synthesized by chemical vapor deposition on iron and cobalt catalysts supported by calcium carbonate substrates. These MWCNT were further treated with nitric acid to produce hydroxyl and carbonyl functional groups on the carbon nanotubes (CNT) surfaces. Of particular importance is that these functionalized CNTs were found to exert profound influence on the elastomeric matrices, particularly the vulcanization activation energy, resistance to solvent swelling, enhanced glass transition temperature, and improved storage and loss moduli. POLYM. ENG. SCI., 49:866–874, 2009. ª 2009 Society of Plastics Engineers
INTRODUCTION In practical applications, elastomers have been invariably reinforced with mineral fillers such as silica and/or carbon black to improve their physical and mechanical properties. The extent of property improvement depends on several factors, notably the size of particles, their aspect ratio, and the strength of interactions between the filler and the matrix polymer. The filler-matrix interactions are especially crucial to improve the filler dispersion as well as the adhesion of the matrix polymer to the filler surface, which in turn increases the effective filler volume and also promotes the stress transfer from the matrix to the filler when the material is subjected to mechanical deformation [1]. Correspondence to: Leon Perez; e-mail:
[email protected] Contract grant sponsor: National Science Foundation (Polymers Program, Division of Materials Research); contract grant number: DMR0314760; contract grant sponsor: The University of Antioquia. DOI 10.1002/pen.21247 Published online in Wiley InterScience (www.interscience.wiley.com). C 2009 Society of Plastics Engineers V
Recent advances in nanotechnology have shown the effectiveness of novel nano-sized particles in reinforcing the polymeric materials, and thus nanoparticle-reinforced composites have gained considerable interest, especially when this goal can be achieved at unusually low filler loadings. The typical reinforcing fillers include layered silicate (clays), expanded graphite, polyhedral oligomeric silsesquioxane (POSS), and single-wall and multiwall carbon nanotubes (MWCNT) [2–6]. Recently, the carbon nanotubes (CNTs) have become increasingly attractive because of the possible large scale production of these materials. Moreover, MWCNT have several advantages in industrial applications over other fillers because their production is less complex and more cost effective due to the requirement of the low loading to achieve comparable composite properties [7]. In addition, syntheses routes based on the chemical vapor deposition (CVD) processes are currently well established that can produce high yields of acceptable CNT products [8]. One of the challenges in dealing with the neat CNT is its inherently weak interactions with the matrix polymers, because the pristine CNTs are made-up of cylindrical graphite layers. These graphites do not have the functional groups required to promote filler-polymer interactions to enhance the reinforcement effects of the nanoparticles in the CNT/rubber composites. However, CNTs can be functionalized during the purification process by exposing them to strong oxidizing acids to introduce hydroxyls and carboxyl groups on the CNT surface. These functional groups of the CNT promote the strong interactions with the elastomer matrix such as electron donor– acceptor interactions, particularly those containing oxygenated and nitrogenated functional groups interacting with unsaturated carbon bonds of the elastomers. Although it is possible to incorporate MWCNT into an elastomer via solution blending [9], this approach is not suitable for industrial practice due to its higher environPOLYMER ENGINEERING AND SCIENCE—-2009
mental and economic costs. Melt blending is the most convenient and efficient technique for the preparation of CNT reinforced elastomers, but there are only limited studies reporting on the physical and mechanical properties of these composite materials produced by melt processing. In the present article, we begin with the synthesis of CNTs by CVD using iron and cobalt catalysts supported by calcium carbonate substrates, followed by the preparation, characterization, and properties of the nanoparticlesfilled composites using styrene-butadiene rubber (SBR) and nitrile-butadiene rubber (NBR) as elastomeric matrices. The emphasis is placed on the determination of vulcanization behavior, morphology, thermal, mechanical, and electrical properties.
