NANO LETTERS
Tough Nanocomposites: The Role of Carbon Nanotube Type
2009 Vol. 9, No. 4 1423-1426
Xiaomeng Sui and H. Daniel Wagner* Department of Materials and Interfaces, Weizmann Institute of Science, RehoVot, Israel 76100 Received October 27, 2008; Revised Manuscript Received January 27, 2009
ABSTRACT Unusually large deformation is observed in poly(methyl metacrylate) (PMMA) electrospun fibers under tension when multiwall or single-wall carbon nanotubes (MWCNTs and SWCNTs) are included as a second phase in the fibers. These distortions are virtually absent in pure PMMA fibers and stem from markedly different energy dissipation mechanisms and necking modes arising from the dissimilar nanotube morphologies. Thus, both nanotubes types are effective tougheners of PMMA fibers, with an advantage for MWCNTs over SWCNTs.
Composite materials based on nanometer size reinforcing particlessmainly in fibrillar or platelet shapesare considered to be a promising new brand of engineering composites. The structural properties of their thriving predecessors, the micrometer size fiber-based composites, indeed remain well below theoretically achievable values. The mechanical strength, stiffness, and toughness of composites based on carbon nanotubes (CNTs) are generally predicted to be much highersby at least an order of magnitudesthan those of micrometer fiber based composites. The walls of CNTs are made of graphene cylinders of very high aspect ratio (length over diameter), in the range of 102-106. Single-wall carbon nanotubes (SWCNTs) simply consist of one of these long and thin cylinders, whereas multiwall carbon nanotubes (MWCNTs) have up to 50 of them, arranged concentrically. It is this basic elongated cylindrical microstructure based on the carbon-carbon bond, and on various possible values of a lengthwise structural twist (chirality), that gives nanotubes their unusually attractive mechanical and electrical properties. Yet, significant property improvements are occasionally observed in composites because it is currently difficult to avoid random aggregation and the formation of clusters upon mixing of large area-to-volume ratio nanoparticles within a polymer solution. Such formations are unwanted, as they readily lead to very high mixture viscosities and void content, andswhen reaching a particle content of only a few weight percentsto consequent mechanical properties that are either not improved or even reduced.1 Despite the large number of recent studies, contradictory data abound in the literature and, experimentally, the role of carbon nanotubes as a mechanical reinforcement of polymer fibers has not been unequivocally demonstrated.2 More specifically, the effect of the nanotube structural characteristics (for example, single wall vs multiple * Corresponding author,
[email protected]. 10.1021/nl803241y CCC: $40.75 Published on Web 03/02/2009
2009 American Chemical Society
walls) on the mechanical properties of the resulting composite has not been suitably addressed. Thus, a major present-day challenge in nanocomposite preparation is the development of efficient preparation procedures leading to well-ordered tubular alignment and compaction, both being necessary for maximizing the mechanical properties. Novel methodologies are currently envisioned to address these complicated issues.3,4 A particularly simple and appealing approach to prepare large-scale, aligned, hierarchical nanocomposites in fibrous form is the electrospinning process, a technique used in applications such as filters, membranes, scaffolding for biotissue build-up, clothing, and more.5 In recent years electrospinning was used to prepare SWCNT and MWCNT reinforced polyacrylonitrile fibers,6,7 MWCNT reinforced poly(ethylene oxide),8 and MWCNT reinforced poly(methyl metacrylate) (PMMA).9,10 This submicrometer fiber formation process has the marked benefit that carbon nanotubes are necessarily confined to an aligned configuration parallel to the fiber axis. A wide range of polymers can be used,5 and achievable diameters range from several nanometers to a few micrometers. Submicrometer fibers made of pure PMMA, and of PMMA reinforced with both pristine SWCNTs and pristine MWCNTs (CNT/PMMA weight ratio of 1.5% in the initial composite dispersion), were electrospun (ES) in our laboratory under the conditions described in the Supporting Information, using the procedure described in our previous studies.9,10 The morphologies of as-prepared pure and nanotube-reinforced PMMA ES fibers can be seen in Figure S1 in the Supporting Information. The surface of the fibers is generally smooth and the cross section is uniform along the fiber length. Most nanotubes embedded in the fiber are indeed aligned along the fiber axis; SWCNTs exist as long and thick ropes within the fiber, whereas MWCNTs are well separated
Table 1. Summary of Mechanical Properties for PMMA and Its Nanocomposite Fibers and Filmsa sample type
failure strain (%)
tensile strength (MPa)
Young’s modulus (GPa)
toughness (MJ/m3)
PMMA PMMA/MWCNTs PMMA/SWCNTs
23.0 (34) 80.6 (38) 85.0 (70)
ES Fibers 75 (13) 193.8 (23) 65.7 (91)
0.82 (11) 2.59 (22) 0.78 (81)
12.6 (39) 133.1 (41) 41.1 (73)
PMMA PMMA/MWCNTs PMMA/SWCNTs
4.7 (16) 5.6 (56) 7.2 (44)
Films 52.8 (12) 47.7 (18) 55.5 (15)
2.25 (8) 2.24 (13) 2.26 (8)
1.5 (28) 1.9 (76) 2.7 (57)
ESEM ES Fiber Testsb PMMA 51.3 (60) 30.9 (40) PMMA/f-MWCNTs 71.8 (78) 79.5 (42) a The data are mean followed (in parentheses) by the coefficient of variation in percent.