EXPERIMENTAL Carbon Nanotube Synthesis The MWCNT were synthesized by CVD according to the procedure described by Giraldo et al. [10] using the cobalt (12 wt%) and iron (6 wt%) catalyst supported by calcium carbonate. The catalyst was prepared by impregnation of carbonate from an aqueous solution of the metallic salts, viz., Co(NO3)26H2O and Fe(NO3)39H2O. The chemical deposition reaction was carried out at 7508C for 30 min. Ethylene, the source of carbon, was mixed with nitrogen at a 3:1 ratio of ethylene/nitrogen under a constant nitrogen flow rate of 400 mL/min into the reactor. The material thus obtained from the CVD reaction was treated with the 30 wt% nitric acid solution for 72 hr to remove the support and the metal, and also to reduce the amorphous carbon by-products. This procedure presumably caused some oxidation to the CNTs. The resulting product was then thoroughly washed using distilled water and finally dried at 1258C for 24 hr. Appropriate amounts of CNTs corresponding to 2, 5, 10, and 15 phr were compounded into the elastomers by melt–mixing using a two-roll open mill. The elastomer compounds were vulcanized with the aid of sulphur according to the conventional vulcanization method; their compositional recipes are shown in Table 1. The vulcanization times for each composition were determined by differential scanning calorimetry (DSC) using a Q100 TA instruments calorimeter under isothermal conditions at 1508C during 1 hr to ensure that the vulcanization was completed. The degree of vulcanization a was calculated based on the heat of vulcanization reaction in what follows, a¼
DHt DH1
(1)
where DHt and DH1 are the reaction enthalpy at time, t and the total enthalpy change for the completed reaction, DOI 10.1002/pen
TABLE 1. Sample compositions in phr.
SBR NBR CNT Zinc oxide Stearic acid Agerite Resin Da Sulphur CBSb ZDECc a b c
NX
SX
0 100 0–15 5 1 2 1.5 2 0.8
100 0 0–15 5 1 2 1.5 2 0.8
2,2,4-trimethyl-1,2-hydroquinoline. N-cyclohexylbenzothiazole-2-sulphenamide. Zinc diethyl dithiocarbamate.
respectively. The samples were vulcanized at 1508C and at 3000 psi using a compression molding to obtain thin sheets (0.8 mm thick). The morphological features of the carbonaceous materials and their dispersion in the elastomeric composites were characterized by scanning electron microscopy (SEM) (Model XL30 FEI, Phillips), and the composites were frozen in liquid nitrogen, fractured, and shadowed with gold in vacuum. The cross sections were analyzed. The carbon nanotube sample was dispersed in propanol, a small droplet was put in the holder, and the solvent was evaporated under vacuum. Structural characteristics of the CNTs were determined by TEM, the instrument utilized was Model CM20, Phillips, and the CNTs samples were prepared in the same way mentioned before for the SEM analysis. The values of equilibrium swelling ratios, Qr were determined using small rubber pieces (weighed 0.5 g), which were placed in a steel mesh cage and immersed in toluene for 72 hr. The swollen pieces were then blotted and reweighed, and the values of Qr were calculated in accordance with the following equation,
Qr ¼
8 9 :m=dr þ ðms mÞ=dt ; m=dr
(2)
where m and ms are the rubber weights before and after the swelling, respectively, and dr and dt are the densities of the rubber and toluene, respectively. The activation energies for the vulcanization were calculated based on the Ozawa equations [11], using the temperature of the vulcanization peak (Tmax) obtained at four heating rates, viz., 5, 10, 15, and 208C/min. In the first thermogravimetric analyses (TGA), thermal stabilities of the CNTs were determined via weight loss by ramping from room temperature to 8008C at 108C/min in air. For the purpose of comparison, another sample was annealed at 4808C for 1 hr and then analyzed by TGA scanned at 108C/min in air. Similarly, the TGA scans of the CNTs-filled rubbers were acquired by increasing temperature from ambient to 7008C at a heating rate of 108C/ POLYMER ENGINEERING AND SCIENCE—-2009 867
FIG. 1. Electron micrographs of the carbon nanotubes. A: SEM showing curly CNT and some particles of amorphous carbon, and B: TEM showing CNT with diameter from 20 to 60 nm.
min under N2 circulation. For all the analysis, a thermogravimetric analyzer (Q500, TA) instruments was used. Electrical conductivities were measured on the flat vulcanized rubber sheets using two parallel flat electrodes operated at various voltages ranging from 5 to 15 volts. The sample thickness was 0.7 mm. The dynamic mechanical experiments were conducted from 2100 to 508C at a frequency of 1 Hz in tensile mode using a DMA (Diamond, PerkinElmer). Heating rate was 38C/min. For the stress–strain measurements, the dumbbellshaped specimens were cut from the vulcanized rubber sheets. Tensile tests were carried out using a tensile tester (MTS TEST/5) at a strain rate of 50 mm/min; each reported value corresponds to an average of at least three measurements. RESULTS AND DISCUSSION
The unannealed CNT manifested dual weight losses in the DTG curve. The first DTG peak, located at 4428C corresponding to a loss of 22.3 wt%, may be attributed to the degradation of the less stable amorphous carbon. The second weight loss peak appearing at 5548C corresponds to a loss of 64.2 wt% which is attributable to the CNTs themselves. In contrast, the annealed nanotubes showed only a single degradation peak (at 572.58C) in the DTG curve, the increasing in the degradation temperature is due to the elimination of amorphous carbon and functional groups on the nanotubes. Infrared Spectroscopic Characterization. Figure 3 exhibits the infrared spectrum for the CNTs in which the characteristic band at 3432 cm21 corresponding to hydroxyl groups and the bands at 1650 and 1730 cm21 corresponding to carbonyl groups suggest functionalized CNT surface. These functional groups are presumably
Structural and Physical Characterization of Neat MWCNT Electron Microscopy. Figure 1A shows the SEM image for the as-synthesized carbon nanotube materials, showing the appearance of loose curly CNTs together with some amorphous-like particles. These nanotubes seemingly have a broad length distribution. The same CNT appearance can be identified in the corresponding TEM image of Figure 1B with some residual metal particles in the dark background, which may be a consequence of low metal/ support interactions. The nanotube diameters range from 20 to 60 nm. Thermogravimetric Analysis. The TGA and the derivative of thermogravimetric (DTG) thermograms of the unannealed and annealed CNTs are depicted in Fig. 2. 868 POLYMER ENGINEERING AND SCIENCE—-2009
FIG. 2. TGA and DTG for the decomposition of the nanotubes in air as prepared and annealed in a previous stage.
DOI 10.1002/pen
the maximum temperature of the DSC thermogram at each rate was collected and hereafter designated as Tmax. The Kissinger model was utilized for the calculation of the activation energy in what follows: Kissinger’s equation reads Ea ¼ R
FIG. 3. Infrared spectra of the nanotubes treated with nitric acid, indicating the presence of carbonyl and hydroxyl groups.
formed due to the chemical treatment with the strongly oxidizing nitric acid during the purification of CNTs.
Characterization of MWCNT/Elastomeric Composites SEM Analysis. The SEM photographs for the NBR and SBR composites containing 15 phr of CNTs are presented in Fig. 4. The CNTs in both elastomeric composites are seen to be randomly dispersed and well wetted by the elastomers. However, some large CNT bundles are visible in the zoomed part of the CNT/SBR composite (Fig. 3B), protruding out from the rubbery matrix. It appears that the CNT dispersion in the NBR composite is finer than that in the CNT/SBR. Determination of the Vulcanization Activation Energy. In this study, the activation energy was calculated from the DSC experiments under dynamic conditions, that is, vulcanization reactions were performed at four different heating rates (b), 5, 10, 15, and 208C, and
2 d lnðb=Tmax Þ dð1=Tmax Þ
(3)
Figure 5 shows the relative activation energies for the NBR and SBR matrices reinforced with various amounts of CNTs. The observed increasing trend in the activation energy of vulcanization may be attributed to the adsorption of the accelerator compound utilized. Also, the adsorption of polymer chains at the CNT filler surface may reduce the mobilities of the rubbery chains, making them less reactive and thus requiring more energy for the vulcanization to trigger, therefore a complete vulcanization is hard to come by. The large increases in activation energy in N15 may be explained by the very strong adsorption of vulcanizing polar reagents because NBR is a polar polymer. The nonpolar additives such as the accelerators are less soluble in polar matrix and thus easily absorbed by the oxidized nanotubes. Among the experimental methods for determining the activation energy, torque vulcanization rheometry and isothermal or/and nonisothermal DSC are the most widely used techniques. The values obtained from those techniques showed consistent trends, and the same behavior was observed for nanoparticles-filled rubbers [12–14]. Swelling Behavior. Table 2 shows the comparison of the swelling behavior of the unfilled elastomers and the CNT-reinforced elastomers. The reduction in the equilibrium swelling ratio Qr is a measure of the degree of adhesion between the polymer chains and the filler particles. The swelling is usually reduced in the case of adsorbing filler surfaces when compared with the case of nonadsorb-
FIG. 4. SEM photographs for samples (A) N15 and (B) S15, showing randomly dispersed CNT and wetted by the polymer.