and dispersed lengthwise, in the fiber bulk. Agglomerates and misaligned MWCNTs are occasionally observed, as seen elsewhere.8,9 The fibers were tested in tension using a homemade nanotensile tester, Figure S2 in Supporting Information, mounted on an inverted optical microscope. We tested 17, 20, and 19 electrospun fibers of PMMA, PMMA/ MWCNTs, and PMMA/SWCNTs, respectively. The strength data were fitted to a standard two-parameter Weibull distribution,9,11 as described in the Supporting Information. The resulting Weibull scale parameters were 118, 148, and 92 MPa, and the Weibull shape parameters were 2.0, 1.7, and 1.7 for PMMA, PMMA/MWCNTs, and PMMA/SWCNTs, respectively. The low values of the shape parameters reflect the large variability in strength in all cases. The slightly lower shape parameters of the composite fibers indicate a larger strength variability compared to that of pure PMMA fibers. Arinstein et al.12 have recently observed an unusual dependence of ES fiber modulus on diameter when the latter is smaller than 500 nm. We observed a similar size effect for other mechanical properties as well, see Figure S3 in the Supporting Information. Therefore, to eliminate in the present work unwanted variability due to any diameter effect, we selected a subset of specimens with a diameter restricted to 500-750 nm only. The addition of CNTs causes a striking, visible transformation in the deformation mode of PMMA ES fibers. In pure PMMA fibers, sparse and unstable polymer necking occurs under increasing tension, leading to failure at relatively small strains. However, the presence of either SWCNTs or MWCNTs causes the failure strain to reach comparatively enormous values (see Table 1), due to the overwhelming occurrence of stable polymer necking all along the fibers. This is clearly observed in movies of the tests; see Supporting Information. This effect seems counterintuitive at first, as one would expect a strong and stiff oblate reinforcement to compel a polymer specimen to fail at smaller rather than larger strains. Transmission electron microscopy (TEM) provides clues to the root causes of the much larger fiber deformation observed when CNTs are present in the PMMA fibers, as compared to PMMA-only specimens. In PMMA/ MWCNT ES fibers, Figure 1 shows that following localized polymer necking leading to polymer failure, extensive MWCNT pull-out takes place. Such double mechanisms necking followed by pull-outsinvolves large inelastic strains and extensive energy dissipation. However, in PMMA/ 1424
0.07 (41) 0.27 (50) b Data from Liu et al.9,10
n/a n/a
Figure 1. TEM images of stretched PMMA/MWCNT ES fibers. Extensive nanotube pullout is observed subsequent to the necking stage. Images shown represent (a) aligned and (b) occasionally agglomerated MWCNTs. Scale bars are 200 nm.
Figure 2. TEM images of stretched PMMA/SWCNT ES fibers. Necking starts at a local perturbation but is arrested when reaching strong and tough SWCNT ropes (a). Such “crack arrest” is followed by extensive longitudinal propagation of the necking region as PMMA molecules slip along the polymer/CNT rope interface upon further fiber stretching (b and c). Scale bars are 100 nm in (a) and (b) and 50 nm in (c), respectively.