DOI 10.1002/pen
POLYMER ENGINEERING AND SCIENCE—-2009 869
tested. This finding is consistent with the notion that the NBR contains polar nitrile groups which strongly interact with the hydroxyl functionalized CNT surfaces via the electron donor–acceptor interaction. One cannot rule out the fact that when the elastomer chains interact more strongly with the filler, a single macromolecular chain can cover sizable numbers of active sites on the filler surface and therefore only a smaller number of chains can be anchored at the surfaces [18]. It is likely unanchored NBR chains may be contributing to the swelling of the elastomeric composite network, but to a lesser extent with increasing the CNT loading level.
FIG. 5. Relative activation energies for the SBR and NBR reinforced with nanotubes, measured by dynamic DSC and obtained using Ozawa equation.
ing fillers [15]. Table 2 shows the values of Qr and Qr/ Qr0, where Qr0 is the swelling ratio for the unfilled elastomer. The swelling ratio for the unfilled SBR is larger than for the corresponding NBR sample because the toluene is a better solvent for the SBR as evidenced from the magnitude of the respective Flory-Huggins interaction parameters [16, 17]. For both the elastomers, the ratio Qr/Qr0 decline with increasing CNT loading, suggestive of the strong filler-elastomer interactions. However, a larger reduction in Qr is seen in the CNT/SBR composites relative to the polar NBR. In view of the nitrile electron acceptor groups in the NBR which can interact strongly with the electron donating hydroxyl groups on the CNT surface, a stronger interaction is anticipated between the nanotubes and the NBR matrix. On the other hand, when the elastomer chains interact more strongly with the filler, a single macromolecular chain can cover sizable numbers of active sites on the filler surface, and therefore, only a smaller number of chains may be anchored at the surfaces [18]. Presumably, unanchored NBR chains may be contributing to the swelling of the elastomeric composite network. It should be pointed out that the swelling ratio of the CNT/NBR is slightly reduced when compared with those of the CNT/SBR composites in all the compositions
Electrical Conductivity. Figure 6 shows the electrical conductivity for the SBR and NBR nanocomposites as a function of the nanotube concentration. The conductivity of the CNT/NBR is orders of magnitude higher than that of the CNT/SBR composites. With increasing the CNT loading, the conductivity value increase for approximately two orders of magnitude and levels off at around 10 phr CNT, which will be regarded as the NBR the percolation threshold. In the case of the SBR, there is no discernible percolation threshold, except for a minor increase in the conductivity with CNT loading. Thermal Stability. The CNT/SBR composite showed only a single drop in the weight loss curve with the increasing temperature. The addition of nanotubes to the matrix polymer raises the onset of degradation as well as its maximum rate. In the case of the CNT/NBR composites, the unfilled elastomer as well as the composite containing 2 and 5 phr of nanotubes revealed dual drops in the weight loss curve, and the corresponding DTG thermograms showed two degradation peaks. However, when the CNT content is increased to 10 and 15 phr, only one degradation peak can be discerned in these CNT/NBR composites, as shown by the electrical conductivity test; when the content of CNT is 10 phr, the percolation threshold was reached. Probably, the established filler net-
TABLE 2. Equilibrium swelling ratios Qr for the NBR and SBR elastomers reinforced with multiwall carbon nanotubes. SBR
NBR
Composition
Qr
Qr/Qr0
0 2 5 10 15
3.21 2.8 2.59 2.4 2.11
1 0.87 0.81 0.75 0.66
a
a
Qr
Qr/Qr0
2.2 2.05 2.12 2.04 1.61
1 0.93 0.96 0.93 0.73
Qr0 is the swelling ratio for the unfilled elastomers.