SWCNT ES fibers a different double mechanism arises, as shown in Figure 2. In this case, multiple necking of the fiber proceeds until it is prevented from further growth by the presence of SWCNTs ropes which act as a deflecting wall. The polymer molecules then likely slip and align by shearing along the ropes, leading to the bridging structures observed by TEM and againsbut for a different reason than that with MWCNTssto very large inelastic deformations. Interestingly enough, increases in deformation and toughnesssbut not strength or stiffnesssdue to the presence of CNTs are also observed in millimeter-size PMMA films (Table 1) albeit to a lesser extent because of the absence of polymer necking in films. Thus, significant toughness enhancements, in films and mostly in nanofibers, due to the presence of both types of CNTs, are the truly significant observation here. QualitaNano Lett., Vol. 9, No. 4, 2009
tive observations of deformation by necking and occasional fibrillation were also reported during the preparation (rather than during controlled tensile testing) of ES poly(ethylene oxide) fibers13 and by crazing followed by pull-out in nanotube-reinforced polyacrylonitrile fibers.14 No crazing or fibrillation was observed here. As seen from Table 1, the mechanical properties of the fibers measured by nanotensile testing are impressive especially in viewsand perhaps becausesof the fact that the nanotubes of both types possess no adhesion-enhancing functional groups at their surface. To verify that no functional groups (such as carboxylic acid) were present on the surface of the MWCNTs, X-ray photoelectron spectroscopy (XPS) data for pristine MWCNTs were compared with those for carboxylated MWCNTs (obtained from the same source, the surface treatment being based on the same pristine MWCNTs). A significant difference in the oxygen content could be observed between the pristine MWCNTs ([O] ∼ 0.95%, most probably originating from residual water absorbed on the surface) and the carboxylated MWCNTs ([O] ∼ 3.5%). In a separate study (not discussed in the present paper) we performed mechanical tests with PMMA/COOH-MWCNTs. The results were significantly different than those presented here with pristine MWCNTs. For example, the average strain was 43.3 ( 15.2% (compared to 80.6% here) and the average tensile strength was 78.3 ( 14.2 MPa (compared to 194 MPa here). This is an additional indication that COOH groups were not present on the tubes used here. The Young’s modulus of MWCNT-based fibers is three times as high as that of the PMMA fibers (it is even slightly above the value obtained for macroscopic films). The modulus of SWCNT-based fibers is about the same as that of PMMA fibers, most likely because in this case the reinforcement consists of thick and tight ropes of SWCNTs rather than of well-dispersed tubes. Indeed, it is known that the modulus of SWCNT ropes strongly decreases with increasing rope diameter (a 30 nm diameter SWCNT rope has a Young’s modulus of less than 70 GPa,15 whereas in comparison, well-dispersed individual SWCNTs and MWCNTs have moduli of ∼1000 and ∼500 GPa, respectively). As detailed in the Supporting Information, we used bath sonication for 1 h for both nanocomposite types as an attempt to break down the agglomerates of SWCNTs and MWCNTs and to debundle the SWCNT ropes. This was generally enough to deagglomerate the CNTs but not enough to debundle the SWCNTs ropes. Longer/stronger sonication is known to induce severe defects and to break the CNTs, leading to a significant degradation of the mechanical properties, and thus was avoided here. Compared to tensile tests previously performed in the environmental scanning electron microscope (ESEM) chamber with the same materials,9 the specimens tested here in standard laboratory conditions (in air, at room temperature) all show significantly higher strength and stiffness, and generally lower strain to failure; see Table 1. Moreover, the relative variability in these parameters, reflected in percent by the coefficient of variation ()standard deviation divided by the mean), which is always high when testing submiNano Lett., Vol. 9, No. 4, 2009
crometer fibers,9,16 is significantly lower in the present case compared to tests performed in the ESEM. These observations are obviously due to the absence of any of the destructive effects of the electron beam on the polymer structure which are observed when tests are performed in an ESEM. If absolute values of mechanical properties are desired, this effectively precludes testing of ES fibers in the ESEM. A final remarkable point is the difference in properties and in variability between submicrometer ES fibers and millimeter-size films made with the same PMMA-based materials. This issue is somewhat more complex. As seen in Table 1: (i) The strength of ES fibers (of all three types) is significantly highersby a factor of almost 4 when MWCNTs are presentsthan the strength of films. Apart from the distinct preparation methods, this difference is likely due to two causes, the widely different scales of the specimens (fiber diameters are about 500 nm, film thicknesses are about 100 µm), as thinner specimens usually have much larger strength, and the fact that CNTs as well as the polymer molecules are well-aligned within the electrospun fiber bulk, whereas they are randomly dispersed in the films. (ii) The failure strain of ES fibers is always much higher than that of films, due to the nucleation and growth of localized, stable polymer necking regions forming in fibers but not in films. (iii) The resulting toughness of ES fibers, calculated here simply as the area under the stress-strain curve, is always much higher than the toughness of films, reflecting the differences in strain. (iv) However, the Young’s modulus of ES fibers is almost always smallersor at best equal, when MWCNTs are presentsthan the film modulus. For all properties, the relative variability is generally slightly higher for fibers than for films, mainly in the case of SWCNT-based specimens but the cause of this is not clear. Points ii-iv deserve some further clarification. The origin of extensive necking in CNT-reinforced PMMA fibers is in part mechanicalsarising from perturbations caused by local stress concentrations at the edges of the aligned nanotubes in the fiberssand in part molecularsarising from molecular entanglements which serve as virtual crosslinks17sboth being a source of local perturbation. Extensive necking leads to enhanced toughness in ES fibers but not in films because no necking is present in these. At the same time a lower Young’s modulus is generally observed in ES fibers, compared to films. This can be explained by a simple molecular argument based on recent thermal tests performed in our laboratory (but not shown here in extenso, see Figure S4 in the Supporting Information). In addition to a typical glass transition temperature (Tg), differential scanning calorimetry (DSC) tests of PMMA ES fibers also reveal a significant endothermal peak at about 60 °C (the β transition18-20), whereas this peak is altogether absent in films of the same material. Andrews et al.20 interpret this as dipole-dipole association bonding between ester side groups. Assuming that, through the electrospinning process, the molecular backbone of PMMA is stretched and oriented along the fiber axis, the probability of slippage between parallel polymer chains increases, compared to the entangled 1425
morphology in films. In films, no β transition is observed in the DSC curve and no dipole-dipole association exists, and no interchain slippage is present: Young’s modulus is higher and the strain is lower because they are dominated by the breakage of C-C bonds. Note that the DSC curves of the PMMA/MWCNTs and PMMA/SWCNTs composite ES fibers are identical to that of the pure PMMA fiber because the low filler content (1.5 wt %) does not affect the thermal properties. Acknowledgment. We acknowledge support from the NOESIS European project on “Aerospace Nanotube Hybrid Composite Structures with Sensing and Actuating Capabilities”, and the G. M. J. Schmidt Minerva Centre of Supramolecular Architectures. The electron microscopy studies were conducted at the Irving and Cherna Moskowitz Center for Nano and Bio-Nano Imaging at the Weizmann Institute of Science. Professor M. Levy and Mrs. E. Wiesel are thanked for helpful suggestions. This research was made possible in part by the generosity of the Harold Perlman family. H. D. Wagner is the recipient of the Livio Norzi Professorial Chair. Supporting Information Available: Description of materials, electrospinning and post-treatment, morphological and thermal characterization, and nano- and microtensile tests, figures showing TEM images of fibers, the nanotensile testing setup, the size effect observed for modulus, strength, and toughness, DSC curves for PMMA fiber and film, stressstrain curves for fibers and films, and Weibull plots of nanofibers and films, and videos showing nanotensile testing. This material is available free of charge via the Internet at http://pubs.acs.org. References (1) Gorga, R. E.; Cohen, R. E. Toughness enhancements in poly(methyl methacrylate) by addition of oriented multiwall carbon nanotubes. J. Polym. Sci., Part B: Polym. Phys. 2004, 42, 2690. (2) Coleman, J. N.; Khan, U.; Blau, W. J.; Gun’ko, Y. K. Small but strong: A review of the mechanical properties of carbon nanotube-polymer composites. Carbon 2006, 44, 1624. (3) Mamedov, A. A.; Kotov, N. A.; Prato, M.; Guldi, D. M.; Wicksted, J. P.; Hirsch, A. Molecular design of strong single-wall carbon nanotube/polyelectrolyte multilayer composites. Nat. Mater. 2002, 1, 190. (4) Podsiadlo, P.; Kaushik, A. K.; Arruda, E. M.; Waas, A. M.; Shim, B. S.; Xu, J. D.; Nandivada, H.; Pumplin, B. G.; Lahann, J.; Ramamoorthy, A.; Kotov, N. A. Ultrastrong and stiff layered polymer nanocomposites. Science 2007, 318, 80.