870 POLYMER ENGINEERING AND SCIENCE—-2009
FIG. 6. Electrical conductivities for the NBR and SBR elastomers reinforced with the nanotubes.
DOI 10.1002/pen
TABLE 3. TGA results for the SBR and NBR reinforced with multiwall carbon nanotubes under N2 atmosphere. SBR
NBR
First loss
First loss
Second loss
CNT (phr)
Maximum (8C)
Residual (wt%)
Maximum (8C)
Residual (wt%)
Maximum (8C)
Residual (wt%)
0 2 5 10 15
447.4 450.0 451.3 454.8 455.1
5.0 6.4 8.8 12.9 16.3
423.9 424.1 425.8
64.9 64.1 61.7
445.7 446.5 425.8 443.5 452.7
9.2 12.0 11.8 17.7 20.6
work can act as a barrier to the diffusion of gaseous subproducts generated during the thermal degradation, which in turn can shift the first degradation of matrix to higher temperatures, such that both the degradations are overlapped. Table 3 summarizes the TGA results for the CNT/SBR and CNT/NBR elastomeric composites at different levels of the CNT loading. In both the SBR and NBR elastomers, the percentage of residual weight increased with increasing amount of the CNT. The CNT filler’s ability to suppress thermal degradation of the elastomers may be attributed to barrier effects, as is the case for clay fillers, or a reduction of the pyrolysis rate, due to the decrease of the polymer global mobility. In fact, it has been shown that polymer chains confined in the mesoporous structures show greater thermal stabilities [5]. Dynamic Mechanical Analysis and Differential Scanning Calorimetry. Table 4 shows the Tg values as determined by DSC in comparison with the E00 and tan d peaks of the DMA measurement. The Tg value of DSC is closer to that of E00 relative to tan d. The DSC results of the CNT/NBR composites show only a slight increase in the Tg, but there is no discernible movement in the Tg of CNT/SBR composites. In contrast, the DMA results reveal definite increase in Tg with increasing filler content in the both composite, but this change was more pronounced for the CNT/NBR composite. This discrepancy in the Tg behavior between the two techniques may be attributed to the different sensitivity of each technique utilized. In principle, the glass transition
temperature Tg of a polymer is a manifestation of the chain mobilities reflecting the glassy-rubbery transition in the bulk. The addition of the filler particles affects the polymer two ways. First, the filler particles tend to restrict the polymer chain mobility through the filler-matrix interaction. In view of the fact that the polar NBR interacts more strongly with the functionalized CNT surface relative to that of the nonpolar SBR/CNT, the Tg of the CNT/NBR is expected to increase more drastically with increasing filler content. Second, as discussed earlier in the TGA studies, the thermal stability of the matrix polymer can be enhanced by the addition of the mineral filler. Because the DSC technique is an indirect measure of the Tg through the enthalpy change, the fillers can suppress the heat flow and thus the DSC thermal signal may be affected showing poor sensitivity to the Tg change in the CNT/elastomeric composites with the filler loading level. An alternative account of the different Tg behavior in the two composites is that the chain mobility depends on the polymer cohesive forces, as advocated by Lipatov [18]. In nonpolar elastomers, the cohesive forces are low and therefore only minor changes in Tg are expected for the CNT/SBR composite. In addition, Table 4 shows the relative areas of the tan d peaks, which are normalized by the elastomer content in the composite. According to Arrighi et al. [19], the normalized intensity of the Tg relaxation peak is found to be inversely proportional to the bound rubber fraction. The suppressed Tg peak is indicative the restrictive chain mobility of the bound rubber fraction. It has been estab-
TABLE 4. DMA and DSC results. NBR
SBR
DMA 00
DMA Tan d
E
CNT (phr)
DSC Tg (8C)
Max. (MPa)
Tg (8C)
Max
Rel. area
0 2 5 10 15
220.8 220.5 220.3 219.7 219.1
16.1 17.1 24.7 28.1 31.9
223.9 220.0 220.8 218.2 218.5
1.58 1.43 1.39 1.11 0.98
1 0.90 0.89 0.83 0.78
DOI 10.1002/pen
00
Tan d
E
Tg (8C)
DSC Tg (8C)
Max. (MPa)
Tg (8C)
Max
Rel. area
Tg (8C)
214.2 212.2 213.7 212.7 211.8
247.4 247.4 246.9 246.1 247.0
54.7 51.7 47.7 60.2 66.2
250.6 249.5 249.9 248.4 247.6
1.64 1.63 1.53 1.24 1.12
1 0.92 0.89 0.80 0.76
241.6 240.5 241.6 240.0 239.2
POLYMER ENGINEERING AND SCIENCE—-2009 871
FIG. 7. DMA curves for the composites SBR/carbon nanotubes, (A) tan d and (B) storage modulus.
lished that the conformational change is more restricted for the chains in close contact with the filler surface. As a result, a higher activation energy is required for those chains to undergo over all motions and consequently, these chains may not contribute significantly to the tan d peak intensity. The areas under the Tg peaks are similar for both elastomer matrices, but this does not necessarily mean that the interactions of the two elastomers with the nanotubes are the same. Figures 7A and 8A compare the trends of tan d for the CNT/SBR and CNT/NBR composites, as a function of amount of CNT. A significant reduction of the tan d peak intensity is clearly seen in the case of the nanocomposites. Figures 7B and 8B compare the values of the storage modulus E0 for the SBR and NBR when unfilled and when filled with the nanotubes. The filled samples show higher storage and loss modulus at each temperature. This could be due to the hydrodynamic effect and the adsorption of polymer chains on the filler surfaces that concurrently increases the filler effective
volume [20] (with reductions of the bulk mobility, as already mentioned). To interpret the variation of the elastic moduli of polymeric composites with the CNT amount, the values of elastic modulus are fitted to the Guth model [1] because this model is the most relevant for this kind of composite, specially for anisotropic filler. More detailed description of this model can be found elsewhere [21, 22]. The Guth equation reads, E0 ¼ ð1 þ 2:5f f þ 14:1f 2 f2 Þ E00
(4)
where E0 and E0 0 are the elastic modulus of the unfilled and filled polymer, and f and f are the filler volume fraction and the aspect ratio, respectively. In the present study, the Guth model (Eq. 4) was compared with the relative increase in the modulus as determined by DMA at 208C (see Fig. 9). The effect of the filler aspect ratio and the volume fraction on the elastic modulus as well as the
FIG. 8. DMA curves for the composites NBR/carbon nanotubes, (A) tan d and (B) storage modulus.
872 POLYMER ENGINEERING AND SCIENCE—-2009
DOI 10.1002/pen
Values of the aspect ratio obtained by fitting the experimental results to the Guth model are lower than expected. This could be due to agglomeration that reduces the effective anisotropy of the filler. A higher aspect ratio was needed in the case of the CNT/NBR nanocomposites, which probably is a result from the better filler dispersion mentioned earlier. Moreover, the nitrile groups of the NBR are good electron acceptors and therefore can interact more strongly with the electron-donating hydroxyl groups of the CNTs.
FIG. 9. Elastic modulus measured from DMA results for the NBR and SBR elastomers reinforced with the nanotubes, and fitted to the Guth model.
interaction between the dispersed particles are considered by the inclusion of the quadratic term into the origin expression of the Einstein’s equation [23].
Tensile Tests. The effects of the nanotubes on the mechanical properties of the SBR and NBR are illustrated in Fig. 10. The tensile strengths of the elastomers increase gradually with increase in the amount of CNTs. When compared with the unfilled elastomers, the stress at 100% strain, the stress at rupture, and the tensile modulus of the composite filled with 15 phr CNTs were increased by 262, 293, and 230%, respectively, for the CNT/SBR composites. The corresponding values for the CNT/NBR ones were 170, 192, and 229%, respectively. Such improvements should attract greater interest in the application of nanocomposite technology of the present kind.
FIG. 10. Tensile test results for SBR and NBR as a function of the nanotube concentration. A: Stress at 100% strain; B: Stress at rupture; and C: Tensile modulus.