1426
(5) Ramakrishna, S.; Fujihara, K.; Teo, W. E.; Lim, T. C.; Ma, Z. An Introduction to Electrospinning and Nanofibers; World Scientific: Singapore, 2005. (6) Vaisman, L.; Wachtel, E.; Wagner, H. D.; Marom, G. Polymernanoinclusion interactions in carbon nanotube based polyacrylonitrile extruded and electrospun fibers. Polymer 2007, 48, 6843. (7) Ko, F.; Gogotsi, Y.; Ali, A.; Naguib, N.; Ye, H. H.; Yang, G. L.; Li, C.; Willis, P. Electrospinning of continuons carbon nanotube-filled nanofiber yarns. AdV. Mater. 2003, 15, 1161. (8) Dror, Y.; Salalha, W.; Khalfin, R. L.; Cohen, Y.; Yarin, A. L.; Zussman, E. Carbon nanotubes embedded in oriented polymer nanofibers by electrospinning. Langmuir 2003, 19, 7012. (9) Liu, L. Q.; Tasis, D.; Prato, M.; Wagner, H. D. Tensile mechanics of electrospun multiwalled nanotube/poly(methyl methacrylate) nanofibers. AdV. Mater. 2007, 19, 1228. (10) Liu, L. Q.; Eder, M.; Burgert, I.; Tasis, D.; Prato, M.; Wagner, H. D. One-step electrospun nanofiber-based composite ropes. Appl. Phys. Lett. 2007, 90, 083108. (11) Barber, A. H.; Andrews, R.; Schadler, L. S.; Wagner, H. D. On the tensile strength distribution of multiwalled carbon nanotubes. Appl. Phys. Lett. 2005, 87, 203106. (12) Arinstein, A.; Burman, M.; Gendelman, O.; Zussman, E. Effect of supramolecular structure on polymer nanofibre elasticity. Nat. Nanotechnol. 2007, 2, 59. (13) Zussman, E.; Rittel, D.; Yarin, A. L. Failure modes of electrospun nanofibers. Appl. Phys. Lett. 2003, 82, 3958. (14) Ye, H. H.; Lam, H.; Titchenal, N.; Gogotsi, Y.; Ko, F. Reinforcement and rupture behavior of carbon nanotubes-polymer nanofibers. Appl. Phys. Lett. 2004, 85, 1775. (15) Salvetat, J. P.; Briggs, G. A. D.; Bonard, J. M.; Bacsa, R. R.; Kulik, A. J.; Stockli, T.; Burnham, N. A.; Forro, L. Elastic and shear moduli of single-walled carbon nanotube ropes. Phys. ReV. Lett. 1999, 82, 944. (16) Ayutsede, J.; Gandhi, M.; Sukigara, S.; Ye, H. H.; Hsu, C. M.; Gogotsi, Y.; Ko, F. Carbon nanotube reinforced Bombyx mori silk nanofibers by the electrospinning process. Biomacromolecules 2006, 7, 208. (17) Nielsen, L. E.; Landel, R. F. Mechanical Properties of Polymers and Composites, 2nd ed.; Marcel Dekker: New York, 1994. (18) Hamieh, T.; Schultz, J. New approach to characterise physicochemical properties of solid substrates by inverse gas chromatography at infinite dilution II. Study of the transition temperatures of poly(methyl methacrylate) at various tacticities and of poly(methyl methacrylate) adsorbed on alumina and silica. J. Chromatogr., A 2002, 969, 27. (19) Diazcalleja, R.; Ribesgreus, A.; Gomezribelles, J. L. Study of Structural Relaxation by Dynamic-Mechanical Methods in Poly(Methyl Methacrylate). Polymer 1989, 30, 1433. (20) Andrews, R. D.; Hammack, T. J. Theoretical Interpretation of Dynamic Mechanical Loss Spectra and Transition Temperatures. J. Polym. Sci., Part B: Polym. Lett. 1965, 3, 655. (21) Li, D.; Wang, Y. L.; Xia, Y. N. Electrospinning of polymeric and ceramic nanofibers as uniaxially aligned arrays. Nano Lett. 2003, 3, 1167. (22) Kaplan-Ashiri, I.; Cohen, S. R.; Gartsman, K.; Ivanovskaya, V.; Heine, T.; Seifert, G.; Wiesel, I.; Wagner, H. D.; Tenne, R. On the mechanical behavior of WS2 nanotubes under axial tension and compression. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 523.
NL803241Y
Nano Lett., Vol. 9, No. 4, 2009