DOI 10.1002/pen
POLYMER ENGINEERING AND SCIENCE—-2009 873
CONCLUSIONS Nitric acid treatment of the CNTs introduced the functional groups on the CNT surfaces such as carboxyl and hydroxyl groups. The nitrile groups of the NBR elastomeric chains interact strongly with these functional groups of the CNT filler surface through the electron donor– acceptor interaction. Such filler-matrix interaction showed the improved properties such as resistance to solvent swelling, enhanced glass transition temperature, and improved storage and loss moduli of the CNT/NBR relative to those of the CNT/SBR composites. Although the nanocomposites showed some improvements in the mechanical properties, the electrical conductivities showed minor improvement in the CNT/NBR composite (i.e., two orders of magnitude), but virtually no improvement in the CNT/SBR composite, presumably because of structural defects of the CNT fillers introduced, viz., oxidations during the nitric acid treatment. ACKNOWLEDGMENTS It is a pleasure to acknowledge Colciencias (Colombia) to the program ‘‘Apoyo a la comunidad cientı´fica Nacional a trave´s de las becas para estudios doctorales.’’ REFERENCES 1. L. Bokobza, Polymer, 48, 4907 (2007). 2. B. Lopez, L. Perez, M. Mesa, L. Sierra, E. Devaux, M. Camargo, C. Campagneand, and G. Giraud, ‘‘Effect of a Mesoporous Silica and Silane Coupling Agents on the Reinforcing of Styrene Butadiene Rubber,’’ in Annual Technical Conference—ANTEC, Conference Proceedings (2005). 3. J.E. Mark, Acc. Chem. Res., 39, 881 (2006). 4. X.W. Zhou, Y.F. Zhu, and J. Liang, Mater. Res. Bull., 42, 456 (2007).
874 POLYMER ENGINEERING AND SCIENCE—-2009
5. L.D. Perez, L.F. Giraldo, W. Brostow, and B. Lopez, e-Polymers, 29 (2007). 6. L.D. Perez, L.F. Giraldo, B.L. Lopez, and M. Hess, Macromol. Symp., 245–246, 628 (2006). 7. M. Moniruzzaman and K.I. Winey, Macromolecules, 39, 5194 (2006). 8. C.E. Baddour and C. Briens, Int. J. Chem. React. Eng., 3, 1 (2005). 9. D. Yue, Y. Liu, Z. Shen, and L. Zhang, J. Mater. Sci., 41, 2541 (2006). 10. L. Giraldo, B. Lo´pez, and G. Barrera, ‘‘Thermal Stability of Carbon Nanotubes Produced by Catalytic Chemical Vapor Deposition of Ethylene on Different Supports,’’ in 2007 NSTI Nanotechnology Conference and Trade Show—NSTI Nanotech 2007, Technical Proceedings (2007). 11. T. Ozawa, Polymer, 12, 150 (1971). 12. M.A. Kader and C. Nah, Polymer, 45, 2237 (2004). 13. G. Sui, W.H. Zhong, X.P. Yang, and Y.H. Yu, Mater. Sci. Eng. A, 485, 524 (2008). 14. W.H. Zhong, J. Li, C.M. Lukehart, and L.R. Xu, Polym. Compos., 26, 128 (2005). 15. L. Bokobza and M. Kolodziej, Polym. Int., 55, 1090 (2006). 16. K.N.A.-E.-N.S.L. Abd-El-Messieh, J. Appl. Polym. Sci., 88, 1613 (2003). 17. J. Yang, M. Tian, Q.-X. Jia, J.-H. Shi, L.-Q. Zhang, S.-H. Lim, Z.-Z. Yu, and Y.-W. Mai, Acta Mater., 55, 6372 (2007). 18. Y.S. Lipatov, Polymer Reinforcement, ChemTec Publishing, Canada (1995). 19. V. Arrighi, I.J. McEwen, H. Qian, and M.B. Serrano Prieto, Polymer, 44, 6259 (2003). 20. A.V. Shenoy, Rheology of Filled Polymer Systems, Kluwer Academic Publishers, London (1999). 21. M.D. Frogley, D. Ravich, and H.D. Wagner, Compos. Sci. Technol., 63, 1647 (2003). 22. Y.P. Wu, Q.-X. Jia, D.-S. Yu, and L.-Q. Zhang, Polym. Test., 23, 903 (2004). 23. H.S. Hsich, J. Mater. Sci., 17, 438 (1982).
DOI 10.1002/pen
